History of Analytic Philosophy Series Editor: Michael Beaney, University ofYork, UK

lt is usually assumed that Wittgenstein's philosophical development is determined either by one dramatic or one subtle change of mind. This book challenges the one­change view. Wittgenstein had many changes of mind and they are so substantial that he can be understood as holding severa[ different philosophies in the late twenties and early thirties. Early in 1929, Wittgenstein envisaged a complementary (phenomenological) symbolism in arder to carry out the Tractarian task of giving the limits of language and thought. lhe symbolism failed and he then developed a comprehensive notion of'grammar' that, he hoped, would fulfill the task. This notion of 'grammar' led in 1930-1 to the calculus conception of language, which is still defended in the Big Typescript (1932-3). As a complementary too[ of the calculus conception, Wittgenstein invented the genetic method, which aimed at dissolving philosophical puzzles by the understanding of how they carne about.After the Big Typescript, Wittgenstein assimilated an anthropological view and put the genetic method at the center of the stage of his philosophy. The use of the genetic method (associated with an anthropological view) developed gradually, taking various forms of application: in the Blue Book, in the versions of the Brown Book (1934-6), and in the Philosophical lnvestígations.

Mauro Luiz Engelmann is Professor of Philosophy at the Federal University of Minas Gerais (UFMG), Brazil.

pai grave macmillan

ISBN 978-0-230-28256-8

11 ~ 111 90101

9 7so2La2ss~ 11 1

www.palgrave.com

:e --1 -1 C') m ::z: ~ -1 m

-~ -e ::e r-0 ~

o -e ::e n l> r-c m < m r-0 -e 3:: m ::z: --1

3: CI e: ...... o ~ e: N m ::::1

co CD

3 CI ::::1 ::::1

~ HISTORY OF ANALYTIC PHILOSOPHY

r w 1 TT G EN s TE 1 N Is PH 1 LOSO PH 1 CAL DEVELOPMENT

PHENOMENOLOGY, GRAMMAR, METHOD, ANO THE ANTHROPOLOGICAL VIEW

Mauro Luiz Engelmann *

History of Analytic Philosophy

Series Editor: Michael Beaney, University of York, UK

Titles ínclude:

Stewart Candlish THE RUSSELL/BRADLEY DISPUTE ANO ITS SIGNIFICANCE FOR TWENTIETH-CENTURY PHILOSOPHY

Siobhan Chapman SUSAN STEBBING ANO THE LANGUAGE OF COMMONSENSE

Annalisa Coliva MOORE ANO W,ITTGENSTEIN Scepticism, Certainty and Common Sense

George Duke DUMMETT ON ABSTRACT OBJECTS

Mauro Luiz Engelmann WITTGENSTEIN'S PHILOSOPHICAL DEVELOPMENT Phenomenology, Grammar, Method, and the Anthropological View

Sébastien Gandon RUSSELL'S UNKNOWN LOGICISM A Study in the History and Philosophy of Mathematics

Anssi Korhonen LOGIC AS UNIVERSAL SCIENCE RusselI's Early Logicism and Its Philosophical Context

Gregory Landini FREGE'S NOTATIONS What They Are and What They Mean

Sandra Lapointe BOLZANO'S THEORETICAL PHILOSOPHY An Introduction

Omar W. Nasim BERTRAND RUSSELL ANO THE EDWARDIAN PHILOSOPHERS Constructing the World

Ulrich Pardey FREGE ON ABSOLUTE ANO RELATIVE TRUTH Douglas Patterson

Alfred Tarski PHILOSOPHY OF LANGUAGE ANO LOGIC

Erich Reck (editor) THE HISTORIC TURN IN ANALYTIC PHILOSOPHY

Graham Stevens THE THEORY OF DESCRIPTIONS

Mark Textor (editor) JUDGEMENT AND TRUTH IN EARLY ANALYTIC PHILOSOPHY AND PHENOMENOLOGY

Nuno Venturinha (editor) WITTGENSTEIN AFTER HIS NACHLASS

Pierre Wagner (editor) CARNAP'S LOGICAL SYNTAX OF LANGUAGE

Pierre Wagner (editor) CARNAP'S IDEAL OF EXPLICATION AND NATURALISM

Forthcoming:

Andrew Arana and Carlos Alvarez (editors) ANALYTIC PHILOSOPHY AND THE FOUNDATIONS OF MATHEMATICS

Rosalind Carey RUSSELL ON MEANING The Emergence of Scíentific Philosophy from the 1920s to the 1940s

G!useppina D'Oro and Constantine Sandis (editors) REASONS AND CAUSES Causalism and Non-Causalism in the Philosophy of Action

Sandra Lapointe (translator) Franz Prihonsky THE NEW ANTI-KANT

Consuelo Preti THE METAPHYSICAL BASIS OF ETHICS The Early Philosophical Development of G.E.Moore

Maria van der Schaar G.F. STOUT: ON THE PSYCHOLOGICAL ORIGIN OF ANALYTIC PHILOSOPHY

History of Analytic Philosophy Series Standing Order ISBN 978-0-230-55409-2 (hardcover) Series Standing Order ISBN 978-0-230-55410-8 (paperback) (outside North America only)

You can receive future titles in this series as they are published by placing a standing arder. Please contact your bookseller or, in case of difficulty, write to us at the address below wíth your name and address, the title of the series and one of the ISBNs quoted above.

Customer Services Department, Macmillan Distribution Ltd, Houndmills, Basingstoke, Hampshire RG21 6XS, England

Wittgenstein's Philosophical Development Phenomenology, Grammar, Method, and the Anthropological View

Mauro Luiz Engelmann Universidade Federal de Minas Gerais, Brazil

,11/--9 ~A ~91 ., f ~·1L- ~ ;1)~.

/l~~J~ ~ -v-Q-U- ~ ~ t~ .

pai grave macmiHan

//~1 ~~

* © Mauro Luiz Engelmann 2013

Ali rights reserved. No reproduction, copy or transmission of this publication may be ma de without written permission.

No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6-10 Kirby Street, London EC1N 8TS.

Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages.

The author has asserted his right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988.

First published 2013 by PALGRAVE MACMILLAN

Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS.

Palgrave M acmillan in the US is a division of St Martin's Press LLC, 175 Fifth Avenue, New York, NY 10010,

Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world.

Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries

ISBN: 978-0-230-28256-8 hardback

This book is printed on paper suitable for recyding and made from fully managed and sustained forest sources. logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin.

A catalogue record for this book is available from the British Library.

A catalog record for this book is available from the Library of Congress.

10 9 8 7 6 5 4 3 2 1 22 21 20 19 18 17 16 15 14 13

Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne

For my mother and to the memory of my father

Contents

Preface X

Series Editor's Foreword xiil

List of Abbrevíations xvi

Introduction 1

1 Phenomenology, 'Grammar,' and the 'Limits of Sense' 6 1.1 Elementary propositions and the nature of necessity 6 1.2 The ultimate analysis and the

phenomenological language 13 1.2.l A complementary notation without hypotheses 14 1.2.2 Objectivlty of the notation and the

determination of sense 19 1.2.3 A complementary notation grounded in the

structure of phenomena 23 1.2.4 Verification and sense 25

1.3 Phenomenological (primary) language: a draft and a method 28

1.4 Phenomenological language: the end of the project 35 1.5 The comprehensive 'grammar' 43

1.5.1 Phenomenology, verification, and the comprehensive 'grammar' 43

1.5.2 Classification of words 51 1.5.3 'Grammar' of phenomena and geometry 54 1.5.4 Arithmetic as 'grammatical' systems 56 1.5.5 Arbitrariness of 'grammar' and the

"essence of the world" 60

2 Russell's Causal Theory of Meaning, Rule-Following, the Calculus Conception, and the lnvention of the Genetic Method 65 2.1 . Causal theories of meaning 65

2.1.1 Russell's causal theory of meaning 65 2.1.2 Ogden and Richardson meaning 67

2.2 Wittgenstein's critique of the causal theory of meaning 69 2.2.1 Wittgenstein's arguments against the casual theory 73

vii

viii Contents

2.2.2 The Tractatus and the absent psychological intentional element 77

2.2.3 Intentionally, 'attltudes,' and what makes a thought true 82

2.2.4 The intangibility of intention and its temptations 88 2.3 Language as a calculus: autonomy and the role

of a system of rules 93 2.4 Genetic method: rules, analogles, and

the physiognomy of errors 99 2.5 A note on the first project with Waismann 111

3 The Big I)!pescrípt, the Tractatus, Sraffa, and the Anthropological View 113 3.1 What is the Big I)!pesaipt (BT)? 113 3.2 The Tractatus revisited and language as

an autonomous calculus 115 3.2.1 Autonomy of 'grammar' 120 3.2.2 A calculus without an a priori structure 124 3.2.3 The Tractatus read afresh (a minimalist reading) 131

3.3 'Grammar' and some tensions 139 3.4 Sraffa's fruitful criticism and the anthropological view 148

3.4.1 The calculus conceptíon revisited 148 3.4.2 Primitive languages and 'grammar' 151 3.4.3 An anthropological view: purpose, point,

and form of life 160 3.5 Concluding remarks: putting two and two together 167

4 The Road to the Philosophical Investigations (Blue Book, Brown Book, German Brown Book, and MS 142) 171 4.1 'Grammar' and genetic method revisited 171 4.2 The Blue Book: 'grammar,' 'calculus,' use and analogies 172

4.2.1 'Grammar' in the Blue Book 173 4.2.2 'Calculus' in the Blue Book 176 4.2.3 The genetic method and grammatical mistakes 179

4.2.3.1 Augustine and time: a false analogy 183 4.2.3.2 Words, sentences, things and something

between (and the Tractatus) 184 4.3 The Brown Book; the anthropological view and

the genetic method 190 4.3.1 From the Blue Book to the Brawn Book 190 4.3.2 An example: language-games, comparison,

and recognition 192

Contents ix

4.3.3 'Grammar' is synonymous with 'use' or 'descriptions of use' 201

4.3.4 The German Brown Book (BrBG) 205 4.4 The first version of the Philosophical lnvestigations 207 4.5 Wittgenstein's second project with Waismann 213

5 The Philosophical Investigations 221 5.1 Preliminaries: genetic method, anthropological view,

and surveyable representation 221 5.2 Preliminaries: the genetic method and the Tractatus 231

5.2.1 A step-by-step procedure 233 5.2.2 The assumptions of the Tractatus 234

5.3 The genetic method applied to the metaphysics of logic of the Tractatus 237 5.3.1 Simple names, primary elements, and analysis 237 5.3.2 Analysis, determination of sense, and the

"crystalline purity of logic" 244 5.3.3 Logic and the general form of propositions 251

5.4 An old remark about grammar in a new context 261

Notes 273

Bibliography 304

lndex 313

Pref ace

Wittgenstein's philosophical project is remarkable for many reasons, but two of them are especially striking. First, in the preface of the PI he recognizes 'grave mistakes' in his first book, the T, which is not just any book, but a classic, one of the most important and influential books of the 20th century. Second, the PI einbodies a reconsideration of his philosophy as a whole. Both books have received a lqt of attention in the literature. However, comparatively, little effort has been made to unravel the thoughts underlying Wittgenstein's recognition and reconsideration. Those thoughts must be found in his philosophical development. My effort in coming to terms with it is presented in this book. ·

Years ago, 1 became puzzled by Wittgenstein's philosophy in the T. The options of interpretation of the book, so went the mantra, were either to take it as an obvious failure in which interesting views about logic are introduced, or to take it as a kind of mystical plot conveying the most preposterous metaphysical theses, like 'realistic solipsism.' One might call this the 'traditional reading.' I was about to give up on my intention of studying the T when Diamond's work indicated a new direction in the research.1 ln her work, a correct reading of the PI seemed to be united with a plausible, not self-defeating, reading of the · T. After some years studying Wittgenstein's philosophy, however, I had the growing feeling that the 'resolute reading' of the T did not account for the changes that took place in the middle period (1929-37). It could well be that the 'traditional reading' was mistaken, as Diamond had argued; the 'resolute reading,' however, did not seem to fit Wittgenstein's development.

At the time of my PhD, interested in the controversies surrounding the T and prompted by the interest in understanding Wittgenstein's philo­sophical development, I decided to come to terms with his middle-period writings. I decided to study not-quite-finished-books, lectures, testi­monies, related philosophers, and to scrutinize the manuscripts of the Nach/ass according to the arder in which they were written. Now Iam inclined to think that the study of the intervening period might lead us to a new approach to Wittgenstein's work, one that might bring us beyond the now already old 'resoluteness-debate' concerning the T that has dominated Wittgenstein studies in the last decades.

X

Preface xi

ln spite of all their differences, 1 think, traditional and resolute readers share a noticeable trait: they do not pay sufficient attention to Wittgenstein's philosophical development after the T. As a consequence, they miss three important things. First, they miss a way of getting clearer about Wittgenstein's two masterpieces. His slow and intense struggle with his old and newly developed points of view sheds light on his early and later works, for the one is criticized and adapted, and the other, constructed. Second, they miss the philosophies that Wittgenstein developed after the T and before the PI (in SRLF, PR, BT, and the BB); philosophies that are original and important contributions in them­selves. Third, they miss the very adventure of Wittgenstein's develop­ment: the foternal dialectic that brought him to his later philosophy.

Since 1 began my work, 1 have benefited from the criticism and encour­agement of many people. This book grew out of my dissertation (2008, the University of Illinois, Chicago) whose adviser was Peter Hylton. To him, my gratitude is immense. His encouragement after the comple­tion of the dissertation was decisive: without it, this book would not have been written. His insightful advice and criticism, always fruitful, have helped me improve my work from 2004 to the present. His friend­ship is, however, the best fruit of the years of work. 1 thank also my dissertation committee (Bill Hart, Daniel Sutherland, Michael Kremer, and Wolfgang Kienzler) for comments and discussions that led me to improve my work at the time. Special thanks to Wolfgang Kienzler for our discussions about Wittgenstein's middle period and topics related to

it in the academic year 2006-7 that 1 spent in Jena. Encouragement and criticism also carne from people who read drafts

of this book in the past years and discussed them with me. Peter Hylton, again, Soeren Stenlund, Mike Beaney, and Craig Fox were kind to give me comments on the whole book (sometimes comments on more than one draft); Andrew Lugg gave me comments on parts of it. 1 thank them ali. 1 also profited from the presentation of parts of this book in the form of talks, seminars, and workshops in many occasions. 1 thank the audiences, the participants, and specially the organizers of those events. Thanks to Bento Prado Neto, João Cuter, Ludovic Soutif, Luís C. Pereira, Camila Jourdan, Edgar Marques, Mathieu Marion, David Stern, Arley Moreno, Marciano Spica, João C. Salles, André Porto, Sébastien Gandon, Denis Perrin, Jean-Philippe Narboux, Rogério Severo, and Chico Ferraz. If names have been forgotten, I beg for forgiveness. 1 was lucky enough to begin my studies of the Tractatus with Paulo Faria many years ago. Recently, 1 discussed a complete draft of this book in a workshop that he organized in Porto Alegre. I thank him for it, but 1 thank him especially

xii Preface

for the friendship and encouragement through the years. Thanks also to the participants of the workshop Marcelo Carvalho, Jônadas Techío, João Cuter, and Alexandre Machado, for comments and objections.

Juliana gave me ali the support and lave that one could wish. To her, mylove.

I thank the following journals for allowing me to use material previ­ously published:

The central argument of Chapter 2 appeared in 'Wittgenstein's New Method and Russell's The Analysis of Mind' in fournal of Philosophical Research, 2012, Vol. 37: 283-311.

The argument of Sectlon 3.4 of Chapter 3 appeared in 'Wittgenstein's "Most Fruitful ideas" and Sraffa' in Philosophical Investigations Early View, Apríl 2012.

Some considerations concerning the calculus conception of language of Chapter 3 appeared in 'Language Comme Calcul dans le Big Typescript' in Philosophiques, 2012, Vol. 39, n.1: 35-55 and in 'O que é o Big Typescript?' in Dois Pontos, 2009, Vol. 6, n. 1: 35-61.

Some remarks concerning Spengler in Chapter 1, Section 1.5, appeared in 'The Multiple Complete Systems conception as Fil Conducteur of Wittgenstein's Philosophy of Mathematics' (Proceedings of the 32nd International Wittgenstein Symposium).

Series Editor's Foreword

During the first half of the 20th century analytic philosophy gradu­ally established itself as the dominant tradition in the English-speaking world, and over the last few decades it has taken fírm root in many other parts of the world. There has been increasing debate over just what 'analytic philosophy' means, as the movement has ramified lnto the complex tradition that we know today, but the influence of the concerns, ldeas, and methods of early analytic philosophy on contem­porary thought is indisputab!e. Ali this has led to greater self-conscious­ness among analytic philosophers about the nature and origins of their tradition, and scholarly interest in its historical development and philo­sophical foundations has blossomed in recent years, with the result that history of analytic philosophy is now recognized as a major field of philosophy in its own right.

The main aim of the series in which the present book appears, the first series of its kind, is to create a venue for work on the history of analytic philosophy, consolidating the area as a major field of philosophy and promoting further research and debate. The 'history of analytic philos­ophy' is understood broadly, as covering the period from the last three decades of the 19th century to the start of the 21st century, beginning with the work of Frege, Russell, Moore, and Wittgenstein, who are genera11y regarded as its main founders, and the influences upon them, and going right up to the most recent developments. ln allowing the 'history' to extend to the present, the aim is to encourage engagement with contemporary debates in philosophy, for example, in showing how the concerns of early analytic phílosophy relate to current concerns. ln focusing on analytic philosophy, the aim is not to exclude compari­sons with other - earlier ar contemporary - traditions, or considera­tion of figures ar themes that some might regard as marginal to the analytic tradition but which also throw light on analytic philosophy. Indeed, a further aim of the series is to deepen our understanding of the broader context in which analytic philosophy developed, by looking, for example, at the roots of analytic philosophy in neo-Kantianism ar British idealism, ar the connections between analytic philosophy and phenomenology, ar discussing the work of philosophers who were important in the development of analytic philosophy but who are now often forgotten.

xfü

xiv Series Edítor's Foreword

Ludwig Wittgenstein (1889-1951) is one of the central figures in the analytic tradition, although there is controversy as to whether his later work counts as continuing or as rejecting that tradition. There is no doubt at ali that his early work, and the only book he published in his Ufetime, the Tractatus Logico-Philosophicus, which first appeared in German in 1921 and in English in 1922, is one of the classics - perhaps the classic text - of analytic philosophy. Deeply influenced by Frege and Russell, in particular, and yet also offering powerful criticisms of their work, it inspired in tum the two main subtraditions of analytic philos­ophy that developed in the 1930s, logiaal positivism and the Cambridge School of Analysis. Wittgenstein gave up philosophy after completing the Tractatus, believing that he had essentially solved áll problems of philosophy, as he claimed in the preface to the Tractatus. By the end of the 1920s, however, he had been drawn back into philosophy, partly by conversations he had with some of the logical positivists of the Vienna Circle. He began to recognize errors in his Tractatus, and his later work developed in an attempt to correct and diagnose those errors, as well as in responding to other philosophical developments around hlm.

Wittgenstein returned to Cambridge in 1929, and throughout the 1930s he worked on revising his ideas and thinking through a whole range of problems that arose as he did so. This work has generally been seen as culminating in the Philosophical Investígations, which was only published posthumously in 1953, in an English translation by Elizabeth Anscombe. But there is a whole series of preliminary manuscripts and typescripts that have been published subsequently, including Phílosophícal Remarks (dating from 1930, translated in 1975), The Btg JYpescript (1932-3, translated in 2005), Philosophícal Grammar (1933, translated in 1974), The Blue and Brown Books (1933-4 and 1934-5, respectively, published together in 1958), andRemarks on the Foundations ofMathematícs (1937-44, translated in 1956). Many commentators have assumed that most of Wittgenstein's new ideas were formed in the immedíate aftermath of the perceived breakdown of the Tractatus, and that what we have from that point on is deepening of the critique and refinement of formulation rather than any further substantial changes. This is reflected in standard talk of Wittgenstein's 'early' and 'later' philosophy. As Mauro Engelmann shows in the present book, however, the path to the Philosophical Investigatíons was long and complicated, and there is a great deal to say about the evolution of his ideas in the 1930s, in what should really be regarded as his 'middle' period.

As Engelmann tells the story, Wittgenstein went through an initial phase of seeking to supplement the logical notation of the Tractatus

Series Editor's Foreword xv

with a phenomenological language. From this developed Wittgenstein's idea of 'grammar,' an idea, however, that was itself to evolve in ways that many commentators have failed to appreciate. lt carne to be associated with a calculus conception of language, but the rejection of this latter conception in tum gave rise to a view of grammar as describing our use of language in actual practice. Alongside these changes were changes in methodology. What Engelmann calls Wíttgenstein's 'genetic method' was introduced to diagnose his earlier errors (and similar errors by other philosophers), a method that was later supplemented by an 'anthro­pological perspective,' which especially informs the opening sections of the Philosophical Investigations. ln telling this story, Engelmann draws on the full range of sources that are now available - not only the works mentioned above but also correspondence, memoirs, notes made from Wittgenstein's lectures, and most importantly of ali, his entire Nachlass, which was published in an electronic edition in 2000. (This major publishing event, and its implications for our understanding of Wittgenstein's philosophy, was the subject of an earlier volume in this series, edited by Nuno Venturinha.) ln paying careful attention to the exact order in which Wittgenstein's ideas and terminology appeared, Engelmann provides the most detailed account to date of the develop­ment of Wittgenstein's philosophy in the 1930s, not only explaining the various criticisms that were made over time of the ideas of the Tractatus, and the influences upon him, but also shedding a great deal of light on many of the key ideas and themes of the Philosophical Investigations.

Michael Beaney October 2012

List of Abbreviations

Wittgenstein's works:

BB BF

BrB BrBG

BT CL FF

LFM

LWGB

MS

NB NDMN NL PG PI PI II PR PUKGE RFM SRLF T

TS

The Blue Book (1933-4) Bearbeitete Fruehfassung (TS 239). Third version of the PI as presented in PUKGE (below) The Brown Book (1934-5) The Brown Book Written in German (second part of MS 115) is the unfinished German version of the BrB. Published as Eine

Philosophische Betrachtung (1936) The Big Typescript (1932-3) Cambridge Letters Fruehfassung (TS 225, TS 220, TS 221). Second Version (first typescript) version of the PI (in PUKGE) Lectures on the Foundations of Mathematics in Cambridge, 1939 Ludwig Wittgenstein Gesammatebriefwechsel/Complete Correspondence Manuscripts of the Nachlass. The number following 'MS' indicates the number of the manuscript according to Von Wright's catalog and the Bergen Electronic Edition. Notebooks 1914-16 Notes Dictated to Moore in Norway (1914) Appendix II of NB. Notes on Logic (1913) Philosophical Grammar (MSS: 114, 115, and 140) (1932-4) Philosophical Investigations The so-called Part II of Philosophical Investigations (TS 234) Philosophical Remarks (TS 209) (1930) Philosophische Untersuchungen. Kritisch-genetische Edition

Remarks on the Foundations ofMathematics Some Remarks on Logical Form (1929) Tractatus Logico-Philosophicus (1921). I usually use the McGuinness and Pears translation. If not, I indicate it with 'Ogden's translation' or 'my translation'. Typescripts from the Nachlass. The number following 'TS' indicates the number of the manuscript according to Von Wright's catalog and the Bergen Electronic Edition.

xvi

!!

UF WiC WLC30-32 WLC32-35 wvc

YB

z ZF

Listo( Abbreviations xvii

Urfassung (MS 142). First version of the PI (in PUKGE). Wittgenstein in Cambridge (Letters and Documents) Wittgenstein Lectures Cambridge (1930-2) Wittgenstein Lectures Cambridge (1932-5) Wittgenstein and the Vienna Circle. Waissman's notes of Wittgenstein's meetings with the Circle (1929-31) Yellow Book. Selected notes of díctations at the time of the Blue Book. Zettel Zwischenfassung. Fourth version of the PI reconstructed in PUKGE.

Other authors:

AM M MoM PLP

Russell's The Analysis ofMind. Moore's Notes of Wittgenstein's lectures in 1930-3. The Meaning of Meaníng (Ogden and Richards) Principles of Linguistic Philosophy. Waismann's book Logik, Sprache, Philosophie which was based on drafts partially written in collaboration with Wittgenstein.

Introduction

The two most common views concerning Wittgenstein's philosophical development are the following: he either had one dramatic or he had one subtle change of mind.1 The one-change-view is wrong in all its versions. Wittgenstein changed his views substantially more than once. The changes are so substantial that he can be understood as holding several different philosophies in the late 1920s and early 1930s. The T's goal, stated in its preface, of presenting the limits of language and thought remains largely in place until the BT and its revision around 1933, but the methods and tools to carry out the goal evolve through several stages. Early in 1929, Wittgenstein comes to think that the logical symbolism of the T alone is an insufficient tool to carry out the Tractarian task of giving the limits of language and thought. He then envisages a complementary (phenomenological) notation. Even though the draft of the language was promising, it failed. Wittgenstein then (from late 1929 onwards) developed a comprehensive notion of grammar and its rules that, he hoped, would fulfill the task of estab­lishing the limits of language and thought. This notion of 'grammar' leads to the calculus conception of language at the end of 1930. As a complementary tool of the calculus conception, Wittgenstein invents the genetic method, which aims at dissolving philosophical puzzles by the understanding of how they come about. The complemented idea of grammar is further developed until the BT (1932-3). After the BT, as a result of Sraffa's "unceasing criticism," Wittgenstein abandons the task of giving the limits of sense and assimilates an anthropological view into his philosophy from 1934 onwards. The anthropological view becomes an important tool at the service of the genetic method, and characterizes Wittgenstein's later philosophy. The use of the genetic method develops gradually, taking various forms of application, from

1

2 Wíttgensteín's Philosophical Development

1933 to 1937: first in the BB (1933-4), then in the versions of the BrB (1934-6), until it finds completion in the first version of the PI in 1937. lt is only in this later stage that the major goal of the philosophy of the T is abandoned and a completely new philosophy is created.

Thus, Wittgensteln's gradual development expresses a constant struggle, as he adapts and overcomes old conceptions by means of newly invented tools and methods. ln the following paragraphs l present a summary of the four main stages that emerge in this struggle: phenom­enology, grammar, the genetic method, and the anthropological view. This summary corresponds to the five chapters of this book.

Early in 1929, Wittgensteln's goal was to combine a phenomeno­logical language (a complementary notation for phenomena) with the logical notation of the T in order to provide a limit to what can be said meaningfully. The introduction of the phenomenological investigation was meant to fix some of the "grave mistakes" of the T pointed out by Ramsey (PI, preface).

Soon after, the idea of 'grammar' replaces the combined notations of phenomenology and logic. lt is a comprehensive discipline of the rules of language (mathematics, logic, and phenomenology are part of grammar) which gives the limits of language and thought. ln 1929, thus, two adaptations of the T take place by means of two projects: the phenomenological language and the comprehensive idea of 'grammar' that includes phenomenology. These issues are discussed in Chapter 1.

At the end of 1930, his struggles with the pictorial conception of language of the T and Russell's causal theory of meaning lead Wittgenstein in two directions: to the genetic method, and to the idea of the autonomy of grammar grounded in the calculus conception of language. Wittgenstein thought that Russell's theory was wrong, but also realized that his own old pictorial conception of language was unable to explain, for instance, how one means and intends a specific fact by a sentence (this problem arises because the Tractarian idea of analysls had been abandoned in 1929). He then created the calculus concep­tion in order to solve his difficulties. ln the sarne period, Wittgenstein invented the genetic method, as 1 call it, in order to complement the calculus conception of language. He carne to see that the calculus conception alone does not make philosophical worries disappear. He also carne to see that his views from 1929-30 were products of the sarne kind of confusion as his old views of the T and Russell's theory. Thís led Wittgenstein to think that philosophical puzzles (for instance, the so-called rule-following problem, which emerges in this period) could only be dealt with if one investigated how they were generated.

Jntroduction 3

His genetic method consists in discovering how and why a philosoph­ical question is asked and what false steps (usually prampted by false analogies) are assumed in the pracess of its formulation. The goal of the genetic method is to make the reader see, as though in a mirrar, what has led or may lead him to puzzlement. ln order to achieve this goal, no specific philosophical views should be presupposed, i.e., the method asks for philosophical neutrality. I discuss ali these issues, and how they are related, in Chapter 2. 1 finish this chapter with an expla­nation of the end of Wittgenstein's first project with Waismann, a revi­sion of the T.

The presentation of the above-mentioned 'mirrar' was intended to be ln the form oi theory-free synoptic presentation of trivialities (neutrality). This, however, did not quite happen. ln the BT, Wittgenstein changed his conception of 'grammar' from 1929-30 by assimilating to it the idea that language is structured like an autonomous calculus (game) of fixed rules for words and sentences. Wittgenstein took such a conception of 'grammar' as trivial: it seems to assume nothing in relation to reality or the workings of the mind. The rules of grammar, which are always part of a 'system of grammar,' or so Wittgenstein thought, constitute the meaning of words and the sense of sentences as rules constitute a game. What is neither a move in a game nor a rule of the game is nonsense. Those matters are discussed in Chapter 3. ln this chapter, 1 also intra­duce a comparison, suggested by Wittgenstein himself, between the BT and the T. This comparison expresses a critique of the T and, at the sarne time, elucidates the calculus conception of language, which is an adaptation of the old philosophy.

1 conclude Chapter 3 by arguing that Sraffa's "unceasing criticism" of underlying assumptions of the conception of language and 'grammar' as a calculus initiated the last fundamental move toward his later philos­ophy. lt marks the final break with the T; after that, no adaptations of the first philosophy are envisioned. Sraffa's criticism was based on the idea of primitive languages and a braad notion of the use of words. When Wittgenstein abandoned the calculus conception, he introduced an anthropological view in his philosophy. lt is an important tool, espe­cially because it dissolves a tension between the genetic method, which asks for philosophical neutrality, and the idea of 'grammar' as the rules of sense of a language, which was, contrary to Wittgenstein's inten­tions, a philosophical conception assumed as late as in the BT.

ln Chapter 4, 1 show how the genetic method gained praminence over the notion of 'grammar' and gradually carne to characterize Wittgenstein's later philosophy. Here we find, then, the significance of

4 Wittgenstein's Philosophical Development

Sraffa's criticism and Wittgenstein's own, reshaped, "anthropological view." With its help, Wittgenstein could employ the genetic method as a tool that worked without philosophical assumptions. lt is in the BB, I argue, that Wittgenstein's genetic method carne to the center of the stage of his philosophy for the first time. The systematic employ­ment of the genetic method and the anthropological view, however, were completely united only from 1934 onward. The unification of the genetic method and the anthropological view was achieved in the BrB. This transformed the use and goals of grammar, which in the BrB does not characterize a comprehensive discipline of the most general and arbitrary rules of sense any longer. 1 finish Chapter 4 with an explana­tion of the ending of the second project with Waisman:iÍ., an intended co-authored book that was to explain Wittgenstein's philosophy after 1931.

Some strategies adopted in the BrB, and even in its reformulation in its German version (BrBG), however, were still unsatísfactory for Wittgenstein's goals. Wittgenstein needed a real example on which to systematically employ his method and the anthropological view. After BrBG, this goal was achieved by focusing on his own earlier views of the T. It is this application, I argue, which characterizes the first version of the PI (MS 142) and all subsequent versions. The application of the genetic method, grounded on the anthropological view, to the philos­ophy of the T in the PI is presented in Chapter 5, the last chapter of this book. It is a step-by-step procedure that in an exemplary manner teaches us a new philosophical skill, and makes us see the point of Wittgenstein's later philosophy.

Four final notes need to be added here. First, this book points to a reconciliation of Wittgenstein's non-theoretical stance with his 'gram­matical remarks' in his later philosophy.2 This is achieved by the expla­nation of the changing significance of grammar and grammatical remarks in the process of construction of the PI. Second, the adaptations of the T presented in this book might explain the conflicting appear­ances of distance and proximity between Wittgenstein's earlier and later philosophies. Incidentally, they might also be helpful for determining the extent to which Wittgenstein is 'resolute' in the T and in the PI. This emerges from the ways that Wittgenstein adapted and criticized the T. The T is subjected to two critiques which are quite different in kind, as already indicated above. An early critique in the BT is grounded on the calculus conception of language (Chapter 3). It is a critique, nonetheless, that shares with the T the assumption that the rules implicit in language determine the limits of sense. A later critique, present in all versions of

Introduction 5

the PI, is grounded in the genetic method and in the anthropological view (Chapters 4 and 5). This critique does not share anything substan­tial with the T, except for Wittgenstein's suspicion (or strong convic­tion) that philosophical problems are the result of misunderstandings. The later critique is deeper and broader, for it also indicates mistakes in the very tools used in the early critique. Third, the chapters of this book work as a kind of uebersichtliche Darstellung of Wittgenstein's develop­ment and are meant to introduce major not-quite-finished-works from the middle period: Chapter 1 functions as an introduction to SRLF and PR; Chapters 1, 2, and 3 as an introduction to BT and PG; Chapters 1, 2, 3 and 4 as an introduction to BB and BrB. The whole book might be seen as an introduction to the PI. Fourth, during the middle period and also later, Wittgenstein thought constantly about the philosophy of mathematics. 1 do not go into the details of Wittgenstein's thoughts concerning these matters in this book; my account, however, indicates how they are related to his philosophical development.

1 Phenomenology, 'Grammar,' and the 'Limits of Sense'

1.1 Elementary propositions and the nature of necessity

Wittgenstein's re-evaluation of the philosophy of the T is clearly connected to criticisms made by Ramsey. He recognizes his debt to Ramsey in the preface of the PI:

... since 1 began to occupy myself with philosophy again, sixteen years ago, 1 could not but recognize grave mistakes in what 1 set out in that first book [the T]. 1 was helped to realize these mistakes - to a degree which 1 myself am hardly able to estimate - by the criti­cism which my ideas encountered from Frank Ramsey, with whom 1 discussed them in innumerable conversations during the Iast two years of his Iife. (PI, preface)

Even though Wittgenstein mentions only the Iast two years of Ramsey's Iife (he died in January 1930), he first met Ramsey in 1923.1 According to Ramsey's correspondence, he studied the T with Wittgenstein when he visited him in Austria in 1923.2 ln November 1923 Ramsey asked Wittgenstein if he wanted a copy of the Criticai Notice, his review of the T: "Has Ogden sent you my review of Tractatus in Mind? If not, and you would Iike it 1 will send it to you, but it is not at ali good and you must remember 1 wrote it before coming to see you" (CL, 192). Even though Ramsey himself says that his Criticai Notice wasn't good, some of his critiques, as we will see, point to the "grave mistakes" of the T and Wittgenstein's efforts to fix them in 1929.

It is not clear whether Wittgenstein actually read the review, but there is no doubt that he was well aware of its contents.3 Apart from Wittgenstein's recognition in the preface of the PI of the "grave

6

"

Phenomenology, 1Grammar, 1 and the 'Limits of Sense' 7

mistakes" of the T due to discussions with Ramsey, some of his major concerns in 1929 are directly related to Ramsey's review - Wittgenstein, in fact, wrote extensively on themes presented there.4

At the center of Wittgenstein's worries in 1929 is the problem of the analysis of propositions that attribute degrees of a property to objects (for instance, attributions of colors, heights, weights, etc). Wittgenstein calls those propositions "statements of degree" (SRLF, 167). Ramsey's criticism in his Criticai Notice relates statements of degree (in particular those related to color and space) to one of Wittgenstein's more impor­tant insights in the T, namely, his understanding of the nature of infer­ence and logical necessity:

It is a principie of Mr. Wittgenstein's ... that every genuine propo­sition asserts somethíng possible, but not necessary. This follows from bis account of a proposition as the expression of agreement and disagreement with truth possibilities of independent elementary propositions, so that the only necessity is that of tautology, the only impossíbility that of contradiction. (Ramsey, 473)

Ramsey then introduces the problem that needs to be solved if this view should hold:

There is a great difficulty in holding this; for Mr. Wittgenstein admits that a point in the visual field cannot be red and blue; and, indeed, otherwise, since he thinks induction has no logical basis, we should have no reason for thinking that we may not come upon a visual point which is both red and blue. Hence he says that 'This is both red and blue' is a contradiction. This implíes that the apparently simple concepts red, blue (supposing us to mean by those words absolutely specific shades) are really complex and formally incompatible. (Ramsey 1923, 473)

Ramsey's critique showed that central insights of the T were at risk (that ali necessity is logical necessity and that propositions of logic are tautologies), for they were grounded on the truth-functional explana­tion of logical necessity. Ramsey's specific target is the following remark of the T about color exclusion:

For two colors, e.g. to be at one place in the visual field, is impos­sible, logically impossible, for it is excluded by the logical structure of color.

8 Wittgenstein's Philosophical Development

Let us consider how this contradiction presents itself in physics. Somewhat as follows: That a particle cannot at the sarne time have two velocities, i.e. that at the sarne time it cannot be in two places, i.e. that particles in different places at the sarne time cannot be identlcal.

(It is clear that the logical product of two elementary propositions can neither be a tautology nor a contradiction. The assertion that a point in the visual field has two different colors at the sarne time is a contradiction.) (T 6.3751)

Wittgenstein clearly says that it is 11logically impossible" for two colors to be in the sarne place in the visual field. He also states that in physics the incompatibility is expressed as a contradiction. This means that an appeal to physics would not be appropriate. What physics supposedly shows, then, is that the incompatibility is logical, and not physical. Therefore, contrary to what Ramsey suggests, Wittgenstein does not intend to ground in physics his explanation of the contradiction. Here is Ramsey's argument:

He [Wittgenstein] tries to show how this may be by analyzing them in terms of vibrations. But even supposing that the physicist thus provides an analysis of what we mean by 'red' Mr. Wittgenstein is only reducing the difficulty to that of the necessary properties of space, time, and matter, or the ether. He explicitly makes it depend on the impossibility of a particle being in two places at the sarne time.

Ramsey is wrong in assuming that the incompatibility, for Wittgenstein, is physical, as seen above. Perhaps, this is the reason he called his review "not good at all" after discussing the T with Wittgenstein. However, he is right in saying that in the T a colar statement and a statement using the relation 'between' were not reduced to a formal tautology, i.e., to a truth-functional tautology. He continues the passage above:

These necessary properties of space are hardly capable of further reduction of this kind. For example, considering between ln point of time as regards my experiences; íf B is between A and D and C between B and D, then C must be between A and D; but it is hard to see how this can be a formal tautology. (Ramsey 1923, 473).

The relevant issues in Ramsey's paper are thus the 11structure of calor" and the "necessary properties of space." Perhaps, Wittgenstein could

Phenomenology, 'Grammar,' and the 'Limits of Sense' 9

deal with axioms of geometry inside the framework of the T. Axioms of geometry, Wittgenstein argues, can be treated as arithmetical equa­tions, i.e., as rules for inferences among ordinary propositions:

Because of this [the role of equations in inferences] the /apparent/ propositions of geometry cannot really be propositions, but [are] indicated transitions (Uebergaenge) from one proposition about spatial objects to another proposition about spatial objects. Thus, 1 can pass (uebergehen) directly from the proposition "A and B are between C and D" to "A is either between B and C or between B and D." The axiom that seems to allow me this transition is a tautology; or there is something different determining about its form, something that can just make it [the axiom] a criterion for both propositions that it connects (MS 105, 51; from the beginning of 1929).

Rules concerning 'between,' which supposedly work like tautologies, belong to "mathematical geometry." They could be taken, prima facie, as a method of logic, i.e., as principles of substitutional transitions between propositions, as the equations of arithmetic (see T 6.234, 6.24). One should have in mind here geometrical equations (MS 105, 17). ln this case, the transition from "A and B are between C and D" to "A is either between B and C or between B and D," would be similar to the transition from "2+2" to "4." One could then say that the "geometry of a specific space" (the visual space) would behave in relation to "math­ematical geometry" like "the sentence two and two plums are four plums and the sentence 2+2=4" (MS 105, 51). However, it is not clear how "mathematical geometry" relates to "geometry as the theory of one [particular] space" (MS 105, 51), i.e., to the geometry of the visual space (the space of the visual field). ln the visual field we find "something different": mathematical space is not colored, but visual space is. lt is in the visual space that the structure of calor and the structure of space are interconnected or, as Wittgenstein says, "calor and space fill (saet­tigen) each other" (MS 105, 53). ln the visual space, "there is no middle link" between calor and space (MS 105, 53). That is, we cannot talk of an object - say, a patch of calor - having the properties of "being here" and "having a color,11 as in the physical space, for the patch is identical with calor and position. Thus, it is the visual space that needs investiga­tion if one is concerned with Ramsey's criticism. This is precisely what Wittgenstein immediately does after his return to Cambridge (see MS 105, 1-11).5

Wittgenstein does not specify in the T what "the logical structure of colar" is; he does not show how the "concepts red, blue [ ... ] are really

10 Wittgenstein's Philosophícal Development

complex and formally incompatible," as Ramsey correctly points out (Ramsey 1923, 473). lf "A is red & A is blue" is a contradiction, a logical impossíbility, then the conjuncts cannot be elementary propositions (since elementary propositions are logically independent, according to the T). How to carry out the analysis was not an important task in the T. This is because, ln the T, discovering which propositions are elemen­tary is not the subject of logic, but of its application (T 5.557).6 Such an analysis must be possible and its results cannot collide with what we can know a priori concerning the possible results of such an analysis (4.211 and 5.134). We can know a priori, according to the T, that there must be logically independent elementary propositions, but we cannot tell a priori which propositions these are (5.557). Wittgeristein's suppo­sition is that the analysis, therefore, must be possíble, while Ramsey's objection could be paraphrased as "lf it must be possible, it must be shown how it goes.11 Ramsey is thus asking Wittgenstein to introduce a method to determine the elementary propositions that express the incompatíbilities.7

Wittgenstein tries to meet Ramsey's challenge with the elucidation of the supposedly intrinsic, implicit, properties of the visual field, for there "the forms color and visual space penetrate (durchdringen) each other" (MS 105, 41). The symbolism of the T might seem, at first, a promising tool for the task of explaining calor incompatibilities:

Do not calor and visual space relate to each other as argument and function? The forms of argument and function must also penetrate each other. (MS 105, 41).

However, it is immediately clear to Wittgenstein that the forms of argument and function have no intrinsic similarities with the special forms of space and calor. Considering what Wittgenstein says in the MSS from the first half of 1929 (especially in MS 1051 01-100; MS 106, 49-101) and in SRLF, we can see what kind of analysis based solely on the forms function and argument he had in mind at the beginning of 1929, and maybe at the time of the T, for calor statements.8 The íntended analysis will indicate why the particular forms of objects (colar and space, but also time) need to be taken into account, why they need to be introduced in the conceptual notation. That is, the logícal forms, the logical symbols, of functíon and argument will not be enough for the analysís.

Wittgenstein first thought that statements of degree could be analyzed as the logical product of "simple propositions of quantity" and

Phenomenology, 'Grammar,' and the 'Limits of Sense' 11

an extra supplementary clause - "and nothing more" (SRLF, 168). Such analysis should give, then, for instance, the "logical structure of color" (T 6.3751) and show how colors could be truth-functionally analyzed. This analysis, however, doesn't work. Even if one could assume that the correct analysis of "A is blue" showed that "A is not red" is implicitly given in it and, therefore, that the conjunction of "A is red and A is blue" was a contradiction, the very analysis of "A is blue" is problem­atic. Suppose, for instance, that E is an entity E and h is its brightness b.9 If E has two degrees of brightness, we represent it as E(2b). ln this case, however, E(2b) must be analyzed as Eh & Eh, which is obviously absurd since Eh and (Eh & Eh) have the sarne truth conditions. It is also not possible to distinguish the units of brightness in the conjunc­tion (say E(b') & E(b")), argues Wittgenstein, for then, in the case that we expressed one degree of brightness with Eb, we would have to ask whether b=b' or b=b" (SRLF, 168).1º Another problem with this kind of analysis concerns color mixture. If one says, "A is reddish-blue," one cannot analyze it as "A is blue and A is red." Even though reddish-blue is a mixture of blue and red, reddish-blue excludes red and excludes blue (MS 106, 99). What follows from "A is reddish-blue" is not that "A is blue,11 but "the opposite," as Wittgenstein says, i.e. that it is not blue (see MS 106, 101). If there is a hidden product in colar statements, then it follows that the product is not merely a conjunction of units. Since from "A is reddish-blue 11 it follows that A is not blue, one should expect that the quantity of blue in reddish-blue is O. Another problem with the idea of truth-functional analysis is the mixing incompatibility between complementary or opposite colors (for instance red and green) (MS 106, 71). There is no such thing as reddish-green, i.e. a logical product of red and green cannot exist. However, one should expect it to be possible if mixed colors are taken merely as conjunctions of colors. This means that Wittgenstein's truth-functional notation would, again, not express the real multiplicity of calor mixture and exclusion. So Wittgenstein ended up assuming that statements of degree cannot be further analyzed by his old logic and are, therefore, elementary.11 Thus, Wittgenstein concluded, there are elementary propositions that are not logically independent.12

Note that this conclusion allows, now, the introduction of numbers at the elementary levei. This was forbidden in the logical notation of the T, for it entails precisely the non-independence of elementary propositions - since, from "X is 2," it follows that "X is not 3," etc. Numbers, for Wittgenstein, will be an important characteristic of the new elementary propositions (propositions that simply don't contain

12 Wíttgensteín's Philosophical Developme11t

logical connectives) and will determine his treatment of incompatibili­ties (Section 1.3).

The notation (symbolism) grounded on the logical forms of argu­ment and function is, thus, insufficient to express the incompatibility of colors:

lf cp(r) and cp(g) contradict one another, it is because [each of] r and g completely fill in the cp and both cannot be in it. But this does not show itself in our signs.

lt must show itself, however, if we look, not at the sign, but at the symbol. For since this includes the form of the objects, it must show itself there, in this form, the impossibility of cpr · cpg. (MS 106, 79; PR §78; my emphasis).

Nothing in the form q;x or in the rules of conjunctlon shows (presents perspicuously) the incompatibility. That is, the "logical prototype" of the T (the forms of argument and function) does not correctly express the form of the objects described (see also WVC 41, 238, and 251). ln fact, no specific form (e.g., calor and space) is displayed in the symbolism of the T. The only forms that concern the Tare logical forms: variations of function and argument.13 Besides logical independence, the forms argu­ment and function are the requirements imposed on elementary prop­ositions: we know them a priori. The objects described in elementary propositions could not have a specific form as its determinant a priori characteristic; otherwise, elementary propositions would be logically dependent. According to the T, the incompatibility should be reduced to a truth-functional contradiction. This task, however, cannot be accom­plished with the help of the old symbolism alone, as we have just seen. The consequence of this is that the specific forms of the objects (color, space, and time) must be included in the analysis; they must be part of the symbol. They were not part of the symbols in the T. The word 'green' in "Green is green" symbolizes in two different ways because it expresses different symbols, namely, function and argument (see 3.32n, especially 3.322-3, and 4.24). The logical syntax of the T is, thus, a syntax with no specific forms (space, color, time); it is a syntax of the forms function and argument. There are no kinds of simples in the T. This is why simple objects are, in a certain sense, colorless (T 2.0232).

Moreover, the notation of the T is rrow clearly deficient because it allows the attrlbution of truth to propositions such as "The point A is blue & the point A is red", since both sentences must be recognized as elementary propositions:

Phenomenology, 'Grammar,' and the 'Limits of Sense' 13

lt is, of course, a deficiency of our notation that it does not prevent the formation of such nonsensical constructions, and a perfect nota­tion will have to exclude such structures by definite rules of syntax (SRLF, 171; my emphasis).

Thís means that the T doesn't succeed in fixing, for language, "a limit for ... the expression of thoughts" (T, introduction, and 4.114). The rule for the construction of all possible propositions, i.e., the general form of propositions does not draw precise limits to what is meaningful. That is, it allows for the construction of nonsense, as Wittgenstein would see it now, for it allows for the conjunction of "elementary propositions" like "A is red" and "A is blue" to be taken to be true (as is shown in the truth­table of the conjunction). It is classified as 'nonsense' instead of 'sense­less' because this kind of proposition cannot be reduced to a formal, truth-functional, contradiction as was assumed in the T. It must, now, be excluded from language. The rule also cannot construct all logical relationships, for "A is red, therefore A ís not blue" is not an inference based on truth-functionality. This inference, íf conditionalized, is not a formal tautology, as the conception of inference in the T assumes (T 6.126n and T 5.132-4). So what is at risk is also the explanation given in the T for the a priori nature of logic and inference; it seems that one cannot explain purely a priori why certain relationships of implication hold.

1.2 The ultimate analysis and the phenomenological language

Wittgenstein, however, doesn't give up the project of establishing limits for what can be said. This is the project of the T that will survive until at least the BT (1932-3), after many adaptations. ln 1929 he thinks that necessary relations that are not truth-functional need a different symbolism. He still thinks that he can demarcate the limits of language with new rules "which tell us in which connections only a word gives sense, thus excluding nonsensical structures" (SRLF, 162). But in order to lay down those rules, an investigation not pursued in the T must take place: "Such rules, however, cannot be laid down until we have actually reached the ultimate analysis of the phenomena in question. This, as we all know, has not been achieved" (SRLF, 171). This is the task that Wittgenstein thought he had to accomplish early in 1929. "Theultimate analysis" that he had in mínd was a phenomenological one, and the results of such analysis should be presented in a· "phenomenological

14 Wittgenstein's Philosophical Development

language.11 ln what follows 1 explain the central characteristics of the phenomenological language (Sections 1.2.1-4).

1.2.1 A complementary notation without hypotheses

ln his writings from 1929 Wittgenstein does not define the idea of a 'phenomenological or primary language' and never gives a full version of it. Later, in 1932, he says that it was supposed to be "the description (Beschreibung) of immediate sense perception without any hypothetical addition" (MS 113, 123r from 05.19.1932; also BT, 491). The "descrip­tion of immediate sense perception" c0uld be understood in two ways. First, 'description' could mean a set of descriptions of faqs (of externai properties) that are either true or false: things that we say- for instance, "the blue patch is right to the yellow patch." Second, 'description' could mean a description of forms (internai properties), i.e. a presentation of the rules related to colar, space, and time supposedly given implicitly in the phenomena. For instance, the representation of the rule "visual space has absolute direction" in a perspicuous notation that shows such a rule, i.e., such an internai property of visual space.

Note that the two meanings of 'description' are derlved from the philosophy of the T. The first sense of description appears, for instance, in T 4.023: "[ ... ]A proposition is a description (Beschreibung) of a state of affairs" (see also T 2.02331, 3.24, 6.341).14 An example of the second meaning of 'description' is the general form of propositions:

It now seems possible to give the most general propositional form: that is, to give a description (Beschreibung) of the propositions of any language whatsoever [ ... ] It is clear that only what is essential to the most general propositional form may be included in lts description (Beschreibung) - for otherwise it would not be the most general form. (T 4.5; my underlining)

This second sense of description is to be taken as the description of internai properties expressed in the symbols of the notation presented in the T (see also T 3.317, 5.156, 5.4711, 5.472, 5.501, 6.125).

For the project of the phenomenological language, the first sense of description (description of externai properties) is relevant only to the extent that it should be in the phenomenological language that propo­sitions are analyzed and verified. Verífiable descriptions indicate what is possible and, thus, give us a clue about what belongs to the second kind of description, which demarcates what is possible. lt is the second

Phenomenology, 'Grammar,' and the 'Límits of Sense' 15

sense of description that characterizes the phenomenological Ianguage, i.e., the explicit notational presentation of the rules of the forms space, time, and calor that should be grasped in "the ultimate analysis of the phenomena" (SRLF, 171).

ln the next pages, 1 attempt to explain what Wittgenstein had in mind with such a notation. Before doing this, 1 further explain the qualifica­tion "without any hypothetical addition." Assumed (hypothetical) enti­ties of scientific theories are to be avoided in a purely phenomenological description of, for instance, colors (MS 105, 90). Colar is taken as a visual phenomenon and not as a physical ar chemical one. Further, descrip­tions of three-dimensional objects, before the analysis is dane, are also to be takerf as hypothetical. This should not be misunderstood. Objects are not mythological elements of a hypothetical world (MS 105, 108); nor are they inferred things or things constructed by means of classes of sense data. They should be, however, definable in terms of the forms implicit in the phenomena. The indefinables that should ground their definitions are the forms space, calor, and time, and the like (MS 106, 171) presumably, "the like" refers to the forms of sounds, smells, etc.

However, not only are assumptions concerning descriptions of things (ordinary objects and scientific assumptions) taken as hypothetical, but also assumptions concerning the way we represent the a priori, in logic, ordinary objects (by using nouns and predicates, functions and arguments, relations, etc.).15 ln the T it is left to the "application of logic" to determine whether elementary propositions should be expressed by means of unary, binary, ar n-ary functions. Wittgenstein says that it cannot be decided a priori which the elementary proposi­tions are, and that "it would be completely arbitrary to give any specific form" (T 5.554). By "specific form" Wittgenstein means something like a "27-termed relation.'' From the strictly a priori viewpoint of the T, whether such a form is needed for such and such particular case would be an arbitrary decision (f 5.5541-2). This point of view is also reflected in a comparison that Wittgenstein makes between the T and Carnap's project in the Aufbau:

1 wrote, We are unable to specify the form of elementary proposi­tions, and that was quite correct too. lt was clear to me that here at any rate there are no hypotheses and that regarding these questions we cannot proceed by assuming from the very beginning, as Carnap does, that the elementary propositions consist of two-place relations, etc. (WVC, 182; my emphasis).16

16 Wittge11stein's Philosophical Development

But independently of what "specific forro" they míght have (two­place or twenty-place relations), the simple names must be artículated by functions and relatíons in elementary proposltíons in the T. This point is obvlous because of the analysis that Wíttgenstein offers for color incompatíbilities in 1929, as we have seen. Moreover, it is clear, in many passages of the T, that argument and function are the symbols in elementary propositions (T 4.1252, 4.24, 5.441, 5.52-5.534, and 5.47). See, for instance, these examples from the T:

1 write elementary propositions as functions of names, so that they have the forro 'fx', '<ji(x, y)', etc.

Or 1 indicate them by the letters 'p', 'q', 'r'. (4.24). Wherever there is compositeness, argument and function are

present ... (5.47)17

ln the Notebooks from 1914-16, the ideais also explicit:

Every simple propositíon can be brought into the forro <jlx. That ís why we may compose all simple propositions from this forro. (NB, 71; from 04.16.16)

However, even the assumption that the distinction between argument and function (or subject, predícate, and relatíons) is a tool to be used in the analysís of propositions is no more than a problematic hypothesis for Wittgenstein in 1929. As we have seen, such a distinction on the elementary levei was not helpful for showing how statements of degree are to be dealt with. It turns out that the a priori use of such a dístinc­tion could be, in principie, an obstacle for a perspicuous representa­tíon of the relatíons of inference and dependency among statements of degree. The subject/predicate and the function/argument forros are both símplifications used in the mode that we project reality.

This is why Wittgenstein introduces the símile of the projection of geometrical figures in MS 106, 107, translated by Wittgenstein in SRLF:

Let us imagine two parallel planes, 1 and II. On plane 1 figures are drawn, say, ellipses and rectangles of different sizes and shapes, and it is our task to produce images of these figures on plane II. Then we can imagine two ways, among others, of doing this. We can, fírst, lay down a law of projectíon - say that of orthogonal projectíon or any other - and then proceed to project all figures from 1 ínto II, according to this law. Or, secondly, we could proceed thus: We lay

Phe11ome11ology, 'Grommar,' and the 'Limíts of Se11se' 17

down the rule that every ellípse on plane l is to appear as a circle in plane II, and every rectangle as a square in II. Such a way of represen­tation may be convenient for us if for some reason we prefer to draw only circles and squares on plane II ... If the facts of reality are the ellipses and rectangles on plane 1 the subject-predicate and relational forms correspond to the circles and squares in plane II (SRLF, 164).18

Thus, the problem with our usual ways of projection ("subject-predi­cate and relational forms") is quite obvious for him:

These forms are the norms of our particular language into which we project in ever so many different ways ever so many different logical forms. And for this very reason we can draw no conclusions - except very vague ones - from the use of these norms as to the actual logical form of the phenomena described. (MS 106, 109; SRLF, 164-5; Wittgenstein's emphasis)

The forms that we use to describe reality are hypothetical and as such are not to be assumed as correct in the phenomenological investigation. As we have seen in Section 1.1, the forms of function and argument did not account for the incompatibilities in the analysis of statements of degree. Wittgenstein's point is that the distinction concept/object and the distinction subject/predicate are both assumptions concerning how (in which form) we represent reality, i.e., they are "norms [of representa­tion] into which we have pressed the sentence," and not the "the results of an analysis" (MS 107, 13). Because they are simplifications, they may hide important differences that we may be able to distinguish in the analysis of phenomena. One obvious difference that the old notation hides is that the forms of color, space, and sound are different symbols (MS 106, 83).

Wittgenstein's worries concerning function and argument also help us to understand why he was in the search of a phenomenological language. One might think of two possible ways of expressing the anal­ysis of phenomena: (1) either as a notation that perspicuously shows what is logically relevant in the phenomena - i.e., shows visually how the incompatibilities work and show how truth-functionality can be added to elementary propositions - and can rule out nonsensical formulations (something like a begriffsschrift for phenomena); (2) or as the written rules that characterize the description of phenomena. Waismann gives a good example of these two possibilities: one can either state that one cannot say "A is north from B and B is north from A" or present

18 Wittgenstein's Philosophícal Development

a map, where this nonsense simply cannot be expressed because the map has the right multiplicity (it expresses ali essential elements of the phenomena with their relevant relationships in a simple way and excludes what in principie cannot take place - see WVC, 79). These two ways of presenting the results of the phenomenological investigation are not necessarily exclusive, even though a "system of signs with the right multiplicity makes the syntax unnecessary" (WVC, 80). 19

Wittgenstein's phenomenological language should be a notation, as described in 1 above. It was supposed to be a description in the sense that the general form of propositions: is a description. A notation for phenomena should "show" in its symbols that the color incompatibili­ties are excluded from language as nonsense, and "show'' that "A is not blue" follows from "A is red". It should "show" these in the sarne sense (in a visual and perspicuous way) that a truth-table shows the relations of dependency and inference among propositions, and "two proposi­tions 'fa' and 'ga' show that the sarne object is mentioned in both" (T 4.1211; my emphasis).2º It is also in this sense that a map shows spatial relations. 21

The phenomenological language (notation) complements the nota­tion of the T. This gives us an extended version of the Tractarian search for the correct begríffsschriff. lt does not present, thus, a break with the goals and tools of the early philosophy. Neither is it the expres­sion of the abandonment of the early views concerning the delimita­tion of language and thought. The T presents bis own earlier version of a correct begriffsschriff, a notation that can be used to express any thought and its relationships with other thoughts. The idea of such a notation is, of course, Fregean (see, for instance, the introduction to Begriffsschrift). Wittgenstein's version of the correct begriffsschríft in the T is presented always with criticisms of Frege's and Russell's notations. They are notations of the kind Wittgenstein was looking for, but not yet free ofmistakes (T 3.325).22 Wittgenstein's begriffschriff in theT includes truth-tables that operate by means of logically independent elementary propositions (those explain the nature of necessity and inference), a new account of generality, the elimination of the identity sign, and a general form that generates ali possible propositions (thus, also shows the limits of language). The fundamental mistake of Russell and Frege in Wittgenstein's eyes was not seeing the tautological nature of Iogical 'laws' and inferences and the non-representative character of logical connectives. That is, both did not have a criterion for logical laws, and so they did not understand the difference between saying and showing (see the two meanings of description given above: the description of

Phenomenology, 'Grammar,' and the 'Limits ofSense' 19

tnternal properties of language and the description of facts). Becaúse they were unable to see this, they could also not see that identity, for instance, is not "an essential constituent of the begríffsschríft" (5.535) -if ali necessity is tautological truth-functionality, 'a=a' is not a law of logic and must not be presented as such in the correct begriffsschríft (5.534).

Ali those Tractarian notational improvements are assumed as given, as a priori primitives given implicitly in the nature of propositions. 23

However, in 1929, the general form of propositions, which assumes the existence of logically independent elementary propositions, cannot determine by itself the nature of proposltions anymore, i.e., it does not determine by itself what can be written in the language of the concep­tual notation. At the elementary levei, then, Wittgenstein has to deter­mine what counts as a proposition and how the necessary relations on this level work. Therefore, in 1929 the search for a correct conceptual notation is broadened: Wittgenstein needs a new and compatible nota­tion for the atomic parts of propositions (1 come back to this issue in Section 1.2.3).

Wittgenstein's philosophy of logic ln 1929 has, then, as it were, two levels: one related to the correct notation of what is truth-functional, and the other related to a notation for relations of necessity that are not truth-functional, but connected to the given forms of colar, space and time or "internai structure of propositions" (WVC, 74). 24

1.2.2 Objectivity of the notation and the determination of sense

Wittgenstein's phenomenological language was neither meant to be a private language, nor a solipsistic description of supposedly 'inner objects.'25 A description of inner objects would not even be a descríp­tion in the relevant sense (in the sense of a notation, a symbolism). Moreover, such a phenomenological language is intended to be an objec­tive language in the following sense: it should be connected to ordlnary language and ordinary descriptions, its main goal is the perspicuous presentation of the rules of the visual space (not mine or yours), and it is not meant to construct or explain inter-subjectivity - inter-subjectivity is a given; it is not questioned right from the beginning. These points will be further elucidated in the next pages.

The phenomenological language (elementary levei) must be clearly connected to ordinary language and its descriptions of facts. The logical necessity contained in the "internai structure of propositions" and the tautological necessity already presented ln the T must have some kind of unity; otherwise, the very project of the clear presentation of the logic

20 Wittgenstein's Philosophical Development

of our language and its limits cannot take place. Another obvious reason for the need of this connection is that the phenomenological language must also be the place where the analysis of ordinary language can be completed. The complex expressions of our language are translated into (or defined by means of) immediate elements of the phenomenological language, which shows what the determinate sense of these proposi­tions is. It is in this language that the "complete analysis" assumed as possible in the T is to be represented (T 3.25). We should be able to show ln it, thus, that "a proposition that mentions a complex will not be nonsensical, if the complex does not exist, but simply false" (T 3.24). The phenomenological language must, thus, show how the sense of propositions is determinate (T 3.23) by presenting the prímitive signs that "cannot be dissected any further by means of a definítion" (T 3.26). It is, then, intended to reveal what underlies the "clothing" of ordi­nary language, which "disguises thought" (T 4.002). As in the T, ordi­nary language is in perfect order (T 5.5563), but needs to be correctly expressed in a notation that can help us to see the disguised thoughts. Thus, it is still part of the project of the T of expressing a notation that avoids misunderstandings suggested in ordinary language:

... the phenomenological language represents the sarne as our ordinary physical modes of expression and has only the advantage that we can express some things in a shorter way and with less danger of misun­derstandings. (MS 105, 122; my emphasis).

The objectivity of the language lies, thus, in the fact that the sarne things are represented by ordinary modes of expression and in the phenomenological language. The particular misunderstanding that Wittgenstein has in mind ln this passage shows another important aspect of the objective nature of the phenomenological language. Wittgenstein críticizes the tendency to subjectívize the visual field:

It could be, for instance, practical to give names to my and other people's hands and, when talking about them, not need to talk about their relation to a person, because it is itself not essential for the hands and because the ordinary mode of expression could make it appear as if the relation to an owner of the hand was something lying in the essence of the hand itself.

The visual field has essentially no owner. (MS 105, 122; my emphasis; in PR §71 only the last two sentences appear).26

Phenomenology, 'Grammar,' and the 'Limits ofSense' 21

The point of the analogy between hands and the visual field is to make clear that even if we usually make reference to the subject when talking about the visual field with "the ordinary mode of expression,11

it is nota fundamental part of the description of what is immediately given. Neither the subject nor the body are "purely visual" (MS 105, 135), i.e. they are not intrinsic characteristics of the isolated visual field. Wittgenstein's goal is to grasp the visual field objectlvely, l.e., without reference to a subject. The passage just quoted above continues in the foilowing way:

Let's now assume that l assume that 1 always see one object among others in my visual field- namely, my nose. Somebody else obviously does not see this object in the sarne way. Does it mean, however, that the visual field that I am talking about belongs to me? That it is, there­fore, subjective[?] No. lt has only been grasped subjectively; and an objective space was opposed to it, which is, however, only a construc­tion wíth the visual space as basis. ln the - secondary - language of the "objective - physical - space" the visual field is called subjec­tive, or rather, whatever in this Ianguage immediately corresponds to [the] visual space is called subjective. As if we would say: in the language of the real numbers that which corresponds directly to the cardinal numbers is called the "positive integers" (MS 105, 124; PR §71; my italics; modified translation).

This passage is interesting for two reasons. lt shows, first, that one way of looking at phenomenological descriptions is by taking them as a part of ordinary language, as the positive integers are included in the real numbers. Ordinary language and phenomenological language, in this sense, are supposed to form one 'system' (a unity). One can, however, look at the positive integers as a system in itself, isolated, as one can, supposedly, isolate the phenomenological language. ln this isolated language, instead of showing the position of the eyes and making refer­ence to one subject, for instance, one can shade the representation of the objects in such a way that the shading makes clear what is reaily seen without making reference to the eyes and the subject (MS 105, 126; PR §71). If we, then, "determine the position of the eyes" through what is shaded, argues Wittgenstein, then this is "only the translation of one mode of expression into another" (MS 105, 126; PR §71). A trans­lation can be achieved by means of rules of translatíon, the definitions (T 3.343). A 'translation' is, in fact, very important if both languages are

22 Wittgensteín's Philosophica/ Development

related in one system and if the phenomenological language is a part of a broader system of language.

The other important point made in the passage above is that we should grasp the vísual field objectively, wíthout reference to a subject, in order to understand and explain the nature of necessity involved in phenomenological statements. As he observed, the vísual field as such is to be taken as "autonomous," independent of a subject (MS 107, 1). An essential property of thls space is, supposedly, that "the descríp­tion of the visual field describes an object and has no suggestion of a subject" (MS 107, 126). This is why Wittgenstein from the beginning of 1929 onwards talks about "our visual space" and "the visµal field" (MS 105, 31-9), and not about the visual field of a solipsist' or a "private" visual experience. Wittgenstein was interested in the visual field as such, primarily in its color and space forms, and not in a private realm of prlvate experiences.27

When Wittgensteín talks about the 'hypothetical' heis not doubting or seeing as uncertain the existence of objects. This because "the hypothesis is only an assumption about the practical/correct/ mode of the representation" (MS 105, 108). Thus, we don't talk about a "hypo­thetical world" ln ordinary language (MS 105, 108). The physical world is not a world that we barely recognize:

What we think of in physical space is not the primary that we can more or less recognize; but, rather, what we can recognize from the physical space shows us to what extent the primary goes and how we have to interpret the physical space. (MS 106, 177)

Note, " ... the phenomenological language represents the sarne as our ordinary physical modes of expression" (MS 105, 122). The relative subjectivity of the phenomenological lies in the fact that we always see the world perspectively, which does not imply that the phenomeno­logical is private:28 "Somebody else obviously does not see this object [the nose] in the sarne way. Does it mean, however, that the visual field that 1 am talking about belongs to me? That lt is, therefore, subjective[?] No" (MS 105, 124; PR §71; quoted in full above). The real question is, then, whether we can find the structure, the forms, of the world seen perspectively in order to correctly interpret the physical space (MS 106, 177). Such a structure is to be found in the part of language that corre­sponds to the description of phenomena.

Thus, the possibility of existence or the possibility of construc­tion of the 'externai' or 'inter-subjective' world is not in question in Wittgenstein's phenomenological project. It is taken for granted.

Phenomenology, 'Grammar,' and the 'Limits of Sense' 23

Wittgenstein is interested in the visual field, and not in "mine" or "yours."29 Therefore, even the reference to the visual is misleading in the expression 'visual space.' It might lead to solipsism:

Even the word 'visual space' is unsuitable for our purpose, since it contains an allusion to a sense organ which is as inessential to the space as it is to a book that it belongs to a particular person; and it could be very misleading if our language were constructed in such a way that we couldn't use it to designate a book without relating it to an owner. It might lead to the idea that a book can only exist in relation to a person. (MS 107, 4; translated in PR §74; my emphasis)3º

The reason Wittgenstein appeals to a language of phenomena is the need to complement the notation of the T with a clear presentation of the necessity involved primarily in the forms of space and color. There is no other reason for it. But why, in the first place, find such a language and not use ordinary language or the language of science? Wittgenstein never explicitly says why, but at this point it should be clear that it is in the visual field (space) that we fínd the most elementary forms that need scrutiny (space, color, and time).31 Another reason is that Wittgenstein thought that the direct apprehension of phenomena could not be elimi­nated, even in ordinary language or scientifíc analysis. When discussing the recognition of a color as the sarne in the visual field, Wittgenstein makes the point:

Of course, it would also be possible to say that the color is the sarne because chemical investigations do not disclose any change. So that if it doesn't look the sarne to me then 1 am mistaken. But even then there must still be something that is immediately recognized. (MS 107, 236; translated in PR §16)

Whatever we use as a criterion to determine the color, the criterion will still depend on our recognition. Therefore the appeal to sense percep­tion seems to be needed. Thus, it is natural, even though not necessary, to think that the examination of forms given in the phenomena has some kind of priority. ln the end they are the 'things' that we refer to when we speak.32

1.2.3 A complementary notation grounded in the structure of experience

Wittgenstein is, of course, aware that his investigation concerning phenomena is not strictly a priori, as his insights concerning the nature

24 Wittgenstein's Philosophical Development

of logic in the T were meant to be. These had the character of the absolute a priori, i.e., insights that are completely independent of any facts in the world, that something specific existed (for instance, colors). He clearly states that his investigation in 1929 is "in a certain sense a posteriori" (SRLF, 163). lt is a posteriori in the sense that it deals with phenomena, which could possibly be different or even not exist - "whatever we see could be other than it is" (T 5.634). Wittgenstein, however, says that only in a "certain sense" is it a posteriori. This suggests that in a certain sense it is not a posteriori. The odd status of Wittgenstein's investigation seems to be derived from his belief, in 1929, that the investigation of forms is an a priori investigation, even if it is dependent, .in a sense, on experience. This puts him very close to saying that he~ is discovering necessities in the phenomena themselves (SRLF, 169). The reason for this is that the "actual form of the phenomena" (SRLF, 164), he thought, makes clear that the notation of the T is wrong, i.e., that it does not have the right multiplicity of specific forms of phenomena. Suppose "RPT" means "The color R is in P at time T" and "BPT" means "The color B is in P at T". ln the notation of the T, we would have:

RPT

T T F F

BPT

T F T F

RPTandBPT

T F F F

The first line is obviously wrong, for the conjunction is not only not true, but also impossible. What if we write 'F' in the first line? The truth-table would, then, express a contradiction. This, once the truth-functional analysis fails, obviously violates the rules of conjunc­tion and, more important, does not express the multiplicity of the phenomena, because according to the rules of color and space the first line is an ímpossibility (it should not exist). The old notatíon neither shows that the conjunction above is a real contradiction, nor that the first line is nonsense. Thus, the notation of the T "gives the proposition a greater logical multiplicity than that of the actual possibilities" (SRLF, 170; my emphasis). Since "actual possibilities" given by the forms of space and color show what is wrong with the notation of the T, one might hope that their analysis can show how a complementary notation with the correct logical multiplicity might be achieved.33

The notation is, thus, meant to be complementary, for it is meant to show how the incompatibility of the first líne of the truth-table above should be explained. lf the explanation works, the T might need only

' Phenomenology, 'Grammar,' and the 'Limits of Sense' ZS

a revision. Explaining this line with a phenomenological language is, in any case, what the notation of the T needs in order to demarcate the limits of sense.

The phenomenological investigation is a priori, even though it depends on what is empirical, not only because it deals with the form of phenomena, but also because it is supposed to present the conditions of sense given by a priori forms, as opposed to the factual conditions of truth (the comparison of proposition with reality). ln 1929 Wittgenstein believes himself to be dealing, as in the T, with the rules that allow us to distinguish what it makes sense to say from mere nonsense - even if the phenomenological a priori is not as "pure" as logic in the T. This is why Wittgenstein's phenomenological language is also called 'primary.' lt expresses, supposedly, what comes before truth:

... physics strives for truth, i.e., right predictions of events while the phenomenology doesn't do this - it strives for sense not truth.

Physics differs from phenomenology in that it wants to establish laws. The phenomenology establishes only the possibilities.

So phenomenology would be, therefore, the grammar of the descríption of those facts on which physics built up its theories (MS 105, 3)

11The grammar of the description of the facts" is, as we have seen, the notation in which the actual descriptions take place. One important trait of the a priori character of the proj ect of the T, then, remains in the phenomenological project: the investigation is still considered an inves­tigation of sense (of "structure" or forms) free from presuppositíons: forms are sense determining and meant to be hypotheses-free.

The search for a phenomenological language is, thus, a good expres­sion of Wittgenstein's method in 1929: "this method is essentially the transition from the question for the truth to the question for the sense" (MS 105, 46). The method consists basically in finding the conditions of possibility, the structure of the forms, given in the phenomena (time, space, and color). I give examples of the application of this method in Sections 1.3 and 1.5.5.

1.2.4 Verification and sense

ln 1929-30, the verifiability of propositions becomes a central issue in Wittgenstein's philosophy. According to the T, to "understand a propo­sition means to know what is the case if it is true" (T 4.024). One can understand a proposition without knowing if it is actually true or false,

26 Wittgenstein's Phi/osophical Development

but one understands it only if one knows under which circumstances it can be true or false. This view remains central in 1929:

My central thought (Hauptgedanke) is that one compares the proposi­tion with reality (MS 107, 155 from 10.06.1929).

But, in 1929 and after, how a proposition is compared with reality is a central notion. The "comparison with reality11 is taken to be deter­mined by the mode or method of verification: a proposition is consid­ered comparable to reality only if it can be, in principie, verified. ln the case of mathematical propositions, a proposition is "verified" if a method of calculation justifies a given equation (MS 105;8-10). A prop­osition indicates, thus, not only its truth conditions, but also how one can establish its truth:

A proposition tells how it is verified. Compare the universality of the genuine propositions (eigentliche Saetze) with universality in arith­metic. lt is verified in a different way and is therefore a different [universality].

The verifícation is not an indication of the truth, but the sense of the proposition. (MS 107, 143; Wittgenstein's emphasis)34

A proposition p has sense only if it indicates a method of verifying it, only if it "tells how it is verified." By verification Wittgenstein means the "method of verification," and not an actual verification: "The sense of a proposition is the method of its verification" (WVC, 79 from 01.1930; see also MS 105, 10). 35 lf a proposition has sense, then it has a method of verification, i.e. it is verifiable.

The verifiability of propositions is a new, and maybe more precise, expression of the requirement of the T that each proposition must have a determinate sense. ln the T Wittgenstein writes: "A proposition must restrict reality to two alternatives: yes or no. ln arder to do that, it must describe reality completely" (T 4.023). ln the T, he does not give a criterion for what exactly counts as a complete description of reality, except that the truth-value of the description must be determinate. Presumably, a completely analyzed proposition would also, as it were, clear up ali the vagueness of ordinary sentences. ln 1929, Wittgenstein still holds to the "requirement (Forderung) of the determinacy of anal­ysis" (MS 106, 49), but heis also explaining in what sense a proposition must describe reality completely: it must be decomposed in such a way

Phenomenology, 'Grammar,' and the 'Limits of Sense' 27

that its content is expressed by propositions that describe immediate experience.

There is, then, a relation between complete analysis and phenomeno­logical language, the language of immediate experience, which gives the ultimate specific forms (color, space, and time) of propositions and determines their content. lf it must be possible to analyze the content of any meaningful proposition into components that are directly related to immediate sense experience, then direct and complete verification takes place in a phenomenological language. The content of proposi­tions of the specific forms (color, space and time) must be, presumably, somehow analyzed; it must be translatable into this language.

To dose this section 1 want to suggest that the post-1929 equivalence of a proposition having sense and having a method of verification should be seen as a response to the problems related to Ramsey's objec­tion. Because the notions of 'color' and 'phenomenological space' make problematic the explanation of the nature of logic and inference of the T, the idea of a general "logical space" has to be revised. Wittgenstein must now deal with the specificity of sense and nonsense in the case of notions related to color, space, and time instead of talking about sense in general terms as in the T (there he talks about the logical space; for instance in T 2.202, 3.4, 3.42 and 4.463). Specific forms or "specific spaces" in the T, as we have seen, were supposed to be reducible logi­cally to the function/argument primitive form (prototype); otherwise elementary propositions would not be logically independent. ln the project of the phenomenological language those specific forms or 'spaces' need to be investigated so that we can understand what it makes sense and what it doesn't make sense to say in these 'spaces.' This asks for particular methods of investigation in each of these spaces, which implies an approach different from the absolutely a priori approach of the T concerning the logical space. Each particular "space" may present particular conditions of sense that are different from the other "spaces."

The particular conditions of sense are made explicit in the way that we verify a propositíon. Propositions about colors ask for methods of verification that differ from those related to sounds, and these methods show also what cannot, in principie, be verified, i.e. what is nonsense in color attributions. If one understands the particularity of color-form, one understands why one cannot apply the conjunction to excluding propositions that are part of this space. It is, then, the particular method of verification of propositions with reference to colors that shows, for

28 Wittgenstein's Philosophical Development

instance, that we cannot make true the statement "A is red and A is blue;" for to make true "A is red" is, at the sarne time, to falsify "A is blue." "A is blue and A is red" is unverifiable when both conjuncts are supposed to be true; therefore, the first line of the truth-table of this conjunction is nonsense. When we make clear why it is unverifiable, we express new conditions of sense, which the rules of conjunction (the rules of the logical space) are insufficient to express. Thus, 1 suggest, the principle is introduced because the purely a priori limitation of sense of the T fails as soon as the analysis of phenomena is taken seriously.

1.3 Phenomenological (primary) language: a draft and a method '

The project of a phenomenological language doesn't last long. Already in October 1929, more or less six months after Wittgenstein started thinking (and writing) about the project, he gives up his search for a phenomenological language. Its purpose was clear (to create a nota­tion for logical relations that are not explained by the notation of the T); however, its details were not. Here 1 present its dra~, and how Wittgenstein arrived at it by means of his 'transcendental method.' ln Section 1.4, 1 explain why it did not work.

The first step to be taken in the construction of a phenomenological language should be the analysis of visual phenomena. 36 The idea is to investigate how the essential constituents of the visual field, calor and space, work and relate to each other: "We have to remember that each part of the visual space must have a calor and that each colar must take a part of the visual field" (MS 105, 41). The "must" here is, arguably, the "must" of a rule of the structure of calor and space.

This structure is intrinsic because it is constituted by internal proper­ties, necessary properties implicit in the visual space. It is not the result of, for instance, corporeal feelings, the "space of feelings."37 It is the isolated apprehension of the visual field and the isolated apprehension of the space of feelings that make the correspondence of both spaces possible:

... the position in the feeling space (so 1 want to call it) has nothing to do with the position in the visual space, both are independent from each other, and if there were no absolute direction in the visual field, then one could not make the correspondence between it and the direction in the feeling space. (MS 105, 41)

Phenomenology, 'Grammar,' and the 'Limits of Sense' 29

The absolute direction must be taken as implicitly given in the visual field, arguably, because this is the only way to make possible the corre­spondence of what is visual to certain feelings. One can, presumably, correlate feelings with what is seen in the visual field only if the direc­tions of space are implicitly given in it (presumably, because a correla­tion presupposes two correlated things). Absolute direction is, thus, a condition of possibility of recognizing the position of feelings. This, as one might call it, 'transcendental standpoint' grounds the following 'transcendental argument':

One can . also say that the visual fiel d is an organized (lined up -gerichtet) space, a space in which there is above and below, right and left. And this above and below, right and left has nothing to do with gravitation or the right and left hand. lt would still have its sense, for instance, if we looked ali our lives at the stars through a telescope. (MS 105, 31)38

The point of the argument is that even if there were no body feelings but only a visual field constituted by what is seen in a telescope, we could still distinguish directions in it. Moreover, argues Wittgenstein, we can determine a location in the visual field when it has only one calor. We can say (i.e. it makes sense to say) that "This part of a red patch is red" or say, of the uniformly yellow field, that its center is yellow (MS 106, 67). Since only one calor is present, there are no referential points that might help us to locate a part of the field. Thus, it seems we can conclude, the directions that allow us to locate a part of the field are intrinsically given in it.

Wittgenstein also argues that one would be able to recognize the position of, for instance, stars that disappear and reappear simply by looking at them. This because, so he thought, one can know the place where a star was in relation to the border of the field merely by looking at the star (MS 105, 33). lf there were no absolute direction in the visual field, argues Wittgenstein, it would not be possible (it would not make sense) to talk about the sarne or different places in it. ln this case, it would not be possible to say that "two stars that alternately appear and disappear at the sarne distance from the border [of the visual field]" (MS 105, 37). Given that we can see and say that two stars appear and disappear at the sarne distance, we must assume places or positions in the visual field; therefore, as a condition of possibility of our talking about places, the directions of the visual space must be intrinsically

30 Wittgenstein's Philosophical Development

given. Thus, concludes Wittgenstein, absolute directions are given in the visual field; they are internai properties of the visual space.

The argument above shows us the centrality of transcendental argu­ments, grounded on what it makes sense to say, for Wittgenstein at the time. They are, for Wittgenstein ln 1929, what characterizes "the method of philosophy." ln the sarne context, as above, he says:

We could not only "not know whether", but it would make no sense to talk about the sarne or different places in this context ... The crite­rion of the structure is precisely which sentences in it make sense -not which are true. To look for this is the method of philosophy. (MS 105, 37; PR §206)

A phenomenological investigation has a non-hypothetical character, as already argued above, because it is an investigation of what makes sense, and not of what is true. ln order to limít what it makes sense to say, one should look at the "conditions of possibility" of what is said. Thus, according to Wittgenstein's method, the descriptions of facts related to the visual field indicate the structure of the visual field (its forms) and how it is to be represented. The conditions of possibility of what is said with sense about the visual space give us the criterion for finding its structure and its representation.

Wittgenstein also introduces, in agreement with his method, a tran­scendental argument connected with a reductio ad absurdum argument in order to show that "the concept of 'distance' is directly given in the structure of the visual field" (MS 105, 64). Consider the following line:

A B e

Let's call the segment AB 'a' and the segment AC 'b'. Wittgenstein argues:

It seems to me that the concept of distance is given immediately in the structure of visual space. Were it not so, and the concept of distance only associated with the visual space by means of a correla­tion between the/a visual space without distance and another struc­ture that contained distance, then the case would be conceivable in which, through an alteration in this association, the segment a, for example, will appear greater than the segment b, even though we

Phenomenology, 'Grammar,' and the 'Limits of Sense' 31

still observe the point B as always between A and C. (MS 105, 64; translated in PR §208; slightly modified translation).

The argument assumes that we meaningfully talk about distance in the visual field - we say, for instance, as indicated above, that two stars appear at the sarne distance from the border of the visual field. lts point is to prove that distance is intrinsically given in the field. This because it makes no sense (it is absurd), Wittgenstein argues, to talk about a structure of distance externa! to the field in arder to explain the sentences about distance in the field that make sense. Suppose that we are aware that B is between A and C when we look at b, which is of course usually true. lf it is not the case that what makes this possible is the distance intrinsic to the visual space, then we have to accept that it would be possible, i.e., we could meaningfully say, that a is greater than b, whlch is absurd. Why so? lf there is no intrinsic distance in the visual space, it is only by means of another structure (say, a ruler) that we can determine distances. Suppose the ruler gives us the measure 10 units for the distance b (this is my example, not Wittgenstein's). Suppose now that the ruler shrinks (say, because it is elastic) and the association ruler/visual space changes, so that it gives us the measure 15 for the distance a. lf the visual space had no intrinsic distances, then a could appear longer than b. Since this is absurd, argues Wittgenstein, we have to conclude that distance is immediately given in the visual space.

Now, if distance is immediately given in the visual space, it seems that the possibility of representing such distances by means of numbers must "arise from the structure of the visual space" (MS 105, 100).39 How could we otherwise represent distances? This implies that a segment unit (Einheitsstrecke) containing the "specific spatial element" (MS 106, 45) will be part of the method of projection. This unit, whatever it might be, is not expressed by means of "rigid rulers" (starre Massstaeben), which are physical tools constructed according to Euclidean geometry.40 What exactly such a purely visual segment unit would be, however, is left unspecified by Wittgenstein.

Wittgenstein's transcendental arguments grounded on the sense of propositions, if sound, allow him to assume that a "system of coordi­nates is contained in the nature (Wesen) of the [visual] space" (MS 105, 35). These arguments supposedly guarantee the central elements of such a system, namely, position (right-left, above-below) and distance. The system of coordinates is also presented in SRLF:

32 Wittgenstein's Philosophical Development

10

9

84-~~~~~~~---~~~...i--~~~-

7

6

5

4

34-~~~~~~~-t<--'--L....-'--'f--~~~-

2

2 3 4 5 6 7 8 9 10 11 12

Wittgenstein's arguments given above, if sound, give us two space dimensions, a system of coordinates with two axes. This is one of the reasons why, in SRLF, he says that his analysis is far from being complete (SRLF, 166).41 Wittgenstein, as far as l know, never presented an argu­ment to show that three-dimensionality is intrinsically given in the visual space or that it can be added to it.42

ln his drafts on the phenomenological notation, descriptions of phenomenological facts related to the visual field are determined by means of a two dimensional system of coordinates (MS 105, 70, and SRLF). A 'name' would be, then, a complex composition of points in the system of coordinates of the visual field (i.e., a 'name' would be a complex of numbers). Given the arguments for the structural char­acterization of distance and directions in visual space, Wittgenstein assumed that numbers must be part of the phenomenological notation: "The occurrence of numbers in the forms of atomic propositions is, in my opinion, not merely a feature of a special symbolism, but an essen­tial and, consequently, unavoidable feature of the representation" (SRLF, 166; my emphasis).43 With the system of coordinates and its numerical representation, Wittgenstein thinks to have in hands the "right multi­plicity" of the visual field:

Imagine a system of rectangular axes, as it were, cross wires, drawn in our field of vision and an arbitrary scale fixed. lt is clear that we then can describe the shape and position of every patch of calor

Phenomenology, 'Grammar/ and the 'Limits of Sense' 33

in our visual field by means of statements of numbers which have their significance relative to the system of caordinates and the unit chosen. Again, it is clear that this description will have the right logical multiplicity, and that a description which has a smaller multi­plicity will not do. (SRLF, 165)44

A description of the visual field with the "right multiplicity" is a description that "can describe all configurations that we can think of" (MS 106, 69) - nonsense is, thus, excluded in such a description. If we use "statements of numbers," we can indicate any possible patch in the visua) field. Only in such a system, supposedly, can we express the diversiiy of calors in visual space by assuring that any possible "object" would have a possible camplex name given by tuples of numbers and, thus, canstitute the carrect "preparation for a sentence" (MS 105, 70). We could say, for instance, following the system of coor­dinates introduced by Wittgenstein, that a certain spot is red in this way: "[6-9, 3-8] R, where 'R' is yet an unanalyzed term" (SRLF, 166; my emphasis). ln this way, the notation shows mathematically that "A is red and A is blue" is nonsense, since 'A' could not have the sarne coor­dinates twice (this must be taken as an interna! property of space). Also, the coordinate system allows us to represent the relation between in visual space and shows how to express basic inferences using that concept. Moreover, cancerning the division of parts in the visual field, it expresses its "final reality" (mínima visibilia) as well as its "infinite possibility" (MS 106, 189), since the intervals between numbers are continuous (SRLF, 166).

