Mind, Brain, Behavior: The Mind-Body Problem and the Philosophy of Psychology 9783110883381, 9783110128765

Table of contents :
Preface
Table of Contents
Introduction
I. The Philosophical and Scientific Career of the Mind-Body Problem
II. What is Mind-Body Identity? Reconstruction of the Identity Theory
III. The Contingent Nature of Mind-Body Identity
IV. The Self as Pilot
V. Science and Mind: Explanations of Behavior and Mentalistic Language
VI. Philosophy and Mind: Psychophysical Dualism
VII. Psychology and Consciousness
VIII. Epistemic Limitations of the Mind-Body Relation
IX. The Mind-Body Relation and Human Self-Understanding
Conclusion
Bibliography
Index of Names
Index of Subjects

Citation preview

Martin Carrier and Jürgen Mittelstrass Mind, Brain, Behavior

w DE

G

Martin Carrier Jiirgen Mittelstrass

Mind, Brain, Behavior The Mind-Body Problem and the Philosophy of Psychology

1991 Walter de Gruyter • Berlin • New York

Geist, Gehirn, Verhalten revised and expanded English edition

© Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication

Data

Carrier, Martin. [Geist, Gehirn, Verhalten. English] Mind, brain, behavior : the mind-body problem and the philosophy of psychology / Martin Carrier, Jürgen Mittelstrass. p. cm. Translation of: Geist, Gehirn, Verhalten. Includes bibliographical references and index. ISBN 3-11-012876-4 : DM98.00 (est.). ISBN 0-89925-828-X : $68.00 (est.) 1. Mind and body. 2. Philosophy of mind. 3. Dualism. 4. Psychology — Philosophy. I. Mittelstrass, Jürgen. II. Title. BF161.C3713 1991 128'.2-dc20 91-8492 CIP

Die Deutsche

Bibliothek

—

CIP-Einheitsaufnahme

Carrier, Martin: Mind, brain, behavior : the mind-body problem and the philosophy of psychology / Martin Carrier ; Jürgen Mittelstrass. [Aus dem Dt. übers, von Steven Lindberg]. — Rev. and expanded Engl. ed. - Berlin ; New York : de Gruyter, 1991 Dt. Ausg. u . d . T . : Carrier, Martin: Geist, Gehirn, Verhalten ISBN 3-11-012876-4 NE: Mittelstrass, Jürgen:

© Copyright 1991 by Walter de Gruyter & Co., D-1000 Berlin 30. All rights reserved, including those of translation into foreign languages. N o part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Printed in Germany Typesetting and Printing: Arthur Collignon G m b H , Berlin Binding: Lüderitz 6C Bauer G m b H , Berlin

To Johanna Sophie and Marcus Benedict

Preface This book began with a discussion between Jiirgen Mittelstrass and Ernst Pôppel within the framework of the "Dortmund University Debates" (May, 1986). The discussion was preceded by two lectures — under the thematic heading "Information and Consciousness" — in which the neurophysiologist Pôppel espoused a strict monism and the philosopher of science Mittelstrass a pragmatic dualism. In this book the approach of a pragmatic dualism is further developed in the context of philosophy of science in the direction of an interactionist dualism as a theoretical position and is supplemented by a practical dualism. This is carried out in particular by Carrier's joining of the interpretation of psychological terms as theoretical terms to Quine's conception of the ontological commitments that theories bring with them. Carrier also supplied the demonstration that all monistic positions display the same conceptual and logical structure (reduction structure) as well as the differentiations applied to the concept of physicalism. The systematic ideas of the empiricist Carrier and the constructivist Mittelstrass meet in the methodological concept of explanatory construction. Above and beyond the treatment of the mind-body problem, the analyses and constructions developed here are to be understood as a contribution to the philosophy of psychology. The book was originally published in German in 1989. In general, the present English version closely follows the German text. We have, however, corrected some flaws and introduced some minor modifications where we felt that things had not been spelled out as clearly as they could have been. The only exceptions are sections VII.2 and VII.4 which have been revised more extensively. In particular, we have added a discussion of "connectionism" (VII.2.§ 6) and of the information-theoretic account of psychosemantics (VII.4.§ 6). In these sections, too, the main thrust of the argumentation has remained unaffected. The book owes much to a long-standing cooperation among philosophers and philosophers of science at the university of Konstanz that found a concrete and institutionalized expression in the foundation of the "Center for Philosophy of Science." We would like to thank again

viii

Preface

all the friends, colleagues, and Center co-members who carefully read the original version before it had been completed. Their criticism, recommendations and advice did much to further and improve the manuscript. In particular, we are grateful to Hubert Schleichert and Gereon Wolters for important objections to and suggestions on several parts of the text. Heinrich Kehl's critical remarks and bibliographical hints contributed a great deal to the final version of Chapter III. Intensive discussions with Alexander Riiger led, among other things, to essential improvements in section VIII.2. Gabriele Heister and Peter SchroederHeister carefully examined section V.2. We gratefully acknowledge the benefit of their cooperation. After the publication of the German edition we received various remarks, objections, and comments that occasioned some of the modifications and alterations indicated above. We are especially grateful to Gad Freudenthal for an examination of our account of Aristotelian psychology in 1.1 and to Peter McLaughlin who, while editing the translation, came up with several valuable suggestions. We heartily thank all of them. The translation is the result of a joint effort of Steven Lindberg and Peter McLaughlin, and the final editing was done by the latter (in cooperation with the authors). Erika Fraiss carefully arranged the word processor files and adapted them to printer's needs (which proved to be a much more laborious job than expected in advance). It is thanks to their kind cooperation that this edition could see the light of day. Konstanz, June 1990

Martin Carrier Jürgen Mittelstrass

Table of Contents Preface

vii

Introduction I. The Philosophical and the Scientific Career of the Mind-Body Problem 1. Aristotelian Psychology 2. Descartes and the Aftermath 3. Monistic Research Programs II. What is Mind-Body Identity? Reconstruction of the Identity Theory 1. 2. 3. 4. 5. 6. 7.

The Theory of Type-Type-Identity The Reduction of Theories Routes to Mind-Body Identity The Functionalist Interpretation of Psychological Concepts Functional Materialism and Human Psychology Mind-Body Identity and Eliminative Materialism Monism and Reductionism

III. The Contingent Nature of Mind-Body Identity 1. Psychophysical Identification and Leibniz's Principle . . . 2. Necessary Identity and the Contingent Nature of Psychophysical Identity 3. Intentionality of the Mental and Anomalous Monism . . . IV. The Self as Pilot

1 9 10 16 28 35 36 42 51 58 66 74 78 81 81 92 99 114

V. Science and Mind: Explanations of Behavior and Mentalistic Language 126 1. Reduction Programs in Psychology and the James-Lange Theory 127 2. The Example of Cognitive Psychology: Mental Events and Behavior 131

x

Table of Contents

VI. Philosophy and Mind: Psychophysical Dualism 1. 2. 3. 4.

T h e Logical Status of Psychological Terms The Ontological Status of Mental States Dualism and its Problems Dualism, Physicalism and the Unity of Science

VII. Psychology and Consciousness 1. 2. 3. 4.

142 142 148 154 167 180

The Empirical Basis of Psychology Intentionality and Cognitive Science Intentionality and Syntactic Psychology Psychosemantics and Folk Psychology

181 189 208 223

VIII. Epistemic Limitations of the Mind-Body Relation

251

1. Epistemic Limitations? Principles of Finitism and Godelian Incompleteness 251 2. Epistemic Limitations? The Brain and the Theory of SelfOrganization 257 IX. The Mind-Body Relation and H u m a n Self-Understanding . . 271 Conclusion

283

Bibliography

287

Index of N a m e s

307

Index of Subjects

311

Introduction Schopenhauer called it a "world knot," the mind-body problem, the enigma of the relation between psychological and physical states. In its prevalent philosophical form the knot was tied in the philosophy of Descartes at the beginning of the modern age, and the attempt has been made since then to unravel it - up to the present day. The history of the mind-body problem is also the history of unsuccessful philosophical endeavors to resolve it. Thus it is no surprise that the mind body problem seems to those, who consider it anyway to be a philosophical invention, to be an example of the superfluous character of philosophical endeavors in the border areas of philosophy and science, while for others it provides the proof of the lack of fruitfulness of philosophy. In either case one might share the evaluation of Du Bois-Reymond given in 1880 who placed the mind-body problem - formulated as the question of the origin of sense qualities and states of consciousness — among the seven riddles of the universe and considered it to be insoluble. 1 In the present, too, such resignative positions can be found. Thus Shaffer declares: It may well be that the relation between mind and body is an ultimate, unique, and unanalyzable one. If so, philosophical wisdom would consist in giving up the attempt to understand the relation in terms of other, more familiar ones and accepting it as the anomaly it is. 2

The study presented here shares this resignation neither in respect to philosophy nor to science. On the contrary we shall attempt with new means and methods of philosophy of science to contribute to a solution to this problem that has become a riddle. In the process we want to develop and defend in a novel way the position of interactionist dualism, which takes mental states and events and physiological processes as different in kind but as entities that interact with one another. 3

1 2 3

Du Bois-Reymond 1880, 85, 88 - 89. Cf. Kuhlenbeck 1982, III, 4 2 5 . Shaffer 1 9 6 7 , 3 4 5 . In the following we shall not distinguish conceptually between " m e n t a l , " "psychologic a l , " and "cognitive," nor between " e v e n t , " " p r o c e s s , " and " s t a t e " when related to psychological or mental and physical o r physiological objects. T h u s , we shall speak

2

Introduction

The opposing, monistic interpretation of the mind-body relation is concretized in various versions of the identity theory. This theory holds that brain states and psychological processes (in a sense that will be articulated and clarified) are identical with one another. We consider this position to be meaningful, contingent, but probably empirically false. The questions we are pursuing are basically epistemic in nature. On the basis of the current state of knowledge and research we shall examine which interpretation of the mind-body relation is philosophically and scientifically best supported and most convincingly grounded. We shall use the instruments of modern philosophy of science and shall focus on a reconstruction or explication of the relevant scientific theories. The point of contact is modern cognitive psychology, whose scientific importance for the further development of a philosophy of mind has not yet sufficiently been taken into account. It is characteristic of our approach that we orient the philosophy of mind on the interpretation of the sciences of the mind — especially psychology. This means, philosophically speaking, a demarcation both from phenomenologically oriented philosophical conceptions, in which insights on the nature of the mind are to be gained out of the depth of inner experience of consciousness as well as from those conceptions — mainly fixated on Wittgenstein — that think they have found the key to almost all problems in the relation of mind and body in the analysis of everyday language. We shall neither take the purported wisdom of an inner perception as the basis of philosophical thought nor give the final word to the "language games" of everyday understanding. Nor shall we erect conceptual barriers between science and the philosophy of mind in a fearful attitude of "noli me tangere" or in a jealous guarding of a philosophical preserve — only to wonder later that philosophical reflection remains alone with itself. When philosophy withdraws to problems it can handle exclusively on its own, science plods onward philosophically unimpeded. In opposition to this already traditional division of labor between science and philosophy, usually expressed in the (comfortable) notion that scientific and philosophical modes of research can be mutually isolated, we shall here advocate a philosophy of mind (or of consciousin general of " s t a t e s " and " e v e n t s . " O n the conceptual distinction between event and process, cf. K. Lorenz 1980, 568.

Introduction

3

ness) based on a critical interpretation and a philosophical treatment of the relevant scientific theories. To the extent that the question, whether the course of all events is determined by strict laws or contains essential elements of accident, is to be answered in an interpretation of quantum mechanics and the problem, whether a relational or a substantivalist account of space (or space-time) is appropriate, is to be decided through the explication of General Relativity Theory: to this extent, the possible peculiarities of psychological processes are to be explicated on the basis of an interpretation of cognitive psychology. 4 This means that the relation of mind and body is translated into the problem of the relation between neurophysiology and psychology. With this approach by way of the philosophy of science, we hope to free dualism from the stigma of obscurantism that has attached to it (to a certain extent justly) in the discussion of the mind-body problem up to the present. We take mind to be a natural phenomenon. The core of the question that constitutes our point of departure is thus: what ontological commitments are implied by the acceptance of modern psychological theories, which are essentially characterized by mentalistic vocabulary. These commitments are not as a rule thematized by the scientific theories themselves; their explication is the job of philosophy or philosophy of science. The prerequisite, once again, is that one bases philosophy of mind on the present state of research, on the best available theories — or their foreseeable continuations — and dispenses with the attempt to base arguments on fictitious (philosophical) improvements on this state of research. One would not, after all, orient an analysis of the nature of space on an estimate of the physics of the 25th century. Our plea is thus to take the state of science seriously in the philosophy of mind, too, and to distinguish properly between possibility and reality. In the framework of this approach philosophy of science has a double task. On the one hand, conceptual analysis is needed insofar as it is necessary to clarify the logical structure of the various positions through an analysis of the language of science. On the other hand, a discussion of the criteria for judging the adequacy of claims is needed. Here, it is important to realize that the philosophical interpretation of the various theories cannot simply be deduced from them, but rather itself rests on philosophical presuppositions. These presuppositions are primarily those 4

By this we mean theory explication in the sense of Mittelstrass 1988, 315.

4

Introduction

of a theory of the semantics of science and the method of science, that is, notions of the means by which scientific concepts acquire their meanings and of what characteristics a good theory should display. In philosophy little is self-evident. Therefore, analysis without presuppositions cannot be the goal, but rather analysis supplemented by a reflection about its own presuppositions. Beyond this, let us point out at the start, that for us, too, the relationship of the mind to the body seems unclear and confused, and that this study will not be able to offer any smooth solutions. Philosophy — aside from strong self-convictions that quickly lead to dogmatic philosophical schools — cannot provide firm ground or "ultimate foundations;" and we are not in a position to imagine with visionary powers the ultimate results and to present with deductive stringency the ultimate solution of the mind-body riddle. Accordingly, most of the arguments that follow are characterized by claims to plausibility and by the conscientious weighing of quite legitimate assertions. The reader who feels even after working through this study on the philosophy of mind that he has not attained absolute clarity about the nature of the mental has our sympathy. What we nonetheless hope to have achieved is to have thrown a little (new) light on an old puzzle. The argument will be carried out in the following steps. We begin (Chapter I) with a short survey of the history of the mindbody problem. We trace the origins of some elements of the problem in classical Greek thought but also show that the problem really starts its philosophical and scientific career only with modern dualism, meaning the Cartesian dualism of substances. The road leads from Aristotelian psychology and its scholastic reformulations — by way of Descartes' formulation of the problem and his attempted solution based on dualism — to monistic conceptions (among them Fechner's double aspect theory in the nineteenth century). The history of the mind-body problem issues in the history of modern (empirical) psychology but remains present in the form of cognitive psychology with its philosophical orientations. Chapter II reconstructs the logical structure of the currently most widespread monistic positions, namely the theories of type identity, functionalism, and eliminative materialism. Each of these positions can be treated as the assumption of a particular intertheoretical relation between psychology and neurophysiology. Our reconstruction concentrates on the concept of reduction and attempts to show that all these

Introduction

5

positions can be grasped as assertions of a reduction relation between psychology and neurophysiology. The versions of monism bear a stronger resemblance in their logical structure than is commonly thought. Furthermore, we want to show that such a reduction of psychology to neurophysiology represents the only legitimate ground for an identification of mind and body, and in particular that there is no direct experimental way to justify this kind of identification. In Chapter III we defend the view that the identity thesis is in fact meaningful and contingent, and that it can therefore neither be proved nor refuted from the philosophical armchair. In the philosophical discussion a great number of logical and conceptual objections have been advanced that are supposed to show that the impediments to an identification of mind and body are of a fundamental nature. A large number of these objections are based on Leibniz's criterion of identity, which only allows the identification of two entities if all of their features agree. Psychological and physical phenomena (at least to all appearances) clearly display different features, so that an identification seems to be excluded. We attempt to show that such arguments are ill-supported. We also reject Kripke's objection to mind-body identity, which is based on the premise that in the world of the mental, sensation is the fundamental reality, whereas in the physical world it is not. Finally, we also reject Davidson's reservations about the possibility of psychophysical laws. In Chapter IV we extend the discussion to the dualistic theory of Popper and Eccles. We present and submit to critical scrutiny both Popper's anti-reductionism thesis and the closely connected emergence thesis in the context of Popper's three-world theory, which provides the framework for an anti-reductionist, interactionist dualism. It will be seen that both theses are insufficiently grounded: the anti-reductionism thesis is too narrow and in fact permits a replacement of psychology by neurophysiology; the emergence thesis in the form in which it is presented cannot be protected from being co-opted by monism. The same holds for the attempt to present the identity theory as incompatible with a Darwinist viewpoint. Finally in Eccles' endeavor to ground dualism neurophysiologically, we witness the return of Cartesian dualism in modern terminology based on pseudo-physiological hypotheses ("liaisonbrains"). Nonetheless, it becomes clear that there is a positive connection between this dualistic conception and the conception we advocate: Popper, too, considers the identity theory to be a meaningful and contingent but empirically implausible assertion; and both Popper and

6

Introduction

Eccles advocate the methodological idea of introducing psychological states or events as empirical, testable hypotheses or explanatory constructions. Chapter V is given over to the analysis of some reduction programs in psychology and to the presentation of some aspects of the state of research in modern psychology. We develop an argument here by analogy with the history of science, asserting that psychological reduction programs have too often failed in the course of history to justify (so to speak inductively) the hope of future success for such reductive claims on the basis of this history. We also give a sketch of some recent developments in a special field of modern cognitive psychology. This presentation of psychological theories serves two purposes. It provides the factual base for the philosophical discussion of cognitive psychology in the subsequent Chapters VI and VII. Furthermore, it should make it plausible that judgments about the adequacy of psychological models rest on the same methodological base as do the corresponding evaluations of theories in the natural sciences. In Chapter VI we develop our central argument in favor of a dualistic interpretation of the mind-body relation. It is shown that psychological concepts display all the features of theoretical terms as these are described in the philosophy of the natural sciences. Therefore, since the logical characteristics of psychological concepts correspond to those of natural scientific concepts and since (as shown in Chapter V) the methodological features of psychological theories basically agree with the corresponding characters of the natural scientific theories, it seems natural to interpret psychological and natural scientific entities ontologically in the same manner and thus to ascribe to both of them the same reference to the real world. This speaks in favor of the assumption of an independent level of mental entities, thus providing an essential prerequisite for a dualist interpretation. However, such an inference from mentalism to dualism could be countered by independent grounds, lying either in the internal incoherence of a dualistic position or in its incompatibility with other well-founded views. We attempt to dispel such reservations. The point here is to make it clear that a dualistic interpretation of psychophysical relations is compatible with a scientific view of the world. In Chapter VII we examine the consequences of some aspects of modern psychology for the philosophy of mind. Here we are dealing with the relation of psychological entities to subjectively experienced contents of consciousness. The structure of introspective reports is