The structure of calor cauld also include numbers in its description in arder to show calor exclusion. This Wittgenstein does not do in SRLF, where he treats red, for instance, as "a yet unanalyzed term" (SRLF, 166). One cauld, however, express the calor distribution of a rectangular patch of one colar by giving the caordinates of the patch and by attributing a number to the calor (MS 105, 70; MS 106, 101). Other camplexities cancerning calor, however, must be accounted for: possible mixtures (red and yellow), impossible mixtures (red and green), and proximity of colors (that orange is dose to red and yellow, for instance, and distant from blue). Those camplexities are internai relations among primary calors (red, blue, green, and yellow, plus black and white) that can be immediately recognized (MS 105, 88-90). They must be shown in the correct notation (MS 105, 88). In arder to express them, Wittgenstein introduces the calor octahedron

34 Wittgenstein's Philosophical Development

pretty early in his investigation (MS 105, 98; see PR §221). Here is the model:

White

green

black

ln the octahedron, four primary colors (red, yellow, green, and blue) are taken as primítive elements forming a square in the middle - black and white are represented as the two vertices of the octahedron. Each possible coloris a mixture of a hue (four primary colors and their possible links in accord with their proximity), a certain brightness (white), and a certain saturation (black). Colors that do not combine (colors that don't have intermediary links) occupy opposite positions in the middle square. This is the case of the oppositions of green-red and yellow-blue, which also show, thus, that a "reddish-green" coloris an impossibility (i.e. nonsense). Note that the octahedron roughly expresses, therefore, how colors can mix, how they cannot mix, and how colors exclude each other (the identity of one color in contrast to other colors). ln this way, the octahedron combined with the system of coordinates indicates a path to explain what the old truth-functional analysis did not explain (see end of Section 1.1). Thus, both complement the logical notation of the T. Note that the new notation introduced is intended as a represen­tation of part of our language; it is implicitly given in it, one could say.

Wittgenstein's treatment of color, thus, shows how the complex phenomena of the visual field are expressed by primary colors and their relations (the internai color relations and spatiality). His draft

' Phenamenalagy, 'Grammar,' and the 'Limits af Sense' 35

of a phenomenological language for color and space also leaves open the possibility of a more detailed determination of both forms. The numbering system for colors, which would complete the analysis of colors (MS 105, 70), is not to be found in Wittgenstein's analysis.45 The incomplete treatment of color is accompanied by the absence of time in the system of coordinates (SRLF, 166). What we have seen is, then, a draft of the phenomenological language: the 11direction in which ... the analysis of the visual phenomena is to be looked for" (SRLF, 166; my emphasis).

One can conjecture, however, that this draft would be extended in other ways. A projection into three spatial coordinates could be in prin­ciple introéÍ.uced to connect phenomenological and ordinary proposi­tions about ordinary objects. This would allow, presumably, for a single verification for each proposition of ordinary language, for it would connect descriptions of three-dimensional objects with their possible ways of verification. The system of coordinates would also allow for a phenomenological and physical time systems (perhaps new coordi­nates), which could be used in connection with three physical space coordinates. These complements to the system of coordinates would make clear how the phenomena are connected to the hypothetical (how phenomena and propositions about ordinary objects are related, etc.).

Contrary to what Wittgenstein thinks in MSS 105-7 and in SRLF, however, the draft presented here already does not correctly express the "actual phenomena." As a consequence, Wittgenstein will give up the project of the phenomenological language.

1.4 Phenomenological language: the end of the project

There are two interconnected problems related to the "inexactness of the sense data" (MS 107, 171) that bring Wittgenstein to abandon the project of the phenomenological language. They are introduced in this paragraph and developed in what follows. First, as we have seen, Wittgenstein assumes that distance is an intrinsic property of the visual field and that it is to be expressed numerically ln the notation; he comes to think, however, that the visual field does not allow measurement (it shows an internai ambiguity, inexactness, or "indeterminacy factor"). Second, the precise determination of the internai ambiguity of the visual space and its correct translation into a general conceptual nota­tion turns out to be unlikely. Contrary to the initial expectations, the system of coordinates presented as the guiding idea of the phenomeno­logical language does not seem to present exactly the phenomena. lt

36 Wittgenstein's Philosophical Development

is at best a rough representation. This means that even the draft of the phenomenological language seemed to be flawed. 1 begin with the fírst problem and then move to the second.46

A very ordinary fact is that some things may look a certain way and yet be different. A geometrical figure in physical space may look like a circle, but inspection with a microscope may show that it is actually a 100-sided figure. Two tines on paper may look the sarne, but after measuring them we may conclude that one is 24 and the other 25 units of measure. ln ordinary language we explain the difference as a differ­ence between appearance and reality (be/ng). Should we say that the sarne distinction applies to the visual field, if we describe what we see? Things are more complicated: ,

... should I now, nonetheless, say that in the visual field something can appear differently than it is? Certainly not! Or that a distance in the visual that once turns out to be n and once turns out to be n+l tum out to be the sarne? Just as little that. (MS 107, 29)

We can neither say that n and n+l are the sarne, nor say that the distinction between appearance and reality applies to the visual field. To say that they are the sarne, on the one hand, would be a kind of contradiction. The difference between appearance and being, on the other hand, entails the assumption of a hypothesis to describe what is immediately seen; for in order to explain that something appears in a certain way, we need to assume an externai standard, a hypothesis, that determines how it really is. This, of course, would undermine the whole project of the phenomenological notation.

If we cannot say that n=n+l is valid i.n the visual space, it seems that we should not use the word 'sarne' to talk about lengths that we see in the visual field, for a segment a (n units) and a segment b (n+l units) might be seen as having the sarne length. But if we eliminate the word 'sarne,' and if we see two lines (a and b), in this account, we cannot talk about seeing the number of parts of each. We could not infer that they have the sarne length, even after counting their parts determined by any unit of measurement (assuming the lines have parts). But this implies that if we want to accept that we can count the number of parts, then we have to admit also that what we count is not necessarily what we see (otherwise it would be inconsistent). This because we "cannot notice if the number of parts in [a segment] d changes between 24 and 25" (MS 107, 30). Here we have the first variation of what Wittgenstein will calI "the problem of the 'inexactness1 of sense data" (MS 107, 171).

Phenomenology, 'Grammar/ and the 'Limits of Sense1 37

The problem is that the inexactness of what is seen in the visual field is not correctly represented in a coordinate system. Note that in a numbered system of coordinates we must distinguish n and n+l, precisely because in it the numbers determine the representation. The simple or direct expression of distances in a coordinate system, thus, does not express the correct multiplícity of distances in the visual field. This, to say the least, indicates "how difficult it is to describe what is really seen" (MS 107, 30). Wittgenstein returns to similar problems 130 pages later ln the sarne manuscript and comes to the conclusion that it is better to give up his phenomenological language project. Now 1 tum to those passages, which Wittgenstein wrote right before giviríg up the phenomenological language (he did so in MS 107, 176).

ln what ways can we preserve the non-hypothetical character of the phenomenological notation? Could we translate the indeterminacy of the visual field into a coordinate system? The difficulty here is that in order to describe directly what is seen, we lose the distinction of what is and what appears to be; at the sarne time, we seem to need the distinc­tion when describing the phenomenological, for the words needed to describe our immediate experience in the visual field seem to have meaning only if it is assumed. ln ordinary language, this distinction grounds the use of standards of measurement based on rules of geom­etry; it makes them compatible with what is experienced. The denial of such a distinction implies a complete change in the meaning of the words used (this is the case of the word 'sarne,' as we have seen). Thus, it seems that we need new concepts, or new rules for old concepts, if we want to establish a phenomenological language. What comes to our aid when we try to describe the visual field, however, are only the ones we have, the "old" ones:

We would need new concepts and we take always again the ones from the physical language (MS 107, 163 from 10.11.1929)

Since new concepts are not given, we only have a change of meaning in the words we already know. This change of meaning has paradoxical consequences. If we deny that the distinction between appearance and reality is valid in the visual fíeld, and so change the meaning of our ordi­nary words, we cannot, for instance, classify lines in the visual field any longer, for the words of our ordinary language (which includes geom­etry) would seem to be insufficient. This is because our words are used with certain standards (hypotheses), i.e. in a comparative way: "there

38 Wittgemtein's Philosophical Development

where there is a certain degree of imprecision, there is also complete precision possible" (MS 107, 163).

Think, for instance, of the following definitions grounded on Wittgenstein's phenomenological primitive elements, the colors:

"Straight [line] is what is not curved." And "circle is a line of constant curvature". (A line is the limit of two colors. A point the place where three colors meet each other). (MS 107, 163)

These definitions seem to be, as it were, minimal. Consider, however, the following drawing:

a

The seen segment a-b does not correspond to the definitions. Wittgenstein points this out in the following way:

My elucidation of circle and straight line presupposes that it makes sense to say of each segment that it is either straight or curved. If 1 look at a short piece of a curve, then 1 cannot see that it is curved. And thus, it seems that a piece of a curve could be straight ...

Again and again we need an expression like "I don't see that this line deviates from a circle, but 1 cannot say that 1 see the circle." And, yet, we cannot say it if we regard the visual field as absolute. But it shows that our mode of expression is completely insufficient. (MS 107, 164-5; Wittgenstein's emphasis)

To regard the visual field as absolute means to consider that, when isolated, absolute directions and distances are intrinsic to it (see Section 3.2). If absolute directions and distance are given, and the coordinate system employed by Wittgenstein is justified, such ambiguity or inex­actness must somehow disappear in the representation of the visual space; yet, visual phenomena suggest just the opposite:

A line in the visual field needs to be neither straight nor curved; naturally, the third possibility is not "dubious (zweifelha~)" (this is

Phenomenology, 'Grammar,' and the 'Limíts o( Sense' 39

nonsense), but one had to have a different word for this, or rather substitute the whole mode of expression with 'straight' and 'curved' for something else. (MS 107, 168 from 10.12.1929; my italics)

A third possibility would be a violation of the excluded middle - a violation of the determinacy of sense, of course - and would imply a direct incompatibility between the truth-functional logic and its complement, the phenomenological language. It seems that the phenomenological language is not a part of our language and that it cannot be, so to speak, abstracted from it; rather, it seems to require words that we don't really have. However, we don't have any good clue about what they would look llke, and the concepts that we really use are the ones that we usually use.

Just having new concepts or finding visual space's own rules, however, would not be sufficient for the phenomenological notation. For íf these mies or concepts are not translatable into the rules of the system of coordi­nates notation (and, thus, into ordinary and physical languages), we don't have the tool of analysis and verification of ordinary or physical proposi­tions. ln this case, it would not be the complementarynotation Wittgenstein was looking for, because it would destroy the "fabric" of logic:

... in logic one cannot put something at the si de of truth-functions, something that merely is there, without penetrating them and forming one fabric with them (MS 107, 128)

How, then, to translate the concepts used to describe the visual field into Wittgenstein's extended begriffsschrift? The general strategy for a solution would be to consider that the correct description of the visual space must include vagueness (Verschwommenheit) or Jack of clarity: "inexactness is depicted (wiedergegeben) with inexactness" (MS 107, 162). Only in this way would the representation of the visual field be strictly correct. But how could the vagueness or inexactness exactly be expressed in the notation?

The inexactness, the vagueness, one could think, must be expressed in the construction of new terms for the phenomenological language and a geometry that is not exactly Euclidean:

It is like this, and we see only that the visual space does not follow the rules of - for example - Euclidean space. (MS 107, 28)

But this view also turns out to be problematic. Ih arder to construct a translatable phenomenological geometry (geometry for the visual

40 Wíttgenstein's Philosophical Development

space), one has to be at least clear about how inexact the apprehension of the visual space is, for only this could give the exact parallel of both geometries and would give a hint ofhow the construction could proceed. Wittgenstein comes to believe that he does not have even the means to express the exact limitation (circumseription) of the inexactness.

Before reaching this conclusion, Wittgenstein actually considers a way to circumscribe the inexactness of phenomena based on Klein's idea of the visual threshold and on Hjemslev's "natural geometry."47

The strategy would be to assume the idea that the limit to what is seen is represented by means of double limits ín the representation. Suppose that we take the following correspondence with Euclidean geometry seriously: ,

... what corresponds to a visual circle ín Euclidean geometry is not a cirde but a class of figures that include the circle but also the 100 sided figure, etc. [figures that look like the circle]. (MS 107, 172)

What Wittgenstein has in mind here is the representation of the ambi­guity in the following way:

Note that such a representation would be possible by means of the coordinate system (two circles with the sarne origin and different radii, say, 7 and 8, expressed by two equations (MS 107: 167, 169, and 173)). However, it does not work. First of all, what the representation repre­sents is not exactly what we see in the visual field. That is, we don't "distinguish both circle limits" in what is seen (WVC, 58); we see one circle. Thus, we don't get the exact multiplicity and the drawing does not express correctly the phenomena. Second, it is unclear what would precisely characterize the class of circles. Wittgenstein suggests that the class could "be formed by the vibration of a circle" anq immediately points out that this does not work:

... for why should 1 take the lines that come to being through the vibratíon of a circle and not the ones that come to being through the vibration of the 100-sided figure? (MS 107, 172 from 10.15.1929)

Phenomenology, 'Grammar,' and the 'Limits of Sense' 41

Wittgenstein concludes:

And here 1 bump into the biggest difficulty, for it seems as though even the exact circumscription of the inexactness would be impos­sible. (MS 107, 172 from 10.15.1929)

lf even the inexactness of what is apprehended in the visual space (the basis for the phenomenological language) cannot be determined inside the phenomenological notation, then there is little hope of relating the phenomenological conceptual notation to the conceptual notation of our language that is already expressed in the T - the "one fabric" of logic would be destroyed (MS 107, 128; quoted above). Thus, Wittgenstein comes to think that the vagueness intrinsic in the phenomena would not be translatable:

But the big question ls: can one translate the "vagueness" of the phenomenon into an imprecision of the drawing? lt seems to me no. (MS 107, 165)

The problem is that the vagueness of the phenomenological apprehen­sion, taken in isolation and without hypotheses, is "different in kind" (MS 107, 166) from the vagueness of a drawing, and it does not matter which effects we use to express the vagueness. The system of coordinates would always, as it were, incorrectly determine the intrinsic indeterminacy, inexactness, or vagueness of the visual field. As happens with a concept like 'heap,' there is no satisfying analysis of the concepts in the visual field that can precisely demarcate them (see MS 107, 168 and 173).

Thus, Wittgenstein comes to the conclusion that the project of a phenomenological language is not worth pursuing anymore. It would not fulfill its purpose. A few pages after the quotations above, Wittgenstein explicitly gives up the phenomenological language:

The assumption that a phenomenological language would be possible and that it would first say what we have to/want to express in philosophy is - 1 believe - absurd. We have to manage with our ordinary language and just understand it correctly. l.e., we are not allowed to be induced by it to talk nonsense. I mean: what 1 call sign must be that which one calls sign in grammar. (MS 107, 176; from 10.18.1929)

Wittgenstein takes the phenomenological language to be "absurd" because we don't have the appropriate concepts to build it. It cannot,

42 Wittgensteín's Phí/osophícal Development

therefore, fulfill its purpose as the tool for verification (and the explici­tation of the condítíons for the determination of sense) and, thus, its purpose as the complementary notation to determine the limits of sense (the llmits of sense, 1 take it, are "what we have to/want to express in philosophy"). lt is not, however, absurd in the sense of being in prin­cipie impossible. It only forms a system that is different from severa! uses of concepts in ordinary language, in which a distinctlon between appearance and reality is fundamental and in which the standards of measurement are derived from Euclidean geometry. What we are bound to accept is the fact that we give preference to the language of Euclidean geometry, which is part of our 'grammar,' in relation to. the language of the visual field (WVC, 59). ln PR, when he restates the abandon­ment of the phenomenological language, Wittgenstein does not call the project 'absurd' anymore (as in the passage above). There he says that it is a goal that he no longer takes to be needed (noetig) (PR §1; WVC, 45). This means that a complete notation (purely logical, as in the T, or logical-phenomenological, as in 1929) is not needed for bis goals.

ln order to avoid confusions, it is important to be clear about why the phenomenological language (its draft) does not work. The problem is that the numerical representation of the phenomena by means of the system of coordinates does not represent them correctly. Strictly speaking, it does not correctly represent the phenomena even lf we "translate" the ambiguity of phenomena by the means at our disposal. That representation was, of course, fundamental, for it would allow the representation of any possible state of affairs (any thinkable presenta­tion of phenomena) and, thus, also make clear what is impossible (the first line of truth-tables concerning conjunctions of color statements, for instance); moreover, it would correlate. the descriptions of phenomena wíth our physical language and, thus, with truth-functional logic. We can say, thus, that although the representation expresses the multiplicity of phenomena (in the sense that we can describe any phenomena with it), it does not express their intrinsic distance and does not circumscribe correctly their ambiguity.

Wittgenstein still uses the transcendental arguments for absolute direction and distance in the structure of the visual field (seen in Section 1.3) after giving up the phenomenological notation. He reintro­duces those arguments some months later in PR (§§206-8) andas late as the BT (BT, 454-461). They are used to show that visual space is struc­turally directional and contains implicitly a kind of distance but not a precise numerical distance, for we cannot "apply exact concepts of measurement in the immediate experience" (MS 107, 212).48 Thus, the

Phenomenology, 'Grammar/ and the 'Limits of Se11se 1 43

arguments have a new function and they don't guarantee a numerical representation of phenomena anymore. The system of coordinates is rather a rough representation of the visual field.

What does Wittgenstein have in mind, however, in the passage quoted above, when he says that "we have to manage with our ordinary language" and that we should not be "induced by it to talk nonsense"? We have to manage ln our ordinary language because there is nothing else to lean on, for logic and phenomenological notatíons could not take care of the task of the T (determining the limits of language and the nature of necessity). When Wittgenstein abandons the phenom­enological language, he thereby abandons the idea that notations (symbol­isms) have /a fundamental role in philosophy. This is an important step in Wittgenstein's philosophical development. However, the step is less large than it might at first appear, for the role of a phenomenological language is transferred to 'grammar.' ln the passage quoted above, Wittgenstein also says that we have to avoid being "induced by ordi­nary language to talk nonsense." This means that ordinary language itself, with its rules of 'grammar,' fills the gap that was supposed to be filled by the phenomenological language. ln what follows, the general traits of 'grammar' in the PR are presented.

1.5 The comprehensive 'grammar'

Wittgenstein introduces a new style of investigation later in 1929: grammar (syntax)49 as the guardian of sense. 'Grammar' (or 'syntax'), as it is shown in the next pages, must be understood as a substantially extended version of the 'Iogical syntax' of the T. The logical syntax of the T had two basic symbols (function and argument), the forms implicit in anything thinkable that were expressed in variables of the notation. The phenomenological investigation introduced specific forms (time, space, calor, etc.) in the notation. 'Grammar,' thus, should account for all forms without relying on a complete logical-phenomenological notation. ln arder to unify his philosophical inquiry, Wittgenstein also included formal concepts of mathematics in his 'grammar.' Common to all these branches of 'grammar' is the goal of making explicit the limits of language and the elimination of "wheels turning idly" (WVC, 48; PR §1).

1.5.1 Phenomenology, verification, and the comprehensive 'grammar'

Wittgenstein makes clear that his new goal, after giving up the phenom­enological language, was to get rid of confusions in language without

44 Wittgenstein's Philosophical Development

constructing a new symbolism. lnventing a new language is not needed: "we need not invent a new language or construct a new symbolism, but our everyday language is the language, provided we rid it of the obscuri­ties that lle hidden in it" (WVC, 45). This does not mean that the aban­donment of the phenomenological language was also the abandonment of the phenomenological investigation. lt is only the fundamental role attributed to a complete notation that is abandoned.

Wittgenstein still thlnks in PR (from 1930) that the method of verification is the sense of empirical propositions and that the direct inspection of phenomena is the basis for the methods of verification of propositions: "lt is only the findings in the course of .a verification, i.e., the phenomenological statements, that are true or fa1se" (WVC, 101 from 03.22.1930). The abandonment ofthe phenomenological language only shifts Wittgenstein's idea of what counts as verification. The idea of a complete, final, and determinate, verification is no longer in place, for the notation in which this would be accomplished is no longer a project. The possibility of verification, however, remains a criterion of sense. Propositions need only to be related somehow to phenomena in order to have sense:

Each sentence is an empty game of strokes or sounds without the relationship to reality and the/its/ only relationship to reality is the mode of its verification.

Ali that is essential is that the signs it does not matter in which complicated way- at the end relate to the immediate experience, and not to a middle Iimb (a thing in itself). (MS 107, 177)

The phenomena themselves do not determine sense, for they are ambiguous. However, a criterion of the sense of signs is to refer to the phenomena, even if this is done in a "complicated way." The expression "complicated way" indicates that it is to be done without a complete logical-phenomenological notation, which could presumably guarantee a completely determinate verification. Note that Wittgenstein writes the passage above right after giving up the phenomenological language (see also below). This means that he still considers the verífiability of sentences a fundamental criterion of sense (even without a unifying tool of verification). This criteríon applies also to physics:

If physics describes a body of a specific form in physical space, it has to (even if implicitly) assume the possibility of verification. The

1 r

Phenomenology, 'Grammar,' and the 'Limits of Sense' 45

places where the hypothesis connects with the immediate experience must be designated (provided for). (MS 107, 255, from 01.25.1930)

Signs have meaning, then, "in as much as - and only in as much as immediately observable phenomena (such as points of light) do or do not correspond to them" (MS 107, 223; PR §225). 50 Thus, propo­sitions and hypotheses work in a similar way: propositions "have the character of hypotheses" (MS 107, 249). 51 The character of a hypoth­esis, its essence, is the generation of "an expectation by admitting of future confirmation" (MS 107, 253; PR §228). ln order to see why the comparison of sentence and hypothesis holds, it is useful to think of the undetermined many perspectives that may confirm that "the book is on the table" is true; also, the many possible positions of the book on the table. One could say that ordinary propositions are like hypotheses because they have a "greater multiplicity" than the one of a "single verification" (MS 107, 255). A phenomenological language, presumably, if completed, would allow such a single verification, for it would allow each possible description of the visual field expressed in the system of coordinates to be projected in three dimensional space. Presumably, all verifications could possibly be expressed in a unified phenomeno­logical notation. Any single determination of the coordinate system would, then, guarantee the verification of complex sentences describing three-dimensional objects. Without a phenomenological language, however, a proposition acquires the multiplicity of a law, since many verifications (many perspectives, many positions of the object) might be needed to confirm the expectations that a proposition generates.

Each hypothesis or proposition (expectation), however, must be capable of being compared with reality (phenomena) in the present. It is the comparability with the present "alone that makes the expectation into a picture" (MS 107, 288; PR §229). This is so because a proposi­tion must make sense at any time, and not only when it is verified or confirmed. It is, thus, the present possibility of comparison of proposi­tion (expectation) with reality qua phenomena that guarantees its sense: "ln phenomenology it is always a matter of possibility, i.e. of sense, not of truth or falsity" (WVC, 63, from 12.20.1929).

The comparison between hypotheses and propositions should not, however, suggest skeptical doubts concerning ordinary objects, even though the verification of a hypothesis in physics ''is never complete" (MS 107, 254; PR §228; see also WVC, 101). 'Never' because a hypothesis points to an undetermined future (MS 107, 283; PR §228). A proposition,

46 Wittgensteín's Phílosophica/ Development

say, 111 see pairs of shoes (now)" expresses, as a hypothesis, an undeter­mined number of possible expectations that will be confirmed or not in a definitive way, but it is confirmed and we do find out its truth:

If 1 say: "What I see here in front of me is a pair of shoes11, and this

is a sentence at all, then there must exist a possibility of finding out for sure whether it is like this or not. If there were not such a possi­bility, then I could not teach a child the language, for 1 would not be allowed to say 11 You see, these are shoes11

, but \only\ "these seem to be shoes". (MS 107, 177)

We do say, of course, "these are shoes" (here, now) and we establish the truth of sentences (this is, indeed, quite trivial when they don't point to the undetermined future). Wittgenstein continues the passage quoted above in this way:

Where illusion is possible, there also seeing the truth is possible/seeing the truth must also be possible/.

In all philosophical theories we find words whose sense is well known to us from the phenomena of daily life employed in an ultra­physical way, therefore false. How is it, in this sense, with the expres­sion 111 am certain that"[?]. (MS 107, 177)52

An 11ultra-physicaI11 way would beta employ words like "Does the box still exíst when I'm not looking at ít?" (MS 107, 160; PR §57). For Wittgenstein, "the only right answer would be 'Of course, unless someone has taken it away or destroyed um (MS 107, 160; PR §57). The appearance of uncer­tainty of verifícation lies in the fact that words in ordinary Ianguage 11oscillate between different meanings;" once meaning is fíxed, we have a criterion for the truth that is certain (WVC, 48).

The 11ultra-physical" way of using words reminds one of Russell's use of 'know' and 'certain', and his doubts concerning the existence of the externai world, as expressed in his hypothesis of a solipsísm of the present moment ín, for instance, An Outline of Philosophy (from 1927). Russell misuses language when he entertains the skeptical idea of an illusion that cannot be, in principie, verified (whether the world carne into existence five minutes ago, for instance). It is in our language and in its relation to what can be verified that we will ask for the verifica­tion - and therefore about the sense of sentences and avoid "ultra­physical" phílosophlcal nonsense:

l 1 Phenomenology, 'Grammar,' and the 'Limits of Se11se' 47

Doubtless: l compare the proposition with reality. Someone tells me this wall is white; I look at it, and say No. (MS 107, 241)

The important changes, then, at the end of 1929 are: the abandon­ment of the search for a complete notation (phenomenological and logical), the consequent loosening of verification requirements, and the introduction of a general 'grammar' of language. From these changes, others unfold (as I intend to show in what follows).

Also the idea of analysis as a decomposition of ordlnary sentences is modífied:

A propofüion is completely logically analyzed if its grammar is made completely clear: no matter in what way it is written or expressed. (MS 108, 88; PR §1- my translation)

The expression of the rules of 'grammar' ('syntax') replaces the old idea of analysis: the way rules are written (the notation, stríctly logical, as ln the T, or partly phenomenological, as in 1929) is not the most relevant tool for it. Verification is still a requirement of sense, but what shows the determinacy of sense is the rules of the grammar of the existing language: "Ali that is possible and necessary is to separate what is essential from what is ínessential in our language" (MS 107, 205; PR §1). Analysis, in the old sense, is not determinate, but rules are. Those rules, supposedly, are given implicitly in our language; we only need to make them clear. We don't see them immediately beca use "our grammar lacks, above ali, perspicuity" (MS 108, 31; PR §1).

Analysis of phenomena is now taken to be the correct expression of the rules of 'grammar' of the phenomena: either represented in a perspicuous notation that "expresses the rules of grammar through íts externa! appearance" (MS 108, 88) or in the form of rules of syntax expressed by sentences. We might express the 'grammar' of colors wíth the position of colors in the color-octahedron (for instance, the distance of colors indicate that there is no possible combination of red and green), or we simply express the grammatical rule that "we can speak of a reddlsh blue, but not of a reddish green" (MS 107, 282; PR §39). Both presentations are correct. We might, thus, even use the tools developed for the phenomenological language in order to express the 'grammar' of language. This is precisely what Wittgenstein does with the coordinate system, with which he shows how space exclusions work, and the color octahedron, with which he shows how the combination and exclusion

48 Wittgensteín's Philosophícal Development

of colors work in our language (see WVC 73-80; PR §§1 and 218-224). Rules of 'grammar' might complement them (WVC, 80).

If, then, the rules of language are expressed and the limits of sense are made clear, we achieve the sarne result that we would achieve with a phenomenological language:

A recognition of what is essential and what is inessential in our language if it is to represent, a recognition of which parts of our language are wheels turning idly, amounts to the construction of a phenomenological language. (MS 107, 206; PR §1)

Note that the purpose of language is,. as in the T, the representation of reality. The new grammatical project has also the sarne sense-limiting function as a phenomenological language, for the purpose of both is to exclude the "wheels turning idly.11 This was also the purpose of the Tractarian philosophy of logic and its begriffsschrift, but this one turned out to be insufficient, for it allowed the expression of nonsense (or wheels turning idly) - as we saw at the beginning of this chapter.53

An interesting issue that comes out of the new conception is the status of logic. At the end of 1929 (12.22.1929) Wittgenstein writes:

How far does logic become uncertain (unsicher) through the uncer­tainty about the analysis of the elementary propositions? - What remains fixed? (Was steht fest?). (MS 108, 30)

What "remains fixed" is what does not need to be changed even after the recognition of the mistakes of the T and the abandonment of the phenomenological language.54 The views of the Tare supposed to still be valid wherever the truth-functions apply:

There where the propositions are independent everything remains the sarne [as in the T]: thus, the complete theory of inference and so on. (WVC, 76 from 12.30.1929)

This means that the notation of the Tis nota perspicuous representa­tion of the whole 'grammar' of language, but a part of it: "the rules for 'and,' 'or,' 'not1 etc. which 1 represented by means of the T-F notation, are a part of the grammar of these words, but not the whole" (PR, §83).

Analysis in the sense of the T cannot show, however, in a precise way the determinate sense of sentences. Here lies the uncertainty. The precise limits of language, given in the T by the general form of propositions,

r 1

Phenomenology, 'Grammar,' and the 'Limits of Sense' 49

cannot be given once and for all, since "language means the totality of propositions" (MS 108, 55; PR §85; Wittgenstein's emphasis). The anal­ysis of "elementary propositions" showed the vagueness or incomplete­ness of propositions that describe the immediately given, and not how the sense is determinate (compare PR, §87 with T 5.156). What we need, thus, is not a special phenomenological notation; we need a "perspic­uous representation of the grammatical rules" (MS 108, 89; PR §1). Grammatical rules of our existing language show the determinacy of sense and the logical role of words. This opens the possibility, however, of regions in our language in which inexactness and ambiguity are not easy to eliminate. Thus, it is a natural step for Wittgenstein to begin to investigáte the nature of vagueness also in ordinary language after 1930.55

The important change where the propositions are not logically inde­pendent is that they should be seen as a system of excluding proposi­tions (MS 108, 35; PR§82; from 12. 24.1929). A system of propositions indicates a type, which is expressed by a particular kind of word in language. Ordinary language and the function/argument logic conceal those types by employing a too general rule of projection. Nouns and what takes the place of an argument are expressions whose meaning have particular rules. lt is the task of philosophy to clearly present those types, their rules, and the necessary relationshíps of the propositions inside each system. Therefore, Wittgenstein says: "Grammar is a 'theory of logical types"' (MS 108, 105; PR§ 7). 56 Types and propositional systems should be described in the chapters of the "book of grammar":

The words 'Color', 'Sound', 'Number' etc. can appear in the chapter headings of our grammar. They need not occur within the chapters but that is where their structure is given. (MS 108, 99; PR §3; slightly modified translation)

lt is the book of 'grammar' that presents the rules of language, i.e., the structure of the types, and, thus, presents a "perspicuous representa­tion" of what makes sense to say in it. The book of 'grammar,' thus, rein­corporates the goal of the T: the determination of the limits of language and elimination of nonsense.

The idea of various systems of propositions suggests a comprehensive 'grammar' that takes the place of the logic of the T. This means, of course, that what was said in the T "doesn't exhaust the grammatical rules for 'and', 'not', 1or1

1 etc." (MS 108, 31; PR §82; my emphasis). Thus, the idea of a tautology in the truth-functional sense is not fundamental

50 Wittgenstein's Philosophical Development

anymore, for not all logical relationships have a truth-functional tauto­logical nature. As Wittgenstein says, "only the rules of syntax are essen­tial and they have always been applied, a long time before anybody knew what a tautology was" (WVC, 91).

The completely a priori, tautological, logic of the T is clearly insuf­fícient to account for all kinds of inferences. Inferences like "John is 5 meters tall, therefore heis not 61 4, 3 ar 2 meters tall,11 "the book is green, therefore it is not red," "This note is a C, therefore it is not a D,11 etc., are all non-truth-functional and escape the philosophy of logic of the T: "the rules for the logical constants form only a part of a more compre­hensive (umfassend) syntax" (WVC, 74 from 01.02.1930): They do not escape, however, Wittgenstein's comprehensive 'granímar' (syntax), for it indudes propositional systems and specific types. According to Wittgenstein's 'grammar,' the analysis of propositions of degree shows that they form systems of propositions, and each system is compared to reality as a whole. ln the sarne way that we use a ruler to measure the length of an object (and see that a given measure excludes all others), all propositions of degree implicitly have such a system of exclusion:

The statements describing for me the length of an object form a system, a system of propositions. Now it is such an entire system of propositions that is compared with reality, not a single proposition. lf 1 say for example, that this ar that point in the visual fíeld is blue, then 1 know not merely that, but also that this point is not green, nor red, nor yellow, etc. 1 have laid the entire color-scale against it at one go. This is also the reason why a point cannot have different colors at the sarne time ...

All this 1 did not yet know when 1 was writing my book: at that time 1 thought that ali inference was based on tautological form. At that time 1 had not yet seen that an inference can also have the form: This man is 2m tall, therefore he is not 3m tall. This is connected with the fact that 1 believed that elementary propositions must be independent of one another, that you could not infer the nonexist­ence of one state of affairs from the existence of another. (WVC, 64, 12.20.1929)

Thus, the systems of propositions are envisaged to explain the infer­ences that are not truth-functional, those that escape the logic of the T. One important question concerning these different spaces that is not answered by Wittgenstein is how they relate to the truth-functional space of the tautological inferences based on the logical constants. He

Phenomenology, 'Grammar,' and the 'Limits of Sense' 51

thinks, however, that there is something common to both, namely, we don't have to look at how things are in the world in order to make infer­ences (MS 108, 53; PR §84). This point should become clearer in Section 1.5.5.

1.5.2 Classification of words

· Each propositional system is inside a particular space of 'grammar.' Wittgenstein explains what he means by "space": "l call space that which we have to be certain about when we search" (MS 108, 155). 1 can only look for a red book or expect it, for instance, if 1 know that 1 look visually for the color. 1 cannot look for the color in the hearing space (usin'g my ears). Each space has its own exclusive rules (rules of 'grammar') that are independent of other spaces or systems. Visual, audible, and tactile spaces (WLC30-32, 6), for instance, have different rules and when we mix spaces in our sentences we say nonsense: ·"A black color can become lighter but not louder. This means that it is in light/dark space but not loud/soft space" (MS107, 279; PR § 45). ln this way, 'grammar' does not allow certain combinations of kinds of words. Note that grammatical 'kinds' or 'types' are direct descendants of the phenomenological forms, which already broadened the Tractarian idea of a symbol.

ln order to determine which kind of word a word is (its "type") one observes how it fits under different substitutions: "To be of the sarne logical kind two words/expressions must be able to be substituted for each other" (WCL30-32, 5). This indicates that they have the sarne grammatical rules. Two words with the sarne kind of rules are the sarne kind of word (MS 108, 171).

When we more closely inspect the 'grammatical' rules (or syntactic rules), we see that our language conceals different kinds of words under general names as 'noun' or 'object':

If someone confronts us with the fact that language can express everything by means of nouns, adjectives and verbs, we can only say that then it is at any rate necessary to distinguish between entirely different kinds of nouns etc., since different grammatical rules hold for them. This is shown by the fact that it is not permissible to substi­tute them for one another. This shows that their being nouns is only an external characteristic and that we are in fact dealing with quite different gemes of words (Wortgattungen). The part of speech is only determined by ali the grammatical rules which hold for a word, and seen from this point of view our language contains countless

52 Witts!'11stei11 's !'llilusupllirnl 1 Jevrlop111e11t

different species of words ( Wortarte11). (MS 107, 210; PR §92; slightJv modified translation; Wittgenstein's emphasis) '

Even though such a classification of words reminds us of the classifi­cation of words in gran1111ars of English, German, etc., it is very different from them, for the most basic classifications of ordinary grammars are suspicious according to Wittgenstein's 'gram mar.' Classifications of ordinary grammars are based on "externai characteristics" and 50 they conceal the real kind of word a word is, which is determined by ali grammatical rules that apply to it. The sarne happens, according to Wittgenstein, with the basic concepts of Jogic:

There is nothing wrong with a symbol like 'qlX,' if it is a matter of explaíning simple logical relationships. This symbol is taken from the case where 'qi' designates a predicate and 'x' a variable noun. But as soon as we look at the real state of affairs, we see that the symbolism is at greJt d isadvantage compared with our real language. Jt is of course completely false to speak of 017<' subject-predícate form. ln reality there is not 011e1 but very many (sl'hr zohlrl'iche). If there were only one, then all nouns and ali adjectives would have to be inter­substitutable. For all intersubstitutable words belong to one class. But already ordinary language shows that this is not the case. On the face of it 1 may say: "The chair is brown" and "The chair's surface is brown." But if 1 replace 'brown' by 'heavy', 1 can utter (aussprechen) only the first proposition and not the second. This proves that also the word "brown" had two different meanings. (WVC, 46; slíghtly modified translation)

ln this passage, Wittgenstein seems to be restating the poínt made in MS 105 and SRLF that subject/predicate and function/argument don't capture the differences that we rnay need to emphasize (see Section 1.2 of this chapter). However, Wittgenstein's point now is somewhat different. The problem is not that these distinctions, these hypotheses concerning our modes of projection, may stop us from building the right phenomenological language; rather, they may stop us from seeing correctly the rules of 'grammar.'

That whicl1 allows us to discover the real kind to which a word belongs is, as seen above, the possibility of substitution. lt is not clear, nonethe­less, what according to Wittgenstein makes a substitution permissible. Very líkely it is the sense of the resulting substitution: if it makes sense, the substitution is allowed. But what determines whether it makes sense

r 1

Phenomenology, 'Grammar,' and the 'Limits ofSense' 53

or not? Bipolarity is not a criterion that is sufficient, for grammatical rules are, as nonsense, not bipolar (see below). The only answer that Wittgenstein can give here, 1 think, is "the grammatical rules of our Ianguage." As an account of sense and nonsense, it is circular. lt indi­cates, however, that the question of sense is the question of the rules of language that must be made explicit. We have been employing these rules (WVC, 91) implicitly and we can only express them, make them explicit, in our book of grammar:

ln grammar we cannot discover anything. There are no surprises. When formulating a rule, we always have the feeling: that is some­thing yóu have known all along. We can do only one thing, [i.e.] clearly articulate the rule that we have been applying unawares. (WVC, 77)

According to Wittgenstein (in 1930), the rules that we "unconsciously employed" are conditions of understanding the words. So rules of grammar are rules that we know, but we haven't explicitly formulated yet:

lf I understand what the specification of a length (Laengeangabe) means, I also know that, if a man is 1.6m tall he is not 2m tall. 1 know that a measurement determines only one value on a scale and not severa! values. If you ask me, How do 1 know that, I simply answer, Because 1 understand the sense of the statement. It is impos­sible to understand the sense of such a statement without knowing the rule. (1 can know the rule in its use without expressly formu­lating it). (WVC, 78, from 01.1930; slightly modified translation)

This answer indicates that the rules, supposedly, are a condition for under­standing the words that we use. One may use the rules implicitly without being aware of them. Of course, 'understanding' is playing a central and not so clear a role in this answer, for if one understands a word, then one must also supposedly understand the syntax of the word:

Thus if I understand the sense of a proposition at all, I must also understand the syntax of an expression occurring in it. (WVC, 78)

Sounderstanding is closely connected to the rules of syntax (grammar), since these rules are the conditions necessary for it to work. It is far from clear, nonetheless, how this implicit understanding takes place.

54 Wittgenstein's Philosophical Development

5.3 'Grammar' of phenomena and geometry

The major goal of the 'grammar' of phenomena (or phenomenology as 'grammar') is to avoid some confusion. When thinking of physical and phenomenological spaces, one has to have ln mind that the words used in descriptions of each don't mean the sarne. This is the case of the geometrical concepts of 'line,' 'curve,' 'circle,' etc., as already seen in Section 1.2.3, and the notion of 'time.' Phenomenological 'time,' even though infinite (MS 106, 29-35; PR §140), is nota time of past, future, and present as ordinarily understood. lt is a 'time' of possible infinite extension in which the only source of the past is memory. The phenom­enological 'past time' is simply the ordering of events given by memory: "My memories are ordered. Time is the way memories are órdered" (WVC, 98). Time in ordinary language means something different because the ordering is controlled (verified) in a different way, "for example, by means of statements made by me or another person11 (WVC, 98). This implies, for instance, that the idea of memory as a picture can be very misleading. A picture in physical time can be seen as fading because we have other means to verify the past and, thus, other means to compare (verify) the picture of the past event. ln a phenomenological system this is not the case, for there is no other standard of 'past time' except memory itself. Both modes of expression, the pictures in the phenomenological and in the physical systems, "are in order and are equally legitimate, but cannot get mixed together" (MS 108, 33; PR §49). We have to be careful, therefore, not to be misled by the analogy of memory and picture: we have to avoid, when using the analogy between memory and pictures, that "the analogy tyrannizes us" (MS 108, 33; PR §49).

The idea that a misleading analogy underlies a philosophical mistake is of enormous significance in Wittgenstein's later philosophy. lt will constitute one of the fundamental features of his genetic method. Here we can see one of its seeds. This method will be worked out some months later (at the end of 1930 and beginning of 1931), as will be explained in the next chapter.

Geometry is certainly part of the 'syntax' ('grammar') of ordinary language, for it gives the rules for the propositions about spatial objects: ''Geometry is the syntax of the propositions of the spatial objects" (MS 107, 218). Since geometry and its rules tell us what is allowed and what is not concerning our measurements, they work like rules of 'grammar,' which show which combination of words make sense and which don't. lf, for instance, a measurement brings results contrary to our expecta­tions based on geometry, they are wrong:

f

Phenomenology, 'Gmmmar,' and the 'Limits of Sense' 55

The proposition 'Vertically opposite angles are equal' means that if they tum out to be different when they are measured, 1 shall declare the measurement to be in errar ... ln this way, the proposition is a postulation concerning the way to describe the facts: and soa propo­sition of syntax. (MS 108, 25; PR §178)

Thus, geometry gives us the standards that we use in our measurements:

No measurement can tell us the value of n or between which values it is to be found, the number n is rather the standard by which we judge the quality of a measurement. The standard is given to us before we start measuring; this is why I cannot alter the measurement. Thus when we say, n has such and such a value, e.g. n = 3.1415265 ... , this cannot mean that we want to say anything about the actual measurements, but only that we are stipulating when a measure­ment procedure is to be counted as correct and when not. Thus the axioms of geometry have the character of stipulations concerning the language in which we want to describe spatial objects. They are rules of syntax. The rules of syntax are not about anything; they are laid down by us. (WVC, 62)57

But what is the status of the geometry of the visual field itself? A correct apprehension of the visual field seems to depend on the seeming/ being distinction, for we accept as standards, in our ordinary talk about objects of perception, some basic arithmetic and geometric rules that tell us what is the case in opposition to what seems to be the case. As a matter of fact, we give priority to the language of Euclidean space, and not to the isolated language of the visual space:

The essential thing is that we use two languages, a language of visual space and a language of Euclidean space, giving the language of Euclidean space priority. Language indicates this difference by using 'being' and 'appearing'. Thus we say of two stretches in visual space that they appear but are not equal. Or of a short are of a circle that it appears straight, although it is curved. And so on. (WVC, 59, from 12.1929)

The priority of Euclidean geometry, however, gives us two possibilities for understanding the geometry of the visual space. If the visual field is taken in isolation, geometry of the visual field forms a system that

56 Wíttgenstein's Philosophical Development

is different from Euclidean geometry - even though in some aspects analogous to it - and complete in itself:

The geometry of visual space is the syntax of the propositions that deal with those objects in visual space. (MS 107, 212-13)

ln talking about the visual field words like 'line' and 'circle' have a meaning that does not correspond to the meaning they have in Euclidean geometry, for, as we have seen, there is no standard of meas­urement given in the visual field. Taken in isolation, the visual space has in it essentially the "indeterminacy factor" (WVC, 55). This implies that the equality of segments in the visual field taken rn isolation has peculiar rules. We might accept, thus, that "the geometry of the visual field and Euclidean geometry have different multiplicities" (WVC, 60); for instance, '"a = b, b = c ~ a = c' may be true or not" for segments in the geometry of the visual field (WVC, 60). The geometry of the visual space is, thus, a compound of Euclidean geometry with the "indetermi­nacy-factor" (WVC, 55). Wittgenstein does not explain in details the "syntax of the indeterminacy factor."

1.5.4 Arithmetic as grammatical systems

By the sarne token that geornetry is the grarnrnar of space, arithmetic is "the grarnrnar of numbers" (MS 108, 116). Arithmetical equations are rules for the manipulation of signs (MS 108, 26) that also exclude cornbinations of signs when they are not rneaningful.

Wittgenstein's conception of mathernatics has two basic character­istics: the dissociation of a calculus and its application (its autonomy) and the plurality of complete systerns. Those characteristics are trernen­dously irnportant, for Wittgenstein will apply thern to 'gramrnar' in general later.58 ln what follows, l explain the first characteristic, in contrast to the philosophy of rnathernatics of the T, and then move to the second. What follows is not rneant to be a complete account of the philosophy of rnathernatics of the T, of course; it is rneant to show only its point.

The T presents two general clairns concerning rnathematics.

[A] "Mathernatics is a logical rnethod" (T 6.2).

This means that we can apply rnathernatical equations in arder to rnake inferences in our ordinary life (T 6.211). Those inferences have the forrn of rules of replacement (similar to tautological rules of

l Phenomenology, 'Grammar,' and the 'Limits of Sense' 57

replacement like 'De Morgan's laws'). What, however, makes this appli­cation possible? It is the second claim that answers this traditional phil­osophical concern.

(B] The essence of number, the fundamental nature of any number, is given by means of a general operation, "the more general form of transition from one proposition to another" (T 6.01). Numbers are the exponents of this operation.

What the general operation makes clear is that the central feature of numbers is t,he order that starts with a base x and develops as xO, xOO, xOOO ... in í:he application of the operation. This is the only essential characteristic of numbers given in this a priori definition (see T 6.02). It provides us with the exponents, which determine the definition of the cardinal numbers (O, O+l, 0+1+1 ... ). This means that the determination of a numerical quantity and the idea of classes are superfluous in the definition of numbers (6.031).

Since numbers are already given in the most general form of transi­tion of propositions, they have a status similar to logical connectives. They are not given inside elementary propositions, but result from the application of the operation. Because numbers origlnate with the sarne operation that generates complex propositions with the base of elemen­tary propositions, they can be applied to them. They allow, therefore, the transition from one proposition to the other, as logical connectives do. It is, therefore, the general form of propositions that guarantees that numbers can be applied (applying numbers means using them in propositions).

This view might be better understood in contrast with logicism. In his critique of formalism, Frege says that the formalists are wrong to say that mathematical rules serve as rules of a game. For him, they don't see that signs express thoughts. An explanation of the application of mathematics is, thus, something that formalists cannot explain, thinks Frege. lt is only the sense of mathematical propositions and, thus, the meaning of numbers, Frege argues, that guarantees thelr application: "Why can't we apply a position of chess pieces? Evidently because it does not express any thought [ ... ] Why can we apply arithmetic equa­tions? Just because they express thoughts [ ... ] It is only the applica­bility that raises arithmetic above the game to the status of a science" (Grundgesetze der Arithmetik II,§ 91).

Russell thinks simílarly. When dealing with the five primitive propo­sitions of Peano, he argues that it is unacceptable that there are "an

58 Wittgenstein's Philosophical Development

infinite number of interpretations" that satisfy them (Introduction to Mathematícal Philosophy, 7).59 We can, for example, take "O" to be 100, or take "number" as if it were even number; we could also take 11 0 11 as meaning 1 and "successor" as meaning "half" and form the sequence 1, 1 and a half, etc .. This means, thinks Russell, that Peano "does not succeed in indicating any constant meaning for O, number, and succes­sion" (The Principies ofMathematics, 126). If numbers don't have constant meaning, it is not possible "to apply them in the right way to common objects" (Introduction to Mathematical Philosophy, 8). For him, the remedy is to define 'O', 'number', 'successor', ,and so on. Like Frege, he takes the extensional way: the relation 1-1 (similarity) is given in any set of objects and is more primitive than the notion of order. '

The remedy of the T, however, is to make clear that numbers don't mean "something." Numbers are simply the result of an operation, the generated arder. This arder, for Wittgenstein, is not even axiomatic, as it is generated by an operation. If we have the system generated by the fundamental operation, we have the essence of number, which is simply the general progressive arder that defines O, O+ 1, etc. The Logicist problem concerning the determínation of the meaning of numbers is, for the T, grounded on the misapprehension of the absolutely general nature of numbers. ln the T, it is precisely this general nature that is the only ground for their application. Thus, Wittgenstein accepts an essen­tial similarity between logic and mathematics ("a method of logic"), but denies, at the sarne time, the relevance of the definitional project of logicism: numbers are simply the points of the progressive system. Note that this entails that there is one single number system - extensions would preserve the essence of number.

ln the PR, Wittgenstein breaks with the ideas of the T in two ways. First, the application of mathematics is not dependent on a general form of propositions. It is autonomous:

On the one hand it seems to me that you can develop arithmetic completely autonomously and its application takes care of ítself, since whenever it's applicable we may also apply ít. On the other hand a nebulous introduction of the concept of number by means of a general form of a proposition - such as I gave - is not needed. (PR §109; slightly modified translation; my emphasis)

Thls conception assumes, as in the T, that "propositions" of math­ematics are neither true nor false, but they are rather mies of substitu­tion that allow us to derive propositions. Autonomy means that what

Phenomenology, 'Grammar,' and the 'Limits of Sense' 59

gives meaning to signs in mathematics is the rules of the systems. This makes mathematics similar to a game (PR §109). One could also say that the possibility of application is a question of sense, which is given by the rules. Thus, Wittgenstein's early logicist worry concerning the possibility of the application of mathematics disappears: the general form is not needed. If the application takes care of itself, the Tractarian "nebulous" introduction of numbers is useless.

But what does it mean to "take care of itself"? It means that no a priori preparation or justification can guarantee that a future applica­tion will be successful. ln the end, mathematics is applied when it is applicable. The point is that it is contingent whether a mathematical system car{ be applied or not. Moreover, the justification or prepara­tion for the application of mathematics is not needed: "we understand and apply the propositions of arithmetic perfectly well without adding anything whatever to them" (PR, §107). We applied those propositions, for instance, without any definition of number being in place. Such a definition was, therefore, never needed for the application.

The second break with the philosophy of mathematics of the T is the idea that the system of numbers is itself composed of various complete systems.60 The system of the natural numbers, for instance, is different from the system of the integers, for the latter includes rules of syntax (the rules for the negative numbers) that are not included in the system of positive integers.

ln Wittgenstein's idea of multiple systems in mathematics we can find a similarity with Spengler's view, even if that is not necessarily the root of Wittgenstein's idea. ln MS 183, 16-17 (from 05.06.1930), Wittgenstein refers to Der Untergang des Abendlandes and claims that Spengler had "many really important thoughts" that he himself had already thought. ln the passage, Wittgenstein is referring to the idea that mathematics is not to be understood in the singular (as one body of knowledge built through historical accumulation), but should be taken as several inde­pendent developments through history. Spengler talks about the math­ematics of the Greek world, mathematics of the Arab world, etc., as independent closed developments. However, Wittgenstein's interest in multiple systems was not properly historical. While Spengler talks about "a multiplicity of independent, closed in themselves developments" in Der Untergang des Abendlandes, Wittgenstein talks about "multiple closed systems" (MS 183, 16-17).61 ln Wittgenstein's conception, "closed" means "complete" and expresses the idea that a system does not need to be complemented. No improvement is required for a system to stand on its own. This is implied by the idea that the autonomous rules of a system

60 Wittgenstein's Philosophical Development

are what characterize it, and not something independently given like the Fregean notion of 'meaning.' The meaning of a mathematical sign is its position inside the system it belongs to, i.e., the rules that apply to it exhaust its meaning:

The system of rules determining a calculus thereby determines the 'meaning' of its signs too. Put more strictly: The form and the rules of syntax are equivalent. So if 1 change the rules - seemingly supple­ment them, say- then 1 change the form, the meaning. (PR §152)

The meaning of mathematical signs is not, thus, an object of any kind (platonic or mental); it is equivalent to the rules of the system. Thus, the system of naturals is not completed by the system of negative numbers. According to Wittgenstein, "there are no gaps in mathematics" (PR §158), i.e., each system is an independent system of rules (complete in itself). When we think that new rules were added to a given system, we have not, in fact, added rules to the old system, but invented a new one. So there is not one single system in mathematics, but various related systems.

1.5.5 Arbitrariness of 'grammar' and the "essence of the world"

The most general characterization of 'grammar' in PR is that its rules cannot be justified. According to Wittgenstein's vocabulary in the BT, this means that its rules are arbitrary. Note that 'arbitrary' means exactly that 'it cannot be justified.' The point of the arguments for the arbitrariness of 'grammar1 seems to be the elimination of the ambig­

uous status of the early phenomenological project and the grammatical investigation. Since what in the T was called the 'application of logic' had shown that the a priori structure of logic of the T does not fit its application, one could think that truths concerning the application (or worse: empirical truths) could justify logic or 'grammar' (the struc­ture of language). Wittgenstein's arguments for the "arbitrariness of grammar" are intended to block this move. They are intended to show that questions of sense (that concern what is necessary) are completely independent of questions of truth (contingent facts).

Wittgenstein calls the rules of 'grammar' conventions in 1929-30. 1 think that this emphasizes that rules of language are not to be reduced to rules given a priori, as in the T. This is a consequence of the comprehen­sive notion of 'grammar.' However, this should not imply that the rules of 'grammar' aren't rules that express necessary (interna! or formal) relations among propositions. After all, as rules, they determine what

l Phenomenology, 'Grammar,' and the 'Limits of Sense' 61

follows frorn what. For Wittgenstein in PR, the fact that they cannot be justified by ernpirical propositions shows that they are necessary: if they were ernpirically justifiable, they would depend on contingent facts. Therefore, they could not express necessary relations (what follows frorn what).

Wittgenstein offers the following argurnent for the arbitrariness of 'grarnrnar':

1 don't call the rules for representing sornething a convention that can be justified by propositions, propositions that describe what is represented and show that a representation is adequate. The conven­tions of grarnrnar cannot be justified by a description of what is repre­sented. Any description of that kind already presupposes the rules of grarnrnar. That is, what counts as nonsense within the grarnrnar that is to be justified cannot count as rnaking sense in the grarnrnar of the justifying propositions, and so on. (MS 108, 104; PR §7; translated in BT, 238; rny ernphasis)

The rules of representation are conventions that cannot be justified. Presurnably, if they could, they would be justified by rneans of true propositions frorn which they would supposedly follow. The justifica­tion would, then, be grounded on a description of facts in the world. The justifying propositions, however, argues Wittgenstein, would already presuppose the conventions. Wittgenstein clairns, then, that a justifica­tion cannot be given for the rules of 'gramrnar' because it would already presuppose the rules that it should justify. lf we suppose that proposi­tions that belong to a different 'grarnrnar' justify the current 'grarnrnar,' we cannot forget that the lirnits of sense in the 'grarnrnar' that should be justified are already in place ("what counts as nonsense within the gramrnar11

). Thus, if the propositions that belong to a presumably different 'grarnrnar' have the sarne lirnits of sense of the current one, i.e. the sarne rules, it is, actually, the sarne 'grarnrnar.1 ln this case, the propositions already presuppose the rules to be justified. If those propo­sitions that belong to a really different 'grarnrnar,' however, don't have the rules of the current 'grarnrnar,1 what is described in the different 'grarnrnar' does not count as rneaningful in the current 'grarnrnar' (the 'grarnrnar' to· be justified).62 As nonsense, the "propositions11 cannot justify 'grarnrnar' either. This argurnent works, then, only if we take 'gramrnar' as the systern of rules that gives the lirnits of sense.

ln another passage Wittgenstein exernplifies the argurnent given above with the tentative justification of a 'grarnrnar' of colors. Again,

62 Wittgenstein 1s Philosophícal Deve/opment

the point is whether the description of facts can make rules of 'grammar' necessary. However, in this passage Wittgenstein introduces the idea of the 'purpose of grammatical conventions.' He does so in arder to get rid of what might appear as a paradoxical consequence of the view that the rules of 'grammar' are conventions: if they are conventions, then anything is possible. This is not the case, argues Wittgenstein, because the conventions (rules of 'grammar') have certain purposes. However, it is important to understand, according to Wittgenstein, that we can create new rules of 'grammar' without having an application or a purpose in mind. Moreover, a purpo~e does not imply one specific set of rules (the purpose does not oblige us to create a specific set of rules). Thus, thinks Wittgenstein, even if we have the purpose óf certain rules for a specific form (for instance, colar), the rules remain arbitrary, i.e., they stíll are not empirically justified. Here is the argument:

If 1 could describe the purpose of grammatical conventions by saying that 1 had to create them because, for instance, colors have certain properties, that could make these conventions unnecessary, because then 1 could say precisely what it was the conventions were excluding. Conversely, if the conventíons were necessary, i.e. if certain combi­nations of words had to be excluded as being nonsensical, then for that very reason 1 couldn't name a síngle feature of the colors that would make the conventions necessary, for then it would be conceiv­able that the colors may not have the feature, and that could only be expressed by contravening the conventions. (MS 108, 98 from 02.28.1930; PR §4; translated in BT 238; Wittgenstein's emphasis)

Can conventions (grammatical rules) be made necessary by real prop­erties of the objects that the rules regulate? No, argues Wittgenstein, because the very supposition that conventions were empirically deter­mined (if we had to create them), i.e., that the truth of certain descrip­tions implied them, would make these conventions unnecessary (not need.ed). These conventions would be unnecessary because one could meaningfully describe what the very conventions (the rules of grammar) would have to exclude. On the other hand, ifthe conventions do exclude "certain combinations of words ... as nonsensical," then one cannot give any property that makes the conventions necessary. Why not? Because, argues Wittgenstein, this assumes as conceivable that the colors don't have the property that is supposed to be necessary, and this contradicts the present convention (the 'grammar' that was supposed to be justified). The contingency of the property must be assumed, presumably, because

Phenome110Iogy, 'Grammar/ and the 'Limits of Sense' 63

Wittgenstein thinks that the present conventions (rules of 'grammar') determine everything that is necessary concerning colors. Thus, the attri­bution of the necessary property is either already necessary or nonsense according to the present 'grammar'. 'Grammar,' therefore, determines what is 'necessary,' but nothing makes 'grammar' itself necessary.

However, note that those arguments don't show that there might not be various presentations of 'grammar.' What we wish of a presen­tation of 'grammar' is, for instance, that it expresses the right multi­plicity of the phenomena (WLC30-32: 2, 5). This multiplicity is already implicit in language. This is why the presentation of the 'grammar' of colors by means of the calor octahedron is better than its presenta­tion by meâns of the colar circle or the calor scale (see PR §§218-24). Moreover, the arguments above don't show that language and reality need not be connected (WLC30-32, 10 and 12). As we have seen, even after giving up the phenomenological language, Wittgenstein holds to the idea that phenomena verify sentences and connect language with reality (see Section 1.5.1). It is the system of coordinates that "shows the connection between language and reality" (PR §46). It does so because it indicates the states of affairs that verify our propositions. Note that a proposition is similar to a hypothesis in that it generates an expecta­tion. The expectatíon will be confirmed by phenomena (Section 1.5.1). Verification requirements related to phenomena wíll be present even in the later conception of the autonomy of 'grammar.'63

The arguments for the arbitrariness of 'grammar' have an obvious transcendental flavor: as a condition of possibility of sense, 'grammar' cannot be justifíed. This means that the 'method of philosophy' used in the phenomenological investigation is still the method used in the early 'phenomenological-grammatical' investigation (see PR §§ 60 and 206). It is important to notice that the transcendental arguments for the arbitrariness of 'grammar' will still be used in the BT, some years later. We will see why in chapter 3.

Arbitrariness of 'grammar' means solely that "grammar cannot be justified by propositions.11 Its point, as we have seen, is to preserve the necessary status of rules of inference in general. Thus, it seems natural to think of the rules of logic in the Tas arbitrary in the sarne sense: they cannot be justified by anything, for they are completely independent from anything that happens in the world. The arbitrariness of the rules of logic in the T is, however, obvious, for they are given a priori. It is because rules related to phenomena (of things that exist) are intro­duced in 'grammar' that arguments for the arbitrariness of 'grammar' are needed in 1930 (in PR).

64 Wittgensteín's Phílosophical Development

Note that in the T tautoiogies indicate or show (anzeigen) the essence of ianguage and the worid:

[ ... ] lt is clear that something about the world must be indicated (anzeigen) by the fact that certain combinations of symbois - whose essence invoives the possession of a determinate character - are tautoiogies. (T 6.124)

This view is, as one wouid expect, broadened in 1929-30 according to the comprehensive idea of necessary propositions. Philosophy, Wittgenstein argues in 1930, "if it was to say anything, would have to describe the essence of the world" (PR §54). Philosophy éannot describe it because it is not a thing to be described: "language can only say those things that we can also imagine otherwise" (PR §54). But there is another way to grasp the essence of the world:

But the essence of ianguage is a model (Bíld) of the essence of the world; and philosophy as custodian of grammar can in fact grasp the essence of the world, only not in propositions of Janguage, but in rules for this language which exclude nonsensical combinations of signs (PR §54; slíghtly modified translation).

Thus, tautoiogies are only a part of what indicates the essence of the world, which philosophers wrongly try to describe. lt is rather the rules of the comprehensive 'grammar' that fulfill this task. As in the T, it is a description of the internai properties of language that indicate the essence of the world (concerning the two meanings of 'description,' see Section 1.2). ln the T, essence is expressed in Iogicai notation; in the PR, it is expressed in mies of 'grammar.' Metaphysicians are wrong in the first place, according to the T and PR, because they try to describe the essence, which is given in mies of language (internai properties), as if they were describing externai properties of objects. From such a misstep only more confusion can aríse.

' 1

1 2 Russell's Causal Theory of Meaning, Rule-Following, the Calculus Conception, and the Invention of the Genetic Method

2.1 Causal theories of meaning

The major goal of this chapter is to explain the nature of Wittgenstein's genetic method, and how and why it was created around 1930-1 in the context of rule-following worries. This will also show how the calculus conception is invented. All these significant moves can only be understood with the background of the causal theory of meaning.1

Russell published his theory in The Analysis of Mind (AM) in 1921, two years earlier than another then-famous book on the subject: Ogden and Richards' The Meaning of Meaning (MoM). 2 I briefly present both in what follows and then explain the nature of Wittgenstein's struggle with them and how it led him to the calculus conception and to the genetic method. I finish the chapter with a note on Wittgenstein's first project with Waismann.

2.1.1 Russell's causal theory of meaning

Russell claims in the preface of AM that he wants to make compatible a materialist tendency in psychology with an immaterialist tendency in physics. Psychology, according to Russell, was gradually taking solid bodies and their behavior as the fundamental object of study (behav­iorism), while physics was in many ways dissolving the solidity of matter (relativity and Quantum physics). Russell's solution is neutral monism: matter and mind are constructed out of the sarne stuff. 3

The only kind of 11stuff11 that exists, in Russell's monistic view in AM, is sensations (AM, 69 and 121). Sensations are 'neutral' because grasped

65

66 Wittgenstein's Philosophical Development

in one way they obey physical causal laws and grasped in a different way they obey psychological causal laws (in this case they are images, i.e., faded sensations). Meaning, as a relation between a word and an object (the meaning of the word), is explained causally: "The relatlon of a word to its meaning is of the nature of a causal law governing our use of the word and our actions when we hear it used" (AM, 198). Roughly put, when I call "John! 11 John comes to me and when 1 see him, for instance, the word 'John' comes to my mind. When the objects are not present, Russell argues, images replace them and so they cause and are caused by words (someone tells me a story and 1 imagine a situation, for instance). Thus, the 'stuff' that constitutes the world appears as sensa­tion and as image. But images, according to Russell, cán be partially eliminated in the process of understanding when the speaker is suffi­ciently well trained in the use of words: "the more familiar we are with words, the more our 'thinking' goes on in words instead of images" (AM, 206).

Russell's causal account of language parallels his causal account of "mental" activities. Russell's major claim concerning desire in AM is that it is a "fiction for describing shortly certain laws of behavior ... the thing which will bring a restless condition to an end is said to be what is desired" (AM, 32). Combined with this behavioristic claim, Russell introduces Freud's notion of unconscious in behavioristic terms: " ... an unconscious desire is merely a causal law of our behavior, namely, that we remain restlessly active until a certain state of affairs is real­ized, when we achieve temporary equilibrium" (AM, 38). For Russell, the unconscious desire is the "natural primitive form of desire, 11 since "wishes are all unconscious and only become known when they are actively noticed11 (AM, 38-9). According to Russell, all desires are first 'unconscious' (as seems to be the case with other animais) and they become 'conscious' only because of habitual observation of what brings rest to a restless condition.

With these new theoretical tools, Russell thinks that he can explain that desire is not "something actually existing in our minds'' (AM, 60), but "a causal law of our actions." Desire is to be explained in terms of an impulse instead of an attraction to something in the future (AM, 66). Some disagreeable sensation (discomfort) starts a 'behavior-cycle' (a series of actions that tend to stop the discomfort) that leads to its cessa­tion (this we call 'pleasure') (AM, 68). Russell says:

If our theory of desire is correct, a belief as to its purpose may very well be erroneous, since only experience can show what causes a discomfort

1 t f f

The Invention of the Genetic Method 67

to cease. When the experience needed is common and simple, as in the case of hunger, a mistake is not very probable. (AM, 72)

What tells one whether the belief of desire ('conscious' desire) is fulfilled is a feeling of satisfaction that takes place with the fulfillment of the desire. Similarly to desire, 'mental attitudes' (memory, expectation and assertion, etc.) are also explained by Russell in AM. Ali the attitudes are, according to Russell, kinds of beliefs. A belief has a content (images and/or words, but sometimes also sensations) and an "actual experi­enced feeling" (AM, 233), which is the attitude related to the content. Russell uses the example of the visual image of a breakfast table to show that the difference among the attitudes does not consist in a difference of content but of feeling. One can remember that image, doubt that it was seen in the morning, desire it or expect it for the next day, and so on. Russell also suggests that "there may be other belief-feelings, for example in disjunction and implication; also a dísbelief-feeling" (AM, 250). Russell's analysis shows, according to him, that a given content (called 'proposition' or 'image-proposition') can always be related to different attitudes (feelings of hope, fear, doubt, expectation, remem­bering and asserting) (AM, 243). What is particular to this theory, then, is that caused feelings indícate whether a belief is a memory belief or an expectation of a future event. This will be an ímportant target of Wittgenstein's criticisms at the begínning · of 1930 and wíll play an important role in Wittgensteín's new philosophy (Section 2.2 of this chapter).

2.1.2 Ogden and Richards on meaning

Wittgenstein not only knew Ogden, who translated the T into English, but certainly had a copy of Ogden and Richards' MoM. He disliked it very much, as can be seen from the following passage in a letter to Russell from 1923: "A short time ago 1 received The Meaning ofMeaning. Doubtless it has been sent to you too. Is it not a miserable book?!" (CL, 184).4

Ogden & Richards change the vocabulary used by Russell, which is rather puzzling since they call thought 'reference'.5 Meaning should be entirely explained by causal relations: "Between a thought anda symbol causal relations hold. When we speak, the symbolism we employ is caused partly by the reference ['thought' and 'reference' are synonyms] we are making and partly by social and psychologícal factors" (MoM, 10). Ogden & Richards don't define 'symbol' explicitly, but they seem to think that a symbol is any sign involved in causal relations and used in language. The social and psychological factors mentíoned above are

68 Wittgenstein's Philosophical Development

"the purpose for which we are making the reference, the proposed effect of our symbols on other persons, and our own attitude1' (MoM, 10).

Language is seen, thus, as "an instrument for the promotion of purposes" (MoM, 16). Social and psychological factors give language its purpose and they constitute the context of the use of words. Words cause certain thoughts (references) and behaviors, and certain behaviors cause the use of certain words: "When we hear what is said, the symbols both cause us to perform an act of reference and to assume an attitude which will, according to circumstances, be more or less similar to the act and the attitude of the speaker" (MoM, 10.,11).

Ogden & Richards' major goal seems to be the elimination of the Russellian use of images and introspection in the explanation of the nature of meaning: "We should develop our theory of signs from obser­vations of other people, and only admit evidence drawn from intro­spection when we know how to appraise it" (MoM, 19). They give three reasons for the elimination of Russellian images from an account of meaning. First, they argue, it is not clear whether images occur in all minds. Second, in many cases introspection doesn't show the presence of images. Third, mental images are not needed to explain meaning (MoM, 60). Their account should be a proof of this third and, perhaps, only relevant reason.

Thinking, in their view, is interpreting signs. So the word 'interpre­tation' designates the most general "thinking" activity. They explain 'interpretation' in terms of recurrence of contexts and causal relations:

... when a context has affected us in the past the recurrence of merely a part of the context will cause us to react in the way in which we reacted before. A sign [presumably, symbol] is always a stimulus similar to some part of an original stimulus and sufficient to call up the engram [11residual traces"] formed by that stimulus. (MoM, 53)

One of Ogden & Richards' examples is the dog that comes to the dining roam when it hears the bell. According to them, it interprets the sound of the bell as the possibility of food. So /1 thinking" is the habit of expecting contexts that are similar to those that have already happened. The dog interpreted the gong because of his past experience with bells and dinners. ln the sarne way, according to Ogden & Richards, we asso­ciate words with contexts and interpret them, i.e., words create (cause) expectations and direct us to similar facts that have already happened. To think of 'A' is to be in a certain relation to A or to what is causally related to A.

l ! ~ 1

The Invention of the Genetíc Method 69

2.2 Wittgenstein's critique of the causal theory of meaning

The central role that psychology should play in the development of philosophical views, especially as a tool for the theory of meaning, is an ímportant assumption shared by Russell and Ogden & Richards. They believed that the "science of symbolism" could considerably improve with the new developments of psychology, which they mainly attrib­uted to behaviorism and to a lesser degree to psychoanalysis. As Ogden & Richards say: "The analysis of the process of communication is partly psychological, and psychology has now reached a stage at which this part may be successfully undertaken" (MoM, 8). Wittgenstein explic­itly denied that an explanation based on psychology, like the one from Ogden and Richards, could further explain anything he had said in the T. Wittgenstein thought that the problems relevant for him, problems of logic, were very different from the problems of psychology. This can be seen in a letter that Wittgenstein wrote to Ogden in 1923:

"The Meaning of Meaning" reached me a few days ago ... i have not yet been able to read your book thoroughly. 1 have however read in it and 1 think 1 ought to confess to you frankly that 1 believe you have not quite caught the problems which for instance - I was at in my book (whether or not 1 have given the correct solution). (Letters from Ludwig Wíttgenstein to C. K. Ogden, 69)

At this time, Wittgenstein saw the causal investigation of meaning as completely irrelevant for the topics he had discussed in the T: the deter­mination of the limits of sense and the correct characterization of the nature of logic and its "propositions.11 How words get their meaning was not part of the investigation of the T, for logic simply presupposes "that names have meaning and elementary propositions sense" (T 6.124). Even more important, of course, is that "logic precedes every experience that something is so" (T 5.552); after all, "logic must take care of itself" (5.473). For Wittgenstein at the time of the T, psychology was "no more closely related to philosophy than any other natural science" (4.1121) and an investigation in the style of the causal theories of meaning could only be an example of an /(unessential psychological investiga­tion" (4.1121).

After Wittgenstein's return to philosophy in 1929, his views on the relevance of the investigation of psychological concepts for questions of meaning in the style of Russell and Ogden and Richards do not change.

70 Wittgenstein's Philosophical Development

ln MS 107, 235 (01.14.1930)1 he writes:

There is a kind of phílosophy - we could call it psychological philos­ophy, but l haven't yet found the exact good na me for ít - wh ich talks about assocíations and the simultaneous or more or less simultaneous appearance of events A, B and C in which the similar elements of two events have as a consequence that the total comes in when a part appears in front of our eyes. A typical philosophical dead-end street (Sackgasse). The mixture of aspired precísion and real irrelevance.

A 11philosophical dead-end street 11 is,· according to Wittgenstein, the "danger of reflecting on something that is not one's business (concern)" (MS 108, 64 from 01.03.1930).6 The "aspired precision11 of Russell and Ogden & Richards is of course related to their belief in extending the use of scientific exactness to philosophical matters. Wittgenstein's view at this time is that the relevant investigation stops where discussions of psychology begin.

However, Wittgenstein will immediately see that he himself needed to make clear some psychological concepts in arder to save his pictorial conception of language. ln what follows we will see why. His worries about the causal theory of meaning increase in 1930. He several times mentions and discusses this theory in manuscripts and classes at Cambridge.7 Less than a month after the remark about the 'írrelevance' of the causal theory of meaning quoted above, Wíttgenstein says the following about Russell's theory:

A wrong conception of the way language functíons destroys, of course, the whole of logíc and everything that goes with it, and doesn't just create some merely local disturbance.

If you exclude the element of intention from language, its whole function then collapses. (MS 107, 289 from 02.08.1930; translated in PR §20)

lt is not immediately clear why Wittgenstein thinks that the causal theory of meaning would be so destructive. Why isn't Russell's account of meaning something outsíde the scope of Wittgenstein's interests or simply irrelevant? The problem wíth Russell's theory according to Wittgenstein is, 1 think, that the description of language based on a causal mechanism subordinates questions of logic and language to this mechanical explanation (one could perhaps say that it violates the arbi­trariness of 'grammar'). So Russell's "conception of the way language

The Invention of the Genetic Method 71

functions" is incompatible with Wittgenstein's own general views about the way language works; in fact, it is incompatible with ít and with any reasonable account of language, according to Wittgenstein, for with it the "whole logic and everything that goes with it" is at risk.

The incompatibílity can be expressed in, at least, three points:

(a) The causal theory of meaning makes the sense of a sentence depend on its own truth and therefore, it seems to transform the logic of language ('grammar') into an empirical discipline. ln the causal theory of meaning, the sense is not independent of the truth of a proposition, for it is a future feeling that determines what is expecteêl. This also implies that one could expect a nonsense, because the criterion of verification is merely the feeling of expect­edness. Supposedly, one could have expected that "John passed the Abracadabra" if the feeling of expectedness appeared. Moreover, the theory seems to imply the nonsensical consequence that we never know what we expect or think: 11We can not be sure whether we really expected this" (MS 107, 247). If we cannot be sure about our own expectations, we are also unsure about understanding our thoughts (point c below).

(b) The function of language, according to the causal theory, is to cause behavior, while the function of language according to the picture conception is to describe reality. Thus, the central aspect of the picture conception is that we compare proposition and reality directly. The comparison of thoughts or propositions and reality is not an essential feature of the causal theory of meaning. It makes the comparison secondary, for it is a feeling that guarantees the corre­spondence of proposition and reality, and not the comparison itself. This contradicts the T and its reformulation ln 1929-30. After 1929, as seen in Chapter 1, how one compares proposition and reality (veri­ficationism) became also an issue. At the end of 1929 Wittgenstein defended that propositions work like hypotheses that generate expec­tations to be verified by phenomena. This makes it clear that the essence of his views on the comparability of proposition and reality have not changed. 8 His Hauptgedanke is "that one compares the prop­osition with reality" (MS 107, 155; from 1929). ln 1930, Wittgenstein incorporates to his views various kinds of sentences (arder, descrip­tion, question, propositions combined with attitudes):

The sarne requirements apply to a description as to a command or a prescription. A description is verified or falsified by comparison with

72 Wittgenstein's Philosophical Development

reality, with which it may correspond or not, and so be true or false. This is true of propositions generally. (WLC30-32, 2)

ln the sarne key, he says:

The command must be a case (Art) of the comparison of a proposi­tion with reality. (MS 108, 190 from 06.20.1930)

The relation to reality is similar in different kinds of sentences, for a fact correlates to the descriptive sentence when it is true and a fact decides whether one is carrying out an order. The descriptive (representational) function of language is, thus, a feature of ali kinds of seri.tences.

(c) The causal theory of meaning is incompatible with Wittgenstein's non-causal notion of understanding. For Wittgenstein, on the one hand, understanding is translating (interpreting) mies of projec­tion: "To understand a thought means to be able to translate it according to a general rule. For example, playing a piano from a score" (WCL30-32, 44). This fits pretty well the Tractarian notion of projection (T 4.014-1). One should keep in mind that definitions are rules of translation (T 3.343). Those rules, presumably, already given in the use of language, could help us to reach elementary propositions in the T. After 1929, they should help us to reach phenomenological descriptions. According to the causal theory of meaning, on the other hand, understanding is, of course, causally determined. lt is clear that what Ogden & Richards, for instance, call 'interpretation' is too weak to deal with interpretation or trans­lation accordíng to general rules. ln the example of the dog given above, for instance, nothing could count as incorrect. The notion of understanding in the causal account, thus, does not give the right multiplicity to our notion of understanding, which is bipolar. This is why Wittgenstein says: "But the score does not cause us to play as we do; if it did there would be no right and wrong way of playing" (WLC30-32, 44).

lt is important to notice that understanding different kinds of sentences is similar: understanding a descríptive sentence (proposition) is knowing how it is verified and made true; understanding a command is knowing how to act to fulfill the command (to make it 'troe'). This means that we know what fulfills an order or an expectation before­hand, as we know beforehand what makes a sentence true. ln other

fl

2

1 d V

I

e s a t f

The lnvention of the Genetic Method 73

words, what actually guarantees the comparison of proposition and reality is that the sense is independent of the actual truth or falsity of a proposition (see T 4.024). This is valid for all kinds of sentences.9

I introduce now Wittgenstein's critique of the causal theory of meaning. This critique is utterly important, for arguments similar to the ones used against the causal theory will be also applicable to Wittgenstein's own conception of language (Section 2.3).

2.2.1 Wittgenstein's arguments against the causal theory

The most obvious argument against the causal theory of meaning is directed at Russell's version. The reason for this is Russell's peculiar views concérning the nature of desire and the verification of beliefs. As we saw above, for Russell that which causes the end of a behav­ioral desire-cycle is what is desired. Wíttgenstein objects to this in the following way:

(lf I wanted to eat an apple, and someone punched me in the stomach, taking away my appetite, then it was this punch that I originally wanted). (MS 107, 290; PR §22)

This reduces Russell's view to absurdity. It shows that what Russell calls 'desire' is simply a construct that has little relation with our under­standing of what desire is. If one desires something spedfíc like an apple, a punch in the stomach won't do, even if the "behavior cycle" of hunger stops. Russell's theory, thus, violates the requirement of multi­plicity: "Language must have the sarne multiplicity as a contrai panel that sets off the actions corresponding to its sentences" (MS 107, 231; PR §13). That is, Russell's theory presents our language as if it had fewer options than it actually has.10

But Wittgenstein thought that what was wrong with the causal theory of meaníng was not simply its bad explanation of desire. What was deeply wrong, according to Wittgenstein, was the very idea that a feeling of some kind was a criterion for the verification of beliefs in general. It is very useful, therefore, to go into more details of Russell's views on the matter.

ln AM, p. 269, Russell explains what counts as a verification of expectations:

When an image accompanied by an expectation-belief is thus succeeded by a sensation, which is the 'meaning' of the image, we say that the expectation-belief has been verified.

74 Wittgenstein's Philosophical Development

1 expect, say, that John will play the piano. This expectation is a cornposition of irnages of John and the piano (proposítion) and a feeling of expectation. According to Russell, the expectation is verified if 1 see John (the 'rneaning' of the irnage of John) playing the piano (the 'rneaning' of the irnage of the piano). But this direct cornparison prornpts a problern for Russell. He asks:

How do we know that the sensation resernbles the previous irnage? Does the irnage persist in presence of the sensation, so that we can compare the two? And even if some irnage does persist, how do we know that it is the previous irnage unchanged? lt does not seern as if this line of inquiry offered rnuch hope of successful issue. (AM, 270)

ln Russell's view, we have to compare irnages in our expectation with sensations that correspond to thern. The first problern with the direct cornparison is with the irnage that we want to compare: do we have the original one, which is present in the expectation, at the time of the sensation? Second, even if it is the sarne, how could we determine whether the present irnage has not changed? Given these difficulties, Russell's view is that a criterion associated with rnernory-beliefs has to be introduced. His suggestion follows irnrnediately:

lt is better, 1 think, to take a more externai and causal view of the rela­tion of expectation to expected occurrence. lf the occurrence, when it comes, gives us the feeling of expectedness, and if the expectation, beforehand, enabled us to act in a way which proves appropriate to the occurrence, that rnust be held to constitute the rnaxirnurn of verification. We have first an expectation, and then a sensation with the feeling of expectedness related to rnernory of the expectation. This whole experience, when it occurs rnay be defined as verifica­tion, and as constituting the truth of the expectation. Appropriate action, during the period of expectation rnay be regarded as addi­tional verification, but is not essential.

I think that all verification is ultirnately of the above sort. (AM, 270; rny ernphasis)

Russell's solution is, then, to observe "a fee!ing of expectedness related to rnernory of the expectation" as a criterion for the fulfillrnent of the expectation or "truth of the expectation." There is, then, a feeling of expectation related to the content of expectation (which distinguishes

r 1

r

The Invention of the Genetic Method 75

an expectation from a mernory, for instance) and a feeling of expect­edness, that is, the feeling of a satisfied expectation. The feeling of expectedness relates the rnernory-belief of the expectation to the orig­inal expectation. So, supposedly, we know by rneans of the feeling of expectedness that the present sensation relates to the previous irnage involved in the expectation. This view seems to apply to other proposi­tional attitudes, each of which rnust be verified by a particular feeling (of expectedness, fulfillment, etc.).

lt rnust be the above passage of Russell's book that Wittgenstein has in rnind when he criticizes Russell's theory in MS 107:

I.e. for m'é there are only two things involved in the fact that a thought is true, narnely, the thought and the fact; for Russell, on the contrary, three, namely, thought, fact and a third event that, if it occurs, is the recognition. This third event, as the satisfaction of hunger (the other two being hungry and eating a specific food), could be, for instance, a feeling of pleasure. Here it is cornpletely irrelevant how/ as what we describe this third event; for the essence of the theory this is irrel­evant. (MS 107, 290; in PR §21; rnodified translation)

So according to Wittgenstein the essence of Russell's theory is the incor­poration of a third elernent to explain the relation of a sentence ('word or irnage proposition' for Russell) to the recognition of what rnakes it true or what fulfills it (in the case of propositional attitudes). ln Russell's theory, as quoted above, the verification does not take place in the direct cornpar­ison of proposition and reality, but when a feeling takes place. This third elernent, the feeling, characterizes an 'external relation' (Russell says 'external and causal view'). ln Russell's theory, recognítion is sirnply a causal effect. According to Wittgenstein, for Russell the recognition of the fulfillrnent of an expectation is like recognition of a headache when "l expecta headache if l arn hit on the head" (MS 108, 59; PR §11).

If Russell's theory were correct, therefore, it would be a future feeling that would tell what one had really expected. By the sarne token, ín such a view, if one has a wish and gíves a corresponding order to sorneone else, then it rnust be a corresponding future feeling (fulfillrnent, happi­ness, or a feeling of 'wishiness') that the order was carried out, the wish or expectation fulfilled, that indicates whether the arder was really carried out. ln Russell's view, the content of the expectation, order, etc., thus loses what is spedfic (definite) in it: "Russell treats wish (expecta­tion) and hunger as if they were on the sarne level. But several things will satisfy rny hunger; rny wish (expectation) can only be fulfilled by

76 Wittgenstein's Philosophical Development

something definite" (WLC30-32, 9, Lent Term 1930). Russell's theory, therefore, gives Ianguage the wrong multiplicity: expressions of expec­tations are fulfilled by specific state of affairs, and not by severa], like various kínds of food and hunger.

Thís Jack of definiteness leads to two problems. First, when one is hungry there is, of course, a variety of food that could stop the hunger. One does not know ln advancewhat will satisfy the hunger, but one knows in advance what will fulfill one's expectation. Russell's view contradicts thís obvíous fact. For one needs to wait for the feelíng to know what the (prior) expectation was. It contradicts, thus, the obvious multipllcity implicit ín our language, for we can express our specific wíshes and see they fulfilled by the situation that we wished for. The second problem related to the Jack of specificity of the content of attitudes is that what­ever brings the feeling must be taken as what fulfills, say, the command: "I believe Russell's theory amounts to the following: lf 1 give someone a command and 1 am happy with what he then does, then he has carried out my order" (MS 107, 290 from 01.1930).11 Whatever brings about the feeling has to be taken as what corresponds to my command or my expectation. The problem is, of course, that a feeling of expectedness (say, a kind of happiness) may be caused by something other than what we usually understand as the fulfillment of an order or expectation. 1 may feel the feeling x connected with the fulfillment of the expectation "l expect Brazil to wín the 2014 World Cup," but it may be the effect of a memory image from my Iast vacation.12

Russell's theory also implies an infínite regress. lf we suppose that a feelíng is the criterion of recognition, we stíll would need something else to play the role of a criteríon for recognízing that feeling. Besides the feeling of expectedness (say, the feeling of a kind of happiness), it seems that I also would need a criterion to recognize that this feeling is the right feeling (PR §22). Wittgenstein in his Cambridge Lectures makes this explicit at the beginning of 1930:13

On Russell's view you need a tertium quid besides the expectation and the fact fulfillíng it; so if I expect x and x happens something else is needed, e.g. something that happens in my head, to link expectation and fulfillment. But how do 1 know that it is the right something? Do 1, on the sarne principie, need a fourth something? lf so, we have an infinite regress, and 1 can never know that my expec­tation has been fulfílled. (We can always ask for a further description of any description given of meaning or fulfillment; which produces an infinite regress.) (WLC30-32 p. 9; from 02.13.1930)

r 1

The Invention of the Genetic Method 77

Jf 1 need a feeling to tell me whether or not an expectation is fulfilled, 1 also need a feeling (or a different criterion to play the sarne role) to tell me whether the feeling of, say, expectedness is what matches the fact that was supposedly expected. Thus, Russell's view leads to an infinite regress and it implies, therefore, that "1 can never know that my expec­tation has been fulfilled. 11

2.2.2 The Tractatus and the absent psychological intentional element

An issue that might seem puzzling in the relationship between Wittgenstein's views in 1930 and Russell's causal theory of meaning concerns "the way that language functions": "If you exclude the element of intention its [language's] whole function then collapses" (MS 107, 289 - see the whole quotation above). The function of language is, presumably, to describe truly or falsely the world. The 'element of intention' in the T is absent; there, the word 'intention' does not even appear, for the only relevant aspect of intention according to the T is the method of projection. All psychological aspects of intention are irrelevant. ln 1930 the worries with Russell's causal theory of meaning, however, will bring Wittgenstein to think that intention and notions connected to it, like understanding, meaning, and thinking, need to be further explained. One of the reasons for this inclusion is that Russell's verification problem does not arise as soon as one has it clear that one knows what one expects, because one knows what one meant with the sentence - i.e., one remembers the intended sense of the words (the intention behind the words, as it were). Before looking at this, we need to understand why notions such as 'to mean something' and 'to intend something' are not present in the T.