Introduction

7

discussed and compared to the structure of observation statements in natural science. There is extensive agreement. We also take up the subject of intentionality, which is attributed great importance in modern philosophy of mind. We discuss the program of cognitive science and of syntactic psychology, arriving at the following results: (1) Intentional characteristics, that is, the semantic features of psychological states are (at present) not specifiable in physical terms. (2) Intentionality is an essential feature of mental states. (3) The content of the mental states considered relevant by psychology can nonetheless diverge from the content of introspectively accessible mental phenomena. The psychodynamics specified by science need not agree with the mental life accessible to our consciousness. In light of these results, doubts once again arise as to the identification of neurophysiological states or events with experienced states of consciousness. Chapter VIII deals with the question of possible limits to knowledge of the mind-body relation. We reject the view that the principle of finitism or Godel's incompleteness theorem stand in the way of the knowability of this relation. We then move to a discussion of the possible impact of new developments in the theory of complex systems on the feasibility of a reduction of psychology to neurophysiology. Our particular focus is on the consequences of deterministic chaos, to which some neurobiologists ascribe an essential function in the workings of the brain. We attempt to show that — if such conjectures turn out to be true — the reduction of psychology to neurophysiology is not feasible even if there is in fact an identity between mind and body. Thus, we would lose the only possible ground for accepting the identity theory. Chapter IX takes a look at the significance of the interpretation of the mind-body relation for human self-understanding. Noting that concepts like "consciousness," "self-consciousness," and "ego" are primarily philosophical concepts that articulate the self-understanding and situational understanding of humans and only secondarily general scientific concepts (with an empirical background), we argue for a cooperative relationship between science and philosophy with regard to the solution to the mind-body problem. Philosophy cannot set itself in the place of science, that is, it should not explain the consciousness that science strives to explain. And science on the other hand cannot set itself in the place of philosophy, that is, it does not describe (explain) the life world that philosophy describes (and explains) with its concepts. Furthermore, science, too, knows that the symbolic representation of the world in life-

8

Introduction

world and science is not identical to the representation of the world in the brain, and that our theoretical representations, too, are thus constructions. Corresponding to the theoretical dualism (of science or philosophy of science) expressed here, there is in philosophy a practical dualism. In a critical analysis of the conceptions of Wittgenstein and Ryle it becomes clear that such a dualism can be close to science, for instance in the form of a practical folk psychology (alongside the folk psychology in an intuitive and in a technical sense as presented in VII.4.2). There exists a cooperative complementarity between the scientific and the life-world philosophical description of man, expressed here in the concepts "consciousness," "self-consciousness," and "ego." Our dualistic interpretation constitutes our motive for developing the ideas on the mind-body problem presented here and also provides the material coherence of the argumentation. Nonetheless, the viability of the philosophical examination of psychology presented here does not depend on the adequacy of dualism. Independent of our intention to help to clarify the mind-body problem on the basis of interactionist dualism, this study also attempts to make a contribution to the philosophy of psychology in general. The philosophical analyses and constructions undertaken in the thematic context of the mind-body problem deal not only with one particular problem of modern psychology, whose roots reach far back into the history of a philosophical psychology, but also with its status as science and furthermore with the scientific status of the social sciences as a whole. This is true especially in questions of methodology. The methodological discussion in psychology and in the social sciences sometimes lags behind the state already reached by the philosophy of the natural sciences. Therefore, — independent of our analytical concentration on the solution to the riddle that the mind-body problem has become — if we have at least succeeded in contributing in a productive manner to the methodological discussion in psychology and the social sciences, we will have achieved one essential constructive goal of this study.

I. The Philosophical and Scientific Career of the Mind-Body Problem Once were the inner and the outer world In harmony united. T h e clear, distinct philosopher Discovered this excited. The inner world (It had a scare) Took refuge in the subject. T h e outer world, of this aware, Retreated to the object. Philosopher was overjoyed In light of this dissension. Forevermore held be employed In this polite profession. (after R . Gernhardt,

Philosophie-Geschichte, 1981) There are problems that only really become visible as problems and occasion scientific research programs after they have found philosophical expression. In this way science gains a philosophical dimension, and philosophy, if it is lucky, gains a realistic relation to its problems. One of these problems is the mind-body problem. In the form influential on philosophy and science the mind-body problem is a consequence of the dualistic construction of reality in Descartes and Cartesianism. This does not mean that there were no reflections before Descartes on the relation of mind and soul, but it does mean that in such earlier reflections the problem did not possess the pressing character that necessitated its scientific treatment and resolution. Whereas before Cartesian dualism, it was possible at least superficially to avoid this exertion by deciding in this case to become an Aristotelian or an atomist, after the advent of Cartesian dualism the situation was completely transformed. Internal and external world had been established; the philosophical and scientific explanation of each of these depended on

10

The Career of the Mind-Body Problem

the solution to the mind-body problem, for so was the relation of the two worlds to one another now formulated. A philosophical research program became a scientific research program in the modern sense.

1. Aristotelian

Psychology

§ 1 The historical background of the problem lies in the notion of a temporally limited connection of body and (immortal) soul which can be traced back to classical antiquity; in fact, Plato in an epistemological context but with allusion to Orphic notions interprets the body as the "prison" of the soul.1 Plato attacks the Pythagorean notion that the soul is the harmony of the body,2 and develops on the one hand a three-tier model of the soul (the highest part being the immortal soul)3 and on the other hand a distinction between life or the soul as the principle of self movement and matter as what is "animated" (8(iV|/u%ov)4 (in some cases) by this principle. At the same time he articulates conceptually the so called hylozoism of Pre-socratic philosophy, which conceived life or the power of self-movement to be a property of matter or the "stuff" of which the things are made. As a philosophical category applied to Pre-socratic philosophy of nature, however, the name "hylozoism" is somewhat problematical, in as much as this philosophy of nature is not characterized more closely by an (explicit) negation of the opposition between animated and inanimate matter or between living and non-living stuff, but rather by the nonexistence of such an opposition. Furthermore the designation "hylozoistic" could also be applied to the Aristotelian conception of simple bodies (within the framework of the Aristotelian doctrine of elements), in as much as these are defined as objects that have a source of motion within them.5 We can see how problematic it is to undertake classifications of historical positions on the backdrop of later conceptions — here, the

1 2 3

4 5

Phaed. 82e, cf. 66d/e; Crat. 400b/c. Cf. Plato, Phaed. 8 5 e - 8 6 d ; Aristotle, De an. A4.407b27 - 408a30. Rep. 439c - 441b; Phaedr. 246a - d, 253c - 254e. On this part of the history of the mind body problem cf. Specht 1980; Mittelstrass 1984b. Cf. Phaedr. 245eff, Laws 896a. Cf. Phys. B1.192bl3 - 14.

Aristotelian Psychology

11

dualistic conception of body (matter) and soul (mind) as well as the Platonic distinction just mentioned. § 2 Accordingly, we find in Aristotle neither a dualistic nor a monistic version of the mind-body problem (monistic, e.g., in the sense of the notion adhered to in ancient atomism and in the Stoa that the soul is material or composed of bodies). 6 Aristotle, on the contrary, emphasizes the unity of body and soul on the background of his theory of form and matter, which — later called hylomorphism — states that finite substances represent a conceptual unity of the "principles" form and matter. For Aristotle, besides the hierarchy of different parts of the soul or powers (8uvd(j.ei m a n y relation (and the corresponding rejection of type-identity). F o r example, the micro-organization of the same person's brain changes in time, making it implausible that one type of mental state could be correlated to precisely one type o f neurophysiological state. M o r e o v e r , there are indications that the organization of different people's brains differs from one another, so that it seems plausible from this perspective that the same mental state is implemented in different neurophysiological ways in different people. Cf. Wilkes 1980, 114-116.

51

Brentano 1874, 8 8 - 8 9 .

68

What is Mind-Body Identity?

logical acts, and only psychological acts, are related to contents in this way; one believes or knows something. Husserl raised the objection that states of pain, although psychological, are not intentional in this sense. This objection can be generalized to all sensory qualities and perception states. Apparently, the phenomenal sensory perception "red" is not directed towards anything. Thus, intentionality would be only a necessary but not sufficient criterion of the mental, unless, one (with Sellars) resorted to giving Brentano's characterization of the mental definitional status, thereby declaring perception states ex vi terminorum to be physical states. On the other hand, Brentano did, in fact, anticipate Husserl's objection and countered with the contention that even pain and sensory qualities were intentional. After all, something hurts, and something is perceived as red. 52 Brentano's thesis about the peculiarities of mental phenomena was transformed into a linguistic claim by Chisholm. Chisholm no longer speaks of intentional acts, but of intentional sentences. Intentional sentences express, as a rule, propositional attitudes and are of the type "A believes that a" or "B thinks of b." The defining feature of intentional sentences is that the truth or falsity of the sentence implies neither the truth or falsity of the content sentence nor that the concept appearing in the content sentence refers to something. It does not follow f r o m "A believes that it is raining" that it is actually raining, and "B is dreaming of Lake Constance" does not imply the existence of Lake Constance. Chisholm's criterion obviously expresses Brentano's intentional inexistence in a linguistic form. Brentano's demarcation thesis then takes the form: Only physical phenomena can be comprehensively described in non-intentional language; intentional sentences can be avoided only at the price of accepting serious limitations on the descriptive power of language. 53 § 2 Functionalism, too, emphasizes the significance of intentional sentences for psychology and this emphasis is based particularly on a consideration of statements about propositional attitudes. Psychological generalizations apply to propositional attitudes in many cases by virtue of the contents of the corresponding propositions. If A perceives the 52 53

Cf. Brentano 1874, 9 0 - 9 1 . On the characterization of intentionality, cf. Sellars 1953, 46 — 47, 74 — 77; Chisholm 1955/1956, 1 2 5 - 1 3 0 ; Bieri 1981, 1 3 9 - 1 4 1 ; Gethmann 1984, 2 5 9 - 2 6 2 .

Functional Materialism and H u m a n Psychology

69

observable feature a, then A also believes that a; if B wishes that b, then B sets out to produce b (as long as nothing prevents him from doing so). The essential point in these schemes is that the same content sentence (i. e., a or b) appears in the corresponding position. Naturally, synonymous content sentences can also be inserted, and this shows that contents are semantically interpreted entities, that they have meaning. By means of their semantic interpretation, contents denote abstract mental states, which cannot be adequately captured by a neurophysiological approach. For example, the simple intentionally formulated statement "C believes that there is a fire and calls the fire department" corresponds to a conglomerate of completely heterogeneous neurophysiological statements. The reason is that C can become convinced there is a fire in a variety of physiologically different ways (by smelling smoke, seeing flames, etc.); and since different systems of perception come into play here, the neurophysiological description must differ in all these cases. In this respect, the neurophysiological approach fails to grasp precisely the characteristic aspect of the mental phenomenon which exists on the abstract level of the semantic contents. The homogeneity of the psychological belief state does not correspond to anything in the incoherent jumble of the neurophysiological pattern of descriptions. 54 Neurophysiological and psychological categories overlap one another in a peculiar way, which will now be described in detail. But first, a few preliminaries. § 3 Every science specifies a certain set of predicates such that events to which these predicates are ascribed fall under the laws of the science in question by virtue of these ascriptions. Thus, certain regularities are stated concerning, for instance, the properties "is a lion" or "is an electron" in the appropriate sciences. This is not true, on the other hand, of properties such as "located within 3 km of the University of Constance." Though it might happen that the property so defined only applies to lions, i. e., the predicate so defined coincides extensionally with the predicate "is a lion;" this would be a mere coincidence and, moreover, an improbable one.

54

Cf. Fodor 1981, 25 - 26; P. S. Churchland 1986, 378 - 380. The problem of the semantic interpretation of mental states will be treated more thoroughly in VII.2 — VII.4.

70

What is Mind-Body Identity?

Predicates of the first type, as opposed to those of the second type, are called natural kind terms. For every natural kind term, there is at least one law which holds for the corresponding phenomena, because they are instances of the property in question. Essentially, this means that there is, indeed, a law involved — i. e., a valid, lawlike statement — and not merely an accidental generalization (as in the case of the lion population of Constance). Natural kind terms are specified by natural laws, natural laws connect natural kind terms with one another. 5 5 The reason for the overlapping of psychological and physical concepts can now be more precisely determined. The truth of intentional sentences, i. e., of sentences about propositional attitudes, is, in most cases, independent of the truth of the corresponding content sentences. Nothing about the truth-value or the reference of a follows from the statement "A believes that a." This peculiarity precisely characterizes Brentano's intentional inexistence. In the same way, the factual equivalence of two entities or circumstances does not permit one to assign the corresponding belief states analogously. "A believes that b" does not follow from "A believes that a" and a*-+b. This makes it clear that psychological natural kind terms — in this case, types of beliefs — are to be conceived of as independent of physical natural kind terms. Types of belief are not modeled on physical types. Despite some physical natural kind terms' being extensionally equivalent, the corresponding beliefs need not be of the same type. The criteria for individuating physical and psychological states do not coincide, and this is the basis of the autonomy of psychological language.56 55

Cf. Fodor 1975, 14. T h e precise terminological determination of "natural kind term" depends accordingly on the clarification of the concept of "lawlike." T h i s involves distinguishing between sentences such as "the charge of all electrons is equal to the elementary charge" and "all the screws in Smith's car are rusty." Only the first sentence would be classified as a law; the second sentence, on the other hand, w o u l d be an accidental generalization. T h e criteria for such a distinction have not yet been strictly and universally clarified, though the following t w o criteria are generally accepted as promising. (1) Confirmability: in the case of lawlike statements, the degree of confirmation increases with the number of cases examined; accidental generalizations, by contrast, can only be based on an exhaustive examination of the realm of application. (2) Support of counterfactuals: laws can (in contrast to accidental generalizations) support subjunctive and contrary-to-fact sentences. Thus, the following statement is true: "if this particle were an electron, then its charge would be equal to the elementary charge." But it is not true that: "if this screw were in Smith's car, then it w o u l d be rusty." O n the concept of lawlikeness, cf. Rescher 1970, 9 7 - 105; Janich/Mainzer 1980, 762.

56

For the elaboration of this viewpoint, cf. III.l, V I 1 . 2 - V I I . 4 .

Functional Materialism and Human Psychology

71

§ 4 Fodor's effort to explain the anti-reductionist position of functionalism (in the intra-species version) and to make it more precise will now be considered against this background. Fodor considers the following conditions necessary and sufficient for theory reduction: (1) Every predicate of the reduced law must be assigned to a predicate of the reducing theory by means of a connecting principle. (2) The connecting principles must themselves be laws, i. e., they must (a) express a relation between natural kind terms and (b) be empirically supported.57 Fodor's condition can thus be stated as follows: The reduction of psychology to neurophysiology demands that psychological and neurophysiological natural kind terms can be assigned to one another in a one-to-one fashion. A reduction claim of this sort, however, is much stronger than a token-identity theory. From the assumption that every psychological phenomenon is also a physical phenomenon, it does not at all follow that every psychological type coincides extensionally with one physiological type. A token-identity theory does not, therefore, imply a reductionism. On the contrary, because of the peculiarities of intentional terminology, it can be supposed that the functional — i. e., the psychological — organization of the nervous system does not correspond to its neurological organization, but rather that the respective types crossclassify one another. 58 Consequently, a reduction claim is not appropriate to this situation. Rather, one must think of the relationship of psychology and neurophysiology in a different way. Each psychological predicate should not be correlated to precisely one physiological predicate ->2 V ... V i>2

neurophysiological laws

Fig. 2 The cross-classifying of psychological and physical types that Fodor assumes can be recognized from this figure. For Fodor the essential point here is that the neurophysiological expression for the psychological law -

»

(

2

)

takes the form $11

V O12 V

...

V ( D i m -»• ® 2 1

V O22 v

...

V ®2N

(3)

and, hence, no longer represents a law. Neither conditions of antecedence nor conditions of consequence can be considered neurophysiological natural kind terms. 61 Fodor also believes it is possible, using this model, to give a satisfactory explanation of the exceptions to the reduced law *Fi —* ¥ 2 . Fig. 2 shows such a case where the predicate ®im is, indeed, connected to but not to 4*2. If one wished to maintain a reduction claim, this could 59

60 61

Cf. Fodor 1975, 20.

Ibid.

Cf. Fodor 1975, 21

Functional Materialism and Human Psychology

73

only be done by conceding that the laws of the reducing theory are equally flawed by such exceptions. Under the assumptions of the present model, according to Fodor, no such concession is called for since expression (3) does not represent a law. 62 § 5 Fodor's treatment of the psychophysical relationship is interesting and informative. Most apparent is that his requirements for a reduction are significantly more stringent than the conditions we have specified. Fodor demands that natural kind terms be assigned in a oneto-one fashion, while we merely require that the terms of the reduced theory be explicitly defined by logical functions of the terms of the reducing theory. Our conception of reduction in no way necessitates that the translation rules (a) establish a one-to-one relation between the predicates in question, (b) are lawlike or even (as Fodor requires in addition) (c) are empirically supported laws. Rather, these translation rules (as already mentioned) merely have the status of definitions. Fodor's sharpening of the concept of reduction is inadequate even when limited to physical situations. If one looks at the reduction of thermodynamics to statistical mechanics, it is obvious that a variety of mechanical states (namely, permutations of molecular energy states) are assigned to just one thermal state (i. e., a particular temperature). A given temperature merely corresponds to a particular energy distribution; how this energy distribution is realized microscopically is not determined (cf. II.2.§ 3). Thus, a fixed value of temperature corresponds to a disjunction of micro-states which are tied together by their functional equivalence. This situation is formally exactly analogous to Fodor's bridge statement (1). From this follows that reductions in physics don't meet Fodor's requirements. In addition, the situation described by Fodor can be easily accounted for by the conception of reduction presented here. The set of translation rules D is represented by definitions of the structure of statement (1); statement (3) is obtained by a translation of the psychological law (2) into the neurophysiological theory. This is obviously a regular Tarskiinterpretation. The problem of possible exceptions to the psychological law is also easily handled in this framework without having to make the special assumption (as Fodor does) that logical functions of natural kind terms no longer represent natural kind terms. To manage this, one need only make use of the concept of relative interpretability. The 62

Cf. Fodor 1975, 2 2 - 2 3 .