It is important to notice that Wittgenstein doesn't use the word "intention" at ali in the T. The general reason for this has been already stressed: Wittgenstein saw investigations related to psychology as irrel­evant to his views on logic. So my aim in the following is to show what Wittgenstein considered the relevant aspect of intentionality in the T. lt is not much, which is consistent with what Wittgenstein believed concerning psychological investigations, but it may illuminate his discussions in the period discussed in this chapter.

ln T 3.5 Wittgenstein says, "the applied, thought, propositional sign is the thought" (Ogden's Translation). So the applied proposi­tional sign is the thought sign. ln arder to explain what a thought is, then, we need to be clear about what it means to apply signs. ln 3.11 Wittgenstein gives us a further hint concerning the activity of

78 Wittge11stein 1s Philosophical Development

applying signs:

We use the perceptible sign of a proposition (spoken or written, etc.) as a projection of a possible situation.

The method of projection is to think of the sense of the proposition.

If the thought is the applied propositional sign and to think is the method of projection, to get clear about what thinking is for Wittgenstein in the T, we need to understand in what projecting signs consists. The simplest case of projection is the projection of an image of a plane figure onto a parallel plane (projection plane) at an incidence of 90 ôegrees. Those are the roles of this projection. As a result, the ratios of the figure to be projected are preserved in the projection plane and the figures are similar (qualitatively identical). With such rules, we can easily recog­nize what the original figure is by means of the result of the projection. We can, nevertheless, change those rules and create different kinds of projection'. If we change the incidence angle or do not require the parallel planes, the result of the projection will be figures that differ from the figures that we start with. With an incidence angle of 110 degrees, for instance, we would not know immediately what the figures used as the starting point were. We could only tell what the figures were if we knew the rules of projection used. We would, thus, understand the figures after a translation or derivation according to the projection rules.

Unlike the depiction of a real landscape, which presents the elements as they appear in the combination seen in the landscape, the projected signs of a proposition (Satz) don't look like the described situation and its elements. But if we know the rules used to describe the situation, we can recognize how the proposítional signs depict what is described. ln other words, if we understand the role of mies in language, we don't have to assume that entities of processes in the mind are the determi­nant aspect of meaning something. The pictorial conception is, thus, designed to avoid a psychological explanation of sense and meaning. Thus, the notion of thought was meant to explain how the proposition is a picture of what it describes by means of the notion of projection. ln 4.0141, Wittgenstein gives an example of a projection:

There is a general rule by means of which the musician can obtain the symphony from the score, and which makes it possible to derive

'I r

e a r t t 4 V

e r };

a i t fi ç V

a 3

t

e f1

The Invention of the Genetic Method 79

the symphony from the groove on the gramophone record, and, using the first rule, to derive the score again ... And that rule is the law of projection which projects the symphony into the language of musical notation. It is the rule for translating this language into the language of gramophone records.

To project a propositional sign - to think the sense of a proposition -we need general rules of projection: rules used to interpret the signs and to relate them in a certain way. What are those rules? We need two kinds of rules to represent facts by means of signs: general rules of correlation and general crules of articulation.14 We can think, for instance, that a match on a"box of matches represents the fact that Jack is at home. ln this case, we could say that the match would designate Jack while the box would designate Jack's home (here we have a correlation rule) - see 4.0312. But since any object or word can stand for any object (as Iong as we use the given object or word in projection), we need more to make clear what we are representing. We can also say that names don't have meaning in isolation, for they have meaning only once rules of articu­lation are in place (see context principie in 3.3). We need a rule that articulates the signs in a way that is similar to the articulation of objects in the situation that we want to project. Those are the rules prescribed by the forms function and argument (logical prototypes). We can say, for instance, that a rule prescribes that the position of the signs in the propositional sign expresses the relation between the objects that we want to represent.15 This means that the rule of projection constitutes a fact: '"That "a" stands to "b" in a certain relation says that aRb"' (T 3.1432). The sarne ideais what explains propositions that contain unary functions:

What symbollzes in <PI; is that <P stands to the left of a proper name [ ... ] What is common to all propositions in which the name of a property (to speak loosely) occurs is that this name stands to the left of a name-fonn. (NDMN, 116)

Note that the correlating rules used are grounded in the rules of func­tion and argument. Those are the most general rules of projection, for they determine any "concatenation of names" (T 4.22), simple or not. If we accept that negation is given when function and argument are given (T 5.47), we also understand that function and argument are the basis for all correlation rules of projection among propositions, i.e., they are

80 Wíttgenstein's Philosophica/ Development

fundamental for the construction of all complex propositions:

Every simple proposition can be brought into the form <l>x. This is why we may compose ali simple propositions of this form. Suppose that all simple propositions were given: then it can be simply asked what propositions I can construct from them. And these are ali prop­ositions and this is how they are bounded. (NB, 71; from 04.16.1916)

The forms of function and argument are not rules determined arbi­trarily, for they are the symbols given a priori in the very nature of propositions (T 3.315). They, at least partially, determine .the projection in any particular case and cannot, therefore, be gíven á posteriori, say, from "particular cases":

We portray the thing, the relation, the property, by means of vari­ables and so show that we do not derive these ideas from particular cases that occur to us, but possess them somehow a priori.

For the question arises: If the individual forms are, so to speak, gíven me in experience, then I surely can't make use of them in Iogic; in that case, I cannot write down an x ora í/Jy. But this I can surely not avoid at all. (NB, 65; from 06.19.1915)

The speciflc rules of articulation and ali rules of correlation are not, of course, given a priori; they are arbitrarily determined conventions (T 3.315; 3.322). However, since those conventions are already in place, we are committed to follow them and extract their consequences (T 3.342). This because they are grounded ln the a priori forms (rules) of logic that determine any projection.

The mies of projection determine and express under which condi­tions a proposition is true, for they give us the rufes to compare a proposition with reality. They determine, therefore, the sense of a prop­osition. To think the sense of a proposition is, therefore, to apply rules of projection (essentially, rules of logic). Thus, following rules of projec­tion is what essentially characterizes thinking given that a "thought is a meaningful proposition" (T 4; Ogden's translation). And what about the intentional element? What one intends is expressed by the rules used in each projection. The method of projection and its general mies are the only logícally relevant aspects of 'intention'. These aspects are logically relevant because they are part of the determination of sense and the understanding of propositions.

However, what guarantees, according to the T, that our thoughts don't miss the target, that we talk about specífic things wíth our propositions?

r

The Invention of the Genetic Method 81

Is there a projection when we use expressions such as "the present king of France"? In these cases, and in cases in which names don't name, we have an apparent violation of the excluded middle (see Russell's "second puzzle" in On Denoting). Russell's solution for the problem, as is well known, is to consider that the real logical form of propositions like "The present king of France is bald" isn't Fa, but ::J x (Fx & (y )(Fy ~ x=y) & Gx), i.e., it's said that only one object characterized by the predicate F exists and that that object is also characterized by predicate G. The important point in this theory, for Wittgenstein, is that something that apparently was a sense condition of the proposition is one of its truth conditions. We can see. a kind of "importation" of the sense presuppositions of the proposfüon; the presupposition now, because it's expressed in the proposition, is part of the proposition itself, of its own sense.16 Its sense is completely expressed at the end of analysis, when we should reach, supposedly, logically independent elementary propositions, in which simple names name simple objects. We must reach simples, otherwise the problem is reintroduced. ln this process, the question 11who or what do you mean ?" disappears, since "the so and so" is eliminated by anal­ysis and definitions (T 3.261). In this way, the need for Fregean senses is also eliminated.

In the T, thus, it is analysis that supposedly shows how world and language ultimately connect by means of names that appear in elemen­tary propositions, in which simple names must name simple objects. But, as we have already seen in Chapter 1, the final analysis of proposi­tions is not presented in the T. Analysis would show, however, in the application of logic, what is speâfic in each propositíon, which partic­ular state of affairs each of them describes. The meaning of simple names was supposed to be part of the complicated "tacit conventions" of language (T 4.002). Those "tacit conventions" grounded on logical syntax determine the "inner similarity" (4.0141) between propositions and states of affairs at the elementary levei. Note that ordinary and elementary propositions work in the sarne way, i.e., obeying the most general rules of projection.17 However, only at the elementary levei are propositions logically independent and the guarantee that names really name objects is ln force. Elementary propositions, thus, suppos­edly guarantee that we understand and mean something specific by our names when we use propositions.

It is important for us here that Wittgenstein's understanding of anal­ysis and elementary propositions in the T implies that there is no need of a notion of 1intention1 in this conception, for the meaning and, thus, the connection of thought and reality, is already given in the proposi­tion. We only need to find it out by means of analysis. If anaiysis, thus,

82 Wittgenstein's Philosophical Development

reveals the logically independent elementary propositions, it also reveals their "tacit conventíons11; it also reveals, then, how a picture (Bild) is a specific picture (Portraet) of a specific fact (a state of affairs). Thus, there is an assumption in the T that one cahnot go wrong in understanding elementary propositions (otherwise, w~ would have an infinite regress). lf we reach elementary propositions there is nothing else to do. This is what allows Wittgenstein to equate thought and proposition with sense (T 4) and to give a minimal characterization of 'understanding' by means of truth-conditions (4.024).

lf one knows what is the case if p is true, one implicitly knows that elementary propositions are revealed in analysis (T 5.5562), that they are given in the further specification of the truth-conditións (or 11impor­tation11 of sense conditions). There is no roam for the idea of further interpretation of elementary propositions because there is no roam for the further specification of truth-conditions once analysis has reached them. The logical forms that they express (function and argument) are given a priori and the arbitrarily determined conventions (tacit conven­tions) make themselves manifest in analysis simply because no further specification is possible. Once we have reached elementary proposi­tions, we have reached the indefinable primitive signs, the simple names (T 3.26). lf we reach elementary propositions, the analysis is complete. Since each proposition "has only one complete analysis11 (T 3.25), and this analysis brings us to simples, no question of ambiguity can arise at the end of analysis and any further specification is superfluous.18 That is, elementary propositions are self-explanatory.19

ln the T's treatment of pictures and projection there is, then, no use of intention as a psychological process. Nothing that happens inside someone's head or mind is relevant for the Tractarian conception of sentences as pictures or the explanation of thought as a rule-following activity, for the rules of projection that are thought "in the mind" would be the sarne if expressed on paper, for a thought is nothing more than a "meaningful proposition" (T 4). 20 Thus, the metaphysical assumption of the T concerning meaning, intending, and understanding something is not that mental processes underlying those activities are essential; it is, rather, that those processes don't need scrutiny because analysis brings to líght what is essential in them (namely, conventions grounded in primitive a priori rules of logic and simple names).

2.2.3 Intentionality, 'attitudes,' and what makes a thought true

In 1929 Wittgenstein abandoned his old conception of function and argument as the fundamental norms of representation at the elemen­tary level (see Chapter l); however, the assumption of general rules of

ç ç k a t s

s e

The Inventio11 ofthe Genetíc Method 83

projection (not necessarily bound to functíon and argument) are still in place (MS 107, 247). Those rules are now merely the rules given by various kínds of words of our existíng language. Wittgenstein also abandons, as we have seen in Chapter 1, his old idea of analysis. Analysis is not the search for logically independent elementary propositions which are supposedly given implicitly in ordinary propositions. Analysís is simply the investígation and perspicuous presentation of rules of 'grammar' of sentences and words in the existing languages without the assumption of the simple elements of the T:

Expecting is connected with looking for: lookíng for something presuppÓses that I know what I am looking for, without what I am looking for having to exist.

Earlier I would have put this by saying that searching presupposes the elements of the complex, but not the combination that I was looking for.

And that isn't a bad image: for, in the case of language, that would be expressed by saying that the sense of a proposition only presupposes the grammatically correct use of certain words. (MS 107, 256; PR §28)

Thus, Wittgenstein does not assume self-explanatory elementary propositions anymore or the "elements of the complex" as understood in the T. He needs, thus, to establish the specificity of each propositlon by ordinary means. Wittgenstein's hope is, at first, that his old concep­tion will need only a change of angle:

It is likely that the angle of my whole conception ofthe propositions (to the present) has to be, as it were, turned a little bit (um einen kleinen Winkel gedreht) to be right ín order to really fit. (MS 107, 247)

Thus, the general problem, after the abandonment of the Tractarian ideas of general modes of representation and analysis, is the following: if analysis does not reveal self-explanatory logically independent elementary propositions and the simples, how do we understand that 'p' describes exactly the fact p? The first answer given by Wittgenstein is that we intend or mean 'p' in a certain way, i.e., we somehow express that 'p' must/should be understood in this or that specific way.

Wittgenstein introduces the notion of 'intention' when distinguishing his and Russell's accounts:

If you exclude the element of intention (Intention) from language, its whole function then collapses.

84 Wíttgenstein's Philosophical Development

What is essential to intention is the picture: the picture of what is intended.

It may look as if, in introducing intention, we were introducing an uncheckable, a so-to-speak metaphysical element into our discus­sion. But the essential difference between the picture conception and the conception of Russell, Ogden and Richards, is that it regards recognition as seeing an internai relation, whereas in their view this is an externai relation. (MS 107, 289; PR §21)

Wittgenstein claims severa! times that Russell left out of his analysis the element of intention (MS 107, 289; PR §21; MS 108: 259, 261, 262), without which language "collapses" (MS 107, 289; PR §21). Even though the psychology related to intention was irrelevant for the T, the relation between sentence and reality cannot be, in Wittgenstein's Tractarian view, mechanical (as we have seen, it is normative). The sarne view is in place in 1929-30. The recognition of a picture as a picture of a fact (an order as the arder for a specific action, etc.) must be directly seen in the comparison between picture and fact. The recognition of what makes a sentence true is not a causal, externai, relation as Russell and Ogden & Richards argued, but an "internai relation." But what does "internai relation" mean here and how is intention related to it? Moreover, how are we supposed to see the internai relation? Wittgenstein uses the term 'internai' here in opposition to 'externai' in his description of the causal theoretic view. Wittgenstein sees the causal theoretic view as assuming something different from the picture, and what is pictured as what needs to be seen in the recognition of a sentence and what corresponds to it. So Wittgenstein argues that what is seen is something relating picture and fact itself; something that is not, however, different from or outside picture and fact (therefore not 'externai'). 'Internai' indicates that the relation is rule-governed, namely, by the rules used in the projection. But what is supposed to be seen is the intention by means of which we understand what is meant:

All my remarks are based on (the feeling) the insight (Einsicht) that thought has an internai relation with the world, and not externai. That we, therefore, mean what we say. (MS 108, 194)

Thís, of course, raises the question of how we mean what we say and how íntentions are directly recognized without an 'externai' relation. How do we understand what is meant by a sentence written by some­body else? How can we mean 1p1 ín a specific way, if what we mean is

r 1

1

!

1 The Invention of the Genetic Method 85

expressed in general rules of projection? Note that the supposition that the specificity of sentences is to be found by Tractarian/Russellian anal­ysis has been abandoned. How can we understand 'p' in a certain way? ln arder to understand 'p' correctly, we have to interpret 'p' correctly, i.e., interpret 'p' in the way that it was meant. What is the criterion for the correct interpretation of 'p'? Of course, whatever the answers for those questions might be, the notions of 'understanding', 'meaning some­thing', and 'intention' are to be understood against the background of the "grammatically correct use of certain words" (MS 107, 256; PR §28; quoted above).

We need to be clear, so thought Wittgenstein, about the 'specific way' propositiorís are meant, how they are intended, in arder to establish whether a proposition corresponds to reality, whether it is true or not. But where is the intention? There is an obvious objection to the suppo­sition that it is in the picture:

How is a picture meant? The intentlon never resides in the picture itself, since, no matter how the picture is formed, it can always be meant in different ways. But that doesn't mean that the way the picture is meant only emerges when it elicits a certain reaction, for the intention is already expressed in the way 1 now compare the picture with reality. (MS 107, 292; PR §24)

Each picture (sentence) is somehow ambiguous conceming its veri­fication because it can be meant in different ways, i.e., its rules can be meant in various ways (I will come back to this point in the next pages). So one could think that the causal account at least offers an answer to this problem: you wait and see what brings the feeling of fulfillment and this is what was expected ('meant' as the expectation). Wittgenstein denies this altemative, of course, for one knows exactly what one means with a command, for instance, before it is fulfilled (now). But then Wittgenstein's account needs to make clear how some­thing specific is meant by each sentence, and how this specificity is understood when one compares it (now) with reality. ln fact this is even more so, because1 as we have seen in Chapter 1, such a specificity must somehow relate to phenomena. What is meant ultimately is connected with phenomena:

Whatever someone can mean by a sentence (Satz), he also may mean by it. When people say, by the sentence (Satz) 'There is a chair here', I don't merely mean what is shown me by immediate experience, but

86 Wittgenstein's Philosophical Development

something over and above that, you can only reply: whatever you can mean must connect with some sort of experience, and whatever you can mean is unassailable." (MS 107, 209; PR §230; translation slightly modified)

The difficulty of finding a connection is aggravated by the Jack of understanding of how one refers to the objects present in a specific fact:

The language of the score, the instruction for playing an instrument.

But the problem is precisely: how is an instruction possible? ... How can the instruction relate to its object, for it [the object} is

only there when it is there and cannot be substituted. If 1 say what the sign is an instruction for, so I only say something, 1 give further instruction.

Is it the sarne or different if 1 introduce a language with real pictures? Isn't it like this: picture and what is pictured are different in some relation, otherwise they would be identical; and here the element of substitution must enter. (MS 107, 243)

How is this substitution possible? Wittgenstein sometimes suggests that a specific substitution of a description is possible because reality and sentence must be in the 'sarne space' (MS 107, 263). This idea is grounded in the comparison of a proposition with a measuring rod. A measuring rod, however, is certainly in the sarne physical space as the object to be measured. But propositions and reality cannot be in the sarne space in the sarne sense (MS 107, 269; PR §45). Another símile used is the one of "fitting objects." A proposition seems to fit reality like a "hollow form" fits a "full form" (MS 107, 265; MS 108, 214). Such a símile seems to be suggested by the incorporation of the idea of expec­tation and fulfillment (Erwartung and Erfuellung). Here, again, however, we only have a símile. How can sentences, or thoughts, be the "hollow form" of a fact? Is the fact something amorphous that fit the linguistic (or mental) form? (MS 108, 213). Whatever símiles we try to use, it is clear that we need to explain how we can mean or intend propositions in such a way that they correspond to specific facts.

The first difficulty (as expressed in the passage above) with finding such an explanation is that the intention is not in the picture itself, since each picture can be understood in different ways. But if it is not in

f 1

The Invention of the Genetic Method 87

the picture, then where is it? It is also not outside the sentence, at least not in the sense of Russell's externa} relation:

To compare the symbol - I mean that which is used as symbol - with reality is easy. The difficulty consists in comparing it with reality together with the symbolizing relation.

Can the interpreted symbol be something different from that which is employed in the present reality?

The causal explanation of meaning does not help because in it we talk again hypothetically about a state of affairs, which is exactly what is problematic. (MS 1081 199 from 06.29.1930)

/

But, again, if Russell is wrong, what is correct? The 11symbolizing rela­tion" must be somehow explained for, after all, we recognize it when we compare a sentence to reality. Russell is wrong, but the pictorial concep­tion of sentences does not bring us further as well:

... the sentence (command) is not simply put together with reality, but it is compared with it in/with a specif!c tendency. (MS 108, 192)21

The 11symbolizing relation", the "specific tendency", the "intention" all seem to point to the need for further explanations and, at the sarne time, seem to be intangible. Wittgenstein's trouble can also be seen when he explains the origin of the "pictorial conception" and "the meaning" that underlies a picture:

At that time I was brought to the picture theory of meaning by means of a newspaper item where it was said that in Paris, in a triai about a traffic accident, this accident was presented by means of dolls and buses. How, then1 such a presentation with dolls is distinguished from a play with dolls? (Of course, through the meaning), but what does it consist in? (Some would say: its effect, it alone is its meaning). (MS 108, 203-4)22

Now the nature of meaning seems to be the unexplained element that needs explanation: 11in the words 'to mean this and this/some­thing' is the whole problem encapsulated" (MS 108, 195). It seems that complete specificity can only be achieved when the meant object is present, when it replaces the word (MS 107, 243). Otherwise, it seems to be possible that the object meant is not exactly this one. Meaning

88 Wittgenstein's Philosophícal Development

something seems to consist in pointing to something present: this doll means this person. This seems to suggest, ín íts tum, that a language of gestures is the primitive form of language on which our language is grounded. As Augustine said, a language of gestures is "the natural language of all peoples" (quoted in Pl, §1). However, meaning some­thing is also meaníng that this and this is the case (MS 108, 205). This form of meaning seems to assume that one means only when one has complex sentences. Moreover, one must also mean that "this doll means this person." And if the meaning depends on a specific final gesture, how can one understand what is meant beforehand? lt seems, thus, that mental processes must replace the object beforehand.

2.2.4 The intangibility of intention and its temptations

Wíttgenstein does not take a definitive step ín the direction of an assumptíon concerning mental pictures or processes in MSS 107-8. Thís is a very significant fact. Wittgenstein thought that he had to avoíd two things (as already in the T): speaking nonsense and íntroducing a psychological investigation in the field of logic (or 'grammar').

The element of intention, or how we recognize the connectíon between sentence and reality, cannot be further explained because, when we try to explain it, we reach the "limits of language":

The sentence, the command - as it were [quasi], carries on reality (Wirklichkeit) as it links with reality (Realitaet) and represents a change. As if we had a doll that represents the present situation of my body and with it the planned changes would be carried out - in effigy.

Consequently, the problem would then be again the nature of the representational relation if we say, for instance, "this dali/figure/ should be you, etc."

1 feel that to answer 1 will here, again, run into the limits of language. (MS 108, 192)

lf the relationship between sentence and reality is internai (MS 108, 195), and it cannot be meaningfully expressed as a fact, we cannot really talk about the intentional without saying nonsense. The only thing to be done in a situation in which we tend to talk nonsense is to "elucidate the impossíbility that we run against" (MS 108, 197). ln MS 108, 265, this elucidation is still not achieved (07.29.1930):

... this impossibility of expressing through language the condition of the agreement of the significant proposition - thought - and reality

r The Inventíon of the Genetic Method 89

(Wirklichkeit) is the solution of the riddle (though 1 cannot see it clearly).

Since the internai relation is an important part of the elucidation of the relation between sentences and reality, it seems that Wittgenstein has a big problem, for he cannot describe what the relation is, but he also needs to elucidate his own views further if he wants to dismiss Russell's account completely. He may end up saying "too little" (as presumably in the T):

It is difficult to speak about the relationship of language to reality without talking nonsense or saying too little. (MS 107, 205)

A "psychologization" is also to be avoided, thought Wittgenstein: ln philosophy one is always in danger of giving a mythology of

symbolism or [a mythology] of psychology. (MS 108, 104; from 02.1930)

The "mythology of psychology" was avoided in the T by means of a strictly a priori investigation that, however, introduced a mythology of symbolism (as seen in Chapter 1). This symbolism suggested also a mythology of the link between reality and language (names work like effigies (MS 108, 192)). The threat of a mythology is now present in the investigation of the "proper rules of language":23

The psychological abyss is always a threat in this investigation [of the proper rules of language]. But one can overcome this danger only if one has seen it clearly (klar ins Auge gefasst hat). (MS 109, 38)

We may not need to investigate the mind (say, by introspection) in order to discover what the intentionality present in language is, even when this seems to be suggested by the normal steps of our investiga­tion. Wittgenstein's strategy, again, is to isolate what is logically relevant and not to deal with the psychological:

ln "understandingwhat someone means" we have to be able to sepa­rate the specifically human behavior, all that is psychologically and physiologically interesting, from what is objective (sachlich), logical. (MS 108, 261)

One of Wittgenstein's major worries in 1930 is thus to resist the normal steps (the temptations) of a suggested psychological investiga­tion and, at the sarne time, to explain intention sufficiently in order to

90 Wittgenstein's Philosophical Development

block Russell's account. The fundamental Tractarian point that must be kept ln mind in arder to get out of these difficulties is that thought is "an articulated process" that has an "articulated expression" of the right multiplicity (MS 107, 262; PR §32). The Tractarian articulation might have been a simplification (or a "mythology of symbolism"); nonethe­less, the idea of the centrality of a logical articulation has not been given up. Such an artículation, from the T to 1930, was not of a psycho­logical nature. ln 1930, the articulation is supposedly also logical, but not grounded on a symbolism anymore.

The most puzzling fact about the inv.estigation of intention, meaning and understanding is the impression of intangibility of those processes. The element of intention seems to be concealed and véry difficult to grasp, if one takes it as "a phenomenon, a fact" (MS 108, 186). lf one sees a picture or a group of signs that look like a sentence, one doesn't know how to compare it with reality, if one doesn't have the method of projec­tion. It is the method of projection, the general rules of translation, that makes the comparison possible (MS 108, 219). ln the sarne way that what "constitutes the inner similarity" between a symphony, a score and the music is a general rui e- "a rule of projection" (T. 4.0141)-what constitutes the similarity between fact and sentences is also a mie of projection. A rui e of projection is a mie of translation of given signs into another group of signs. ln other words, a mle is a method of interpretation. This is the element of intentíon present in thought that distinguishes Wittgenstein's non-causal approach from Russell's causal approach.

ln MS 108, 219-20, Wittgenstein explains that someone may dream that A visits B, but that this is not an expectation even if A does visit B in the next day. According to Wittgenstein, the intention that makes the dream an expectation is missing. The dream has become an expec­tation only if we can compare it now (when it is dreamt) with reality, which will make it tme or false (fulfill it or not).z4 ln the sarne way, "a picture (in the strict sense) is not sufficient, for it is not given in it how it should be compared with reality" (MS 108, 219). Wittgenstein goes on to explain what is missing, what counts as 'intention': "The method of projection must also be there; but then the picture reaches to the place where the object of the picture is". Therefore, says Wittgenstein, "we could say that the intention is the method of projection" (MS 108, 219; my emphasis).

The problem is that the mles of projection are general mies and so apparently insufficient to determine completely in advance a specific situation.zs Since rules of projection can always be interpreted and understood in various ways, lt may seem that the method of projection

r 1

!

The Invention of the Genetic Method 91

or process of projection is not intention itself, but only part of a "process":

The method of projection must be contained ln the process of projecting; the process of representation reaches up to what it repre­sents by means of a rule of projection. lf l copy anything the slips ln my copy will be compensated for by my anger, regret, etc. at them. The total result - i.e. the copy plus intention - is equivalent of the original. The actual result - the mere visible copy - does not repre­sent the whole process of copying; we must include the intention. The process contains the rule; the result is not enough to describe the process. tWCL 30-32, 36; from 11.10.1930)

The temptation26 here is to take the process as something more than the rules of projection and to start an investigation of the process of representation (MS 107, 245). What could it be? It could be that the process of thinking and meaning itself needs to be explaíned further and that the elucidation by means of rules of projection leaves some­thing very important out - perhaps something behind or beyond the rules.

The sarne temptation is also present in the activity parallel to projecting, namely understanding (or interpreting). In Wittgenstein's Tractarian and post-Tractarian account understanding must be the translation according to the rules of projection. As Wittgenstein says:

How do we know that someone had understood a plan or order? He can only show his understanding by translating it into other symbols. He may understand without obeying. But if he obeys he is again translating - i.e. by co-coordinating his action with the symbols. So understanding is really translation, whether into other symbols or into action. (WLC30-32, 24; 10.1930)27

Essential for the understanding of a picture and what is depicted is its translatability, and what is common to different kinds of sentences is the possibility of translation from one into another. Even action is a kind of translation. The translation between different kinds of sentence or translation of fact and picture is similar, according to Wittgenstein, to translation between different languages:

You can draw a plan from a description. You can translate a descrlp­tion into a plan.

92 Wíttgenstein's Philosophical Development

The rules of translation here are not essentially different from the rules for translating from one verbal language into another. (MS 107, 243; PR §20)

But what about the understanding of my own thoughts? It does not seem plausible to think of any translation going on when I think. It is clear that, in the case of a thought, "we don't have the feeling of inter­preting" (MS 108, 224). This seems to suggest, however, that when one thinks one can see the rules of projection immediately, while a plan or sentence on the paper needs to be interpreted or translated. Thus, it seems that I don't need to interpret what l think, but need to interpret what others say: "Thinking means operating with plan~. A thought is not the sarne thing as a plan because a thought needs no interpretation anda plan does" (WLC30-2, 24 (10.12.1930)). This seems to bring us to the "uniqueness of thought" (MS 108, 208).28

The next step here (the next temptation) is to make a distinc­tion between what is "outside" and needs interpretation, and what is "inside" and does not: "The interpreter sees the thought, in fact, from outside and not from inside; ali that we see, we see from the outside: i.e. ali that we experience is phenomenon" (MS 108, 187-8). It seems that to understand the intention we have to see it from within, while we can only see it from without. But we may also have a problem when trying to get within what is meant. How can we do it? The "element of intention" seems to slip away every time we get close to it. For if we ask someone to explain what is meant only new signs that need interpretation again will come to us. If you instruct someone about what is meant, "only" signs are given: "If 1 say what the sign is an instruction for, so 1 only say something, I give further instruction. ( ... )To me there is nothing left to answer the question concerning to what p is an instruction for except saying i.e., giving a further sign" (MS 107, 243-4).

But the problem also seems to appear when one considers how he, for instance, is the one from whom something is demanded. ln the suppo­sition that one doesn't follow the demand, it seems that something else has to take its place:

But how is it with myself when the demand is directed at me? If 1 understand the demand and don't follow it, then understanding can only consist in a process that replaces (vertritt) the execution; there­fore, in a process other than the one of execution. (MS 107, 245)

So here we will be tempted to assume a psychological process to explain how a proposition relates internally to reality. Understanding

r

The Inventíon of the Genetic Method 93

may be seen as a process that can take the place of the execution, and it is seen as the process that can bring us closer to reality by means of a substitution. One is tempted here to postulate a kind of shadow between sentence and fact. 29

The investigation in 1930 was, then, pointing in two directions: it could lead to an explanation showing that the rules of projection, which are rules of 'grammar,1 and their interpretation are suffícient to characterize what is intentional in language wíthout any appeal to mental acts; this could also show that the causal theories of Russell and of Ogden & Richards misrepresent an 'internai' relation by means of an 'external' one. However, the investigation could also show that WittgenstefÍ:l needed to go beyond the rules of projection and assume a fundamental mental act or mental process to explain the compar­ison between picture and reality, in this way introducing psychological processes not present in the T. This second alternative is Wittgenstein's temptation around 1930.30 A temptation, for sure, that he wanted to avoid, as we have seen. ln what follows, 1 explain how the calculus conception and the new method solve Wittgenstein's problems and avoid the temptations.

2.3 Language as a calculus: autonomy and the role of a system of rules

At the end of MS 108, Wittgenstein makes clear what the solution of his difficulties should look like: "Philosophy will consist, at the end, of trivial sentences, remarks" (MS 108, 238). The trivial sentences, remarks, as understood by Wittgenstein at the tlme, are the guidelines of the calculus conception of language. The triviality of the conception should show how to avoid getting into the doubtful fiel d of psychology, and also explain how we can compare sentence and reality in a specific way. Here 1 summarize the view. To take language as a calculus means, first of all, to explain the essence of language as a system of signs: "One could explain the essence of language with a system of signs'1 (MS 109, 249).31 The sense of sentences of the system and the meaning of its words are determined by rules, which operate always in contrast to other rules. It is the contrasting alternatives given ln the calculus of language that guarantees that the rules of 'grammar' have the right multiplicity in the expression of what is meant. Rules exhaust and constitute the meaning of words: meaning is nothing more than the position of words in a grammatical system. Meaning, thus, is neither an object, nor some­thing mental. The meaning is part of the interpretation of language, and not of its application. When language is applied, new meanings are

94 Wittgenstein's Philosophical Development

not created. lt is the usual meaning that makes clear that we don't need a substitution or words and objects (this book, for instance) each time we express a sentence.

Note that this new view of meaning reflects Wittgenstein's idea of meaning defended for mathematics in PR, í.e., the meaning as equiva­lent to and determined by the rules of the system. lt also reflects the idea of autonomous systems in mathematics, as presented in Chapter 1 (Section 1.5.4). If one looks at language as a game or calculus, one looks at it solely as a set of rules and one does not assume any particular appli­cation of the rules. One tabulates and studies the rules of language as a system and one does not take into account why people play the game (its function or purpose), as one can tabulate the rules of chess without worrying about why it is played (for pleasure, to win, to win money, etc.). This means that the function of language (be it 'to cause behavior,' as for Russell, of 'to describe reality,' as for Wittgenstein earlier) is not taken to be the fundamental characteristic of language.

Wittgenstein's application of the idea of a system of rules of grammar (a game or calculus) to the problem of the relation between sentence and reality in 1930-31 consists in two moves. The first involves explaining the relation between sentence and reality as taking place inside language, í.e., inside the system of signs or calculus:

The expectation waits till the moment of decision. But then it touches the decision. - - Like the calculation [touches] its result. The fitting of expectation and decision expresses itself in the common feature of the word expression [ ... ] Event and expectation touch in language. (MS 109, 59-60)

What is relevant when we describe the expectation is that it and the event that fulfills it (or when we express a sentence and the event it describes) are expressed with the sarne words: 110f course, the expecta­tion and the fact must have something in common. But this is expressed in that the language describes both with the sarne words." (MS 109, 217).32 They are described wíth the sarne words because both belong to the sarne grammatical system:

The meaning of a sign is the complete symbol to which the sign belongs. Or we could say that it is the place where it is in the gram­matical space. (MS 109, 175)

The grammatical space is "given by the grammatical rules" (MS 110, 135). Rules constitute the meaning of words and are, therefore,

The Invention of the Genetíc Method 95

autonomous in a strong sense, i.e., the word has meaning only if it has rules of correct use. There is nothing beyond the rules that could be called meaning:

There cannot be a discussion about whether these or other rules are the correct ones for the word 'not.' For the word has as yet no meaning without these/the rules; and if we change the rules, then it has a different (orno) meaning and we can just as well change the word. Hence these rules are arbitrary, for the rules first make the sign. (MS 1101 133)

So if the-'rules constitute the meaning, and it is the position of a word inside a system what should be called 'meaning,' then 'grammar' is autonomous, or "self-meaning", as Wíttgenstein also says (MS 111, 111). Note that Wittgenstein derives the arbitrariness of 'grammar' from its autonomy, in the passage above ("Hence ... "). The autonomy of 'grammar' ímplies, ít seems, the arbitrariness of 'grammar,' i.e., it implies that the rules cannot be justified. Since the rules determine the meaning, the rules "make the sign," there is no previous or externai description of facts that determines the rules. Without the present rules, a word does not have a meaning and new rules would simply change the meaning of a word. The description that would supposedly justify a rule would simply be the indication of a new rule, and the old rule would not be justified.

The explanation of meaning is always a sentence inside an autono­mous system of rules. Thus, the apparent triviality that the meaning "can only be what the explanation of meaning explains" (MS 109, 140) means that the meaning is what an explanation of meaning inside a system explains. In this case, the meaning is not an object but íts positíon inside a system of grammatical rules: "The place of a word in language is its meaning'' (MS 110, 74; BT, 30). Sometimes we explain the meaning by pointing to an object (ostensive definition), but not always. When we do it, the object works as a sample that is part of the language (MS 109, 144). Actually, we can explain the meaning of a word whose named object does not even exist and the name does not lose its meaning once the object named is destroyed. The idea that the meaning of a word is the object that a name replaces in language must be, therefore, wrong:

One always has the false idea that the important thing concerning the meaning of a word is an object, i.e., a thing, in the sense that the sword Nothung was the meaning of the word 'Nothung.' But

96 Wittgenstein's Philosophical Development

there is something wrong here, for I can say, "Nothung doesn't exist anymore"; and is 'Nothung' meaningless here because the sword doesn't exist anymore? (MS 109, 140)

The ostensive definition, which is central in a causal theory of meaning, and could be taken as the last interpretation of what is meant, is a rule among others; for "meaning is fixed inside language" (WLC30-32, 62), and not by objects outside it. Even if one equates meaning and object, the ostensive definition gives one nothing more than a kind of table that indicates a rule. For instance: red - red object, blue - blue object, etc. The object, thus, does not serve as more than a sample. That is, at the end the role of the 'meaning' is like the role of anothér sign (sample) inside the language (the object, the 'meaning,' can be replaced by a picture of the object without more ado). One can compare a table that ascribes meaning to sigos or ostensive definitíons with the method of projection in spatial projection, as Wittgenstein in the T, but one needs to be clear that it is "nothing more than a comparison" (MS 112, 84r; BT, 53). lt is not, however, a bad comparison, but it should not lead (as Wittgenstein in the T) to stopping an investigation of the variety of functions of words: "[lt is] a pretty good comparison, but it does not spare us from having to investígate how words function, leavíng aside the case of spatial projection" (MS 112, 84r; BT, 53).

Ostensive explanations, as any other explanation of meaning (a verbal definition, for instance) are part of the structure of language, its grammar, its interpretation, and not part of the application of language (say, a descríption of a fact) (MS 109, 178). 33 This already indicates that one ís puzzled about íntention and meaning (what is meant) because one assumes that one intends, and means, and makes the final inter­pretation of sigos by means of an object 'outside' language. The object, as the fact, is only 'outside' language, however, in the determination of the truth, in the application of language. ln the interpretation of language, it is a part of language, for ostensive defínitions are part of the mies, which we must understand before the application of language takes place. The notions of 'intention' and 'meaning something,' as we will see, are also part of language and its interpretation.

The second move introduced in the calculus conception is the explanation of the content of the mental notions of 'understanding,' 'meaning something,' and 'thought,' by means of a system of rules for sentences. The relevant content of mental notions is expressed in the sentences of a language, i.e., ln a system of rules, and is determined

The Invention of the Genetic Method 97

inside it: "a sentence/sign has sense only in the system of a language'1

(MS 109, 175). ln Wittgenstein's view, the primary condition for under­standing a sentence is not the knowledge of how it is to be compared with reality or how it can be made true, but the knowledge of its rela­tion to other sentences of a system of language:

... it is impossible to understand only one sentence of a language. I.e., a sentence can only be understood as part of a system. (MS

109, 175)

Understal).ding a sentence is nota matter of isolated insight, but "anal­ogous to uri.derstanding a move in chess" (MS 109, 182), i.e., it assumes the background of a system of rules, a multiplicity of contrasting alter­natives. For someone who does not know the rules of the game, an isolated move of the queen in a game is incomprehensible. One under­stands a move of the queen when one contrasts its move with the moves of the pawns, the king, etc.

'Understanding,' 'meaning something,' etc., understood as psycholog­ical processes, can be dismissed. ln order to understand what someone means by a sentence it is not necessary for me to know his state of mind or how the sentence is psychologically related to a possible fact ln our projections. Even if one assumes that what one understands is ln any sense mental, the mental has the multiplicity of a sentence, an explana­tion: "Understanding is not only aroused by means of explanations, but must (also) itself be of the multiplicity of a/this explanation" (MS 110, 54). Since the articulated thought "in the mind" can be written on paper, the "thought as psychical process" cannot do more than signs on paper (MS 109, 26). lmages in the mind could also be equated, in prin­cipie, with signs (for instance, a red mental image can be replaced by a red sample). Wittgenstein, thus, treats the 'inner' and the 'outer' as not essentially different. This, he says, is the "behavouristic" in his concep­tion (MS 110, 296).34 Anything that is relevant in the mental character­izes a symbol inside the system of rules; if something is not a symbol, it is "amorphous" (MS 108, 216). What is "amorphous" is completely arbitrary (ad hoc). Meanings are not, however, invented ad hoc:

'l understand these words' (which l, for instance, say to myself), 'l mean something with this', 'they have a sense' must always mean something like: 'they are not sounds invented ad hoc, but signs from a system'. (MS 109, 285; BT, 199)

98 Wittgenstein's Phílosophical Development

What we need to look at, therefore, when we want to understand what is relevant in regard to mental notions is how they are expressed in a sentence, an explanation, and how this sentence relates to other sentences of the sarne system of language:

The answer to the question "what do you mean [?]" should be the explanation of the system of signs to which the given sign belongs. -And only this is done through the answer "I mean that you should come,11 for the sign was now translated into a sentence of a language that we know. And a language is qnly understandable because we know it, its system. For ali explanation can do nothing except teaching us the language. (MS 110, 294)

Now we have the elements to show how the calculus conception deals with the problem of the specificity of sentence and fact. What does it mean for a sentence not to determine specifically a state of affairs? It means that the sentence is ambiguous, for it allows more than one interpretation among the possible interpretations inside the system of rules; that is, there is room for doubt concerning what is meant. However, language itself has the means (the "sufficient multi- · plicity") to specify interpretations, to exclude alternatives: /1

••• the word-language has at its disposal the sufficient multiplicity in order to exclude a doubt [concerning what is meant], since we would say {if needed] something different in a different way" (MS 110, 72). The specificity of a sentence, thus, depends solely on other sentences of the sarne system of rules.

Since one always means something "in contrast to" other given possi­bilities (MS 109: 170, 184, 220, 222), determining the specificity of a fact by means of a sentence is tantamount to determining one inter­pretation among others already given inside the system. When one explains what one means, one explains it in contrast to other possible explanations in the system. When I expect John to come, for instance, I expect him in contrast to, say, Peter or Deborah. Those names are explained by different ostensive definitions, which are rules inside a system of language. When he arrives, I use the name 'John,' one of the signs used to express my expectation, to describe the fact that he has come. Meaning something or intending something is, thus, choosing one of the given possibilities of the system: one rule in contrast to other rules (MS 109: 222, 280). One specifies a fact, thus, when "the descrip­tion leaves nothing open" (MS 109, 298-9), when ali other alternatives inside the system have been excluded.

The lnvention of the Genetic Method 99

If we have the impression, then, that somethlng is only hinted at in the words we use, for instance, in an intended command, and that there must be something going on beyond the words, this is because the system of rules is not fixed:

The command x 1 2 3 4 xz

seems to be incomplete. lt seems to us as if something was only hinted at, something that is not expressed.

But something is only hlnted atas far as a system is not explicltly / or incompletely/ fixed. (MS 109, 276-7)

'

lf we know the system of numbers and the rule expressed in 'x21, then we know how to derive the second line; we know what specifically is intended. lf we don't, then the rules connected with 'x2' need to be explalned. What disambiguate sentences, thus, are the rules of language in the sarne sense that rules disambiguate a calculus. Ambiguities exist only because many rules of the system of language are not explicit. This is why the Wittgensteinian philosopher tabulates rules in the book of 'grammar' (MS 110: 221, 228, 233; MS 111, 77). lf one wants to know how the command is meant, one has to know the fixed rules of the system of signs (calculus), which intrinsically possesses the multi­plicity of excluding alternatives. If one does not understand the specific command, then one is not clear about various alternatives inside the system or allowed by the system. Reasons, justifications, and explana­tions come to an end, thus, when all alternatives grounded on the rules of the system are made explicit (MS 110, 96; 111: 35, 42; Ms 112, 97r).

Note that the calculus conception expresses the anti-psycholo­gistic stance of the T and, at the sarne time, avoids the "mythology of symbolism." The old anti-psychologism in logic finds its place ln 'grammar.'

2.4 Genetic method: rules, analogies, and the physiognomy of errors

The calculus conception of language may explain how the specificity of each sentence works and how to avoid the introduction of psycho­logical machinery without assuming the Tractarlan idea of analysis. However, it does not stop worries concerning the whole system of rules; it does not stop by itself questions that seem to ask for 'abso­lute,' system-independent, answers. The problems of the specificity of

100 Wittgenstein's Philosophical Development

a sentence and the infinite regress will reappear in MSS 109-10 in the forro of the problem of the specification of rules of 'grammar' and in the determination of which rule is followed in each action. The problems found in the causal theory of meaning, thus, reappear. This will make Wittgenstein tum his investigation to what generates the philosophical puzzlement in the first place (first of all, his and Russell's).

Note that, according to the calculus conception, language is to be seen as an autonomous calculus determined by its rules. lf disputes concerning what is meant by contrasting sentences should arise, the rules of the calculus, the grammatical1rules, are our last resort:

The only things that are exact and unambiguous arid indisputable are the grammatical rules, which ín the end must show what is meant. (MS 109, 90; BT, 374)

The problem is that one can, ín principie, have doubts concerning those rules. The rules themselves don't contaín theír own applicatíon:

"To know what is the case, if the sentence is true'' can only mean to know the rule according to which it is controlled. - But how is it depicted that it is to be controlled by this rule? The rule is only added to the sentence, but where is its application to the sentence depicted? But wherever it would be depicted, it would be depicted by another picture, and so we would get (into) an infinite regress (MS 109, 78; my italics).

The rules, as it were, cannot do the work by themselves, one might think, for they also need to be explained in a particular application. But once the rules are explained, the words used to explain the rules may also need an explanation. 35 Even if explanations in the system work in contrast to other explanations, it seems there is no guarantee that we will reach an end when we apply rules. lt seems that we will get into an infinite regress. Here, then, we find Wittgenstein's critique of Russell arising in his own view, again.

lmmediately after the passage quoted above, Wittgenstein introduces the seeds of what could be thought of as a new puzzling situation. A rule not only has different interpretations, but also seems to be in accord with various actions:

One could also say this: a drawing is not a plan because someone once - accidentally- walked in such a way that his path corresponded

The Invention of the Genetic Method 101

to the plan, but rather because he followed the plan according to a definite rule. lncidentally, otherwise every path would correspond to the plan (according to some rule). (MS 109, 86-7)

The point is also made elsewhere in the sarne MS: Does "to follow the rules of grammar" mean to think, in any

sense, about these rules while talking? No. - Does it mean, always to talk according to certain rules? No. lt means to follow rules. - But anyone who does anything at all, does this; for there will be a rule that corresponds to what he does." (MS 109, 281; my emphasis)

This sounds similar to what Wittgenstein says later: "whatever l do is, in some interpretation, in accord with the rule" (PI, §198). One is, thus, tempted to think that no course of action "can be determined by a rule, because every course of action can be brought into accord with the rule" (Pl § 201). Something is needed, so it seems, to fix the rule, i.e., to determine which rule is followed. If something does not fix the rule that expresses the correspondence of plan and action, the distinc­tion between accidentally acting in accord with a rule and following it disappears.

Also here, then, a problem with Russell's theory reappears in Wittgenstein's account: it seems that we don't have a way to completely determine what makes a proposition the description of something specific. lt seems that anything goes.

The difficulty is that it seems that we never get to the real rule, the completely definite rule. This difficulty arises when Wittgenstein discusses 'understanding a command.' Since understanding a command is a kind of translating (interpreting) a sentence eíther into different words or into action, the understanding of what is meant seems to be something that is added to the signs that we hear or see. ln MS 109, 273, Wittgenstein observes:

1 tell someone: 'Go call Peter'. - How is he supposed to know what 1 mean? It must have been explained to hlm. But still with signs. He follows this explanation now. If understanding was a necessary prep­aration of the following [the command], it must have added some­thing to the sign; but something that is not the execution.

What is this "something added"? What is it that 1 understand? Given the in-principle ambiguity of signs, one gets the impression that it is only the interpretation of signs that gives us the command that was

102 Wittgenstein's Philosophica/ Development

meant. But, apparently, the rule that grounds the description of what fulfils the command is "still with signs" and, therefore, different from the execution itself. This all seems to suggest that a sentence inside a system could be understood as meeting us in an incomplete manner; as though there was something that is hinted at (see MS 109, 277); or even worse: something that cannot be conveyed (MS 109, 169; MS 110, 55). One (presumably a philosopher) may not see or accept that language works as a system of excluding and exhausting alternatives, which reach an end. The temptation arises again that it is something in our minds, a mental state that corresppnds to understanding, or is the understanding, that guarantees that we can follow a command. Here we find, then, another element of Wittgenstein's critique of Russell in the form of a temptation. The only thing that seems to stop a regress is a mental process, the self-explaining intention (a kind of self-explaining third element, similar to Russell's third element).

What Wittgenstein needs, thus, in order to solve his difficulties is not merely the supposedly trivial idea that language works as a calculus. The idea of language as a calculus may be a short and trivial answer to the concerns Wittgenstein had, but it does not eradicate the puzzlement. One could also say, therefore, that showing the truth is not sufficient when one is dealing with philosophical puzzlement. One has to find a way to bring someone away from the mistake:

(ln order to convince someone of the truth, it is not sufficient to establish the truth, but one has to find out the road from the miscon­ception to the truth.)

(One has to start with the misconception and bring him to the truth).

(I.e., one has to expose the source of the misconception; otherwise hearing the truth is not helpful for us. It cannot penetrate as long as something else takes its place.) (MS 109, 58)

This passage shows that in the MSS 109 and 110 Wittgenstein is not simply reformulating the problems of his own explanations of the diffi­culties concerning projection, understanding and thinking that he had already dealt with in MSS 107 and 108. Heis also not simply explaining his conception of language as a calculus, which may be taken as the trivial truth concerning those matters. In the passage above, the emphasis lies in the exposition of the "source of the misconception." This, we will see, is the guiding line of the genetic method.36

l The Invention of the Genetic Method 103

Here we have to pay attention to two things. First, the similarity between the recurrent problems in Wittgenstein's and Russell's views is striking. lt would be even more striking, or extraordinary, if Wittgenstein were not aware of it; for he himself presented the infi­nite regress and the lack of specificity arguments agalnst Russell in his MSS and lectures of the time. Second, in many passages after the one quoted above, Wittgenstein makes clear that the interpretation of a rule is something that comes to an end, even though one may be inclined to forget it while philosophizing. Contrary to all philosophical paradox­ical conclusions, we do follow rules and understand what other people mean. As a matter of fact, the infinite regress ends with my acting:

/

... I don't need another model that shows me how /the depiction goes and, therefore/ [how] the first model has to be used, for otherwise I would need a model to show me the use/application of the second and so on ad infinitum. That is, another model is of no use for me, I have to act at some point without a model. (MS 109, 86)

lf we needed a new interpretation and a new model, it would be, presumably, impossible to act. Our problem, if real, would stop us from acting, which is obviously not the case. What, then, is the nature of a problem that doesn't really stop what it seems to stop? Wittgenstein thinks that problems that supposedly should make a difference to our action, but do not, are the product of some confusion in our reasoning. ln arder to get rid of the problem, first, we have to be clear that suppos­edly self-explanatory devices like mental processes don't really help us:

No psychological process can symbolise better than signs on the paper.

The psychological process also cannot achieve more than the char­acters on paper.

For always again are we tempted to want to explain a symbolic process with a particular psychical [mental] process, as though what is mental could do more in this matter than the sign. (MS 110,18; my emphasis)

As the passage indicates, Wittgenstein is well aware of his temptations and the way that they relate to Russell's views. His new approach to the problem is to investigate what is misconceived in the train of thoughts that gives rise to the temptation to use subterfuges self-explanatory

104 Wittgenstein's Philosophical Development

mechanisms - in philosophy. He intends neither to assume "some­thing" self-explanatory (as a condition of possibility of understanding mies, say), nor to accept Russell's account:

Ali difficulties of philosophy can ónly be based on misunderstand­ings. A discovery is never needed, is not necessary to solve them. It is [they are] a misunderstanding and can only be solved as such. I.e., without violence. For the door gets open and everything in it is ali right; you only have to understand the lock and move in the correct way. (MS109, 298)

The misunderstandings that Wittgenstein has in mind ate grounded on false analogies. To be led byfalse analogies into mistakes, and to be unable to get rid of them is, for Wittgenstein, "the morbus philosophicus":

We go astray through false analogies and we cannot get out of this entanglement. This is the morbus philosophicus. (MS 110, 86-7)37

One of his new tasks is precisely to warn against false analogies that can bring one to philosophical puzzlement:

lt is one of the central activities of philosophy to warn against false comparisons. To warn against false comparisons/similes that underlie our modes of expression - without our complete consciousness of them. (MS 109,174)

This central activity is, for Wittgensteín, 'grammatical', for at the bottom of our mistakes there might be analogies suggested by our modes of expression (místakes caused by "false grammar" or "false compari­sons/símíles", as he also says). For ínstance, one might be misled into assuming self-explanatory mental images ín the following way:

How strange that if he isn't here 1 can look for hím, but not poínt to him. That is really the problem of looking for something, and it shows the misleading comparíson.

One might want to say: If I'm looking for him then surely he too has to be part of the process. Then he also has to be part of the process if 1 don't find him, and even if he doesn't exist. (MS 110, 274; BT, 367)

We might be tempted to postulate mental images (ora mental 'shadow' between sentence and fact) because we are misled by the similarity

The lnvention of the Genetic Method 105

among the forms of expression. When one explains who John is, and John is ln the roam, one can point to John. When John is absent, and one looks for John, John seems to be somehow present. After ali, we say: "l am looking for John." So it seems that our words indicate that the object needs a replacement in the mind when it is absent.38 Underlying this whole train of thought, one could think, is a "confusion between the meaning and the bearer of a name" (MS 110, 276). Because, suppos­edly, if 'John' means the object John, when John is absent we seem to need a new bearer for the name (after all, 'John' still has a meaning in the sentences in which we employ this name).

Wittgenstein considers philosophical mistakes "difficult grammatical illnesses" (MS 110, 247). "The cure", says Wittgenstein, "is the pointing out (Aufzeigen) of the misleading picture" (MS 110, 248) that leads to the philosophical question. 39 The idea is also to show the steps that lead to puzzlement:

One of the most important tasks is to express all false trains of thought in such a characteristic way that the reader says 'yes, I meant it in exactly this way.' To portray the physiognomy of each errar (MS 110, 230; BT, 410; modified translation) 4º

'To portray the physiognomy' of an errar means expressing the thoughts that lead to the errar and finding the source of the train of thoughts that leads us to the formulation of the problem. Later in 1931, Wittgenstein points out that once this is made explicit, the reader can see for himself what is wrong:

l should be only the mirrar in which my reader sees his own thinking with ali its deformities (Unformigkeíten) and the reflection according to which he can set his thinking correctly/ with its help set his thinking correctly. (MS 153a, from 1931; also in MS 112, 225, from 11.1931)

The reader will recognize Wittgenstein's writings as a mirrar only if he sees them as an expression of his own tendency, as an expression of how he (and not only Wittgenstein) is inclined to think:

We can bring someone away from a mistake only when he recognises that this is the expression of his feelings / ... if he (really) recognises this expression as the correct expression of his feelings. (MS 110, 230)

A kind of key should be given to the reader to enable him to get out of the philosophical problem (MS 110, 98). 'Get out' means to give up the

106 Wittgenstein's Philosophícal Development

problem as a problem; to understand that his problem is not a real one, but the result of a confusion that takes place already when the problem is first formulated: "Problems are literally dissolved like a lump of sugar in water11 (MS llO, 99; BT, 421). The key is, of course, the correct expression of the train of thoughts and the analogies that bring us to puzzlement. The reader recognízes himself ín what is presented to him by Wittgenstein and so unlocks the doar, i.e. he changes his attitude (does something different):

As I have often said, philosophy does not lead me to any renun­ciation, since 1 do not abstain from saying something, but rather abandon a certain combination of: words as senseless. ln another sense, however, philosophy does require a resignation, but one of feeling, not of intellect. And maybe that is what makes it so diffi­cult for many. It can be difficult not to use an expression, just as it is to hold back tears, or an outburst of rage. (MS 110, 189-90; BT, 406)

The 'difficulty of philosophy' finds expression in Wittgenstein's own thoughts. So severa! of his remarks in his later writings are confessional. Wittgenstein himself was tempted by many thoughts expressed by interlocutors from the Pl. He himself in his middle period was tempted by severa! of the moves that he blocks in this book (especially in 1930, as we have seen in MSS 107-8).

Wittgenstein in 1930-31 is clear about where to find the genesis of philosophical confusions, namely, in philosophical misleading trains of thought and false analogies. He is not completely clear, however, about how to explain precisely the mistakes in the trains of thought concerning rules of projection, facts, intending, and understanding. ln MS 110, 95, Wittgenstein explicitly says that he "cannot find the central grammatical mistake11 that underlies his worries concerning the infi­nite regress:

What we want is to clearly express [analyse] the grammar of the expression 'the command is followed.'

"Yes, but how do I know, then, that I have followed the command?11

(I cannot find the central grammatical mistake on which ali these problems rest).

Naturally, it is the sarne question as this one: How do 1 know that this sentence describes this fact. (MS 110, 95)

The Invention of the Genetic Method 107

Note that in this passage Wittgenstein is obviously aware that diffi· culties here are merely a version of the old problems related to the specificity of sentence and fact. Confusion is present right when the philosophical question is first asked:

When l say "l talk about the fulfillment of a sentence ín general terms,11

then l mean that l talk with words that are not produced/made for this specífic/specíal occasion. (MS 110, 247)

Wittgenstein thinks that he needs to "keep away false interpretations of this senter,ce11

1 i.e., "to exclude false comparisons that push their way (sich zudraéngen)." (MS 110, 247). With such a sentence, one has the impression that what one expects is the fulfillment of the expectation, as if it was a unique thing, simple or complex, that was expected (only the thing that we can point to). However, when one expects someone to arrive, "the expected is not the fulfillment, but that he has come" (MS 110, 248). That is, what we expect is the fact, and nota thing. What, thus, generates the wrong question concerning the specificity of the descrip· tion is the collapsing of fact and complex, a false comparison between fact and complex (MS 110, 249). One thinks that a fact is described in a "general way" and that the fact fulfils the description, as if it was a complex object that could amorphously fill an empty space. This is, of course, a clear reference to the analogy of hollow form (sentence) and full form (fact), which was indicated in Section 2.3. One compares, in this train of thought, a "fact with a house or another complex" (MS 110, 249). A fact, however, is nota complex object, for "l say of a complex that it moves from one place to the other, but not of a fact" (MS 110, 249). A complex is a spatial object, and it is a fact that such an object is here or there. One, thus, points to an object, and not to a fact (MS 110, 251).41

This false comparison is not the only one that underlies Wittgenstein's puzzles in 1930. ln one passage he points out the analogy between thinking and the workings of a mechanism that might take place when one uses expressions like "expecting p" and "thinking p":

One is (by means of false grammar) tempted to ask: how does one think the sentence p, how does one expect that this and this take place (how does one do this) [?] And this false question contains already the whole difficulty in nuce.

The false comparison consists, in this case, in the fact that we think/represent the thing as a mechanism whose externai part we

108 Wittgenstein's Philosophical Development

know and whose interna! working is still hidden from us. (MS 109, 174)

Of course, if one asks "How do 1 do this?11 one is tempted to think that introspection will reveal the nature of thinking and expecting, and how they differ. Then it is almost natural to think that a feeling underlies the difference between "expecting p 11 and "thinking p" (this could have been Russell's train of thought). Wittgenstein finds another misleading analogy at the beginning of MS 110:

The first thing that we want to say about the thought is that it is an activity. A comparison that immediately forces upon us is the one with the digestion. (MS 110, 1)42

Here, of course, we are using a picture:

We simply employ a picture if we talk about the activity of thought. Thinking is not to be compared wíth the workings of a mechanism,

which we see from the outside and whose inside we must however gaze at in arder to understand its activity. (MS 110, 4)

The mind, like a machine, one might think, produces thoughts by means of its activity. This tempts us to think in terms of a mechanism, whose real workings are concealed: "it is like a mechanism, but we cannot see its inside" (a very peculiar mechanism, indeed, for usually we can see the inside of a mechanism).

Wittgenstein returns several times, with different perspectives, to suggestions of false analogies that may have engendered bis tempta­tions and Russell's theories. He mentions, for instance, that what misleads us is the fact that we can have an intention without expressing it (MS 110, 230). This may suggest that we cannot definitely interpret what is said; it may also lead us into thinking that the meaning of a word is a concealed mental representation (MSllO, 230). ln MS 110, 234, he suggests that the false picture is that thoughts must accompany sentences. If we believe that a thought always accompanies sentences, we are tempted to think that we may never reach the thought itself (we get many expressions of thoughts in our sentences, but not the thought itself). On page 236 of the sarne MS the suggested misleading analogy is the idea that understanding is a mental state like toothache. The grounds for such an analogy could be expressions such as "we grasp the meaning ata stroke" (compare with PI, §138).

The Invention of the Genetic Method 109

Given tbe multiple sources (false analogies, misleading trains of tbougbts) of tbe puzzles surrounding Wittgenstein's worries in 1930, it is reasonable to tbink tbat only at a later stage will be fully understand tbe genesis of bis and Russell's confusions, and tben express tbem in a clear way. Tbe task is very complex and pbilosopbical puzzlements bave many sources. Moreover, Wittgenstein soon realizes tbat puzzles, 'grammatical problems/ are embedded in "immensely diverse associa­tions" (MS 113, 23v). However, tbe task later, as is to be sbown in tbe following cbapters, is tbe task of tbe genetic metbod, namely, clearing misunderstandings away until complete clarity is acbieved. ln tbe BB, for instance, all tbe misleading analogies mentioned above are discussed (see Cbapter 4). Note tbat complete clarity, as Wittgenstein says later, "simply means tbat tbe pbilosopbical problems sbould completely disappear" (PI §133). ln arder to acbieve tbis goal, "a metbod is now demonstrated by examples" (PI § 133; from MS 112, 47v; also BT, 431). Tbe metbod, even tbougb it bas more elements later, is tbe genetic metbod (some of tbose elements will be introduced in tbe next cbapters).

lf tbe reading of MSS 109-11 presented bere is correct, Wittgenstein's genetic metbod can be summarized in five points: 1) Wittgenstein bas tbe suspicion, grounded on coincidences between tbe puzzles tbat arise in bis and Russell's accounts, tbat pbilosopbical problems are mainly tbe product of misunderstandings grounded on false analogies, compar­isons, or "pictures," and misleading trains of tbougbt; 2) Wittgenstein's task is to indicate tbese false analogies or pictures and clearly express tbese trains of tbougbt ("portraying tbe pbysiognomy of errors"); 3) tbis task is only successful if be can make tbe pbilosopbical reader acknowl­edge ber own tbougbts and inclinations expressed in bis indication of analogies and descriptions of trains of tbougbt (mirrar metapbor); 4) Since Wittgenstein is investigating tbe genesis of pbilosopbical puzzle­ment, be does not intend to offer or discuss developed doctrines. His metbod works at a pre-tbeoretical level;43 S) Tbe point of bis metbod is to dissolve - in contrast to solving - pbilosopbical problems.

It sbould be clear by now tbat Wittgenstein is not inventlng tbe so-called 'rule following problem' as a variation of a skeptical argu­ment - as Kripke argued (1982). Tbe problem comes more or less naturally into existence. lt is precisely tbis naturalness tbat needs scru­tiny, according to Wittgenstein. Tbe rule-followlng problem, like any pbilosopbical problem, sbould be dissolved by tbe genetic metbod: "Problems are Iiterally dissolved - like a lump of sugar in water" (MS 110, 99). If Kripke bad investigated MSS 109 and 110, 1 tbink, be would

110 Wittgenstein's Philosophical Development

agree that Wittgenstein, right from the beginning, was not interested in answering 'rule-skepticism,' but in finding out what brings one to formulate such a puzzle. The real issue behind the 'paradox' in the PI is 11where did these strange ideas come from?" (PI, §194). It is the issue of how come we, while philosophizing, are prone to 11draw[ing] the oddest conclusions" (PI, §194)- one could add 11such as Kripke's". The problem of the determination of rules, of how we follow specific rules, is a ques­tion that "contains a mistake" (PI, §188). Therefore: we have to investi­gate how a "mode of expression suggests itself to us" probably because it is "a result of the crossing of different pictures" (PI, §191).

lt is clear, thus, that the later rule-following considerations don't express the 'new kind of skepticism' that Kripké attributes to Wittgenstein. However, it should also be clear that those later consid­erations are important and, perhaps, connect many aspects of the later work because they are related with Wittgenstein's temptations in 1930-1, the invention of the calculus conception, and the invention of a new method.

After MS 110, Wittgenstein reevaluates some ofhis views from1929-30. This has become known as the Wiederaufnahme (Kienzler, 1997 and 2001). Examples of the Wiederaufnahme can be found for instance in MS 111: 28-31, MS 112: 238-70, MS 113: 128-284 and MS 114: 1-56. ln those passages, Wittgenstein goes back and criticizes §§1-117 from TS 208 (which is the first version of TS 209, the PR). This self-critique of views held during especially the years 1929 and 1930 is certainly important, for he applies his method to his prior, yet after-1929 views. The major points of this chapter agree with Kienzler's discovery of the Wiederaufnahme, but they also explain haw Wittgenstein got to his change in perspective and why he looks back to his old ideas.

l won't analyze the Wiederaufnahme here. However, 1 need to point out an important issue related to it that escapes Kienzler's analysis. What Wittgenstein looks at right before beginning the Wiederaufnahme is the inexactitude of sense data, its ambiguity, etc. (see Chapter 1, Section 1.5.4). Note that the "biggest difficulty" in establishing a phenomeno­logical language was the 11impossibility of the demarcation of the inex­actness" in the visual space (MS 107: 29-30, 160-7). ln MS 111, he takes into account the use of the word "approximately" in ordinary language and concludes that further specifications of the word, i.e. establishing precise limits for it, are not needed. It is a characteristic of ordinary language "that I am not in a position to give 'precise' limits" when 1 use the word 'approximately"' (MS 111, 21). Wittgenstein argues, then, that like in 11heap of sand,'' the word 'approximately' belongs to the

The lnvention of the Genetic Method 111

11geometry of our space" (MS 111, 24), i.e., it is not only the visual space that is inexact, for such a characterízation can be used also when we talk about positions in the physical space in ordinary language. This brings Wittgenstein to the insight that a precise definition of 'plant' in sentences like "the ground was completely covered by plants'' would not be helpful. ln fact, 11with such a definition we wouldn't be able to under­stand each other better in ordinary cases"; it seems that in such cases "in a certain sense it is the undefined that belongs to our language" (MS 111, 82). This indicates that we may not find strict and fixed rules in all fields of our existing language (see BT: 136-141). The calculus concep­tion, thus, allows concepts that don't have precise limits.

/

2.5 A note on the first project with Waismann

With the explanation of the formulation of Wittgenstein's calculus conception and the genetic method given in this chapter, one can easily understand why Wittgenstein distances himself from the first project of presenting bis views in collaboration with Waismann. Wittgenstein had agreed, in 1930 or earlier, with the publication of a kind of new version of the T in the series of publications of the Vienna Circle, the Schriften zur wissenschaftlichen Weltauffassung.44 Waismann would present the basic ideas of the T in a small book in arder to further the understanding of Wittgenstein's philosophy. The closest that Waismann got to the accom­plishment of the task is an appendix printed in WVC called "Theses," where the T and some relevam changes in Wittgenstein's philosophy from 1929-30 are presented. All the changes are related to the efforts made by Wittgenstein after his acceptance of non truth-functlonal necessary relationships. The most obvious changes introduced there are the idea of a "system" of propositions (of colors, for instance, WVC, 260), the idea that "the sense of a proposition is the way it is verified" (WVC, 244), that analysis leads us to the path of the verification of a sentence, that elementary propositions describe phenomena (WVC, 249), and that sentences of ordinary language are similar to hypotheses (WVC, 256). These changes are tentatively unified in a comprehensive notion of 'syntax' (WVC, 239). ln December 1931 Wittgenstein stopped this project. His reasons are given in a meeting of the Vienna Circle on the 9th:

As regards your Theses, l once wrote, If therewere theses in philosophy, they would have to be such that they do not give rise to disputes. For they would have to be put in such a way that everyone would say,

112 Wittgenstein's Philosophical Development

Oh yes, that is of course obvious. As long as there is a possibility of having different opinions and disputing about a question, this indi­cates that things have not yet been expressed clearly enough [ ... ] 1 once wrote, the only correct method of doing philosophy consists in not saying anything and leaving it to another person to make a claim. That is the method 1 now adhere to. (WVC, 183-4)

Wittgenstein says that he now adheres to this method.45 lt expresses Wittgenstein's new approach to philosophical problems, as presented in this chapter. lf his goal is to show how misleading analogies and confusing trains of thought give rise to philosophical problems, he has to abstain from having philosophical opinions. Therefore, he will mainly analyze what other people say or are inclined to say. The point of doing so is to avoid opinions about the solution of philosophical questions or substantive assumptions concerning the dissolution of such problems. The T assumed a given a priori structure of thought and language that is expressed, among other things, in the assumption that there must be logically independent elementary propositions and simples. The assumption of logically independent elementary proposi­tions is now seen as dogmatic (WVC, 182). Wittgenstein's goal now is to avoid such assumptions. Whether this is also his practice after 1931 is a different question. He certainly thought that his calculus concep­tíon was one of the obvious things, trivialities, to be said that would occasion the answer "Oh yes that is of course obvious" (see quotation above). ln the next chapter we will see that this is not the case.

After the meeting mentioned above, Wittgenstein did not gíve up working with Waismann, however. They still planned to write a book together expressing Wittgenstein's new ideas. This new book is the second project that they began and never finished. The new book was intended to present Wittgenstein's new method, it seems, and his new conception of language (a calculus). I discuss the second project in Chapter 4, Section 4.5.

3 The Big Typescript, the Tractatus, Sraffa, and the Antl1ropological View

3.1 What is the Big Typescript (BT)?

The BT (TS 213) is very close to the typewritten TS 212, an extensive collection of remarks written between 1929 and June 1932. Those remarks origínally belong to MSS 105-114, which were first organízed in TSS 208-211 before they were ordered in TS 212. With the exception of small modífícations from 1932-3, the BT (TS 213), in its clean copy, is a final arrangement of TS 212. There is also a copy with Wittgenstein's handwritten remarks on the text and on verso pages. This version of the BT has also rnany remarks crossed out.1

The BT, it seems, was the first version of the book that Wittgenstein intended to write at the time (1932-3), but it was also used as the rough material of a book that he was co-authoring with Waismann. The second project with Waismann began in 1931-2, at the time that Wittgenstein collected remarks in TSS 211-2, and probably ended around 1935-6.2

ln a letter to Watson from 04.1932, Wittgenstein writes that he lacked the power of "condensing" what he had written and did not mind "if someone else published it" (WiC, 199). ln this letter, he must have had the project with Waismann in mind, who would condense the almost 800 pages of the BT with the help of Wittgenstein's dictations and "constant detailed oral communications" (LWGB, letter to Schlick from 06.05.1932). Thus, he had two projects at the time: his own book and the book with Waismann.3

One year later, in a letter from 04.1933, again to Watson, Wittgenstein says: " ... I was busy dictating about 800 pages of my bl. philosophy. They contain all I want to say but very badly said and now I have begun to

113

114 Wittgemteiri's Philosophical Developmerit

rewrite the whole business" (WiC, 208). With "rewriting" Wittgenstein probably meant MS 114 (its second part), MS 140, and the beginning of MS 115 (Rush Rhees published this material as Philosophical Grammar). It seems, therefore, very unlikely that Wittgenstein thought that the BT was ever anything dose to a book ready for publication. Later, when he senta sample of his work to Russell, he did not send the BT, but the BB (which he dictated in 1933-4). A good reason for this, we will see, is the change that begins to take place in his philosophy after clictating the BT.

The BT did not have a title, but it wpuld probably have been named "Philosophical Grammar" if Wittgenstein had developed the project to its conclusion (MS 110, 254; from 1931). This is the titlé of MSS 113-4 from 1932 and ofthe "first revision" ofthe BT (MS 114). The "revisions," í.e., PG (MS 114, 140, and the first part of 115), show dose proximity to the BT, but also indicate substantial changes. Considering that PG partly overlaps with the BT and with the clictation of the BB, it is to be taken as a transitional work.

The significance of the BT lies, first, in the fact that Wittgenstein in 1932-3 saw it as his best collection of remarks since his return to Cambridge. Second, it can be taken as an intermediate philosophy between the PR and the PI: it is the only work where Wittgenstein systematically presents the calculus conception of language, which clearly characterizes a specific, and defensible, philosophy in its own account. Third, it shows the applícation of the genetic method in the context of the calculus conception. Finally, it shows an adaptation of the T that is to be understood with the background of the first systematic critique of the book. This critique is grounded on the calculus concep­tion of language and is quite clifferent from the critique presented in the PI, which is grounded on the genetic method (this last point will be shown in Chapters 4 and 5). Attention to those facts indicates a path for the understanding of subsequent changes in his philosophy.

ln my analysis that follows, I will give priority to the calculus concep­tion and I will not discuss the details of the application of the genetic method in the BT. However, one should not underestimate its use in the BT. lt can be found throughout the whole typescript. ln the BT, the clearing up of misunderstandings of 'grammar' is only achieved with an investigation of the genesis of those problems; otherwise they don't completely clisappear, even if the system of language is understood (as seen in Chapter 2). The genetic method is, as it were, at the service of the calculus conception. The genetic method presupposes that concep­tion, for the former is applied in accord with the basic workings of the

The Big Typescript, Sraffa, and the Anthropological View 115

rules of the calculus and used when the understanding of the workings of language is not sufficient to eradicate philosophical puzzlement. This presupposition is a major reason why the analysis of the autonomous calculus must have priority in our analysis here. Another reason for this priority is that the calculus conception is the necessary background for us to understand the critique of the T (Section 2.2) and further develop­ments in Wittgenstein's philosophy (Sections 3.4 and 3.5).4 1 will present many aspects of the calculus conception of language in what follows. However, 1 will presuppose the results of Chapters 1 and 2, especially Sections 2.3 and 2.4 of Chapter 2.

/

3.2 The Tractatus revisited and language as an autonomous calculus

ln 1931, Wittgenstein also had in mind calling his book "Philosophical Reflectíons Alphabetically Ordered/Gathered Accordíng to Their Objects/ Themes - Ordered According to Headwords" (MS 154, lr). ln the BT, divisions and subdivisions are not alphabetically ordered, but gathered according to main themes: psychological concepts, logic, mathematics, and phenomenology. This is the old idea of a "book of grammar, 11 this time as a unifying "pure calculus" (MS 154, 9r), in which headwords (divisions and subdivisions) indicate the topics handled.

The divisions and subdivisions of the BT suggest interesting aspects of Wittgenstein's general goals at the time. Divisions ("chapters") like "Understanding," "Meaning," 11Thought," "lntention and Depiction," are directly related to his struggles with Russell and the development of a new method in 1930-1. Other divisions, however, show that Wittgenstein's 'grammar' in the context of the calculus conception of language was intended to be "the complete ledger book of language" (BT, 526), in which the broad topics of 'grammar' are represented: 11Proposition," "Logical Inference," "Generality," "Phenomenology,11

"Mathematical Proof,11 etc. ln those divisions Wittgenstein seems to be giving an explanation of the nature of necessity grounded in the calculus conception of language.

The organizing idea in the BT is that depicting language as a calculus gives usa perspicuous representation of 'grammar.' Language, calculus, and system of rules are taken to be equivalents in the BT:

What holds for the word "language" must also hold for the expres­sion "system of rules. 11 And therefore also for the word "calculus." (BT, 66)

116 Wittgenstein's Philosophical Development

The genetic method and 'grammar' are united in the BT by means of the old idea of a perspicuous representation of 'grammar' modified by the notion of 'calculus.'5 The genetic method is explaíned in a division called "Philosophy. 11 ln the next subdivision, it becomes dear how the genetic method and the old idea of 'grammar' are united in the calculus conception of language. There, the idea of a perspicuous representation of the PR is reintroduced:

The proposition is completely logically analyzed whose grammar has been completely clarifíed. No matter how it is expressed in wríting or in speech.

Above all, our grammar is lacking in perspicuity (Uebersichtlichkeit). (BT, 417; modified translation; already in PR §1; originally in MS 108, 88-9, from 02.1930)

What lacks perspicuity must be elucidated: "To elucidate grammar means to bring it in the form of a game with rules" (MS 113, 24r; from 12.1931). One supposedly tabulates the rules of the calculus, the 'gram­matical game' (BT, 77). 'Perspícuous representation' in the BT consists in making clear the limits of sense by means of the rules of "our grammar" (in all its subdivisions) and in pointing out how and where language may mislead us. lt indicates, thus, the areas of the comprehensive 'grammar' in which the genetíc method is applied and how language works as a calculus (game). As in the 'grammar' of PR, the 'mies of language' drive the investigation:

The description of a new notation, say one that is more perspicuous (perspicuity is our concem), is like the description of one of those languages that children invent or leam from each other, ln which for example each vowel of our ordinary words/in our ordinary language is doubled and a 'b' is inserted between the two parts or the doubling. Here we have come very dose to a game. One can take such a descrip­tion or list of rules as the definiens of the name of a language or a game. (BT 245; modified translation; my emphasis)

The "new notation" that Wittgenstein has in mind here is not a complete logical notation, as in the T, nor a phenomenological comple­ment, as in 1929 (Chapter 1). We might introduce partia! notations, for instance, the notation for the elimlnation of identity (BT, 244) and the color octahedron (BT, 441). However, like in PR, the "perspic­uous symbolism" (BT, 3), the notation of 'grammar,' is a description in the sense of being a list of rules, and not a set of descriptions of

f, l:

f,

í:

i: f1

1

" e

1

e ç 1 t 2 t r i-

e

t ]

t

e

The Big Typescript, Sraffa, and the Anthropological View 117

facts, propositions (BT, 245). Language is understood as the existing language, "as a temporal phenomenon" (MS 110, 152) and "grammar is the description of language" (MS 110, 110; BT, 192v).

This ideais the focus of an early critique of Sraffa. ln TS 212, 1108, the forerunner of the BT, the critique is presented:

[Sraffa] Is grammar ... only the description of the actual transactions of language? Languages? So that its sentences could be grasped as sentences of a natural science?

This remark was first written in MS 109, 281, from 01.31.1931, i.e., right after Wittgenstein began meeting Sraffa regularly at the end of 1930. ln this MS Sraffa is not mentioned the sarne happens ln TS 211 and in the BT, 408. ln the BT, however, a whole subdivision is related to this issue (BT: 240-7). The critique was certainly important for Wittgenstein, for already in 1931 he includes Sraffa in a list of influences (MS 154, 16r, from 1931; see Von Wright 1980, 45). This partially explains why Wittgenstein says that Sraffa's criticai discussions of his thoughts went on for "many years" (PI, preface). The topic was further discussed in 1932, as a note written by Sraffa to Wittgenstein seems to attest (WiC, 196). Unfortunately, there is no further material directly related to their discussíons ín 1932-3, a period in which a decisive step ín Wittgenstein's philosophy begins to take place due to the criticai stimulus of Sraffa. ln 1934, Wittgenstein decides to stop their meetings with the allegation that he had "learnt an enormous amount" from Sraffa 11durtng the last 2 or 3 years," but that he had "leamt most of what at present [1934] can be assimilated.'1 Wittgenstein also says in the sarne letter: 111 have just now very little or practically no power to spare, my work takes up what 1 have" (WiC, 222). 1934 is the year in which Wittgenstein introduces an "anthropological view" in his philosophy (1 discuss this in Section 3.4).

The point of Sraffa's early critique quoted above is that if we take seri­ously Wittgenstein's supposedly purely descriptive standpoint in the BT, 'grammar' must be a kind of natural science and grammatical descrip­tions merely descriptions of facts. ln a note to Wittgenstein from early 1932, Sraffa writes: "lf the rules of language can be constructed only by observation, there never can be any nonsense said. This identifies the cause and the meaning of a word" (WiC, 196). The consequence of the descriptivist standpoint, thus, would be that there is no clear-cut distinction between sense and nonsense and, apparently, no bounda­ries between necessary and contingent propositions (l come back to this issue in Section 3.4). 'Grammatical' sentences would be at the levei of a description of facts. ln Wittgenstein's vocabulary in the BT,

118 Wittgenstein's Philosophical Development

'grammar' would not be arbitrary, i.e., empirical facts could justify 'rules of grammar,' since they would be on the levei of empirical descriptions. Moreover, according to Sraffa's early critique, the descriptive standpoint, as it were, assimilates Wittgenstein's descriptions of meaning to Russell's causal theory. Wittgenstein's answer at the time (around 1931-2) is that the discipline of the rules of language, in Sraffa's sense, would not be the science of thought, in which he was interested, but merely the "descrip­tive sdence of speaking" (TS 212, 1108; BT, 408). Wittgenstein argues that there is an essential difference between a description of facts and a 'descrlption' of the rules of a calculus or game. The difference between bis and Sraffa's sense of 'description,' according to Wittgenstein in the BT, is that when we describe the real, existing, language (the "spatial and temporalphenomenon of language"), we talk about it "as we do about the pieces in a chess game, by tabulating rules for it, not by describing its physical properties'' (BT, 71). Nonsense is, then, lack of conformity with such tabulated roles (BT: 176-7, 188, 192, 245). Wittgenstein thus thinks that Sraffa confuses the double meaning of 'description':

What seems to lead us astray here is the ambiguity of the word 'description.' Sometimes we speak of the description of a real house or tree, etc., and sometimes of the description of a shape, a strocture, etc., of a notation, a game. (BT, 245)

The double meaning of description is already used in the T (Chapter 1, Section 1.2). ln the BT, the descriptions of 'grammar' are to be found in the roles present in the actual language. Wittgenstein's point is that rules implicit in language are part of the descriptions of grammar:

ln descrlbing a notation one might say: "ln this book I shall write 'p V q' instead of 'por q' ", and of course that is a complete sentence. But what I want to call a "rule" that might be written "por q. =r. p V q" is not a complete sentence. - What 1 am calling a "rule" must not contain anything about a particular (or even general) time or place for its application, and must not refer to particular people (or people in general); it is to serve only as an instrument of representation. (BT, 246)

So the rules tout court, which are part of the "science of thought" (BT, 408), have no reference to time, place, or people (thus rules are under­stood abstractly). So they are not descriptions of facts (even though one might take them as such). Reference to space, time, and people concems

1

g l

The Big Typescrí pt, Sraffa, and the Anthropological View 119

the application of 'grammar.' Supposedly, grammatical descriptions contain a role and they merely indicate that a rule is the instroment of representation that shall be used when we give the strocture (grammar) of language (MS 109, 210). This answer, as we will see in Section 3.4.3, will not satisfy Wittgenstein after the BT.

The description of the "roles of our language" is important, for Wittgenstein, because we fail to see the system of Ianguage (BT, 257v). Failing to see the system leads to misunderstandings conceming the 'grammar' of language - one could say the broad "logic of our language." ln the BT, as in PR, 'grammar' takes the place of the "logic of our language" (T, preface). Wittgenstein's idea in the BT is that the most substântial assumptions of the T are abandoned, while its task

remains: expressing the limits of sense (now, by means of 'grammar') and disqualifying nonsense (one applies Occam's razor to nonsense, to what is idling in language). The adaptation of the task of T should not assume anything that could be disputed. This should be accomplished by means of trivial remarks leaving "everything as it is" (BT, 418). Reflections should be "homespun 11

:

All reflections can be carried out in a more homespun (hausbacken) way than 1 used to do. And therefore no new words have to be used in philosophy, but the old, ordinary, words suffice/ the old ones suffice. (BT, 420)

The sarne point is made in a MS, where the "homespun" is understood as "grammatical trivialities":

ln my old book the solution of the problems is still presented in a far too Iittle homespun (hausbacken) way; it has still too much the appearance as if discoveries were needed to solve our problems and everything is still too little brought in the form of grammatical trivi­alities in ordinary language. Everything looked still too much like discoveries. (MS 109, 212-13)

Note that in the T "everything is still too little brought in the form of grammatical trivialities, 11 which indicates that the remarks there were at least meant to be trivialities, even though they were, in fact, dogmatic. ln a letter to Schlick, Wittgenstein writes:

Perhaps the main difference between the conception of the book [the T] and my current one is that 1 realized that the analysis of the

120 Wittgensteín's Philosophical Development

sentence is not [like] finding hidden things, but tabulating - the perspicuous representation of - grammar. As a result, falis everything dogmatic that 1 said about 'object,' 'elementary proposition,' etc. If we want to understand the word 'object,' for instance, we look up how the word is really used. (LWGB; Ietter from 11.1931)

Notice that Wittgenstein thinks that he has dropped his Tractarian dogmatism. The old reflections, when freed from the dogmatic stance of the T, can be carried out in a more homespun way (BT, 420). The tabulation of rules of 'grammar' gives us a "perspicuous representa­tion," which replaces the a priori logic of language and _the idea of an implicit connection with the world revealed in analysK Wittgenstein, thus, gives up what had been found wrong in the T (Chapter 1); but he also remedeis, ln essence, the old "solution of the problems" in the new frame of the calculus conception of language, for 'grammar' takes the place of the old Iogic.

3.2.1 Autonomy of 'grammar'

Wittgenstein's 'grammar' is supposed to be homespun, i.e., a set of trivi­alities lacking metaphysical commitments concerning world and mind. What makes this view possible is the idea that the rules of 'grammar' are the rules of an autonomous calculus. The idea of 'grammar' and language as a "calculus'1 (or "game11

) is thus grounded on the idea that the operations of the Ianguage (the workings of words and all kinds of sentences) are determined by fixed rules:

1 view Ianguage and grammar from the point of view of a calculus// in the forro of a calculus/ J as a calculus, i.e. as operating according to fixed rules// as a process that follows fixed rules. (BT, 258; emphasis and all variations in the original)6

Wittgenstein's calculus in the BT is autonomous because "everything is carried out in language" (BT, 383). This means that grammatical rules operate as rules of an autonomous calculus or game: "Language func­tions as Ianguage only in virtue of the rules that we follow in using it, just as a game is a game only by virtue of its rules" (BT, 196). As in a game, supposedly, the rules of 'grammar' can be investigated independ­ently from its application or particular pwpose (BT, 52).7 As we can distinguísh between the rules of a game and what happens when people play the game, we can dístinguish between the rules of 'grammar' and their application or particular purpose (to describe reality, to express

The Big Typescript, Sraffa, and the Anthropological View 121

commands, to influence people's behavior, etc.). When we are dealing with the roles of language ('grammar'), we are dealing with the interpre­tation of signs in language, the conditions of understanding and sense, and not with its application (BT, 43). The roles of 'grammar' prepare any application (BT, 43)i thus, they show what makes the application of sentences and words possible (BT, 143). The task of the 'grammarian' is thus the description of the interpretation of language, i.e. the tabulation of roles before their actual application: "To describe the calculus we have to tabulate the roles, but to apply it we proceed now in accordance with one, now in accordance with another role" (BT, 155).8

Since 'grammar' constitutes the meaning of words (the meaning of a word Is equivalent to its roles), when meanings are discussed in commands, descriptions of facts, etc., they are already 11completely determined within grammar" (BT, 43). Thus, what is relevant in the relation between calculus and world and in the explanation of thought and mental concepts (the relation between language, world, and "the mental") is part of the calculus of language, is part of 'grammar.' Thoughts are grammatically articulated, and not merely unarticulated or amorphous pictures in the mind (they belong to a given grammatical system). The sarne holds for the notions of 'meaning something' and 'understanding,' etc. Wittgenstein argues:

If you ask someone "How do you know that the words of your description capture what you see?", he could answer, for instance, "I mean that by my words". But what is this 11 that'' if it isn't articulated in tum, and is therefore already language? So "I mean that" is not an answer at all. The answer is an explanation of the meaning of the words. (BT, 190)

If the mental experiences or processes of meaning, understanding, etc. are articulated, they can be rendered in the form of an articulated sentence: 11

... understanding begins only with a sentence" (BT, 1). If they are not articulated, they are at best amorphous signs, i.e., signs that don't belong to a system of signs; as such, they are superfluous and arbi­trary.9 An amorphous, thus Humpty-Dumpty arbitrary style, injection of meaning cannot be of interest:

If meaning something is to have any significance or importance for us, then a system of meanings must be assigned to the system of propositions, no matter what sorts of processes meanings might be. (BT, 203)

122 Wíttgenstein's Philosophícal Developme11t

If someone means something or understands a sentence, he can, in prináple, explain its sense: 11 'Understanding1

- by this I mean a correlate of an explanation of sense" (BT, 11). This sense is a sense only by virtue of belonging inside the calculus:

"Sois 'understanding a sentence' like 'having command of a calculus'? And so like: being able to multiply? I believe so." (BT, 143)

Any interpretation of signs is still correlated with an explanation inside the calculus. One explains the sense of a sentence or interpre'ts it by laying out the mies of grammar employed in it (BT 79). This is how what is meant and the intention are also explained (BT: 374, 272-92). What is meant is given in an explanation of meaning, which is a sentence (BT, 2). Such a sentence specifies what is meant because it works in contrast to other sentences inside the system (BT: 2-3, 93-4). If one means some­thing (this or that), one means words of the system in contrast to others; one chooses one alternative already given by the roles of the calculus. The explanation of meaning thus merely indicates the meaning (BT, 34); it gives us the meaning itself, i.e., the roles that determine the place of a word inside 'grammar.'