74

What is Mind-Body Identity?

validity of the psychological law (2) is to be relativized to the conditon that Oim does not obtain. Fodor speaks in a way that makes it seem as if he precluded corrective reductions and that he required that a reduction deduce the laws of the reduced theory in unaltered form, i. e., including all its shortcomings (namely, the exceptions). In any case, in the framework of the concept of reduction used here, Fodor's psychophysics can be expressed naturally without any additional hypotheses. Another interesting aspect of reduction is revealed by Fodor's example, namely, the asymmetry of reduction. Though the psychological law is, indeed, interpretable in the neurophysiological theory, the neurophysiological law is not interpretable in the psychological theory. The logical content of the neurophysiological theory contains theorems which have no counterpart in the psychological statement. While it remains open in statement (3) which of the Oii's is lawfully connected to which of the ®2j's, this can indeed be derived from the neurophysiological statement (as can be seen in figure 2). One obtains the expression: (Oil

O22) v

(0>12 - » ® 2 N ) V . . . v

(0>1N - > ® 2 i )

(4)

This makes it clear that the neurophysiological theory is logically more stringent here than the psychological law; the former contains statements which cannot be expressed in the latter. This is also a typical characteristic of successful reductions in the sciences. Thus, it is evident that, in Fodor's conception of the psychophysical relationship, the token-identity theory is only compatible with antireductionism because Fodor advocates an inadequately strict conception of reduction or (couched in complementary terms) an extraordinarily weak concept of anti-reductionism. If the requirements are sensibly adjusted, then psychology is, indeed, reducible to neurophysiology in Fodor's model. This means, however, that the identity thesis, even in the weakened form of token-identity — despite all declarations to the contrary — is a version of reductionism. It must make reduction claims, if it is to be substantially defended.

6. Mind-Body

Identity and Eliminative

Materialism

§ 1 In the framework of the type-type-identity theory, it is considered probable that psychological and physiological terms can be correlated one-to-one. Although Feigl, too, favored this option, he also

Mind-Body Identity and Eliminative Materialism

75

examined the possibility of a different way of correlating terms. It could turn out that several physiological processes correspond to just one introspectively identified mental event, because the ability to distinguish events introspectively is limited.63 What Feigl articulates here is apparently the possibility of correcting the linguistically expressed content of psychophysical experience through science. Although two mental events may seem to be of the same kind from an introspective point of view, science can teach us that this is not the case. It is this possibility that eliminative materialism emphasizes and develops systematically. Eliminative materialism arose as a reaction to those objections to mind-body identification which took certain characteristics of mental events to be logically incompatible with features of material processes.64 The elimination theory counters by declaring that, if sensory qualities and intentionality are, in fact, incomprehensible from a materialistic perspective, then the problem lies with talk about sensory qualities and intentionality, not with the materialistic perspective. Eliminative materialism suggests that our usual way of speaking about psychological phenomena should be replaced by an improved, materialistically oriented way of speaking. This view attributes the same status to talk of sensory qualities as to predicates such as "possessed by demons." Just as the latter was shown to be inadequate by the progress of science, our present way of describing psychological phenomena will also be surpassed. Instead of talking about an identity of the phenomenological features of experience or qualia with brain processes, one will speak, more correctly, about an identity between brain processes and that which is now commonly called "mental qualia." If psychological identities are expressed using the psychological categories of everyday life, then they lead to sentences of the same status as the statement "Zeus's thunderbolts are (identical to) electrical discharges." It would be more accurate to say that Zeus's thunderbolts do not really exist, though lightning used to be referred to in that way.65 Eliminative materialism asserts (in Churchland's words): that our common-sense conception of psychological phenomena constitutes a radically false theory, a theory so fundamentally defective

« 64 65

Cf. Feigl 1958, 381. For some of these objections, cf. III. Cf. Feyerabend 1963, 5 6 - 6 1 ; Rorty 1965, 2 7 - 2 8 .

76

What is Mind-Body Identity? that both the principles and the ontology o f that theory will eventually be displaced, rather than smoothly reduced, by completed neuroscience. 6 6

The point of view of common-sense or folk psychology is, therefore, fundamentally false and not to be reduced to anything; it is to be eliminated. Churchland's position, along with the arguments advanced in its support, will be sketched more closely in the following. § 2 Folk psychology is characterized (1) by its use of intentional concepts, and (2) by the fact that it generalizes in its laws over the propositional attitudes that occur in intentional sentences. Typical laws of folk psychology take the following form, for example: VxVp (x hopes that p) A (X discovers that p) —> (x is happy that p)

(1)

VxVpVq [((x wishes that q) A (x believes that (p —• q)) A (X can produce p))] —> (if no other wish gains the upper hand, then x sets about to produce p)

(2)

or:

From the point of view of folk psychology, mental activity consists in dealing with propositional attitudes according to this model. The essential aspect of the eliminative materialistic position is the claim that folk psychology is, in fact, an empirical theory. After all, we explain human behavior by reference to this model; and the statements invoked here (like (1) and (2)) meet the criteria for lawlikeness. They are empirically confirmable and can support counterfactuals. 67 Folk psychology thus characterized as an empirical theory is, according to eliminative materialism, very probably false. First, the empirical failure of this psychology speaks for this negative evaluation. Namely, folk psychology fails to illuminate certain problems such as mental illnesses, sensory deception or the nature of memory. Moreover, it is heuristically unproductive, since it has not made any progress since the time of Sophocles. Finally, it does not fit in with the framework of the theoretical unification of the physical sciences. 68 « P. M . Churchland 1981, 67. 67 For this presentation of folk psychology, cf. P . M . Churchland 1981, 68, 7 0 - 7 1 , 78; Horgan/Woodward 1985, 197; P. S. Churchland 1986, 2 9 9 - 3 0 0 . 68 Cf. P . M . Churchland 1981, 73 — 76. We will discuss the adequacy of this evaluation in VII.4. For a telling critique of Churchland's arguments, cf. Stich 1983, 2 1 2 - 2 1 4 ; Horgan/Woodward 1985, 2 0 0 - 2 0 4 .

Mind-Body Identity and Eliminative Materialism

77

If, however, folk psychology is said to be fundamentally wrong in this way, the only possible recommendation is to abandon it, as with other failed theories in the history of science — such as Aristotelean physics or alchemy — and replace it with a better theory. From the perspective of eliminitive materialism we should not expect a reduction of folk psychology — with its intentional conceptual apparatus — to a future neurophysiology, since folk psychology is said to be fundamentally defective. Theory reduction cannot, therefore, be a reasonable goal of scientific progress, but rather theory replacement.69 § 3 The elimination theory does not, to be sure, militate against psychophysical reduction in general. Only in conceiving of the relation between folk psychology and (future) scientific psychology, does it deny the possibility of reduction — and blames this on folk psychology. On the contrary, the view that the future scientific psychology will be reducible to neurophysiology is vigorously defended. At this level, mindbody identity does exist; here we have an ontological reduction as a consequence of a theory reduction. This reduction relationship is conceived, moreover, more or less in the usual way, insofar as the elimination theory rejects the concept of the cross-classification of psychological and neurophysiological concepts that functionalism puts forward. As presented in II.5, functionalism defends the position that, on the level of neurophysiology, a heterogeneous collection of disparate processes corresponds to the — from a psychological viewpoint — simple category of belief. This stands in marked contrast to the contention of the elimination theory which says that — if it should prove to be true that the concept of belief is a psychologically meaningful category, i. e., that the beliefs characterized by a particular content sentence actually prove to have theoretically meaningful things in common — these common features would also be accomodated on the neurophysiological level. Thus, Churchland expresses the expectation that if there really is a commonality of psychological state in the heads of all who come to believe there is a fire in the building, then there is 69

Although Wilkes shares with an eliminative materialism the view that folk psychology is surpassed by scientific psychology and c a n n o t , therefore, be considered a reasonable starting-point for a psychophysical identification (cf. Wilkes 1 9 8 0 , 1 2 5 ) , she nevertheless believes that folk psychology cannot be adequately characterized as an empirical theory. T h e concepts of folk psychology do not merely serve descriptive o r predictive purposes, but are used in other contexts as well. Cf. Wilkes 1981, 1 4 9 — 1 5 0 . F o r the significance o f folk psychology in non-descriptive contexts, cf. I X . § 5.

78

What is Mind-Body Identity?

every reason to expect that at some appropriate level of neurophysiological organization, this commonality corresponds to a common neurobiological configuration. 7 0

This makes it clear that eliminative materialism sees the relation between scientific psychology and neurophysiology as one of reduction by means of correlating one to the other predicate by predicate (i. e., ^one — ^one). 71 In other words, on this level, eliminative materialism makes the same claim as the type-identity theory.

7. Monism and

Reductionism

§ 1 The reconstruction of three variants of monism presented here — namely, of the type-identity theory, of functionalism (in the form of the token-identity theory) and of eliminative materialism — reveals the significance of the concept of reduction for the philosophy of psychology in general and for the mind-body relationship in particular. For that reason, in conclusion, we summarize once again the essential aspects of the intertheoretical relations that are assumed by these positions. These relations are to be seen from two points of view or in two respects; on the one hand, from the viewpoint of the relation of folk psychology to scientific psychology and, on the other hand, from the viewpoint of the relation of scientific psychology to neurophysiology. Feigl unquestioningly identifies the entities of the everyday, intersubjective psychological manner of speaking with entities of an empirical psychology, as if this were unproblematic. Obviously, he never considered the possibility of a further division of these concepts (cf. II.1.§4). In functionalism, too, an identification is supposed. Psychological laws are interpreted as generalizations about those propositional attitudes which are also assumed by folk psychology (cf. II.5.§ 2). By contrast, eliminative materialism does not presume an identification, it asserts instead that the concepts of folk psychology be replaced by the categories of a future scientific psychology (cf. II.6.§ 1—2).

70 71

P. S. Churchland 1986, 381 (emphasis in the original). T h e Churchlands' general reduction concept also presumes this assignment is predicate by predicate. Cf. P. M. Churchland 1985, 1 0 - 1 1 .

79

Monism and Reductionisra

On the level of the neuropsychological relationship, functionalism's claim — despite its self-assessment — can be reconstructed as a reduction claim (cf. II.4.S 5; II.5.§ 5.). The concepts of the future psychology can be defined by logical functions of neurophysiological concepts in such a way that the psychological laws become translations of neurophysiological theorems. The type-identity theory, moreover, favors a particular form of possible definitions. Namely, they should assign psychological and neurophysiological terms in a one-to-one fashion and, hence, take the form of an equivalence relation (cf. II.3.§ 1). Naturally, the type-identity theory also assumes that the neuropsychological reduction must be limited by certain relativizing conditions, since not all neural phenomena are psychologically relevant. But, within the domain staked out by the relativizing conditions, the translation rules are equivalence relations. This means that definitions of psychological terms by neurophysiological terms are, at the same time, definitions of neurophysiological terms by psychological terms. The type-identity theory predicts reciprocal reducibility of neurophysiology and psychology. Eliminative materialism takes over this radicalized claim of the type-identity theory (cf. II.6.§ 3). Taken as a whole, this results in the following tabular overview of the intertheoretical relations envisaged: type-identity theory

functionalism

elimination theory

reciprocal reduction

reduction

reciprocal reduction

identification

identification

justified theory replacement

neurophysiology

scientific p s y c h o l o g y

folk p s y c h o l o g y

Table 1 § 2 The essential conclusion of this chapter is that all monistic approaches can be reconstructed as reduction claims. The exhaustive — and perhaps a little exhausting — analysis in light of the concept of reduction has paid off: We now have at hand a unified conceptual framework for the entire spectrum of monistic positions. From this framework, we can derive the important result that the structures of the

80

What is Mind-Body Identity?

positions treated here resemble one another much more closely than their respective self-assessments would lead one to believe. Moreover, all these monistic positions not only make claims to reduction in fact but also necessarily have the structure of reduction assertions if they are to be defended substantially. A successful reduction of psychology to neurophysiology is the only legitimate route to mind-body identification (cf. II.3.§ 3 — 5).

III. The Contingent Nature of Mind-Body Identity 1. Psychophysical

Identification

and Leibniz's

Principle

§ 1 The identity thesis, in the reconstruction proposed here, can be understood as asserting the ontological reducibility of psychological to neural entities. This assertion has the status of a sensible, contingent claim whose validity can only be determined empirically. Despite this, it is often assumed that the identity theory could be proved or disproved from the philosopher's armchair. In the following, we want to examine some examples of these arguments, in order to support our assertion of the contingent nature of psychophysical identity. To that end, we must first evaluate the logical-semantic objections to the identity theory. 1 Most prominent among these is the reproach of category mistakes. It is based, as a rule, on Leibniz's criterion of the identity of indiscernibles. The following objection by Cornman is characteristic of this type of argument: In general we accept the principle of the identity of indiscernibles as the criterion of identity. [...] Thus if the Identity Theory is correct, it seems that we should sometimes be able to say truthfully that physical processes such as brain processes are dim or fading or nagging or false, and that mental phenomena such as after-images are publicly observable or physical or spatially located or swift. [ . . . ] This is surely a conceptual mistake. 2

In its modern reconstruction, Leibniz's principle of indiscernibility ("principium identitatis indiscernibilium") contends that the identity of two entities implies the sameness of their respective properties; identity implies indiscernibility.3 In logical language, this means that, for possible properties F, the following equation holds: x = y —> VF(F(x) F(y)). 4 If 1

2 3

4

F o r a discussion of further objections of this type, cf. Smart 1959, 146 —154; von Kutschera 1982, 2 9 8 - 3 0 1 . Cornman 1962, 4 9 0 - 4 9 1 . In logic, the reciprocal case, i. e., deriving identity from indiscernibility, is often denoted as Leibniz's principle. Leibniz presents the theory o f indiscernibility in both an ontological version and a logical version. Cf. K. Lorenz 1984a, 1 8 9 - 1 9 0 .

82

The Contingent Nature of Mind-Body Identity

two entities are identical, and one of them has the property F, then the other must display this property as well. From which the reciprocal can be concluded, namely, if the properties do not coincide, then the corresponding entities are not identical. From discernibility follows nonidentity. Accordingly, Cornman's objection is supposed to show that Leibniz's condition of identity is not fulfilled in the case of the mind-body relation. On the contrary, if one wanted to ascribe to physiological processes the property of confusion (which so often rightly characterizes psychological processes), then this would not only be false, it wouldn't make any sense. It would be a category mistake, comparable to a localization of abstractions. In other words, to speak of confused brain processes would make about as much sense as trying to find where brotherly love is located. If two entities are conceptually so disparately determined, then an identity of these entities is precluded. Polten, too, bases his refutation of the identity theory on Leibniz's criterion. If one takes the criterion seriously, Polten argues, then concepts of different intension never have the same extension; identity of reference only exists in the case of synonyms. Even in Frege's famous example of the extensional equivalence of "evening star" and "morning star," no such equivalence exists in fact, since the first notion denotes Venus under certain conditions in the evening and the second one designates Venus under certain conditions in the morning. 5 In Polten's model, every empirical identification is precluded, since some phenomenal characteristics always remain different. 6 Polten's semantic thesis has several implausible consequences, making it clear (unintentionally), what results from an unlimited application of Leibniz's principle. N o t only is every ontological reduction, such as the identification of temperature and average molecular kinetic energy, ruled out but also every possibility of referring to the same object by different descriptions. For reasons of logic, if one follows Polten, the tree in the garden when viewed from the north cannot be identical to the tree in the garden when viewed from the south. Similarly precluded are all forms of genidentity which express a continual process of development. The Kant of The True Estimation of Living Forces (1746) is in no way identical with the Kant of Opus posthumum. The expressions "the 5 6

Polten 1973, 28. Polten 1973, 8 4 - 8 7 .

Psychophysical Identification and Leibniz's Principle

83

University of Constance on April 30, 1987 at 4 : 1 5 p.m." and "the University of Constance on April 30,1987 at 4 : 1 6 p.m." refer to different objects. In that case, there would be an incessant transmutation of all things; only change would be constant. N o t only the identity theory, but also a good part of all sensible speech falls prey to Polten's Heraclitean semantics. § 2 If one pushes, as Polten does, Leibniz's principle to its extremes, then unacceptable consequences result. Criteria for identity should reconstruct everyday discourse about the identity of people and things, as well as scientific theoretical identifications, not do away with them. The validity of Leibniz's principle must, therefore, be limited; and, in fact, in general usage, its scope is restricted to extensional contexts. Extensional contexts are characterized by the fact that a component proposition of a compound statement can be replaced by another expression of the same extension or the same truth-value, without affecting the truth-value of the compound statement. If such a substitution is not possible, the context is intensional. In intensional contexts, the truthvalue of the corresponding statement is determined not only by the extension of the terms appearing in it, but also by their meaning, their intension. For example, all modal contexts that refer to necessities or possibilities are intensional. Thus, it is necessary that 9 = 9. If one substitutes for "nine" the extensionally equivalent expression "the number of planets," then one obtains the result that the number of planets necessarily equals nine — which is obviously false. In the same way, one can speak of Venus' diameter or its orbit, instead of the diameter or orbit of the morning or evening star; if, however, one substitutes the subject in the statement "the evening star is in the evening sky" with the extensionally equivalent expression "morning star," the statement is false. Obviously, the case of diameter and orbit is an extensional context, while talk of the nature of the evening sky is an intensional context. 7 For theoretical identifications, Leibniz's principle must be limited to extensional contexts. After all, what is at issue in such identifications is always extensional equivalence, i. e., that several statements are true of the same objects. To cite a standard example, man can be described both as a rational animal and as a featherless biped; both descriptions 7

For this presentation, cf. Meehl 1966, 1 6 0 - 1 6 1 ; Fodor 1968, 100; K. Lorenz 1984b, 257; Wolters/Schroeder-Heister 1984, 666.