To check whether an expectation has been fulfilled, one does not leave the calculus: "expectation and event make contact in language" (BT, 371). Suppose that I expect someone called 'Peter' to arrive. What characterizes "I expect him to come"? Wittgenstein argues:

Surely l could pace back and forth without expecting him to come, as well as look at my watch, etc.; so that isn't what characterizes expecting him to come. What does characterize it is given just by these very words. And 'he' means the sarne thing as in the assertion "He is coming", and "is coming" means the sarne thing as in that assertion, and their combination means nothing different. That is to say: one ostensive explanation of the word "he" is good for both propositions. (BT, 373)

The phenomenon of "paáng back and forth" surely does not suffi­áently characterize the expectation, for one can go back and forth for other reasons. This does not mean, however, that we have to postulate an entity "inside the mind" in arder to characterize expectation. What matters is neither those actions nor the "mental activity" (BT, 355); what matters is that expectations are articulated, that they are "the articulated thought" (BT, 355), i.e. the result of a choice ofwords (or also images: BT,

The Big Typescript, Sraffa, and the Anthropological View 123

86) from many contrasting possibilities inside the system of rules (BT, 93).10 If 1 say to myself before Peter arrives "I expect that he's coming" and after he arrives "He has come," the pronoun in both sentences refers to one ostensive explanation, a rule of language that contrasts to others in the sarne system. The sarne rule ostensively explained can be used in both, the expectation and the description of its fulfillment.11

If one compares a proposition with reality, one can describe reality and check the description with the first proposition:

If 1 want to test reality to see whether it agrees with a proposition, then 1 can also do this by describing it and seeing whether the sarne propositlon results. Or: 1 can -in accordance with grammatical rules -translate reality into the Ianguage of the proposition and then cany out the comparison within the domain of language. (BT, 204)

One does not leave the calculus when one checks whether a proposi­tion is true, even though the truth of a proposition is determined by facts. One has simply two propositions inside the domain of language when one wants to establish an agreement between proposition and reality, for one can describe reality first and then compare the descrip­tion with the proposition. Reality is part of the equations of language: "The sentence is as similar to reality as '5' [is] to the expression '2+3111

(MS 110, 221). This means that reality is translatable, i.e., it is already articulated by rules of 'grammar' in the form of a "projective relation­ship" when we make the comparison: "the transposition is from one articulated thing to another" (BT, 190). We thus begin in an existing calculus, in a system of rules to which we are committed:

It is also nonsense/incorrect to say that the agreement (and lack of agreement) between proposition and world/reality is created arbi­trarily by coordinatíng them. For how is this coordination to be expressed? It consists of the proposition "p" saying that this partic­ular thing is the case. But how does this "this particular thing" get expressed? If through another proposition, then we gain nothing in the process; but if through reality, then the latter must already have been articulated - understood - in a particular way. (BT, 189; see also BT, 200)

'The world is grammatically articulated' means that the 'given,' where we begin, is the system of rules of language artlculated with the world. We don't have, thus, 'grammar' on one side and world on the other with

124 Wittgenstein's Philosophical Developme11t

thought in the middle. The contact of world, thought, and language is already given in one system.

The articulation of world, thought and language makes the derivation of the fulfillment of the expectation work, argues Wittgenstein, like the derivation of a new step in a calculation:

It is a step in a calculation that leads from expectation to fulfillment. lndeed the calculation 25 x 25

50 125 is related to the result 625 exactly as

expectation is to fulfillment. And to the extent and only to the extent - that this calculation is a picture of the result~ is expectation also a picture of fulfillment. And the extent to which the result is determined by the calculation - that's the extent to which fulfillment is determined by expectation. (BT, 3 7 6)

ln both cases, in the mathematical derivation and the derivation of the fulfillment of the expectation, derivations are made according to a rule (we act "in accordance with a rule" (BT, 3 78)). What relates the first step (the expectation or 25 x 25) with the second step (the fulfillment of the expectation or 625) is a rule inside the system of 'grammar.' If one expects Peter, it is Peter and not John or Paul that one expects, in the sarne way that the result of the calculation will be one among several numbers inside the numerical system. ln both cases, there is a result that is correct and others that are incorrect (if John arrives instead of Peter, the expectation is, of course, not fulfilled). The implicit rules in the expression of expectation, thus, reduce the number of possible facts that will fulfill it and determine what to expect (the "picture of fulfill­ment"). ln the sarne way, the rules of multiplication will determine one among many possible, contrasting, results.

3.2.2 A calculus without an a priori structure

The BT and the T have the sarne goal: "The goal of philosophy is to erect a wall at the point where language ends anyway" (BT, 425; my emphasis).12

In the BT, this is achieved when we "define the grammatical game, state its rules and leave it at that" (BT, 77). That is, we begin and end in the existing language, and not in a notation: "Philosophy is concerned with existing languages and shouldn't pretend that it has to deal with an abstract language" (BT, 72). In this view, without the requirements of an "abstract language" (names with reference, their meaning, and bounded concepts), we have to accept that some concepts (for instance, 'plant'

The Big Typescript, Sraffa, and the Anthropologícal View 125

and 'ovoid') don't have sharp boundaries and names might not have a clear set of roles as their meanlng (BT, 251-2). We have to accept that certain regions of the existing language may not have the roles that we are looking for:

Let's investigate language with regard to its roles. lf here and there it doesn't have any roles, then that is the result of our investigation. (BT, 254)

Wittgenstein also accepts in the BT that there is no definition of prop­osition (a general form) and that, thus, proposition is a blurred concept (BT, 60). TlÍUs, it is not logic, logical syntax, and the general form of propositions to which Wittgenstein appeals in arder to make explicit the limits of language (as in the T). However, the roles of 'grammar' in the BT and the general form of propositions in the T, have the sarne goal of determining what counts as a proposition:

What a proposition (Satz) is, is determined by grammar. That is, within grammar. (This is also what my 'general form of the proposi­tion' was aiming at). (BT, 77)

The 'fixed rules' ofthe 1calculus11 in the BT, are supposed to tell us "what

is to count as a proposition" (BT 76), í.e. which combinations of words say something and which are nonsense: "Grammatical roles determine the sense of a proposition; and whether a combination of words makes sense (BT, 79)." 'Proposition' or 'sentence' has its system-relative roles, its "determinate grammar" (BT, 62). lt is inside 'grammar,' the structure of the system of language, that a sentence is 11completely analyzed," í.e. its 'grammar' is "completely clarifled 11 (BT, 417).

The mies of 1grammar1 share with tautological roles of language in the T the characteristic of "saying nothing": 11When one adds the rule to the proposition, the sense of the proposition doesn1t change" (BT, 241). ln this passage, Wittgenstein has in mind the rule "lm=the length of the standard meter in Paris" (BT, 241). As with the tautologies in the T, "we could imagine the rule conjoined to every proposition11 (BT, 241).13

The focus on 11existing languages 11 and the idea that language works as a cakulus make the idea of indefinables in a priori logic (the primitive signs of logic) and in logical analysis (the simples) disappear. ln the T, Wittgenstein thought that some roles are arbitrarily determined (they could be in principie different), namely, the roles that connect sign and

126 Wíttgensteín's Philosophical Development

meaning (the object)1 and that other rules, the a priori roles of logic, are not arbitrarily determlned. Common to both kinds of roles in the T is the appeal to indeftnability.

The a priori rules of logic are primitives supposedly given in the essence or "nature of the proposition" (T 5.471). This is expressed, first, in the idea that propositions have a "logical prototype":

lf we tum a constituent of a proposition into a variable, there is a class of propositions all of which are values of the resulting vari­able proposition. ln general, this class too will be dependent on the meaning that our arbitrary conventions have given ~o parts of the original proposition. But if ali the signs in it that have arbitrarily deter­mined meanings are tumed into variables, we shall still get a class of this kind. This one, however, is not dependent on any convention, but solely on the nature ofthe proposition. lt corresponds to a logical form -a logical prototype. (3.315; my emphasis)

The "logical prototype" that depends solely on the nature of the prop­osition could be expressed by real variables (rpx) so that 1p is any n-ary function and x its n arguments. Such a prototype, such a minimal form, is given a priori, independent of any experience:

We portray the thing, the relation, the property, by means of vari­ables and so show that we do not derive these from particular cases that occur to us, but posses them somehow a priori. (NB, 65)

The prototype expressed by variables, the form, is common to ordinary non-quantified and elementary propositions. It is not sufficient, however, to express their unity according to the T. It is also not adequate, of course, to express the unity of language, i.e., the construction of any proposi­tion, for neither quantification nor logical constants are expressed by it. Otherwise, the general form of propositions could be rpx ar even p - "But surely no one is going to call the letter 'p' the general form of proposi­tions" (PI §134). ln order to express this unity in the T, we need elemen­tary instances of the logical prototype and an operation. According to the T, once we have those, we have the basis for the general form of propositions and, thus, a way to generate ali possible complex forms of propositions. Again, it is a primitive sign, a variable (f 4.53), given in the "very nature" of propositions, that gives the unity of language:

One could say that the sole logical constant was what all propositions, by their very nature, had in common with one another. (T 5.47)

The Big Typescript, Sraffa, and the Anthropological Víew 127

ln the presentation (or 'description') of the general form of proposi­tions it is assumed that the logical prototype determines the possible forms of elementary propositions; the general form shows, in its tum, that the operation N can generate all complex forms (T 6). ln this way, according to the T, we have all essential elements of any proposition. Note that nothing here is arbitrarily determined, for the logical proto­type and the operation N are given a priori. They are given, supposedly, "in the very nature of proposition11 (T 3.315). Somehow, Wittgenstein thought, any competent speaker understands implicitly the prototype and the operation. It is that implicit understanding that supposedly explains how there is no proposition whose form cannot be foreseen or constructed (T 4.5). "Logic, 11 Wittgenstein says in the T,

is not the field in which we express what we wish with the help of signs, but rather in which the nature of the absolutely [intrinsically or essentially] necessary signs (die Natur der naturnotwendigen Zeichen) speaks for itself. (T 6.124)

Such a view suggests that the general form of propositions is self-ex­planatory, it speaks for itself. According to the BT, it assumes that when we understand the operation N, we grasp its meaning (or essence, or nature) "all at once" (BT, 159). From the perspective of the BT, the T's notation assumes the myth of self-interpreting signs that supposedly express the implicit knowledge and implicit nature (or essence) of prop­osition, logical forms, and negation (but also simple names, as below). ln the T, it is assumed, for instance, that when the "logical syntax" is given, "then we have already been given all propositions of logic" (T 6.124). This because if function and argument are given, "ali logical constants" are also given (T 5.47). The elucidations of the book lead one to understand a notation, in which the meaning of the primitives is assumed. There is, however, something very misleading in this idea:

That tautology and contradiction say nothing in no way emerges from the T-F schema, but has to be stipulated. And all that the schema does is to make the stipulation of the form simple / to make the form of a general stipulation simple. (BT, 168).

Given the emphasis of the T on the notation that shows the nature of signs, one gets the idea that meaning emerges out of signs. This gives us the wrong idea of what is really going on: "It is as if one could infer from the meaning of negation that --p means p. As if the rules for the negation sign followed from the nature of negation" (BT, 164; my

128 Wittgenstein's Philosophical Development

underliníng).14 However, what we really have here, accordíng to the BT, is not somethíng that emerges from the nature of signs, but a stípulation of a rule: "ln '-p & (--p = p)' all the second part can be is a rule of the game// (BT, 165).15

The myth of self-interpreting signs (of indefinables or primitives) in the T is also expressed in the idea of analysis. Simple names are also primitive signs (T 3.26). The specific real simple signs, however, are not given in the very nature of 'proposition' logical syntax and the nota­tion only show that they must be found in analysis. The meaning of a sign that is not given in the "very nature" of propositions, that is not "intrinsically necessary" (T. 6.124), and yet determines the comparison of proposition and reality, is the result of an arbitrary convention or, as Wittgenstein also says ln the T, "an arbitrary determination" (T 5.473). Arbitrary conventions should present the rules that give the ultimate content of all propositions; they are supposedly the expression of the connection among language, thought, and reality. The arbitrarily deter­mined conventions, "tacit conventions, 11 Wittgenstein thought at the time of the T, "are enormously complicated" (T 4.002). Those conven­tions, which could be revealed by analysis, would show us the simple names in the "calculus with indefinables" of the T (MS 111, 31). Since the central task of the T is to present the a priori structure of language by means of a conceptual notation, the explanation of how the compli­cated underlying conventions work is not the subject of the book.16

However, the assumption was that simple names name simple objects in fully analyzed elementary propositions (Chapter 2, Section 2.2.2). Those names, thought Wittgenstein, could not be further explained or defined (T 3.26); they were supposedly understood in eluàdatory elementary propositions, i.e., propositions that contained those names (T 3.263). They are thus understood because they are already impliàtly known (T 3.263 and 5.5562).

Accordíng to the calculus conception in the BT, however, we are always dealing with defined and undefined terms, but not with indefina­bles (MS 111, 31). One could say that analysis is horizontal and takes into account the whole system of language, and not vertical, i.e., it is not meant to indicate simpler elements, i.e., indefinables supposedly implicit in the sense of propositions. lt is horizontal beca use the meaning of a word is completely exhausted by its rules. The meaning of a word is not an object; neither is it a mental entity; the meaning of a word, according to the BT, is merely its position inside a grammatical system, i.e. the meaning is nothing more than the set of rules that constitute it inside 'grammar.117

The Big Typescript, Sraffa, and the Anthropological View 129

If one would like to find the best way to determine the meaning of simples in the T, one would perhaps think of ostensive definitions (if they are understood as propositions like "This is red"). However, the relation of pointing to an object or attaching a label to it, says Wittgenstein, is only a rule inside a system of rules; as such it "is only meaningful because of the system to which it belongs" (BT, 290), as parts of algebraic formulas in a mathematical system.18

What the idea of both kinds of indefinables assumes in the T, according to Wittgenstein in the BT, is that "language essentially seems to be some­thing that is given a structure and then superimposed on reality" (BT, 54; quoted in full in the next paragraph). The a priori structure of language of the T (its "abstract language" (BT, 72)), however, does not fit the real language (for the reasons given in Chapter 1). Thus, no a priori structure is to be presupposed in the BT. 19

The absence of an a priori structure and, consequently, of indefinables (logical a priori primitives and simple names), bring the "arbitrary deter­minations" and the "intrinsically necessary rules" of the T to the sarne levei of mies. Of course, definitions in logic (also in a scientific theory) and ostensive definitions (or ostensive explanations) are different, but not because there is an essential or in-principle difference between the two: they are only defined differently (BT, 56). It is a misconception, according to the BT, to think that there is a difference in principie:

Rules of grammar that establish a "connection between language and reality," and those that don't. "1 call this calor 'red"' is an example of the first kind, for instance. -"--p=p" is of the second. But there is a misconception about this difference: the difference seems to be one of principie; and language seems to be something that is given a structure and then superimposed on reality. (BT, 54)

Wittgenstein's point here is that "in giving an ostensive explanation of signs one does not leave grammar" (BT 43), for all rules of language "belong to the basic stock of explanations that prepare the calculus" (BT 45). They belong to the interpretation, and not to the application of language. As part of the "basic stock," the interpretation, they are simply rules and are not essentially different from other rules of language, for any rule simply prepares the application of the calculus.

What Wittgenstein's reformulation of the T shows, thus, is a shift conceming the significance of the a priori. The idea of the a priori does not seem to bring any relevant element into the explanation of neces­sary relationships in the calculus of language. Whatmatters are the very

130 Wíttgenstein's Philosophical Development

ideas that mies are the conditions of representation (or instruments of representation) and that, as such, they are not temporal or spatial descriptions. Considering that mies are always system-relative, one could say that we have a relativization of what counts as a necessary proposítion. ln this way, the Tractarian notion of absolutely, purely, a priori mies vanishes and necessity is understood as the expression of rules ín general. zo

lf there is no a priori stmcture of language, the difference between rules related to phenomenology and tmth-functional rules related to logic, one might think, is not a difference of kind either. But what is the difference between them? One wonders, here, whether Wíttgenstein has the means to make such a distinction. Wittgenstein asks:'

And how can we express our feeling that the tmth-functions are more fundamental than what is phenomenological? For, l believe, only in grammar must this be expressed. (MS 109, 120 from 09.1930)

One could think that the difference is between form and content, as Wittgenstein suggests (BT, 114). This distinction, however, seems incompatible wlth Wittgenstein's idea that rules constitute meaning: the meaning of formal and 'real' concepts is given by general rules. But even if there were room for such a distinction in the BT, how would it be "expressed11 ? How is the "feeling11 that there is a difference "expressed"? The difference should be shown in "the book of grammar":

lf grammar were laid in front of us ín the form of a book, ít would not consist (merely) of a series of equivalent items one after another, but would show a different structure. And it is in this structure if l am right that we would have to see also the difference between what is phenomenological and what is not phenomenological. There would be, for instance, a chapter about colors where the use of color-words would be regulated; but somethíng comparable about the words not, or, etc. (the 'logical constants') would not be said in the grammar.

lt would, for instance, emerge from the rules that these last words can be applied in/at/ each sentence (but not the color words). And this 'each' would not have the character of an empirical universality, but of the incontestable (innapellablen) universality of a highest rule of the game. (BT, 113; modified translation; my emphasis)

So what is phenomenological would be regulated, whíle the logical would emerge as applicable to any sentence. Note the Tractarian

The Big Typescript, Sraffa, and the Anthropological View 131

self-explanatory vocabulary ("emerge from the rules" and "show") appearing here. This is a sign, I take it, that things are not clear. The answer indicates that even if rules of 'grammar' are to be taken as auton­omous, the analysis of the application of rules must be part of 'grammar'. The application, in this case, will show a difference.

The metaphor of "the incontestable universality of a highest rule" seems to assume that there are rules of a game that are essential, and that others aren't. Wittgenstein suggests an analogy: one can play chess "without certain pieces, but never without a chess board" (BT, 113). This is, however, false, as Wittgenstein admits in a handwritten remark (BT, 114). And what to say about the difference between truth-functions and propositiorís that make reference to time? Is the difference grounded on the applicability of truth-functions in mathematics? (BT, 114). Wittgenstein does not solve the problem and the subdivision of the BT where those issues are discussed ends with a typed remark "Discuss: The difference between the logic of content and the logic of propositional form in general[ ... ]" (BT, 116).

The distinction between phenomenological and logical rules can only be found in the application of 'grammar,' and not merely in its interpre­tational rules. They are, after all, rules with the sarne status. ln Section 3.3, 1 return to this issue. ln the next section I present the draft of a reading of the T that is compatible with Wittgenstein's works in the middle period.

3.2.3 The Tractatus read afresh (a minimalist reading)

It seems that, in some aspects, the T is not very far from the calculus conception of language in Wittgenstein's own terms. This is because there too language, world, and thought are already given in an articu­lated form (in the harmony between language, thought, and reality, if you wish). One could say that the harmony between language and the world is, already in the T, the articulation of both in one system:

Every picture has to have something in common with the world of what it represents in order to represent [crossed: a picture ofj some­thing in this world. But this means only: The picture and what it represents have their method of projection in common, so to speak. (BT, l 88v)21

ln the T it is already an implicit system of rules with the right multi­plicity that articulates the world by means of a universal method of projection, namely, "the self-contained system" of logic (T 5.4541). lt is

132 Wíttgenstein's Philosophical Development

this system that explains how the logical form is the form of reality (T 2.18). Thought, language, and world are in a sense one because they are all articulated in language by means of the roles of projection (T 4.014). Projections, one must keep in mind, are partly conventional, but they are not made ad hoc; except maybe by philosophers, who tend not to accept the projective relations grounded on logic which are implicit in language. Language is in order as it is, according to the T, for we don't need to add anything to ít, if we perspicuously represent its logic and assume Russell's theory of descriptions as the means for finding the given implicit connection between languag,e and world. Russellian analysis supposedly shows usa way to reach the complicated tacit conventions of language (T 4.002). These conventions guarantee refeí:ence. ln the T Wittgenstein is not explaining how the connection between thought, language, and reality is established (say, by a transcendental subject). The connection is already given in language. Before we begin analysis we are already committed to ali conventions previously made. Thus, the question about the relatlon of priority concerning simple names and simple objects (whether names come fir.st (idealism) or whether objects come first (realísm)) is not answered at all in the T.

Not only reality, but also thought is articulated in language, for a thought is a meaningful proposition (T 4). Whatever the components of thoughts are, they are expressed in a proposition (T 3.2; also letter to Russell from 08.19 .1919 in CL, 124 ). Thought in the T, thus, is not amor­phous, but also an articulated sentence, as in the !ater calculus concep­tion (see Chapter 2, Section 2.2.2).

Peculiar to the articulation of thought and world in language are the underlying indefinables to be discovered by analysis and the suppos­edly a priori structure of language (logic) to be expressed in a perspic­uous truth-functional notation. The meanings of the indefinables, the objects, lie outside language, in spite of their being already articulated in language. The thing Iies outside, but not its projective artículation, which is implicitly given in language by means of tacit conventions. lf meaning were not articulated in language, the discovery of simples by means of analysis could not be accomplished in the application of logic, as Wittgenstein urges in the T. Note that analysis does not 11resolve the sign for a complex in an arbitrary way" (T 3.3442). It is, however, because of this hidden indefinable connection that 'grammar' (logic) is not autonomous in the T; and not, say, because the reality of simples determines thought in a realístic fashion. It also lacks autonomy, from the point of view of the BT, because of the logical indefinables ("the real

The Big Typescript, Sraffa, and the Anthropological View 133

primitive signs" (T 5.461)) assumed there: it is as if, also in this case, meaning were not equivalent to a set of rules.

The comparison between the BT and the T pursued here indicates a way out of the apparent self-defeating (paradoxical) structure of the latter. It is important, first, to distinguish between the actual notation of the T and "propositions" that supposedly elucidate such a notation. The a priori strocture of language expressed in the notation of the T is given by means of a priori rules (T 3.334, 3.343-41, 5.476, 5.512-4). Note that the expressions of roles in the symbolism of the T (identity notation, troth-tables, general form of propositions, and the role expressed in the troth-table: '.'elementary propositions must be logically independent") are not ndnsense, simply because they don't purport to describe anything. Note that the whole notation of the T is an articulation of roles expressed by means of variables (T 4.1272-4, 4.53, 5.24-2). Very early on, contra Russell, Wittgenstein thought that "it is the variables and not the sign of generality that are characteristic of logic" (NB, 11). The notation itself makes no claims; it shows visually, speaks for itself. Moreover, the roles expressed in the notation are, supposedly, special roles: the simplest roles, given in the nature of propositions. Since those roles determine what is sense and nonsense, the limits of language, the T is not self-defeating when Wittgenstein calls his propositions or sentences unsinnig (T 6.54). After all, they cannot be generated by the central rule of the book (the general form of propositions), since they are neither logical nor bipolar propositions (they are not comparable with reality). "He who understands me ... 11 (T 6.54) means "he who understands my notation throws away all my nonsensical propositions. 11 A general claim of self-defeat is, thus, incorrect.

Doesn't Wittgenstein, however, use his nonsensical "propositions" in arder to teach us his conceptual notation? Yes, but we must understand two aspects of nonsensical propositions in arder to see things clearly here. First, if nonsense helps us to have a clear view of the implicit roles of language (roles that are not nonsense), it can only bring us to what we supposedly already know, for any competent speaker must already under­stand and master the simple a priori roles of logic.22 As Wittgenstein says later, when talking about the T: "it was essential that we, in an impor­tant sense, did not want to find out anything new, but wanted only to understand thewell known" (MS 157a, 47; from 02.1937). Everyone has an implicit understanding of the "well known" roles of logic. One only lacks the right notation for them. This notation, however, speaks for itself. Nonsensical elucidations in the book are, thus, in a certain sense,

134 Wittgenstein's Philosophícal Development

superfluous, since they simply bring one to see explícitly in the notation what one implicitly knows.

ln dictations to Moore in 1914 Wittgenstein makes the following point:

Even if there were propositions of the form "M is a thing, 11 they would be superfluous (tautologous) because what this tries to say is something which is already seen when you see 11M. 11 (NDMN, 110)

Note that Wittgenstein himself before the T called some of the nonsense there 'superfluous' or 'tautologous.'23 His use of 'tautology' is, however, even more striking in this passage: "Is it a tautology to say: Language consists of sentences? It seems it is" (NB, 52; from 05.29.1915). ln the Tone reads, "language is the totality of propositions" (T 4.001). This should be readas a superfluous, tautologous, triviality (one that is shown in the general form of propositions). The "propositions" of the T that help us understand logic are, thus, as superfluous as 11 1 is a number" or "there are objects" (T 4.1272). Note that in 4.1272 Wittgenstein writes unsinnig, but does not correct Ogden's translation in which one reads senseless. One already knows 111 is a number" when one understands the form 'number'; one already knows "there are objects" when one understands the existential quantifier. The "propositions" of the T are parasitical upon its notation. No content is added by those "proposi­tions." One must then apply Occam's maxim (T 5.47321 and 3.328) to the T itself, to its superfluous propositions, if one understands the book's notation. One, as ít were, eliminates the parasites.

Second, the superfluous nonsense of the propositions of the T is not necessarily "important nonsense, 11 even though the way Wittgenstein presented his remarks might give such an impression; after ali, as he says later, they are not homespun enough.24 ln the BT, Wittgenstein gives usa clue to the idea of important nonsense:

How can one reinforce the nonsensicality (Unsínnígkeit) of a sentence (say: "This body is extended") by saying "I can't imagine how it could be otherwise"? For can 1 possibly try to imagine it? Doesn't it mean: To say that 1 am imagining it is senseless (sinnlos)? So how then does this transformation from one [piece of] nonsense into another help me? -And why dDes one just say: "I can't imagine how it could be other­wise" and not "l can't imagine what that would be like" -which, after ali, amounts to the sarne thing? Seemingly one discovers something like a tautology, as opposed to a contradiction, in the nonsensical

The Big Typescript, Sraffa, and the Anthropological View 135

sentence (im unsínnigen Satz). But that too is false. One says, as it were, "Yes, it is extended, but how could it be otherwise? So why say it!" It is the sarne tendency that causes us to respond to the sentence "This rod is of a certain length" by saying "Certainlyl" rather than "Nonsense!." (BT, 96; modified translation)25

Note that 'nonsense' is used to characterize a trivial sentence and imag­ining what the sentence says can be taken to be senseless. One thinks that the 'grammatical sentence' is always verified and its negation is never verified (something like tautology and contradiction). This is why we are prone to thinking that a nonsensical sentence is an important truth, insteád of simply saying 'Nonsense!' Wittgenstein himself during his career was tempted to take nonsense to be important. ln the PR, for instance, he says that philosophy, as "the custodian of grammar," grasps the essence of the world in the "rules that exclude nonsensical combina­tions of signs" (PR, §54). This, by all means, gives at least the impression that something important has been said. An example of such a sentence could be "everything is in flux" (PR, §54; BT, 427). The sentence that excludes nonsense, however, is itself senseless or nonsensical.

Beca use his remarks are superfluous trivialities in the T, and not descrip­tions to be compared with reality, Wittgenstein says "Nonsense!" - rather than "Certainly!" or "Senseless!" at the end of the book. It is also clear that the distinction between senseless (Certainlyl) and nonsense has, in fact, a clear use in the T: it distinguishes the real tautologies given in the essence of proposition - for which we have notational criteria - from the expressions that are tautologous merely because they are in prin­cipie superfluous and, supposedly, obvious trivialities. The superfluous remarks are Iike tautologies (they say nothing), but they are not really tautologies (part of the symbolism). The talk conceming the "essence of the world" in the Tis also determined by this distinction: "something about the world must be indicated by the fact that certain combina­tions of symbols [ ... ] are tautologies" (T 6.124). ln them, it is clear that the "nature of the absolutely necessary signs speaks for itself" (T 6.124). The "propositions" of the book help us to see this by indicating the "logical syntax of any sign-language" (T 6.124), i.e. the logical syntax that is presented in the book and its notational improvements (logically elementary propositions, the logical form of propositions, truth-tables, and the elimination of identity). Presumably, those improvements speak for themselves and indicate something about the world. Since thought, language, and world are given in an artículated form, the rules of the notation also express (show) the essence of the world (the logical form

136 Wíttgenstein's Philosophical Development

it shares with language). Note, however, that one can quickly expand the 'trivial facts' that are supposedly grounded on a logical notation and analysis:

The trivial fact (Die triviale Tatsache) that a completely analysed propo­sition contains just as many names as there are thíngs contained in its reference; this fact is an example of the all-embracing representa­tion of the world through language. (NB, 11; my emphasis)

This means that Wittgenstein takes, even the remarks on 'ontology' of the T as trivial. After all, they seem merely to express the condition of reference implicit in Russell's idea of analysis and its ríotation: at the elementary levei, variables must vary over names that really have refer­ence, if analysis is to have an end. Thís expansion of what is 'trivial' is the ultimate hybris of the T.

However, Wittgenstein's nonsense is dífferent from other people's nonsense (say, Russell's). lt is supposedly harmless, for íts nature is strictly elucídatory (T 4.112). Note, again, that nonsense only bríngs us to nota­tional devices that express the essence of proposition and language, which any speaker knows implicítly. Wittgenstein's nonsense is thus harmless because with it "in a certain sense we can talk about formal properties" (T 4.122), i.e., in an elucidatory sense. What Wittgenstein's nonsense does not do is "to assert by means of propositions that such internai properties and relations obtain" (T 4.122). This, however, was precisely what Russell thought philosophy had to do. He defended the view that philosophical propositions have two essential characteristics: a) that they "must be general," and b) that "they must be a priori" (On Scientific Method in Philosophy, 85-6). Someone who understands the notation of the T knows that Russell's scientific philosophy is einfach Unsinn. When we put Russell's criteria for "philosophical propositions" together, we merely get empty tautologies. Thus, only someone who has not grasped the notation shown in the T remains a Russellian meta­physician (see CL, 124). Note that after finishing the T, Wittgenstein thought that he had solved "our problems," i.e. his and Russell's (CL, 111). However, the clear as crystal solution of the T, writes Wittgenstein to Russell, "upsets all our theory of truth, of classes, of numbers and all the rest" (letter from 03.13.1919; CL, 111). Now "our theory" is merely nonsense. lt is, índeed, misleading nonsense, for it distorts the "correct logical point of view" expressed in "our sign-language" (f 4.1213). Our clear as crystal sign-language does all the work that needs to be done by showing the logic of our language (l come back to "clear as crystal" in

The Big Typescript, Sraffa, and the Anthropological View 137

Chapter 5). To a Russellian who still clings to the old problems, suppos­edly, we have to apply "the only strictly correct method" that remains after the T is understood (T 6.53).

Metaphysical (or Russellian) nonsense and Tractarian nonsense are nonsense because they are not descriptions expressible in the notation and are not tautologies. However, metaphysical nonsense, or Russelllan nonsense, cannot be assimilated to Tractarian nonsense: it does not elucidate the logic of our language.

There is, thus, nothing extraordinary about the last remarks of the T (say, in 6.53-4 and 7). There is neither a mystical plot, nor a deep metaphysics! nora deceiving strategy in the T. Wittgenstein could have simply calléd his propositions 'superfluous' or 'trivialities' ('grammat­ical remarks'). Take, as an example of nonsense in the T, the context principie:

If one says: only in the context of a proposition has a word meaning, this means that a word has its function as word only in propositions; and this can be said just as little as (laesst sich ebensowenig sagen wie) that a chair fulfils its task only in a space .... "(MS 107, 229)

The context principie is one of the pseudo-propositions, a superfluous triviality, that cannot be said (as presumably ali the pseudo-propositions of the book). However, it does not need to be understood as a proposi­tion. lt might be taken, as Wittgenstein makes clear, merely as a trivial 'rule of language':

The proposition The word has only meaning in the context of a prop­osition (Satzverband) must also read, naturally, correctly understood, quite differently (naturally, as a rule of language). (MS 110, 225)

One could say, thus, that the context principie is an example in the T of what Wittgensteín calls later 'grammatical remarks.'z6 He could have called the principie a 'rule of language,' or a triviality, instead of 'nonsense' in the T. The advantage of calling it 'nonsense,' as we have seen, is that in doing so Wittgenstein could, in principie, stop us from thinking that rules expressed as pseudo-propositions introduced in the T describe something.

The T, however, is not homespun enough, as we have seen. Wittgenstein's trivialities tumed out being less trivial than he thought they were. Such a story might reappear later in the BT. The talk about homespun and trivial in the BT must make us, readers, attentive to

138 Wittgenstein's Philosophical Development

its possibly misleading character. The fact that Wittgenstein dassifies 'grammatical rules' as trivialities might conceal assumptions that heis making, as in the T.27 'Trivialities,' 'tautologies,' or 'elucidations' might tum into "grave mistakes'' (PI, preface). This fact is aggravated in the BT by a comprehensive notion of 'grammar,' which includes philosophical 'trivialities,' logic, phenomenology, and mathematical equations.

There are some other points concerning the T that can be made clear in line with the interpretational draft that I am presenting here. The symbolism shows the true nature of the a priori logic: simplex sigillum veri (5.4541). The results of the simple truth are the nonsensicality of philosophical propositions (4.003), and the elucidation that which can be said yields propositions of natural science (6.53). "Little is achieved" (T, preface) by solving the problems of philosophy because they are not real problems. The results are achieved when we think about how the "truth itself" is communícated (as said in the preface) in the book. The "truth itself in its entírety" (T 5.5563), i.e. the logic expressed in the T and its philosophical results, is 1communicated, 1 and not said, because it is shown perspicuously in the conceptu;il notation, and not in descrip­tions of facts.

The simple truth is the ultimate expression of the minimalism intro­duced already in the motto of the book by means of Kuernberger's remark: " ... and whatever a man knows, whatever is not mere rambling and roaring that he has heard, can be said in three words" (T, motto).28

ln the T, therefore, what is not communicated in and by the logical notation (for instance, the nonsensicality of philosophical propositions) is only "rambling and roaring".

The nonsensical propositions are only means to teach the a priori symbolism, but not part of it. We can climb the ladder without para­doxical consequences, as we have seen, because rules of logic expressed in the notation of the T that determine the limits of language are nat nonsense. At the end of the book, however, Wíttgenstein asks us to throw away the whale ladder.29 What sense can we make of that? What is Wittgenstein's ultimate resoluteness? He gives us a good reason to throw away even the most relevant rule expressed in the book in a remark written early in 1929:

Can one say: it is possible to construct the general form of proposi­tions as a variable, but one can not use it in a proposition, since in the proposition ea ipso it cannot have ali propositions as arguments, if the proposition should have a sense at ali. The general form of propo­sitions is also one of those devices of philosophy that we dismiss as soon as it has brought its demonstration to an end. (MS 106, 254)

The Big Typescript, Sraffa, and the Anthropologícal View 139

After the general form of propositions accomplishes its results or demonstration (it shows the limits of language, of sense, and how the logic of language can be constructed, and that phllosophical proposi­tions are nonsense), it is also a device to be dismissed. This is because it cannot be used in propositions, for the rule inside a sentence makes this very sentence a piece of nonsense. The general form, the rule, can only be used to draw the limits of language. Thus, once we have climbed the ladder, i.e. reached the conclusion that philosophical propositions are nonsense, we throw away the whole ladder, which includes even the general form. However, we don't do it because the general form is nonsense (again, it is not; it is an a priori rule, a variable). We do it because there is no íurther use for it, because the deed is done. lf we don't throw it away after reading the T, we can only demonstrate to the one who says something metaphysical "that he had failed to give meaning to certain signs in his propositions" (T 6.53). The best, however, is to remain silent (T 7), for we know that in philosophy we don't say anything (T 6).

3.3 'Grammar' and some tensions

One míght well accept that 'grammatical rules' or 'remarks' are trivial, once one accepts the calculus conception and the idea of a comprehen­sive 'grammar.' However, to what extent is Wittgenstein's conception of 'grammar' itself a set of trivialities? ln what follows 1 argue that, as a discipline of sense, 'grammar' is far from being a set of trivialities that no one would dispute; it expresses a disputable philosophical conception, which is neither trivial nor neutral, as he wished.

According to the BT, philosophers might not misunderstand or be puzzled by all the rules of language. Thus, Wittgenstein's grammatical investigations can focus only on rules that are in fact transgressed (BT, 255). ln the BT, the book of 'grammar' (the "account book of language" (BT, 58)) could be written as the "complete account book of language" (BT, 526; see also BT, 115). However, it is not needed as a whole, for we supposedly need only the basic operations of the calculus of language that enables us to elucidate 'grammar' and the rules of language that are, in fact, transgressed.

"To elucidate grammar," writes Wittgenstein, "means to bring it ín the form of a game with rules" (MS 113, 24r; from 12.1931). The philosopher describes the "structure of language" (see BT, 195, marginal comment), expresses grammatical rules implicit in language and, in this way, "creates a clear order" (BT, 416). This structured game gives usa perspicuous representation of language (BT, 414). The game (or calculus) of 'grammar' is the game of the rules of sense. What goes against these

140 Wittgenstein's Philosophical Development

rules is nonsense and cannot be understood. ln the game of chess, for instance, the bishop can move diagonally, as long as there is no other piece in the way of the diagonal movement. Moving the bishop vertí­cally or horizontally is not playing chess. Analogously, in the BT, if a combination of words doesn't agree with the rules of 'grammar,' it is not part of the language. However, what goes against the rules of ordinary grammar is not necessarily nonsense and can expressa correct descrip­tion (for instance, "Three men was workingwhen 1 arrived"). Moore used this ungrammatical sentence when he expressed serious worries about Wittgenstein's understanding of 'grammar' in one of Wittgenstein's lectures (02.26.1932). He even wrote a short paper on the subject in which he says:

" ... he [Wittgenstein] often makes statements of the form 'So & sois a grammatical rule', where the so & so is not the sort of thing you would find in any grammar book as an example of a grammatical rule [for instance: "You can't say 'Three men was working111

] ( ••• ] he is using the expression 'rule of gran;1mar1 in some different sense." (Moore 2007, 202-3)

ln which sense Wittgenstein used 'rule ofgrammar' remained unclear to Moore. Moore had certainly a point, for ungrammatical sentences don't seem to violate Wittgenstein's 'grammar' (as above) and nonsense can very well be grammatical in the ordinary sense. For instance, "Delicious sevens have more lions in their fronts than the prime dinosaurs whose square roots are -1" doesn't seem to make sense, even though it doesn't violate the rules of the grammarian's grammar. Note that one can under­stand sentences that are grammatically ill-formed, while, supposedly, one cannot understand nonsense. These differences make 'grammar' suspiciously look like designating a philosophical conception or thesis. Wittgenstein in the BT, however, tries to approximate 'grammar1 from the grammarians' grammars. The grammarians' grammars, according to the BT, classify words in very rough groups when compared to a philo­sophical 'grammar.' Wittgenstein, in the BT, thinks that heis doing the sarne kind of dassification as the grammarian, but he wants it in more detail:

I could explain things in this way: the calor of this patch is called 'red', the shape 'circle'. And here the words 'color' and 'shape' stand for kinds of application (grammatical rules) and really signify/really are kinds of words, such as 'adjective', 'noun.' One could perfectly

The Big Typescript, Sraffa, and the Anthropological View 141

well introduce 'color-word,' 'shape-word' into (ordinary) grammar to stand alongside the latter terms ... (BT, 32)

Wittgenstein's classification gives us the old idea of types of words from PR.3º Whereas the grammarian ends his work with the classifíca­tion of nouns, adverbs, verbs, and so on, the philosopher-grammarian starts his work with kinds of words that haven't been yet sufficiently classified (according to the purposes of the Wittgensteinian phlloso­pher). However, they could have been so classified, according to the BT, if the interests of the grammarians were those of Wittgenstein: "In general the rules that the philologist totally ignores are the ones that interest us"/ (BT, 413). ln this view, some rules of language are excluded from the grammarian's investigation because of lack of interest, and this supposedly makes both investlgations (the ordinary grammatlcal and the philosophically grammatical) of the sarne kind.

Thus, Wittgenstein thought in the BT that his 'grammar' was merely an extension of the grammarian's grammar; therewas, supposedly, "only an externai differentiation" between them (BT, 413). This is doubtful in at least five different ways, corresponding to five differences between Wittgenstein's 'grammar' and the grammarians' grammars: first, Wittgenstein's 'grammar' is a kind of comprehensive discipline of neces­sary sentences; second, it deals with the limits of sense and assumes the distinction between questions of truth and questions of fact; third, it assumes phenomenological verificationism; fourth, it is arbitrary in a non-trivial sense; fifth, it is taken to be a necessary condition for something to be a language and, as such, used to oppose Russell's causal theory of meaning. I deal with those characteristics below in topics a, b, e, d, and e respectively. My point in what follows is not that Wittgenstein must have used 'grammar' in the sense of ordinary grammar. The point is, rather, that the differences between both grammars indicate the assumption of non-trivial philosophical conceptions or theses, which is ln tension with the genetic method and, of course, with the idea that 'grammatical descriptions' are trivial. Further reasons why Wittgenstein's idea of 'grammar' is problematic are given in Section 3.4.

a) Wittgenstein gives an interesting list of grammatical rules in the BT, where he obviously intends to deflate the notion:

We need to call to mind how we actually talk about rules in philos­ophy, i.e. when we are clarifyíng grammatical questions; so that we keep our feet on the ground and don't construct castles in the air (nebelhafte Konstruktionen). For example, I set up roles such as: (::lx).

142 Wittgenstein's Philosophical Development

Fx : V: Fa : V: Fb = (3x). Fx or --p=p, or 1 say that it makes no sense to speak about a 'reddish green' or about a 'blackish black', or 1 say that 'a=a' makes no sense, or 1 describe a notation that avoids this formation, as well as '(3x).x=x', or 1 say that it makes no sense to say that something 'seems to seem red/ or that it makes sense to say that in the visual field a curved line is made up of straight pieces, or that it makes equally good sense to say 'The stone falis because it is attracted by the earth' and 'The stone has to fali because it ... by the earth, etc. (BT 243-4; modified translation)

Besides the old rules of the T (identity, quantification, and negation), the interesting aspect of this list is that.it shows that rulés of 'grammar' are nothing dose to rules of ordinary grammars. Of course, the very idea that there is an extended logic or comprehensive 'grammar' (say, implicit ln language) is a disputable philosophical position. It is indeed hard to see how the comprehensive rules of 'grammar' are merely an extension of ordinary grammar of a specific language. It is rather a kind of universal 'grammar' and something more: propositional logic, phenomenology, geometry, and physics. Wittgenstein's 'grammar' is a comprehensive discipline of ali propositions (or rules) that are taken as necessarily true by philosophers.

b) ln order to give the limits of sense and demarcate thus the task of 'grammar,' Wittgenstein needs the idea of 'conditions of sense':

The only thing that doesn't belong to grammar is what consti­tutes the truth and falsehood of a proposition. That's the only thing grammar is not concerned with. Ali conditions of comparison of the proposition with reality/with the facts belong to it. That is, ali condi­tions of understanding. (All conditions of sense). (BT, 43; modified translation)

So it seems that Wittgenstein is suggesting that, on the one hand, some things constitute truth and falsehood; on the other hand, there are conditions that need to be met for propositions to be compared with reality and be judged either true or false. Further, those conditions that allow the comparison are the conditions of understanding, i.e., if we understand a proposition we may be able to decide whether it is true or false. The idea of a structure of language of the T, thus, reappears in 'grammar.' The difference is that it is not strictly logical and is not given in purely a priori terms.

The Big Typescrlpt, Sraffa, and the Anthropological View 143

Such a view assumes a difference in kind between two groups of sentences: sentences that express 'grammatical' roles (related to ques­tions concerning sense) and propositions of language that are regulated by the rules of 'grammar' (related to questions concerning truth). But, then, Wittgenstein is assuming a distinction that is itself a philosophical position, namely, a Carnapian one (think of the distinction between internai and externai questions, for instance). Moreover, the need for a criterion that distinguishes those kinds of propositions arises. Such a criterion must be argued for. But if there is such an argument for a crite­rion to distinguish grammatical sentences from other kinds of sentences, it is suspiciously philosophical (see below).

e) Wittgeilstein's old phenomenological investigation in the style of 1929-30, in PR, as seen in Chapter 1, is still present in the BT. There are severa! reasons for his interest in phenomenology: the mistakes surrounding a phenomenological investigation, Wittgenstein's verifíca­tlonism, and the application of 'grammar.' 1 deal with those reasons in what follows.

First, Wittgenstein investigates the mistakes surrounding phenom­enology in arder to clear up his and other people's mistakes. lf one deals with mistakes surrounding phenomenology, one needs to be clear, presumably, about the phenomenological rules in 'grammar.' Some of the mistakes were already dealt with at the time Wittgenstein was constructing his phenomenological language (for instance, the assump­tion of a subject in the description of visual space (BT, 438)); others, were not, for instance, that "distance is already contained in the structure of the visual field" (BT, 458). Wittgenstein discusses also the problem of a 100-sided figure and tries to show that there is a confusion in the discussion of the problem (BT, 433-4); he discusses the different multi­plicities of visual and physical spaces (BT, 446); he defends the use of a system of coordinates grounded in its simpliàty (BT, 453). Thus, even if Wittgenstein's major motivation in his phenomenology in the BT is in accord with the goals of the genetic method, his practice indicates commitment to old views.

Second, his phenomenological stance is preserved because of his old idea concerning the verification of hypotheses; this idea also finds its place in the BT (BT, 117-21). This is dane in the division called 'Idealism,' and in the subdivision called 'The Nature of Hypotheses.' The verifica­tion principie, which Wittgenstein calls "our old principie" (BT, 126), is present in many passages of the BT and expressed in the following way: "How a propositlon is verified, that's what it says [ ... ] Verification

144 Wittgenstein's Phílosophical Development

isn't merely an indication of the truth; it determines the sense of the proposition" (BT, 266).31

Wittgenstein's phenomenological verificationism in the BT supposedly expresses the truth in ídealism: "the truth in idealism is really that the sense of a proposition (Satz) emerges completely from its verification" (BT, 500).32 ln a MS from 1932, Wittgensteín is even more emphatic:

The hidden truth in idealism was that idealism recognized the essen­tial connection between a statement about the physical world and a statement about our direct experience which is said to support the fírst statement. (MS 155, 4lv; from 1932)

A "hypothesis" such as "There is a book lying here" (BT, 117) indicates whether an "experíence may tally with it or not" (BT, 118). The imme­diate experiences correspond, thus, to the "facets of a hypothesis that are verified" (BT, 120). Each time that a specific "facet of a hypothesis is laid alongside reality, the hypothesis turns into a proposition" (BT, 121), i.e., gets a "Yes/No" answer. A facet of ~he hypothesis "There is a book lying here" would be presumably tested by means of the visual or tactile phenomena related to a book.

The calculus conception, thus, does not rule out the significance of phenomena (the truth of idealism). The meaning of a word is not an object, phenomenological or not; however, our physical statements must refer to phenomena, otherwise they don't have meaning. Thus, even if meaning is not what words refer to, sentences can only have sense because words have a possible reference (BT, 122; quoted below).

The third reason is the application of 'grammar.' Even though Wittgenstein thinks that "grammar is apure calculus (not the applica­tion of a calculus to reality)" (BT, 558), he sees the need to investigate its application. ln Section 3.2.2, we have seen that the appeal to the appli­cation seems unavoidable in order to give plausibilíty to the distinction between rules of logic and phenomenological rules, for they cannot be different in kind (pure rules of 'grammar' are general and contain no reference to time, space and people).

The 'application of language,' i.e., the connection between language and reality, is to be taken, then, as part of 'grammar' (even if ít is not pure grammar). The application of language is, for instance, "the description of what we see" (BT, 42). The rules used in this description are part of 'grammar;' they belong to the interpretation of language, but the actual employment of roles in descriptions is an 'application' (BT: 42v, 53). This, of course, parallels the distinction between questions conceming

1 The Big Typescript, Sraffa, and the A11thropological View 145

understanding (sense) and questions of truth, as above. Wittgenstein argues:

ln grarmnar the application of language is also described - what we would like to call the connection between language and reality. lf it weren't described then on the one hand grarmnar would be incom­plete, and on the other it couldn't be completed from what was described. (BT, 441)

A complete 'grammar' must thus also describe the application of language, the connection between language and reality. We describe this connection'phenomenologically, for our description at the end must refer to phenomena: "[ ... ] signs have meaning in so far and only in so far - as a phenomenon that has been observed corresponds to them or doesn't, no matter how roundabout the way" (BT, 122). ln the passage above (BT, 441), Wittgenstein discusses the fact that "our visual field is in calor." He claims that '11being colored' is contained in the definition of the concept 'visual space,' i.e. in the grammar of the words 'visual space"' (BT, 441). Thus, any visual space description will refer to colors; in arder to do so, we employ, for instance, the rules of the 'grammar' given, before its application, in the octahedron, which is the rough representation of colar space (BT, 441). The reference to colors in the visual space is, supposedly, the application of those rules.

The assumption that a phenomenological vertficationism determines the sense of sentences is, so it seems, dose to Wittgenstein's old concep­tions from the PR and expresses a philosophical conception that is far from trivial. One should also note here that the old arguments for the 'absolute directions' of the visual space are also reintroduced in the BT (Chapter 1, Section 1.3). The absolute direction of the visual field is defended with the old transcendental arguments (BT, 456-61).

d) The whole set of 'grammatical rules' is said to be arbitrary in a non-trivial sense in the BT: it is not possible to justify 'grammar.' ln fact, Wittgenstein reintroduces the "transcendental arguments'' used in the PR to show that rules of 'grammar' cannot be empirically justified, i.e. that they are arbitrary (BT, 238). The old transcendental arguments for the arbitrariness of 'grammar' of PR are even supplemented with the claim that 'grammar' is autonomous and that rules constitute meaning (BT, 234-5). 1 will not discuss the old arguments from PR again here, for they are the sarne arguments discussed in Chapter 1 (Section 1.5.5). One might think that transcendental arguments are the most common philosophical tool in the history of philosophy. To say the least, they

146 Wittgenstei11's Philasaphical Develapment

conceal or affirm disputable philosophical claims or explanations. It is not different in the case of Wittgenstein. The claim that empirical propositions cannot justify 'grammatical rules' might be correct, but it is certainly highly controversial, and not something that no one would dispute.

e) One could expect that the purpose or purposes of our use of language could be employed to justify its 'grammar.' This is not, however, quite Wittgenstein's view in the BT. If we look at the remarks in the clean copy of the BT, the ones that were typed and originally belong to the BT, we find that Wittgenstein defends a strong, position concerning the purpose of games and languages:

Couldn't 1 look at language as a social institution that is subject to certaín rules because otherwíse it wouldn't be effective? But here's the problem: 1 cannot make thís claim; I cannot give any justificatíon of the roles, not even like this. I can only describe them as a game that people play. (BT, 191)

Another passage makes a similar point by emphasizing that 'grammar' is a necessary condition for language:

Is everything 1 can say said by: One cannot say of grammatical rules that they are an instrument that enables language to fulfill its purpose? As one, for instance, says: If a steam engine didn't have a governar, the piston couldn't go back and forth as it is supposed. As if one could imagine a language without grammar. (BT, 193; modifted translation)

Both passages were crossed out (as many other remarks in the BT), which índicates that Wittgenstein was not satisfied with them after a while. One wonders, however, what motivates the claims that one cannot imagine a language without grammar or why one cannot consider the rules of grammar as a mean to achieve certain purposes.

The causal theory of meaning is, once more, a motivating background for Wittgenstein's position. According to the causal theory, the impor­tant aspect of language is what one achieves with it: it is "an instru­ment for the promotíon of purposes" (MoM, 16). Language looks like a useful tool to make people act accordíng to our desires and needs. For Wittgenstein in the BT, where the descriptive purpose of language is not assumed anymore, there are apparently two options: either we look at language as autonomous, or we are committed to the mistakes of

The Big Typescript, Sraffa, and the Anthropological Víew 147

the causal theory when taking it as a mechanism for the promotion of purposes. Language as a mechanism makes Wittgenstein's philosophical task of identifying the limits of sense impossible. lt seems that talking about the rules of language, Umits of sense and the tabulation of roles that make these limits dear excludes the possibility of viewing language in causal terms. Wittgenstein is led to think that the correct point of view is that "grammar is nota mechanism justified by its purpose" (BT, 191). He goes so far as to claim that something is a language only if it has rules of grammar: "Language functions as language only by virtue of the rules we follow in using it, just as a game is a game only by virtue of its rules" (BT 196, also 197; my emphasis). ln this view, 'grammar' is a necessary cóndition for language. Thus, to invent a language one has to set up its rules, to "compose its grammar" (BT, 65). This excludes taking language as a device for particular purposes:

lnventing a language does not mean inventing a device for a partic­ular purpose on the basis of laws of nature, or in agreement with them. (BT, 193)

Wittgenstein thinks that if we don't have rules of grammar, we don't have a language at all: "without grammar it isn't a bad language, but no language" (BT, 194). We cannot, in this view, "imagine a language without a grammar" (BT, 194). This, of course, sounds very much like a disputable philosophical conception, one that Wittgenstein will not accept later (Section 3.4).

Many of Wittgenstein's views related to 'grammar' are associated with the analogy between game/calculus and language or rules of a game and roles of grammar: "Someone describing the game of chess neither lists the properties of the chess men nor does he talk about the usefulness and use of the game of chess11 (BT, 192). We have already seen that Sraffa expressed objections related to such a notion (Section 3.2). Wittgenstein, however, defends the central analogy of his calculus conception with a not-quite-convincing "sure instinct that there is an affínity11 (BT, 239; modified translation). lt does not seem to be clear to Wittgenstein that the "sure instinct" might be grounded on an analogy extended too far (one could say, a false analogy). This, in spite of Wittgenstein being aware of the dangers of using this and other analogies:

(A símile (Gleichnis) is part of our edifice; but we cannot draw any conclusions from it either; it doesn't lead us beyond itself, but must remain standing as an analogy. We can draw no inferences from it. As

148 Wittgenstein's Philosophical Development

when we compare a proposition to a picture [ ... ] or when I compare the application oflanguage with thatofthe calculus ofmultiplication, for instance.

Philosophy simply sets everything out, and neither explains nor deduces anything.) (BT, 418; my emphasis)

The analogy between proposition and picture did not "remain standing as an analogy" in the T: it led Wittgenstein beyond itself. We have seen already that aspects of the analogy between language and calculus are defended as if they were necessary in the BT. Next, we will see in more details that the analogy, indeed, leads Wittgenstein "beyond itself."

3.4 Sraffa's fruitful criticism and the anthropological view

Wittgenstein acknowledges a profound debt to Sraffa in the preface of the PI:

... Iam indebted to that [criticism] which a teacher of this university, Mr P. Sraffa, for many years unceasingly applied to my thoughts. It is to this stimulus that I owe the most fruitful ideas of this book.

There is no doubt that Wittgenstein meant it seriously: an equally emphatic statement can be found in all versions of the preface of the PI (PUKGE: 207-9, 565-8, 741-4). Note that he does not say that Sraffa was one influence among others, but that he owes to the stimulus of Sraffa's criticism "the most fruitful ideas" of PL We will see that Sraffa's criticism helped Wittgenstein to finally break with the project of the T of estab­líshing the limits of sense; this break brought Wittgenstein to an anthro­pological view, which takes the centre of the stage of Wittgenstein's philosophy under the guidance of the genetic method thereafter.

3.4.1 The calculus conception revisited

The calculus conception assumes, as we have seen, that language has a certain complexity; it assumes the central role of sentences and their compositionality in any calculus of language. This, as we have seen (Chapter 2, Section 2.3), grounds Wittgenstein's explanation of how the description of specific facts is possible without an appeal to mental acts. Mental concepts (how one means something specific, etc.) are explained as part of the calculus of language. The correlation of mental contents and sentences of the calculus assumes that they have the sarne

The Bíg Typescript, Sraffa, and the Anthropological View 149

multiplicity as word-language: "One [representation] is as good as the next, as long as the multiplicity is correct; that is, so long as there's a dífference in representation corresponding to every difference in what is represented" (BT, 285). "The thought/' thus, 11is essentially what is expressed by a sentence (Satz)" (BT, 222); thus, to understand a thought is to understand a sentence (BT, 1). If someone understands a sentence, he can explain its sense (BT, 11). One explains the sense of a sentence or interprets it by laying out the rules of 'grammar' employed in it (BT, 79). This is how what is meant and the intention are also explained (BT, 374 and 272-92). What is meant is given in an explanation of meaning, which is a sentence (BT, 2). This explanation must contain a rule, one of the rules oHhe system of 'grammar' (the calculus) that is already given.

A very important feature of the conception of language as a calculus is that whatever is not articulated in a sentential system is not part of language. Because of the calculus conception, Wittgenstein does not take seriously what does not count as an articulated sentence and is correlated with one. The actual learning of language, something that will be tremendously important for him later, is not one of his concems. Since "the learning of language is not contained ln its use" (BT, 175), it "falis out of the present as mere cause, hlstory" (BT, 5). Only the learned rule remains relevant, for it justifies the future employment of words (BT, 199).

Moreover, gestures, and a language of gestures or signs, are only rele­vant as far as they belong to a system that allows back and forth transla­tions with articulated language (MS 110, 127; BT: 9, 176, 183).33 Those languages have, supposedly, an implicit 'grammar' (MS 110, 109-10). Any slgn occupies a position inside a grammatical system in opposition to other signs; therefore, says Wittgenstein, "a sign without grammar cannot exist" and there is a 11grammar of gestures" (MS 110, 126). Even if we imagine a language without words, the language still has a complex structure that allows translation into word-language. It can be explained with a word-language (BT, 186). What we need for our definitions and explanations is only 11a system with the right degree of multiplicity" (BT, 89). Psychologically, a language of gestures may be more primitive (BT 55), but logically (or grammatically) it is not.

The language of children and a language of gestures are languages only to the extent that they are incorporated in a calculus of sentences by means of translation or correlation. They are, of course, translatable into a complex calculus of sentences. Wittgenstein expresses this point by means of language-games. The language-game that Wittgenstein introduces in arder to show how Augustine thinks about language and

150 Wittgensteín's Philosophical Development

"the primitive picture" (BT, 26) guiding it, begins with the sentences "That's a pillar", "That's called slab", etc. (BT, 25). That is, ostensive definitions work essentíally as sentential explanations. They are mles like any other mies. When Wittgenstein talks about restricted languages (and not merely language-games) without propositions, he suggests that we can imagine, for instance, a language of commands in the form of sentences and a language of questíons and Yes or No answers (BT, 209 and 207, respectively). He thus translates the primitive forms into complex combinations of words. This supposedly shows that primitive forms are not the foundations of language. ln fact, Wittgenstein thinks in the BT, they cannot be the prímary ground for language, for, as with sentences, we míght misunderstand gestures or we míght not under­stand them at ali: "We don't understand Chinese gestures any better than Chinese sentences" (BT, 10).

Note that a 1grammar1 of gestures, or the translatability of gestures, according to the calculus conception, avoids taking gestures as funda­mental or primary signs. Without a 'grammar' they would have the status of indefinables. But the very idea of indefinables is precisely what the calculus conception of language is supposed to eradicate: language is a system of rules of 'grammar' that work as contrasting defined or unde­ftned alternatives given in sentences. With an ostensive explanation one "doesn't leave grammar" (BT, 43). ln a language of gestures, we "could replace the ostension with the sign, 11 which delivers a "representation in the form of a definition" (BT, 44). This is also why the language of chil­dren is seen from the perspective of a complex language with 'grammar.' One could say therefore, that even though Wittgenstein supposedly looks at the "existing language" (BT, 72) in his investigation, there is an idealized or "general concept of language" operating in the background. This general concept of language finds its support on the "autonomy" of language or 'grammar.'

While still collecting the remarks for the BT (see Section 3.1), Wittgenstein started reworking most of the material (in fact, all of it, except for the part on the philosophy of mathematics and logic). He crossed out many of the remarks and wrote revisions on the margins of severa] pages and reworked the material in different manuscripts published as part I of PG (MS 114, MS 140, and the beginning of MS 115). The PG, taken as a whole, seems to be written from the perspective of the BT: the emphasis of the project is still on the idea of language as a set of grammatical mies that constitute a calculus. The idea that 'grammar' is what determines the "conditions of sense" still prevails (PG, 88; from MS 114, 58). A sentence, says Wittgenstein in the PG, must be part of

a 5

r g s

E

\

e E

e

1

e l

t

e

t

t s ~

1

f f

The Big Typescrípt, Sraffa, and the Anthropological View 151

a system (calculus): 11A sentence is a sign in a system of signs" (also PG: 55, 128, 149, 152-3) and "the role of a sentence in a calculus is its sense" (PG, 130).34 This means that "what gives a proposition sense is its agree­ment with gramrnatical rules" (PG, 127; see also BT, 81). Likewise, what gives meaning to words, according to the PG, is the role they play in the systern or calculus of language: "The meaning is the role of the word in the calculus" (PG, 63). How a word is used in the calculus is the key to explaining its meaning.

However, as we will see, many new elernents are introduced in the PG. This fact qualifies it as a transitional work between the BT and Wittgenstein's later work: it looks like an effort to make the calculus conception 'cornpatible with the anthropological view.35 Those new elernents and the handwritten rernarks in the BT will show us Sraffa's criticisms and the fruitful ideas that were developed thereafter.

3.4.2 Primitive languages and 'grammar'

The rnost farnous anecdote concerning Wittgenstein and Sraffa is the latter's Neapolitan gesture "of brushing the underneath of his chin with an outward sweep of the finger-tips of one hand" (Malcolrn 1989, 58). Malcolm says that, for Wittgenstein, it showed the "absurdity in the insistence that a proposition and what it describes must have the sarne 'forrn111 (Malcolrn 1989, 58). lt is very unlikely that this is the correct description of what Wittgenstein had in mind. Wittgenstein had given up the idea of a comrnon form of what is described and the proposi­tion that describes it already in 1929, when he gave up the "analysis of phenornena" in the construction of a phenomenological language (see Chapter 1). The comrnon form expressed in the notation of the T, i.e., logical forrn, was insufficient to explain the logic of our language (as seen in Chapter 1). Also the notation that would show the specific forms (space, color, and time), as we have seen, did not work. Sraffa thus did not have to tell Wittgenstein that that view was wrong, since he already thought so by 1929. ln fact, Wittgenstein does not mention Sraffa at all in his MSS of 1929-30; moreover, when he began meeting Sraffa regularly he was certainly already assuming a cornprehensive 'grammar' while thinking inside the framework of the calculus concep­tion of language.36

Malcolm also talks about 'logical multiplicity' and connects it to the T. This is not quite correct, for a kind of multiplicity requirement is still in place in the calculus conception of language of the BT, while the logical forrn requirement was already abandoned before the PR. However, there is sorne truth in Malcolm's claim related to multiplicity (as we will see).

152 Wittgenstein's Philosophical Development

The anecdote of the gesture has also a different version. ln the sarne page where Malcolm mentions it he gives, in a footnote, von Wright's version:

The question at issue, according to Wittgenstein, was whether every proposition must have a 'grammar', and Sraffa asked Wittgenstein what the 'grammar' of that gesture was. ln describing the incident to von Wright, Wittgenstein did not mention the phrases 'logical forro' or 'logical multiplícity'. (Malcolm 1989, 58; footnote 3)

This version of the story fíts better severa! aspects of Wittgenstein's development, as we will see, and was. confirmed by Sraffa himself in a conversation and in a letter. Alessandro Roncaglia, a Sraffa-scholar, writes:

ln a conversation (21 December 1973) Sraffa confirmed the anecdote, telling me that von Wright was right. The correctness of von Wright's interpretation is also confírmed in a letter by Sraffa, dated 23 October 1974 (now in the Sraffa Papers, C 303) quoted in Bellofiore and Potier (1998: 73). (Roncaglía 2009, 27)

The mentioned paper has now an English translation where we find the following:

The correspondence preserved at the Wren Library in this respect [the Malcolm/Von Wright versions] gives far-reaching evidence from Sraffa. lndeed in a letter from 22nd October 1974, Ulrich Steinvorth asks the ltalian economist about this point (SP-C 303), after a meeting on 16 October to which he got a verbal answer. Sraffa annotates two times in the margin, "I had not then nor have since read the Tractatus"; and he peremptorily asserts that the word criticized by him was not 'logical form' but 'grammar' - therefore supporting von Wright and dissenting from Malcolm. (Bellofiore and Potier 2011, 342; my emphasis)

Note that also in the early critique by Sraffa, discussed in Section 3.2, the issue was 'grammar' (TS 212, 1108). Von Wright's version also conforms to what Wittgenstein himself says in 1937:

It showed itself that I did not have a general concept of sentence and language. 1 had to recognize this and this as signs (Sraffa) and could not, however, give a grammar for them. Understanding and knowing the roles. (MS, 157b, Sv; my emphasis)37

\

\

t

1

The Big Typescript, Sraffa, and the Anthropological View 153

Wittgenstein is not specifically concemed with a Tractarian logical form in this passage, as Malcolm suggests, but with 'grammar,' as claimed von Wright. He is talking about a "general concept of sentence and language, 11 which is, however, related to the general form of proposi­tions (Saetze) in the T and to the idea of 'grammar' as the structure of language. Sraffa's criticism is not a critique of the idea of logical form, but it is also a critique of the T, where a stronger version (a completely a priori version) of his later 'grammar' was defended. There, he even offered a definition of what counts as a proposition and hence what counts as language: "The totality of propositions is language" (T 4.001). ln the PR, the sarne idea of equivalence between language and totality of propositidns is still expressed in the context of 'grammar': "language means the totality of propositions" (PR §85). But Wittgenstein explicitly says in the BT, in a mil der form, that "what a proposition is is determined by grammar"; this, he makes clear, was also what the "'general form of the proposition' was aiming at" (BT, 77). Sentences, we have seen, are fundamental in the calculus conception, for they are the complex units that explain how the multiplicity of language accounts for the descrip­tion of a specific fact and explains the use of mental concepts (Section 3.2.1). Sraffa's criticisms were, then, broad and directed at an aspect of the T that survived in the BT: the idea that language must be a system of fixed rules that function as a calculus whose limits are given by mean­ingful sentences.

Why, then, could not Wittgenstein give a 'grammar' to Sraffa's signs? As will be shown, what was at issue was not Sraffa's isolated gesture (its translation or its underlying 'rule'), but its possible surround­ings: languages of gestures as well as many primitive languages taken in isolation don't fit into the central role that Wittgenstein ascribed to sentences in the logical calculus of the T and of its revised grammar-version in the BT.

That primitive languages don't fit the conception of 'grammar' of the BT can be seen, first, in the conception of 'understanding' parallel to 'explanation' that underlies the explanation of mental concepts (Section 3.2.1). Sentences have a central role in the calculus (the system of gram­matical rules), for, according to the calculus conception, "the under­standing begins only with a sentence" (BT, 1).38 Therefore, Wittgenstein thinks in the BT, one does not have to be interested in the unarticu­lated (or 11amorphous11

1 as he also says) understanding of supposedly causal relations or mental images (as Russell was). The amorphous might be ignored because it does not have the multiplicity of a contrasting alternative in the system of rules of 'grammar'. Sentences are the corre­late of explanations of sense and, therefore, play a central role in the

154 Wittgenstein's Philosophica/ Development

explanation of 'understanding':

Understanding correlates with explanation; and in so far as it doesn't, it is unarticulated and therefore does not interest us; or it is articu­lated and correlates with the sentence itself, whose sense we want to render. (BT, 11)39

When Wittgenstein discusses understanding and explanation in a handwritten remark added to the BT, he makes a direct reference to Sraffa's Neapolitan "derisive gesture." The long marginal remark is preceded by the indication "To 'Leaming the Language"' and should be inserted after the sentence "We don't understand Chinese gestures any better than Chinese sentences":

[fhat is, the failure to understand isn't limited to sentences. For how do we leam the language of foreígn gestures? They can be explained to us in words. We can be told "Among these people this is a deri­sive gesture (eine hoehnische Gebaerde)", etc. Or, on the other hand, we learn to understand these gestures the way we learned as chíldren to under­stand the gestures and facial expressions of grown-ups - wíthout explana­tion. And in this sense learning to understand does not mean learning to explain, and so we understand the fadai expression, but can't explain it by any other means.] (BT, 10; handwritten remark; my emphasis)

This is the sole passage where Wittgenstein directly mentions Sraffa's "derisive gesture11 in an appropriate context. Note that its context suggests that several concepts are related to the gesture: form of lífe ("among these people"), primitive or simple languages (children's language), understanding, meaning, explanation and leaming. Note also that heis clearly denying his claim in the BT according to which "under­standing correlates with an explanation'' (BT, 11). In fact the heading of subdivision 3 of the BT was called "Understanding as a Correlate of an Explanation" (BT, 11).

An ltalian can explain to us that Sraffa's derisive gesture could be translated as "I don't care about what you have just said." Gestures can be translated into words (say, contemporary English). This translation would, then, give those gestures a 'grammar' in the sense of the BT. This is not, however, Wittgenstein's poínt, as he himself makes clear. The point is, first, that "the failure to understand isn't limited to sentences" (for one may not understand Sraffa's derisive gesture); second, that "learning to understand does not mean learning to explain"; third, and

1

1

The Big Typescript1 Sraffa, and the Anthropologícal View 155

this is essential, inside a primitive or simple language there may not be sentential explanations ("we understand the facial expression [as chil­dren], but can't explain it by any other means"). What Wittgenstein is thus taking into account is an isolated primitive language. ln this case, it is not a calculus, a game with fixed mies, that is the determining aspect to be looked at, but the environment where the language functions (the form of life).

The understanding of a gesture in our language may come before the capacity to give explanations according to mies of 'grammar.' This is what we do as children when we "understand the gestures and facial expressions of grown-ups - without explanation. 11 Thus Wittgenstein's view concefoing understanding in the BT excludes not only causal and hidden self-explanatory psychical entities (which he thought were a by-product of the false analogy between language and a mechanism), but also the possibility of understanding in primitive (or simple) languages. Wittgenstein's notion of 'understanding' as a correlate to 'explanation' in the BT does not correctly express the complexity of our language, since ours begins with primitive forms of communication and under­standing. This means that we cannot exclude the learning of language from our investigation as mere history (as urged in BT 5). The claims, then, in the BT, that "understanding begins with a sentence" (BT, 1) and that "understanding correlates with an explanation" (BT, 11) are false or, at best, idealizations of our language.40 Those daims were requirements imposed on our language derived from an analogy between mies of a game and 'mies of grammar.'41

But what about the general idea that languages are languages only to the extent that they have a 'grammatical stmcture'? If we take primitive languages ln isolation, we find that there is no set of mies of 'grammar1

internai to them. They don't have the required multiplicity to have a 'grammar,' since without sentences it seems useless to ta1k1 for instance, about contrasting possibilities that make explicit what is meant, or even about possible combinations of words that make sense. When taking Sraffa's gesture as part of a primitive language, Wittgenstein thus "had to recognize this and this as signs (Sraffa) and could not, however, give a grammar for them" (MS, 15 7b, 5v; quoted above). It is not only Wittgenstein's notion of 'understanding,' therefore, that does not fit primitive languages, but also his idea of 'grammar' as a system of mies that limits sense. Even if sentences play a central role in complex languages, they might not have that role in primitive languages, for it is not clear that they are even part of those languages in the first place. Primitive languages, thus, make the concepticin of language as the

156 Wittgenstein's Philosophical Development

"totality of propositions" of the T and the conception of language as a system of rules of 'grammar' of sentences lose their point. The idea that language is the totality of propositions is not an obvious tautology or a trivial 'grammatical remark,' but a misleading assumption (see Section 3.2).

According to the BT, as we have seen, someone who understands something understands a sentence (ora correlate of a sentence, some­thing with the sarne multiplicity). If one understands a sentence, one can explain what he means by the rules of grammar that he uses to formulate it. If he can do it, what he says makes sense. Thus, the goal of outlining the limits of sense, in the BT, makes the spedfication of rules for 'sentences' (Saetze) in general and propositions a ceritral matter. ln this way, the presentation of the "rules of language" took the place of the Tractarian idea of the general form of propositions (BT, 77, quoted above). Thus, in the BT, "what is to count as a sentence is", of course, "determined in grammar'' (BT, 76). The rules of grammar "determine the sense of a sentence" (BT, 79), and the sense is the "role the sentence plays in the calculus" (BT, 81). Therefore, Wittgenstein claims:

"Sentence" is obviously the heading for the grammar of sentences. But in a sense ít is also the heading for grammar in general, and is there­fore equivalent to the words "grammar" and "language." (BT, 63)

The introduction of primitive languages into Wittgenstein's investiga­tion, however, suggests that the equivalence in significance of sentence, language, and grammar is based on a misleading view of language. ln one passage of the PG this point is quite explicit:

- Imagine a language in which all sentences are commands togo in a particular direction. (This language might be used by a primitive kind of human beings exclusively in war. Remember how restricted the use of written language once was.) Well, we would still call commands "come here", "go there11

1 "sentences". But suppose now the language consisted only in pointing the finger in one direction or the other. - Would this sign still be a sentence (Satz)? - And what about a language like the early speech of children whose signs expressed only desire for particular objects, a language which consisted simply of signs for these objects (of nouns, as it were)? Or consider a system consisting of two signs, one expressing acceptance and the other rejection of preferred objects. Is this a language, does it consist of sentences? (PG, 112-3; from MS 114, 99; my emphasis; slightly modified translation)

The Big Typescrípt, Sraffa, and the Anthropological View 157

The point of this passage is to suggest that language is not something restricted to a system of sentences. A language with only a few signs will be, well, a language, but it is not clear that there are sentences in this language. If we ourselves call some sets of signs (a set of signs like the one introduced to Wittgenstein by Sraffa) 'languages without sentences,' ar languages in which it is unclear that there are sentences, then the idea of a general form of propositions that determines the bounds of language is flawed. More important, however, is the blow to the aspect of the T that had still survived in the BT: the determination of sense by means of fixed rules of 'grammar' grounded on sentences. What is the 'grammar' of children's primitive language taken in isolation? lt is a language that does not seém to consist of propositions ar sentences. Therefore, it does not seem to have a 'grammar' in the sense of the BT. Note that children's primitive language is part of our language and part of their learning process; a general 'grammar' for our language, thus, is an idealization of our language.

The assumption that "without grammar it isn't a bad language, but no language" (BT, 194), then, is flawed. Remarks such as "as if one could imagine a language without a grammar1

' (BT, 193) turn out to be very misleading indeed.42 ln fact, one can imagine a language without a 'grammar' in the sense of the BT. This is precisely what Wittgenstein is doing when he describes, in the passage quoted above, "a language like the early speech of children whose signs express only desire for partic­ular objects" (PG, 112-3). ln the PG, then, Wittgenstein begins to give up 'grammar' as a necessary condition for language.

Primitive (ar simple) languages are not, of course, like our language -that is why we call them 'primitive.' However, we are not really author­ized to think of them as, say, "essentially" different. Note that our language has some primitive parts (fundamental indeed to its learning process) and it is unclear whether we should stipulate sharp bounda­ries between primitive and complex languages. lt is clear, however, that we don't have to. Why, in the first place, should we exdude primitive languages from our understanding of 'language'? These issues come out in a discussion of ostensive explanations (definitions) in the PG:

Think also of teaching a child to understand words by showing it objects and uttering words. The child is given ostensive explanations and then it understands the words. - But what is the criterion of understanding here? Surely, that the child applies the words correctly. Does it guess the rules? - lndeed we must ask ourselves whether we are to call these signs and utterances of words "explanations" at ali. The

158 Wittgenstein's Philosophical Development

language game is still very simple and the ostensive explanation has not the sarne role in this language-game as in more developed ones. (For instance, the child cannot yet ask "What is that called?"). But there is no sharp boundary between primitive forms and more compli­cated ones. 1 wouldn't know what I can and what I can't still call "explanation". I can only describe language games or calculi; whether we still want to call them calculi or not doesn't matter as longas we don't let the use of the general tenn divert us from examiníng each particular case we wísh to decide. (PG, 62; slightly modified translation; emphasis added; from MS 140, 18)43

The boundaries between primitive and complex languages are not easily drawn. They are a matter of degree, and not a matter of something to be assumed 'in prinàple.' Hence, the very analogy between language and a calculus or game with fixed rules like chess can be misleading. The underlying point in this passage is that the exclusion of languages that don't have a 'grammar' was a requirement imposed on language and that the use of the general term 'calculus' was diverting Wittgenstein from seeing the variety of primitive languages. (The "general term" calculus, as Wittgenstein says, "may divert us from examining each particular case".) This seems to be the point of what Sraffa says in a note from 1932 that he wrote for Wittgenstein: "We should give up the generalities and take particular cases, from which we started" (WiC, 196). ln this note, Sraffa is expressing doubts concerning the possibility of finding general rules to classify sense and nonsense, which was the goal of Wittgenstein's calculus of fixed rules of 'grammar' (BT, 425) - see Sections 3.1and3.2.

ln the BrB (1934-5), it is clear that Wittgenstein gets rid of the require­ment that language must be a set of sentences, that language must have a 'grammar' in the sense of the BT. There, he introduces a language of four words and asks us to imagine a "society in which this is the only system of language" (BrB, §1). In his 1934-5 lectures Wittgenstein uses a similar language as an example to discuss the nature of negation. He says:

"Not Brick" would be translated into our language as "Don't bring me a brick", and 11 Apple" into "Bring me an apple." But not in the primitive language. ln the game ín whích one says "Not apple" there is nothing one could call a proposition. (WLC32-5, 77, my emphasis)44

Prtmitive languages, thus, may Jack propositions; they may Jack orders. Thus, they may lack complex units grounded on compositionality and

The Big Typescript, Sraffa, and the Anthropological View 159

multiplicity.45 lf one wants to call them sentences (and, of course, one can doso), then one has to think that sentences are not just sequences of combinatíons of words. This, of course does not fit the idea of 'grammar' of the BT:

If a sentence were not one possible combination among others, then it wouldn't have any function. That is to say: lf a sentence/descrip­tion was not the result of a decision, it would have nothing to say. (BT, 92)

Since one can take 'Apple' or 1Brick' as sentences, the assumed complexity óf sentences in the BT, as in the T, is not correct. Even more so, if Wittgenstein is willing to call 'language' a system of 11 just one signal."46 ln bis 1934-5 lectures, when discussing "a simple language in a tribe in which only orders are given," he writes:

We, who talk about the tribe having this language, call these "orders" because the role [the original reads 'rule', but this is clearly a mistake] these words play in the life of the tribe is that of orders. The word "order" is not in their language, nor is any such thing as conversa­tion. The whole object of communícation between a builder and his workman. (WLC32-5, 102)

The absence of complexity and of propositions, orders, and sentences in general, inside a primitive language is also a central issue in the PI. Wittgenstein asks us in §2 to consider the language of four words of the builders as a "complete primitive language." ln order to doso, we think about a form of life: "To imagine a language means to imagine a form of life" (PI, §19). ln §6 he says that the primitive language of §2 could be seen as "the whole language of A and B, even the whole language of a tribe" (Wittgenstein's emphasis). ln §5, referring to a child's language, he says that in "primitive forms of language ... teaching of language is not explaining, but training." There can be little doubt, therefore, that Sraffa's critique was fruitful not only because ít is connected with the final break with the T, but also because it helped Wittgenstein to shape the final form of his thoughts. 47

Note that if 'sentence' and 'language' have loose limits, if "there is no sharp boundary between primitive forms and more complicated ones" (PG, 62), then it is quite hopeless to establish the supposed 'limits of sense' by means of 'rules of grammar,' which was the purpose of Wittgenstein's 'grammar' in the BT. Conceming 'sense', according to

160 Wittgensteín's Philosophical Development

Moore, Wittgenstein says the following ín a lecture from 1932-3 (thus, after the corpus of the BT was written):

[ ... ] he [Wittgenstein] went on to say that his present view was that 111 sense' was correlative to proposition" ([ ... ]) and that hence, if 'prop­osition' was not 'sharply bounded,' 'sense' was not sharply bounded either [ ... ] He concluded finally that "'makes sense' is vague, and will have different senses ín different cases,'1 but that the expression 'mak.es sense' is useful as 'game' is useful, although, like 'game,' it "alters its meaníng as we go from proposition to proposítion"; adding that, justas 'sense' is vague, so must be 'grammar,' 'grammatical rule' and 'syntax.' (M, 268)

The lack of sharp limits to 'proposition' and 'language' thus suggests that the idea of 'grammar' as the rules of sense is idle. Tabulating those rules is useless, if one is lookíng for the limits of language or the limits of sense. One should thus expect that Wittgenstein's philosophy after 1933 will not focus on those ideas. Th~ good news is that he does not need them, if he gives priority to the application of the genetic method (l return to these issues ín Chapter 4).

ln arder to see the variety of languages it is better not to impose requirements on the concept 'language', but to look at ít as "a family of structures more or less akin to one another" (MS 142, §109; PI, §108).48

Once we look at language as a "family resemblance" concept, we will not be tempted to take the "calculus of fixed rules of grammar" as a common property of language or as a transcendental condition for language. ln this case, we recognize that the primitive languages are "forms of language not separated by a break from our more compli­cated ones" (BB, 17). That is, the idea of "family resemblance" is meant to break the spell of requirements that we are tempted to impose on language (1 will come back to this in Chapter 4, Section 4.3.2).

3.4.3 An anthropological view: purpose, point, and fonns of life

The significance of Russell's causal theory of meaníng is re-examined by Wittgenstein after the idea of primitive languages is taken seriously. Once 'grammar' is not a necessary condition for something to be a language, the in-príncíple exclusion of language as a mechanísm is abandoned. Let us compare a passage of the BT that is reintroduced in the PG with a dramatic change. First, 1 quote the passage from the BT:

Inventing a language does not mean inventing a device for a partic­ular purpose on the basis of the laws of nature, ar ín agreement with

The Big Typescript, Sraffa, and the Anthropological View 161

them. As is the case, for instance, with the invention of the gasoline engine or the sewing machine. The invention of a game is also not an invention in this sense, but is comparable to the invention of a language. [ ... ] one cannot say of grammatical rules that they are an instrument that enables language to fulfill its purpose. (BT, 193; Wittgenstein's bold; my italics)"

The rewritten passage in the PG makes the opposite point:

Inventing a language could mean inventing a device for a particular purpose on the basis of the laws ofnature (or in agreement with them); but it alsó has the other sense, analogous to that in which we speak of the invention of a game. (PG, 192; my italics; see also PI, §492)

The invention of a language as a device for a given purpose on the basis of the laws of nature, that is, a language as a mechanism based on the laws of psychology in the style of Russell, tums out to be possible after all. This is because primitive or simple languages can be understood as a mechanism:

One can of course consider language as part of a psychological mech­anism. The simplest case is if one uses a restricted concept of language in which language consists only of commands.

One can then consider how a foreman directs the work of a group of people by shouting. (PG, 187; my emphasis; MS 114, 162)

Wittgenstein, thus, is giving sense to the idea of "language as part of a psychological mechanism." One gives it sense, if one restricts it to a language of commands, for instance. But this might be all the sense that we are able to give to the idea (in Chapter 41 return to the idea of 'giving sense'). The very idea of a "mechanism" is, thus, not einfach Unsinn (it is not excluded in principie); it is, however, an analogy that suggests a very simplified account of our language and can lead to assumptions that we might do better to avoid (for instance, the self-explalnlng "images" in the mind). Such slmplifications are better understood lf we imagine primitive languages ln two ways: elther as a part of our way of living (for instance, the language children speak) or as a description of a different form of life (for lnstance, "a primitive kind of human beings exclusively ln war" (PG, 112)).

The use of primltlve languages suggests, then, an anthropological view. Rhees said, in fact, that Wittgensteln told him that "with Sraffa he learnt above ali to look at philosophical problems with an anthropological

162 Wíttgenstein's Philosophical Development

view."49 ln order to fully appreciate what the role of signs in a language is, one should look at how those signs relate to the form of life of which they are part. 50

Because of the new perspective, Wittgenstein steps away from some of his comments in the BT, for example this one:

"A sign is also intended for a living being, so that must be something essential to a sign." Certainly: a chair too is always intended just for a human being, but it can be described without our talking about its purpose. A sign has a purpose only in human society, but this purpose is of absolutely no concem to us." (BT, 192; my emphasis)

Wittgenstein's new trend is to consider the purpose and the point of languages and language-games precisely in "human societies." This implies a broader notion of use, which is not restricted to the posi­tion of words in a calculus; a use of words that "meshes with life": "Is meaning then really only the use of a word? Isn't it the way this meshes with our life? But isn't its use a part of our life?" (PG, 65; MS 140, 22). The broader notion of use understood as part of a form of life (or our life) is also clearly related to Sraffa's criticisms. Besides the remark about the "derisive gesture" (BT, 10) quoted above, the connection between Wittgenstein's anthropological viewpoint and Sraffa's criticism can be seen in a passage of the BT that he crossed out. ln this passage, we can also see why Wittgenstein did not accept Sraffa's point immediately and, maybe, this is why he thinks Sraffa's criticism needed to be "unceasingly practiced" (PI, x). Wittgenstein first introduces Sraffa's view:

(Sraffa) An engineer is building a bridge. To do this, he checks in several handbooks; in technical handbooks and in legal ones. From the one he finds out that the bridge would collapse if he were to make this part weaker than etc., etc.; from the other that he would be locked up if he built it in such and such a way. - Now are the two books not on the sarne levei? That depends on what kind of role they play in his life. After all, for him the legal handbook can simply be a book about the natural history of the people around him. Maybe he also has to check in a book on the life of beavers in order to fínd out how he has to paint the bridge so that the beavers won't gnaw on it ... (BT, 242; MS 113, 24r, from December 20, 1931; my emphasis)

Note that this remark suggests that the descriptive grammar criticized by Sraffa (Section 3.2) should take into account the various roles of rules

1

1 i 1

\ 1

The Big Typescript, Sraffa, and the Anthropological View 163

in life. If one takes into account the introduction of the anthropological view in the PG (MSS 114, 115 and 140), one can understand that the reason Wittgenstein may have had for crossing out that passage lies, in fact, in its continuatlon (in Wittgenstein's answer, which was also crossed out.) There Wittgenstein denies Sraffa's approach:

-But isn't there still another way of looking at laws? ln fact don't we even have the distinct feeling that this is not the way we look at them? - lsn't this the sarne questionas: - Is a contract merely a state­ment that it is useful for the parties to act in such and such a way? Don't we in some cases (even if not in all) feel "bound by a contract" in a différent way? Now one can say: "Whoever feels bound by a contract or a law erroneously plctures the law as a person (or God) who constrains (zwingt) him with physical force"? - No; because even if he acts as if someone were constraining him, then still his action is real ... The words "the contract binds me" are, to be sure, a pictorial representation, and are therefore a false proposition, given the usual meaning of the word "to bind"; but understood correctly, they are true (or can be true), and they distinguish the one case from another in which a contract merely tells me what is useful for me to do. And even if one objects to the words "the contract (or the law) binds me", one can't object to the words: "l feel bound by the contract." (BT, 242-3)

Wittgenstein, in the BT, does not yet accept an anthropological view and denies that the distinction between the technical and the legal book "depends on what kind of role they play in life." He bases his denial on the "distinct feeling that that is not the way we look at them" and that this "distinct feeling" does not imply that the binding character of laws assumes the picture of God constraining us. The relevant point here is that an anthropological view, in Wittgenstein's view in the BT, seems to imply that we are not bound by roles at all. lt seems to imply that roles are merely descriptions of facts (as seen in Section 3.2).

Even though Wittgenstein speaks of a contract and legal law in the passage above, one can see that the danger of an anthropological view lies in what it seems to imply in our understanding of the status of the laws of logic and mathematics (MS 117, 172; quoted in the next para­graph). The binding power of logic and mathematics seems to be lost when we emphasize "the role they play in our lives." Such a view seems to put at risk the "hardness of the logical must" (MS 117, 121) expressed in the fixed rules of language or 'grammar.1 It see:rrts to imply that the

164 Wittgenstein's Philosophical Development

rule of 'grammar' 2 + 2 = 4 is a special case of descriptions of how we calculate.51

After the BT, however, Wittgenstein comes to think that it is exactly the role that mathematics and logic play in our lives that makes them binding, it shows how logical laws "constraln" us. He does not simply follow Sraffa's opinions, but he, as it were, reshapes the anthropological view. On the one hand, attention to what we do with mathematical sentences makes clear that they are not anthropological descriptions; on the other hand, the anthropological view allows us to see what "binding" can mean, as longas we pay attention,to a broad notion of use. Around 1940, Wittgenstein refers to a conversation with Sraffa .similar to the passage mentioned above: ..

Are the sentences of mathematics anthropological sentences that say how we, human beings, infer and calculate? - ls a book of laws a work of anthropology that tells us how persons of this people treat a thief, etc? - Could one say: "The judge consults in a book of anthropology and, after it, sentences the thief to a p,rison sentence"? Now, the judge does not USE the book of laws as a manual of anthropology (conver­sation with Sraffa). (MS 117, 172; my translation)

To take mathematical sentences as "anthropological sentences" (say, as descriptions of actual calculations) is a misunderstanding of the role that such sentences play in our form of life. We don't use math­ematical books as anthropological manuais. The use of law books, mathematics, logic, etc., that is, the use understood as the role that they play in our lives (which includes their purpose, their function, and their practical consequences), give usa clue to why and how they are binding:

... one can say that the laws of inference constrain us (uns zwingen); namely, in the sense that other laws [constrain us] in human society ... But the one who infers differently, to be sure, gets into conflict: for instance, with soàety; but also with other practical consequences. (MS 117, 30-1; my translation; see RFM 1, §116)

It is not God that constrains (zwingt) us; but it is also not an internai property of rules or laws that make us "feel bound by the contract" (BT, 243-4; quoted above). The constraint is related to the central role that certain rules or laws (say, the "laws of thought") have in our practices. Many of our interests and purposes are connected to those roles.

The Big Typescript, Sraffa, and the Anthropologica/ View 165

The use of mathematics and logic in our life also indicates that we mostly agree about the results of inferences and calculations. We can say, then, that we agree in the language we use, which is "not an agreement in opinions but in form of life" (PI, §242).52 lt is a fact about human beings (of natural history, if you wish) that we agree in mathematics and logic almost always (one should not forget, for instance, intuitionist logi­cians); it is also a fact that the application of mathematics has been very helpful and important in human societies; however, it is not a fact that mathematical equations describe how we calculate ar describe an agree­ment of opinions conceming calculation. They don't work as descrip­tions at ali in our form of life; they have a different role, which is to be described by philosophers who wish to understand the compelling character of practices related to our calculations.

Note that the anthropological view is neither meant to lay down the foundations of necessary propositions, nor to suggest, say, the "frailty of mathematics":

If we employ the ethnological viewpoint, does it mean that we declare philosophy to be ethnology? No, it only means that we take up our position far out in arder to be able to see things more objectively. (MS 162b, 67v; from 1940)

The ethnological viewpoint brings more objectivity to our investigation because it may avoid mythological assumptions (such as Platonism, Idealism, or Grammatlcism) and because it may help us to reach a clear view of how mathematics is binding in our practices - our prac­tices might also get clearer if we contrast them wíth imaginary forros of life. Thus, Wittgenstein is not trading the hardness of the rules that are supposed to be, in the BT, implicit in language for the hardness ar frailty of ethnological facts. The point is: look at what we do with mathematics and logic and how we do it; look at their central role in our líves in contrast with other forros of life if you want to be clear about why their laws are binding and why we say things like "math­ematical equations are necessary." ln this case, we may not need to assume that there is something that makes them necessary (be it eternal objects, or the structure of the mind, or essence in language, or human nature).

One should certainly ask oneself to what extent philosophical founda­tions formathematics reallymakemathematics more secure by explaining the true nature of 'necessity.' If we assume that intuitions a priori ar a platonic realm (ar 'rules of grammar') are a necessary condition for

166 Wittgenstein's Philosophical Development

mathematical practice, have we really achieved somethíng? Accordíng to that kind of account, mathematics depends on somethíng outside itself that can easíly be seen as more doubtful or less clear than mathematical practice itself. lf mathematics depends on philosophical grounding, one can think, so much worse for mathematics. And to what extent do phíl­osophical foundations really work as explanations? lt is clear that we go a step back when we assume that mathematics depends on a priori intui­tions, a theory of types, or on a platonic realm (or even on grammatical rules). What kind of step is it? lsn't it like when people assume a creator in order to explain the beginning of toe world? ln fact, such a step (a

priori intuitions, platonic realm, God) "doesn't really explain anything, but does create a beginning that is acceptable to humans" (BT, 230). lt is a psychological device, not a logical one. lt is not a logical consequence of our point of departure: necessary propositions of mathematics or, in the case of God, the nature of existing things, say, the material world. Note that a priori intuitions, a theory of types, or platonic realms would merely do the job of an unanalyzable something, a self-explaining thing that would, perhaps, satisfy us for a whil~. Such a temporary satisfaction is obvíously controversial and not really realistic (in the ordinary sense of the word).

As a later remark makes clear, the anthropological view is intended to change our perspective concerning those matters:

Instead of the unanalysable, specific, indefinable: the fact that we act in such-and-such ways, e.g. punish certain actions, establish the state of affairs thus and so, give orders, render accounts, describe colors, take an interest in others' feeling. What has to be accepted, the given it might be said - are facts of Iiving. (RPP 1, §630)

"lnstead" is not indícating a new foundatlon of language. The expres­síon only makes plain that we should look at facts, which we have prob­ably forgotten while philosophizing, instead of looking for foundations. Unanalyzable terms, indefinables, etc., in philosophical prose, conceal as much as 'God' in theological theorizing.

Note that the genetic method is first used in order to find the origíns of the supposítíon that something mental (mental states, mental images) really explains lntention, expectatíon, rules etc. With the anthropological view ín place, the use of the genetíc method in the dismantling of the quest for foundations thus finds an important complement.

r 1

The Big Typescript, Sraffa, and the Anthropological View 167

3.5 Concluding remarks: putting two and two together

We have seen in this chapter that Sraffa's criticisrn broke the last surviving bond with the T: the idea of language as a calculus grounded by 'grammar.' This reshapes Wittgenstein's descriptive stance: after the BT, a description is not the tabulation of implicit rules of language anyrnore. ln order to dissolve philosophical puzzlement, we should rather describe (after rernernbering) what people do in certain surroundings when they use language.

The articulation of language, thought, and world is to be accepted as a fact of natural history. This articulation is not to be found merely inside thought or language, but also in the surroundings of language. The articulation varies according to the form of life in question (how people are trained, the purposes they have, the actions that are signifi­cant, etc.). This fact is forgotten when we assume an articulation given in the 'system of grarnrnar' or in a priori logic.

At the end of 1933, Wittgenstein asks: "What is then our concept of systern if it is not a cloud cuckoo land?" (MS 146, 45). He asks this in a context in which he criticizes his own tendency to "systernatize" (MS 146, 30).53 Theidea of system is, for sure, not used in connection with the requirement of rnultiplicity after the BT. There is a loose use of 'systern' (see, for instance, PI, §6). 'Multiplicity' requirernents vanish. 'Calculus,' 'game,' 'systern' might be useful because they partially elucidate, for instance, the specificity of descriptions of facts or what is rneant.

'Calculus,' 'game,' and 'autonorny' of rules also find their way back horne: to the philosophy of rnathematics. They rnight be useful in those contexts because we, indeed, distinguish between the calculus and its application in rnathematics. We apply rnathematics when we build bridges, for instance. To say that we apply language in general (or 'rules of grarnrnar'), however, is misleading, for it gives the irnpression that language has a structure ('gramrnar') that rnight be dissociated frorn its surroundings.

The contexts of the rernarks in which Wittgenstein's jargon in the BT ('grarnmar,' 'calculus,' 'system,' 'verification,' for instance) is used later are, of course, fundamental. Not paying attention to those might lead to wrong conclusions concerning what the words mean. Phrases which introduce those words after the BT (for instance, "in a certain sense" and "one could say") should be read very slowly.

An important feature of Wittgenstein's philosophy after the intro­duction of the anthropological view is the disappearance of the worries

168 Wittgensteín's Philosophical Development

conceming the 'application of language' (as opposed to its under­standing). The distinction between the 'interpretation of language' and its 'application' has no fundamental role once primitive languages are introduced. Those features are obviously absent in primitive languages. ln those languages, one cannot even talk about rules of 'grammar' and the application of those rules, for they don't have this sort of complexity. Thus, the distinction between language and application is not even tenable in the terms of the BT. The idea of verification by means of phenomena, as applied in the PR and in the BT, is therefore completely abandoned. Verificationism is replaced (and thus deflated) by a question conceming criteria. After the BT, asking for the verification of a sentence (i.e. asking for a method that determines its truth) is, redlly, merely one way of determiníng criteria to what is meant with a sentence:54

The question of the possibility of the mode of verifícation of a sentence is only a special form of the question "How do you mean that?" The answer is a contrlbution to the grammar of the sentence. How do we know if it rains? We see, feel, the rain. The meaning of the word 'rain' was explained to us with those experiences. I say, they are 'criteria' for [the fact that] it is raining. (MS 115, 72; from 1933-4)

The transcendental arguments and the talk about 'conditions of possi­bility,' distinctions in principie, the philosophical idea that 'things must be like this,' are gradually taken out of circulation. The transcendental arguments for the arbitrariness of 'grammar' are not used after the BT. lt is a matter of fact that we don't justify mathematical sentences empiri­cally (they are not used as empirical descriptions in human sodeties). The notion of the 'use' of mathematical equations broadly understood (it includes their purpose, their point, their role in life) is simply different from the use of descriptions of facts - even though this distinction is not clear-cut and there are changes of status. This means that a "tran­scendental argument" for the "arbitrariness of grammar" does not add anything to the description of our practices.55 Such arguments, actually, are predsely one of those philosophical devíces that make us forget to look at our practices. They give the impression that we can know the meaning of the binding character of mathematics and logic a priori, as if the use of words embedded in their surroundíngs was a detail, something that could be looked at later. This gives us, in its tum, a sublime picture of the inexorability of logic and mathematics. To break such a picture is a major goal of Wittgenstein's later philosophy of mathematics (RFM 1, §§4-16, 113-53). He intends to break such a picture with reminders of

l )

~ 1

' 1

(

~

i

f

\

~ {

The Big Typescript, Sraffa, and the Anthropological View 169

general facts of natural history (RFM 1, §§ 63---4, 111, 142), descriptions of our practices (RFM l, §§1-15), and descriptions of alternatíve practices (RFM 1, §§5, 147-51).

The transcendental arguments, still used in the BT for the existence of absolute position in the visual field, which was the last surviving aspect of the "sense datum notation," are also finally left behind:

Now the danger we are in when we adopt the sense datum notation is to forget the difference between the grammar of a statement about sense data and the grammar of an outwardly similar statement about physical objects. (From this point one might go on talking about the misundefstandings which find their expression in such sentences as: "We never see an accurate circle," /1 All sense data are vague." Also, this leads to the comparison of the grammar of "position," 11motion, 11

and "size" in Euclidean and in visual space. There is, e.g., absolute position, absolute motion and size, invisual space.) (BB, 71)

Note that 'grammar' can be replaced by 'use' in this passage (l come back this issue in Chapter 4). Note also that all of Wittgenstein's allegedly crit­icai talk conceming vagueness, absolute position, etc., after abandoning the project of a phenomenological language, is itself the product of a confusion. Such talk arises at the point where we have already forgotten something, where we are misled by a false analogy between statements concerning sense data and physical objects. The very talk about "abso­lute position" that survived in the BT needs, thus, the scrutiny of the genetic method.

The old idea of a perspicuous representation (uebersichtliche Darstellung) of PR and BT is profoundly modified. Here is the idea expressed in the BT:

The investigation of the rules of the use of our language, the recog­nition of these rules, and their clearly perspicuous representation (uebersichtliche Darstellung) amounts to, i.e. accomplishes the sarne thing as, what one often wants to achieve in constructing a phenom­enological language. (BT, 437; modified translation)

What the phenomenological language was supposed to achieve, we have seen in Chapter 1, was the demarcation of the limits of language (or sense). Thls, we have seen in this chapter, is nota goal to be pursued once the anthropological view is taken seriously. The tabulation of rules is also no longer a goal. Therefore, 'tabulated rules' will not be

170 Wíttgenstein's Philosophical Development

equivalent to a perspicuous representation anymore. When discussing the infinitesimal calculus and Hardy's idea of the infínite, Wittgenstein states his goal:

The goal is a comparative perspicuous representation of all appli­cations, illustrations, conceptions, of the calculus. A general illumi­nation. (For each one-sided illumination also sheds shadows). The complete perspicuous view about everything that can create lack of clarity. And this perspicuous view must cover a broad area, for the roots of our ideas reach very far. (MS 116, 55)56

Note that the perspicuous representation will not be centéred in the idea of rules (not even in the mathematical calculus), but in many aspects of mathematical practice. At the sarne time, it will be a tool that sheds light on possible shadows that Wittgenstein's own investigation might cast. Note also that the focus of the perspicuous representation will be the ''roots of our ideas, 11 the roots that generate philosophical confusion, the very first steps into philosophical re&soning.

We have in hand now the seeds of the most relevant features of Wittgenstein's later work: the genetic method (Chapter 2) and the anthropological view (this chapter). ln the next chapters we will see how they lead to further developments in Wittgenstein's philosophy. The anthropological view will be in service of the genetic method from the BrB onwards. The sarne is to be said of the new notion of grammar. 'Grammar' will shift its meaning: the comprehensive discipline of 'grammar' will have no place in Wittgenstein's philosophy after PG. It will be mostly used as being synonymous with use or descriptions of uses. The philosophical activity or investigation is, thus, grammatical in a new sense because a philosopher, in Wittgenstein's sense, describes the uses of words in their environments (the form of life) in arder to avoid philosophical theorizing right from the beginning. This kind of descrip­tion delivers the new idea of a perspicuous representation (uebersichtliche Darstellung), which could be correctly translated now as surveyable repre­sentation of our practices (not just of the 'grarnmar' of our language).

,... 1

4 The Road to the Philosophical Investigations (Blue Book, Brown Book, German Brown Book, and MS 142)

4.1 'Grammar' and genetic method revisited

After the PG, Wittgenstein changes the meaning of the word 'grammar.' The new meaning wíll be introduced in this chapter by means of the elucidation of the new context in which it appears from the BB onwards. 'Grammar' is to be seen as the description of the use of words broadly understood, i.e., the description of practices related to words and their surroundings (and notas the discipline that tabulates the rules of language and presents the "limits of sense11

). Such a description ís at the service of the genetic method. Even though the genetic method was used with the presupposition of the old 'grammar' in the BT, from the BB onwards this is not the case anymore. The application of the genetic method and the use of an anthropological view dispense with any relevant role for the old views on 'grammar.'

Besides the dispensability of the old 'grammar,' which will be eluci· dated in this chapter, there are many reasons for Wittgenstein not to use the word with its old meaning (until the BT). First, as we have seen in the last chapter, if 'sentence' and 'language' don't have clear limits, then neither does 'sense.' The whole project of drawing the limits of sense, which was the original purpose of Wittgenstein's notion of 'grammar' since PR, is abandoned. Second, because of the tension between Wittgenstein's neutrality stance and the assumptions that surround the idea of 'grammar' as the discipline of the limits of sense (Chapter 3, Section 3.3). A third reason for the abs.ence of the old 'grammar' is that it generates lack of satisfaction and, thus, resistance of the reader. A philosopher is inclined - correctly, I think - to ask for a justification

171

172 Wittgenstein's Philosophical Development

of the use of the word 'grammar.' This point is evident in the case of Sraffa's and Moore's worries concerning 'grammar' as the discipline of the limits of sense (as seen in Chapter 3).1 ln the BT, one gets the impres­sion that Wittgenstein wants to stop a philosopher from theorizing just to make him believe in a particular theory concerning 'grammar' and necessity. This means that Wittgenstein's strategy of being a mirror for the reader fails with the old idea of 'grammar.' The old view of 'grammar' might generate a meta-grammatical dialectic and philosophical resist­ance instead of the acknowledgment of philosophical mistakes. But this acknowledgment is a central goal of the genetic method:

lndeed we can only prove that someone made a mistáke if he (really) acknowledges this expression as the correct expression of his feelíng. For only if he acknowledges it as such, is it the correct expression (Psychoanalysis). What the other person acknowledges is the analogy I'm presenting to him as the source of his thought. (BT, 41 O)

ln the BT, some 'grammatical' sources, of philosophical problems are indicated (some false analogies), but the reader is unlikely to recognize these as the real sources of his confusions. One of the reasons for this is precisely the suspicion that Wittgenstein himself assumes a disputable philosophical conception in the BT, as we have seen.

Moreover, one can easily imagine many questions that a phíloso­pher who reads the various divisions of the BT might be inclined to ask before he acknowledges a mistake. One might thus think that there is no space for a dialogue in the BT, while the genetic method creates the urge for it. lf the reader is supposed to recognize his own tenden­cies and mistakes in Wittgenstein's writings, the reader needs a voice. So it could be profitable to use imaginary interlocutors to express their voices until complete clarity is achieved. ln Wittgenstein's later philos­ophy the reader's temptations will find expression in interlocutors who always come back with new questions and doubts. It is such a dialectíc that will help Wittgenstein to create a surveyable representation instead of the tabulated rules of 'grammar,' which might generate the meta­grammatical concerns. The reader's possible dissatisfaction might thus be seen as one of the reasons for Wittgenstein to pursue new writing styles after 1933.

4.2 The Blue Book: 'grammar,' 'calculus,' use and analogies

A new writing style appears fírst in the BB. The old idea of 'grammar' plays no relevant role in the book and it is the first work that expresses

The Road to the Philosophical lnvestigations 173

the centrality of the genetic method in Wittgenstein's philosophy (see the description of the genetic method in Chapter 2, Section 2.4). His primary concern in the BB is to apply the genetic method consistently and exhaustively: "by our method we try to counteract the misleading effect of certain analogies" (BB, 28). The BB is not yet a complete expres­sion of Wittgenstein's later philosophy, but it is certainly closer to it than the transitional work in PG (or the MSS that were used by Rhees to compose it). At the end of 1935, it was the BB that Wittgenstein sent to Russell as a presentation of his philosophy. At that time, Wittgenstein was still "correcting mistakes and misprints" in the BB (CL, 269). This indicates th~t he did not take the BB as a bad expression of his thoughts some years' after its dictation. More striking, however, is that Sraffa wrote comments on the BB in 1941, many years after the dictation. Wittgenstein would not give a copy of the BB to Sraffa, his more impor­tant critic, if he did not think that central aspects of his later philosophy were presented there.

4.2.1 'Grammar' in the Blue Book

ln the BB Wittgenstein is no longer concerned with the limits of sense discovered and presented in the form of tabulated rules of 'grammar,' although there are passages in the BB that could appear to be similar to the views from the BT. Actually, they express a significant move from those views when read more carefully. This does not mean, however, that they cannot be understood as philosophical claims. After all, the BB is partially contemporaneous with the PG. To make this point clear, it is useful to take a look at passages where Wittgenstein discusses 'grammar' and 'calculus'.

'Grammar' in the BB is mostly used as synonym for 'use.' On page 23, for instance, this is explicit: " ... it was a way of examining the grammar (the use) of the word "to know" ... " (my emphasis).2 There is, however, a passage in the BB where 'grammar' could mean something similar to 'the discipline of the limits of sense':

But we are, as a matter of fact, misled into thinking that our expres­sion has a meaning in the sense in whlch a non-metaphysical expres­sion has; for we wrongly compare our case with one in which the other person can't understand what we say because he lacks a certain information (This remark can only become clear if we understand the connection between grammar and sense and nonsense). (BB, 65)

The context of the remark is the discussion of solipsism. Wittgenstein examines in this context solipsistic remarks like "Only what 1 see

174 Wittgenstein's Philosophical Development

(or: see now) is really seen.11 What is peculiar to such a claim is that it could apparently only be understood by the solipsist himself, given that according to him only his experience is real. The solipsist thinks, nonetheless, that "Only what 1 see is really seen" is a statement like "Only 1 have seen Mr. Bush talking to Mr. Blair.11 The analogy between the statements is, however, misleading. The first statement suggests that, if it made sense at all and were true, it could only be understood by one person, while anyone can understand the statement about Bush's supposedly secret meeting. So the quotation above is meant to eluci­date that when a solipsist comes to t.he conclusion that nobody can understand him, he is prone to thinking that nobody understands him because of Jack of relevant information. But, Wittgensteih says, it is not because other people lack information that they cannot understand the solipsist philosopher (under the assumption that only what he sees is real); it is because it is meaningless for other people in the solipsist's own terms (people would not have any use for such a sentence ín the solipsistic scenario).

The question is, now, how can this point become clearer if we "under­stand the connection between grammar and sense and nonsense"? What is this connection according to the BB? Actually, Wittgenstein does not explicitly explain in the BB what the connection is. Should we, there­fore, conclude that Wittgenstein has in mind his views on grammar expressed in the BT in the quoted passage? Here, l think, we can go both ways. lt is not clear enough whether Wittgenstein has completely aban­doned his views on 'grammar' from the BT. The BB gives us, however, a way to deflate that remark. Since Wittgenstein does not further explain "the connection between grammar and sense and nonsense," he might have thought that the point was not relevant in the course of his dicta­tions. Wittgenstein's use of 'grammar' as synonym with 'use' already in the BB (BB, 23) also indicates that he is not interested in 'grammar' as the discipline that tabulates rules of a 'calculus' and establishes the limits of sense. This, in fact, flts with what takes place later in the BrB (section 3).

Moreover, it is striking that in the BB 'nonsense' appears only as an indication that we had better exclude a useless expression from language to avoid confusion. Thus, 'nonsense' means pretty much 'useless' in the BB. The word 'nonsensical' appears for the first time in the BB in the context of the discussion of the different uses of the words 'thought' and 'sentence' (BB, 7). One may think that given that sentences can be located, thoughts must also have a place where they occur (a misleading analogy between thought and sentence). Wittgenstein asks in the

\

! ) i

ti

[

l r 1

The Road to the Philosophical Investigations 175

context of this discussion: 11Now does this mean that it is nonsensical to talk of a locality where thought takes place?"; the answer follows imme­diately: "Certainly not. This phrase has sense if we give it sense" (BB, 7). We need to be clear about what uses we could make of the locality of thought. Wittgenstein himself imagines an experiment in order to give sense to the expression 'locality of thought.' He supposes that an experimenter observes his brain working by means of a mirrar while he thinks (thinking "may consist of a train of images, organic sensa­tions, or on the other hand of a train of the various visual, tactual and muscular experiences which he has in writing or speaking a sentence"; BB, 8). What the observer sees going on in the brain may be called the expression 'Of thoughts and, thus, the brain may be called the locality of thoughts. ln this context the word 'nonsensical' is used again. Wittgenstein says the following about the two phenomena:

Both these phenomena [images that appear while one is writing a sentence, etc., and the brain working] could correctly be called "expressions of thought"; and the question "where is the thought itself?" had better, in arder to prevent confusion, be rejected as nonsen­sical. (BB, 8; my emphasis)

It is very significant, first, that Wittgenstein is making a suggestion, as is indicated by "had better." Second, the reason to exclude "where is the thought itself?" is to avoid more confusion in the case that we try to answer the question. Since what is observed (the brain working) was considered the expression of a thought and the images, sensations, etc., were also so considered, there would be nothing else to be taken as the "thought itself." Our tendency here would probably be to go ahead and postulate "something" (it could go on like this: "there must be some­thing, after all the thought is not a nothing and it is different from the expression of a thought; therefore, the something must be the thought itself"). So the word 'nonsensical' indicates a suggestion of exclusion of a combination of words that is not doing any relevant work anyway. But even though there is a suggestion of exclusion, there is no indication of a transgression of the limits of sense. lt seems, rather, that sense and nonsense are a matter of degree and that it is always important to try to find, flrst, a grammar (a use) for a combination of words that seem difficult to understand or seem to be nonsense. But we can also exclude a combination of words if it is not doing any relevant work and may confuse us. Instead of using the idea of 'limits of sense,' Wittgenstein tries thus to give sense to philosophical sentences; If this fails because

176 Wittgenstein's Philosophical Development

the sense we can give is, as it were, modest and local, while the philo­sophical claim is general, we exclude the sentence from language in arder to avoid further confusions.

The strategy of 'trying to give sense' is very much in line with the idea of dissolving philosophical puzzlement. If we cannot give sense to a philosophical worry, we better give it up as a real worry. It might be the case that a philosophical worry has a kind of internai instability: if we give it a sense, it gets an ordinary sense, and it Iases its puzzling character (here I have in mind radical views such as solipsism and skepticism).

So, there is no good reason to suppose that the connection between sense, nonsense, and grammar in the BB is the connectión suggested in the BT. It is more fruitful to think that the connection is the following: a philosopher (say, the solipsist) believes that he has a valuable piece of information because he does not pay attentíon to the use of his words (grammar); if he paid attention to it, he would concJude that what he saíd was not a piece of information (sense) but something that could be excluded from language as a useless com,bínation of words (nonsense) in arder to avoid further confusions. Thus; we don't have to, and perhaps should not, take Wittgenstein's statements about grammar in the BB as statements about the discipline that prescribes the limits of sense.

4.2.2 'Calculus' in the Blue Book

One might think that Wittgenstein is still defending the view of 'grammar' of the BT in the BB because of his use of the word 'calculus' in the latter as well (BB: 13, 25, 42 and 65). His use indicates, however, that he is moving away from his old conception of language as a calculus with fixed rules. ln the passage that could best support the old conception of 'grammar' Wittgenstein says that "the sentence has sense only as a member of a system of language; as one expression within a calculus" (BB, 42). This certainly sounds like his old view on 'grammar' as the rules of a calculus. However, if we look at the whole passage this impression might disappear:

I have been trying in all this to remove the temptation to think that there 'must be' what is called a mental process of thinking, hoping, wishing, believing, etc., independent of the process of expressing a thought, a hope, a wish, etc. And I want to give you the following rule of thumb: If you are puzzled about the nature of thought, belief, knowledge, and the like, substitute for the thought the expression of thought, etc. The difficulty which lies in this substítution, and at

l l

The Road to the Philosophical Investigations 177

the sarne time the whole point of it, is this: the expression of belief, thought, etc., is just a sentence; - and the sentence has sense only as a member of a system of language; as one expression within a calculus. Now we are tempted to imagine this calculus, as it were, as a permanent background to every sentence as written on a piece of paper or spoken stands lsolated, in the mental act of thlnklng the calculus is there - all in a lump. The mental act seems to perform in a mlraculous way what could not be performed by any act of manipu­latlng symbols. Now when the temptation to think that in some sense the whole calculus must be present at the sarne time vanishes, there is no more point ln postulating the existence of a peculiar kind of mental áct alongside of our expression. (BB, 41-2; my emphasis)

ln this passage, the analogy between language and calculus may stop another analogy (counteract its effects), namely, seeing the words on paper as dead and the mental as the propelling englne behind them. Wittgenstein is pointing out that the attempt to justify the "life of symbols" by a mental act that "seems to perform in a miraculous way what could be performed by any act of manipulating symbols" is only a postulation, and not an explanation. The only point of using the calculus metaphor is to stop seeing the postulation as a necessary step. lnstead of thinking that mental acts or a mental model grounded either on images or mental states give life to signs in a mysterious way, one can think that our operating with the 'calculus' of language is the life of signs. The passage quoted above ends in the following way:

This, of course, doesn't mean that we have shown that peculiar acts of consciousness do not accompany the expression of our thoughts! Only we no longer say that they must accompany them. (BB, 42; Wittgenstein's emphasis)

Acts of consciousness, mental images, etc. may accompany thought and its expression; however, the assumption that they must only appears to be a necessary step in our investigation, before the alternative picture offered (language as a calculus) dissolves this apparent necessity (the assumption is not logically necessary, even though we might be strongly inclined to postulate it). It seems, thus, that the analogy between calculus and language is used in arder to dissolve a certain conception that we believe must be the case. What Wittgenstein is doing, thus, is presenting an alternative picture (or analogy) so that the reader stops thinking that a certain conceptlon is necessary.

178 Wittgenstein's Philosophical Development

Such an alternative conception is better than the one criticized to the extent that it avoids a particular puzzlement, to the extent that it stops theorizing, but it is not meant to be, say, absolutely better, for it itself could lead to dubious postulations; it might also mislead. This was, for example, the apparent consequence of the reconsideration of the pictorial conception of language of the T around 1930, when the "permanent background of a calculus" seemed to imply the necessary existence of mental processes (see Chapter 2). Also misleading was the view that language must have a grammar in the sense of the BT. This is why, in the long passage quoted abqve, Wittgenstein mentions "this calculus ... as a permanent background to every sentence" (BB, 42). A misleading view of the calculus is derived from a generálization based on mathematícs and scíence:

When we talk of language as a symbolism used in an exact calculus, that which is in our mind can be found in the sciences and in math­ematics. Our ordinary use of language conforms to this standard of exactness only in rare cases. (BB, 25) .

ln the BT, this "standard of exactness" should be established always when possible. Wittgenstein had already recognized that not all concepts can be seen in such a way (BT, 249). This was already a step beyond the T, where the standard was assumed to be always underlying language. There, analysis would show the "determination of sense" in the form of well bounded concepts and simple names that have reference. ln the BB, however, a new step away from this view is taken. Note that Wittgenstein derived the calculus conception of language from his views concerning mathematics (see Chapter 1, Sections 1.5.4 and 1.2, Section 1.3). Now, one of the points of comparing language with a calculus is precisely to show that the very view that language works as a calculus creates philo­sophical puzzles. The last passage quoted continues like this:

Why then do we in philosophizing constantly compare our use of words with one following exact mies? The answer is that the puzzles which we are trying to remove always spring from just this attitude towards language. (BB, 25-6)

The expression "always spríng" might be an exaggeration. An exaggera­tion that did not escape Sraffa's criticism. ln his comments to the BB, he quotes "language as exact calculus" and "always spring from just this attitude towards language" and gives the following comment: "[ ... ]

The Road to the Philosophical Investigations 179

Have you [Wittgenstein] found out whether these puzzles have in fact arisen out of this attitude to language, have you made sure that they did not exist before anyone took that attitude, etc?" (in Venturinha (2012)). However, for us here the relevant point is that Wittgenstein's purpose in the passage is to criticize "the attitude" of taking language as a calculus. This was also his own attitude in the BT, of course: "l compare the application of language with that of the calculus of multiplication" (BT, 418). Note that now his own attitude in the BT is taken as the source of the puzzles that heis "trying to remove." Thus, the views of the BT are also under attack.

4.2.3 The-genetic method and grammatical mistakes

ln arder to understand Wittgenstein's application of the genetic method in the BB, one needs to get clear about the general target of the method there. Philosophy does not begin, of course, with empirical investiga­tions; it begins usually with what seem to be necessary postulations concerning "the mind." The first step into philosophical theorizing generally takes the form of an assumption concerning "a peculiar mental activity" (BB, 39); this quickly bring us to the postulation of "a gaseous or ethereal" mental world (BB, 47), a "psychical mechanism" (BB, 12), or a "mind-model" (BB, 4). 3 Wittgenstein wants to show that it is the intro­duction of "the occult character of a mental process" (BB, S), which is nothing more than the use of '"ethereal objects' as a subterfuge" (BB, 47), that fuels philosophical theorizing. lf one gets clear about this, one may see "how the problem of the two materiais, mind and matter, is going to dissolve" (BB, 47; my italics). lf such a problem dissolves, one can well imagine that the problems related to idealism, realism, and solip­sism might also dissolve (BB 47-74). lt is this first step in the direction of a "mind model," and its subsequent unfolding (idealism, solipsism, etc.), that Wittgenstein will try to dismantle with the application of the genetic method. First, it must be noticed that it is not a necessary step, for it does not follow from the supposed premises that lead to it, even though we are temped to think that it follows. Philosophical 'musts' are suspicious (BB: 12, 25, for instance).4 We are tempted to take the first step into philosophical theorizing because we are very often misled by analogies suggested by the use (grammar) of expressions (BB: 7, 9, 23, 26, 29, for instance) and peculiar trains of thought. They are 'pecu­liar' because they seem to be prompted by logical compulsion, while the compulsion is actually psychologícal, the result of a puzzling situ­ation. Second, the temptation is to be overcome - thus, the problem dissolved - by remembering how we actually use our expressions and

180 Wittgenstein's Philosophical Development

how those expressions are interconnected (BB: 12, 25). ln the BrB and later, the interconnection of words (the system of language, one could say) will be shown to have a background of practices in our lives, a background that we are, again, prone to forgetting (this is also why the anthropological view enriches the genetic method). 1 come back to this issue in the next section.5

One can, thus, think that one of the major results of Wittgenstein's application of the genetic method is that philosophers might see their "explanations" in a different way (say, the aspect might change). They might see the mythological charact~r of their postulations (be it a mental model, a platonic realm, etc.). Philosophers are satisfied with their explanations only at the price of not being clear ab6ut them, since they merely push the questions one step back. A parenthetical remark in the PG makes the point:

(What kind of propositions are these [Wittgenstein's elucidations]? They are like the observation that explanations of signs come to an end somewhere. And that is rather, like saying "How does it help you to postulate a creator, it only pushes back the problem of the beginning or the world". This observation brings out an aspect of my explanation that 1 hadn't noticed. One might also say: "Look at your explanation in this way - now are you still satisfied with it?"). (PG, §53)

lf one arrives at the postulation of contents of the mind in arder to explain meaning, one could also ask for the conditions of possibility of the workings of those, and so on (more about this in section 3.2). Wittgenstein's strategy in the BB is to make us see some traits of philo­sophical "explanations" in arder for us to become suspicious of them and ultimately, give up the old philosophical "explanatory" goals.

The quick assumption of a mind model, which seems to ask for philo­sophical explanations, is dealt with by means of various indications of misleading analogies that lead one to it. Wittgenstein suggests, for instance, that "we easily forget that the word 'locality' is used in many different senses" (BB, 8) and because we forget this we are led to think that we can, analogically, always ask for the locality of something as if there was only one sense of locality. We might ask, for instance, "where do you see the visual field?" based on the apparently analogous ques­tion "where do you see the image of a tree in your visual field?" This move based on an analogy is called a 'grammatical misunderstanding' by Wittgenstein (BB, 8). The misunderstanding consists in proceeding

The Road to the Philosophical Investigations 181

by "a grammatical analogy without having worked out the analogy in detail'1 (BB, 9). The misunderstanding is called 'grammatical' because what misleads us is a fixed way of looking at certain uses of words in language (Wittgenstein also calls what generates the misunderstanding a 'picture', a 'metaphor' or a 'símile'). We believe that because a word has a certain use in a given context it preserves its sense in all other contexts, and so we extend its use based on what looks similar to us:

When words in our ordinary language have prima facie analogous grammars we are inclined to try to interpret them analogously; i.e. we try to .make the analogy hold throughout. (BB, 9)

Note. the plural of the word 'grammar' in this passage. lt suggests that it is merely a synonym of use; the plural makes clear that we have certain intervals of use, as it were. Sometimes those intervals are bridged by means of the extension of analogies. lt is the analogy that is at the bottom of the question that was asked. The problem is that when we extend an analogy we may not know what we are talking about any longer; we may get lost in our ways of expression. Therefore, it is important to indicate when such analogical moves happen and look at them more carefully. So, for instance, the problem of the "locality of thought" has, according to Wittgenstein in the BB, various analogies at its origin. One may come to believe that the "head is the place of thinking" because of the false analogy between thinking and body activities:

Perhaps the main reason why we are so strongly inclined to talk of the head as the locality of our thoughts is this: the existence of the words "thinking" and "thought" alongside of the words denoting (bodily) activities, such as writing, speaking, etc., makes us look for an activity, different from these but analogous to them, corre­sponding to the word "thinking". (BB, 7)

Since one doesn't see a single activity that we can call 'thinking' and because of the distinction between "really thinking" and "looking like thinking", one is inclined to assume that thinking is like running, but not visible (since the visibility of thinking only indicates that someone looks as if she was thinklng). This may be the case. Perhaps it happens like this. Another possibility is the following:

We say, "the thought is not the sarne as the sentence; for an English and a French sentence, which are utterly different, can express the

182 Wittgenstein's Phílosophical Development

sarne thought". And now, as the sentences are somewhere, we look for a place for the thought. (BB, 7)

Our train of thought starts with a 'grammatical remark,' i.e. a trivial distinction between thought and sentences in different languages, and lead us to assume a thought as 11something" that underlies sentences in different languages and remains the sarne under different mutations, even though this does not follow from the distinctlon we made. Then, a new step could extend the analogy further: since sentences are some­where thoughts also must be somewhere.

As the examples above already suggest, Wittgenstein's strategy in the BB is to describe the genesis of philosophical mistakes in terms of misleading analogies. The topics discussed in the BB are always related to misleading analogies that supposedly lead philosophers to formu­late philosophical problems (see BB: 1, 7, 9, 23, 26, 29, 35, 48, 56). The book opens with the description of the misleading comparison between no uns that reter to o bjects and the ones that don't. The second discussed analogy is the one between the locality

1 of sentences and thought (BB,

7); the third analogy discussed is the one between being conscious and knowing something (BB, 23); the fourth is the example of the meas­urement of time from Augustine (BB, 27); the fifth is between 'to say something' and 'to mean something' (BB, 35); the sixth is related to different uses of 'subjective' that are taken as the sarne (BB, 48); the seventh analogy discussed is the one between metaphysical and empir­ical propositions (BB, 56).

This strategy is combined with the indication of moments ín the phil­osophical reasoning where one is inclined to formulate a philosophical question based on unclear assumptions and a 'grammatical remark' that is, as Russell writes, "so simple as not to seem worth stating." Wittgenstein tríes to indicate the moment when one takes the first of a series of false steps. These are indicated by expressions like "tempted to ... " (for instance, BB: 3, 5, 23, 33, 39, 56, 64, 71) or 11one is inclined to ... " (BB: 2, 39, 41, 73), or 11why one is irresistibly tempted to ... /1 (BB, 60) or "the source of puzzlement is ... " (BB, 59), or 11this creates the illu­sion that ... " (BB, 69), or "connected to that peculiar temptation is ... " (BB, 72) etc. Wittgenstein also describes his philosophical activity as "removing the temptation" (for instance, BB, 41).6

Thus, Wíttgenstein's genetic method is designed to bring back, for instance, Russell to where philosophy starts. Russell says: "the point of philosophy is to start with something so simple as not to seem worth stating, and to end with something so paradoxical that no one will believe it" (The Philosophy ofLogical Atomism, 52). Wittgenstein's point is

The Road to the Philosophical Investígatíons 183

that we only end up stating something "so paradoxical that no one will believe it" because, right at the beginning, we are puzzled by false anal­ogies, comparisons, or pictures. The only way to get to a philosophical paradox, in Wittgenstein's view, is through a false train of thoughts.

The genetic method in the BB also indicates why Wittgenstein insists on describing and scrutinizing the uses of words (their grammars). The point is not to tabulate rules in order to discover the limits of sense. It is, rather, to avoid confusions, to avoid temptations in our bad reasoning:

Let's sum up: If we scrutinize the usages that we make of such words as "thinking", "meaning", "wishing", etc., going through this process rids us of the temptation to look for a peculiar act of thinking, inde­pendent of the act of expressing our thoughts, and stowed away in some peculiar medium. (BB, 43)

Severa! uses of a word are in question as attested by the plural ín the passage above (usages). We could understand, thus, grammar as use in the BB (BB, 23), or as the overview of various uses or, like a little later in the BrB, as the details of usage (BrB, 155) or variety of uses (BrB, 116; §64). Such an overview can be called 'surveyable representation' (ueber­sichtliche Darstellung), and not 'perspicuous' as in the PR and in the BT; for 'perspicuous' expresses the idea of a notation of a list of rules that replaces it (Chapters 1-3).

One investigates the grammar of words (various usages) in order to avoid philosophical puzzlements generated by false analogies and mísleading trains of thought originating in "certain fixed standards or our expression" (BB, 43). ln order to make this point clearer, let's look at two examples of false analogies and misleadíng trains of thought that Wittgenstein uses in order to illustrate his method in the BB. The first example (Augustine and time) is simplest and illustrates the operations of Wittgenstein's method in a nutshell. The second example is more complex and points to Wittgenstein's temptations in 1930. After this, in section 3, I discuss some shortcomings of the method in the BB and how they might have led to Wittgenstein's innovations in the BrB.

4.2.3.1 Augustine and time: a false analogy

An example of a misleading analogy that can engender a philosophical problem is present in the Augustinian worry concerning the measurement of time. Wittgenstein describes Augustíne's worry in the following way:

How is it possible that one should measure time? For the past can't be measured, as it is gone by; and the future can't be measured because

184 Wittgemtein's Philosophícal Development

it has not yet come. And the present can't be measured because it has no extension. (BB, 26)7

This is, for Wittgenstein, a simple example of what is very common in philosophy: misleading analogies create puzzles. As it happens so often in philosophy, after the presentation of the "problem" one could feel tempted to give a counterargument to prove that we can measure time; a different person (say, Moore) could point to a watch and argue that this shows that Augustine's skepticism conceming time must be wrong, and so on. So a "philosophical dialectjc11 could develop. Wittgenstein's strategy is to bring attention to the use of words that may not be present to the reader when she first feels puzzled by the argumerit. So she needs to attentively remember the uses of 'measurement.' One can easily remember, ln this case, that we measure time and length in different ways (say, with a watch and with a ruler). Once this is remembered, it is easy to see that Augustine seems to have mixed these different ways of measuring, for he seems to want to measure time w ith a ruler, which would be possible only with an object 1 in front of hlm. So Augustine wrongly uses an analogy between measuríng time and space.

4.2.3.2 Words, sentences, things and something between (and the Tractatus)

According to Wittgenstein, a common misleading analogy that is at the bottom of philosophical problems is the one between groups of nouns that are explained in different ways. 8 Some nouns are explained by verbal definitions, others by 'ostensive definitions,' i.e., by pointing to something. lfwe had only verbal definitions, one could think, we would not be able to define everything, for we would get involved in an infi­nite regress. Thus, one might think, ostensive definitions are somehow more fundamental. However, not all words can be explained in this way (one does not point to the concept 'number'). But one may expect that there is "something" that corresponds to nouns, the 'meaning,' even when one cannot point to any object, even when one cannot give an 'ostensive definition1 for it:

(We are up against one of the great sources of philosophical bewil­derment: a substantive makes us look for a thing that corresponds to it.) (BB, 1)

As Frege does ln the Die Grundlagen der Arithmetik, one can actually start a philosophical investigation with the question conceming which kind

The Road to the Philosophical Investigations 185

of thing a number is.9 Of course, what is assumed by Frege as unprob­lematic, namely, that numerais mean numbers and that numbers are 'things,' is seen by Wittgenstein as already an unclear step.10 The 'thing' is considered the meaning of the noun in a way analogous to which John is considered the meaning of 'John.' Once we assume that there must be a 'thlng' that corresponds to the numeral, it seems evldent that we have to flnd out what the 'thing' is.

Frege's justiflcation for this step is that mathematics does not deal with slgns on paper (BB, 4), but with numbers or with what gives life to the signs, the meaning (Fregean 'sense'). So someone like Frege comes to believe that what gives life to mathematical propositions and to the signs of númbers on paper must be "something" immaterial, for the addltion of more signs on paper could not "make the propositlon live":

And the conclusion that one draws from this is that what must be added to the dead signs in order to make a líve proposition is some­thing immaterial, with properties different from ali mere signs. (BB, 4)

Frege purported to explain "the life of signs" by assuming that it was the objective thought the source of thís life. For Frege, it is a condition of logic that thoughts are objective.U But Frege leaves unexplained how a thought can give lífe to signs without being psychological or subjec­tive. It is an assumption, one that apparently we have to make.

One can also investigate the meaning or the relation of meaning using tools of psychology. This is what Russell does in The Analysis of Mind: supposedly, what constitutes the meaníng is a "mechanism of association" (BB, 4).12 Since one cannot find objects that correspond to some words, one is led to think that what gives meaning to words are the images in the mind. If someone gives me a word or a sentence, it seems that 1 don't have enough to go on if 1 don't know how to interpret the words:

If I give someone the order 'fetch me a red flower from that meadow', how is he to know what sort of flower to bring, as I have only given him a word? (BB, 3)

Because the words seem to be insufficient ("signs are dead") to convey exactly what we mean, it seems that we need to postulate images or other so to speak mental entitles in order to explain how someone can follow an order:

186 Wittgenstein's Philosophical Development

We are tempted to tbink tbat tbe action of language consists of two parts; an inorganic part, tbe bandling of signs, and an organic part, wbicb we may call understanding tbese signs, meaning tbem, inter­preting tbem, tbinking. (BB, 3)

Witb tbis move, we seem to be involved in an investigation of very mysterious entities tbat need to be added to signs in arder to give tbem meaning: tbe mind, images, intention, etc (or "abstract entitíes," like Frege). Wittgenstein's suggestion at tbis point is to replace tbe image tbat supposedly gives meaning to signs by outward objects like "a painted or modeled image" (BB, 5). Tbis can bring us to see tbe nature of tbe "occult cbaracter of tbe mental process":

... wby sbould tbe written sign plus tbis painted image be alive if tbe written sign alone was dead? - ln fact, as soon as you tbink of replacing tbe mental image by, say, a painted one, and as soon as tbe image tbereby Iases its occult cbaracter, it ceases to seem to impart any life to tbe sentence at ali. (lt was in fact just tbe occult cbaracter of tbe mental process wbich you needed for your purposes). (BB, 5)

Tbe "occult cbaracter" of tbe mental process was important as an unex­plained, but supposedly self-explaining, sometbing tbat sbould guar­antee a way-out for our difficulty of identifying wbat gives meaning to dead signs (as tbe last possible step of regression). Tbe "occult cbar­acter," bowever, does not fare better tban tbe idea of a creator: "it only pusbes back tbe problem" (PG, §53; quoted above). Of course, wbat Wittgenstein tbinks one needs to jettison in tbis wbole train of tbougbt is tbe very idea tbat nouns always stand for sometbing (tbe 'meaning') and tbat from tbe distinction between meaningful and non meaningful signs it "follows" tbat signs in tbemselves are dead.

Wittgenstein bimself in tbe T, bowever, tbougbt tbat tbe meaning of a word is an object (T 3.203). Taking tbe meaning to be an object migbt lead one to postulate simple objects, wben one tries to understand bow one can say sometbing false. lf we describe a fact tbat is not tbe case, if we imagine wbat does not exist, "' ... we imagine non-existent combina­tions of existing elements"' (BB, 31), one is inclined to tbink tbat "tbe elements, individuais, must exist" (BB, 31).

Wittgenstein does not furtber discuss tbe details of bis position in tbe T ln tbe BB. His interest lies somewbere else: to discuss tbe tempta­tions, similar to bis own in 1930, that migbt lead anyone to postulate a

The Road to the Philosophlcal Investigations 187

shadow between fact and sentence. This is the subject of a long part of the BB (pp. 31-44). lf the simple elements are not postulated, as was the case with Wittgenstein ln 1929-30 (Chapters 1 and 2), one might then be tempted to think that the explanation lies not in the elements, of course, but in something between thinking and the fact:

We may now be inclined to say: "As the fact which would make our thought true if it existed does not always exist, it is not the fact which we think." (BB, 31)

Here we have already a confusion. We think that the distinction we are making is ánalogous to "it isn't Mr. Smith who hangs ín the gallery, but his picture" (BB, 31). ln philosophy, however, we are inclined togo on (we take the next step). What is it, then, the thing that we think (the "object of thought")? It seems that something might take the place of the fact:

The next step we are inclined to take is to think that as the object of our thought isn't the fact it is a shadow of the fact. There are different names for this shadow, e.g. "proposition", "sense of a sentence." (BB, 32)

Such a shadow is called by Wittgenstein a "picture of similarity" (BB, 36). Perhaps, the mentioned shadow could be, for instance, Russell's "image-proposition" or "belief" (AM, chapter XIII) or even Frege's "sense." Wittgenstein's example in his Iectures was the Cambridge Iogi­cian W. E. Johnson. Moore writes:

One chief view about propositions to which he was opposed was a view which he expressed as the_ view that a proposition is a sort of 'shadow' [ ... ] He attributed this view to W. E. J ohnson, and he said of it that it was an attempt to make a distinction between a proposition and a sentence. (M, 260)13

The shadow that Wittgenstein has in mind is not a Tractarian Satz. A Satz is a sentence. The point of projections in the T was precisely to avoid such a shadow by means of a shared forro between a Satz and a fact. This is clear in the following passage of the BB:

A plane projection of one hemisphere of our terrestrial globe is not a picture by similarity ... It would be conceivable that I portrayed

188 Wittgenstein's Philosophical Development

some one's face by projecting it in some queer way though correctly according to the adopted rule of projection, on a piece of paper, in such a way that no one would normally call the projection "a good portrait of so-and-so" because it would not look a bit like him.

If we keep in mind the possibility of a picture whích, though correct, has no símílaríty wíth its object, the ínterpolatíon of a shadow between the sentence and reality Iases ali point. For now the sentence itself can serve as such a shadow. The sentence is just such a pícture, which hasn't the slightest símilaríty with what it represents. (BB, 37; my emphasis)

The whole passage expresses agreement with and is anobvious refer­ence to the T. Before explaining that the internai refation between propositlon and reality (T 4.014) consists in rules or laws of projection, Wittgenstein observes in the T:

At first sign a proposition (Satz) - one set out on the printed page, for example - does not seem to be a picture of the reality with which it is concerned. But neither do written .notes seem at first sight to be a picture of a piece of music, nor out phonetic notation (the alphabet) to be a picture of our speech.

And yet these sign-languages prove to be pictures, even in the ordi­nary sense, of what they represent. (T 4.011; my emphasis)

Thus, in the T Wittgenstein does not defend the criticized idea of a 'picture by similarity.' Wittgenstein goes on ln the T to explain the notion of projection by means of a "piece of music" (T 4.011) in the remarks that follow the one quoted above (4.012-5). Thus, projection obviously is intended to explain how words can represent things and how sentences (which are facts) can describe facts. Hence, the point, already in the T, is that "the interpolation of a shadow between the sentence and reality loses ali point," for the proposition ("one set out on the printed page, for example") is itself a picture. We don't need a "picture by similarity" or a "shadow" in the T because sentences are pictures - even if a sentence "does not seem to be a picture of reality" (T 4.011; quoted above). We have to note, thus, two things. First, it is not correct to think that the T is assuming any kind of mentalism in arder to explain how propositions depict reality.14 Second, that the target of the BB in the discussion of a "shadow" is not the T. The obvious target is, as Wittgenstein indicates, Johnson (M, 260) - but also for instance Frege in Der Gedanke.

The Road to the Philosophlcal Investigations 189

More important, however, is the broader target: a way of thinking, of reasoning, and some steps that Wittgenstein was inclined or tempted to take in 1930 (see Chapter 2, Section 2.2). A major problem with Russell's causal theory was that it excluded the normativity of language: the internai relation between thought, proposition, and reality. ln 1930 Wíttgenstein called it "the intentional element." One of the difficulties Wittgenstein found was that rules of projection might not be equiva­lent to what is meant. Thus, it seemed that intention was an intangible element. Moreover, rules of projection did not seem sufficient to deter­mine the specificity of a described fact and they seemed to lead one to an infinite regress of interpretations. With those things in mind, it is not difficult to understand Wittgenstein's point here:

Suppose we said "that a picture is a portrait of a particular object consists in its being derived from that object in a particular way". Now it is easy to describe what we should call processes of deriving a picture from an object (roughly speaking, processes of projection). But there is a peculiar difficulty about admitting that any such process is what we call "intentional representation". For describe whatever process (activity) of projection we may, there is a way of reinterpreting this projection. Therefore - one is tempted to say - such a process can never be intention itself. For we could always have intended the opposite by reinterpreting the process of projection. (BB, 33; my emphasis)

What underlies such a misleading train of thought is that "we imagine the processes of saying and meaning to take place in two different spheres" (BB, 33). What "tempts us" to think this is "the analogy between the forms of expression" (BB, 35). The forms of expression that Wittgenstein has in mind are "to say something" and "to mean something," which, he claims, "seem to refer to two parallel processes" (BB, 35). Now, the many forms of postulation of various kinds of shadows might be prompted by the question "What is the object of a thought?" Such a question, Wittgenstein points out, "is already the expression of several confusions" (BB, 35), whose origins he, of course, investigates in the BB. Note that this question was one of the questions that prompted Wittgenstein's investi­gation in 1930. Thus, Wittgenstein is applying the genetic method to his own views from 1930 in the BB. His own temptations work as exemplars. He expects us to give up similar temptations once we see the misleading analogies and misleading series of thoughts that prompt them.

190 Wittgensteín's Philosophical Development

4.3 The Brown Book: the anthropological view and the genetic method

4.3.1 From the Blue Book to the Brown Book

The BB was an interesting experiment, for Wittgenstein developed in it a systematic application of the genetic method, as seen above. It was also important ln Wittgenstein's development, as we will see, because it indicated some shortcomings of the application of the method. One shortcoming is trivial; one, non-trivial. The trivial shortcoming relates to undeveloped tools that could be used in the application of the genetic method. Wittgenstein introduces the idea of 'primitive l~nguages,' but does not apply it (BB, 17). He first applies the idea in his 1934-5 lectures, at the time that he was already dictating the BrB. The BB, as he writes in a letter to Russell from 1935, was "meant only for the people who heard the lectures"; many points in it, therefore, "are just hinted at" (CL, 269). Since the discussions surrounding the BB involved themes specific to the BrB (primitive languages), it seems plausible to think that it works as a bridge to the later book. It indicated a new strategy that was worth following.

The non-trivial shortcoming of the BB is related to the very idea that misleading analogies underlie philosophical questions or puzzlements. It is difficult to determine exactly which analogies underlie a philosoph­ical puzzlement. Some misleading analogies pointed by Wittgenstein may be more interesting than others, depending on the philosopher's recognition of them as the source of bis missteps. There may also be various analogies or pictures operating in the background of philosoph­ical puzzlement. What is the best way to indicate a supposedly major misleading analogy? Another difficulty is that a philosopher may not see the false analogies pointed out as relevant for his case and so he may not see himself reflected in the "mirror" that Wittgenstein's writings are supposed to be. Moreover, the mirror only works if Wittgenstein makes clear where exactly a misleading analogy drives one astray. He recognizes some of the mentioned difficulties:

When we say that by our method we try to counteract the misleading effect of certain analogies, it is important that you should understand that the idea of an analogy being misleading is nothing sharply defined. No sharp boundary can be drawn round the cases in which we should say that a man was misled by an analogy. The use of expres­sions constructed on genetic patterns stresses analogies between cases often far apart. And by doing this these expressions may be

The Road to the Philosophical Investigations 191

extremely useful. It is, in most cases, impossible to show an exact point where an analogy begins to mislead us. (BB, 28, my emphasis)

How can one correctly identify the analogies that mislead us and correctly present the genesis of philosophical confusion if it is "in most cases, impossible to show the exact point where an analogy begins to mislead us"? This question leads us, 1 think, to one of the central moti­vations for the style of the BrB. ln the BrB, Wittgenstein begins with the description of a simple primitive language in which one builder calls for slabs, bricks, columns and cubes and another brings the objects when they are caHed for. This language is gradually extended (BrB §2). One of the reasons' to start with a very simple language and gradually extend it in the BrB is, 1 think, that this procedure better indicates where the misleading analogies take hold or "begin to mislead us," as Wittgenstein says (BB, 28). Each extension of the language introduces new grounds for puzzlement.

The anthropological view in the BrB fills the gaps of the use of the genetic method in the BB (this I called the trivial shortcoming of the BB). Even the more complex languages of the BrB that gradually appear are associated with tribes or societies (BrB §§1, 5, 30, 32, 39, 40, 46, 48, 49, 58, 59). ln fact, ali sections of part I of the BrB are directly introduced in a minimally described form of life ar indirectly associated with one. Those primitive languages are imbedded in a form of life simpler than ours. They are, therefore, easily surveyable because in them "forms of thinking appear without the confusing background of highly compli­cated processes of thought," as Wittgenstein suggests already in the BB (BB, 17). The language-games, thus, are used to frame possible objec­tions, which are examples of how philosophical reasoning comes into being by false analogies and misleading trains of thought in our complex languages. Note that each new extension of the simpler first language, i.e., each new levei of complexity, introduces new philosophical tempta­tions connected to the new kinds of words, the new tools, introduced. This means that genetic method and anthropological view are working in association. Besides gradually introducing leveis of complexity in the BrB, the anthropological view indicates that the use of words is always part of an environment: we should not forget to pay attention to the surroundings where words and human activity takes place.

The indication of misleading analogies appears in the BrB in the form of parenthetical remarks after the presentation of the first language game and its extensions (BrB: 77, 79, 80, 82, 94). The parenthetical remarks, however, don't cover the whole book.15 After some language

192 Wittgenstein's Philosophical Development

games are introduced, Wittgenstein simply indicates the point where philosophical problems might begin (as in the BB) with the expressions "inclined to ... 11 (BrB: 80, 86, 88, 113, 120) "tempts us ... " (BrB, 114), "tendency" (BrB, 117), "it seemed to us ... 11 (BrB, 124), "tempted to ... " (BrB: 78, 82, 125), etc.

ln the BrB Wittgenstein presents again a reconstruction of the "phys­iognomy of his temptations" related to hís efforts to understand the peculiarity of attitudes, intention, and understanding and meaning around 1930. He thinks, of course, that his temptations might appear in the 'mirrar' because they are exemplary. But this does not mean that Wittgenstein is exclusively examining his míddle period views. One interesting aspect of the genetic method is that it allaws an imagina­tive use of philosophical reasoning (i.e., one can imagine how a philos­opher comes to think what he thinks). Based on his own tendencies and the examples of other philosophers (mostly Russell and Frege)16,

Wittgenstein can imagine how philosophers are inclined to think, what would they be tempted to ask or object, etc. This possibility is explored in the BrB (and even more so,in his later works) in the form, for instance, of interlocutors that try to construct philosophical ques­tions and justify their positions. The confessional and the imaginative aspects of the method can be seen in the example below.

4.3.2 An example: language-games, comparison, and recognition

ln this section I present Wittgenstein's genetic method in the BrB by means of an example: the problem of 'direct recognition.' ln this example one can see how his philosophy in the BrB is a philosophy grounded on the genetic method with a more dialogical style than before.17 The example might also be relevant for us to see the point of the language-game of the shopkeeper in PI §1 discussed in Chapter 5. Moreover, it also illustrates the use that Wittgenstein makes of the idea of "family resemblance."

As already indicated, Wittgenstein in the BrB strategically starts with the description of a very simple (or primitive) use of language (language-game) and gradually introduces more complicated ones. The introduction of new elements, the extension of language-games, also introduces new opportunities for the formulation ofphilosophical prob­lems (new temptations for a philosopher to begin theorizing). Consider, for instance, the move from the first to the second language game in the BrB. The first language-game is a language of only five words ('brick', 'slab', 'cube', 'beam' and 'column') used by builders A and B. If one finds

The Road to the Philosophical lnvestlgations 193

it difficult to think of it as a language, say, because it is too primitive, one should imagine it as the "entire system of communication of a tribe in a primitive state of society" (BrB, §5). Builder A calls, for instance, 'Brick' and B brings him a brick. A philosopher may be tempted to think that when A calls 'Brick' he must mean "Bring me a brick!" and so she may believe that there must be a complex mental state that accompanies the word 'brick' when A says it. The second game introduces numerals and now the philosopher might be tempted to think that there is a "something" that must be referred to when numerals are used - in a way similar to 'brick' in the first language-game.

Wittgenst~in goes on and introduces other language-games with other kinds' of words and sentences (words and sentences with different uses, with different functions in the form of life described): indexicals, proper names, samples, questions and answers, etc. ln §13 Wittgenstein introduces orders that are given by means of symbols. Someone B is shown by a person A, for instance, a colored geometrical figure and B is supposed to bring A an object. Here one is inclined to ask: how can B catch the right object? How can he compare them? What makes it possible for someone to compare the object to the given pattern?

First, Wittgenstein counteracts the temptation of thinking that when we compare a sample to an object there must be always one single "thing" that characterizes all comparisons that we make. Suppose that B has to bring a cloth of the color of the sample that is shown to him by A. B compares the sample to the cloth and recognizes it as the correct one. Various things may happen when B does it:

1) A shows the sample to B. B fetches the material 'from memory' (BrB, §14)

2) "A gives B the sample and B looks from the sample to the materials on the shelves from which he has to choose" (BrB, §15).

3) "B lays the sample on each bolt of material and chooses that one that he cannot distinguish from the sample, for which the differ­ence between the sample and the material seems to vanish" (BrB, §16).

These three examples show already that 'comparing' is used in various ways. This should be a warning to the philosopher who wants to find the meaning of comparing - to the philosopher who assumes that 'comparing' must always mean something particular in all uses of the word. Now, cases 1-3 will not certainly stop the philosopher from looking for the common element or to the essence of comparing, for he

194 Wittgensteín's Philosophícal Development

might think, for instance, that in ali the cases there is an underlying image used ln the comparison. ln the first, he may think, there is a memory image; in the last two examples, a sample, but he may go on and thínk that the image of the memory is what is really important and that it helps B even when he uses samples. The samples in 2 and 3 were only externai superfluous toais, a philosopher could think, while the mental image was the real criterion.

But what happens if the material to be brought by B should be slightly darker than the sample (BrB, §17)? This could, again, be seen as supporting the supposition that the irpage in the mind was doing the real work in the comparison, for how could the samples by themselves be the guides if one looks for something darker? So orie could think that the image "in the mind" could be the standard of the comparison, something líke the ultimate interpretation. Thís supposedly ultimate standard would be accessed, then, by memory.

But when we say that memory accesses the standard of the compar­ison, the image, we are inclíned to think that there is, again, one single mental activity that underlíes fetching the material from memory. It is important for the philosopher that. there is (must be) one single activity that ís 'catching the material from memory,' for he wants to give an explanation of what comparing consists in. He is inclíned to assume that the mental image gives us a necessary condition for comparing. This would then justify its postulation: "in arder to explain the comparison, we must postulate the grasping of an image or mental act, etc."

When we look at cases that are described as "catching the material from memory" we encounter, again, a variety of things. Wittgenstein gives three cases of catching something "from memory":

1. "B has a memory image before his mind's eye when he goes for the material. He alternately looks at materiais and recalls his. He goes through this process with, say, fíve of the bolts, in some instances saying to himself "Too dark", in some instances saying to himself, "Too light". At the fifth bolt he stops, says, "That's it" and takes it from the shelf" (BrB, 85)

2. "No memory image before B's eye" (BrB, 85). He shakes his head and feels a mental tension each time that he looks at a bolt. His tension relaxes when he fínds the corresponding material. He nods his head and takes it.

3. "B goes to the shelf without a memory image, looks at five bolts one after the other, takes the fifth bolt from the shelf" (BrB, 85)

r The Road to the Philosophical Investigations 195

At this point an imaginary interlocutor objects: "But this cannot be all comparing consists in" (BrB, 86). The interlocutor feels that the most important part is missing in this description of cases 2 and 3, namely, the "essential feature", "a specific experience of comparing and recog­nizing" (BrB, 86). So he says that it cannot be all comparing consists in because the examples don't seem to point to a common element that supposedly must be there.

The problem with the tendency of looking for the "essential feature" is that one concentrates on one example (what is taken to be the "full­fledged case" in BrB, 150-1) and assumes that other examples must work in the sarne way. When one thinks about comparing from memory, one brings to ml.nd a case in which one has an image in one's mind, looks at different objects and chooses. One tends to be fixed on one particular case and one then extends what seems to be characteristic in it to all cases (they seem to be ali analogous, similar; they are part of the sarne picture). But if one pays attention to each particular instance, one may see an unexpected variety of cases:

We hold pieces whose colors we want to compare together or near each other for a longer or shorter period, look at them alternately or simultaneously, place them under different lights, say different things while we do so, have memory images, feelings of tension and relaxation, satisfaction and dissatisfaction, the various feelings of strain in and around our eyes accompanying prolonged gazing at the sarne object, and all possible combinations of these and many other experiences. (BrB, 86)

Of course, the variety of cases and the differences among all the exam­ples does not entail that it is impossible to find one particular experi­ence that is characteristic of all cases. One should not forget, however, that the philosophical issue is not the mere possibility of finding a particular experience. lts need, and then its necessity, is assumed. One should, at this point, at least become suspicious of finding such a partic­ular experience:

The more such cases we observe and the closer we look at them, the more doubtful we feel about finding one particular mental experi­ence characteristic of comparing. (BrB, 86)

But the attention to particular cases brings also another interesting thought. lnstead of the picture of looking for one essential feature of

196 Wittgenstei11's Philosophlcal Development

comparing, we may think that it is enough finding a varlety of "more or less characteristic experiences" (BrB, 86). We may find that "what connects all the cases of comparing is a vast number of overlapping similarities" (BrB, 87) - a "family resemblance" among them. Once we see thís, "we feel no longer compelled that there mustbe some one feature common to them all" (BrB, 87, my emphasis). Instead ofbeingcompelled by the picture according to which all cases of comparing have to have a common element (like, for instance, alcohol in ali bottles of wine) one can think that it is exactly the variety of more or less related examples what characterizes the meaning of some words. One can think of the following metaphor (family resemblance):

What ties the ship to the wharf is a rope, and the rope consists of fibres, but it does not get its strength from any fibre which runs through ít from one end to the other, but from the fact that there is a vast number of fibres overlapping. (BrB, 87)

The role of this last comparison (picture) is simply to break the spell of the analogy with which the interlocutor is working (he assumes that mental experiences and recognition function like alcohol and wine). Wittgenstein wants to point out that "an analogy made our thinking captive and brings it away without resistance11 (BrBG, 156) and that a different analogy may break the spell of what we see as an obvious necesslty. So the new analogy should show that we don't need to think according to the old picture. lf one needs a picture, one may take the new one and 11 feel no longer compelled" by the old one. But Wíttgenstein is not saying that concepts must be family resemblances. He is also not saying that because they are family resemblances we cannot fínd the defínition by means of an essential feature. There is no such argument here. Wittgenstein is simply opposing pictures, for the interlocutor is much too impressed by one specific simile. Wittgenstein's picture, however, is a picture that the interlocutor might not want to deny. How could he deny, for instance, that there are various experiences related to comparing?

The argument until this point does not show any impossibility, any conclusion that the imagined interlocutor must accept. He should only feel doubtful about bis search for the characteristic feature of comparing and recognizing. But he may also feel suspicious about Wittgenstein's procedure. Perhaps, he might think, Wittgenstein is simply lazy and does not want to look for what is important with more determination. A philosopher can always continue looking for bis definitions, no matter

r i

1

The Road to the Philosophical Investigations 197

what is said by Wittgensteín, for one can always believe that one might somehow, somewhere, someday find an essential characteristic for comparing and recognizing. But Wittgenstein could also tum the tables and ask: why should we accept the philosopher's picture of the common feature underlying the use of a word? Well, we don't have to. And why should we think that it is a mental image that guides the comparison if we have examples in which no image is, apparently, required in the comparison? Again, we don't have to; this is not a necessary step to be taken.

The interlocutor's view is that Wittgenstein is using examples of comparing and recognizing that are not appropriate and so making it difficult to find out what the essential feature of ali cases is. If one includes, for instance, !ions among domesticated animais, it may make a definition of 'domesticated animal' difficult or impossible. How can Wittgenstein include case 3 as an example of comparing and recog­nizlng? Maybe it is not. The interlocutor objects:

'But surely in case 14c) [case 3] B acted entirely automatically. If ali that happened was realiy what was described there, he did not know why he chose the bolt he did choose. He had no reason for choosing it. If he chose the right one, he did it as a machine might have done it'. (BrB, 87)

So the interlocutor's strategy is to exclude the example that seems harder to fit in with the cases that he sees as standards. Thus, the interlocutor now wants to defend an essential difference among the examples given. Wittgenstein answers this objection in three steps. First, we don't need to consider B a machine. We can well imagine that he had "personal experiences" when comparing (say, he was feeling pain). Maybe he also had other sensations (muscular, tactile, etc). Second, to what extent giving a reason is a necessary step in the comparison? Why should every choice have a reason? (BrB, 88). Before asking for a reason, we need to be clear about what a reason that makes the choice non-automatic could look like, i.e., "What do we imagine it to be like?" (BrB, 87).

It seems that the interlocutor would only be satisfied with some­thing that we can call "conscious comparing'' of a sample or a memory image with an object whatever "conscious comparíng" exactly means. But, in this case, the interlocutor needs to minimally explain how the comparing takes place. Even the "ideal case" of "conscious comparing" seems to suppose a criterion for the recognition of the correct mate­rial. So what counts as the criterion in the "full-fledged example"?

198 Wittgenstein's Philosophical Development

What seems to give a necessary condition for the recognition is a "specific feeling" that underlies it: "a specific feeling of not being able to distinguish in a particular way between these samples and the mate­rial chosen" (BrB, 87). But the problem with this view of conscious comparing is that it does not look at all like a reason anymore. The obvious problem is the correctness of the feeling: what if I take the feeling of toothache to be the recognition of the cloth? One needs to ask "what connects this particular experience with either?" (BrB, 87), i.e., what connects the experience with the image and the material? Now, if we needed an experience in the1 first place, we would need a new experience to guarantee that the experience we have is the right one. This generates a regress (see Chapter 2)" The difficulty he're makes clear that what we took as a necessary condition (there must be something) was not grounded on logic (the necessary condition does not follow from the analyzed state of affairs). It was not logical compulsion that brought us here, but psychological compulsion (we merely cannot accept that there is no such step). For if we logically needed a 'something,' we would need also 'logically' a someth).ng else, and so on. We could, thus, in principie, think that we must assume not a mental image, but the ímage of the agreement between the mental image and the object recognized by means of the mental ímage (BrB, 88) - one might be very inclined to introduce "private reasons" here.

Third, it is not clear that we can easily and not arbitrarily draw a line between automatíc and non-automatic comparison. lf we look at the examples above (1-3) and to the difficulty of using a "particular experience" as what characterizes non-automatic behavior, we may be inclined to think that the distinction automatic/non-automatic is not clear-cut. ln other words, it is not clear that we can tell in general what is automatic and what is not, even though we can clearly use the distinc­tion in particular cases. One may say, for instance, that B did not act automatically because she thought about different possibilities, called an image of the material in her head, and finally uttered her decision. But in a different circumstance one may distinguish between images that carne to mind automatically and those that did not. So if images come automatically to mind, images cannot be the criterion for the non-automatic choice.

Of course, if the interlocutor counts cases of recognition as cases in which there is a reason to justify the choice, then he can say: "But B didn't really recognize the material as the right one" (BrB, 88). But this is also a very dubious move by the interlocutor. It seems to be merely arbi­trary to simply exclude 3 as a case of recognition. What was B doing ln

s t

e e

a a i. e e 13

f e 1 t E t

5

The Road to the Philosophical Investigations 199

the example if not recognlzing? Would someone explicitly deny that B recognized the material after looking at 5 bolts one after another before taking the fifth (the right one)?

The interlocutor can still come back and argue that his point was that no criterion of recognition was given in the example 3: "But doesn't B lack the criterion by which he can recognize the material?" (BrB, 88). Since Wittgenstein is not making any assumption of a hidden 'reason' or hidden cause, it seems, still, that something is missing, namely, a criterion. As bad as it was, the feeling was at least the semblance of a criterion. Besides this, the interlocutor may refuse the idea of a neces­sary feeling as the criterion for the recognition. He could stop a step earlier and assume, as in case 1, that the image itself is the criterion: "ln 14a) [case 1 above], e.g., he had the memory image and he recognized the material he looked for by its agreement with the image".

Now, if an image is a necessary condition for the recognition, then one can ask for an image of the agreement between the image and the material, and so on. Here we have, again, a regress. If we assume that there must be a last link between the order and the execution of the order that works like a criterion (image, feeling, concept, representa­tion, etc), then we are only assuming something, but we are not proving anything. We can assume whatever we want, but this does not bring us further, for the last assumed link is not necessarily the last one. Why should it be? Assuming a last link is not logically necessary; it is merely the expression of a psychological need (it works like God as the first cause). We can very well go back to describing examples that are cases of family resemblance.

What seems to be peculiar about cases 1 and 2 is the apparent guar­antee of success that the images or feelings can give. They seem to guar­antee a last move, a last interpretation, a criterion that is something inherently correct. Based on this appearance the interlocutor strikes back once more: "But supposíng B bríngs the bolt like in 14c) [case 3] and on comparing it with the pattern it turns out to be the wrong one?" (BrB, 88). Of course, it could have happened, but it could have happened in all the cases described. The image itself does not have any magicai property that guarantees the correct recognition. The idea of inherent correctness is, say, quite convenient, but only because it is a chimera. The interlocutor leaves unexplained how it works and simply assumes that it must work: it must be "self-interpreting" and free of mistakes. But, of course, we don't know any images with such fantastic proper­ties. The interlocutor is simply forgetting an obvious fact about his own supposition that images play the role of a criterion: an image is just an

200 Wittgenstein's Philosophical Development

image. What the interlocutor really wants is something analogous to an image, but does not specify what it is. The supposedly self-interpreting character of mental images does not specify mental images precisely because it is this characteristic that makes it completely different from what we usually call images.

The image itself is, then, no guarantee at all and B may make a místake even in case 1. Again, here, various things may happen:

Suppose in 14a) [case 1] the bolt which B brought back was found not to match with the pattern. Wouldn1t we in some such cases say that his memory image had changed, in others that the pattern or the material had changed, in others again that the light had changed? lt is not difficult to invent cases, imagine circumstances, in which each of these judgments would be made. (BrB, 88-9)

The interlocutor comes back once more. lf all cases are so similar concerning their fallibility, still, it seems that there is an essential differ­ence between them and that Wittgenstein is denying it. Is Wittgenstein denying the difference among the cases? It seems that he is trying to argue that 1, 2 and 3 are the sarne, when they seem essentially different, for different activities take place in them. So the interlocutor objects: "'But isn't there after all an essential difference between the cases 14a) and 14c) [1 and 3]?" (BrB, 89). The essential difference that the interloc­utor has in mind is "internal,/J i.e., given the description of the cases, he thinks that the cases are different because nothing is going on "inside" B's brain, mind or head in case 3.

But Wittgenstein does not want to deny nor assume anything inside B1s mind or brain. Concerning the question above, Wittgenstein answers that surely there is a difference: "- Certainly! )ust that [the difference] pointed out in the description of these cases. -" (BrB, 89). The exam­ples already show that they are different. We can tell the differences by means of the descriptions that were given. But they show their differ­ence in what is described, and not in what supposedly underlies their description. The examples may have more differences than the ones described, but the ones described already indicate how different the three cases are and precisely because they are different the inclination to explain them in the sarne way (as the interlocutor wanted in the first place) is misleading.

lt is clear that the interlocutor can always come back. Wittgenstein changes the topic after p. 89 of the BrB, but the interlocutor still may have different questions to ask, which need to be looked at. One obvious

e a

t e

ç i

The Road to the Philosophical lnvestigations 201

point connected to the discussion above is the question "What happens when B recognizes something as a pencil?" (BrB, 127). This kind of recognition seems also to suggest the necessity of a mental image (one might think: 1 can recognize a pencil as a pencil only because 1 have a memory image of a pencil). This possibility of coming back, looking from a different angle is the reason why Wittgenstein writes later, in the preface of the PI, that the need to go in different directions in his discussions is "connected with the very nature of the investigation" (PI, preface). The philosophical discussion does not necessarily have an end. This is why Wittgenstein's new idea of a surveyable representation (after the BT) is so. important: we must shade light from various angles, for a light also cásts shadows (MS 116, 52).

4.3.3 'Grammar' is synonymous with 'use' or 'description of uses'

False analogies and misleading trains of thoughts take place usually because only one use (grammar) of a word is considered 'standard' and this use is assumed in other contexts. So the attention to the use of words in their surroundings may avoid philosophical puzzlement. Wittgenstein uses the word 'grammar' as a synonym to 'use' in the BrB to indicate the linguistic character of a good number of such puzzles. ln this sense, his investigation is grammatical.

The usage of 'grammar' until the BT pointed to something presum­ably fundamental, at the center of a general idea of language. The usage of 'grammar' after the BB conveniently points to the fact that we are dealing with something rather plain. It indicates that there is not really a discovery to be made when we are dealing with a good number of philosophical problems, for attention to what we do with words in our lives and remembering different uses of words might be sufficient to dissolve them. This was perhaps Wittgenstein's intended use of the word already in 1931; however, if so, his practice did not match his intention (see Chapter 3, Sectlon 3.3). Because the word 'grammar' shows up only in four different passages in the BrB it is useful to look at them. This wíll make clear that ln the BrB Wittgenstein has left behind completely his old views on grammar from the early thirties. 1 indicate the four different contexts where 'grammar' appears with capital letters (A-D).

A) The first passage's context is the discussion ofpropositions about the past and the future. ln the language-game of §56 of the BrB Wittgenstein introduces forecasts. He supposes a simple use oflanguage. One observes traffic lights changing colors and, after a while, makes forecasts about the changes of color. ln this game sentences about the future ("The light

202 Wittgenstein's Philosophical Development

will tum green after turning red") don't look mysteríous or problematic. But under certain circumstances one may be puzzled and ask questions like this: "Where does the present go when it becomes past?" (BrB, §561

107). Under certain circumstances such a question seems very important, but under different circumstances we Would "wave it away as nonsense" (BrB, 107). We come to these puzzling circumstances because "we are led into puzzlement by an analogy that irresistibly drags us on" (BrB, 108). What we do before getting into this puzzle is to look at forms of expression like the following: "we talk of the future event of something coming into my room, and also of the f,uture coming of this event" (BrB, 108). This may bring us to another metaphor: we may imagine that the coming of events is analogous to an object arriving (say/a floating log) on the water. The log still to come is something in the future, the log we see in front of us is the present and the logs down the stream are the past. One may be inclined to bring this analogy a step further again and become really puzzled about the events to come:

... comparing future events with pa,st events, one may almost be inclined to say that though the past events do not really exist in the full light of day, they exist in an underworld into which they have passed out of the real life; whereas the future events do not even have this shadowy existence. We could, of course, imagine a realm of unborn, future events, whence they come into reality and pass into the realm of the past; and, íf we think ín terms of this metaphor, we may be surprised that the future should appear less existent than the past. Remember, however, that the grammar of our temporal expres­sions is not symmetrical with respect to an origin corresponding with the present moment. This is the reason why it has been said that propositíons concerning future events are not really proposi­tions ... (BrB, 109; my emphasis)

One can clearly substitute 'use' for 'grammar' above. The use of words related to past and future are not symmetrical to an origin, the present moment. We can obviously remember that Brazil won the World Cup in 2002, but it is nonsense to say in 2013 that one remembers that Brazil will win the World Cup in 2022. Now, based on this asymmetry, one may also argue that there are not really propositions about future events. Wittgenstein responds to this philosophical thesis with the following:

And to say this is ali right as longas it isn't meant to be more than a decision about the use16 of the term "proposition"; a decision which,

The Road to the Philosophical Investígations 203

though not agreeing with the common use of the word "proposi­tion", may come natural to human beings under certain circum­stances. (BrB, 109; my emphasis)

So the philosopher may say (contrary to ordinary use) that propositions about the future are not really propositions (for instance, because we cannot tell whether they are true or false), and could have the appear­ance of an important scientific discovery. We could, however, obviously say that there are propositions (sentences) about the future ("Brazil will win the 2022 world cup"). According to Wittgenstein, the only danger in the obser.vance of the "grammar of temporal expressions" would be the philosópher's belief that "he imagines that he has made a kind of scientific statement about 'the nature of the future"' (BrB, 109).

B) The second passage indicates that the use of a word may bring one to postulate a necessary "condition of possibility" for something to take place. The use (grammar) of a word may suggest a necessary "interme­diary step":

We are treating here of cases in which, as one might roughly put it, the grammar of a word seems to suggest the 'necessity' of a certain intermediary step, although in fact the word is used in cases in which there is no such intermediary step. Thus we are inclined to say: "A man must understand an arder before he obeys it", "He must know where his pain is before he can point to it,11 "He must know the tune before he can sing it", and suchlike. (BrB, 130)

One of the uses of 'understand' is, for instance, the following. Somebody tells me "You have to go to the basement and find five pictures like the one hanging on the corridor and then bring one to Mr. Olms, two to Mr. Brick and three to Mr. Sachser". Before following this arder, one would think about the picture on the corridor, people to whom one should bring the pictures, etc. So it seems natural to think that one must understand an arder before following it when we look at this specific use (grammar) of understanding, in this specific context. Of course, one can say that one has to understand an arder before following it. But one may be inclined to take this example as the "full-fledged case" (see BrB, 150-1) and think that the intermediate step of understanding (imag­ining, comparing and thinking about what is involved in the arder) musttake place. But in such a situation we easily forget "cases in which there is no such intermediary step11

1 for instance, in the arder "Turn right!11 of a sergeant to a private. So in the passage above Wittgenstein

204 Wittgenstein's Philosophical Development

is warning us of the danger of being impressed by one use of a word instead of looking at its multiple uses.

C) The third passage is the following:

lt is one of our tasks here to give a picture of the grammar (the use) of the word 'a certain'. (BrB, 135; my emphasis)

ln this passage it is very clear that 'grammar' and 'use' function as synonyms and there is no explanation needed here.

D) The subject of the last passage! where 'grammar' occurs is the various uses of the expression 'the picture hasn't changed':

... the expression "The picture hasn't changed" is used in a different way when we talk of a material picture on the one hand, and of a mental one on the other. Just as the statement "These ticks follow at equal intervals" has got one grammar if the ticks are the tick of a pendulum and the criterion for their regularity is the result of meas­urements wh ich we have made on our,apparatus, and another grammar if the ticks are ticks that we imagine. (BrB, 171; my emphasis)

Again in this passage the word 'use' can clearly be used in the place of 'grammar' and the word 'grammar' can be used in the place of 'use'. Note, however, that here we can count grammars, as we supposedly can approx­imately count uses of an expression (one grammar, another grammar).

Since the four passages are the only passages where 'grammar' occurs in the BrB, we can conclude that Wittgenstein's now old conception of grammar as the discipline of the 'limits of sense' from the BT is completely absent in the BrB. It plays no role in the BrB. 'Grammar' is always under­stood as the use that the expression has in linguistic practices. No philosophical conception about the 'arbitrariness' of 'grammar' or the claim that a language is only a language if it has a set of fixed rules of 'grammar' is present in the BrB (for such views, see Chapter 3). There is no talk at all about the tabulation of such rules anymore. Rule is not at the center of the investigation, as was the case in the BT:

It is the rule that stands at the centre of our examination; not the fact that 1 offer it to someone, not the fact that someone uses it, etc .. (BT, 244)

The facts ignored ln the BT are now very relevant facts. The description of uses of words in their environments characterizes now Wittgenstein's method:

The Road to the Philosophical Investigations 205

Our method is purely descriptive; the descriptions we give are not hints of explanations. (BrB, 125)

"Our method" in practice consists in identifying the orígin of the formulation of philosophical prablems in false analogies (misleading pictures, misleading símiles) and misleading trains of thoughts, and presenting those to the readers in the form of a mirrar in which the readers can see their own tendencies and mistakes; a mirrar that also indicates how to avoid the traps. The mirrar contains, however, not only philosophical missteps, but also a strategy to avoid them, which consists of r.eminders of the multlple uses of words embedded in their surraundirígs. It is purely descriptive (or really descriptive) because the old idea of 'grammar' has no place and because practices are described and not the old 'rules of grammar'. As Wittgenstein says later:

When I describe the language, 1 describe the conduct (Handlungsweise) of human beings seen ethnologically. (MS 124, 253; my emphasis)

What we have now, thus, is a break with the two old meanings of 'description' that Wittgenstein already had in mind at the time of the T (Chapters 1 and 3). Describing the language is not tabulating rules, but really providing a description of the background of our practices with words. This means that Wittgenstein does not want to provide anything like a theory of language, meaning or concepts and what he says is not intended as a prato-theory that the reader is supposed to extract from his writlngs either. His descriptions, he writes, are "not hints of explanations" (BrB, 125) - as they actually were in the BT. He is not indicating the first steps of a theory of language, meaning, etc., but examining the suspicious first step into such theories. Thus, one should not say that one is building a "theory of meaning as use," a "theory of language graunded on language-games," or a "theory of concepts as family-resemblances" as an "extension" of Wittgenstein's philosophy. Such "extensions" are not just non-Wittgensteinian; they go against his goals.

4.3.4 The German Brown Book (BrBG)

After writing the BrB Wittgenstein attempted to revise it twice (both in German) between August and November of 1936: in MS 141 (only the first 15 sections are rewritten) and in the second part of MS 115 (118-292).19 ln MS 141 Wittgenstein leaves out the parenthetical remarks from the original English BrB and enumerates only para­graphs that intraduce new language-games (in the BrB some sections

206 Wittgenstein's Philosophical Development

don't introduce new language-games). ln the second part of MS 115 (BrBG), the enumeration of paragraphs follows the idea of MS 141 and instead of using parenthetical remarks Wittgenstein distinguishes language-games from the comments on them meant to indicate where analogies and false trains of thought mislead us by means of different widths for the two kinds of paragraphs. ln this respect, then, BrBG is more systematic than the BrB. lt also has more unity, for the first and second parts of the BrB disappear: there is one sequence of languages that combines both parts.