84

The Contingent Nature of Mind-Body Identity

are extensionally equivalent (even if their meanings differ). T h e logical structure of theoretical identifications is of precisely this kind, even though the latter are differently grounded. Intentional contexts are a special case of intensional contexts. This can be seen in that intentional statements do not allow for substitutions of extensionally equivalent propositions. As presented above (II.5.§ 2 — 3), despite the extensional equivalence of two statements, the truth value of the corresponding intentional sentences can be different. If A knows that there is a university in Constance and if it is true that Constance is the biggest city on Lake Constance (i. e., is identical to the biggest city on Lake Constance), it in no way follows that A knows that there is a university in the biggest city on Lake Constance. Correspondingly, it is not a sound objection to psychophysical identification to point out that the statement "B knows that he is perceiving the color red" may be true, even if the statement "B knows that his R-neurons are being stimulated," which would express the neurophysiological correlate of this perception, is not true. 8 Using this as background, Fodor dealt with the objection that a category mistake is implicit in the identity view, as expressed here by Cornman's argument. Fodor reconstructs claims that statements are meaningful or meaningless as intentional speech of the form "we say that ..." or "we don't say that . . . . " 9 The fact that it makes sense to speak of confused thoughts, but not of confused neural processes, therefore, would merely be understood as a peculiarity of everyday speech that does not necessarily have an objective significance. This reply is already implicit in Feyerabend, who drew attention to the fact that such category mistakes, in the first place only indicate a conflict between the materialistic and the everyday manner of speaking. To turn this conflict into an argument against materialism, it must also be shown that materialistic talk is inferior to everyday linguistic practice. Such a demonstration is not, however, in sight. 10 In fact, it cannot be denied that certain ways of speaking that now seem to be category mistakes could possibly be made unproblematic by a change in linguistic practice. Perhaps the now current statement "yellow light has a wavelength of 580 nm" would also have seemed to be a category mistake to earlier eras, since the phenomenally determined color is identified with 8 9 10

Cf. Smart 1959, 1 4 6 - 1 4 7 . Fodor 1968, 1 0 3 - 1 0 4 . Cf. Feyerabend 1963, 5 0 - 5 3 .

Psychophysical Identification a n d Leibniz's Principle

85

an (obviously colorless) wavelength. Analogously, Cornman's "confused brain processes" might appear equally unproblematic to a future linguistic practice. This example shows that such category mistakes are not peculiar to psychophysical identification. While it is completely sensible to speak of a comfortable temperature, for instance, talking about a comfortable mean kinetic energy would surely have to be considered a category mistake of the kind described here. Another example can be derived from the ontological reduction of ferromagnetic phenomena to the interaction of electron spins. Thus, the sentence "sailors used to orient themselves with the help of magnetic compasses" makes sense, while its micro-theoretical counterpart "sailors used to orient themselves with the help of spin-spin interactions" makes as little sense as talk of confused brain processes. This means, however: If one grants the category-mistake objection to psychophysical identification, then many — if not all — theoretical identifications would be sacrificed as well. If this objection is, in fact, a consequence of Leibniz's principle, then this would surely do more to discredit Leibniz's principle than to argue against the claim of (for example) statistical mechanics to deal with thermal phenomena. § 3 A further type of argument seeks to preclude any psychophysical identification by recourse to specific peculiarities of psychological states. But no one has yet succeeded in clearly stating these peculiarities, i. e., in developing necessary and sufficient conditions for characterizing a state as "psychological" or "mental." 1 1 The "privacy" of psychological events, which appears at first to be a likely criterion turns out on closer inspection not to be a particularly promising candidate. In a certain sense, an exclusive access to our own psychological states does exist in fact, but on this path we acquire, at best, acquaintance of them, not knowledge. In this distinction, which goes back to Schlick, acquaintance (Kennen) is conceived as what is intuitively given to perception; knowledge (Erkennen), on the other hand, is conceived as ordering and classifying by thinking. 12 Only the former, not the latter, is produced by introspection. Conscious awareness does not imply knowledge. We stand to [our own ego] in just the relationship that the mystics so greatly desired for cognition, that of full identity. We are, in the strict 11 12

Cf. D a v i d s o n 1970, 2 0 9 - 2 1 2 ; R o r t y 1982, 333. Cf. Schlick 1925, 8 2 - 8 3 .

86

The Contingent Nature of Mind-Body Identity sense, completely acquainted with it. Hence anyone w h o ignores the distinction between acquaintance and knowledge will have to believe that w e possess absolutely perfect knowledge of the nature of the self, and in fact this is a widely held view. [... But it is false.] Consciousness experiences [...] can become k n o w n in the proper sense of the term only through a scientific psychology, one that classifies and constructs concepts. 1 3

Rorty (though without explicit reference to Schlick) presents the relation of introspection and conceptual knowledge of the mental virtually along the same lines: We no more k n o w "the nature of m i n d " by introspecting mental events than w e k n o w "the nature of matter" by perceiving tables. T o k n o w the nature of something is not a matter of having it before the mind, of intuiting it, but of being able to utter a large number of true propositions about it. 1 4

Introspection does not permit us to recognize the nature of mental states and events; moreover, on this path we often do not even become acquainted with their content. Not only is conceptual understanding of psychological states and events unattainable by introspection, but even simple acquaintance with them cannot be securely obtained in this way. Who can reliably find out just by listening in on himself what really moves him? It is well-known from psychological research that so-called introspective reports commonly merely reflect wide-spread theories and generally accepted prejudices. Moreover, introspection is connected with retrospection, i. e., with forms of a "repeating" representation of mental states and events: Memory colors self-perception. Just as with observation in general, there is also a tendency in selfobservation to see only that which one already knows. This is especially true in explaining the causes and motives of one's own actions; the judgment of others may well be more accurate here than one's o w n . 1 5 Thus, it is not even beyond doubt that introspection is a privileged path — and not simply a special kind of access — to the realm of the mental. Those who are close to us can know us better than we know ourselves. Taken as a whole, this means that mental events are of a special kind insofar as they occur within us. They are subjective primarily in the sense of appearing within a person, and not so much in the sense « Schlick 1925, 84. 14 Rorty 1982, 331. 15 Cf. Griinbaum 1987, 3 8 - 4 8 , 1 2 1 - 1 2 2 .

Psychophysical Identification and Leibniz's Principle

87

of an epistemological characterization. This special localization has no serious epistemic significance. 16 The non-spatiality of the mental is no more suitable as a weapon against identification than the privacy argument. Feigl already emphasized that perceptual impressions are ordered in a phenomenal space and, hence, are spatial in this sense. Colors seem to us to exist outside of ourselves, tastes in our mouths, and pain (for example) in our teeth. 17 For the identity theory, all sensations must also be located in our heads. Thus the theory leads to the conclusion: The mental event a is in person A's head, and this event has the phenomenal characteristic of being outside of A's head. This sounds like a contradiction, but it is not. In physical space, the event is found in A's head; one characteristic of this mental event is that it is located outside of A's head in phenomenal space. 18 § 4 Another logical-semantic objection to a psychophysical identification seeks to build up a conflict between phenomenal qualities and physical quantities and to derive from it the impossibility of extensional equivalence. In response to that, Feigl had already indicated that, on the one hand, phenomenal states or events can exhibit not only ordinal relations (such as color intensity or musical pitch) but also metrical relations (such as estimates of distance or time); they are quantitative in this sense. On the other hand, qualitative differences do, indeed, appear between physical entities (such as those between air pressure and current intensity). It makes sense to speak of mental quantities and physical qualities. 19

16

17 18

19

We will return to the privacy problem in the context of the problem of the empirical basis of psychology in VII.1. Cf. Feigl 1958, 4 0 7 - 4 0 8 ; Rorty 1982, 333. Cf. Meehl 1966, 161 —162. Schlick is also aware of the problem of a contradiction lurking here and attempts to resolve it by distinguishing between phenomenal and physical space. Though he does give priority to phenomenal space: the place of sensory perception is where it is experienced (cf. Schlick 1925, 303 — 307). This leads Schlick to conclude: "the mental is not located in a man's head; rather, a head is itself only a representation in consciousness" (Schlick 1925, 306). Seeing the psychological as equally well localized in one's head, leads Kuhlenbeck to formulate the "brain paradox." It says that the brain on which scientists experiment is itself a manifestation of consciousness, although consciousness itself is produced by the brain. In this sense, the brain it self is a brain phenomenon. Cf. Kuhlenbeck 1982, I, 189 — 190. Feigl 1958, 4 1 0 - 4 1 1 .

88

The Contingent Nature of Mind-Body Identity

Moreover, the objection is clearly based on the idea that some phenomena are quantitative by their very nature, while others are naturally qualitative. In fact, however, numbers do not simply lie around among things, as it were, waiting to be collected. Quantities are not attributes of phenomena, but only of our descriptions of phenomena. Quantities do not divide reality into distinct spheres, but only different approaches to reality. If, in fact, we only had qualitative descriptions of the mental at our disposal, this would be a challenge to develop a quantitatively operating psychology and not an excuse to meditate on the essentially qualitative nature of the mental sphere. This second argumentative strategy, which asserts the existence of allegedly exclusive properties of the mental, is also based, like the set of arguments based on category mistakes, on Leibniz's principle. Properties are given which are said to be solely attributable either to the realm of the psychological or the realm of the physical world, so that the path to identification would be blocked by Leibniz's principle. The critique of these objections lies in showing that the impression of the existence of different properties was deceptive and that it disappears under closer examination. § 5 The "grain objection" and the "constancy objection" are of a kind different from those we have discussed up to this point. Both objections run largely parallel and should, therefore, be analyzed together. The grain objection is already found in Schlick. He emphasizes that psychological experiences such as the perception of a tone or a red surface are simple, while the assigned physiological processes in each case contain an immense number of molecular processes and are, therefore, complex. These antithetical properties preclude any identification. 20 Schlick's reply to this objection is that a partial event of the complex brain process taken as a whole is to be identified with the psychological sensation. This psychological correlate of the sensation may be a very small partial process, one that is extremely simple. The most we can conclude from the objection we are examining is that the process must indeed be a quite simple one. 21

20

21

Cf. Schlick 1925, 320. The argument is also found (without reference to Schlick) in Meehl 1966, 177. Schlick 1925, 321 (emphasis in the original).