There is no significant change concerning the use of the word 'grammar' in BrBG. The few times that the word appears in a new context are not relevantly different from the BrB (for instance, BrBG, §112). What is relevantly new in BrBG, if compared to the BrB, is the change of language, of course. Wittgenstein feels at home when he uses German and it is certainly a far better tool of expression for him than English. The change of language allows him to introduce the personal pronoun 'Du'. The use of the pronoun 'Du' brings the reader (or the German reader) doser to the author. 1. think that this might involve the reader with the topics discussed and so facilitate the mirroring of the reader in what is said. The pronoun seems to shorten the distance between the writer and the reader: they, as it were, are thinking together. The German pronoun 'Du' thus indicates a step further in the dialogical style of Wittgenstein's philosophy.

The presence of interlocutors is nota novelty of BrBG, for it is already present in BrB (as we have seen above), BB and, to a smaller extent, in the remarks from the BT. But this feature of Wittgenstein's style is more relevant in BrBG and in the various versions of the PI that follow it because it appears quite more often and is meant, 1 think, to dose the gap in the genetic method felt by Wittgenstein at the time of the BB. With such a strategy, the genetic method can have better results.

ln the BrBG, Wittgenstein gives an important step in the direction of finding another way to dose the gap between the presentation of misleading analogies and false trains of thought and the reader's under­standing that her thoughts and tendencies are expressed by his analysis of the genesis of philosophical problems. This step is the direct refer­ence to the T. While the BrB does not contain direct references to the T, the BrBG has three. The references are related to the idea of simple objects in the T (BrBG, §5), to what logicians say about the construc­tion of all propositions (BrBG, §11), and to the idea that propositions are pictures (BrBG, 288). ln BrBG §60, Wittgenstein introduces the discussion of analysis and simples. Once the origin of the conception

The Road to lhe Philosophical Investigations 207

is presented there, he claims that the "raísonnement involves different mistakes" (BrBG, §60). The mistakes that he has in mind are the ideas that a name must mean an object, that logical and chemical analysis are analogous, and that there must be "logical exactitude" in our concepts. Wittgenstein also quickly suggests that such views are connected to the context principie: they might arise if we "look only at the position of words in our sentences" (BrBG, §60).

As one can attest in MS 142 (UF), these references to the Tare impor­tant because they indicate a new tendency in Wittgenstein's philosophy after 1936, namely, to apply the genetic method to his own early philos­ophy, i.e., tomake his own old publíshed views an example of how philo­sophical problems are engendered and how they are supported by false trains of thought.

4.4 The first version of the Philosophical Investigations

MS 142 is a preliminary version of most of the first 188 sections of the PI. ln November 1936, in a letter to Moore, Wittgenstein explains why he decided to stop the revision of the BrB (BrBG) and started writing MS 142:

I don't know if I wrote to you that when I carne here I began to trans­late into and rewrite in German the stuff I had dictated to Skinner & Miss Ambrose [BrB]. When about a fortnight ago, I read through what I had done so far l found it all, or nearly all, boring & artificial. For having the English version before me had cramped my thinking. I therefore decided to start all over again and not to let my thoughts be guided by anything but themselves ... and so I'm writing now a new version and I hope I'm not wrong ín saying that it's somewhat better than the last. (CL, 283; letter from 20.11.1936)

This letter explains why Wittgenstein thought his revision of the BrB "no worth" (MS 115, 292): it was "ali, or nearly all, boring & artificial". 20

Wittgenstein's solution was to work without the English version, which was "cramping" his thinking. Thus, Wittgenstein does not seem to believe that there were huge changes in the new version of his work, for he says "I hope I'm not wrong in saying that it's somewhat better than the last." lt is unlikely that he thought that the BrB or BrBG were wrong expressions of his philosophy; his goal in the new version was to improve the previous version by making the new version less boring and less artificial.

208 Wittgemtein's Philosophícal Development

There are important differences between MS 115 (BrBG) and MS 142 that need to be taken seriously, however. Concerning the anthropolog­ical view, it must be noticed that each language-game in BrB and BrBG is explicitly connected to a primitlve tribe. After BrBG this is not the case. Why does Wittgenstein stop associating each language to a tribe? One reason is obvious: boringness and artificiality. The exhaustive use of new languages embedded in primitive tribes creates an unnecessary repetition. However, not only boringness was at issue. Wittgenstein was probably also concerned with some worríes that the way the anthro­pological view was presented could ar~se. The tribes associated to each language might give the impression that he is dealing with something distant from our form of life. This issue appears when one reflects on the supposed translation of primitive languages into our complex languages. Such a worry is explicitly stated by Wittgenstein (BrB §48; BrBG §52). Why do these tribes speak English? If they don't, how can we translate their words into English, if the background of our language is míssing?z1 Are we presupposing the background of our language when we describe those tribes? We could cal~ these questions "anthropolog­ical worries." They arise when one does not see that the relevant point is that the role of a expression "plays in the whole life of the tribe" (BrBG, 103) is to be compared with the role it plays in our whole life, i.e.1

how it is interconnected with other expressions, our goals, purposes, institutions, etc. The problem is precísely that the structure of the BrB invites the anthropological worríes when tribes are emphasized in each section. One míght focus on the exotic character of each of the tribes introduced. Such worries might create distance from, instead of an interest in, our practices.

Wittgenstein solves those worries ín ali versions of the PI by not exag­gerating the use of examples of tríbes. Severa! language-games are intro­duced at once (see PI, §8) and not ali games make explicit reference to primitive people. ln the PI Wittgenstein rarely uses examples of tribes (PI: §§6, 200, 282, 285, 419). Instead of tribes, his emphasis is on human beings (Pl: §§6, 53, 64, 157, 189, 241, 243, 257, 281, 325, 331, 344, 360, 415, 491, 528, 554). The human beings are ourselves (PI, §241) - say, human beings who share chemistry, mathematics, and logic - or human beings that have different forms of life. The degree of similarity of those forms of life with ours varies: some are very simple (for instance, PI: §§2, 528), some more sophisticated (Pl, §243).

The most important change after BrBG, as already mentioned, concerns the application of the genetic method. ln a conversation with Moore reported by Rhess (BrBG, introduction), Wittgenstein said that

\

s f

1 1

1

The Road to the Philosophical Investigations 209

in MS 142 he employed the "right method." Moore said that he did not know what Wittgenstein was talking about. Wittgenstein, in fact, refers to his method in the (irst preface for the PI (from 1938):

I begin this publication with the fragment of my last attempt to arder my philosophical thoughts. This fragment has perhaps the advantage of, comparatively, easily convey a concept of my method. (TS 225; PUKGE, 208)

Wittgenstein does not talk about the right method, but a.bout conveying a concept of his method. This is dane in two different ways, both absent in the BrB and in the BrBG. First, by explicitly stating his method:

One of our most important tasks is to express all false trains of thought so true to character that the reader says, "Yes that's exactly the way I mean it". To trace the physiognomy of every errar.

We can never bring someone away from a mistake unless he acknowledges this expression as the correct expression of his feeling.

Only if he acknowledges it as such is it the right expression. (Psychoanalysis).

What the other acknowledges is the analogy that l'm presenting to him as the source of his thought. (MS 142 (UF) §§121; TS 220 (FF), §106; TS 239(BFF), §139)

This paragraph is connected with the following:

We get rid of the spell of the ideal by recognizlng it as a picture and by giving its origin. How have you come to this ideal; out of which stuff have you formed it? Which concrete image (Vorstellung) was its original archetype (Urbild)? We have to ask ourselves this, otherwise we cannot get rid of its misleading aspect. (MS 142, §122;TS 220 (FF), §107; TS 239 (BFF), §140)

Other statements of the genetic method are included in the first versions of the PI (for instance (TS 220, FF, §106 and BFF §140), but those passages above suffice to make the point clear here.

Second, Wittgenstein conveys a concept of his method by applying it systematically to the T. Note that the critique of the T in PI is different from the critique ln the BT. In the BT, the critique is grounded on the

210 Wittgenstein's Philosophical Development

calculus conception of language (as was shown in Chapter 2). ln the PI, it is grounded on the genetic method. Why did Wittgenstein decide to apply his method to his own work in a systematic manner after BrBG? There are two reasons for this. To apply the genetic method to his work would give the reader a good example of the method; second, the examination of old views could make clear how his then current views were different from the views expressed in the T. 22 If Wittgenstein succeeds, the reader can see in front of him a true example of how philosophical questions and confusions arise. Wittgenstein had already applied the genetic method to his own views, but the appl~cation of the method was mostly related to his tendencies and temptations from the period 1929-30, i.e., to unpublished work. This could not be the best example (a mirror) to a reader, for the reader would not have the reference of the critique. lt is interesting to note that his own views in the T are reflected in works of other philosophers (Russell, Frege, and Plato, for instance). Many refer­ences to other philosophers, 1 take it, are intended only to make clear that bis case is, in many ways, exemplar in the history of philosophy.

One can see in Wittgenstein's notebooks from the end of 1936 (MS 152) and the beginning of 1937 (the second part of MS 183, the second part of MS 157a and MS 157b) that his new goal at the time that he wrote MS 142 was to apply the genetic method to bis philosophy in the T (MS 157a, Slv) and to show how bis views had changed sínce then (see MS 157a: 48v and 54r). ln 02.09.1937, in MS 183, Wittgenstein gives the first characterization of the sublime conception of logic of the T:

The "sublime conception" forces me to go away from the concrete case, for what 1 say does not fit it. 1 make my way to an ethereal region, speak about proper (eigentliche) signs, about rules that must exist (although 1 cannot say where and how) - and 1 get 'on to slippery ice'.

The "proper signs" are the signs of the notation of the T. The rules that "must exist" are presumably rules for simples, which are supposedly implicit in language (see Chapter 3, section 2). Wittgenstein continues this remark in MS 157a, 47r, in the sarne day (02.09.1937):

For the interest that drove [me] to it ["sublime conception"] was not the demand for new facts - for knowledge of nature - but the demand for learning to understand the essence, the structure as we say - of facts. It was essential that we, in an important sense, did not want to find out anything new, but wanted only to understand the well known.

(

The Road to the Philosophical Investigations 211

The well known was the a priori rules of logic, l.e. the a priori rules of language and thinking:

Language is language through the rules by which it is operated Logic is not allowed to have anything empirical in it. -

Or also:

"thinking must be something unique" (Why?) (Here is the root of the problem) (Hier líegt der Hund begraben). (Ms 157a, 49r)

lf logié should report about the essence of language (of thinking) so 1 must be in possession of a transparency (a crystal purity) that we don't find in the sciences.

Thinking/the concept of thinking is involved ln a nimbus. (MS 157a, 49v)

Soon after this remark (and similar ones that follow it) Wittgenstein points to what heis looking for in the manuscripts from the beginning of 1937:

We are under the illusion (1 want to report under which illusion we are!) that the sublimity of our question must consist in asking about the essence of the unique - of the world, of experience, of thinking. (MS 157a, p. Slv)

As is clearly stated by Wittgenstein, he wants "to report under which illusion" he was at the time of the T. Besides making clear what the illu­sion in his description is, he is also looking for what engendered it. This can be seen in severa! passages of the manuscripts (notebooks 157a and 157b) from the time of Wittgenstein work on MS 142:

We have to acknowledge (eingestehen) in which representations the ideal forms originate. (MS 157a, 54v)

Or:

Where do you have this ideal [ideal name] from? What is its arche­type? For this is what gives it life. (MS 157a, 56r)

How have you arrived to this ideal? (MS 157a, 57 r)

212 Wittgenstein's Philosophical Development

Besides the a priori character of logic and the ideal simple names, the Tractarian idea of analysis is under scrutiny. And as in the other cases, Wittgensteín is interested in the physiognomy of the mistake:

Role of the false conception of analysís. We labor under a mistake. (That what 1 want to explain and cannot is: how are we laboring under a mistake.) (MS 1S7a, S9v)

The most interesting aspect of the remarks of the notebooks from the beginning of 1937 is Wittgenstein's interest in pointing out the path that brought him to his assumptions in the T. Here another passage that clearly states that he is investigating how he arrived to,'his ideal, what underlies his old beliefs:

How have you arrived at this ideal [the crystal purity of logic]? Out of which material have you formed it? Which concrete representation was its archetype/was its model?

You have to ask yourself this, othe.rwise you will not get rid of the fascination of the ideal. (MS 157b, 13r)

The significance of looking at what engenders the philosophical pictures, trains of thought and doctrines lies in the struggle to free oneself of the "fascination of the ideal." After the passage quoted above, Wittgenstein makes reference to some passages of the BT. It is very significant that the pages of the BT that he indicates are related to the genetic method. He indicates precisely pages between 409 and 431 of the BT.23 These pages are part of the sections on philosophy. BT 409, the first page that he refers to, is the page where he states that philosophers are misled by analogies and in BT 410 he presents a description of the genetic method.24 After mentioning the passages from the BT, Wittgenstein makes clear, again, that he is using his genetic method:

The solution of the problems consist in eliminating the worrying aspect prompted by false analogies in our grammar. (MS 157b, 14r)

The questions asked in MSS 183, 157a and 157b about the underlying motivations of the T and the investigation present in them informs what is truly new in the first (MS 142) and subsequent versions of the Pl. It is exactly the particular application of the genetic method to the T that is not present in the two versions of the BrB (BrB and BrBG).

Here we have reached all-important steps that lead to the construc­tion of the PI. Our last task, in the next chapter, will be identifying how

The Road to the Philosophical Investigations 213

the traits of Wittgenstein's philosophy that we have been discussing operate in the Pl. Before doing this, we will look at Waismann's second project with Wittgenstein.

4.5 Wittgenstein's second project with Waismann

At the end of Chapter 2 l explained what led Wittgenstein to stop his collaboration in the first project with Waismann, i.e., the project of a new version of the T. Now l turn to the collaboration with Waismann on the second project. The second project started probably at the beginning of 1932 and ,ended completely, very likely, only in 1936, when Schlick died and the strongest bond between Wittgenstein and Waismann was broken. To look at this project is lmportant, here, for two reasons. First, because one should not underestimate Wittgenstein's serious work with Waismann and his proximity to the Vienna Circle. 25 Second, because we will see how Wittgenstein's philosophical development descríbed in this book connects with his collaboratlon with Waismann.

According to Baker, after giving up the project of the reformulation of the T to be published by Schlick in the Circle's series Schriften zur wissentschaftlichen Weltauf(assung, Wittgenstein began working with Waismann on a new project: an alternative work was to take the place of the old project. Wittgenstein and Waismann had many meetings after 1931. 26 Their collaboration was full of personal difficulties through the years. The major reason for difficulties was Wittgenstein's changes of mind. As we have seen, there were many. ln a letter from Waismann to Schlick, from 08.09.1934, he complains about Wittgenstein "demol­ishing" what had been written:

He [Wittgenstein] has the great gift of always seeing thlngs as if for the first time. But it shows, l think, how dífficult collaboratíve work with him is, since he is always following up the inspiratíon of the moment and demolishing what he has previously sketched out ...

Ali one sees is that the structure is being demolished bit by bit and that everything is gradually taking on an entirely different appear­ance, so that one almost gets the feeling that it doesn't matter at ali how the thoughts are put together since in the end nothing is left as it was.27

This letter shows that Wittgenstein's ideas concerning how to present his philosophy in 1934 were changing. ln fact, the BB and the BrB have an "entirely different appearance" from the BT and the PG. However, the BB and the BrB also look quite different from what

214 Wittgenstein's Philosophical Development

survived from the collaboration between Wittgenstein and Waismann. Waismann himself could not keep up with Wittgenstein's pace after 1933. This can be seen in what remains from Waismann's collabora­tion with Wittgenstein as presented in his Logik, Sprache, Phílosophie, translated as The Principies of Linguistic Philosophy (PLP). Waismann's published book must have been in many ways transformed after the end of the project and it is not completely clear what in the book are Wittgenstein's and what Waismann's ideas. However, various character­istics of Wittgenstein's philosophical development are clearly present in it.28 ln fact, Waismann's book maybe seen as a kind of fully devel­oped version of the BT with the incorporation of aspects of the anthro­pological view (a kind of PG) and, at the sarne time, /as a summary of Wittgenstein's developing ideas from 1930-35. One finds in PLP discussions concerning phenomenological language and visual space (PLP: 44-50), verification (PLP: 73-5, 323-37), color incompatilibili­ties and the grammatical understanding of inferences (PLP: 59, 364-7), classification of words (PLP, 189-90), and 'grammar' (PLP: 13-4, 35-40, 64). A whole chapter is dedicated to tlj.e presentation and critique of Russell's causal theory of meaning (chapter VI). Waismann says, for instance, that he is not interested in a causal account of language, but in a conception of language as a calculus (PLP, 122-4). This, as we have seen, was Wittgenstein's position in the BT.

Both Wittgenstein's genetic method and his views on grammar are described in PLP. About the method, for instance, Waismann says:

An important part of our task is to scent the analogies which lead to philosophical problems. Our method is in some respects similar to that of psycho-analysis. Using this terminology, we might say that an unconsciously active analogy becomes harmless as soon as it is brought into consciousness ... Analogies used intentionally can do no harm ... This procedure becomes dangerous only if the metaphor is embodied in language and fossilized in such a way that we do not realize that it is a metaphor. (PLP, 179)

ln several passages of PLP Waismann points out the significance of the role of false analogies in philosophical mistakes (for instance, PLP: 42, 45, 60, 81, 82, 102, 109, 168 and 176). He also tries to make sure that heis not expressing any philosophical thesis or assertion: "I really state the point of view from which 1 look upon the matter, but 1 do not make any assertion; 1 merely suggest one aspect which is neither true nor false" (PLP, 133; also 72).

The Road to the Philosophical lnvestigations 215

About grammar Waismann says:

Grammar is, as it were, the installation and adjustment of a system of signs, in preparation for their use. If we take the word 'grammar' in this wide sense, it includes not only the rules it usually includes -rules of accidence and syntax - but also, for example, the definition of simultaneity, the rules governing mathematical calculation, the rules of logical inference, and ostensive definition of the sort 'this color is yellow' (pointing at a yellow patch). For this pointing too is part of the preparation of language. In short, grammar includes all the enormous .number of conventions which, though nowhere expressly formu­lated, are' presupposed in the understanding of everyday language. (PLP, 14; my emphasis)

This is certainly a clear statement of the broad understanding of 'grammar' defended by Wittgenstein in the BT, since it deals with the conventions supposedly presupposed and not expressly formulated of everyday language which include "the rules governing mathematics, the rules of logical inference and ostensive definitions" (see Chapter 3). It also presents the dubious role played by grammatical rules in the "preparation of language" and in the explicitation of the bounds of sense. The task of philosophy as 'grammar,' in Waismann's account, is to reach a classification of words that reflects the "fine details of the architecture of concepts" (PLP, 99). Such a classification would make clear the rules of the sentences of language that express the "conditions of meaning" (PLP, 145). If a sequence of words is "against the rules" (PLP, 64), it is meaningless. These rules are, as in the BT, arbitrary. ln Waismann's wording, they are "arbitrary conventions" and "grammar is autonomous and not dictated by reality" (PLP, 40; Waismann's emphasis).

Waismann's book, then, presents central ideas underlying the BT and PG. However, it also presents some important worries concerning the use of the word 1grammar1 and some of the changes in Wittgenstein's reformulations of the BT. Waismann is aware that the use of 'grammar' present in bis book goes against the ordinary use of the word. He says:

We have to be careful as to the form in which we sum up our point of view. Saying 'The sentence "red and green exclude one another" is really a rule of grammar' looks suspiciously like a metaphysical statement. (Remember, for example, 'A physical object is really a class of sense-data1

). For, according to the ordinary use of language, this sentence is not a rule of grammar, and nearly everybody whose

216 Wittgenstein's Philosophical Development

opinion could be asked would be strongly disinclined to describe it as a rule. Thus what we are doing is apparently violating the ordinary use of language. (PLP, 65-6)

Waismann defends that it only looks as if the ordinary use was violated by making a distinction between a conservative and a reformist view of the use of language. The conservative view says that "red and green exclude one another" is not a rule of grammar, while the reformist says it is. For the reformist, we can "change or reform language arbitrarily in such a way as to obtain the simplest se.t of grammatícal rules" (PLP, 66). The reformist tendency, Waismann claims, is to be found in science. This seems to flatly go against Wittgenstein's view (at léast after 1934) that he does not want to reform language (PI, §132). Indeed, 'grammar' for Waismann turns out to be the totality of the 'rules' of science plus other 'necessary' rules like "red and green exclude each other." Thus, it's a kind of overall science. In Waismann's view, it is a mistake not to be a reformist concerning 'grammar,' for "we get a more simplified picture of grammar" with it (PLP, 67). He gives t.he example of projective geom­etry to sustain his claim:

It would, for example, be quite impossible to obtain the simple and clear system of laws which we find in projective geometry if we persisted in distinguishing between points in the ordinary sense and points in the new sense. Similarly, it simplifies the picture of the system of grammar to agree to the convention whích 1 have suggested. (PLP, 6 7)

Waismann is, then, certainly correct to say that his, or Wittgenstein's at the time of the BT, view of 1grammar1 is not metaphysical in the sense of being a description of how things, supposedly, really are. For, after all, he is making a terminological recommendation, a "recommenda­tion to stretch the use of the word 'rule"' (PLP, 66). As a recommenda­tion or reformation, the extended idea of 'grammar' is not, arguably, a violation of the ordinary use. This extended view assumes, however, as Wittgenstein's own view in the BT does, various dubious ideas concerning the nature of 'grammar' (see Chapter 3, Section 3.3).

Apart from Waismann's philosophical opinions derived from Wittgenstein's in the BT, PLP presents also some of the anthropological aspects that Wittgenstein introduces in the manuscripts collected in PG (especially MS 140). Waismann says, for instance:

" r

' s

e \

r e a

a i

t

The Road to the Philosophical Investlgations 217

Imagine the case of a primitive trlbe which uses only gestures, and for whom such a thing as a word does not yet exist. Could we not stíll speak of the meaning of the gestures? For it would be natural to say that the meaning of the gesture is its part in the life of the tribe. This is a sense rather like that in which we speak of the 'the meaning of an event in our lives.' ln any case it is only we, as spectators of the tribe who can speak of such a meaning. It is mostly taken for granted that in every language one can ask a question about meaning, and answer this question. Wrongly, however, for why should there not be a language in which expressions like 'explain', 'mean~ 'denote' are unknown? ln such a Ianguage the question about meaning cannot even be asked. (PLP, 126; Waísmartn's bold, my ítalícs)

Note that "in the life of the tribe" expressions like 'explain' and 'mean' are unknown and the "the question of meaning cannot even be asked."29 The anthropological víew (taking seriously the role of words in our life and primitive, or simple, languages with their envi­ronment) introduced in corrections of the BT and in the PG is present in Waismann's book because he had access to the material used in the PG. He had, for instance, access to MS 140, the second revision of parts of the BT (parts 1-III of the PG) and even access to later material (Baker 2003, xxvi-vii). ln the passage above, as 1 pointed out, it is clear that in primitive languages "the question about meaning cannot be asked." Waismann, however, does not seem to notice that 'grammar' has the sarne fate in such languages.

Waismann's book, then, presents the most important characteristics of the 'grammar' of PG including at least aspects of the anthropological view and the genetic method. What is, however, not clearly present in PLP is the priority of the genetic method over the old conception of 'grammar' - as is the case in the BB and in the BrB. One has the impres­sion that Waismann, probably for his own reasons, did not want to leave the philosophy of the BT behind. This point is clear, if one thinks that there are two other elements of Wittgenstein's later philosophy missing in Waismann's book: the use of language-games of increasing complexity and the application of the genetic method to the T. The application of the method to the T is absent, I think, simply because in 1937 the collaboration with Waismann had ended. Moreover, such a task Wittgenstein had to accomplish alone, especially because of its confessional tone. The increasing complexity of language-games, however, was part of what Wittgenstein wanted Waismann to change

218 Wittgenstein's Philosophical Development

in the book. ln a letter to Karl Menger, probably from 1936, in which Waismann explains his parting ways with Wittgenstein, his preference for and his insistence on the old book model is clear:

He [Wittgenstein] was, in fact, in agreement with content and form of presentation (he call it a "valuable work"), but invoked Schlick and me not to publish the work in this form, since in his opinion the book would gain a lot, if it took a completely dífferent route. I took as the starting point in my presentation the elucidation of philo­sophical problems by carrying out, a linguistic investigation of the expressions in terms of which the problem is formulated. The solu­tion of the philosophical problem was the main thing; the grammat­ical investigation the means to achieve it. But Wittgenstein suggested [me] to write a book that contained nothing from philosophy, [a book] that constructs the grammar systematically; 1 should, so he thought, think of how I could build a series of examples, which lead from the simplest to the most complicated concepts in philosophy. lf such a repre­sentation succeeds, then we don't have to make any effort concerning the solution of philosophical problems: they fali in one's lap like a ripe fruit. (quoted in Manninen 2011, 263-4; my translation)

lt seems that a "valuable work" is not exactly the best expression of agreement with content and form. Note that Wittgenstein suggested a "completely different route" in arder to make the book "contain nothing from philosophy" (as his method required). This different route (the route present in the BrB) was first indicated by Wittgenstein in a letter to Schlick from 07.31.1935: "As you know, there is a MS [BrB] which 1 dictated in the first two terms of this year and which shows the way how 1 want the whole stuff to be dealt with" (LWGB). ln the new project, the goal of 11containing nothing from philosophy" would be achieved by a 11series of examples," as the anthropological view in the BrB: "from the simplest to the most complicated concepts in philos­ophy." ln the end, philosophical problems disappear in the light of the method employed. This, of course, according to Wittgenstein's, and not Waismann's, representation.

This important idea, developed gradually by Wittgenstein in BrB, BrBG, and PI, is not followed by Waismann, as he himself indicates in the sarne letter:

1 said at the time that I liked this thought, but that 1 considered it very difficult to carry it through; I said that 1 did not feel to be a

The Road to the Philosophical Investigations 219

match to such a task. I also thought that my original presentation had certaín advantages that would be lost in the revision. Shortly, I vehemently resisted this proposal. [ ... ] At last l agreed under the condition that I could stop working and that l would not bear responsibility, if the book either were published late or not published at all. (sarne letter as above; my emphasis; my translation)

Waismann does not explain which the advantages of his presentation were, but one can surmise that many of the discussions of Wittgenstein concerning topics doser to his own interests were not dealt with in Wittgenstein's presentation. This, in Waismann's view, should be part of the "systématic part" of the book:

I did not want to give up my old ideas completely, but thought the laying out of the book in this way: ln detailed introduction - approx­imately one fífth of the book - a number of examples should show philosophical problems come into existence through lack of darity in thinking and how they dissolve as soon as we dearly recall the sense of the used expressions. After the significance of the gram­matical investigation had become dear in this way, the systematic part should begin. (sarne letter as above; my translation)

PLP actually retains Waismann's own ideas expressed in this letter concerning the book. He begins the book with a discussion of the nature of philosophy and then moves to examples. Those examples, however, are merely part of his introduction (Part I of the book), which is in fact one fifth of the book and is supposed to show "The Transition from the Classical to the Linguistic View of Philosophy." The linguistic view of philosophy, the "systematic part" of the book, is still compat­ible with Waismann's own views and could be seen as an interesting development of logical positivism. This compatibility, however, seems to disappear when the genetic method and the anthropological view frame Wittgenstein's philosophy. 30

One cannot, of course, blame Waismann for not finishing the project in the way Wittgenstein wished. PLP is in itself a very interesting testa­ment of all the themes that build Wittgenstein's philosophical develop­ment between 1929 and 1934. One can, however, and one should, be aware of Wittgenstein's insistence in changing the whole representa­tion of his philosophy, even though the book with Waismann was very dose to being finished - i.e., dose to being finished, if one did not want to express the centrality of the genetic method, Jf one did not want

220 Wittgenstein's Philosophical Development

to teach a new method of philosophy framed in an anthropological view. Those, however, were Wittgenstein's major goals from the BrB onwards.

All those facts indicate that the best way to look at Waismann's work is already given full expression by Wittgenstein himself in a letter from 05.1936. ln thís letter, Wittgenstein complaíns about the expressíon "valuable stimulations" used by Waismann to express his debt to him in the paper Ueber den Begriff der Identitaet. "Valuable stimulations," Wittgenstein thinks, distorts what really took place; "a theme from Wittgenstein" would be better:

Jf a composer A has written varíatíons about a theme from composer B, he does not write 'I was stimulated to this piece from B,' but 'Theme from B.' [ ... ] with the statement that the theme is from B nothing is said about the value of the variations; this [value] can be as great as the value of the theme, even greater. (letter to Waismann from 05.19.1936 (LWGB))

'Theme' and 'variations' seems really a better way to make clear the relation between Waismann's work and Wittgenstein's. ln his answer to Wittgenstein from 05.27.1936, after a conversation with Schlick, Waismann explains that he was constantly in doubt about how to mention Wittgenstein's work and makes clear that he obviously recog­nizes that he owns Wittgenstein "the decisive stimulus, the direction of the whole investigation" in the mentioned paper. It seems to me that this is true of most of what Waismann wrote thereafter, including PLP, and that Wittgenstein's metaphor of theme and variations correctly expresses the situation. They are, however, mostly variations of Wittgenstein's philosophy at the time of the BT and PG.

Approximately one month after the letter just quoted Schlick died. This fact, the tension created through the years of dose work, and Wittgenstein's new ideas and strategies after 19331 1 conjecture, are the reasons for the end of his cooperation with Waismann.

t

a a

1

a

e e f V

s t

e o

5 The Philosophical Investigations

ln this chapter, 1 intend to apply results of the four previous chapters in order to elucidate main traits of the Pl. l do not intend to give an account of the whole book, nor get into the detaíls of the sections discussed. ln parts 1-3, 1 focus on Pl, §§1-136, where the genetic method, embedded in an anthropological view, is applied to the T. These sections give us the "right light" in which the book is to be read, i.e., "by contrast and against the background of [Wittgenstein's] older way of thinking" (PI, preface). ln part 4, 1 explain the role of an old remark (PI, §372) in the new context of the PI. This 1 do in order to avoid the misleading view according to which the 'old grammar' of the BT is still in place in the PI, and to elucidate the role of "grammatical remarks. 11

5.1 Preliminaries: genetk method, anthropological view, and surveyable representation

The sections of the PI (§§1-39) that precede the most detailed applica­tion of the genetic method to the T after §40 introduce the method and the anthropological view. A characteristic of the method is intro­duced right at the beginning. One must bear in mind that Wittgenstein operates at a pre-theoretical level, where philosophical questions are first asked and phílosophical problems arise. That is, 1 think, the reason why Wittgenstein begins his book with a description of how Augustine thought he had learnt language, and not with any established philo­sophical thesis.1 He begins the book with an ordinary description that might induce someone to formulate a philosophical theory (one is tempted to generalize to all kinds of words what seems to work for concrete nouns). This descríption expresses a rough idea (a picture) of the workings of language. Such a rough idea also operated to some

221

222 Wittgenstein's Philosophical Development

extent in the T, of course, for in the T the meanings of simple names are simple objects.

The pre-theoretical levei strategy is also clear in the introduction of the fírst language-game of the PI in §1. One is inclined to think that the way the shopkeeper and the buyer operate is quite amazing: one has the impression that some things that guide their behavior (the sign for apple and the color chip) are in the wrong place, i.e., open to view, in front of them. But what is actually wrong with the way they operate? One is tempted to object: 11 But how does he [the shopkeeper] know where and how heis to look up,the word 'red' and what heis to do with the word 'five'?" (PI, §1). Only in rare cases do we operate like him. However, what the philosopher is really missing iri the picture is not the normal operations of the shopkeeper in a normal circumstance in natural surroundings. He misses a description in which something analogous to the sign for apples and the color slips are taken to be in the mind; i.e., he misses what he wants to input into the description.2

Obviously, Wittgenstein's description of the situation makes no refer­ence to the shopkeeper's mind. To what,extent, then, is such a descrip­tion incomplete? For a philosopher, presumably, the most important thing is missing: an explanation of behavior grounded on the opera­tions of the mind, the store of meanings. Here we thus have a descrip­tion of a situation that might easily bring us to require a supposedly explanatory mind-model. Wittgenstein's point is: how will we do it? How will we make our first step into what we would happily call an explanation of the operations of the mind?

The language of §2 is also an important part of Wittgenstein's general strategy in the PI because in that there is no grammar or rules of language in the sense of the T and in the sense of the BT. There are no sentences in this language. Thus, the primitive language of the builders undermines the very goal of the T (the understanding of the essence of propositions and the expression of the limits of language, therefore of sense) and the goal of PR and BT (the presentation of the rules of grammar which give the limits of sense). This language indicates that in both cases Wittgenstein was operatíng under an idealized view of language. Before looking for the essence of language or its structure, he had already excluded many languages. Can we talk about grammar at ali in that language-game? We might try to give it a grammar analogous to the grammar of the BT, but it would be at best a degenerate grammar. To what extent can we talk about propositions and calculus in the first language-game? Those are questions that Wittgenstein wants us to ask, I think.3 Note that even if someone takes the words (slab, pillar, etc.)

The Philosophical Investigations 223

as sentences (say, as commands), they are not complex as our sentences are; they are not grounded on compositionality (as the T, Frege, Russell, for instance, assumed propositions should be). They also don't describe reality; therefore, they don't express the multiplicity of facts.

However, Wittgenstein's goal is not to present new necessary condi­tions for a language to be a language (say, necessary conditions that once in place can be used to include primitive languages in our defi­nition). He is only giving us a reason to think about which condi­tions we are inclined to impose on 'language' when we think about the logic or the 'grammar' of language. A philosopher is probably inclined to say things like: "these builders are automata; they don't seem to think, therefore it is not a language;" "they don't seem to have a mental life, therefore this Is not a language;" "language must connect with mental states, otherwise we have only empty signs;" "there is no sign for negation, therefore no possible description; there­fore no language in §2;" "there are no sentences, no possible complex combination of words, therefore no logical operations, therefore no. language in the language-game of §2;" "these builders are too primi­tive; they cannot have a language; therefore, there is no language in §2;" "there are no grammatical or syntactical rules in §2, therefore no language" etc. Such reactions, I presume, are expected, but they need to be looked at closely. They give rise, say, to 'logical' or 'grammatical' requirements, psychological mind-models, and models of rationality. These, of course, are the very beginnings of philosophical theorizing; they are the source of what philosophers call 'philosophical expla­nations.' Wittgenstein's questions are, however: why do we want to impose necessary or "transcendental conditions" on what counts as a Language? How do we expect to construct our philosophical explana­tions? How do we get to our first assumptions that presumably ground our theories? Perhaps our worries arise because we are also inclined not to take into account the surroundings of a language and the form of life in which it is embedded.

The introductory remarks also begin the critique of the most funda­mental view of the T, namely the idea that there is a general form of propositions that determines what counts as a part of language (given that language is the totality of propositions). The language of §2 is a language that has no propositions. What is then the use of the bounda­ries of language presented in the T by means of the essence of prop­ositions? This first criticism of the view of language of the T comes full-circle in §136, where the notion of the general form of propositions is trivialized and reduced to a misleading ideal.4

224 Wittgenstein's Philosophical Development

The sections of the PI prior to the direct discussion of Tractarian topics also have the function of making the reader attentive to the "multiplicity of tools in language" (PI, §23). Attention to different uses of words and sentences may be the best way to avoid confusions right from the beginning of our philosophical investigations. This is also the reason why Wittgenstein presents an analogy between use and meaning (PI, §43). The 1picture1 (rough idea) of meaning as use is supposed to counteract the picture that the meaning of a word must be an object. The role of this picture, this analogy, is thus that of a suggestion to avoid confusion; it is not intended to be a proto-theory of meaning. It is an object of comparison meant to "break. the spell" (TS 239, BFF, §140). The analogies, metaphors, símiles or pictures (rough ideas) that Wittgenstein uses to break the spell of misleading analogies are nothing more than analogies, and cannot be used as pictures to take the place of the old ones or as general explanations out of which we would derive other explanations: "A símile belongs to our editice; but we cannot derive any conclusions out of it; it does not lead us beyond itself, but must remain there as símile. , We cannot derive conclusions out of this" (TS 239, BFF, §135; also FF §102 and BT, 418).5 Wittgenstein is not saying, then, that ''the meaning of a word is its use, therefore it is not the object that it refers to," but rather saying something along these lines: "if you have difficulties in getting rid of the idea that the meaning of a word is what it refers to, and see the obvious limitations of such a view in the case of words like 'but,' 'therefore,' '5,' 'nothing,' etc., then look at the meaning as the use of words in their surroundings."

The broad idea of use that Wittgenstein employs is meant to make us attentive to the fact that it is our forgetting how words are used in certain surroundings that makes us look for philosophical theories of meaning and mind. He wants to remind us of the uses of words in order to solve philosophical puzzlement: "The work of the philoso­pher consists in marshaling recollections for a particular purpose" (PI, §127 - concerning recollections see also PI: §§30, 8911611 194, 217, 257, 269, 322, 591). The step towards a "mind-model" is taken, for instance, because we are puzzled by very limited uses of words and systemati­cally forget their surroundings. When we forget the surroundings of the use of words, we are prone to postulate a spirit in our explanations (PI, §36). So, for instance, he talks about the variety of language-games as "part of an activity, or a form of life" (PI, §23) that are meaningful in their surroundings ("Giving orders, and acting on them," "reporting an event,11 "guessing riddles," etc.). The variety of linguistic and mental

The Phllosophical Investigations 225

activities, he makes clear, "are as much part of our natural history as walking, eating, drinking, playing" (Pl, §25).

We have already seen how the genetic method gradually comes to play a more prominent part in Wittgenstein's thought after 1931 and how it has a central position in the BB, in the BrB and in BrBG. ln fact, it gains significance without being substantially changed. There is, however, an important change in the expression of the method. lf Wittgenstein ever assumed that all philosophical problems have a false analogy at their bottom, then this assumption is abandoned by the time of the Pl. Note the italicized expression below:

Our inquiry is therefore a grammatical one. And this inquiry sheds light on our problem by clearing misunderstandings away. Misunderstandings concerning the use of words, brought about, among other things, by certain analogies between the forms of expres­sion in different regions of our language. (Pl §90, my emphasis)

The part that 1 italicized is a small but significant addition to the previous versions of the Pl. ln the last version of the PI (TS 227), Wittgenstein clearly avoids a dogmatic position according to which all philosophical problems are prompted by false analogies. With the new formulation, he is not committed to the view that there is only one source of philo­sophical confusion; philosophical problems may have various sources and there may be thus various ways (methods) of dealing with them.

ln MS 142 Wittgenstein wants to clear away misunderstandings caused by false analogies. The false analogies are indicated and the false trains of thoughts that lead to philosophical questions are made explicit. This is still, 1 take it, the sarne method, broadly conceived, that Wittgenstein wants to show by examples in the final version of Pl:

It is not our aim to refine or complete the system of rules for the use of our words in unheard-of ways. For the clarity that we are aiming at is indeed complete clarity. But this simply means that philosoph­ical problems should completely disappear. The real discovery is the one that enables me to break off philosophy when 1 want to. - The one that gives philosophy peace, so that is no longer tormented by questions which bring itself in question. - lnstead, a method (eine Methode) is now demonstrated by examples, and the series of exam­ples can be broken off. - Problems are solved (difficulties eliminated), nota single problem. (PI, §133)

226 Wittgensteín's Philosophical Development

Wittgenstein believes that he is teaching a method broadly conceived (more about this below), and not severa! methods to deal with phil­osophical problems. lf Wittgenstein had various methods in mind, he would have said that methods are shown by examples, and not a method (eine Methode). He, however, seems to contradict this right after the passage just quoted:

There is nota philosophical method (eine Methode der Philosophie), but there are methods, different therapies, as it were. (PI §133; slightly modífied translation)

This remark sounds out of place. ln fact, it might be:' Its ancestry is a slip of paper on which one can also find the observation "p. 91 foot­note" (PUKGE, 817; TS 227a). Page 91ofTS227 is, indeed, thepage where §133 is. The remark above is Wittgenstein's, for he originally wrote it in MS 116, then in TS 228 (a TS used in the elaboration of TS 227), but it was not he who wrote the indication "p. 91 footnote." This indication is from somebody else's handwriting CPUKGE, 817). ln any case, the remark should not be in the text properí but at most in a footnote. 6

There is no contradiction in §133, however, even if the remark on methods appears there. Parallel to his non-dogmatic view concerning the origin of philosophical problems seen above (maybe not only analo­gies prompt them) is Wittgenstein's non-dogmatic view concerning the dissolution of philosophical problems. Both remarks, the one concerning analogies in §90 and the one concerning method, are absent in early versions of the PI. The point is that philosophical problems need not originate in false analogies. There may be, then, different methods. Wittgenstein is thus not excluding other methods while presenting a method (or therapy, as he also says).

One must also note that there are severa! ways (methods) to point out how philosophical conceptions originate. ln this sense, there are many methods related to Wittgenstein's very broad methodological guidelines of the genetic method. Another method subordínated to the general guidelines would be, for instance, describing language-games that fit certaln pictures of how language works ln order to show that they are, in fact, very limited. This could be called "the method of §2" (see PI §48). One could also think that the post-BT idea of a surveyable repre­sentation is a method in itself or a kind of device that works with or in service of the genetic method (more about this in what follows). Even the genetic method, however, might be carried out in various ways, for instance, by delineating false trains of thoughts and by pointing

The Philosophical Investigatlons 227

out false analogies. Whether we think of Wittgenstein as having one method or several methods is thus not really relevant, for his method can itself be split into various methods.

There is another change in the last version of the Pl concerning the description of the genetic method. The remarks that most explicitly describe lt are present in the first three versions of the Pl, but not in the last two. The dearest description of his method is the following:

One of our most important tasks is to express all false trains of thought so true to character that the reader says, "Yes that's exactly the way l mean it". To trace the physiognomy of every error.

We can only bring someone away from a mistake if he acknowl­edges this expression as the correct expression of his feeling.

Only if he acknowledges it as such is it the right expression. (Psychoanalysis)

What the other acknowledges is the analogy that l'm presenting to him as the source of his thought. (MS 142 (UF) §§121; TS 220 (FF), §106; TS 239(BFF), §139; modified translation of BT, 410)

This paragraph is connected with the following in TS 220 (FF), §107 (also present in TS 239, the BFF):

We get rid of the spell of the ideal by recognizing it as a picture and by giving its origin. How have you come to this ideal; out of which stuff have you formed it? Which concrete image (Vorstellung) was its original archetype (Urbild)? We have to ask ourselves this, otherwise we cannot get rid of its misleading aspect.

Neither of these paragraphs was induded in the last two versions of PI (also: FF, §106 and BFF §140). However, those remarks are among the remarks that were introduced in the FF in order, as Wittgenstein says, "to easily convey a concept of my method" (TS 225, the first verslon of the preface of the PI, from 1938). So it is interesting that the absence of the remarks above parallels the fact that Wittgenstein does not mentlon his method in the last version of the preface (concerning its first version, see Chapter 4, Section 4.4). Why did Wittgenstein decide to take these paragraphs out and not to mention 'method' in the preface? I think that Wittgenstein had good reasons for so doing. I begin with an externai reason, which concerns the way people took his remarks, and then move to internai reasons, which concern the way he presents his philosophy.

228 Wittgenstein's Philosophica/ Development

ln 1941, Sraffa read, made comments on, and discussed the BB with Wittgenstein - his criticisms were, indeed, applied for "many years" (PI, preface). He criticized, among other things, Wittgenstein's supposedly psychoanalytic prose in the BB:

When you [Wittgenstein] describe the cause of these puzzles and prescribe the remedy you act as a scientist (like Freud). Have you found out whether these puzzles have in fact arisen out of this atti­tude to language, have you made sure that they did not exist before anyone took that attitude etc? And,also, is it a fact that the disease is cured by your prescription? Even if this is so, you have only based it on your assertion, you have not given the evidence (Cp. the mass of actual examples produced by Freud) (in: Venturinha (2012))

Other remarks make the point that Wittgenstein was overusing psychoanalytic jargon. Sraffa remarks, for instance: "Remedy. Does it in fact cure?" He also stresses Wittgenstein's vocabulary in thís passage of the BB: "They state their case wrongly ... For if they don't wish to talk of ... they should not use." With these BB-passages in mind, Sraffa writes the following remark: "Psychoanalytic Dispute."

It was certainly after 1941 that Wittgenstein cut out the passages that directly describe his genetic method with psychoanalytic overtones; i.e., he does so after Sraffa's criticisms of passages that have a similar sugges­tion in the BB. ln order to avoid this suggestion, I take it, Wittgenstein decided to leave out of the last version of the PI some passages that might overstress the therapeutic character of his philosophy.

Keynes also recognized the psychoanalytic flavor of many of Wittgenstein's remarks. According to Bouwsma's report of a conversa­tion with Wittgenstein, Keynes "was much impressed with the idea that philosophy is like psychoanalysis" (Bouwsma, (1986, 36)). Keynes carne to this conclusion after reading at least part of TS 220 (FF) as a member of the committee that nominated Wittgenstein as professor of philosophy in Cambridge in 1939.7 Wittgenstein was not quite satisfied with the comment, in spite of hís indication of an analogy between his method and psychoanalysis (for instance, in FF §106, quoted above). Bouwsma reports that Wittgenstein thought he "had himself talked about philosophy as psychoanalysis, but in the sarne way in which he might say that it was like a hundred other things" (1986, 36; my emphasis). 8 Wittgenstein may have thought that the best way to prevent readers from taking the analogy too far would be to cut out of the book some of the psychoanalytic allusions - as Wittgenstein, in fact, did.

The Philosophical Investigations 229

After all, if one emphasizes too much the therapeutic character of the method, one might end up forgetting the method itself.

Now 1 move to the internal reasons for the absence of some remarks concerning the genetic method. Wittgenstein certainly thought that readers should be able to grasp his method through the several exam­ples that he gives of it in the Pl. When commenting about §133 above 1 did not mention what is perhaps the most important point of that section: the method that he has in mind is shown in the examples that he gives: "a method is now shown by examples and the series of exam­ples can be broken off" (Pl §133). The reader can follow the examples and can also. break them off once she is able to see for herself how to go on and look at philosophical problems in a new perspective. She might also stop doing philosophy when she wants. So, by following examples, one should be able to see how the method used in them works and thus learn a skill.

This is in accord with Wittgenstein's belief that he should leave work to the reader. His intention was that the reader find out for herself the kind of method that he was using: "l should not like my writing to spare other people the trouble of thinking" (Pl, preface). Thus, the examples should be sufficient to an attentive reader to understand his method, while its description could be redundant and "spare other people the trouble of thinking."

Another good reason for Wittgenstein not to state his method explic­itly and directly in the Pl could be the belief that this would not, in fact, teach one how to apply it. Being able to state the method does not mean being able to apply it. One can learn the method only by somehow working through it. lt is a skill that is to be learned.9 Wittgenstein explains his method, then, by means of exercises meant to demon­strate the skills needed to dissolve philosophical worries. There is, then, an obvious parallel between Wittgenstein's strategy of presenting his method by examples and the way we learn the meaning of words: usually, not through definitions, but through explanations in the form of examples.10

1 dose this section with remarks on the idea of a surveyable repre­sentation (uebersichtliche Darstellung) in the new context of Pl. ln the last chapters, we saw that the meaning of the expression changed after the BT. The goal was no longer the tabulation of rules in the "book of grammar" as in the BT. That kind of uebersichtliche Darstellung would be better translated as 'perspicuous representation.' Such tabulation would be useless, since it is not required by the task of the genetic method and could lead to meta-grammatical worries (as seen itl Chapter 4). When

230 Wittgenstein's Philosophical Development

talking about mathematics, Wittgenstein presents our problem with the mies: "The entanglement in our own mies is what we want to under­stand: that is, to survey (uebersehen)" (PI §125). What we need, then, is not a list of mies, as in the BT, for ít is in them that we are entangled. What we need is, rather, an overview of the use of words:

A main source of our failure to understand is that we don't have an overview of the use of our words. - Our grammar is deficient in surveyability.

A surveyable representation (uebersichtliche Darstellung) produces precisely that kind of understanding which consists in 'seeing connections.' Hence the importance of finding and 1nventing inter­mediate links. (PI, §122)

'Seeing connections' is not a question of seeing the underlying mies of language; rather, it is an overview of the use of words in their surround­ings. This passage, 1 think, gains íts tme significance when we take into account what Wittgenstein says about l).is investigation in the preface. There Wittgenstein claims that the "very nature of our investigation11

is what "rompeis us to travei criss-cross in every direction over a wide field of thought" (PI, preface). For this reason, presumably, "the sarne or almost the sarne points were always being approached afresh from different directions, and new sketches made11 (PI, preface). The points need to be approached from different directions, for one approach might illuminate some aspects of a point, but might also cast shadows over other aspects (MS 116, 55). Granted the connection between the preface and Wittgenstein's elucidation of an uebersichtliche Darstellung in PI §§122-5, one can read the whole book as a surveyable representation of how one confusingly gets into philosophical theorizing, and how attention to the use of words in their surroundings might avoid that. The representation is, thus, surveyable if it shows how misleading analo­gies from various directions lead us to misleading trains of thought that seem to ground philosophical 'musts.' This can be dane with interme­diate analogles, intermediate links, which show similarities and díffer­ences among the cases:

The solution of the problem consists in the elimination of the worrying aspect triggered by certain analogies in our grammar. Philosophy changes this aspect. By showing other analogies. Inserting interme­diate links. Etc." (157b, 14r; from 1937)

The Philosophical Investigations 231

Certain analogies present in our use of words naturally suggest philo­sophicaI difficulties. Wittgenstein's philosophy aims at changing the way we reason about those difficulties. Other analogies and/or inter­mediate links might change our way of Iooking at philosophical issues. The longest and more detailed intermediate link in the PI is, perhaps, the reading interpolation (Pl §§ 155-78) - but also the ideas of 'meaning as use' (Pl, §43) and of 'family resemblance' (Pl, §§66-7). We will see examples of "other anaiogies" used as intermediate links in Section 5.3.

5.2 Preli~inaries: the genetic method and the Tractatus

As seen in the last chapter, the application of the genetic method to the T is the most important difference between the Pl and Wittgenstein's previous works. One has to have in mind, however, that some criticisms of the T were modified after the first two versions of the PI (MS 142 and TSS 220-1; or UF and FF respectively).11 Therefore, 1 first explain some aspects of the critique of the T present in the first versions of the PI that were left out in later versions; then 1 come to the organizing ideas of the most thought-out critique in Sections 5.2.1-2. Finally, the critique grounded on the genetic method is presented in more details in Sections 5.3.1-3.

Some of the passages from the FF (TS 220) from 1937-8 in which Wittgenstein criticizes the Twere left out in later versions. Some of those are very dose to points made elsewhere in PI (compare, for instance, FF §93 to Pl: §§92, 97, 114).12 There are, nonetheiess, some critiques of the T that simply disappeared in the Iater versions of the Pl. TWo long passages, which make the sarne or a similar point, were excluded. The first is in FF §108:

1 said once (in the Trac. Log-Phil.) that an 'elementary proposition' is a concatenation of names. For to the name corresponds an object and to the proposition a complex of them. To the proposition "The bottle is right from the glass", if it is true, corresponds the complex consisting of the bottle, the glass and the relation right-left (or how one wants to call it). - The words "object" and "complex" employed against [their use in] language!! A complex of houses 'consists' of houses, and not of houses and their reciprocai positions! And if 1 say that 1 see three objects on the table, then 1 don't mean: the glass, the bottle and their special relationship. [ ... ]

232 Wittgenstein's Philosophical Development

The second follows immediately in §109 of the FF:

But 1 search, I search desperately, after a system, after a UNlTY of all propositions. - And now 1 become the prisoner of certain forms of expression of my language, I stay caught in the net of language. For if we say "The bottle has the quality blue" instead of "The bottle is blue", and if we say "The bottle stands to the glass in the relation­ship right" instead of 11The bottle is at the right of the glass", and so on, - then it can look, to be sure, as if each sentence were a concate­nation of names. For all words with, as it were, material meaning appear here scattered in a net of purely logical relationships. And again: To all words correspond objects; for "Paul" designates this, "is" designates this, "three" thís, and "apple" this. The picture held us captive ... (Wittgenstein's emphasis)

Peculiar to these passages is that Wittgenstein seems not only to be talking about a "picture" underlying the conception of language in the T, but also attributing to the T the defeQ.se of a very narve "Augustinian picture." ln the last version of the Pl "the picture that held us captive" seems to be the general form of propositions and the crystal purity of logic. ln the passage above, however, the picture is that all words refer to things, even relations. This is a distorted presentation of the T. First of all, because if complexes of things and facts collapse, the very first sentence of the T loses completely its point. Why would Wittgenstein write that "the world is the totality of facts, and not of things" (T 1) if a fact is, as he suggests in the passage above, identical to a complex of things? Second, relations don't correspond to things either according to the T. This is actually one of the interesting points of the pictorial conception of language of the T and the reason why Wittgenstein, it seems, calls propositions 'facts' there (T 3.14-3.1432). This insight is already expressed in 1913: "What symbolizes in 'aRb' is that R occurs between a and b. Hence, 'R' is not the indefinable in 1aRb111 (NL, 98). Also: "To symbols of this form [xRy] correspond couples of things whose names are respectively 'x' and 'y 111 (NL, 95). Third, numbers are exponents of an operation (T 6.021 and 6.03) and number is a form expressed by a variable in the T (4.1272 and 6.03). As the exponents of an operation, they don't correspond to anything. Contrary to what the quotation above suggests, numbers are not objects in the T! ln fact, one of the major points of the philosophy of mathematics of the T is precisely to deny that numbers are objects.

The Philosophical Investigations 233

The most obvious reason for the exclusion of FF, §§108-9 from the PI is that those sections are the expression of an unfair critique of the T. ln the preface of the PI Wittgenstein points out, indirectly, the reason for the changes when he writes: "Four years ago 1 had occasion to re-read my first book (the Tractatus Logico-Philosophicus) and to explain it to someone." This is not in the preface of the FF written in 1937-8. The re-reading certainly took place in the forties, be it 1941 or 1943.13 One must conjecture, thus, that in the process of explaining the T to "someone" (probably Bachtin) Wittgenstein had the opportunity to recall his old views and avoid wrongly attributing to the T some views that he did not defend there. The exclusion of these passages, 1 think, shows that Wittgenstein, at least for a while, might have considered the T worse than it in fact was. Before his re-reading, he oversimplified his own early philosophy.

5.2.1 A step-by-step procedure

Wittgenstein applies the genetic method to the T in the PI in six steps:

1. ln the opening sections of PI, Wittgenstein breaks with the idea that the meaning of a word must be an object and that language must be equivalent to the totality of propositions or sentences. This makes the whole project of the T suspicious.

2. He discusses simples: names and objects. Preliminary answer for ''why does it seem that we need them?": propositions make sense only if names name. Before looking at what really motivates the postulation of simples ("the sublimation of the logic of our language", §38), Wittgenstein looks at the belief that we can somehow reach simples.

3. He scrutinizes the view that we can reach simples (indefinables) via "logical analysis." Analysis in the T is taken to be similar to chemical analysis: it seems that we may find the simples by decomposing the sentence.

4. He discusses analysis as a tool that supposedly shows that the sense of propositions is determinate. That the sense is determinate means the following: "A proposition must restrict reality to two alternatives: yes or no. ln arder to do that, it must describe reality completely" (4.023). So it seems that the Law of the Excluded Middle must be valid for each isolated proposition, independent of the sense or truth of any other proposition. This is expressed in three ways in the T. A) A proposition can be analyzed into logically independent elementary

234 Wittgenstein's Phi/osophical Development

propositions. B) A name must name; so there must be simple names that name simple objects (substances). C) Concepts must be sharply defined, otherwise one cannot determine when a proposition is true or false (4.063); analysis supposedly shows the precise boundaries of concepts.

5. He makes clear the belief that sense must be determinate is an expres­sion of the idea of logic as absolutely a priori: the crystalline purity of logic.

6. He finally doses the critique of the Twith an evaluation of the general form of propositions. It was taken to be the result of the investigation based on the absolutely a priori character of logic. It was the zenith of the T, and from it Wittgenstein determined the nónsensicality of propositions of philosophy and his own propositions in T 6.n. ln the PI, it is trivialized (§§134-6). The idea of a general forro of proposi­tions that is put into question right from the beginning of the PI is finally, as it were, brought "back to the rough ground" (PI, §107).

5.2.2 The assumptions of the Tractafus

The assumptions of the T uncovered by Wittgenstein's critique in the PI are not equally significant. The assumption of the existence of primary elements seems for Wittgenstein in the T to follow from the absolute a priori nature of logic and it is supported by a specific view of what anal­ysis should be. These points can be seen when we answer the question of "Why primary (simple) elements are needed?" According to the T, the sense of a proposition is independent of the truth of any other proposi­tion or fact (4.061) and it must be completely determinate, i.e., it "must restrict reality to a yes or no" (4.023). This seems to entail that logically independent elementary propositions must exist and their expression must be part of the conceptual notation (this is shown in truth-tables). If such propositions must exist, it must be possible to find them by analysis. The paradigm of analysis that was in Wittgenstein's mind at the time of the T was Russell's theory of descriptions (T 4.0031; MS 116, 81). ln the T, an elementary proposition is logically independent only if its elements are simple and cannot be further logically analyzed, i.e., if they cannot be defined (T 3.26). These simple names name simple objects (T 3.22). So, again, an analysis must show that we can reach such elements. Similarly, if a proposition restricts reality to a yes or no (if its sense is determined), the concepts in the propositions must deter­mine whether each object falls under them.

The idea of simple names and objects and analysis is shared at least partly by Russell (the first two ideas are shared also, according

The Philosophical Investigatlons 235

to Wittgenstein, by Plato PI, §46), while Wittgenstein's views on the nature of logic in the T are, in one sense, not shared by either Russell or Frege.

Both Wittgenstein and Russell pushed aside the need to give exam­ples of elementary or atomic propositions: "Russell and 1 both expected to find the first elements, or 'individuals', and thus the possible atomic propositions, by logical analysis ... And we were at fault for giving no examples of atomic propositions or of individuals. We both in different ways pushed the question of examples aside" (WLC32-35, 11). Russell is also the philosopher that Wittgenstein mentions in PI, §38-9 as the one who wants to make the word 'this' into the best example of a proper name. Russell says, for instance, in The Philosophy of Logical Atomism, 62:

The only words one does use as names in the logical sense are words like 'this' or 'that'. One can use 'this' as a name to stand for a partic­ular with which one is acquainted at the moment. We say 'This is white'.14

On the other hand, the idea that concepts must be sharply defined certainly has a Fregean origin. ln §71 of the PI, when Wittgenstein criti­cizes the view that a concept must be completely circumscribed, he says explicitly that "Frege compares a concept to an area and says that an area with vague boundaries cannot be called an area at all." Wittgenstein is, in this passage, certainly referring to Frege's metaphor:

A concept, which was not sharply circumscribed, would correspond to an area that did not have complete sharp boundaries, but would become blurred in some places of the surrounding area. This would not be genuinely an area at all; and in this way a concept that is not sharply defined is wrongly called concept. (Grundgesetze der Arithmetik, §56)

The T's idea of the nature of logic is completely original. ln Wittgenstein's Tractarian eyes only the author of the T was, in fact, clear about the significance of the a priori nature of logic, for only he made clear the nature of logical necessity by means of his notation. Wittgenstein's basic insights in the T (the tautological nature of logic and the operational character of logical connectives) allowed him to think that he had showed, in the T, the true character of the a priori nature of Iogic (T. 6.112): logic does not depend bn any 'experience,'

236 Wittgenstein's Philosophica/ Development

self-evidence or empirícal assumptions. Frege (but also Russell) appealed to "self-evidence" as a criterion for logical 'propositions' and so mixed a psychological element with pure logic. Wittgenstein says in the T that "it is remarkable that a thinker as rigorous as Frege appealed to the degree of self-evidence as the criterion of a logical proposition" (T, 6.1271). How bad such a criterion, is shown by the inconsistency in the Grundgesetze that originates in Frege's fifth "logical law." Russell not only supposed that logic depended on 'self-evidence' and 'acquaintance,' an unclear 'experience,1 but also assumed more or less empirical propositions as part of logic.15 ln Wittgenstein's view iR the T, the nature of propositions and the nature of logical rules make clear that logic is free from psycho­loglcal and/or other empirical assumptions. Frege and Russell, because they do not understand the true nature of logic, are still representatives of the "old conception of logic" (T 6.125). Thus Wittgenstein's views on the a priori character of logic are what most clearly distinguish him from Russell (and Frege), while the doctrines of simples and analysis are (in some ways, but not in all ways) shared by Russell.

What has been said lndicates that the conception of logic of the T might be seen as the real metaphysics of the book. The 'essence of the world' is a model of the essence of language, for language, thought, and world are already articulated (Chapter 3, Section 3.2.2). There is an interesting passage in the opening of MS 157a, the manuscript written in 02.1937 used to delineate the first version of the PI, which makes this point:

It seemed that we had to win this understanding [of the structure of facts] by a special insight, but it was clear to us that we didn't want/ try to see through the facts, but through the language in which we talk about them.

And thus we wanted to know the essence of language. (The essence of the proposition, of inference, of grammar). And it seemed that in the answer to this question lay the answer to the question about what could be said about the 'essence of the world' and the answer to our questions.

And it was essential that that answer could be given once for all, therefore independently of any future experience. (MS 157a, 47r)

"Through the language," in its "essence," we see the nature of facts without the need of "future experience." So Wittgenstein's discussion of the essence of language (the "crystalline purity" of logic) indicates the motive behind his efforts in the T. It indicates what led him to make dogmatic assumptions about elementary propositions, logical names,

The Philosophícal Investigatíons 237

understanding and sense: "If we express the essence of the proposi­tion, then everything else must follow. It cannot, then, be anything different 11 (MS 157a, p. 70v). This assumed 'must' was the ground for many supposedly trivial views in the T, including the "trivial fact that a completely analyzed proposition contains just as many names as there are things contained in its reference" (NB, 11).

5.3 The genetic method applied to the metaphysics of logic of the Tractatus

1 present here the steps of the application of the genetic method to the T mentíoned in Section 5.2.1, except for step 1, which was already discussed in Section 5.1. I divide them, however, into three groups. ln Section 5.3.1, 1 go over simples and analysis; in Section 5.3.2, 1 go over the determination of sense and the nature of logic; in 5.3.3, I look at Wittgensteln's critique of the general form of propositions. The result of the whole process is that the fundamental metaphysical picture in the T is the absolutely a priori nature of logic. lts source is the belief that if such a picture does not hold then our rationality is at risk.

5.3.1 Simple names, primary elements, and analysis

Wittgenstein points to what engenders the assumption of slmple names in his discussion of indexlcals. He wonders how someone (Russell) can call 'this' a proper name. Wittgenstein is obviously applying his method and asking for the genesis of the idea. The motivation comes from a 11 logical11 view:

... one is tempted to make an objection against what is ordinaríly called a 'name'; and it can be expressed like this: the name ought really to denote a simple. And one could justify it, for instance, in this way: A proper name in the ordinary sense is, for instance, the word 'Nothung'. But the sword Nothung consists of parts combined in a particular way. If they are combined differently Nothung does not exist. But it is clear that the sentence "Nothung has a sharp biade" makes sense whether Nothung is still whole or is broken up. But if Nothung is the name of an object, this object no longer exists when Nothung is broken in pieces; and as no object would then correspond to the name it would have no meaning. (PI §39)

What lies behind the assumption that simples must exist is the bipo­larity of propositions and the Law of the Excluded Middle.16 Note

238 Wittgenstein's Philosophical Development

that Wittgenstein in the T accepts the elimination of Fregean senses through analysis put forward by Russell in On Denotíng (see Chapter 2, Section 2.2). lf Fa is a proposition, then -Fa is also one, and both must be possibly true. Each of them must be true if the other is false. lf Fa, however, is the form of a sentence like "Nothung has a sharp biade," then the loglcal laws do not seem to apply to it. Suppose Nothung does not exist; so "Nothung has a sharp biade" seems neither true nor false, for in this case -Fa is not true even if Fa is false. lt seems, then, that "Nothung has a sharp biade" must be nonsense. However, as a matter of fact, we do understand the sentence "Nothung has a sharp biade" and it obviously makes sense:

But it does make sense; so there must always be something corre­sponding to the words of whlch it consists.

The "solution" is, then, to assume, as a kind of "condition of possibility of sense," which we take to be a logical step, that there must be simple names and that "something" must correspond to them in arder to "save the appearances.11 Only in this way can we say that "a proposition that mentions a complex will not be nonsensical, if the complex does not exist, but simply false" (T 3.24).

This seems to give us, then, "strictly logical reasons" to denounce the usual proper names of our language as not really names, not names in the "purest sense," i.e., not names with a guarantee of reference. At this point (Pl, §39), nonetheless, the whole picture underlying the apparent necessity of simples based on the nature of logic is not open to view. This is something that is developed in §§40-136.

The next important point in §39, for our purposes here, is the sugges­tion that once it is assumed that simples must exist, one seems to need to find a way to make this claim plausible. We need, so it seems, at least a rough method to find them (even if our method does not clearly show that it works for ali cases). How can we find them? The T left things undecided to a certain degree by indicating only that names must be somehow given implicitly in the sense of sentences and that analysis must find them:

So the word 'Nothung' must disappear when the sense is analyzed and its place be taken by words whlch name simples. (PI, §39)

Wittgenstein does not give us a complete analysis of ordinary sentences in the T. He tells us that such a thing must exist. But in arder to make

The Philosophical Investigations 239

his claim plausible, he points to a possible way to answer the worries. The paradigm of the analysis is Russell's analysis of descriptions (T 4.0031). The reasoning underlying part of the philosophy of the Tis, then, the following. That analysis will reach simple names is a condi­tion of possibility for a sentence like "Nothung has a sharp biade" to make sense; it does make sense; therefore an analysis that shows them must be possible and Russell's theory of descriptions can provide an analysis (one presumes).

ln §40 Wittgenstein points to at least two confusions that bring one to such an "inference." First:

... it is a folecism (sprachwidrig) to use the word "meaning" to signify the thing that 'corresponds' to the word. That is to confound the meaning of a name with the bearer of the name. When Mr. N.N. dies, one says that the bearer of the name dies not that the meaning dies. (PJ §40)

The second confusion that underlies the train of thoughts presented above is an idea of logical analysis. The assumption was that linguistic or logical analysis is a decomposition of sentences into unseen and concealed parts, as if sentences were objects that can be decomposed in unforeseen ways (MS 113, 62r; BT, 101, 557-8).

A third confusion is related to understanding a proposition like "Nothung has a sharp biade." The analysis, supposedly, must show that this sentence is composed of logically independent atomic sentences, which are concatenations of sim pie names. Thus, Wittgenstein concludes in the T: "If we know on purely logical grounds that there must be elementary propositions, then everyone who understands proposi­tions in their unanalyzed form must know it" (5.5562). An "implicit" understanding of elementary propositions must be, then, assumed, as soon as we assume "strict and clear rules of the logical structure of propositions":

The strict and clear mies of the logical structure of propositions appear to us as something in the background - hidden in the medi um of the understanding. I already see them (even though through a medium): for I understand the propositional sign, I use it to say something. (PI, §102)

Even though it is quite doubtful that "everyone who understands propositions" knows that there are or must be elem~ntary propositions,

240 Wittgenstein's Phi/osophical Development

it seems more difficult to deny that anyone who understands a language knows that a proposition is either true or false. Wittgenstein's assump­tion is that we must know about the existence of elementary proposi­tions on purely logical grounds. Anyone can understand a proposition like "Nothung is Siegfried's sword" (to simplify matters, we are assuming here that this is not fiction); but then the names in it must have meaning, otherwise the proposition would not have a sense, which would imply that nobody could understand it. For understanding a proposition is, for the T, knowing what is the case lf it is true or false (T 4.024). So analysls must show what everyone knows "implicitly." This is, of course, a very doubtful assumption.17

ln a similar doubtful way, one assumes the existénce of simple objects. lt seems that if one denies the possibility of an underlying analysis, and the implicit understanding in the terms of the T, one has to deny the Law of the Excluded Middle. One has to deny at least, lt seems, that the law applies to ordinary language. But is not ordinary language logical? Is it not the means of understanding that we have to settle rational disputes? Wittgenstein ex;plains in the T that the proposi­tions of ordinary language are in perfect order as they are (5.5563) - as long as we assume that it must be possible to analyze them and present their concealed elements, the simple names. Now, since, supposedly, the simple names must be assumed, their correlates, the simple objects, must also be assumed (PI, §46).

Wittgenstein opens the sections of the PI dedicated to his discus­sion of simple objects by asking a question that any person applying the genetic method should ask: "What lies behind the idea that names really slgnify simples?" (PI, §46). Wittgenstein does not answer the question, but quotes a passage of the Theaetetus, in which Plato claims that the compositionality of language depends on simple elements, the names. ln the passage it is suggested that simples cannot be defined (they define everything else), but only named; this seems to imply that no precise account of simples can be given. The relevant point here is that simples are indefinables (T 3.26 and 3.261). Simples seem to be a conditlon of speech, since ali speech is a concatenation of names. Wittgenstein makes clear that the 'objects' of the T and Russell's indi­viduais were such primitive elements.

With the quotation from Plato, Wittgenstein wants to point out, of course, that similar ideas were around for many years, as is typical in philosophy (think of the many versions of realism and anti-realism through history). The point is that Plato, Russell and Wittgenstein made similar assumptions. Philosophical problems and "solutions" are partially repeated in the history of philosophy.18

The Philosophical Investigations 241

Even though there may be similarities between Plato's and Wittgenstein's simples (the lack of examples of the elements, for instance), it is not this similarity that is the most important point that Wittgenstein is making by quoting Plato. The point is, l think, to suggest that even though the accounts of simples are different, their roots may be similar. "Logical questions" seem to prompt Plato, Russell and Wittgenstein into the assumption of simples. One is inclined to think that indefinables must exist for logical reasons (one cannot define everything). It is, nonetheless, in Wittgenstein's T where purely a priori concerns about the nature of logic (and thus of rationality) give the grounds for the assumption. So "what lies behind the idea that names really signi!y simples" (PI, §46) is similar, but not exactly the sarne in the cases of Plato, Russell, and Wittgenstein. What lies behind the idea of simples or what brings one to such an assumption is what interests Wittgenstein most. He wants to show the confusions that surround the idea, for sure, but his ultimate target is the engine propelling such assumptions in the T.

Wittgenstein points out, after §47, that the very idea of simplicity and compositionality that is assumed in the claim that simples must existis a philosophical illusion. What is 'simplicity'? The simple nature of "the elements of reality" seems to indicate a very deep problem, but where do we get this idea from, if not from objects and the way we talk about them in our ordinary life? Wittgenstein, thus, confronts our deep ques­tion to a more humdrum one concerning chairs:

But what are the simple constituent parts of which reality is composed? - What are the simple constituent parts of a chair? (PI, §47)

The contrast here is not the one between a difficult question and an easier one, but between a case in which we know what we are talking about anda case of confusion, i.e., a case in which we should reject the question. Here we have an example of an "intermediate link" (PI, §122) in a surveyable representation of the generation of a philosophical puzzle. The unspecified way of talking about complexity and simplicity is linked to more humdrum notions of complexity and simplicity. The simples of reality, according to the T, must be such that they cannot be taken as composite: they are, supposedly, absolutely simple. If they could be composite, they could not be what is named by simple names, for they could be destroyed and so fail in their essential task of securing reference for simple names. The philosopher who defends the exist­ence of simples would have to say that the simplicity he has in mind

242 Wittgenstein's Phílosophical Development

is just like the humdrum one, except that it does not allow any kind of logical compositeness. That is, the philosopher grounds his use of words analogically, but points also to a difference in use. The question is whether such an exceptional use is still coherent.

Supposedly, a simple object canrtot, in its tum, be considered composite in comparison with a third object. The problem is that the analogy of the philosopher with our notion of simplicity conveniently begins with our notion of simplicity and ends up with a notion that has very little to do with it. It is in clear contrast with the way we usually talk about composite objects. The parts of the chairs can be legs, back, and seat, for instance, or bits of wood, or molecules, etc., depending on what the context of the question makes appropriate. A éhair can itself be consídered a simple object in the líving roam, when compared to a composition of the totalíty of the furniture that, for instance, a deco­rator has in mind. What we take as simple in one occasion may be taken as composite in a dífferent one:

(Is the calor of a square ín a chessboçi.rd simple, or does it consist of pure white and pure yellow? And is white simple, or does it consist of the colors of the rainbow? ... ) (PI, §4 7)

The problem with asking for the absolute simple is that in so doing we lose the support for our most basic distinction concerning composite and simple objects. We lose the environments, the surroundings, that help us to understand their meaning. Simple is simple in relation to something that is complex; but any object considered simple can be considered composite in a different sítuation:

To the philosophical questíon: "Is the visual image of this tree composite, and what are its component parts?" the correct answer is: "That depends on what you understand by 'composite'." (And that is of course not an answer but a rejection of the question). (PI, §47)

The question is rejected because no clear sense of the notion of 'composite' is made. Could we, however, make sense of 'composites' and 'simples' with a more detailed background of the views of the T? The remarks up to this point do not show yet what is the misleading pícture generating the view that simples must exist. Before indicating the misleading picture, Wittgenstein looks at various analogies that could make one belíeve that such a view is plausible in the sections following §47. Partia! analogies seem to give plausíbility and support

The Philosophical Investigations 243

to philosophical claíms. The point of indicating such analogíes is to find out its "rough material" and how such analogíes are mísleadingly extended. Wittgenstein indicates some "objects" that we can understand as possessing, in an analogical way, some of the properties of simples. Wittgenstein takes, for instance, Plato's claim ln the Theaetetus (quoted ln PI, §46) that u ••• everything that exists ln its own right can only be named, no other determination being possible" and looks for some­thing that could have analogous characteristics. Again, Wittgenstein is demonstrating "intermediate links," analogies that partially hold (PI, §122). We could say, Wittgenstein points out, that the standard meter in Paris, as a. means of representation or standard of correctness of what counts as one meter, is neither one meter nor not one meter long. Here we could say that no "determination is possible." But this only means that we don't ask how long the standard meter is, for it is used as our "ultimate" criterion to decide whether an object is one meter long (PI, §50). Some color samples, analogícally, may be seen as simples, for we usually don't decompose colors (PI, §48). But again, it does not mean that we could not do it (PI, §53).

We say also that the red flag was burnt, but we don't say that we bum redness (PI, §57). Here, it seems, we have something that is 'indestruct­ible' just like the simples should be. One thinks that since "we bum a red flag, but not redness, or 'red itself1

111 then redness or 'red itself' is

indestructible. It seems that the partial analogy makes the philosophical claim plausible. Nonetheless, when we extend the analogy we forget that we say things like "the red is vanishing" (PI, §57).19 This suggests that the "indestructibility" is only relative to a certain use of the substantive redness. If we paid attention to these different uses of 'red,' however, we would not be inclined to push the analogy further. To each use of simples ln philosophical discussions that seems analogous to some use that we usually have, Wíttgenstein points out how the analogy does not extend any further. The intermediary links show how different, and not how similar, the varíous ideas of simplicity are.

One says sometimes that the whole is destroyed, but the parts are intact. This is the case, for instance, when we talk about chairs: "the chaír ís destroyed, but íts legs are still like new and we may use them to buíld a new chair ora bookshelf! 11 We do not say, however, that the legs are indestructible or absolutely simple because they are intact. While philosophizing, one may be ímpressed by those partial analogies and feel confident that the philosophical talk makes perfect sense (PI, §59). The analogies break down, however, when we remínd ourselves of further examples ln which they don't hold. ln one sense we do talk

244 Wittgenstein's Phi/osophical Development

about simples; we also speak metaphorically of indestructible elements, etc. We do not, however, have any notion of what "absolute" simplicity might be. ln the end, we are merely guided by a partial analogy:

We see constituent parts of something composite (a chair, for instance). We say that the back is part of the chair, but that it itself is composed of different pieces of wood; whereas a leg is a simple constituent part. We also see a whole which changes (is destroyed) whíle its constituent parts remain unchanged. These are the materials from which we construct that picture of reality. (PI, §59; my emphasis)

Our picture seems to make sense because it finds apparent support in one use of language, i.e., it finds analogical support as long as we don't work out the details of the analogy (relative simplicity, destructi­bility and changeability of 'simples,' etc.). One constructs a "picture of realityu grounded on partial analogies. But why does someone insist on holding such a view?

5.3.2 Analysis, determination of sense, and the "crystalline purity of logic"

ln §§60-64, Wittgenstein criticizes the idea that analysis shows some­thing hidden and more fundamental, the real structure of proposi­tions. This connects to the critique of absolute simplicity at the levei of completely analyzed propositions. The completely analyzed proposi­tions supposedly show what we really mean, what we really want to say with our complex sentences. As if the specific state of affairs described in a proposition could be made explicit in analysis (see Chapter 2, Section 2.2.2). If what we really mean is the analyzed form of a sentence when we say it, then it seems that when we say "The broom is in the comer" what we really mean is "the broomstick is there, and so is the brush, and the broomstick is fixed in the brush" (PI, §60) this is the point made by an interlocutor, perhaps the author of the T, one could say, who tries to give analogical plausibility to his claim. However, if we look at what we usually do and say concerning what we mean, we immediately see that there is something wrong here:

If we were to ask anyone if he meant this, he would probably say that he had not specially thought of either the broomstick or the brush. And that would be the right answer, for he did not mean to speak either of the stick or the brush in particular. (PI, §60)

The Philosophical Investigations 245

"l did not mean to speak of the stick or the brush in particular" is what one usually would answer here. Otherwise, one would particularly refer to the brush. Suppose, on the other hand that someone would give the arder "Bring me the broomstick and the brush that is fitted on it!" instead of "Bring me the brush!" ln this case, one would only be using a weird expression (one would be "putting ít oddly"). Obviously, the "analyzed sentence" would not be better understood. Perhaps, it would be misun­derstood: the person obeying the arder might tear apart the broom. We have no grounds, thus, to think that an analyzed sentence is the hidden structure that is really meant and really understood in our sentences. ln this case we.seem to have a good analogy for our idea of analysis, but further uses of it show that the analogy does not hold throughout.

Could we say, however, that the analyzed command is at least a more explicit version of the sarne or at least that the commands command the sarne? lt depends on what we do with them. They may serve the sarne purpose, but they may not. We have to look at the surroundings of the commands in order to answer such a question. We would not think it correct, for instance, if someone brought broomstick and brushes that don't fit together when the command was "Bring me the broom." So we may actually lose something in the analyzed form, contrary to what was first expected. Again, here, an analogy that seems to make a philo­sophical view more plausible turns out to be rather misleading. When we philosophize, nonetheless, the talk about the analyzed form "readily seduces us into thinking" that the analyzed form "is the more funda­mental" (PI, §63), even though we are not always prepared to accept an analyzed form as the best or correct form of expression.

So what makes us forget strong disanalogies and "seduces us ínto thinking that we need to find an ultimate analysis of a proposition?" ln §65, an interlocutor reminds us that the analysis was, in fact, in the service of something deeper and more important in the T. Wittgensteín prefaces the interlocutor's objection by qualifying ít as "the great ques­tion that lies behind ali these considerations" (PI, §65). This indicates that from here on ín the PI we move closer to what was the engine, the picture underlying the philosophy of the T. The whole idea of analyzing propositions into logically independent elementary propositions was at the service of the presentation of the general form of the proposition, arguably the most important achievement of the book. So, the inter­locutor objects:

"You take the easy way out! You talk about ali sorts of language­games, but have nowhere said what the essence of a language-game,

246 Wittgenstein's Philosophical Development

and hence of language, is: what is common to all these activities, and what makes them into language or parts of language. So you let yourself off the very part of the investigation that once gave you most headache, the part about the general form ofthe propositions and of language" (PI, §65)

Wittgenstein, so the interlocutor suggests, does not talk about propositions anymore and seems to concentrate on language-games; contrary to the strategy of the T, however, he does not explain what is common to all language-games or explain what makes a language-game part of language. Wittgenstein neither presents the general form of language-games, nor shows where language-games 'and language are circumscribed by making explicit the supposedly implicit rules of language. Thus, the interlocutor could argue, Wittgenstein is not showing what the limits of Ianguage are, as was presumably the major goal of the T and, therefore, his critique of analysis does not seem to take into consideration what sustained it or what was its major motivation.

Wittgenstein agrees with the interlocutor that he is not looking for the general circumscription of language-games or Ianguage (PI, §65). To oppose the apparent need for the general form of propositions and the precise circumscription of language, Wittgenstein offers an alternative picture or analogy of what may underlie language. He makes the point with 'game.' lnstead of thinking one characteristic underlying ali games, one can think of games as a family of characteristics interwoven, over­Iapping only partially (PI, §67). Wittgenstein is not, however, arguing that the real structure underlying concepts is family resemblances (see Chapter 3, Section 3.4.2, and Chapter 4, Section 4.3.3). He is only offering an alternative, clear to view, for someone inclined to think that there must be a 1general form' - as he was in the T.

lf one looks more closely at the way we apply our concepts, it will look very doubtful that one can finda definition of game (PI, §66), a common feature of games, for one only sees 11a complicated network of similarities overlapping and criss-crossing: 11sometimes overall similarities, some­tímes similarities of detail" (PI, §66).2º The interlocutor suggests, then, two definitions that seem to take into consideration that the definition of concepts must include family resemblance cases. The first kind of definition suggested is a 11disjunction of ali common properties," for instance: Game =def. (amusing v competitive (losing or winning) v use of skills v decided by Iuck). To this suggestion Wittgenstein replies that the interlocutor is only "playing with a word." lf we call a whole set of disjunctions a 1common property,' then we distort what 'common'

The Philosophical Investigations 247

means. This would be like sayíng 11something runs through the whole thread, namely the continuous overlapping of many fibres 11 (PI, §67). However, we can distinguish the case in which there are only overlap­ping fibres and the case in which there is, besides the fibres, say1 a wire running through them. Moreover, such a definition does not do what it is supposed to do, namely1 to delimit the notion of 'game.' We amuse ourselves with movies, we use our skílls to work, and serious matters might be decided by luck.

Pointing out that the interlocutor is misusing language is not a deci­sive move against him, for even though he is misusing the expression 11something in common," he presents a formulation of what a definition should lool<like. ln any case, this kind of definition already incorporates the complexity of family resemblance and, thus, distances itself from what is common to all cases. The interlocutor must then at least admit that definitions would look far more complicated than he wished.

The second kind of definition of 'game' proposed is the "logical sum of inter-related concepts,11 for instance: Game =def. (amusing-games V competition-games V skill-games V luck-games). This definition certainly gives "rigid limits1

' for the concept and may enable us to decide whether each possible activity falls under it. Such a definition would correspond to the requirement of determination of sense of the T (4.023) and to Frege's 11principle of completeness of definitions11 in Grundgesetze der Arithmetik II §54. lt, nonetheless, misses one impor­tant aspect of our actual use of concepts, namely, the extension of the application of concepts to new cases. We may draw a line and give the concept rigid limits "for a special purpose" (PI, §69), but we don't need to do it; and if we do it, we may lose something important: we may not be able to incorporate new cases after accepting the definition.

An important aspect of Wittgenstein's family resemblance picture thus emerges from the second suggestion made by the interlocutor: it expresses the way that we actually apply concepts like 'game to new instances. First, we certainly do not use the definition of 'game' when we use the word. We do not consult the definition (though maybe we could if we had one) to decide whether an application of the word is correct or not. Second, we do consider various similarities between various activities that we call 'games' when we are not sure whether a recently invented activity is a game. We did not look at our definition of game when we called 'games' some of our dangerous time-wasting-activities in front of computers. Ifwe were not sure about calling them 'games,' we would have asked questions like "Should we call those things 'games', if we do not play them with other people? 11

; and answered something like:

248 Wittgenstein's Philosophical Development

''Oh, it does not matter, for we also can play card games, like patience, alone!!"

The way we actually apply our concepts to new cases presents two prob­lems to someone who claims that "there must be something common" (PI, §66) to ali instances of the application. First, do we really need such a definition if we in fact use the concept without one? Second, if we had such a definition, would it be compatible with our practices? The need for such a definition will be, of course, decided according to our purposes: "we can draw a boundary - for a special purpose" (PI, §69). But even if we draw such a boundary in order, say, to apply the excluded middle in a logical system, it does not follow that our cu.rrent ordinary concept is unusable (PI, §69). ln fact,, uncircumscribed concepts may be, in many ways more useful than circumscribed ones. Sometimes we may need the uncircumscribed concepts (PI, §71), for they allow us to extend the concept.

The second questíon presents a more serious threat to someone who wants to find a definition. If the definition is not in accord with our practices, it seems to lose its point and )ose its utility. Would we recog­nize the rigid limits of a concept as a limit that we always wanted to draw? Not necessarily:

If someone were to draw a sharp boundary I could not acknowledge it as the one that I always wanted to draw, or had drawn ln my mind. For I did not draw one at ali. (PI, §76)

Why should we recognize, say, the definition of 'game' as g1vmg boundaries for our concept if we never really worried about finding such boundaries and used the word w.ithout drawing them? We could rather think that the boundaries would only make our practices diffi­cult and confuse us. Perhaps we would have to exclude some things that we call 'games' in the application of the concept and we would not know how to apply the word to new instances. A philosophical definition may then look more like something idling in language. As a matter of fact we do use concepts without rigid limits ali the time and we do understand each other when we do it. We don't need to ask for the definition of 'leaf' when someone tells us that the ground was covered with leaves.

If the ideal of exactness seems rather like an imposition that can have even bad results in practice, why do we come to think that we need the sharp boundaries for our concepts? It may seem that if one does not give precise boundaries for a concept one does not know what one is

The Philosophical Investigations 249

talking about. It may seem that one understands a concept only after eliminating ali possibilities of making a mistake. The sarne seems to be true about proper names: if one does not pick out exactly one object when one talks about someone or something (PI, §79), one seems not to know what one is really talking about. lt seems that if a name does not have a "fixed and unequivocal sense" in ali possible circumstances, it is not really a name. ln fact, however, we use names "without a fixed meaning" (PI, §79). Philosophers cannot go as far as denying that people know the meaning of words. For to say that is like saying that we don't know what we are talking about most of the time. So the way out seems to be to say that the meaning needs an analysis that will show a suppos~ edly implicit understanding. Thus, the idea of the T was to assume that ordinary sentences are in perfect logical arder and that such a logical arder must be shown by logical analysis, an analysis that supposedly shows the implicit understanding; analysis uncovers, supposedly, what is really meant (PI, §63).

We do, in fact, say sometimes that one knows and yet is unable to say something. This ordinary case of knowing and not being able to say seems, then, in a first moment, to support analogically the idea that one knows lmplicitly the meaning that the philosophical analysis makes explicit. However, when we look more closely at those cases, the inter­mediary link makes clear that the analogy does not hold. What is the difficulty involved in knowing and not being able to say something? Wittgenstein suggests several examples:

Compare knowing and saying: How many meters high Mont Blanc is -How the word "game" is used -How a clarinet sounds. Someone who is surprised that one can know something and not

be able to say it is perhaps thinking of a case like the first. Certainly not of one like the third. (PI, §78)

One could know how a clarinet sounds and be unable to say it. ln this case, one could perhaps imitate the sound of a clarinet instead of saying how it sounds. Usually, we do not describe sounds (as we do not describe a primary calor). Thus, in such a case, one is not surprised if someone knows how a clarinet sounds and is not able to say it. One could, perhaps, describe it, however, in a roundabout way. Aaron Copland, for instance, describes the sound of a clarinet in the following way: "the clarinet has a smooth, open, almost hollow sound."21 This,

250 Wittgenstein's Phílosophical Development

one could think, is a specialízed way of describing. However, to know how many feet high Mont Blanc is and not being able to say it would be a very strange case of ignorance, and would surpríse us. Unless the person who knew and was unable to say ít was physically restrained (say, he has been gagged), it would be difficult to make sense of this statement. But isn't the case of the boundaries of a concept and the fixed meaníng similar to the case of the height of the Mont Blanc? After all, in philosophical analysis we are looking for a case of precise meas­ures that can easily be described, if supposedly known. What we do in philosophy, thus, is to apply analogically the third case to the first. This misleading analogy seems to give plausibility to the idea of "implicit knowledge" (see T 5.5562 on elementary propositions). i

But what about our knowledge of how the word 'game' is used? ln arder to express knowledge of how the word is used we simply describe examples of the use of the word. This knowledge is not buried in our understanding of 'game,' it is not waiting to be discovered by philosoph­ical analysis. The philosopher who describes the use of words derives his knowledge from the very language that everybody else speaks. lt might be difficult to say how words are used because their use is vari­egated. This, however, is not the kind of difficulty that the philosopher looking for an analysis has in mind.

lf one looks at how words are really used in language (Pl: §§79-80), one finds that the lack of analysis neither impairs the use, nor commits us to speak nonsense, íf the meaning is not fixed or the named object does not exist (thís was presented as a major worry in Pl §40). Words can be explaíned in various ways and an explanation is not fixed for ali circumstances. This is the point of the example of 'Moses':

Consider this example. lf one says "Moses did not exíst", this may mean various things. lt may mean: the lsraelites did not have a single leader when they withdrew from Egypt - or: their leader was not called Moses - or: there cannot have been anyone who accomplished all that the bible relates of Moses or: etc. etc. (PI, §79)

Depending on what we mean by 'Moses,' depending on which expla­nation we use to explain the name, the sentence '"Moses did not exist' acquires a different sense, and so does every other sentence about Moses" (Pl, §79). ln the use of the word 'Moses' we don't find a "fixed and unequivocal use ... in all possible cases" (Pl, §79). We are not in a position to simply use one definition, one explanation, of the name 'Moses.' One or more explanations might tum out to be false and, in

The Philosophical Investigations 251

such a case, we would simply change our explanation (or definition). If we had thus a supposedly final definítion, it would be unpractical and useless in such a case. One could think that some explanations are more essential than others and that those are to be used in a final definition. However, the very boundaries of the relevant and the irrelevant are not clearly traced; in general, "we are not equipped wíth rules for every possible application" (PI, §80).

How can we then apply logic, if we don't assume an analysis that shows how meaning is fixed and that the sense of each sentence is determinate? (PI, §81) It is better to see logic, then, as an object of comparison imd notas an ideal that underlies language, as was assumed in the T: "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 must be playing such a game" (PI, §81).

Why does one come to think that the role that logic must play is the role of an underlying structure of language such that "if anyone utters a sentence and means and understands it he is operating a calculus according to definite mies" (PI, §81)? The answer to this question shows, at last, the metaphysical picture that underlies the T.

5.3.3 Logic and the general form of propositions

§89 of the PI begins the analysis of the most fundamental assumption of the T with the question "ln what sense is logic something sublime?" The sublimity of logic lies in its apparent universality and its independ­ence of any empirical facts:

For logic seemed to have a peculiar depth a universal significance. Logic lay, it seemed, at the foundation of all the sciences. - For logical investigation explores the essence of all things. It seeks to see to the foundation of things and shouldn't concern itself whether things actually happen in this or that way. It arises neither from an interest in the facts of nature, nor from a need to grasp causal connections, but from an urge to understand the foundations, or essence, of every­thing empirical. (PI §89)

Logic seems to be universally significant because any science seems to need to follow its rules. Its rules are the most fundamental rules of thinking and regulate all rational dispute. Therefore, it seems that logic's concern is only with rules that are given a priori independently of what actually happens. One grasps those rules supposedly a priori by grasping the nature or essence of the proposition, which expresses the essence of

252 Wittgenstein's Philosophical Development

thought (PI: §§92-5); after all, a thought is a meaningful proposition (T 4). The particular depth of logic derives thus from the assumption that it is the "essence of thought." Since an "unlogical world," whatever this could mean, is unthinkable, logic supposedly expresses the essence of the thought and world:

Thinking is surrounded by a nimbus. - lts essence, logic, presents an arder: namely, the a priori arder of the world; that is, the arder of possibilities, which the world and thinking must have in common. But this arder, it seems, must be utterly simple. (PI, §97)22

As Wittgenstein says in the T, "logic sets the standard of simplicity": "simplex sigillum veri" (T 5.4541). The a priori logical arder in the T is "utterly simple" because the fundamental notion of logic, the proposi­tion, is given with one single rule, the only primitive sign in logic (T 5.472): the general form of propositions. lt expresses the "essence of propositions" (T 5.471) and in this way gives also the limits of language. Since logic, thought, and world are intrinsically connected in language by means of a priori rules of logic and tacit conventions used to project the world, logic expresses the 'essence' of the world (see Chapter 3, Section 3.2). The general form thus expresses the goal of the book: it presents the limits of the expression of thoughts and this limit is drawn in language (see T, preface). Since "language is the totality of propo­sitions" (T 4.001) arid "thought is the meaningful proposition" (T 4; Ogden's translation), by giving a rule for the construction of any propo­sition, the limit of what makes sense to say is drawn inside language (see T preface and 4.11, 4.lln). This limit shows that philosophical prob­lems are posed because "the logic of our language is misunderstood" (T, preface). Only one who does not understand that there is a limit for language, one who is not clear about the general form of proposition, thinks that he can answer philosophical questions with propositions that describe essential features of the world.

The a priori arder of logic in the T was thought to be utterly simple also because all propositions of logic can be generated by means of the general form of propositions and all thinking (calculating in language) is reduced to a simple operation. Since all those propositions are true independent of any facts, being tautologies, the propositions of logic finally find a clarification of their a priori status.

Beca use this a priori arder is not immediately found in language, it has to be assumed "on purely logical grounds" that it is somehow under­lying the use of language. Even in the vaguest sentence "there must be

The Philosophical Investigations 253

perfect order" (PI, §98). The existence of elementary propositions, the final points of analysis, was assumed in the T in exactly these grounds: as a necessary condition of the underlying understanding of language. Everyone who understands ordinary propositions knows implicitly that they exist (T 5.5562). Logical structure is, thus, somehow hldden in the understanding of sentences (PI, §102).

Wittgenstein took the logic of language to be hidden because he had an idealized conception (a mythological conception, one could say) of the rules of thinking in the T. The "purely logical grounds" were only the expression of an ideal of logic imagined as being as pure as crystal:

It is prior to all experience, must run through all experience; no empirical cloudiness or uncertainty can be allowed to affect it - lt must rather be of the purest crystal ... (PI, §97)

Note that in a letter to Russell from 03.13.1919 Wittgenstein writes that even though "nobody will understand" the T, it is as clear as crystal (CL, 111). It is as clear as crystal, supposedly, because it expresses the purest crystal in its notation. This analogy, símile, or picture of the /(purest crystal" seems to sustain or be mixed with the idea of a purely a priori structure of language, logic. This ideal does not seem to be an ideal (a construction); it seems to be the /(truth itself" (T 5.5563), simple and concrete, for it must underlie all language and all thinking:

... This crystal does not appear as an abstraction; but as something concrete, indeed the most concrete, as it were the hardest thing there is (Tractatus Logico-Philosophicus No 5.5563)23

The "hardest thing" is what underlies language, that which is suppos­edly the ultimate essence of thinking: logic. Note that in Wittgenstein's early notebooks he wrote that his method was to see the /(hardness of the soft,11 i.e. see the logic in our ordinary language (NB, 44). It appears to be crystal pure, simple and concrete because it seems to regulate everything that is thought, said or done.

The problem with such a conception or picture of the crystalline purity of logic is that one may not see that it is nothing more than a picture, i.e. a rough analogical idea of logic. One is not aware that one is dealing wlth a picture. The kind of investigation pursued in the T required, from the beginning, that logic must be crystal pure. lt /(was a requirement" and "not the result of the investigation" (PI, §107, Wittgenstein's emphasis). ln the T Wittgenstein was held captive by the

254 Wittgenstein's Philosophical Development

picture because "it lay ín our language and language seemed to repeat it to us inexorably" (Pl, §115). lt seemed that since logic grounds all rational dispute it must be prior to, and independent of, any dispute in order to guarantee the possibility of agreement and disagreement. What regulates agreement and disagreement, so it seemed, must be prior to agreement itself.

We can see, however, that it was only a requirement if we examine the actual language: "the more narrowly we examine actual language, the sharper becomes the conflict between it and our requirement" (PI, §107). This conflict becomes clear to Wittgenstein, as shown in this book, when he comes back to philosophy and tries to analyze prop­ositions that seem to go against the requirement that elementary propositions are logically independent (Chapter 1). What elementary propositions could be in reality was not what they had to be according to the requirements imposed on the "application of logic" in the T. The examination of the actual language also shows that the requirement of definiteness of sense was an imposition that does not fit our use of words. We use words without a fixed meaning and this is quite impor­tant for our common understanding.

It is tremendously important for the view of the T that the require­ment of the existence of a pure crystalline logic seemed to be, in fact, fulfilled by the general form of propositions. It seemed to give the key to such a conception. It was as if it could show the nature of language and that logic was completely independent of anything empirical. Thus, it was as if Wittgenstein had found in it the exact expression of the crystal pure logic. It seemed to be "something pure and clear-cut," the expres­sion of "the nature of the real sign" (Pl §105). The 'real sign' is, of course, the logical notation of the T.

But what happens if we give up the picture, "if we see that what we call 'sentence,' 'language' does not have the formal unity" (PI, §108) that Wittgensteín imagined that was given by the general form of prop­ositions? An interlocutor, 1 take it, in §108 asks: " - But what becomes of logic now? lts rigor seems to be giving way here. But in that case doesn't logic altogether disappear?" Those questions are important in two different ways: first, they give one of the best clues to why the picture of crystalline purity is so strong; second, it asks for the real (non-mythological) status of logical necessity. These two points are related.

Why do we, or Wittgenstein in the T, come to believe that logic is "pure"? Or: Why are we, or Wittgenstein in the T, inclined to take for granted the crystalline purity of Iogic? lt seems that if logic is not pure,

The Philosophical Investigations 255

it Iases its rigor and "logic altogether disappears11 (PI, §108). With this in mind, we think that we cannot, however, afford to lose the rigor of logic, for it has a central part in our lives (in our form of life): it grounds rational dispute, it seems to be the basis of ali communication. We might then continue our reasoning and think that the only alternative to the "disappearance of logic" is to take it as an "ideal language" that is "more perfect than our everyday language" (PI, §81). Finally, we might introduce the ideal as implicit, hidden, in the nature of proposition and language. What is in fact peculiar in our quest for the best expression of the ideal of a crystalline pure logic is the reasoning that "the ideal 'must' be in reality" (PI, §101), i.e. in the very nature of proposition:

/

Tractatus Logico-Philosophicus (4.5): "The general form of propositions is: This is how thlngs are." That is the kind of proposition one repeats to oneself countless times. One thinks that one is tracing nature (Natur) over and over again, and one is merely tracing round the frame through which we look at it. (PI, §114; my emphasis)

We, thus, misunderstand "the nature of this 'must111 (PI, §101). We believe that we have a kind of logical inference that leads us to it, while it is no such thing. If it is not a real inference that brings us to the conclusion that the ideal must exist, what is it then? Only "deep disqui­etudes" prompt us; these have "their roots ... as deep in us as the forms of our language and their significance is as great as the importance of our language" (PI, §111). The deep disquietude that induces us to go into a false train of thought and conclude that the ideal must exist in the very nature of proposition is precisely the fear that logic will other­wise have no rigor, that it would otherwise disappear:

To understanding [communication] by means of language belongs not only an agreement in the definitions, but [also] (weird as this may sound) an agreement in judgments. This seems to abolish logic, but does not do so. (PI, §242; my translation)

Because we fear that the need for any kind of agreement, in any sense, prior to logic will abolish or jeopardize any possible agreement in all kinds of argument ar thinking, we come to the 'conclusion' that the ideal must exist. lt seems that there is too much at risk; rather, it seems that everything is at risk. The view according to which logic is purely a priori assumes that "logic is at the bottom of all possible agreement in language and its rules are used as the rules that regulate all possible

256 Wittgenstein's Philosophical Development

dispute that may arise, therefore it must be purely a priori" is a good "transcendental" inference. The nature of this premise and of this tran­scendental 'inference' is psychological, however, rather than logical. lt is the fear of losing the hard core of rationality, logic, that prompts one to find the essence of language and thus the a priori structure of logic.

Here we are forgetting the kind of creatures we are. We do have hard logic grounding our disputes; however, we are not completely clear about what 'hard' means, i.e., we are not clear about the role that logic plays in our forro of life. lt seems that if we don't have a crystal pure core of rationallty we have to allow for a kind of random arbitrariness and make everything depend on arbitrary agreements of what we take to be true or false. lt seems that abandoning the picture óf the pure crys­talline logic entails a forro of relativism. This alternative is, however, again, the expression of false trains of thought:

"So you are saying that human agreement decides what is true and what is false?" - What is true or false is what human beings say; and it is in their language that human beiµgs agree. This is agreement not in opinions, but rather in forro of life.(Pl, §241)

What is in quotation marks above, 1 take it, is what an interlocutor who fears the lack of solid logical ground asks (such an interlocutor could be, of course, the author of the T, Russell or Frege). Wit1:genstein's answer is that what we say is true or false, as opposed to what we agree on. We decide what is true or false based on what we see in the world around us and we do it because we agree in language, i.e. we agree in a forro of life. But could we have a different agreement? Wittgenstein does not say a lot about this in the PI. I take it that this was one of the roles of his philosophy of mathematics, which was originally intended to be the second part of the PI. 24 Wittgenstein in the PI, however, points to one kind of agreement that we have in our forro or life:

Disputes do not break out (among mathematicians, say) over the question of whether or not a rule has been followed. People don't come to blows over it, for example. That is part of the framework on which the working of our language is based (for example, in giving descriptions). (PI, §240)

It is a fact that mathematicians agree, almost always at least, about rules that we follow when, for instance, counting. If they did not, we can imagine that we would have a very different kind of 'mathematics' or

The Philosophical Investigations 257

no such discipline at all. This is an obvious and important fact that we are prone to forgetting (PI II, xi, §341). This fact indicates the surround­ings of our practices. It indicates the significance of concepts of math­ematics and logic:

What we have to mention in order to explain the sígnificance, mean the importance, of a concept are often extremely general facts of nature: such facts are hardly ever mentioned because of their great generality. (PI, §142)

Our agreement is a fact (empirical, if you wish) related to how we live our lives, tdwhat kind of beings we are. Something harder than such a fact, one will not find (see Chapter 3, Section 3.4.3). Now, is that empir­ical fact itself a necessary condition for mathematics without which mathematics would not be possible? It is difficult not to ask for further reasons and see this stop as a full stop (PI, §654). Here one is inclined to claim, for instance, that we only have the concepts we have because of the natural facts in our history. ln TS 229 from 1947, the continuation of TS 228 which Wittgenstein used as the basis for remarks in the PI, he says:

I don't say, however: if the natural facts were different, then we would have different concepts. This is a hypothesis. lt is of no use for me and it does not ínterest me. (TS 229, §716; similar remark in PI part II, xii, §365)

That we have concepts x and y if facts w and t are the case is a scien­tific hypothesis that needs empírica! investígation to be established as true. Wíttgenstein does not want to defend such a hypothesís. What he wants to do is simpler. We in fact agree, when dividing 12 apples between two people, that each gets 6 if they want to have the sarne number; we also agree that a child that skips numbers when learning to count is not counting correctly, and so on. Wittgenstein stops at the agreement in language, the empirical fact, and not in what makes the agreement possíble: 111 have not said why mathematicians do not quarrel, but only that they do not" (Pl II, xii, §341). This fact, when we think wíthout philosophical prejudice, is more certain, more trivial, and more profound (in any sense of those words) than any philosoph­ical claim concerning "conditions of possibility" or foundations. lf one looks for something more fundamental, one might find something else: a philosophícal mythology (say, a platonic realm of logical objects or

258 Wittgenstein 1s Philosophical DeYelopment

the structure of the mind) that makes some of us feel unrealistically better.25

One could, of course, in natural science, present a hypothesis concerning natural history and mathematics, and look at the corre­lation between concepts and natural facts. ln such an investigation, however, we will not find intuitions a priori, platonic numbers or other "conditlons of possibilíty"; rather, we might find, at best, brain cells and evolutionary facts, which will, of course, not satisfy us once more. Wittgenstein is not interested in either philosophical groundings or scientific hypotheses because heis neither doing science nor presenting philosophical theses or explanatlons. ln accordance with his genetic method, he is only looking at what generates philosoplíical questions. Heis also pointing out that the use of concepts in our lives may show us more about the binding character of mathematics than transcen­dental assumptions (be it a realm of weird objects or the structure of the brain, or the mind). We would, l suppose, find out that we always get 6 and 6 apples when we have 12 apples. Then we would like to go beyond the fact and look at their "condltions of possibility," which are supposedly neither physiological, nor sociological, nor biological. ln philosophy, then, we believe to be doing something similar to science, but not exactly the sarne. And here is the problem: what are we doing when we have sentences that look like hypotheses, but don't work like hypotheses? Something like a hypothesis in science, but without the empirical commitments of science, without experimentation, is some­thing that we should be suspicious about. lt is perhaps analogous to the case of someone believing that there is an E sharp in the musical scale, and not analogous to the case of someone believing that a C major chord sounds similar to a C minor chord.

Wittgenstein doses his critique of the T with some final observations about the general form of propositions in §§134-6. This is the most natural way to finish his remarks about the T, since the general form of propositions was its zenith, the last piece of his conceptual notation and the best expression of the crystal pure view of logic. In these sections he gives us the new humble status of the general form of propositions. He begins his remarks ín the following way:

Let us examine the proposition: "This is how things are." How can I say that this is the general form of propositions? (PI, §134)

"This is how things are" is the first formulation of the general form of propositions in T 4.5. This is not, of course, the last and clearest

The Philosophical Investigations 259

formulation of the general form, namely: "[p, t N(Ç, )]" (T 6). "This is how things are" does not give us any clue about how all propositions can be constructed out of elementary propositions. It is not a rule. lt seems to be, however, an ordinary formulation of the general form and it indicates one important aspect of the general form of propositions: it works like a variable (T 4.53). lt seems that in ordinary language itself we could find the clue for a general form, for in some cases we employ "This is how things are" or "Such and such is the case" as a variable. But then, it "would also be possible ... simply to use a variable, as in symbolic logic" (PI §134). However, "no one is going to call the letter 11p 11 the general form of propositions" (PI §134). "P" does not give us, of course, any clue abôut what is essential to a proposition, what unifies proposi­tions with various forms, and what could be the totality of propositions (language) - see Chapter 1, Section 1.2. An important characteristic of "This is how things are" or "Such and such is the case" is that we do not compare them with reality. It is not, therefore, itself a proposition in the terms of the T. To say that "Such and such is the case" agrees or not with reality would be nonsense. This does not show, however, that it is not a sentence of our language. ln fact, it is. And when used in contexts such as "John visited me yesterday, said that such and such was the case and that therefore he was looking for a new job" it refers to things said (propositions) by John yesterday.

"Such and such is the case" has this "one feature of our concept of a proposition (Satz)," namely, "sounding like one" (PI, § 134). Should we be teaching someone what a proposition (Satz) is, we would give exam­ples of what grammatically sounds like a proposition (for instance, a sequence of proper name, verb and noun). ln one sense what a proposi­tion is "determined by the rules of sentence formation (in English for example)" (PI §136). Being true or false, thus, is what characterizes a Satz in logic, a descriptive sentence, and not a Satz in general.

If we are asked "What is a proposition?" we shall give examples of very common, say, English propositions. We may, however, "include what one may call inductively defined series of propositions" (PI §135), i.e. we may include propositions that we can generate with the rule expressed in the "logical form of propositions." This could be useful, for instance, for showing how we can always construct more complex propositions in a series. This would be similar to an explanation of numbers: we could give examples, count objects, but it would also be helpful sometimes to presenta defined series like [O, x, x+l] (PI, §135; see T 6.03). Such a series indicates that in our conception of numbers there is no end in the series, as the rule shows. The fact that the construction

260 Wittgenstein's Philosophical Development

of propositions is endless is important in the T, for it indicates that a general form must be a rule, and not itself a proposition (see Chapter 3, Section 3.2.3).

ln §136 Wittgenstein evaluates the core of the general form of prop­ositions (Saetze), namely, the bipolarity of all propositions: "whatever can be true or false" (my emphasis). lf a proposition is whatever can be true or false, and all necessity is logical necessity (T 6.375), i.e., tautological necessity; all propositions that seem to be necessary and non-tautological are nonsense. This is the fate of metaphysics and of the "propositions" of the T, for they obviously fail to be either tautologi­cally necessary or bipolar (if they are taken to 11describe11 the essence of world). 1f a proposition can be true, it can also be false; if it is true, its negation is false, and vice-versa. The possibílity of negation, then, seems to be given in all propositions. Negation is somehow 'there' where there is complexity (argument and function), where something is said about something (see T 5.47). The peculiar understanding of negatíon of the T, as joint negation andas isolated negation, is the clue for the construc­tion of all propositions. lt is, therefore, fundamental that negation is gíven in the bipolarity of each proposition. One could say that, for the T, negation is given in the essence, in the very meaning, of a proposition. This should mean that the rules of negation are implicit in any proposi­tion and those rules are made explicit by the notation of the T, which perspicuously expresses (or shows) them.

That the bipolarity is such a fundamental characteristic of proposi­tions is seen, in the PI, as amounting simply to saying that a proposition is to what the calculus of truth-functions is applied:

And to say that a proposition is whatever can be true or false amounts to saying: we call something a proposition when in our language we apply the calculus of truth functions to it. (PI §136)

We ca/l something that can be true or false a proposition. We call it something to which truth functions apply in the sarne sense that we call the king of chess that which can be checked. ln other words, calling something that can be true or false a proposition (Satz) is part of the use of 'proposition' in logic. However, this is not the essence of anything, for we also call something a proposítion, a Satz, when it is in accord with "the rules of sentence formation" in English (or German, etc.). ln another sense, yet, what a proposition (sentence) is determined "by the use of the sign in the language-game" (PI, §136). Thus, the T presents one sense of proposition, for "in our language" one of the uses

The Philosophical Investlgations 261

of the word 'proposition' is precisely the use made in logic, and not the essence of anything that we call 'proposition' or the essence of language. ln the language-game of logic, or in the use of language as a calculus, the words 'true' and 'false' have an important role related to proposi­tions. We could say, then, that these words are "constituent parts of this game" and, in this case, we also say that these words belong to our concept of proposition. We had better, however, not say that 'true' and 'false' flt our concept of proposition. This would be a "bad picture" (PI, §136). For then, it may appear that 'true' and 'false' don't have a very central role in logic:

To say tlrat check did not flt our concepts of pawns would mean that a game in which pawns were checked, in which, say, the players who lost their pawns lost the game, would be uninteresting or stupid or too complicated to something of the kind. (Pl, §136)

The logical calculus would be "uninteresting" or "stupid11 without the notions of true and false because it would lose its purpose, for the pres­ervation of truth among propositions that follow from each other is the major goal that we usually have in logic. Moreover, if we say that some­thing fits something else, we assume that we have two different thíngs that can engage with each other (like cogwheels). It is as though we had "a concept of true and false, which we could use to determine what is and what is nota proposition11 (Pl, §136); as ifwe would go on applying the concepts to cars, words, angels, and propositions and then come to see that they only connect to propositions.

The point of the critique, then, is that we take a proposition to be what is true or false when we want to apply the "calculus of truth func­tions." This is an important role, of course, since the purpose of logic is to show how truth is preserved in inferences. But this is only one role of 'proposition' (Satz) among others. It is not an expression of the essence of language. ln fact, we can well imagine languages in which such a role does not even exist, for instance the language of PI §2 and its extension in PI §8.

5.4 An old remark about grammar in a new context

I hope to have shown how the PI has its center of gravity in the genetic method associated with the anthropological view. So 'grammar/ as in the BrB, means merely our use of words in their surroundings or the description of such a use compared to other uses. Nothing of the 'old

262 Wíttgenstein's Phílosophical Development

grammar' is in place in the PJ. The idea of a comprehensive discipline of the rules of language that give the limits of sense was already aban­doned around 1934. Someone might think, however, that passages of the PI appear to indicate that the old idea of 'grammar' is still playing an important role there. l will look at the most striking passage in arder to show that this appearance is deceiving. This will also make clear, l hope, the role of Wittgenstein's old jargon in the Pl.

After havíng written approximately the first 188 sections of PI (MS 142 and FF), Wittgenstein thinks about selecting old remarks for his work (MS 119, 84). He thinks about, selecting remarks that are still "enjoyable":

My difficulty now is to know which selection or quantity of my remarks is still enjoyable (geniessbar). For what is not enjoyable is also not useful. But my judgment fluctuates and 1 don't know where the Umits are to be drawn. (MS 119, 65; from 10.1937)

The old remarks certainly included , remarks from the BT, besides those concerning method that he already used in MS 142 (see Chapter 4, Section 4.4). He considered the old remarks that he did not use in MS 142 "slack," but also the product of much thinking (MS 119, 79rv). This should not lead one into thínking that reading the PI as if it were conveying the sarne views of the BT is correct. This would be like reading Wittgenstein's philosophy through Waismann's glasses (see Chapter 4, Section 4.5), i.e., without taking into account many important changes that took place in Wittgenstein's philosophy after the BT. When reading passages of the BT in the PI, we must take into account their possible "slackness," the thoughts behind them and, more important, the new context in which they appear.

ln what follows l will deal with what is, perhaps, the most puzzling of those remarks (§372). lt is puzzling because it might seem to sustain the notion of 'grammar' as the comprehensive discipline of the mies of sense. This seems to be suggested also by the remarks surrounding it. I quote the remarks:

371. Essence is expressed in grammar. 372. Consider: "The only correlate in language to an intrinsic necessíty

(Natumotwendígkeit) is an arbitrary rule. It is the only thing which one can milk out of this intrinsic necessity into a proposition."

373. Grammar tells what kind of object anything is. (Theology as grammar).

The Philosophical Investigations 263

One may think that 'grammar' is taking over an important task that was misunderstood in the past. This important task would be the expla­nation of the nature of necessity: 'grammar' would be the ground for a linguistic conception of necessary propositions.26 However, we should not read such a strong view of grammar into the discussed sections of the PI. First of all, §§371-3 are less significant than one might think they are. They are simply elaborating a point of §370, namely the nature of Wittgenstein's interest in words. §370 occurs in the discussion of imagination and opens with the following remarks:

One ought to ask, not what images (Vorstellungen) are or what goes on wherí one imagines something, but how the word "imagination" (Vorstellung) is used. But that does not mean that I want to talk only about words. For the question of what imagination essentially is is as much about the word {/imagination" as my question.

So Wittgenstein is pointing to the significance of looking at the use of words when in the grip of a philosophical puzzlement. Is, then, Wittgenstein only interested in words? Words in contrast to what? The things-themselves? Wittgenstein's point is that even the fírst question ("What are images?") is asking flfor a word to be explained" (PI, §370), namely, 'image.' lf we don't pay attention to the use of words, however, we may "expect a wrong kind of answer" (PI, §370) to a question such as "What are images?" The {/wrong kind of answer" that Wittgenstein has in mind in this section was already discussed in the BB, BrB, and in the preceding sections of the PI: {/this ques­tion is not to be declded - neither for the person who does the imag­ining, nor for anyone else - by pointing; nor yet by a description of a process" (PI, §370). At this point of the PI, Wittgenstein has already made clear that introspection and a mind-model are not really good alternatives for answering the question, simply because they don't explain at all.

According to Wittgenstein, looking at the uses of words in their surroundings is a good start to find out what something is, for in the explanatíons of the use of words we gíve such informatíon. This is why he says that "Essence is expressed in grammar" in §371, i.e., we may find what we are looking for when we ask what a thing ttreally is" when we look at how we use words (grammar). Aren't we askíng for the meaning of a word when we ask what an 'image' is? If so, we had better look at what we do with words, which does not mean that we "talk only about words" (PI, §370).

264 Wittgenstein's Phílosophical Development

Once §371 is reasonably deflated, we can better see the point of Pl §373. On §373 Wittgenstein is emphasizing, again, that we should pay attention to the use of words (grammar) in our form of life (the broad notion of use) when in philosophy we ask for the nature of something. This is the case, according to Wittgenstein, even in issues of theology. The reference to theology in §373 may be interpreted as an example of the point from §§370-1: even in questions like "what is God?" or "how does God communicate" one should look at the way words like "God" and "communication with God" are used. This, of course, stops a, say, medieval investigation of the 'idea of God.' One of the uses of 'God,' according to Wittgenstein, is such that "You can't hear God speak to someone else, you can hear him only if you are being áddressed" (MS 130, 7, from 08.1946; also Z § 717). Wittgenstein calls this sentence "a grammatical remark" in MS 130 (Z §717) and this means that it is a remark about the uses (grammar) of the word 'God.' The various uses of 'God' (the various grammars or grammar of the word) will indicate what people (or a people) mean by it. The description of the uses of 'God' also involve making connections, finding links, and, above all, looking at what people do, how they act in the surroundings and circum­stances in which the word is used. The word is, thus, anchored in the practices that surround it, in its purpose, point, etc., and has nothing to do with a rule of an autonomous calculus. Thus, what Wittgenstein says in §373 concerning grammar actually goes against the idea that 'grammar' is a new way of expressing the nature of necessity. If not so, we would have to interpret Wittgenstein as assuming a comprehensive discipline of necessary propositions (or rules) that íncludes religion, which would be at the side of mathematics and logic. This seems to be preposterous.

lt could be argued, however, that the reading offered here is not in accord with §372, where Wittgenstein supposedly defends the arbítrari­ness of grammar and a de dieta conception of necessity. 27 ln this passage, one might think, Wittgenstein expresses an opposition between two conceptions concerning "necessity," supposedly the conception of the T and the conception of the PI: that necessity is given in reality and is something real, 'intrinsic necessity,' or that necessity is something determined by arbitrary rules of grammar.

The point of §372, in my view, is quite different. First, note that Wittgenstein is not endorsing anything in PI §372. The conception presented there is quoted. Wittgenstein asks us to consider the remark that he quotes. lt would be wrong, thus, to attribute the conception to him in the PI. Second, heis quotíng himself (BT, 235). From this, one could well think that he is quoting himself because he does not hold an old

The Phílosophical Investigations 265

view anymore. Precisely this is the case. Third, in the quoted passage he is discussing two notions: "the correlate in language" to "intrinsic [objective or absolute] necessity." He asks us to consider that the correlate is an arbitrary rule. Before we understand this, however, we have to get clear about "intrinsic necessity."

Wittgenstein certainly did not defend the view that there are uintrinsic necessities" in the sense of de re necessity given in simple objects in the T. There he clearly states that "the only necessity that exists is logical necessity" (T 6.37), i.e. the necessity expressed in tautologies of logic. Such a claim (T 6.37) would be quite absurd if Wittgenstein was defending the claim that properties of simple objects determine the logic of language. For in this case, there would be another kind of logical necessity, i.e., logical necessity that is not tautologous. This would undermine the whole point of the T. As we have seen in Chapter 1, Wittgenstein introduces an investigation of forms that are not logical forms (which would be the properties of simples) precisely because the Tractarian account was incorrect: forms of space, color, and time could not be redticed to logical form as understood in the T.

ln §372 Wittgenstein asks us to consider the opinion that "the only correlate in language to an intrinsic [objective or absolute] necessity is an arbitrary rule." He is contrasting, therefore, the opinion presented there with another, according to which the "correlate in language" is not an arbitrary rule. What are then the two views that are under consideration in §372? ln fact, we have already dlscussed both in Chapter 3, Section 3.2.2. The origin of the remark is MS 110, 114-5 (from around 02.20.1931) and it is copied to TS 211, then cut into TS 212 and then it gets its place in the BT, 235. ln the clean copy of the BT, Wittgenstein does not quote the remark because there he defends it. The original context shows that the point was to oppose the view that the possibility of a double negation to produce a positive proposition is already contained somehow in the nature of the negation sign, a primi­tive sign of logic. Here is the context:

"lt lies already in the act of the operation/ of negation/ that, doubled, it cancels itself." That what already 'lies there,' one can always express by means of a rule; for we cannot express it if it lies there, but only detached.

Therefore, in 11--p = p" 1-

1 is nota negation.

The only correlate in language to an intrinsic necessity (Natumotwendigkeit) is an arbitrary rule. It is the only thing which

266 Wittgenstein's Phi/osophical Development

one can milk out of this intrínsic necessity into a proposition. (MS 110, 114-15)

What Wittgenstein criticizes in MS 110, and in the BT, by means of the calculus conception, is the understanding of logic of the T and the weight it puts on its logical notation, which supposedly reveals the hidden logical properties of language:

This contains the decisive point. We have said that some things are arbitrary in the symbols that we use and that some things are not. ln logic it is only the latter that express: but that means that logic is not a field in which we express what we wish with the help of signs, but rather one in which the nature of the absolutely necessary sign (die Natur der natumotwendige Zeichen) speaks for itself. lf we know the logical syntax of any sign-language, then we have already been given all the propositions of logic. (T 6.124)

Thus, Natumotwendigkeit (intrinsic, objective, or absolute necessity) in the BT and in Pl §372 is a reference to T 6.124, and not to simple objects and de re necessity. The target is the idea that "the nature of the absolutely necessary sign (die Natur der natumotwendige Zeichen) speaks for itself" in the correlate to the objective necessity in the T: suppos­edly, the logical notation. ln the T, the "one and only general primitive sign in logic" (T 5.472) is intrinsic to the very essence of propositions (5.471). According to the T, intrinsic in the essence of propositions, "in their very nature," we have already present argument and function; thus also already present "all logical constants" (T 5.47). What is given in the essence of propositions is not expressed with arbitrarily determined signs (T 6.124; 3.315). What is given in the essence or nature of proposi­tions is a primitive sign of logic (fundamental concept). Primitive signs are not defined in the T, but presented by means of the variables and the primitive logical connective of the notation: the general form of propositions, argument and function, truth-tables and the operation N. Those primitives in logic, of course, show themselves and they show the nature of necessity in tautologies. They are, however, not optional, for any language must have them (T 5.47, 6.124) see Chapter 3, Section 3.2. Thus, the target of the quoted critique in Pl, §372 is the idea that a notation might show necessary signs. Such a notation is, thus, a correlate in language that according to the Tis not merely an arbitrary rule.28

The negation, one could think, would present the "property of resulting in an affirmation when doubled" (MS 110, 102; BT, 161). ln

The Philosophical Investigations 267

fact, in the T Wittgenstein says that "an affirmation can be produced by a double negation" (T 5.44). This, however, suggests that "the mies from the negation sign follow from the nature of negation" (BT, 164). With this view concerning logic in the T, in PI §372, Wittgenstein is contrasting the calculus conception of language, according to which there is no primitive or indefinable given from which the mies unfold. ln the calculus conception '--p=p' "is merely a rule of the game" (MS 110, 107; my emphasis). That is, negation is not the fundamental opera­tion in Jogic; it works like any rule of the calculus or game of language. According to the calculus conception, rules of 'grammar' constitute the meaning ofwords and the sense of propositions. There is no such thing as a fundamental concept (Grundbegriff) or a primitive sign (Urzeichen) (see T 5.451 and 5.46, respectively). It is, thus, the idea of an intrinsic necessity of indefinables- of negation, for instance - expressed in a nota­tion that is opposed to the idea of arbitrary rules of the BT. According to the BT, we could have the rule '--p -p' instead of '--p=p/ as some uses in ordinary language seem to justify ("I don't know nothing about the last events"). So in the calculus conception, the rules explain the meaning, and not the other way around. The rule that explains the meaning is not given in the essence of proposition or negation, but is an arbitrary or autonomous rule.

Note, however, that this is the view of the BT (Chapter 3). The point, thus, of the quoted remark is to contrast the, perhaps, less misleading view of the BT with the more misleading view of the T (less misleading because the view of the T might tempt us, for instance, to think that a notation expresses the essence of any language). The view criticized by the "arbitrary rules" view is, thus, indeed from the T; but it has nothing to do with what one might call de re necessity. It is related, of course, to the way that Wittgenstein understood necessity and the logical primi­tives in the T, i.e. as completely shown in a perspicuous notation. 1 suppose that nobody would attribute a de re conception of necessity to the tautological logic of the T.

Thus in §372 Wittgenstein is asking the reader to consider an old idea that he quotes. He is not subscribing to the view. The contrast of ideas is supposed to break the spell of the view of the T, which might be more misleading than the calculus conception of the BT:

We change the aspect by putting a system of the expression along­side of the other. ln this way the spell in which an analogy hold us can be broken, if we put another at its side that is recognized as having the sarne rights ... " (TS 220, FF, §106 and BFF §140)

268 Wittgenstein's Philosophica/ Development

'The arbitrariness of rules' (of 'grammar,' if you wish) is a picture that opposes a view concerning the a priori and the power of a notation in the T. Both pictures, however, "have the sarne rights." They are at the sarne levei, they are "equally arbitrary":

1 may occasionally produce new interpretations, not in order to suggest they are right, but in order to show that the old interpreta­tion and the new are equally arbitrary ... 1 will only make gas to expel old gas. (LFM, 14; my emphasis)

Gas (new interpretation) that expels old gas (old ínterpretation) is maybe a less confusing pícture. lt can also cast shadows;' however, even though it partially illuminates the matter by expelling the old picture. Like any picture, i.e., like any phílosophical rough idea, comparison, analogy, it might be misleading. The view of the BT, quoted in PI §372, might mislead, cast shadows. ln the BT, as seen in Chapter 3, Wittgenstein overemphasizes the role of "arbitrary" rules of language and does not take into account the purpose and point of the game where rules appear. Perhaps, he does so because he thinks there that "if anyone utters a sentence and means or understands it, heis thereby operating a calculus according to definite rules" (PI, §81). The gas "arbitrariness of grammar" or the gas "X is a rule of grammar" might mislead because they are jargon that can be taken as important insights, for instance, into the nature of necessity (perhaps, as important nonsense that suppos­edly grounds a de dieta conception of necessity). Note that Wittgenstein in his Lectures goes as far as denying not only the idea that he defends philosophical theses (PI §128), but any opinion whatsoever:

One of the greatest difficulties 1 find in explaining what 1 mean is this: You are inclined to put our difference in one way, as a differ­ence of opinion. But 1 am not trying to persuade you to change your opinion. 1 am only trying to recommend a certain sort of investiga­tion. If there is an opinion involved, my only opinion is that this sort of investigation is immensely important, and very much against the grain of some of you. lf in these lectures I express any other opinions, I am making a fool of myself. (LFM, 103)

Wittgenstein is also, in his own view, making a fool of himself if he expresses opinions concerning 'grammar' and its arbitrariness. This is reason enough for him to quote the remark in PI §372. It indicates a

The Philosophical Investigations 269

heuristic device in use, which is only used to expel a certain picture that at some point Wittgenstein himself held. ln another lecture from 1939, Lewy implied that Wittgenstein wanted him to believe a partic­ular view concerning the jargon 'grammar' and claimed in class: 111 know what you want me to say.11 Wittgenstein saw this as severe criticism (LFM, 55). His answer is this:

... 1 have no right to want you to say anything except just one thing: 11Let1s see" -One cannot make a general formulation and say that I have the right to want to make you say that. For what could that general formulation be? My opinion? But obviously the whole point is that I mdst not have an opinion ... For instance, I have no right to want to make you say that mathematical propositions are rules of grammar. (LFM, 55; my emphasis)

Note that even 11mathematical propositions are rules of grammar" is not a conception Wittgenstein wants to defend. It does not express Wittgenstein's conception concerning the nature of mathematics or necessity. His 11whole point" is really that he "must not have an opinion."

Wittgenstein's philosophy of mathematics seems the place for us to find the old idea of 'grammar,' since, apparently, "mathematical propo­sitions are rules of grammar" is an important insight. It is, however, only a good picture to expel the bad picture of Fregean Platonism (for instance). The old 'grammar' is not a central conception even in Wittgenstein's later philosophy of mathematics. TS 222, Part 1 of RFM, is Wittgenstein's most thought-through text on the subject. It was intended to be part li of PI until, probably, 1943. There, the word 'grammar' appears only twice. lt indicates, first, the superficial simi­larity of the use of words. Note, how the word is qualified: "[ ... ] this appearance is merely produced by the superficial form of our grammar (as it might be called)" (TS 222, 81; RFM 1, §108; my emphasis). The phrase where 'grammar' appears could be replaced by "superficial form of our use of words"; we, however, 11might call it grammar." A second modest appearance of 'grammar' is this:

The connection which is not supposed to be a causal, experiential one, but much stricter and harder, so rigid even, that the one thing somehow already is the other, is always a connection in grammar. (TS 222, 103; RFM 1, §128)

270 Wittgenstein's Philosophical Development

A connection in grammar is a connection in the use of expressions in language and its surroundings. The surroundings are usually blurred by the prose, and not by the calculus, of logicians and mathematicians.29

It is what they say about what they ·do or prove that needs scrutiny. The "hardness of the logical must" (RFM l, §121), the "inexorability of logic" - suggested by Fregean prose, for instance - are expressions that, in fact, indicate the significance of logical or mathematical compulsion for us. We say things like "lf you accept this and this, you have to accept so and so." The function of those expressions is better understood if we are clear about "the role of thinking and inferring in our life" (RFM l, §116). A connection in grammar is, thus, a connection in the way we use ('use' broadly understood) mathematical and logical expressions in thinking and inferring. The line between thinking and not thinking is not "hard and fast" (RFM l, §116). This seems to be the point of Wittgenstein's weird tribes (RFM l, §§147-53). However, in our form of life, "if you draw different conclusíons, you get into conflict, e.g. with society; but also with other practical consequences" (RFM 1, §116). A "stricter and harder" connection (the h.ardness of the logical must) is a connection in the use we make of inferences (logical or mathematical) in our form of life; such uses, of course, are "harder" than others, for they have a central role in our surroundings (mathematical thinking works very well, we mostly agree about it, we train our children in its use, we don't accept wrong results in calculations as examples of 'thinking,' counting and calculations usually agree, etc). But, again, understanding the connections in uses of our form of life is not intended as a new, fragile, foundation for mathematics, logic, and 'grammar.' Its point is to show that looking for foundations is a confused project. We begin looking for foundations because we insist on not looking at some details of obvious facts of life.

One is naturally inclined to think that the old 'grammar' (from PR and BT) is a fundamental notion in Wittgenstein's later philosophy mainly because one thinks he is saying that the 'rules of grammar,' the 'grammatical remarks,' and prose such as "X is a rule of grammar" are the good guys and that 'nonsensical propositions' are the bad guys. Both, however, when they are part of philosophical prose, are idling in language. Note that mathematical equations and formal logical notations, the calculus, are all rlght as they are. It is the added prose that is under scrutiny in the PI (even if mathematicians and scientists add the prose). Note also that Wittgenstein suggests that a grammat­ical remark (or grammatical sentence) such as "This body has exten­sion" is nonsense (PI, §§251-2; compare with BT, 96; also BB, 30). We

The Philosophical Investigations 271

don't call the sentence nonsense simply because we take it as a kind of tautology to which we are inclined to reply "Certainly!" This inclina­tion is misleading, for we are then inclined to think that we have said something deep and important about language (which is misleading) or about the world or the nature of signs (which is, perhaps, even more misleading). Think of Kant's "All bodies are extended" and how much noise this triviality has generated - and how much noise Wittgenstein's own trivialities have. When Wittgenstein suggests that "this body has extension" is nonsense, he is teaching us to "pass from unobvious nonsense to obvious nonsense" (PI, §464). Nonsensical sentences, things that we could exclude from language, are sometimes used transitorily as grammaticál remarks to expel gas; they are not, however, important nonsense. The good guys are thus also bad guys.

What about the bad guys? "The sense datum is private'' is, of course, the best possible example of nonsense for Wittgenstein and Wittgensteinians. But Wittgenstein says the following about this nonsense:

"The sense datum is private" is a rule of grammar, it forbids to use such expressions as "they saw the sarne sense datum, 11 it may (or may not) allow such sentences as "he guessed that the other had a sense datum of this ... kind". It may only allow expressions of the form: "The other looked round, had a sense datum and said ... " You see that this word is such a case [that it] has no use at all. But if you like to use it, do! (MS 151, 39-40; also MS 181, 1; from 1935-6)

Here, a bad guy is a good guy. Nonsense and grammatical remarks or trivial "rules of grammar" have "no use at all.11 We don't do many things with them, except while philosophizing - we like using them in philosophical prose as big truths. One should remember, however, that "when we are tempted in philosophy to count something quite useless [something without "a use in our life"] as a proposition, that is often because we have not reflected sufficiently on its application" (PI, §520). The difference between a trivial rule of grammar (or a grammatical remark that says that something is a rule of grammar) and nonsense is that the former might expel gas in a particular context (in this context, nonsense is the gas expelled). But in a different context, the gram­matical remark might itself tum into gas. At the end, both should be expelled in the application of the genetic method; both should dissolve along with the philosophical puzzlement. 'Rule of grammar,' 'gram­matícal remark,' and 'nonsense' are expressions, part of the prose, that

272 Wittgenstein's Philosophical Development

one should leave behind: one can throw away the ladder, as it were. ln the end, there is no battle between rules of grammar and nonsense, for they have the sarne more or less useless nature. We can take both out of circulation. They are part of a jargon that indicates that philosophical puzzlement has not yet disappeared; Wittgenstein's method has not yet properly worked. If one clings to the Wittgensteinian jargon (which is, indeed, as attractive and misleading as Kantian jargon), then one is still captive of the idea of important nonsense (one is, as it were, a dose cousin of the metaphysician).

ln a pessimistic mood, Wittgenstein sai d at the end of his 1939 lectures: "the seed that Iam most likely to sow is a certain jargon11 (LFM, 293). He was right, unfortunately, for his jargon impresses many of his readers, who emphasize how important the nonsense of grammatical remarks is in Wittgenstein's philosophy. The old vocabulary - his jargon, his prose - might be what Wittgenstein had in mind when he wrote: "lt will be difficult to follow my presentation: it says something new, but something to which eggshells of the old still stick" (MS 129, 181; from 1944). One should not take the eggshells of the old that still stick for the new egg - and this is a grammatical remark.

Notes

Preface

1. The "resolute reading" of the T ("The New Wittgenstein") has been the focus of many scholars for the last 20 years. The reading is consolidated in Diamond's papers published in The Realistic Spirit. Many defenses and attacks followed; for instance: Diamond (2004), Conant (2000, 2006), Conant and Diamond (2004), Ricketts (1996), Goldfarb (1997), Hacker (1999), Floyd (2000), Kremer (2001), Williams (2004), Sulllvan (2004), McGinn (2006), Mulhall (2007), Kuusela (2008).

Introduction

1. A one mild change of mind is usually assumed by 'resolute readers' (see Conant 2007, Kuusela 2008). The one big change view is defended, for instance, in Baker and Hacker (2005) and (2009), Hacker (1997) and (2000), Hintikka and Hintikka (1986), Pears (1987 and 1988), and Jacquette (1998).

2. Difficulties concerning the reconciliation of Wittgenstein's non-theoretical stance and the interpretatíon of his 'grammatical remarks' are illustrated in the discussions that followed the partíng of the ways of Baker and Hacker (see Baker 2004 and Hacker 2007). Similar diffículties are discussed or at least mentioned by many scholars. See, for instance, the introductions of Lugg (2000), Diamond (1996), Mulhall (2007), and Stern (2004). See also Chapter 1 of McGinn (1997), Chapter 4 of Forster (2004); also Kenny (2004), Hanfling (2004), and Fogelin (1996).

1 Phenomenology, 'Grammar,' and the 'Limits of Sense'

1. Before 1923 Ramsey was already familiar with Wittgenstein's work, for he had "assisted C.K. Ogden with the translation" of the T (see publisher's note in Ogden's translation).

2. See Ramsey's letters to his mother and to Ogden in CL, 186-7. 3. ln MS 112 (written at the end of 1931), for instance, Wittgenstein writes

retrospectively about the problem of incompatlbilities: " ... if 'f(x)' says that x is in a certaín place, then 'f(a) · f(b)' is a contradiction. But why do I call 'f(a) · f(b)' a contradiction when 'p · -p' is the form of the contradiction? Does it mean that the signs 'f(a) · f(b)' are not a proposition in the sense that 'ffaa' isn't? Our difficulty Is that we have, nonetheless, the feellng that here there is a sense, even if a degenerate one (Ramsey)."

4. Wittgensteln also had a great interest in the philosophy of mathematics when he returned to Cambridge. It seems that his interest was a mixture of legitimate, general, philosophical worries concerning mathematics and the intrinsic problems of his philosophy (see Secti6ns 1.3 and 1.5.4). He

273

274 Notes

was especially interested in problems related to infinity and tbe "actual infi­nite," wbicb be tbougbt was a fiction. Tbe reason for tbis interest might bave been Brouwer's lecture in Vienna in 1928 (see WVC, 16). Wittgenstein was certainly an "intensionalist" in 1929, for be accepted tbe infinite only as a rule and not as an extension (MS 105, 23-27; MS 106, 36; PR §§144-5), perbaps because be tbougbt tbat tbe extensionalist conception of tbe infi­nite was tbe cause of tbe "paradoxes of tbe infinite" (see MS 106, 36 and 246). ln fact, be claims: "All paradoxes of tbe infinite dissolve if we under­stand tbat only tbe intensional generality bas sense in tbe domain of the infinite series of numbers" (MS 106, 50). For Wittgenstein in 1929-30, kinds of infinities are forms given by rules (intensions), i.e. "potential infinites" or "possible infinites" (MS 106, 1). Tbis view concerns not only mathematics; it is also expressed in bis understanding of pbenomenology. Time and space are given as infinites by rules of constructlon and not as infinite totallties (MS 106, 1; PR§ 143). Given bis broad opposition to ex tensional conceptions of infinites, it is not clear wbetber Wittgenstein's denial of tbe actual infinite is prompted by a deep conviction about infinity in general or prompted by bis opposition to some 'paradoxical' results in matbematics. ln any case, Wittgenstein certainly opposes any extensionalist view in 1929-1930: 11

••• it is impossible to reter to an infinite extension. Infinity is tbe property of a law, not of an extension" (WLC30-32, 13, from 03.10.1930).

5. He asks, for instance: "What is tbe general form of spatial statements?" (MS 105, 11) and "Can we do geometry of coordinates in tbe visual fleld?" (MS 105, 9). He also writes: "It seems one could not see a patcb [as] composed of one color, except if one imagines it as not of one color" (MS 105, 9). ln a letter to Scblick from 02.18.1929 Wittgenstein explicitly writes tbat be decided to stay in Cambridge in order to work on tbe "visual field and otber tbings" (LWGl3).

6. Ramsey's way-out of the problem is quite different from Wittgenstein's. At first sigo, Ramsey in Facts and Propositions (1927) seems to agree witb Wittgenstein: 11

••• justas in tbe study of cbess notbing is gained by discussing tbe atoms of wbicb tbe cbessmen are composed, so in tbe study of logic notbing is gained by entering into tbe ultimate analysis of names and tbe objects tbey signify" (p. 41). Ramsey, bowever, beld tbat sucb a point of view implied that incompatibílities bave an empirical, and not a logical, nature: " ... Leibniz and Wittgenstein bave regarded 'Tbis is botb blue and red' as being self-contradictory, tbe contradiction belng concealed by defec­tive analysis. Wbatever may be tbougbt of tbis bypotbesis, it seems to me tbat formal logic is not concerned witb it [ ... ] No one could say tbat tbe inference from 'Tbis is red' to 'This is not blue' was formally guaranteed like tbe syllogism [ ... ] This assumption migbt perbaps be compared to tbe assumption tbat tbe cbessmen are not so strongly magnetized as to render some positions on tbe board mecbanically impossible, so tbat we need only consider tbe restrictions imposed by rules of the game, and can disregard any others wbicb might conceivably arise from tbe pbysical constitution of tbe men" (p. 48). For Wittgenstein, tbis is not a solution, for incompat­ibilities bave a formal nature; it follows from 'Tbis is red' tbat 'Tbis is not blue.' Any inference must be formally, and not materially, guaranteed (MS 106, 85).

Notes 275

7. Here we bave tbe roots of Wittgenstein's general principie of verifícation: íf propositions bide elementary propositions, a metbod to find out wbat is bidden must be introduced. ln a !ater remark, this point is explicit: "Can a logical product be bidden in a proposition? And if so, bow does one find tbis out, and wbat metbods do we bave of pulling wbat is bidden in tbe proposi­tlon into tbe open? So long as we don't bave any metbods (for finding it), tben we can't speak of sometbing being bidden or possibly bidden." (MS 111, 12 from 07 .1931; BT, 100).

8. Wittgenstein calls SRLF "sbort" and "weak" in a letter to Mind from May 1933 (tbe subject of tbe letter is Braitbwaite's unautborized use of bis ideas). l use SRLF to point out wbat Wittgenstein's pbilosopby looked like before be gave up tbe pbenomenological language. Tbe sarne or very similar points are also made in tbe manuscripts of tbe period. ln severa! places l indicate tbe manuscripts wbere tbese ideas originate and assumptions made in SRLF tbat are explained in the manuscripts.

9. ln wbat follows I use Wittgensteín's own notation. Since be uses 'E(b)' to express tbat an object E bas a certain degree b of brigbtness, lt seems tbat be is inverting function and argument. Tbis inversion expresses, actually, as already indicated above, an important insigbt concerning logical analysis: " ... I see a red patcb in front of me. Am I now to say tbat red is a property of tbat patcb? What is an object and wbat is a property bere? Tbat question is completely pointless. Tbe trutb is tbat our traditional linguistic forms (noun, adjective, etc.) completely lose tbeir significance as soon as we apply tbem to actual pbenomena. A state of affairs is a combination of elements. But tbere is notbing in tbat combination wbicb indicates tbat tbere is tbing-like or property-like in it" (WVC (Theses), 251).

10. A similar point is made retrospectively in BT, 101 (MS 112, 265-6); and BT, 473-77 (MS 112, 251-255).

11. Note tbat tbe arguments against Tractarian analysis concerning calor mixture and opposition are not presented in SRLF. Tbis migbt partially explain wby tbe paper is 'weak.'

12. It is very likely tbat Wittgenstein bad tbis conclusion in mind already wben be arrived in Cambridge, for bis first written notebook (MS 105) introduces tbe draft of a pbenomenological notation before díscussing tbe problems of colar incompatibility related to tbe notation of tbe T. Tbe first explicit discussion of tbe notation of tbe T takes place ln MS 106, 71-120. It is very likely tbat Wittgenstein was tbinking about writing SRLF wben be introduced tbose issues in bis MSS, for in SRLF be explains (even tbougb in an incom­plete way) wby be decided to begin a pbenomenological investigation.

13. Functions in tbe Tare propositional functions (see Hylton 1997). 14. See also MS 106: 55, 69, 77. 15. For tbis reason, Wittgensteín's non-bypotbetical enterpríse is different from

Russell's, wbo bas only the tbings in tbe world ln mínd wben be lntends to describe wbat is known "witbout any element of bypotbesis" (Our Knowledge of the Extemal World, 84). Ris "inventory of tbe logical forms of tbe facts" (61) won't find forms different from functions and relations.

16. Tbe meeting witb tbe Vienna Circle took place in December 1931, i.e. !ater tban tbe period tbat we are dealing witb bere. lt is clear, nonetbeless, tbat Wittgenstein is referring back to tbe subject of our discussion. Camap's

276 Notes

fundamental 2-term relation is 'memory of similarity', which expresses that "between x and y [memory representations] there is a memory of simlladty" (Aufbau §78). Wittgenstein also makes a similar remark, this time referring to Russell, in one of his classes. Moor~ says: "He [Wittgenstein] said that in the Tractatus he had objected to Russell's assumption that there certainly were atomic propositions which asserted two-termed relations - that he had refused to prophesy as to what would be the result of an analysis, if one were made, and that it might tum out that no atomic proposition asserted less than, e.g., a four-termed relation, so that we could not even talk of a two­termed relation" (M, 290).

17. The intrlnsic complexity of propositions, according to the T, is a complexity of argument and function. The explanaition given of quantification is another place to look for the assumption that argument and function express the form of elementary propositlons in the.T (see T 5.51-4). ·

18. The criticism of ordinary language's predicate/subject form, as argued in SRLF, also applies to the function/argument distinction, for Wittgenstein says: "you can of course treat the subject-predicate form (or, what comes to the sarne thing, the argument-functlon form) as a norm of representation ... " (MS 107, 13; TS 208, 68; PR §115; my emphasis). See also: "Concept and object, this is subject predicate" (MS 106, 112). These two modes of representation are "the sarne" in the sense that both are taken for granted before the inspec­tion of phenomena (and so are not seen as what they are: mere hypotheses). Thus, Wittgenstein only points to one relevant aspect under which they should be considered !dentical.

19. The requirement of the right multiplicity makes sure that a representation of language (a notation or a theory) does not have one of the following faults: that it Is not able to express ali the facts that concern it (ln the case of a theory), or that it employs more expressions than needed in order to express what lt is aiming at (thls would be a wrong begriffsschrift). ln the second case, we have to find a more perspicuous notation and apply Occam's razor to the useless signs (T 3.328). This is precisely what Wittgensteln does with the sign of identity in the T. ln the first case, we can be sure that our symbolism needs to be extended. If we wanted, for instance, to express generality "by writing 'Gen. fx' it would not be adequate: we should not know what was being generallzed" (T 4.0411). Those requirements Wittgenstein takes from Hertz's Principies of Mechauics, as he makes clear ln T 4.04.

20. See MS 212, 251-55 for a retrospectiveview on the matter. Austin (1980), Noe (1994), and Marion (1998), have correctly pointed out that the phenomeno­logical language was intended as a conceptual notation that should express the "limits of language and thought" for Wittgenstein at the time. Rhees (1963), however, was the first one who pointed this out.

21. The abandonment of this notation (language), however, will bring him to the idea that the rules, as in 2 above, are the only rules that we need (even though we may occasionally employ notational devices).

22. Weiner (2001) and Floyd (2001) have argued that Wittgenstein did not present a begriffsschrift ln the T. lt is certainly correct that he did not present a complete one. But why should he, if Frege and Russell had already done it? Wittgenstein was correcting their mistakes because he was otherwise accepting their notations (see T 3.325).

Notes 277

23. There is a strictly logical reason for the assumption of logically independent elementary propositions. We could call it a "bottom-up reason." Frege eluci­dated, roughly, the logical connectives by means of truth-conditions (for instance, the conditional in Begriffsschri~ §5). Wittgenstein presented the explanation perspicuously in the T by means of truth-tables. This presenta­tion, however, shows that the atomic propositions that replace the propo­sitional variables must be logically independent; otherwise the truth-table, and Frege's elucidation, does not work properly. lf we take, for instance, the propositions "A is now in London" and "A is now in Paris," the first line of the truth-table does not work (it cannot be the case that both are true). Thus, the elucidation given by Frege seems to require logically independent elementary propositions. There are also reasons related to Russellian anal­ysis that led Wittgenstein to logically independent elementary propositions (we could call those "top-down reasons"). I discuss the top-down reasons in Chapter 2, Section 2.2.2. Those reasons are related to what is discussed ln the following pages.

24. The distinction between the two leveis is discussed as late as the BT (See BT, 113-16).

25. Contrary to what has been claimed by interpreters (see Pears 1987; Hacker 1997; Stern 1997). The major target of Wittgenstein's !ater criticism of solip­sism and "private language" is not Wittgenstein himself but Russell, who severa! times speaks of the privacy of sense-data (see Russell: Our Knowledge of the Externai World, 101, 104, 109; The Philosophy of Logical Atomism, 59; An Outline of Phi/osophy, 14). Russell also writes about a "private world": "No place in the private world of one observer is identical with a place in the priva te world of another observer" (The Relation of Sense-Data to Physics, 120; also: 124-5). Wittgenstein, in fact, employs the expression "solipsism of the present moment" to characterize Russell's "hypothesis that the world was created five minutes ago" (WLC32-35, 25). He also refers to Russell's "solip­sism of the present moment" in a Notebook (MS 156a, 1 lr), where he says that "it means nothing" (also MS 219, 19; MS from 1932-3). Russell seriously entertained the idea of the sollpsism of the present moment, for instance, in An Outline of Philosophy (from 1927): " ... the world might have sprung iota being five minutes ago, exactly as it then was, full of acts of remembering which were entirely misleading" (p. 5). The expression "prlvate language", as far as l am aware of, was first used by Ogden & Richards (MoM, 210); however, Russell was the one seriously involved with it: "ali the names that it [a logically perfect language] would use would be private to that speaker and could not enter iota the language of another speaker" (The Philosophy of Logical Atomism, 59). A variation of Russell's privacy idea is found ln Schlick: "[ ... ] the undeníable fact that what Is contentful and qualltative in our expe­riences must remaín forever priva te ... " (Erleben, Erkennen, Metaphysik, 38). Moore in bis 1910-11 lectures takes it as a common view that "the sense­given fteld of vision of each of us, at any moment, constltutes a private space of that person's own" (Some Maín Problems of Philosophy, 42).

26. Sometimes ít is important to be aware of the chronological arder of the manu­scripts, for otherwise one may not be able to correctly follow Wittgenstein's changes. Thecorrect arder is the followlng: MS 105 (2,4, 1,3 ... 59, 60, 62 ... 106, 109,111, 131); then comes MS 106 (1,3 ... 111; 112, 114 ... 296; 6, 8 ... 110;

278 Notes

113, 115 ... 297, 298); then comes MS 105, again: (6,8 ... 58; 61,63 ... 105; 108, 110 ... 132, 133, 134). The second part of MS 105, then, immediately precedes MS 107. (Only MS 107 is dated and the first date is 10.06.1929, on page 153). This sequence is the result of Wíttgenstein using first the recto and then the verso pages. The Wiener Ausgabe presents the correct order of the manuscrlpts.

27. The passages where Wittgenstein criticizes the "privacy" of sense data in 1929-30 are not críticísms of bis own views. One of the passages is the following: "ln the sense of the phrase 'sense data' ín which it is inconceiv­able that someone else should have them, it cannot, for this very reason, be said that someone else does not have them. And by the sarne token, lt's senseless to say that 1, as opposed to someone else, have them" (MS 107, 216; translated in PR§ 61).

28. This may be why "what the solipsist means is quite correctí' (T 5.62), i.e., if he means that the world is seen perspectively. lndeed, solipsism and idealism point to an obvious fact that we usually are not aware of: "We do not notice that we see space perspectively" (MS 108, 48; PR §47). However, "the world is seen perspectively" does not mean "only my perspective is the world." Subjects are not in the center of the world, for if they were, the world would not be seen perspectlvely (if this were the case, there wouldn't be "truth" in solipsism). Perspectives, according to the T, are the Iimit of the world (and not its center). My world is not a world that belongs to me, but the world seen perspectívely; my language is not a private language, but the language spoken by many subjects, always grounded on the a priori logic. Thus, my does not exclude others (my and your language is our language). This is what the solipsist fails to see. Since "logic pervades the world 11 (T 5.61) and language, there are no /ogical grounds for solipsism. What the solipsist does, then, is the following: he expresses the right feellng concerning the perspec­tivity of the world in the wrong way (for instance: 11 only my perspective is real"). If this is correct, Wittgenstein was never a solipsist and the following remark in Wittgenstein's notes for lectures from 1935-6 could be seen as ao explanation of ao early insight: '"But aren't you neglecting something- the experience or whatever you might call it-? Almost theworld behind the mere words?' But here solipsism teaches us a lesson; it is that thought which is on the way to destroy this error. For if the world is idea it isn't any person's idea. (Solipsism stops short of saying this and says that it is my idea)" (MS 149, 60; Wittgenstein's emphasis). Already according to the T, 1 suggest here, solipsism "stops short" of saying that: " ... solipsism, when its implications are followed out strictly, coincides with pure reallsm. The self of solípsism shrinks to a point without extension, and there remains the reality co-ordi­nated with it 11 (T 5.64; my emphasis). Solipsism is "on the way11 to destroy the idea of a private realm because it can do so when "its implications are followed out strictly." ln this respect, it is a step ahead of idealism because it is closer to bringing us back to the idea of a unique world: "This is the way I have traveled: Idealism singles men out from the world as unique, solip­sism singles me alone out, and at last 1 see that l too belong with the rest of the world, and so on the one side nothing is left over, and on the other si de, as unique, the world. ln this way ideallsm leads to realism if it is strictly thought out" (NB, 85; 10.1916).

Notes 279

29. Th!s clearly distances Wittgenstein from Carnap's "methodological solip­sism" in the Der Logische Aufbau der Welt (§ 64), which seems to be intended as the basis for the passage from the "private" data to the objective world. The construction of the constitutional system sbould show how to get to an objective world, i.e., a world identical to ali subjects, from the subjec­tive, private, point of departure (§2 and §66). Carnap also argues that the real point of departure is the subjectless "given" ("neutral given"), since the constitution of subjects only takes place at a later stage (§65). However, at the sarne time, he argues that the "neutral given" is materially subjective and structurally objective (§16). One wonders, of course, why the neutral subjectless given is, at the sarne time, subjective. For Carnap, what is "struc­turally objective" is a type-theory Jogic that is the toai for the construction of reality (similarly, but platonically, Russell assumes the objectivity of logical notions and the privacy of the contents of the given in, for instance, Part 1 of Theory of Knowledge). The major difference between Carnap and Wittgenstein is, that the latter does not intend to construct or constitute the objective reality. "How to reach the objective world" (Aufbau §2) is out of question for Wlttgenstein. The problems Wittgenstein is dealíng with are related to the very nature of Jogical inference and Jogical form (those, in Carnap's Aufbau, as in Russell's construction, are taken for granted in a non-frugal form that includes set and type theory).

30. Compare with T 5.6331. 31. The notion of 'time', however, will not be analyzed, for its analysis is not

necessary for my goals here. Wittgenstein does investigate and worries about the distinction between phenomenological and physlcal time, but 'time' is not part of bis draft of the phenomenological Janguage. Problems related to visual space will be sufficient for Wittgenstein to give it up.

32. According to Wilhelm Ostwald, who wrote one of the calor treatises best known in the 1920s and 1930s, one can consider calor from the point of view of physics (characteristics and measures of Jight), physiology (the effects of light in the eyes, nerves, and brain), chemistry (constitution of colors as pigments); however, he claims, "calor is rather sensation" (Einfuehrung ín die Farbenlehre, 15-16). It is from the analysis of sensations, with the help of the sctences,that a mathetic of colors (ordering laws, Jaws of mixture and meas­urement) is supposed to be determined. This approach is shared by other scientists from the 19th and beginning of the 20th centuries: Mach, Hering, Hoefler, Ebbinghaus, Buehler, for instance (one could say that those scien­tists were influenced by Goethe's theory). For an overview of the possible background of Wittgenstein's phenomenological investigation of colors see Rothhaupt (1996).

33. Concerning the visual space, Wittgenstein writes: "One needs, as it were, an elastic sign in arder to represent the space - so it seems to me ... The sign must have the multiplicity and the qualities of the space" (MS 105, 55).

34. Wittgenstein mentions verificationism or uses it as a toai in severa! passages of the midd!e period writings: WVC: 47, 65, 79; MS 105: 8, 10, 16, 40, 42; MS 107: 142-3, 177, 248, 250, 252, 254, 255, 284; MS 108: 1, 7, 93; MS 109: 26, 75, 193, 194; MS 110, 137 (command), 238 (grammar), 240; MS 111, 1 and 5, MS 112, 49v. By following these passages, one sees two things. First, it is a principie that, from the beginnlng, should also apply to mathematics (see MS

280 Notes

105, 10 and MS 107, 142-3). Second, Wittgenstein gradually sees the verifica­tion of a sentence as one way among others to make clear what is meant by it (MS 112, 49v).

35. Schlick already in his General Theory ofKnowledge (second edition from 1925) considered verificatlon as the way to test the truth of scientific hypotheses. A hypothesis must be translated by means of definitions into a long conJunc­tion of definitions and "judgments of perception" (p. 142). ln Erleben, Erkennen, Metaphysík (1926) Schlick comes closer to the principie: "lt is gener­ally accepted that the question whether the red that I experience is the sarne that somebody else experiences quite simply cannot be answered [ ... ] There is not a method (nane is thinkable) with the help of which both reds could be compared and the question answened. Therefore, the question does not have a reportable sense [ ... ] The question is whether one should call such questions that allow in principie no answer senseless or whether one should say: they have a sense, but we can only not report it 11 (34-5). Wittgenstein might be the philosopher who systematically uses the principie for the first time, even if not necessarily its inventor (as states Carnap in Autobiography, 45). He is not, however, responsible for the use made of it by the positivists Ia ter.

36. Visual phenomena have a spatial structure, the visual space. The visual space is, thus, the supposedly intrinsic spatial structure implícitly given in the visual field. '

37. Feelíngs as grounding the geometry of the visual fiel d had been defended by Mach (see Space and Geometry). Poincare, similarly, grounds this geometry on voluntary acts and muscular sensations (see La Notion d'Espace). See Nicod's Geometry in the Sensible World for the negatíon of this thesis. Wittgenstein was well aware of Nicod's work (see MS 105, 43 and 45).

38. Wittgenstein might have Mach and Russell in mind in this passage (perhaps also Poincare). Mach claims that "it is extremely probable that sensations of space are produced by the motor apparatus of the eye" and relates its motor functions to one being right or Ieft-handed and "the direction of gravity" (see Analysis o( Sensatíons, 49-51). ln The Relation of Sense-Data to Physics, 124, Russell argues that "logical economy11 asks for relativity in the "solipsistic experience" of each "private space": "For if we assume, as logical economy demands, that ali position is relative, a place is only definable by the things in or around it, and therefore the sarne place cannot occur in two private worlds which have no common constituent". Concerning right and Ieft ln the visual fiel d, Russell says: " ... right and left are complicated rela­tions involving the body of the percipient" (On the Relations of Universais and Particulars, 117). ln the sarne paper, Russell defends the view that "perceived space is certainly not absolute, i.e., absolute positions are not among objects of perception" (p. 116). It is precisely the opposite that Wittgenstein's "tran­scendental" argument purports to show. He agrees with Russell, however, in that absolute positions are not "objects of perception" - they are forms.

39. Also: "a distance [in the visual space] can be expressed by means of a number" (MS 106, 43).

40. The numerical representation of the visual field is at the centre of Wittgenstein's draft of a phenomenological language. As we will see, it is precisely this assumption that will make this language problematic.

Notes 281

41. The analysis of time is also absent (SRLF, 166). ln SRLF Wittgenstein also says that "the use of two dimensional space [the system of coordinates that he presents there] is not justified even in the case of monocular vision" (166). He simply makes this claim, because there he does not introduce the tran­scendental arguments discussed above that justify the system.

42. Poincaré points out that the third dimension is preclsely one of the character­istics that distinguishes visual from geometrical space: "the third dimension is revealed to us ... by the effort of accommodation, and by the convergence of the eyes" (Scie11ce and Hypothesls, 53-4). Strictly speaking, two eyes always see differently, since they are inches apart. But it is precisely this fact that helps us to see three-dimensionally. The role of the eyes became clear with the invention of the stereoscope. For an account of binocular vision that Wittgenstein might have read see Helmholtz's The Recent Progress of the Theory of Vision..-

43. Since the "analysis of phenomena" requires "rational and irrational numbers" in the phenomenological language (SRLF, 165), Wittgenstein's interest in the philosophy of mathematics ln 1929 is not accidental. ln fact, after his first remarks on visual space and phenomenology, he writes: "Apparently, against my will, 1 will be thrown back to arithmetic" (MS 105, 19). This strongly suggests that one should not overestimate the role that Brouwer's lectures played in Wittgenstein's return to philosophy ln 1929.

44. See also MS 106, 66: "If one makes use of analytic geometry, one gets certainly the right multiplicity ofthe designations (Bezeichnungen)".

45. Like Maxwell earlier, Ostwald grounds bis number system for colors (Einfuerung in die Farbenlehre) on rotating disc experiments with proportion­ally mixed colors. A metric of colors grounded on experiments was not what Wittgenstein was looking for: "1 could carry out the [mixing) experiment, say, with a rotating color wheel. Then it might succeed or not succeed, but ali that shows is whether or not the respective visual process can be produced by this physical means; it doesn't show whether it is possible. Just as the physical dissection of a surface can neither prove nor disprove its visual divis­ibility" (MS 108, 79, from 02.1930; in BT, 480).

46. Both problems are discussed in PR §§205-17. Many passages quoted in this section can be found there. See also WVC, 55-63.

47. Hjemslev developed a kind of rough geometry. ln such a geometry, there is no difference, for instance, between a 100-sided polygon and a circle (Die Natuer/iche Geometrie, 32 and Die Geometrie der Wirk/ichkeit, 43), for both are included inside a "fixed threshold". Klein introduced the idea of the "visual threshold" in the following way: " ... if we think of the distance between two points, we can estimate or measure it only to a /imited degree of exactness, because our eyes cannot recognize as different two line-segments whose difference in length lies below a certain limit" (Elementary Mathematics from an Advanced Standpoint, 35). Wittgenstein does not mention Klein and Hjemslev in his manuscripts, but does so in WVC, 56-7. ln his manuscripts, he writes about "the geometry of thick lines" (MS 107, 167) and mentions concentric circles in MS 107: 167-9, 173 (see below).

48. My interpretation of the nature of a phenomenological language and the reasons that led Wittgenstein to give it up dissent from current interpretations. Usually commentators think that the abandonment of the phenomenological

282 Notes

language (or its supposedly impossibility) is related to phenomenological 'time,' and not space as l argued (see Hintikka and Híntikka 1986; Kienzler 1997; Marion 1998; Prado Neto 2003). I don't deny that the notion of 'time' could have led Wittgenstein to abandon the phenomenological language (but lt did not) l briefly discuss the notion in Section 1.5.3. ln its draft, 'time' was not yet included and the draft itself, as it was, showed to be incorrect (it was1

however, not an "impossible" draft, for it is still used thereafter as a rough gram­matical representatlon). Moreover, as we have seen, right before abandoning the phenomenological language (4 pages earlier in MS 107), the discussion was about visual space as presented in the draft of the language, and not time.

49. Wíttgensteln uses the words 'grammar' and 'syntax' as synonyms through October 1929 and the end of 1930. After 1930 the word 'grammar' occurs far more often than 'syntax.' I think that this indicates that Wittgenstein is takíng some distance from bis views from the T and from eady 1929.

50. Hypotheses and propositlons are discussed in PR §§225-38. Wittgenstein, sometimes, seems to be suggesting that propositions and hypothesis have an identical structure. In the MSS, however, there are clear indications that they work in a similar, but not identical, way (see next pages).

51. This point is made clearer !ater (beginning of 1932): "To be sure, the process that leads to a piece of knowledge in a sdentific investigation (say in experi­mental physics) is not the sarne as one that leads to a piece of knowledge in life outside the laboratory; but it is similar and, when placed/held next to the latter, can shed some light on it" (MS 113, 58v; translated in BT, 120; Wittgenstein's emphasis).

52. "l am certain that" is a reference to Moore's A Defeme of Common Sense (1925). In the sarne MS, there is another: "Would it be possible that everything that 1 believe knowing for certain - for instance, that 1 have had parents, brothers and sisters, that Iam in England that ali that turned out to be wrong?" (MS 107, 248). Note that Wittgenstein's examples are of the kind of sentences that Moore claims to know in the mentioned paper. Wittgenstein's worry in the quoted passage is that the hypothetical character of ordinary propositions could give rise to a doubt concerning the existence of ordinary objects. The !ater discussions of On Certainty are, thus, grounded on old worries, which were refreshed by Moore's Proof of an Externai World (1939) and, perhaps, by conversations with Malcolm and other papers written by Moore (Certainty).

53. The limitations of the begriffsschri~ of the T are also to be noticed in the case of quantified propositions. At the end of 1929 Wittgenstein envisages a new interpretation of those (see PR "chapter" IX). Even though this issue is important and interestlng, I won't discuss it in this book because it is not determining of further developments in Wittgenstein's phílosophy after 1930.

54. For an account of what remains fixed see also Waismann's Thesen {WVC, 233-61). It is an important document because it was, apparently, a common project of Wittgenstein and Waismann intended for publication. In Thesen, the whoie T is reformulated grounded on the new views developed in 1929. ln Chapters 2 and 4 I come back to Wittgenstein's collaboration with Waismann.

55. See, for instance, MS 111, 89 (1931) where Wittgenstein discusses the vague­ness of ordinary concepts (see chapter 2, pp. 110---1).

Notes 283

56. With 'types' Wittgenstein does not have in mind, of course, a hierarchy of types, but kinds of concepts, regions ln language, determined by specific systems of propositions.

57. See also MS 107, 214: " ... Euclidean geometry is the syntax of the state­ments about objects in Euclidean space11

; also WLC30-32, 8: "Euclidean geometry is a part of grammar. lt is a convention of expression and so part of grammar."

58. ln the next chapterwe will see how Wittgenstein applies the idea of autonomy to 'grammar' in general.

59. Peano's primitive propositions are: O is a number, the successor of any number is a number, two numbers are equal if they have the sarne successor, O is not the successor of any number and the principie of mathematical induction.

60. The most obvious reason for Wittgenstein to introduce the idea of multiple systems is that he had to give up the general form of propositions. Thls reason is, however, insufficient, for the form of any number could be given, in principie, by an operation not grounded on the general form of proposi­tions. Another reason might have been the peculiarity of algebra. Algebra can be taken as an independent calculus with letters, whích suggests that numbers don't give the essence of mathematics (see MS 106, 77-113). Here 1 am only speculating.

61. The whole passage in Spengler's work reads: "There is no mathematic [singular], there are only mathematics [plural]. What we call history of "the11 mathematic, supposedly the progressive realization of only one and immutable ideal, is in fact, as soon as we dispose of the deceptive picture of the historie surface, a multipllcity of independent and in themselves closed developments ... 11 (p. 82).

62. Moore says that Wittgenstein claimed that "it is impossíble to 'justify' any grammatical rule" (M, 272; my emphasis). Moore points out that Wittgenstein gave the following argument for this claím: ''(l) that any reason [justifica­tion] "would have to be a description of reality": this he asserted in precisely those words ... (2) that /1 any description of reality must be capable of truth and falsehood 11 (these again were bis own words), and it turned out, I think, that part of what he meant by this was that any false description must be significant. But to complete the argument he had to say something like (what again he actually said in one place) "and, if it were false, it would have to be said in a language not using this grammar"" (p. 272).

63. 'Autonomy of grammar' is a !ater development (end of 1930---33), which will be explained in Chapters 2 and 3. Even though Wittgenstein !ater (after 1931) takes the arbitrariness and the autonomy of 'grammar' as practically synonyms (see, for instance, BT, 233-8) the notions have different roles when introduced. 'Arbitrariness' indicates that rules of grammar cannot be justified by means of propositions (as above); 'autonomy' indícates that the rules of 'grammar' operate like the rules of a game ora calculus: the meaning of a word, as in the autonomous calculus in mathematics, is equivalent to its rules. One could say that 'autonomy' entails 'arbitrariness,' but not vice­versa (the faet that 'grammar' is arbitrary does not entail that the meaning of a word is not an object). ln the next chapter we will see why and how Wittgenstein extends the idea of an autonomous calculus from mathematics to the whole 'grammar.'

284 Notes

2 Russell's Causal Theory of Meaning, Rule-Following, the Calculus Conception, and the Invention of the Genetic Method

1. See MS 107: 235, 290, 292; M5 108: 183, 188, 254, 256, 262; MS 109: 74, 211, 198; MS 110, 94; see also PR §21; and WCL30-32, 9. For discussions related to the theory see MS 107: 229-94; MS 108: 184-280; MS 109, 7; Notebooks 153a (almost the whole notebook), 153b and 154 (ali from 1931).

2. ln fact, the name The Meaning of Mea11i11g comes from the title of a sympo­sium from 1920, in which the major topic was Russell's paper On Propositions: What they are and how they mean (from 1919) see Hilmy (1987, 110). ln this paper Russell introduces the basic ldeas of his causal theory of meaning, which are also present in AM and in An Outline of Philosophy from 1927. Wittgenstein refers to the latter in his lectures from 1932 (see WLC32-5: 25-7; see also M, 313-15).

3. Russell's neutral monism is a result of his long struggles with James' monism (see Essays in Radical Empiridsm). Russell begins with clear opposition (for instance, Theory of Knowledge) and ends up with acceptance in AM.

4. For another unflattering opinion about Ogden and Richard's book see Ramsey (1924).

5. Ogden & Rkhards use of 'thought' and 'r~ference' ís not completely uniform. They do speak about "thought or reference" (MoM: 10-11, 73) indicating that they can be used interchangeably, but they also say that 'reference' means "a set of externai and psychological contexts línking a mental process to a referent" (MoM, 90) (the referent is the object or set of objects referred to by a reference). lt is not clear how this broader idea accommodates the common understanding of 'thought' or even Russell's. The authors say that they are unable to give criteria of ídentíty for thoughts or references (MoM, 91).

6. Pages 1-64 of MS 108 come ríght after page 229 of MS 107. Page 230 of MS 107 comes right after p. 64 of MS 108. Wittgenstein started MS 108 in December 1929 because he dídn't bring MS 107 with him to Austria at the end of that year (see Nedo's introduction of Wiener Ausgabe Band II p. IX).

7. Russell's causal theory of meaning was, for instance, the main subject of his 4th Lecture ln 1930 (WLC30-32, 9). The theory is in the background of severa) subsequent lectures (see WLC:30-2: 9, 23, 25, 27, 30, 31, 33, 36, 40, 41, 43, etc.)

8. See, for ínstance, MS 107, 177: "Each sentence is an empty game of strokes or sounds without the relation to reality and its only relation to reality is the mode of verification."

9. Note that the idea that ali kinds of sentences are pictures could be seen as very doubtful when contrasted with the explanation of orders present in the Causal Theory of Meaning. When we command someone, we want this person to do something. So it might seem that we want to cause a certain behavior in this person. The 'pícture' present in the order (the description of the state of affairs that Is expected to fulfill the arder) seems, then, to be determined by the causal goals present in language.

10. The multiplicity requirement is not abandoned in Wittgenstein's grammat­ical conception, for, ultimately, it only expresses a requirement of perspicuity

Notes 285

(a representation must be as simple and as exact as possible). Language has the right multiplicíty, for we can describe any facts with it (we can also give orders, express desires with it, etc., that are confirmed by facts). The rules of 'grammar' must make this explicit.

11. One might wonder whether Russell really intended to defend that verlflca­tion-feelings are the signals of the verification of all propositional attitudes. There are two reasons for thlnking that he did hold this general view. First, he explicitly held the view in the case of desire and expectation. Second, he seems to say that his claims concerning expectation apply to ali kinds of verification: "I think that ali verification is ultimately of the above sort" (AM, 270 'above sort' seems to refer to his explanatlon of the verification of expectation).

12. ln Notebook 153a, from 1931, Wittgenstein suggests that if I give the order "Bring me something red" and someone brings me something and says that this gave him the feellng of order fulfillment, then I could reply "I didn't ask you to bring something that gives you this feeling" (MS 153a, 23v).

13. The point is also made in MS 107, 291 from 02.02.1930 (also PR §23). 14. Ricketts (1996) makes this point. He calls the two kinds of rules "rules of

designation and rules of agreements for forms" (1996, 74). 15. I am not suggesting here that we change rules of projection at wíll when we

speak a language; quite the opposlte. First, because the most important rules are given a priori (the forms function and argument). Second, because "arbi­trary rules" are already implicitly ln use in ordinary language.

16. If the existence of a were a sense condltion of Fa, this condition would be affirmed by means of other propositions ("externally"); but being a truth conditlon, it's expressed in the proposition ("internally").

17. Note that quantification Is constructed ln the T and, thus, assumes analysls into elementary propositlons of the form function/argument. For instance, -(3x) Fx is N(!;) (T 5.52). The values of; are, in this case, glven by a function fx (5.501) - a particular expression of the logical prototype <Px. This function indicates that propositions of the form fx will take the place of the propo­sitional variable ;. Thus, the construction of quantification begins with a given set of elementary proposltions of a given loglcal form and analysls of quantifled proposltions must end wlth the sarne kind of propositions.

18. Wittgenstein writes in 1915: "When thesense of theproposition is completely expressed in the proposition itself, the proposition is always divided into its slmple components - no further division is possible and an apparent one is superfluous and these are objects ln the original sense" (NB, 63).

19. Concerning the self-explanatory character of elementary proposltions see Winch (1969).

20. ln 09.1916 Wittgensteln writes: "Now it is becomlng clear why l thought that thinking and language were the sarne. For thlnking is a kind of language. For a thought too Is, of course, a logical plcture of the proposition, and therefore lt just is a kind of proposition" (NB, 82). ln a letter to Russell from 1919 Wittgenstein says: "I don't know what the constituents of a thought are but I know that lt must have such constituents which correspond to the words of Language. Again the kind of relation of the constituents of thought and of the plctured fact is lrrelevant. lt would be a matter of psychology to find it out" (CL, 125).

286 Notes

21. The sentence continuous with a sarcastic reference to the Russellian solution: "Is it similar to the case where we described a stomach pain and one asked '1 understand it, but where does the unpleasantness lie [?]"'.

22. It should be clear ín this passage that the worry about meaning is a middle period worry, and not a problem for, the T, where the question "ln what conslsts meaning?" is not asked or answered.

23. Wittgenstein's anti-psychologism has certainly a Fregean inspiration. He says in MS 111, 182: "Frege about psychological language. Ali bis remarks refer to the inexactitude of the psychological investigatlon, in opposition to the logical."

24. So what Wittgenstein means with the intention of the picture or the way we mean a picture seems to be quite strong: we have to take consciously the picture as the picture of something in reality.

25. Wittgenstein's investigation of the "intentional element" of language, method of projection and application of rules is the forerunner of the rule-following discussion of the PI. ln the following pages we will see how different aspects of this discussion arise in Wittgensteln's manuscripts from 1930-1.

26. 'Temptatlon' and 'being tempted to' are Wittgenstein's own expressions. It is interesting that Wittgensteln rarely uses these words in MSS from 1929 (MS 106, 220 and MS 108, 1), but after the middle of 1930 their use is frequent (see MS 108, 279; MS 109, 173, 209, 267; MS 110, 18, 26, 86, 106, 123, 141, 148; MS 111, 176; MS 112, 67v; MS 113,' 40r). The frequent use of the word is clearly related to the invention of the genetic method in MSS 109 and 110 at the end of 1930 and beginning of 1931. The first occurrence that is directly related to the method is the following: "One is tempted (through false grammar) to ask: How does one thlnk the sentence p, how does one wait for this and this to happen (how does one make it) [?] And in this false question is contained the whole difficulty in nuce" (MS 109, 173-4 from 24.10.1930). lt is also interesting to observe that 'temptation' and 'being tempted' are used very frequently once the genetic method is applled more consistently !ater (see Chapter 4 for the use of this word in the BB and the BrB).

27. Also: "To know what the sign means means to interpret it" (108, 256); and WLC30-32, 44: "To understand a thought means to be able to translate it according to a general rule. For example, playing a piano from a score. But the score does not cause us to play as we do; if it did there would be no right and wrong way of playing.1'

28. ln MS 108, 208 is written: "The feeling that 1 now connect to ali my observa­tions is the one of the uniqueness of thought".

29. lt is important to notice that Wittgenstein never defended such a shadow. He only feels tempted to postulate lt in 1930 (I further discuss this in Chapter 4, Section 4.2.3.2).

30. This temptation and the diffículties involved in an explanation of the work­ings of a projection are made expliclt !ater in the BB: "Suppose we said 'that a picture is a portrait of a particular object consists in its being derived from that object in a particular way'. Now it is easy to describe what we should call processes of deriving a plcture from an object (roughly speaking, proc­esses of projection). But there is a peculiar difficulty about admitting that any such process is what we call "inte11tio11al representatian~. For describe whatever

Notes 287

process (activity) of projectlon we may, there is a way of reinterpreting this projection. Therefore - one Is tempted to say - such a process can never be the intention itself. For we could always have intended the opposite by reinterpreting the process of projection" (BB, 33, my emphasis).

31. 'Calculus' and 'system' are taken as synonyms by Wittgenstein (at least before 1934). See, for instance, MS 109: 182, 220--2 and MS 110, 200, where he says: "To invent a language means to construct a language. To set up its rules. To write its grammar ... What holds for the word 'language' must also hold for the expression 'system of rules'. Therefore, also for the word 'calculus"' (BT, 65). Wittgenstein also uses 'calculus' in a purely mathematical sense (for instance, in MS 110, 11.)

32. Also: "Picture and reality must give one system. Like the result of a calcula­tion and the whole rest of the calculation" (MS 111, 109).

33. This, again, indicates that the calculus conception of language is an exten­sion of Wittgenstein's understanding of mathematical calculus in PR. He is, in fact, explicit about it: "[ ... ] r compare the application of language with that of the calculus of multiplication" (MS 110, 217; BT, 418).

34. Note that Wittgenstein thinks that his conception is in one sel!Se behavior­istic, namely, ln the sense that we should think the 'internai' as 'externa!.' He, of course, is not arguing for the reduction of the mental to behavioral dispositions.

35. lt is clear that thls is the context in which the so-called 'rule following problem' (Kripke 1982) comes into being.

36. ln the next pages, l will introduce the various traits of the genetlc method: 1) suspicion that philosophlcal puzzlement originates in false trains of thoughts and misleading analogies; 2) the presentation of those in the form of a mirrar; 3) this presentation is meant to generate in the reader the recog­nition of the genesis of his own puzzlement; 4) the focus of the method is not philosophical doctrlnes, but confusions and temptations that might have generated them; 5) its goal is to dissolve philosophical puzzlement.

37. ln earlier manuscripts the idea that false analogies mislead philosophers is sometimes mentioned (for instance in MS 108: 23-4, 32, 34, 266). It is inter­esting that the early uses are, however, directed to views that Wittgenstein wanted to criticize, and not to his own temptations, as is the case of the passages from MSS 109 and 110. This shows that the idea that false analo­gles underlle the formulation of phllosophical puzzlement is not the conse­quence of an out of the blue insplration.

38. Note that this is similar to what brings Russell to postulate mental images in AM. For Russell, the behaviorist habit-explanation of the use of words works well for the cases in which the objects are present, but inadequate when not. Thls, for Russell, implies the need for images: "Having admitted lmages, we may say that the word 'box' in the absence of the box, is caused by an image of the box" (AM, 204). This, in its tum, seems to imply the following: "When we understand a word, there is a reciproca! assodation between it and the images of what it 'means' [ ... ] Thus speech is a means of producing in our hearers the images which are in us" (AM, 206).

39. 'Misleading picture,' 'misleading analogy,' 'misleading símile,' and 'misleading comparison' work as synonyms here. This is clear in the example of what might be the misleading analogy behind his struggles with the causal

288 Notes

theory of meaning that 1 give in the next pages. Wittgenstein, however, also uses 'plcture' as a thought hablt "engraved in our Janguage itself" (MS 109, 224; BT, 423). A 'picture' is, in this case, a rough idea.

40. Wlttgenstein also says: "Nothing is more lmportant than to pull into light ali false thoughts and present them in an absolutely correct/faithful way11 (MS 109, 99). .

41. Wittgenstein wrltes that Frege knew about his confusion between fact and complex (MS 110, 250) - he might be referring to a conversation or to letter from Frege. Wittgenstein seems to attrlbute the confusion to the T; lt is unclear, however, whether this is a fair reading of the T (I come back to this issue in Chapter 5, Section 5.2).

42. In this case, we may think that we need an account of the thought in general, independently of this or that organism, in order to save the objectivity of thought. At some point, Wittgenstein himself says that,í'thought is not something human11 (MS 108, 217).

43. One could think that already in T 3.323 Wittgenstein points out that linguistic analogies mis!ead us and that in this way "the most fundamental confuslons are easily produced (the whole philosophy Is full of them)". There are, however, clear differences that run deep between what takes place in the T and what is prescribed in the genetic method in respect of analogies and the genesis of philosophical problems. ln fact, Wittgenstein does not talk at ali about misleading traias of thbughts in the T. He neither investi­gates the genesis of the formulation of philosophical problems, nor shows the steps that lead to their formulation there. Thus, the central features of the genetlc method are not present. Moreover, the critique of philosophy in the T is grounded on a conceptual notation. There is no ldea of a concep­tual notation grounding the genetic method. ln T 3.323-5 Wittgenstein is concemed with the correct conceptual notation and points out that one word in ordinary language may belong to different symbols, i.e., it may express dlfferent logical roles in the conceptual notation. Thus, in "Green is green," the first occurrence of the word 'green' expresses one kind of symbol, an argument, and the second expresses another symbol, a function; 'is' can symbolize copula or identity. If we don't pay attention to those roles, philo­sophical confusions are created (T 3.324). In order to avoid them, we need a conceptual notation "that excludes them by not using ln a superficially similar way sigas that have different modes of signification" (T 3.325). In this passage it is clear that, concerning mtsleading analogies, Wittgenstein in the T is closer to Frege's and Russell's ídeas than to the genetic method, for it is a conceptual notation that should avoid them. Wittgenstein explidtly says that Frege and Russell have such a notation (T 3.325). Their notation, however, he also claims, simply "fails to exclude ali mistakes" (T 3.325). Frege and Russell would certainly agree that philosophy is full of mistakes caused by "superficially similar use of signs, 11 for they, severa! times, criticized in similar terms the subject-predicate logic. Russell, on the one hand, crlticizes Hegel's idea of the 11 union of identity in difference": "Hegel's argument [ ... ] depends throughout upon confusing the 'is' of predication, as in 'Socrates is mortal', with the 'is' of identity, as in 'Socrates is the philosopher who drank the hemlock'. Owing to thls confusion, he thinks that 'Socrates' and 'mortal' must be identical [ ... ] vast and imposing systems of philosophy are built

Notes 289

upon stupid and trivial confusions" (Our Knowledge of the Externai World, 48-9, footnote 1). This obviously sounds like T 3.323. On the other hand, Frege says, when criticizing "superficially similar uses of signs," that his conceptual notation should "break the dominance of word over the human spirit by uncovering illusions about the relationships of concepts that, very often, almost inevitably, come into existence" (Begriffsschrift, preface xii). The very idea of the centrality of a notation, as we have seen ln Chapter 1, is abandoned by Wittgensteln already ln 1929, a year before he begins devel­oping the genetlc method.

44. The plan for the book was announced in the fírst volume of Erkenntnis, in 1930. The project is also mentioned in a letter (05.08.1930 (LWGB)) from Schlick and Walsmann to Wlttgenstein.

45. This means, of course, that the "would be" method of T 6.53 was not prac­ticed ln that book: there Wittgenstein does not say propositions of natural science and does not wait for someone else to say something. The dissolution of phílosophlcal problems prescrlbed in 6.53 and the silence prescrlbed in 7 seem to be merely the results of the logical elucldation of "what can be sald" by means of the general form of proposltions (T 6).

3 The Big Typescript, the Tractatus, Sraffa, and the Anthropological View

L The clean text is published ln the Wiener Ausgabe (Michael Nedo (ed.)) and the copy with his handwritten remarks was published with an English trans­latlon by C. Grant Luckhardt and Maximilian A. E. Aue. When I make refer­ence to the BT, it is the clean copy that is meant, but l always use the English translation in quotations when it is possible. I also explicitly note when the remark that l am quoting Is not part of the dean copy, for instance, a verso remark (for this a use simply a 'v' after the indication of the page).

2. l discussed the first project in Chapter 2, Section 2.5. l discuss the second project in Chapter 4, Section 4.5.

3. Rothhaupt (2011) argues that Wittgenstein had another book in mind, the so-called Kringel-Buch, a book consisting of remarks from 1930-31 concerning mostly culture and anthropology. Wlttgensteln, however, always refers to "my book" (MS 110: 10, 13, 18, 178, 184, 243, 254, 258) and it seems un!ikely that he had in mind writing two books himself. Most likely, he wanted to use severa! remarks conceming culture and anthropology in the book he was planning (TSS 211-3). ln fact, some of those remarks appear in the BT and many in TS 211: 313-23. This strongly suggests that Wittgenstein did not have another book in mind, and that he simply gave up those remarks in the process of finishing the BT. ln any case, there is a big díffer­ence between the remarks on Frazer (from 1930-31), when Wittgenstein was interested in anthropology, but had not incorporated ao anthropological view yet, and the !ater remarks (from 1936-7) concerning mathematlcs, mostly of fictional examples, when Wittgenstein had assumed the perspective (see Sections 3.2 and 3.4.3). Wlttgenstein's interest ln real anthropology, it seems, was prompted by Spengler's ideas around 1930-31 (see MS 113, 102v; BT, 543). The sarne can be said about Wittgenstein's general interest in cultural

290 Notes

matters at the time (see MS 110, 13). The adoption of the anthropological view, however, took place after 1933 and it is connected with Sraffa's stimu­lating criticlsm (PI, preface), as is explained in Sectlon 3.4.

4. ln MS 155, from 1932, one finds the following interesting remarks about 'grammar' and method: "1 only want totabulate the use ofwords. 1 ama deaf and dense secretary who asks you 10 times before he puts anythlng down. What 1 want to teach you isn't opinions but a method. In fact the method to treat as irrelevant every question of opinion [ ... ] Refrain from writing down any hypothesis and any vague general statement and you have made a philo­sophical investigation" (MS 155, 40r-v).

5. This point is, 1 think, missed when one takes the therapeutic character of Wíttgenstein's philosophy ln the BT as psychoanalytic and fails to see later developments in his philosophy. Baker (2004), especially Part li, correctly criticized Hacker for not paying attentlon to the therapeutic character of Wlttgenstein's philosophy and for over-emphasizing the role played by 'grammar,' but he did it while not paying attention to the relevance of 'grammar' and its sense determining role in the BT. He grounds his interpre­tation mostly on Waismann's work, which he sees as the correct expression of Wittgenstein's philosophy in the early thirties. As Hacker (2006) correctly poínts out, Baker's psychoanalytic reading misrepresents the works of Waismann (I discuss Waismann's work ln Chapter 4, Section 4.5). Moreover, Baker misrepresents the philosophy of the BT, for he does not account for the calculus conception of language. 1 think that the dispute between Baker and Hacker is grounded on partial emphasis: Hacker emphasizes 'grammar' as a sense determiníng discipline, while Baker emphasizes the therapeutic aspect of Wittgenstein's philosophy. Both aspects are in place in the BT (and in Waismann's work), but Baker and Hacker over!ook their role there and !ater, 1 think. What characterizes both interpretations is the Jack of details in their treatments of Wittgenstein's middle period.

6. In severa! passages the sarne point of view is expressed (BT: 65, 81, 100, 112, 143).

7. "[ ... ] grammar isn't concerned with the purpose of language and whether it fulfils it. Any more than arithmetic is with the uses of addition" (BT, 192v).

8. Concerning the tabulation of mies and the list of mies of language see BT: 68, 71, 153, 155, 178, 188,245,250,253,426,548, 758.

9. BT, 292: "Ali my reflections are always directed towards showing that it doesn't do any good to conceive of thinking as hallucinating. In other words, that lt is superfl-uous, and leaves the problem unchanged. For no image, not even a hallucinatlon, can bridge the gap between image and reality, and no one image is better at this than another."

10. One might wonder, here, whether only the content ofwhat is called "propo­sitional attitudes" is part of the calculus. What about the attítudes them­selves? Even though Wittgenstein does not discuss the issue, and focuses on the propositional content, one mlght think that he is committed to taking attitudes as part of the calculus in the BT, for attitudes are expressed in the system of language and, prima facie, they lndicate moves dífferent from each other in the calculus. "Fearing that p," for instance, usually excludes "hoping that p." Some complícations seem to unfold, however, ln this line of reasoning. "Fearing that p" and "expecting that p," for instance, are not necessarily exclusive attitudes. Moreover, attitudes in general are not chosen

Notes 291

as the words we mean inside a calculus. Thus, even though attltudes are expressed in words, lt is unclear how they precisely work according to the calculus conception of language. Wittgenstein's worries conceming attitudes in the BT are mainly related to the mistakes that we are prone to making when talking about them. The idea that attitudes are transparent to intro­spection, for instance, is not correct, argues Wittgenstein. Introspection is a process of evaluation. We can ask, for instance, "Do I really !ove her?" (BT, 391). The answer to this question will be given in a process of lntrospection that "consists in calling up memories, imagining possible situations and feel­ings that one would have, etc." (BT, 391). ln thls sense, thus, we don't say 111 know by introspection that 1 have toothache." There is no process involved here. However, even saying "I know that I have toothache" is nonsense, for there is no data dlfferent from merely having toothache that would justify the use of "I know." "l know that I have toothache" does not say more than "! have toothache" (they are not contrasting alternatives). We only use the phrase "I know" in those cases because we misleadingly construct an analogy between the questions "How do you know that you have toothache?" as "How do you know that heis at home?" The analogy breaks down, however, as soon as we realize that we can give various reasons in order to answer the second, but not the first question (BT 392-3).

11. Here, presumably, one needs to distinguish between the rule that belongs to the interpretation of language and its temporal-spatial explanation (applica­tion of language). The rui e itself works, then, as a chart ( color red - word 'red' or 2 X 1= 2, 2 X 2= 4, etc.), while its explanation makes reference to time and space.

12. See T, preface: "Thus the aim of the book is to draw a lirnit to thought, or rather - not to thought, but to the expression of thought ... It will ... only be ln language that the limlt can be drawn, and what lies on the other side of the limit will simply be nonsense". This is not the expression of a deceiving strategy, but a clear statement of the goal of the book.

13. This is a little odd. The rule "lm=the length of the standard meter in Paris" conjoined to "Wagner líved in Dresden" seems to be absurd. It is difficult to imagine a situatlon in which such a conjunction would be appropriate. Wittgensteln thinks, however, that the rule does not change the sense of a sentence when it is conjoined in specific contexts (for instance, maps and legends, or measurement and the definition of the unit of measurement). Thls rule is also odd because, supposedly, rules don't make reference to time or space. But the rule in question can also hardly be a rule of the applica­tion of 'gramrnar,' for it defines 'meter,' i.e. it Is part of the lnterpretation of language.

14. "As if" indicates that this was not the idea that moved Wittgenstein in the T; rather, it is an idea that he was not aware of at the time.

15. One might, however, wonder whether Wittgenstein is likewise not simply assuming now that rules implícit ln the existing language also have a self­explanatory character. He assumes, for instance, that "the only things that are exact and unambiguous and indisputable are the grammatical rules" (BT, 374). The idea that at the end we reach rules of 'grammar' that constltute language seems to reintroduce in a milder form the rnythology of the self­explanatory and of the implicitly given. This point is further discussed in the next sections.

292 Notes

16. The conceptual notation of the T includes truth-tables, logically independent elementary propositions, the general form of propositions, the constroction of quantificatlon (that generates truth-functional compounds), and conven­tions that eliminate the ideotity sign (see, respectively, T: 5.101, 5.134, 5.5-5.5262, 5.53n).

17. Moore, in his notes on Wittgenstein's lectures from 1930-3, says that Wittgenstein insisted that "every significant word or symbol must essentially belong to a 'system,' and (metaphorically) ... that the meaning of a word is its 'place' in a 'grammatical system"' (p. 252).

18. See BT, 662: "What makes the calculat!on possible is the system to which the proposition belongs; and that also determines the mistakes that can be made in the calculation. 11 See also BT, 636-7: ": ... the proposition (Satz) must belong to a system of proposit!ons, and the proof to a system of proofs."

19. ln the calculus conception of the BT, elementary proposit1ons have a new role: 11

... an elementary propositíon is cine that, in the ~as Iam rww uslng it, doesn't represent itself as a troth-function of other propositions" (BT, 100; my underlining). They, thus, may imply and be lmplled by other propositions (BT, 101). As longas we don't have a method to finding out the "hidden product, 11 as we have ln a mathematical division, we don't have a hidden product either (BT, 100-1).

20. The consequence of this view is that reasons for choosing 'grammars' can be only pragmatic in nature. It is interesting to notice that, to say the least, the BT is very similar to fundamental distinctions in the structure of the argu­mentation in Carnap's Logical Syntax of Language (see, for instance, pp. 4-5). The similarity I have in mind is not restricted to the distinction between questions of fact and questions of truth and the several comparisons of rules of syntax with rules of a calculus or of a game. The point of similitude, I think, goes deeper: if only pragmatic issues are relevant in the choice of a 'grammar' because there are no questions of facts concerning such a choice, Carnap's "tolerance" in the choice of 'grammars' must be accepted. If no 'grammar' needs to be accepted as true, we have better tolerate different 'grammars' and choose the one with the best practical results: " ... what we call possible and what not depends entirely on our grammar, i.e. on what it permits. But that's arbitraryl - Certainly, but I can't do something with just any structure; that is: not every game is useful ... 11 (BT, 99; see also BT: 124, 453). Note, however, that the philosophy of the BT also coincides with a Quinean view, if we change our perspective from propositions to sentences that express roles: necessity is expressed by roles in general without any intrinsic or in-priociple distinction of kinds of roles. Now, if this is the case, we can only distinguish them because some apply to more, say, regions of language than others. Moreover, as for Quine, pragmatic reasons will determine which rules are more relevant to keep if "surprises do occur in the world" (BT, 63). lt is not an exaggeration, thus, to look at the material of the BT as the forerunner of the Camap/Quine debate.

21. Also: "A proposition is as similar to a fact as the sign "5" is to the sign "3 + 211

• Anda painted picture is as similar to a fact as "11111" is to the sign "li+ Ili"." (BT, 84)

22. Kuusela makes the point (or a similar one) of the assumed implicit knowl­edge of logic in the T when he argues that for Wittgenstein the reader of the T has a "pre-theoretical understanding of language" (2008, 63).

Notes 293

23. He also called 'p or not-p' a "nonsensical function 11 in NL. Note that Wittgenstein calls tautology 'nonsense' and a nonsensical expression 'tautology.' It thus comes without surprise that Wittgenstein did not correct Ogden's rendering of Unsínn as senseless in many remarks of the T, including 6.54 (also: 4.124, 4.1272, 5.473)-see Lugg (2003). It would be quite amazing if Wíttgenstein had not seen many instances of U11sinn in his very careful proofreading of Ogden's translation (Letters to C.K. Ogden). l am not claiming, however, as 1 make clear in the next pages, that there is no point in distin­guishing unsínnig and sínnlos.

24. Ramsey wrote in Philosophy (1929, 1) that if phllosophy is nonsense, "we must then take lt seriously that it is nonsense, and not pretend, as Wittgenstein does, that it is important nonsense. 11

25. This remark is introduced around 06.1931 (see MS 153a, 51 v-54r). 26. My reading is, at least partially, in agreement with Lugg (2003), for whom

se1JSeless and 1101JSense might designate the sarne sentences, and Moyal­Sharrock (2007), who calls the sentences in the T 'grammatical.' These authors correctly deflate the talk about inexpressability.

27. As we will see in Section 3.4.2, the 'tautology' that language is the totality of propositions quoted above (NB, 52; T 4.002) will also tum into something far from obvious or trivial.

28. Kienzler (2009) brought my attention to the significance of the motto. 29. The "resolute readers" (see Conant 2000) incorporated sentences of the

middle of T in the book's "frame," which was first only the preface and the final remarks (Diamond 1996). Wittgenstein, however, ln T 6.54, does not say that 'many' or 'almost all' of the propositions of the book are nonsense. He says "my propositlons" - no exceptions. If one takes the idea of a 'frame' seriously, one should also take the idea of limits of sense introduced in the preface seriously. Wittgenstein gives, indeed, the llmits of language in his notation. This means, I think, that there is no such thing as "piece­meal view of nonsense11 (a case-by-case analysis of nonsense) ln the T, as Diamond suggested (2004). The piecemeal view of nonsense is introduced in Wittgenstein's philosophy only !ater, after he abandoned the idea that a special notation has a central role in philosophy (Chapter 1) and invented the genetic method (Chapter 2). However, even in the BT the genetic method is not completely developed. Resolute readers usually rely on the context principie (certainly not part of the 'frame') in order to argue for a plecemeal view of nonsense. As we have seen, the context principie is as nonsensicaI andas superfluous (or trivial) as any other "proposition" of the book (it does not really matter whether we count it as part of the 'frame' or not). There is also no difference in function between this principie and other propositions of the book, for they ali elucidate, as Wittgenstein makes clear (6.54).

30. See also BT: 42, 61, 263 ("the sarne rules, the sarne type of word11), and 743

("logical types"). 31. See also BT: 266-71 315-61 392, 497, 500, 518, 543, 616, 618, 728. 32. The "old principie" is, however, restrained in the BT in order to avoid obvious

counterexamples. Sentences in works of Uterature have sense, even though they don't have empírica! conftrmatíon: "they relate to sentences (Saetze) for which there is verification as does a genre-paintlng to a portrait" (BT, 85).

33. Wittgenstein shows increasing interest ln gestures from 1931 onwards (see, for instance, MS 153a: 32v, 39v, 55v,154r). This, one can surmlse, is related

294 Notes

to the primitive picture of language that ascribes fundamental signlficance to naming: the meaning of a word is an object. Presumably, one uses gestures to give names. It relates, thus, to the T's assumption of simples and to Augustine's descrlption of how he learnt language. Gestures, says Augustine, are "the natural language of ali peoples" (PI, §1). A child learns a language by means of gestures (BT, 173). It is thus very tempting to connect gestures with the idea of indefinables, of self-explanatory sigos.

34. See also PG: 125, 110 and 160; compare with BT: 81, 76, 79. 35. I won't give an account of PG here. It seems that Waismann tries to make

the calculus conceptlon and the anthropologlcal view compatible in PLP (see Chapter 4, Section 4.5).

36. Wittgenstein met Sraffa in 1929, but had regular meetings with him only after October 1930 (WiC, 191). He mentions him as an influence, however, already in 1931 (about this, see Section 3.2). ·

37. This is a remark from 1937, when Wittgenstein finished the first version of the PI (MS 142). ln 1938 he wrote the first version of the preface, where he thanks Sraffa's crlticism for his "most fruitful ideas" in PI.

38. Note that this remark is within handwritten quotation marks and prefaced by "Check: Consider. 11 As we will see, Wittgenstein abandons the view that "understanding begins only with a sentence".

39. See also (BT: 3, 11, 14, 38, 40, 48, 179, 187, 190-4, 210, 215). Many of these remarks are crossed out.

40. Sraffa's (and !ater Wittgenstein's) polnt that our language includes primitive forms, which are not part of a calculus, is stated several times in later refor­mulatíons of passages of the BT besides the one discussed above. Here is one of them: "I said that the meaning of a word is its role in the calculus of language ... But think of the meanlng of the word "oh!" If we were asked about it, we would say '"oh'! is a sigh; we say, for instance, things like 'Oh, it is raining again already"'. And that would describe the use of the word. But what corresponds naw to the calculus, the complicated game that we play with other words? ln the use of the words "oh!", or "hurrah", or "hm", there Is nothing comparable" (PG, 67; from MS 140, 23; my emphasis).

41. Wittgenstein says ln the PI that it is lack of clarity concerning understanding, meaning and thinking that "may lead us (and did lead me) to think that if anyone utters a sentence and means and understands it he is operating a calculus according to defini te rules" (PI, §81 ). Thus, in the PI he is criticizing not only the T, but also hís own conception of language as a calculus or game with fixed rules of 'grammar' from the BT. Thus, using Wittgenstein's works from 1930-33 to give support to an interpretation of his ideas concerning 'grammar' in the PI is nota good move.

42. Those remarks were both crossed out by Wittgenstein. 43. Note that, agaln, Wittgensteln is pointing out that the correlatíon between

understanding and explanation taken for granted in the calculus conception and incompatible with some surroundings of Sraffa's "derisive gesture" is quite doubtful.

44. ln the original, one reads "Not an apple, 11 which must be wrong, given the parallel with "Not Brick".

45. Sraffa, then, also criticized the idea of 'multiplicíty.' Thls might explain why Malcolm reported hearing the word.

Notes 295

46. ln a handwritten remark in the BT, ln the sarne context, one finds: "A language consisting of just one signal that is always given when a particular action is to be carried out. Animal-training" (BT, 92v).

47. Sraffa begins his book with the description of a primitive society ln which the production for subsistence takes place: "Let us consider an extremely simple society which produces just enough to maintaln itself. Commodities are produced by separate industries and are exchanged for one another at a market held after the harvest" (Production of Commodities by Means of Commodities, 3). Further "complications" (surplus, proportion of labor and production, fixed capital, etc.) are introduced in each subsequent chapter. The use that Wittgenstein makes of the idea of primitive cultures and growing complexity is certainly quite different from Sraffa's. Arguably, it shows immediately that there was something wrong with the conception of Ianguage of the T. Note that the phllosophy of logic of the T mies out primitive languages right from the beginning, for "language is the totality of propositions" (T 4.002). Wittgenstein also has his own use of the idea of growing complexity. He uses it to show how philosophical puzzlement arises as each new element is introduced in language (proper names, numerais, indexicals, tables, etc.). He points out the genesis of philosophical problems at each levei of complexity in arder to scrutinize philosophizing right from the beginning, i.e., before it develops into a more sophisticated theory, in which the obscure beginnings are taken for granted.

48. Wittgenstein probably invented his idea of a family resemblance after reading Spengler's Der Untergang des Abendlades. ln his comparison of various cultures through history, Spengler ascribes a similar meaning to words such as 'king' and 'religion' (BT, 543). Wittgensteln also says that Spengler would be better understood if he said that he is merely comparlng different periods that resemble each other Iike family likeness (BT, 258). Spengler, in fact, does not merely do that. He thinks that he has found a hlstorical pattern that shows that Western Civilization wíll decay. He thus takes the pattern that he supposedly discovered in history and imp!ies that history must behave in accord with the pattern. This, for Wittgenstein, is dogmatism (BT, 258-9) -dogmatism similar to the Tractarian idea that the logical structure of language must be found in its appllcation. The idea of family resemblance arises, thus, out of a critique formulated against Spengler.

49. Quoted in Monk (1990, 261). It is very likely that Wittgenstein read Malinowski. This fact can be attested by common vocabulary, by the fact that Wittgenstein certainly had a copy of Ogden and Richards' MoM (see Chapter 2) where Malinowski published a supplement, and by some remarks conceming the metaphysical use of language. Thls passage of Malinowski is telling: "This attitude in which the word is regarded as a real entity, containing its meaning as a Soul-box [ ... ] is shown to be derived from the primitive, magicai uses of language and to reach right into the most important and influential systems of metaphysics" (MoM, Supplement I, 308). Wittgenstein seems to say similar things ln the middle of a series of remarks on Frazer, for instance, that "ln our language a whole mythology is put down" (MS 110, 206) and that words don't have magicai effects (MS 111, 108). However, Malinowski accepts the causal account of meanlng as a general account of language: "lt is interesting to see·[ ... ] that their [Ogden

296 Notes

and Rícbards'] conclusions, wbicb are arrived at by the study of bigber types of speecb, bold good in tbe domaín of primitive uses of words" (MoM, Supplement I, 326). For Wittgenstein, sucb an account makes sense only in primitíve uses of language, in a "restricted concept of language" (PG, 187; quoted above).

50. Wittgenstein first uses tbe expressions 'Lebensform' and 'Form des Lebens' in the BrBG, §108 (1936). His interest in real antbropological data begins at the end of 1930, wben be began meeting Sraffa. Tbis, bowever, is a coincídence, I conjecture, for bis interest in real anthropology is dueto Spengler (MS 113, 102v; BT, 543). For notes on Frazer's The Golden Bough see, for instance, MS 109: 178-88, 195-200.

51. A connection with Frege's anti-psycbologisrn is also quite obvious bere: at first sigbt, an anthropological view seems to suggest a confusion between taking something to be true and beíng true (psycbologisrn). So it1 Ís quite obvious wby Fregean scenarios from Grundgesetie (buman beings witb different laws of logic) are discussed by Wittgenstein in, for instance, RFM l, §152.

52. Note that, according to Frege, logical psycbologism assumes that it is an agreement of opinions tbat grounds Iogic. He attributes to Erdmann tbe view according to wbicb truth is equivalent to general validity grounded on the "general agreement of tbose wbo make judgments" (Grundgesetze, preface).

53. MS 146 was written between 12.12.1933 and the 02.1934, i.e., tbesame period in whicb the BB was dictated. Tbe context of Wittgenstein's self-criticism is tbis: "'Tbe sentence becomes a sentence by means of tbe language tbat it belongs to.' But tbe concept sentence is given by wbat we call 'sentences.' And, nonetbeless, tbey are always positions of words/signs of a language. And 'sentence' and 'saying' are correlative / words tbat belong togetber (tbink bere of 'bypothesis' and 'conjecture' and analogous ones) (Iam again and again prone to systernatize instead of purely to describe, as it would be correct)" (MS 146, 30).

54. 1 write 'really' because this way of understanding the notion of 'verification' is already suggested in the BT. As seen in Section 3.3, bowever, the principie of verification carries far more weight tban tbis in tbe BT.

55. Compare, for instance, wbat Wittgenstein says about arbitrariness before 1934 and later. ln MS 156b, lOr (before 1934), be writes: "lf l say its [language's] rules are arbitrary, it means "l leave aside tbe purpose or tbe usefulness [of language]"". Around 1938, in MS 117, 140 (also MS 160, 3v), be writes: 1111Grammatical rules are arbitrary" rneans: their purpose is not, for instance, to correspond to tbe nature (Wesen) of negation or colors - but [to] the purpose of negation and the color concept."

56. Tbis is a remark from Part l of MS 116, from 1936 or !ater (see Hilmy (1987) about the dating of tbis MS).

4 The Road to the Philosophical Investigations (Blue Book, Brown Book, German Brown Book, and MS 142)

1. Wittgenstein, in fact, says that bis 'grammar' is "more complete [tban ordi­nary grammar] since it includes ostensive definitions [ ... ], Russell's tbeory of descriptions, etc." (WLC32-35, 31; from 1932-3).

Notes 297

2. But see also: BB:l, 7, 8, 15, 19, 24, 26 51 and 70, where we can always substi­tute 'grammar' for 'use'. In the BrB the sarne equivalence reappears: "It is one of our tasks here to glve a picture of the grammar (the use) of the word 'a certain'" (BrB, 135).

3. Such assumptlons are also usually part of the prose used in the philosophy of mathematics. Wlttgenstein points out, for instance, that for Hardy, Goldbach's conjecture is a proposition "because he can believe that it is true" (BB, 11); on the other hand, Frege thought that "something" grasped by the mind, presum­ably a self-explaining meaning, must be added to the dead signs (BB, 4).

4. Philosophical reasoning, ln general, has the forrn: this is (or must be) the case, as everybody knows; what else must or cannot be the case? Deflatlng the ph!losophical must is one of the major points of Wittgenstein's philosophy. One should be very cautious in arder not to attribute to him after the BB the talk about "conditions of possibility," "transcendental arguments, 11 and other expressions which involve necessity and impossibility.

5. An interesting question is whether the strategy of the BB - where the anthro­pological view is not systematically employed is more promising than the strategy of the BrB. Both books, BB and BrB, need a more careful scrutiny than the one 1 can give here in order to answer this question.

6. Bouwsma (1961) sees the BB as a kind of "intellectual therapy" meant to help philosophers to get rid of certain temptations prompted by false analo­gies (155), which he also calls "grammatical confusion" (157). However, his overemphasis on 'therapy,' even though suggested in the BB, might mislead. The focus should be on how Wittgenstein argues in order to avoid philo­sophical assumptions, and not on the cure of a disease, I think. The great merit of his paper is that he sees correctly that Wittgenstein does not explain how the mind supposedly works, but rather how philosophical confusion concerning the mind arlses. Bouwsma thinks that Wittgenstein operates ln three moments: a) in showing that philosophical questions are weird, b) in presenting the meaning of words involved in the particular puzzlement in order to show its sources or roots, c) in seeking to uncover the misleading analogy that !eads one to puzzlement (158). In my view, a is certainly an important strategy to enrlch the application of the genetic method, while b and e are some of its old traits.

7. See Augustine's Confessions, Book II, Chapter 14. 8. Wittgenstein thought that such an analogy misled himself, Russell and Frege.

Different uses of nouns are described in the first language-games of the BrB and remain important even in the PI. The description of different uses of nouns in the PI shows Wittgenstein's strategy of avoiding false analogies that lead to philosophical questions.

9. Of course, not only Frege uses such "strategies." A more contemporary iden­tical strategy is given, for instance, by Stalnaker, at the first paragraph of his paper Propositions: "Propositions are things people express when they make predictions or promises, give orders or advice. They are also things people doubt, assume, believe to be very likely, and hope are true. What kinds of things are they?"

10. In the PI, when talking about a different problem, Wittgenstein says that "the first step is the one that altogether escapes notice," exactly the one where the "conjurlng trick" takes place (PI, §308).

298 Notes

11. See Grundgesetze der Arithmetik I, introduction, and the "transcendental" argument for the objectivity of thoughts in Der Gedanke.

12. Wittgenstein also mentlons Russell's views from AM on pages 21, 22 and 48 of the BB.

13. For Johnson, a proposítion is an object of thought (an expression that Wittgenstein uses in tbe BB; see above)~ Sucb an object, tbinksJohnson, "can be treated independently of this or that thinker" (Logic 1, 3; see also cbapter XIII). Johnson's sbadow is, tbus, dose to Frege's 'thougbt'.

14. 1 am stressing this point bere because Hacker (1999), following Makolm, uses the BB in order to attribute a kínd of mentalism to the T.

15. But in the German revision of tbe BrB (BrBG), Wittgenstein distinguisbes the language-games from bis comments about the temptations of the philoso­pber by different widths of paragrapbs. Paragraphs witb comments are larger tban paragrapbs witb new language-games.

16. But not only tbese pbilosopbers. Wittgensteln also mentions severa! times in his middle and late periods Plato (PI §§46, 48, 518), Augustine (PI §§1, 32, 89, 436, 618) andJames (PI §§342, 413, 610) among otbers.

17. lf one does not pay attention to tbis fact, and to the use of an anthropo­logical view in tbe BrB, one might end up tbinking, as Ayer, that the BrB is merely a "guerrilla warfare against mental acts" (Ayer 1986, 57). Ayer thinks that Wittgenstein in the BrB ("like Ryle", he says) "would be satisfied with a dispositional account" (1986, 58) of the mental; Wittgenstein, according to Ayer, did not overcome the difficulties of such a reduction. Ayer assumes, of course, that Wittgenstein is after such an account. He observes that for Wittgenstein the intentions of pointing to a calor and pointing to a shirt are not distinguished by mental acts. He claims, then, that ln order to distin­guish them Wittgenstein "gives the unhelpful answer" (1986, 54) that the dlfference lies in tbe surroundings. Ayer does not see that Wittgenstein's point is rather that only by forgetting the surroundings can a philosophical questlon like "What distinguishes two intentions?" arise in the first place.

18. ln this passage we could also clearly substitute 'grammar' for 'use' as we could have substituted 'use' for 'grammar' in tbe former passage quoted. Note specially tbis replaced expression: "a decision about the grammar of the term proposition."

19. The German version of the BrB (second part of MS 115), which 1 call 'BrBG' (Eine Philosophische Betrachtung) is not simply a translation, but a somewhat revised version of the book.

20. ln MS 115, when Wittgenstein decides to give up tbe revision of the BrB (BrBG), he writes: "This whole attempt of a revision from page 118 [the beginning of BrBG] is no worth" (292). This passage prompted commenta­tors (Pichler 2004; Schulte 2006) to think that there is something completely new in Wittgenstein's subsequent work. This, for the reasons given bere, 1 think is incorrect.

21. Problems of translation connected to different cultures were a current topic in Cambridge in the 1930s. This can be seen from 1. A. Richards' Mencius on the Mind: Experiments in Multiple Definition. Concerning bis translation of Mencius, Richards claims that right from the beginning the text "plunges us abruptly into the wíld abyss of conjecture through which we have ... to make our way" (p. 1). He thinks that difficulties are to be

Notes 299

overcome with a "systematic survey of the language we are forced to use in translation" (p. 92).

22. There is also an interesting connection between the application of the method to the T and Wittgensteln's private life. Actually, if one wants to Jook at how hls Jife and philosophy are closely related, the step from BrB and BrBG to MS 142 is perhaps the most striking example. At the time (1936-7), he decided to confess his mistakes and weaknesses to severa! friends (Pascal, 1984). The application of the genetic method to the T is, of course, a sort of confession. Wittgenstein, for sure, took very seriously such a ritual. He also did it in 1931 (Drury, 1984).

23. Wittgensteln refers directly to the BT on MS 157b, 15r, where he writes "See typescrlpt". Further evldence is that Wittgenstein, besides the remarks from MSS 157a and b, uses only remarks from BT 409-431 in the remarks about philosophy in MS 142 (the first version of the PI): MS 142 §108: 157b ... ; §109: 157a-b; §110: 157; §112: BT 412; §114: BT 416; §115: BT 417; §116: BT 417-8; §117: 418; §118: BT 419-20; §119: BT 415, 419 and 422; §120: BT 419; §121: 409-10; §122: MS 157b; §128: BT 412; §129: BT 411, 425; §133: 157b; §134: BT 431-2;

24. The section's heading of BT 409 is "Philosophy points out the misleading analogies in the use of our language". BT 409 reads: "lf l rectify a philosoph­ícal mistake and say that this is the way it has always been conceived, but this is not the way it is, I must always point out an analogy according to which one had been thlnking, but which one did not recognize as an analogy".

25. Wittgenstein was not, of course, a logical positivist, but one can hardly enter­taln the idea that he would be able to work closely to Waismann and Schlick for so many years if there were no overlapping !nterests amongst them. 1 hope that the first chapters of this book give at least strong hints of such proximity.

26. Baker describes the project as" an alterna tive" to the first project and mentions "eleven separate meetings with Waismann during the Eastern vacation of 1932, and many more during vacations in the next two years" (Baker 2003, xxvi). Contrary to what Baker writes, however, meetings also took place after 1934 (see Manninen 2011).

27. Quoted by Baker (2003), xxvii. 28. Waísmann, according to Baker, had also access to !ater works, such as the BB

and the BrB (Baker (2003), xxvii). We will see ín what follows some elements of the anthropological view present in Waismann's book.

29. Note that this confirms an important point made in the Chapter 3, when the anthropological view was discussed: in primitive (or simple) languages the talk about propositions and 'grammar' is very doubtful.

30. ln How 1 See Phílosophy, actually, Waismann distances himself from Wittgenstein. Even though he attributes to "Logical Positivism" the "obses­sion" of a "clarity neurosis" that ends up "sapping any creative power" (pp. 359-60), one can easily grasp that his target is Wittgenstein: "No great discov­erer has acted in accordance with the motto, 'Everything that can be sai d can be sai d clearly.' And some of the greatest discoveries have even emerged from a sort of primordial fog. (Something to be said for the fog. For my part, l've always suspected that clarity is the last refuge of those who have nothing to say)" (p. 360).

300 Notes

S The Philosophical Investigations

1. See Goldfarb (1983) for a very illuminating discussion of this topic. 2. Note that this is similar to what the interlocutor objects to in the BrB when

Wittgenstein discusses 'comparing': 11 But this cannot be ali comparing consists in" (BrB, 86) (see Chapter 4, Section 4.3.2). There the point is also to scrutinize the view according to which the supposedly essential trait of recognition is missing - namely, the mental operations.

3. Note how the central idea of the calculus conception of the BT (namely, the idea that sentences and words operate in contrast to other sentences and words) is deflated in PI: §§10, 20, and 28.

4. I disagree, thus, w:ith Hacker's understanding of PI §§428-65, which he takes to be a critique of the T (see, for instance, Baker and Hacker 2005b, 18-9). In my view, there Wittgenstein is not presenting the "deepest and most far­reaching criticism of the picture theory of representation, li but discussing the extended views of the T and views that he was inclined to defend around 1930 as a kind of answer to Russell (see Chapters 2 and 4). So it was not, as claims Hacker, 11 a tactical errar to bury" the supposed critique of the T in §§428-65 -simply because these sections are not a critique of the T at ali.

5. This is one of the passages of the BT about bis method that Wíttgenstein selected for the composition of the PI at the beginning of 1937 (see Section 4.4 of Chapter 4).

6. The problem of the position of the remark concerning methods has been finally taken into account in the new edition of PI, the Schulte and Hacker revised Anscombe translation, where it appears as a boxed remark. A similar strange inclusion happens in §108 of the PI (PUKGE, 808--9). A slip of paper w:ith paragraphs b, c, d of §108 was included ín the text. The remark was left between two pages and there is no specific indication that they ended up ln the place where Wittgensteln wanted them to go. In fact, it seems that they ended up in the wrong place. Without the insertion one can see the connec­tion between the sections given by the italicized "preconceived idea," but w:ith ít one has the clear impression that the author of the book changed the subject and then carne back to it ln the next paragraph. Schulte made this point in a talk some years ago. This has also been corrected in the new Schulte and Hacker translation.

7. See CL, 304-9. Wittgenstein sent Keynes MS 226 (Rhees' English translation of part of TS 220) and TS 220, but urged him to read the German, since he did not like Rhees' English translation, which Keynes also disliked (CL, 308). The translator (Rhees), wrote Wittgenstein, was 11an excellent man" (CL, 308) who "did his very best"; however, "the stuff is damn difficult to translate" (CL, 306), especially because "nothing's more difficult to translate than colloquial (non-technical) prose" (CL, 308).

8. The exaggeration of the therapeutic aspect of Wittgenstein's method was taken even further later in two papers published in Mind, in 1946, by B. A. Farrell called 11 An Appraisal of Therapeutic Positivism I and II". "Therapeutic Positivism" was taken to be Wittgenstein's philosophy, supposedly in accord w:ith papers written by Wisdom and Malcolm. According to Karl Britton's testimony, "the papers had much annoyed and upset" Wittgenstein (Britton 1967, 62).

Notes 301

9. As one can see frorn Moore's notes: "He said ... that a 'new rnethod' had been discovered ... and that lt was now possible for the fírst time that there should be 'skillful' philosophers ... And he sai d that the required skill could not be acquired rnerely by hearing lectures: discussion was essentlal" (M, 322).

10. One rnight also think that the general organization of the rernarks in PI led Wittgenstein to cut out sorne rernarks concerning the method. He may have felt the need to shorten the number of passages about the way he sees philos­ophy, which were a big part of the book in its early verslons.

11. The FF is composed by the following TSS: TS 225, the preface; TS 220, the revised MS 142 or UF; TS 221, the original second part ofthe PI, Wittgenstein's philosophy of mathematics.

12. Sorne other passages that were left out are related to the method of the PI (FF §§97, 99, 102 and 107). 1 have already dealt with those above.

13. The reason for uncertainty here is that there are two prefaces frorn 1945: the one published and another one where Wittgenstein says that he read the book with someone "two years ago". Schulte conjectures that this rereading happened in 1943 with Wittgenstein's friend Nicholas Bachtin (PUKGE, 23).

14. See also Knowledge by Acquaintance and Knowledge by Description, 177. 15. About Russell and 'self-evidence' seeT 5.4731and5.1363; about 'experience',

T 5.552; about the 'axiom of reducibility', T. 6.1232-3. 16. See Russell's second rlddle in On Denoting. 17. Concerning understanding and its relation to elementary propositions see

Chapter 2, Section 2.2.2. 18. How come? It Is certainly not because Plato had super powers: "l read

' ... philosophers are no nearer to the meaning of 'Reality' than Plato got ... ' What a strange state of affairs. How strange in that case that Plato could get that far in the first place! Or that after him we were not able to get further! Was it because Pia to was so dever?" (BT, 424; originally in MS 111, 134). This Is, perhaps, a reference to Whitehead: "The safest general characterization of the European philosophical tradition is that it consists of a series of footnotes to Plato" (Process and Rea/ity, 39). The riddle has perhaps this solution: "our language has rernained constant and keeps seductng us into asking the sarne questions" (BT, 424).

19. This is one of the reasons why Wittgenstein says that "the work of the philos­opher consists in assernbling reminders for a particular purpose" (PI § 127). See also §§89-90.

20. Goldfarb (1997) guides my interpretation of §§66-77. 21. In Copland (2002, 74). The passage goes on like this: "It is a cooler, more

even-sounding instrument than the oboe, being also more brilliant. Much closer to the flute than to the oboe in quality, it has alrnost as great an agility as the former, singing with an equal grace melodies of all kinds. ln its lowest octave it possesses a unique tone colar of a deeply haunting effect. Its dynarnic range is more rernarkable than that of any other woodwind, extending frorn a mere whisper to the rnost brilllant fortíssimo."

22. Concerning what is cornrnon between world and thought see Chapter 3, Section 3.2.

23. T 5.5563 reads: "In fact, ali the proposltions of everyday language, just as they stand, are in perfect logical order. That utterly sim pie thing, which we

302 Notes

bave to formulate bere, is not a likeness of tbe trutb but tbe trutb itself in its entirety. (Our problems are not abstract, but perbaps tbe most concrete tbat tbere are)."

24. Note tbat as late as 1949 Wittgenstein still wanted to finisb bis pbilosopby of matbematics: "I want to call 'Beginning Matbematics' my observations about matbematics tbat belong to my Pbilosopbical Investigations" (MS 169, 36v). Tbis seems to be a reference to TS 221, tbe early second part about matbematics of tbe PI (publisbed as RFM, I).

25. "'How does it belp you to postulate a creator, it only pusbes back tbe problem of tbe beginning of tbe world. 11 Tbis observation brings out ao aspect of my explanation tbat I perbaps bad not noticed. One migbt also say: "Look at your explanation in tbis way - ºº\'.\ are you still satisfied witb it?'"(PG, §52).

26. Hacker, for instance, says: "Par from reflecting tbe essential natures of tbings, grammar determines tbeir essential nature, in tbe following sense. ln laying down rules for tbe use of words, in fixing wbat is to count as sucb-and-sucb a tbing and in determining its criteria of identity, grammar tells us wbat kind of entity anytbing is (§373)" (Hacker, 2000, 80). Tbe idea bebind Hacker's reasoning is that in §§371-2 Wittgenstein is denying tbat "essential natures" are to be found in reality and that §373 states that tbe rules of grammar constitute meaning (see Hacker 2000, 85) and so determine wbat eacb tbing is. Hacker's move makes grammar a discipline tbat regulates sense, lays down rules for words and fixes wbat counts as "a tbing. 11 Hacker, on the one band, claims tbat, for instance, arbitrariness of grammar is "not a pbilosopbical thesis at ali" (2000, 89); on tbe otber band, be claims tbat Wittgenstein bas a "conception of necessity or essence" and talks about "Wittgenstein's !ater account of necessity" (1993, 238) in terms of tbe "arbitrariness of grammar. 11

Tbose very loaded pbilosopbical assumptions tbat Hacker attributes to Wittgenstein don't reflect tbe development of Wittgenstein's pbilosopby after tbe BT. I discuss related issues in my Engelmann (2011).

27. Wittgenstein's !ater account, argues Hacker, is a de dicto account in opposi­tion to a supposedly de re account of the T (see Baker and Hacker 2005a, 252). It is interesting to observe tbat, at first sigbt, tbe !ater Hacker does not count PI §372 as evidence for the de dicto conception, as tbe early Hacker did (Baker and Hacker 1994, 329). La ter, in the tbird volume of tbe commentary, Hacker recognizes tbat §372 is a quotation, but tben tries to make it fit bis old inter­pretation by assuming tbat in tbe quote Wittgenstein is opposing tbe T to tbe PI. He interprets the first claim of tbe quote as ao allusion "in a general­ized way to tbe position delineated in tbe Tractatus" (Hacker 1993, 237). He also says tbat "at tbat stage [Tractatus] Wittgenstein firmly believed tbat tbere are intrinsic necessities" (Hacker 1993, 237). He opposes tbis view to a view tbat be attributes to tbe PI: " ... we are being invited to consider the quoted remark by way of contrast witb the conception of necessity or essence that char­acterizes Wittgenstein's later philosophy [my empbasis]" (Hacker 1993, 238). His reading of §§371-3 is, tbus, tbe following: in §§371and373 Wittgenstein defends tbe supposedly !ater de dicto conception and in §372 be invites us to consider it in contrast to tbe T. However, Wittgenstein would not invite us "to consider" tbe remark in §372 if be really was defending it in §§371 and §373.

Notes 303

28. Thus, the target of PI, §372 is also the idea of a phenomenological notation, whlch supposedly would perspicuously show the necessary properties of the specific forms of space and calor (Chapter l, Section 1.3).

29. The distinction between the calculus (what is really proved) and the prose (what is said about what is proved) is made already in PR §155 and reappears in various occasions thereafter, for instance: WVC, 149, BT: 374, 629-37; RFMV, §46; LFM 16-7.

Bibliography

Augustine (1960) Confessions. Ryan, J. K. (tr.). (New York: Image Books). Austin, J. (1986) 'Wittgenstein's Solutions to the Color Exclusion Problem' in

Ludwig Wittgesntein II: Criticai Assesments. Shanker, S. (ed.); 207-12. (London: Croom Helm).

Ayer, A.]. (1986) Wittgenstein. (Chicago: University of Chicago Press). Baker, G. (2004) Wittgenstein's Method CI;Veglected Aspects), Morris, K.J. (ed.).

(Oxford Blackwell Publishing). -- (2003) The Voices of Wittgenstein (The Vienna Circle). Ludwig Wittgenstein

and Friedrich Waismann. Original German Texts and English translations. Transcribed, edited and with an introduction by Gordon Baker. (London: Routledge).

(1988) Frege, Wittgenstein and the Vienna Circle. (Oxford: Blackwell). Baker, G. and Hacker, P.M.S. (2005a) Wittgenstein: Meaning and Understanding(Part

I Essays). Vol. 2 of an analytical commentary on the Philosophical Investigations Second, extensively revised edition by Hacker, P.M.S. (Oxford: Blackwell).

-- (2005b) Wittgenstein: Meaning and Understanding (Part II Exegesis: §§1-184). Second, extensively revised editíon by Hacker, P.M.S. (Oxford: Blackwell).

-- (1994) Wittgenstein: Rules, Grammar and Necessity, Vol. 2 of an analytical commentary on the Philosophical Investigations. (Oxford: Blackwell).

Bellofiore, R. and Potier (2011) 'New Perspectives from the Sraffa Papers: A Preliminary Survey' in R. Ciccone, C. Gehrke and G. Mongiovi (eds), Sraffa and Modem Economics, Volume II. (London: Routledge); 338-353.

Black, M. (1982) A Companion to Wittgenstein's Tractatus. (New York: Cornell University Press.

Bouwsma, O. K. (1986) Wittgenstein: Conversations 1949-1951.]. L. Craft and R. E. Hustwit (eds). (lndianapolis: Hackett Publishing Company).

-- (1967) 'The Blue Book' in Ludwig Wittgenstein: The Man and His Philosophy, Fann, K.T. (ed.); 148-70. (New York: Delta Book).

Britton, K. (1999) 'Portrait of a Philosopher' in Ludwig Wittgenstein: The Man and His Philosophy, Fann, K.T. (ed.); 56-63. (New York: Delta Books).

Carnap, R. (1959 [1934]) The Logical Syntax of Language. Smeaton, A. (tr.). (New Jersey: Littlefíeld, Adams & Co. Paterson).

(1998 [1928]) Der Logische Aufbau der Welt. (Hamburg: Felix Meiner Verlag).

-(1963) 'Autobíography'ín The Philosophy of Rudolf Carnap. Schilp (ed.). (Open Court).

Conant, J. (2007) 'Mild Mono-Wittgensteinianism,' in Wittge11stei11 and the Moral Life: Essays in Honor of Cora Diamond, Crary, A (ed.). (MIT Press).

(2000) 'The Method of the Tractatus' in The New Wittgenstein Crary, A. and Read, R. (eds); 174-217. (London: Routledge).

Conant, J. andDíamond, C. (2004) 'On Reading the Tractatus Resolutely: Reply to Meredith Williams and Peter Sullivan', in Wittgenstein's Lasting Significance, Kõlbel, M. and Weiss, B. (eds); 42-97. (London: Routledge).

304

Bibliography 305

Copland, A. (2002) What to Listen For in Music. (New York: Signet Classics). Diamond, C. (1996) The Rea/ístic Spírit: Wittgenstein, Philosophy and the Mind.

(Cambridge, Massashussets: The MIT Press). -- (2004) 'Criss-cross Philosophy' in Wittgenstein at Work. Method in the

Philosophícal Investigations. Ammereller, E. and Fischer, E. (eds). (London: Routledge); 201-220.

Drury, M. O'C. (1984) 'Some Notes on Conversations with Wittgenstein' ln Reco/lections of Wittgenstein. Rhees, R. (ed.); 76-96. (Oxford: Oxford University Press).

Engelmann, M. L. (2009a) 'The Multiple Complete Systems conception as Fil Conducteur of Wittgenstein's Philosophy of Mathematics' in Proceedings of the 32nd Intemational Wittgenstein Symposium; also in 'Agora Project': http://wab.uib.no/ agora/alws/2011/engelmann2009. pdf.

-- (2009b) 'O que ê o Big Typescript?' Dois Pontos Vol. 6, n. 1: 35-61. (2011) 'What Wittgenstein's "Grammar" is not. On Garver, Baker and

Hacker, and Hacker on Wlttgenstein on "Grammar111 Wittgenstein Studien Vai. 2; 71-102.

-- (2012a) 'Sraffa and Wittgenstein's "Most Fruitful ldeas"' Philosophical Investigations Early View (04.2012).

-- (2012b) 'Russell's The Analysis of Mind and Wittgenstein's New Method' fournal of Philosophical Research Vol. 37; 283-311.

(2012c) 'Language Comme Calcul dans !e Big Typescript'. Philosophiques Vai. 39 (1); 35-55.

Farrell, B. A. (1946) 'An Appraisal of Therapeutic Positivism I' Mind, Vai. 55, N 217; 25-48.

-- (1947) 'An Appraisal of Therapeutic Positivism II' Mind, Vai. 55, N 218; 133-50.

Floyd, J. (2001) 'Prose versus Proof: Wittgenstein on Goedel, Tarski and Truth' Philosophia Mathematíca Series III, vai. 9; 280-307.

--(2000) 'Wittgenstein, Mathematics and Philosophy' in The New Wíttgenstein, Crary, A. and Read, R. (eds); 232-261. (New York: Routledge).

Fogelin, R. J. (1996) 'Wittgenstein's Critique of Philosophy' in The Cambridge companion to Wittgenstein. Sluga, H. Stern, D. (eds). (Cambridge: Cambridge University Press).

Forster, M. N. (2004) Wittgenstein on the Arbitrariness of Grammar. (Princeton: Princeton University Press).

Frege, G. (1987 [1884]) Die Grundlagen der Arithmetík. (Frankfurt: Reclam). (1962 (1893]) Die Grundgesetze der Arithmetik I. (Hildesheim: Georg Holms). (1962 [1903]) Die Grundgesetze der Arithmetik II. (Hildesheim: Georg Holms). (1998 [1879]) Begriffsschrift. (Hildesheim: Georg Olms Verlag). (1993 [1918]) 'Der Gedanke' in Logische Untersuchungen. Patzig, G. (ed.);

30-53. (Gettingen: Vandenhoeck & Ruprecht). Goldfarb, W. D. (1997) 'Wittgenstein on Fixity of Meaning' in Early Analytic

Philosophy: Esssays in Honor of Leonard Linsky. Tait, W.W. (ed.); 75-89. -- (1983) '1 Want You to Bring me a Slab. Remarks on the Opening Sections of

the 'Philosophical Investigations" Synthese 56; 265-82. Hacker, P. M. S. (2000) Wittgensteill: Mind and Will (Part I Essays). Vol. 4 of

an analytical commentary on the Philosophical lnvestigations. (Oxford: Blackwell).

306 Bibliography

(2007) 'Gordon Baker's Late Interpretation of Wlttgenstein' in Wittgenstein and His Interpreters, Kanterian, E., Kuusela, O. and Kahane, G. (eds); 88-122. (Oxford: Blackwell Publishlng).

-- (1999) 'Naming, Thinking and Meaning in the Tractatus' Philosophical Investigations; 119-135.

-- (1997) Insight and Illusíon, 2nd ed. (Bristol: Thoemes Press). (1993) Wittgenstein: Meaning and Mi11d (Part II Exegesis: §§243-427). Vol. 3

of an analytical commentary on the Phllosophical Investigations. (Oxford: Blackwell).

Hanfling, O. (2004) 'The Use of 'Theory' in Philosophy' in Wittgenstein at Work. Method in the Phílosophical Investigations. Ammereller, E. and Fischer, E. (eds) ; 183-200. (London: Routledge).

Helmholzt, H. Von (1995) 'The Recent Progress of the Theory of Vision' in Scíence and Culture (Popular and Philosophical Essays). Cahan, D. (ed.);127-203. (Chicago: The University of Chicago Press).

Hertz, H. (1896) Die Prinzipien der Mechanik in neuem Zusammenhange dargestelll, Gesammelte Werke Band III(Leipzig: Johann Ambrosius Barth).

Hilmy, S. S. (1987) The Later Wittgenstein (The Emergence of a New Philosophical Method). (Oxford: Basic Blackwell).

Hintikka, M.B and Hintikka, J. (1986) Investigating Wittgenstein. (Oxford: Basic Blackwell).

Hjelmslev,J. (1916) 'Dle Geometrie der Wirklichkeit' Acta Mathematica 40, 1916; 35-66.

(1923) ,Die Natuerliche Geometrie'. Hamburger Mathematische Eínzelschriften 1. Helft; 1-36.

Hylton, P. (1997) 'Functions, Operations, and Sense in Wittgenstein's Tractatus' in Early Analytic Philosophy: Esssays in Honor of Leonard Linsky, Tait, W.W. (ed.); 91-105.

Jacquettte, D. (1998) Wittgenstein's Thought in Transition. (West Lafayete: Purdue University Press).

James, W. (1996 [1912]) Essays in Radical Empiricism. (University of Nebraska Press).

Johnson, W.E. (1921) Logic l (Cambridge: Cambridge Universlty Press). Kenny, A. (2004) 'Philosophy States Only What Everyone Admits' in Wittgenstein

at Work. Method in the Philosophical Investigations. Ammereller, E. and Fischer, E. (eds); 173-182. (London: Routledge).

Kienzler, W. (1997) Wittgensteins Wende zu seiner Spaetphilosophie 1930-1932. (Frankfurt am Main: Suhrkamp).

-- (2009) 'Dle Sprache des Tractatus: Klar oder deutlich?' in Wittgenstein Philosophie ais Arbeit ,,an Eínem selbst". Gebauer, G., Goppelsroeder, F., and Volbers, J. (eds). (Munich: Wilhelm Fink).

-- (2001) 'About the Dlviding Llne Between early and Late Wittgenstein' in From the Tractatus to the Tractatus and Other Essays, Oliveri, G. (ed.). (Frankfurt: Peter Lang Verlag).

Klein, F. (2004 [1908]) Elementary Mathematics {rom an advanced Standpoint: Arithmetic, Algebra, Analysis, Hedrick, E. R. and Noble, C.A. (tr.). (New York: Dover Publications).

Kremer, M. (2001) 'The Purpose of Tractarian Nonsense' Nous, 35; 39-73.

Bibliography 307

Kripke, S. A. (1982) Wittgenstein on Rules and Private Language. (Cambridge, MA: Harvard Universlty Press).

Kuusela, O. (2008) Wittge11stei11's Struggle Agai11St Dogmatism. (Cambridge, MA: Harvard University Press).

Lugg, A. (2000) Wittgenstei11's Jnvestigatio11s 1-133 (A Cuide a11d lllterpretation). (London: Routledge).

-- (2003) 'Wittgenstein's Tractatus: True Thoughts and Nonsensical Propositions' Philosophical Investigations: 26/04; 332-47.

Mach, E. (2004 [1906]) Space and Geometry in the Light of Physiological, Psychological and Physical Jnquiry. McCormack, T. J. (tr.). (New York: Dover Publications).

Malcolm, N. (1989) Ludwig Wittge11stein: A Memoir. (Oxford: Oxford University Press).

Malinowski, B. (1923) 'The Problem of Meaning in Primitive Languages' in Ogden, C.D., Richards, LA.: The Meaning of Meaning (Suplement J). (New York: Harcourt).

Manninen, J. (2011) 'Waismann's Testimony of Wittgenstein's Fresh Starts in 1931-3' in Friedrich Waisma1111: Causality and Logical Positivism, McGuinness, B. (ed.); 243-266. (Dordrecht: Springer).

Marion, M. (1998) Wittge11stei11, Finitism and the Fou11datio11s of Mathematics. (Oxford: Clarendon Press).

McGinn, M. (2006) Elucidating the Tractatus. (Oxford: Oxford University Press). -- (1997) Wittgenstein and the Philosophical Investigations. (London: Routledge). Monk, R. (1990) Ludwig Wittgenstein - The duty of Genius. (London: Penguin

Books). Moore, G. E. (1962) Some Main Problems of Phílosophy. (New York: Macmillan). -- (1965) 'Wittgenstein's Lectures 30-33' in Philosophical Papers; 252-325.

(New York: Collier Books). -- (1965 [1925]) 'A Defense of Common Sense' in Philosophical Papers; 32-59.

(New York: Collier Books). -- (1965 [1939]) Proof of an Externai World in Phílosophical Papers; 126-48. (New

York: Collier Books). (1965 [1941]) 'Certainty' in Philosophical Papers; 223-46. (New York: Collier

Books). -- (2007) 'Wittgenstein's Expression 'rule of grammar' or 'grammatical rule"

ln Wittgenstein Jahrbuch 2003/2006. Majetschak, S. (ed.); 201-7. (Frankfurt: Peter Lang).

Moyal-Sharrock, D. (2007) 'The Good Sense of Nonsense: A Reading of Wittgenstein's Tractatus as Nonself-Repudiating' Philosophy 82 (1);147-177.

Mulhall, S. (2007) Wittgenstein's Private Language: Grammar, Nonsense, and Jmagination in Philosophical Investigations, §§243-315. (Oxford: Clarendon Press).

Nicod, ]. (1970) Geometry and Induction. (Berkeley, CA: University of California Press).

Noe, R. A. (1994) 'Wittgenstein, Phenomenology and What it Makes Sense to Say' Philosophy and Phenomenological Research Vol. 54, N 1; 1-42.

Ogden, C.D. and Richards, LA. (1923) The Meaning of Mea11i11g. (New York: Harcourt).

308 Bibliography

Ostwald, W. (1919) Einfuehrung i11 die Farbenlehre. (Leipzig: Reclam). Pascal, F. (1984) 'Wlttgenstein A Personal Memory' in Recollections ofWittgenstei111

Rhees, R. (ed.); 34-9. (Oxford: Oxford Unlversity Press). Pears, D. (1987) The False Prison. Vol. I. (Oxford: Clarendon). -- (1988) The False Prison. Vol. 2. (Oxford: Clarendon). Pichler, A. (2004) Wittgensteins Philosophísche U11tersuchunge11. Vom Buch zum

Album. (Amsterdam, New York: Rodopi). Poincare, H. (1952 [1902]) Science and Hypothesis. (New York: Dover

Publications). (1970 [1905]) 'La Notion d'Espace' in La Valeur de la Science; 55-76. (Paris:

Flamarion). Prado Neto, B. (2003) Fenomenologia em Wittgenstein: Tempo, Cor e Figuracao. (Rio

de Janeiro: Ed. UERJ). Ramsey, F. P. (1923) 'Criticai Notice ofWittgenstein's TractatusLogico-Philosophicus'

Mind 32; 465-78. . (1924) 'Review of C. K. Ogden's and I. A. Richards's the Meaning ofMeaning'

Mind 33; 108-9. -- (1990 [1927]) 'Facts and Propositions' ln Philosophical Papers Mellor, D.H.

(ed.); 1-7. (Cambridge: Cambridge University Press). - (1990 [1929]) 'Philosophy' in Philosophical Papers Mellor, D.H. (ed.);34-51.

(Cambridge: Cambridge University Press). Rhees, R. (1963) 'The Tractatus: Seeds óf Some Misunderstandings' The

Philosophical Review, vol. LXXII; 213-20 Richards, 1. A. (1932) Mendus on the Mind: Experiments in Multiple Definition.

(New York: Harcourt). Ricketts, T. (1996) 'Pictures Loglc, and the Limits of Sense' in The Cambridge

Companíon to Wittgenstein. Sluga, H. and Stern, D. G. (eds); 59-99 (Cambridge: Cambridge University Press).

Roncaglia, A. (2009) Piero Sraffa. (New York: Mcmillan Palgrave). Rothhaupt, J. G. F. (2011) 'Wittgensteins Kringel-Buch ais unverzichbarer

Initialtext seines ,,anthropologisches Denken" und seiner ,,ethnologischen Betrachtungsweise111 Wittgensteín-Studien 2; 137-186.

-- (1996) Farbthemen in Wittgensteins Gesamtnachlass (Weinheim: Beltz Athenaeum Verlag).

Russell, B. (1996 [1902]) The Principies of Mathematics (New York: Norton and Company).

- (1992 [1913]) Theory of Knowledge (The 1913 Manuscript}. Eames, E. R. and BlackweII, K. eds. (London: Routledge).

-(1999 [1914]) Our Knowledge ofthe Ex.tema! World. (London: Rout!edge). (1998 [1919]) Introduction to Mathematical Philosophy (London: Routledge). (1998 [1918]) The Phílosophy of Logical Atomism. (Chicago: Open Court

Classics). - (1997 [1921]) Analysis ofMi11d. (London: Routledge). -(1995 [1927]) An Outline of Philosophy. (London: Routledge).

(2004 [1910]) 'Knowledge by acquaintance and Knowledge by Description' ln Mysticísm and Logic: 165-183. (New York: Dover Publications).

-- (1992 [1911]) 'On the Relations of Universais and Particulars' in Logic and Knowledge Essays 1901-50, Marsh, R.C. (ed.); 103-24. (London: Routledge).

Bibliography 309

-- (2004 [1914]) 'On Scientific Method in Phi\osophy' ln Mysticism and Logic. (New York: Dover Publications).

-- (2004 [1914]) 'The Relation of Sense-Data to Physics' in Mysticism and Logic. (New York: Dover Publications).

(1992 [1919]) 'On Propositions: What they are and how they mean' in Logic and Knowledge Essays 1901-50, Marsh, R.C. (ed.); 283-320. (London: Routledge).

Schlick, M. (1985 [1925]) General Theory of Know/edge. Blumberg, A. E. (tr.). (La Salle, Illinois: Open Court).

-- (2008 [1926]) 'Erleben, Erkennen, Metaphysik' in Die Wiener Zeit: Aufsaetze, Beitraege, Rezensionen 1926-36. Kritische Gesamtausgabe Abteilung I, Band 6, Friedl, ]. and Rutte, H. (eds). (Vienna and New York: Springer).

Schulte, ]. (2002) 'Wittgenstein's "Method"' in Wittgenstein and the Future of Philosophy. A Reassesment After 50 Years. Haller, R. and Puh\1 K. (eds); 399-410. (Vienna:.

Spengler, O. (2003 [1922]) Der Untergang des Abendlandes. (Munich: Deutscher Taschenbuch Verlag).

Sraffa, P. (1963) Production of Commodities by Means of Commodities (Prelude to a Critique of Economic Theory). (Cambridge: Cambridge University Press).

Stalnaker, R. (1976) 'Proposltlons' in lssues in the Philosophy of Language, MacKay, A.F. and Merrill, D.D. (eds); 79-91. (New Haven: Yale Unlversity Press).

Stern, D. (1995) Wittgenstein on Mind and Language. (Oxford: Oxford University Press).

-- (2004) Wittgenstein's Philosophical Investigations: An Inroduction. (Cambridge: Cambridge University Press).

Sullivan, P. M. (2004) 'What is the Tractatus about?', ln Wittgenstein's Lasting Significance, Kõlbel, M. and Weiss, B. (eds);28-41. (London: Routledge).

Venturlnha, N. (2012) 'Sraffa's Notes on Wittgenstein's "Blue Book'" Nordic Wittgenstein Review 1; 1-11 (Pre-print draft).

Von Wright, G. H. (1980) Wittgenstein. (Minneapolis: Unlversity of Minnesota Press).

Walsmann, F. (1965) The Principies of Linguistic Phi/osophy. Harre, R (ed.). (London and New York: MacMil\an, St Martin's Press).

-- (1959) 'How I See Philosophy?' in Logical Positivism Ayer, A.]. (ed.); 345-80. (New York: The Free Press).

-- (1973 [1936]) 'Ueber den Begrlff der Identitaet' ln Was ist logische Ana/yse? Gesammelte Aufsaetze, Reitzig, G.H. (ed.). (Frankfurt: Athenaeum Verlag).

Watson, J.B. (1930) Behaviorism. Chicago: Universlty of Chicago Press. Welner, ]. (2001) 'Theory and Elucidation: The End of the Age of Innocence' in

Future Pasts. The A11alytic Tradition in TWentieth-Century Philosophy. Floyd, ]. and Shieh, S. (eds); 43-66. (Oxford: Oxford University Press).

Whitehead, W. N. (1978 [1929]) Process and Reality. (New York: The Free Press, Macmillan).

Winch, P. (2006 [1969]) 'The Unity of Wittgenstein's Philosophy' in Studies in the Philosophy of Wittgenstein, Winch, P. (ed.); 1-19. (London: Routledge).

Williams, M. (2004) 'Nonsense and Cosmic Exile: the Austere Reading of the Tractatus', in Wittgenstein's Lasting Significance, Kõlbel, M. and Weiss, B. (eds); 1-28. (London: Routledge).

310 Bibliography

Wittgenstein, L. (2009) Philosophical Investigations. German-English edition. Revised translation of Anscombe, G.E.M. by Hacker, P.M.S. and Schulte, J. (Oxford: Wiley-Blackwell).

-- (2001) Philosophische Untersuchungen (Kritisch-genetische Edítion). Schulte, J., (ed.) with the collaboration of Nyman, H., Von Savigny, E., and Von Wright, G.H. (Frankfurt: Suhrkamp).

-- (2004) Tractatus Logico-Phílosophicus. McGuiness B.F. and Pears, D. (trs). (London: Routledge).

-- (1999) Tractatus Logíco-Phi/osophicus. Ogden, C.K. (tr.) with an introduction of Russell, B. (New York: Dover Publícations).

-- (2000) Wittgenstein's Nach/ass. The Bergen Electronic Edition. (Oxford). -- (2000) Wiener Ausgabe. Nedo, M. (ed.)1 Zweitausendeins. -- (2004) Ludwig Wittgenstein Gesamtbriefwechsel/Complete Correspondence

Electronic Edition. Past Masters. McGuinness, B. Unterkircher/A. (eds). (1933) 'Letter to Mind', Mind Vol. 42, N. 167.

-- (1979) Notebook 1914-16, 2nd edition. Anscombe, G.E.M. (tr.) and Anscombe and Von Wright (eds). (Chicago: The University of Chicago press).

--(1979 [1913)) 'Notes on Logic,' in Notebook 1914-16, 2nd edition, Anscombe, G.E.M. (tr.) and Anscombe and Von Wright (eds); 93-107. (Chicago: The University of Chicago press).

(1979 [1914]) 'Notes Dictated to Moore in Norway' in Notebook 1914-16, 2nd edition. Anscombe, G.E.M. (tr.) and Anscombe and Von Wright (eds); 108-19. (Chicago: The University of Chicago press).

-- (1929) 'Some Remarks on Logical Form'. Proceedings of the Aristotelian Society, Supplementary Volumes, Vol. 9, Knowledge, Experience and Realism; 162-171.

-- (1989) Phílosophische Bemerkungen. Werkausgabe Band 2. (Frankfurt: Suhrkamp).

-- (1975) Philosophica/ Remarks. White, R. and Hargreaves, R. (trs). (Chicago: The University of Chicago Press).

(1993) Wittgenstein und der Wiener Kreis. Werkausgabe Band 3. Frankfurt: Suhrkamp).

-- (1979) Wittgenstein and the Vienna Circle. McGuinness, B. (ed.) and Schulte, J. (translators). (Oxford: Blackwell).

(2000) The Big Typescrípt. Nedo, M. (ed.). Zweitausendeins. (2005) The Big Typescript: TS 213. German-English Scholars' Edition. Luckardt,

C.G. and Aue, A.E. (eds and trs). (Oxford: Blackwell). -- (1965) The Blue and the Brown Books. (New York: Harper Torchbooks). -- (1989) Philosophische Grammatik. Werkausgabe Band 4. (Frankfurt:

Suhrkamp). (1978) Philosophical Grammar. Kenny, A. (tr.). (Berkely: University of

California Press). -- (1984) Eine Phí/osophísche Betrachtung. Werkausgabe Band 5. (Frankfurt:

Suhrkamp). (1999) Bemerkungen ueber díe Grundlagen der Mathematik. Werkausgabe Band

6. (Frankfurt: Suhrkamp). -- (2001) Remarks on the Fow1dations o{Mathematics. Von Wright, Rhees and

Anscombe (eds.), Anscombe G.E.M. (tr.). (Oxford: Basil Blackwell).

Bibliography 311

-- (1988) Remarks 011 the Philosophy o( Psychology Vai. I. BillnguaJ edition, Anscombe, G.E.M. (tr.).(Chicago: The University of Chicago Press).

-- (2000) Denkbewegungen (Tagebuecher 1930-2 und 1936-7). Somavilla, 1. (ed.). (Frankfurt: Fischer Taschenbuch Verlag).

(1989) Wittgenstein's Lectures Cambridge, 1930-32. From the notes of John King and Desmond Lee (ed.). (Chicago: Un!versity of Chicago Press).

(1997) Ludwig Wittgenstein: Cambridge Letters. Correspondence w!th Russell, Keynes, Moore, Ramsey and Sraffa. MCGuiness and Von Wright, G.H. (eds). (Oxford: BJackwell Publishers).

-- (1983) Letters to C.K. Ogden. With Comments on the English Translation of the Tractatus Logico-Philosophicus. (Oxford: Blackwell).

-- (1976) Wittgenstein's Lectures on the Foundations o( Mathematics Cambridge 1939. From the notes of Bosanquet, R.G., Malcolm, N., Rhees, R., andSmythies, Y. Diamond, C. (ed.). (Chicago: The University of Chicago Press).

(1997) Zettel. Werkausgabe Band 8. (Frankfurt: Suhrkamp).

Index

a priori, 10-28, 50, 57-60, 63, 80, 82, 89, 112, 120, 124-36, 142, 153, 165-8, 212, 234-7, 251-8, 262

analysis, 10-28, 32-5, 39, 47-50, 81-5, 99, 111, 119-36, 178, 206-14,233-53

ultimate, 13-15 anthropological view, 3-5, 113, 117,

148-66, 170-1, 190-1, 208, 214-221,254-61,289,294-9, 312

anthropology, 164, 289, 296 arbitrarily determined conventions,

80-2, 127-9 arbitrariness, see grammar,

arbitrariness of Augustine, St., 88, 149, 182-4, 221,

294,297-8 Austin, ]., 276 Ayer, A.]., 298

Baker, G., 213, 217, 273, 290, 299 Baker, G. & Hacker, P.M.S., 273, 290,

300,302 begriffsschrift, 17-19, 39, 48, 276-7,

282 behaviorism, 65-6, 69, 287 Bellofiore, R. and Potier, 152 Bouwsma, O. K., 228, 297 Brltton, K., 300 Brouwer, L. E.]., 274, 281

calculus, 56, 60, 303 calculus conception of language, 1-4,

65,93-102, 111-30, 139, 144-53, 160-2, 167, 172-9, 210, 214, 222, 251, 260, 264-70, 283, 287, 290-4,300

Carnap, R., 15, 153, 275, 279, 280, 292 color

incompatibility, 7-35 number system, 35, 281 octahedron, 33-5, 47, 63, 116, 145 wheel, 281

complex, 7, 20, 32-3, 45, 57, 80, 83, 88, 126-7, 132, 191-3, 223, 231-2,238,242-4

and fact, 107, 231-2, 288 complexity, 148-59, 168, 191, 193,

217, 223, 241, 259-60, 276, 295 Conant,J., 273, 293 Conant,]. and Diamond, e., 273 confessional, 106, 192, 217, 299 context principie, 79, 137, 207, 293 conventions

arbitrarily determined, 80-3, 127-32, 252

arbitrary, 60-3, 213-15 Copland, A., 249, 301

definable/indefinable, 15, 82, 125-32, 150, 233, 240-1, 267, 294

definition, 15, 20-1, 38, 72, 81, 111, 125, 129, 145, 149, 153, 196-0 223,229,246-9,280

of number, 57-9 ostensive, 95-8, 129, 150, 157, 184,

215,296 description

kinds of, 14-19 of our practices, 168 of uses, 201-4

dialogical style, 172, 192, 206 Diamond, C., ix, 273, 293 dissolution, 112, 226, 289 distance, 29-38, 42, 143, 280-1 dogmatism, 112, 119-20, 225-6, 236,

295 Drury, M. O'C., 299

exactness/inexactness, 35-41, 49, 110

expectation, 45-6, 63, 67-77, 85-94, 107, 122-4, 166, 285

explanation, 70, 86-7, 93-102, 177-80, 187, 194, 204-5, 222-4, 229,250-1,258-9,294, 302

313

314 Index

facts of natural history, see natural history

family resemblance, 160, 192-205, 231, 246-7, 295

Farrell, B. A., 300 Floyd, ]., 273, 276 Fogelin, R. ]., 273 form of life, 154-70, 191-3, 223-4,

255-6,264,270 Forster, M. N., 273 Prazer,]., 289, 295-6 Frege, G., 18, 57-60, 81, 184-8, 192,

210, 223, 235-8, 247, 256, 276-7, 286,288-9,296-8

game, 3, 57-9, 94-7, 116-31, 139-49, 155, 158, 161, 167

language-, 149-50, 158, 191-3, 201-8, 217, 221-6, 245-6, 260-1

genetic method, see method geometry, 9, 31-42, 54-6, 274-83 gesture(s), 88, 149-55, 217, 293-4

derisive, 154-62, 294 Goldfarb, W. D., 273, 300-1 grammar

arbitrariness of, 60-3 autonomy of, 2, 58, 63, 93-100,

120-4, 150, 167, 283 book of, 49, 53, 115, 130, 229 comprehensive, 1-4, 43-64, 111,

116, 138-42, 151, 170,262-4 see a/so use and grammar

grammatical remarks, 4, 137, 221, 261-73

Hacker, P. M. S., 273, 277, 290, 298, 300, 302

Hanfling, O., 273 Helmholzt, H. Von, 281 Hertz, H., 276 Hilmy, S. S., 284, 296 Hintikka, M.B and Hintikka, ]., 273,

282 Hjelmslev, J., 40, 281 Hylton, P., 275 hypothesis, 16-36, 45-6, 63, 144,

257-8, 274-7, 296

intention, 70, 77-96, 108, 115, 122, 166, 186-9, 286-7, 298

Jacquettte, D., 273 James, W., 284, 298 Johnson, W.E., 187-8, 298

Kenny, A., 273 Kienzler, W., 110, 282, 293 Klein, F., 40, 281 Kremer, M., 273 Kringel-Buch, 289 Kripke, S. A., 109-10, 287 Kuusela, O., 273, 292/

ladder, 138-9, 272 language-games, 149-50, 162, 191-3,

205-8,210224-6,246,297-8 limits of sense/language, 1-4, 6, 13,

18-25, 42-9, 61, 69, 88, 110-11, 116-25, 138-42, 147-60, 171-6, 204,222,246-52,262

logic, nature of, 13-27, 69, 234-41, 279 logical must, 163-70 logical syntax, 10-12, 43, 266, 292 Lugg, A., 273, 293

Mach, E., 279-80 Malcom, N., 151-3, 282, 294, 298, 300 Malinowski, B., 295 Manninen, J., 218, 299 Marlon, M., 276, 282 mathematics, 2, 5, 43, 56-60, 94,

115, 131, 150, 163-8, 178, 215, 230, 232, 256-8, 269-71, 273-4, 279-81,283, 289,290301-2

Maxwell,]. C., 281 McGlnn, M., 273 meaning, causal theory of, 2, 65-92,

100, 118, 141, 146-7, 160, 189, 214,284

Menger, K., 218 metaphysics of the Tractatus, 137,

236-7, 260 method

genetic, 1-5, 54, 99-110, 114-16, 141, 143, 148, 160-6, 170-212, 221-71,286-9,293, 297-9

transcendental, 25-32, 42, 63, 145, 160, 168-9,223,256-8,280-1

minimallsm of the Tractatus, 131-8 mirrar metaphor, 3, 105, 109, 172-5,

190-2,205-6,210,287 mlsleading analogy/símile/

comparison/pícture, 54, 104-11, 156-8, 167, 173-212, 237-60, 287-8, 291, 290 299

monism, 65, 284 Monk, R., 295 Moore, G. E., 134, 140, 160, 172, 187,

207-9,276-0 282-3,292,301 Moyal-Sharràck, D., 293 Mulhall, S., 273 multiplicity, 11, 18, 24, 32-3, 37,

40-5, 63, 72-6, 90, 93, 97-9, 131, 149-59, 167, 223-4, 276, 279, 281-5,294

mythological, 15, 165, 180, 253-4

names, 16, 32-3, 51, 69, 79-89, 95-8, 105, 124-32, 136, 178, 193, 207, 211-12, 231-43, 249-50, 274, 284,294

natural h!story, 162-9, 225, 258 necessity, nature of, 6-12, 22, 43, 115,

263-8 Nícod, ]., 280 Noe, R. A., 276 nonsense, 3, 13, 18, 24-8, 33-4,

39-53, 61-3, 71, 88-9, 117-19, 123-5, 133-40, 158, 173-6,202, 238,250,259-60,268-72,291-3

notation complete/partia!, 44-9, 116-18,

124, 142, 183, 276 conceptual/logical, 10-13, 23, 43,

47-9, 64, 127-8, 132-8, 151, 188, 210,234-70,288-9,292

phenomenologícal, 1-2, 13-42, 49, 169, 275, 303

numbers, general form of, 56-60

Occam's razor, 119, 134, 276 Ogden, C.D. and Richards, l.A.,

65-72, 84,93,270 284, 295-6 Ostwald, W., 279, 281

Pascal, F., 299 Pears, D., 273, 277

Index 315

Phenomena, 2, 7-47, 54, 63, 71, 85, 111, 144-51, 168,275-6,280-1

phenomenological language, see notation, phenomenological

physiognomy of errors, see method Pichler, A., 298 pictorial conception of language, 2,

69-92, 184-90 Plato, 210, 235, 240-3, 298, 301 Poincaré, H., 280-1 Prado Neto, B., 282 primary, 14, 22, 25, 28, 150, 234, 237 primitive language, 3, 151-68, 190-1,

208,210222-3,295 primitive signs, see definable/

indefinable privacy, 22, 277-9 projection, 16-17, 31, 35, 49, 52, 72,

77-96, 131-2, 187-9, 285-7 propositions

elementary, 6-19, 27, 48-50, 57, 69, 81-3, 111-12, 126-8, 133-5, 234-59,285,292

general form of, 13-19, 48, 58-9, 125-7, 133-4, 138-9, 153-7, 223-60,283,289,292

psychoanalysis, 69, 172, 228-9

Quine, W. V., 292

Ramsey, F. P., 2, 6-10, 27, 273-4, 284, 293

recognition, 192-201 reference, see names; simples regress, infiníte, 76-7, 82, 100-6, 184,

189, 198-9 representatíon

perspicuous, 12-19, 47-9, 83, 115-16, 120, 132, 138-9, 169, 229, 260, 267, 276, 303

surveyable, 169-70, 183, 191, 201, 221-31

resoluteness, 4, 138, 273, 293 Rhees, R., 114, 161, 173, 276, 300 Richards, I. A., 298 Ricketts, T., 273, 285

316 Index

Roncaglia, A., 152 Rothhaupt, J. G. F., 279, 289 rule-following, 2, 65, 82-110, 286 Russell, B., 2, 18, 46, 57-8, 65-92,

114-15, 132-48, 153, 160-1, 173, 182, 185, 189, 192,234-41,253, 275-80,284-8,296-301

Schl!ck, M., 113, 119, 213, 218-20, 274,270280,289,299

Schulte, J., 298, 300-1 sense

determination of, 19, 42, 80, 157, 178, 237, 244-7

to give, 175-6 limits of, see limlts of sense

showing/saying, 14-18, 136-8 simples, 12, 81-3, 112, 129, 132, 206,

210,233-44 solipsism, 23, 46, 173-9, 277-9 Spengler, O., 59, 283, 289, 295-6 Sraffa, P., 117-20, 147-66, 172-3, 178,

228-30 Stalnaker, R., 297 Stern, D., 273, 277 Sullivan, P. M., 273 symbolism, 10-13

see alsa notation system

of coordinates, 31-45, 63, 143, 281 of propositions, 49-50, 111, 292 of rules, see calculus conceptlon

tacit conventions, see arbltrarily determined conventlons

tautology, 7-13, 49-50, 127, 134-5, 156, 271, 293

therapy, 172, 226-9, 290,290300 theory of descriptlons, 132, 234, 239,

296 time, 10-19, 27, 35, 54, 282, 291 transcendental arguments, 29-31, 42,

63, 145, 168-9, 280-1, 297-8 see alsa method, transcendental

tribes, 159, 191-3, 208, 217, 270 triviality, 3, 93, 95, 102, 112-20,

134-9, 145, 156, 182, 223, 237, 271, 293

type, see words, kinds/species of

uebersichtliche Darstellung, see representation

understanding, 53, 661 71-2, 82-92, 96-7, 101, 294

implicit, 53, 127, 133, 239-40, ! 249

use and grammar, 169-76, 181-3, 201-5

vagueness, 26, 39-41, 49, 169, 282 see also exactness/lnexactness

Venturinha, N., 179, 228 verlfication

and grammar, 43-51, 143-5 and phenomenological language,

! 13-27 principie of, 27-8, 275, 279-80,

296 Vienna Circle, 111, 213, 275

see also R. Carnap; M. Schlick; F. Waismann

visual field/space, 7-42, 45, 50, 55-6, 142-5, 169, 180,274, 280

Von Wright, G. H., 117, 152-3

Waismann, F., 3-4, 17, 65, 262, 289-90, 299

First project with, 111-12, 282 Second project with, 113, 213-20,

290,294,299 Watson, W. H., 113 Weiner, j., 276 Whitehead, W. N., 301 Wiederaufnahme, 110 Williams, M., 273 Winch, P., 285 words, kinds/species of, 12, 51-3,

140-1, 191-3, 221,283

1

1

1

1

1