Psychophysical Identification and Leibniz's Principle

89

The entire complex process in the brain is required to produce the simple psychological event in the right way and in the right context. 2 2 Schlick's reply is not very persuasive, since the reference to the partial event in the context of the entire process merely veils verbally the complexity of the physical correlate which remains unchanged. Obviously, the corresponding neural context is indispensable for the production of the psychological experience and can, moreover, hardly be distinguished objectively from the partial event that is chosen as the crucial feature. By contrast, Feigl attempts to reconcile phenomenal unity and physical diversity by constructing special neurophysiological models. The essential aspect of these models is the postulation of the existence of threshold values or fusing effects. According to this theory, sensory qualities arise in a region of the brain if the intensity of certain neural patterns in another area of the brain exceed certain threshold levels. Or a region of the brain scans the neural pattern of another, as it were, and merely reflects the average values of the other area. Feigl emphasizes here that he is not interested in the factual validity of these explanations; rather, what matters is that neurological models can be conceived which could get rid of this conceptual obstacle. 2 3 From a philosophical perspective, the essential aspect of Feigl's response is that the simplicity of sensory qualities is implicitly challenged. In fact, i. e., from a neurological point of view, the perception is complex. The appearance of simplicity is deceptive and results from the fact that not all neural processes also have phenomenal correlates, i. e., that only the psychological counterparts of a certain portion of neural events (namely, metacritical values or average values) are accessible to introspection. The simplicity of the perception would, therefore, be considered a deception whose unmasking was more an achievement of neurophysiology than a failure. This argument, which is more or less implicit in Feigl, is found explicitly in Armstrong: I think it can be maintained that although the secondary qualities appear to be simple they are not in fact simple. Perhaps their simplicity is epistemological only, not ontological, a matter of our awareness of them rather than the way they are. The best model which I can give for the situation is the sort of phenomena which the gestait psycholog-

22 23

Ibid.

Cf. Feigl 1958, 4 5 7 - 4 5 8 .

90

The Contingent Nature of Mind-Body Identity ists have familiarized us with. It is possible to grasp that certain things or situations have a certain special property but be unable to analyze that property. For instance it may be possible to perceive that certain people are all alike in some w a y without being able to m a k e it clear to oneself what the likeness is. We are aware that all these people have a certain likeness to each other, but are unable to define or specify that likeness. Later psychological research may achieve a specification of the likeness, a specification which may come as a complete surprise to us. Perhaps, therefore, the secondary qualities are in fact complex, and perhaps they are complex characteristics of a sort demanded by Materialism, but w e are unable to grasp their complexity in perception. 2 4

This rebuttal to the grain objection essentially agrees with an argument developed by Feyerabend. Although the grain of the basic terms in both approaches does not coincide, this in no way has to be true of the defined terms. Complex structures can have simple properties. Thus, for example, density is a simple magnitude that does not reproduce the complexity of the atomic structure of the material in question. This means that by drawing on procedures such as computing average values terms can be defined that have a grain different from the grain of the basic terms used for their definition. 23 The constancy objection is conceived analogously to the grain objection. There are sensory qualities, such as those which arise from observing a uniform spot of color or from hearing a constant pitch, which — at least within a certain time-span — do not change any phenomenal property. The uniformity of perception, however, has no neurophysiological correlate. Neurophysiology teaches us that no timespan and no region of the brain can be chosen small enough that no neurophysiological property changes. It does not help to refer to average values or to show that some complex physiological factor remains constant on the average, since the phenomenal claim is that all properties, not some or even just one, remain unchanged. 26 Fodor seeks to restrict the impact of the constancy objection to the type-identity theory and to show that the token-identity theory is not affected. He argues that an organism could, indeed, remain in a given psychological state and still reveal neurophysiological changes. What is essential is that the neurophysiological states between which the organ24 25 26

Armstrong 1977, 94 (emphasis in the original). Cf. Feyerabend 1963, 53. Cf. Meehl 1966, 1 6 7 - 1 6 8 . This argument is attributed to Sellars.

Psychophysical Identification and Leibniz's Principle

91

ism changes back and forth are functionally equivalent neurophysiological states. The constancy problem can be solved, then, by identifying the psychological state with the whole class of these functionally equivalent neurophysiological states. 27 What is interesting here is that the neural counterpart to a psychologically identified state is of an entirely different type for Fodor than for Schlick. Whereas Schlick aims at a simple, physical partial event as the adequate starting point of a psychophysical identification, Fodor seeks a class of complex physiological processes. If one keeps in mind that both Schlick and Fodor are advocates of the token-identity theory, then it becomes clear that there are significant divergences and differences even within the identity-theory factions. It is also important to note that Fodor's rebuttal does not really meet the objection. The objection aims at the temporal variation within a neural counterpart to a psychological event or state, while Fodor's response refers to the variability between different neural correlates of a mental event or state. Even in the framework of the token-identity theory, a singular neural process must be identical to a singular psychological perception; and, for this pair of singular events, the constancy objection can be formulated in the same way as for types of events. It is more promising to react to this objection with an analogous version of the deception argument and to admit that, in the phenomenal event as well, not all features are "really" unchanged. It is merely introspection's inadequate ability to distinguish that produces the illusion of constancy, while "in reality" change prevails. § 6 Taking recourse to the deception argument has important philosophical consequences. These become more clear if one applies the grain objection and the constancy objection to other theoretical identifications. One can argue in the same way against the identification of temperature with average molecular kinetic energy. After all, temperature can reasonably be considered a simple value, both as it is understood in everyday life and within the framework of phenomenological thermodynamics, while from the perspective of statistical mechanics, it is seen as the result of the effects and counter-effects of an immense number of microparticles. Moreover, in the case of thermal equilibrium, the temperature remains constant even if the kinetic energy of the micro-particles is 27

Cf. F o d o r 1968, 118.

92

T h e Contingent N a t u r e of Mind-Body Identity

subject to constant change. It is not sufficient to point to the constancy of the average value, since — analogously to the psychophysical case — none of the macroscopically accessible values of phenomenological thermodynamics show variations at thermal equilibrium. This peculiarity arises from the fact that statistical mechanics does not simply reproduce phenomenological thermodynamics, but corrects it. From the standpoint of statistical mechanics, not only the simplicity but also the constancy of temperature are, as it were, a deception. "In reality," i. e., from the standpoint of the advanced theory, temperature is a complex magnitude and varies even at thermal equilibrium. Transferred to the psychophysical identification, this argument says that perceptual qualities that seem simple and unchanging are, in fact, not so. This strategy obviously coincides with the deception argument described above and goes beyond it in that the psychophysical identification can only be brought about via a corrective reduction. Thus, mind-body identification would be seen as an ontological reduction via a corrective theory reduction. The essential philosophical aspect of this interpretation is obviously that identification of this type does not lead to the agreement of all properties of the entities in question. Phenomenological temperature and phenomenal perception reveal features that are not attributes of the entities to which they are reduced ontologically, or to which they are supposed to be reduced. Correspondingly, the psychological states that are experienced do not coincide in all their aspects with their neurophysiological counterparts. Nevertheless, identifications of this type do not violate Leibniz's principle, since those properties which are not reproduced by the more advanced theory are dismissed as mere deceptions that in reality do not belong to the entities at issue. 28

2. Necessary

Identity and the Contingent of Psychophysical Identity

Nature

§ 1 In 1970, Kripke developed an argument against psychophysical identity that has since received a lot of attention. Kripke's argument appears paradoxical in that it accuses the identity theory of exactly 28

T h u s , Feigl is right in clinging to Leibniz's principle. Cf. Feigl 1958, 4 4 0 , 4 6 3 .

Necessary Identity and Contingent Nature

93

what it never claimed, i. e., that mind-body identification was merely contingent and not necessary. This objection seems very strange at first, but it makes sense given the background of Kripke's theory of synthetic identifications which attributes to them the status of necessary truths a posteriori. In contrast to physical identifications, this kind of necessity cannot be true of mind-body identity. Kripke's objection thus represents an a priori argument against the identity theory. Kripke's model of synthetic identifications is based on a theory of proper names. So we shall start by outlining this theory. This will be followed by an explanation of its consequences for theoretical identifications in general and then by a discussion of this model's application to mind-body identity. § 2 Kripke's central assertion is that all true statements of identity are necessary truths. This thesis is based on a logical and a semantic argument. The logical argument makes use of Leibniz's principle, and of the axiom that every entity is necessarily identical with itself (Vx • (x = x)). From this equation, one can conclude that the identification of two entities, too, if it is at all true, must be necessarily true (i. e., VxVy (x = y) • (x = y)). 29 Kripke's semantic argument rests on his theory of proper names. According to this theory, proper names are not abbreviations of definite descriptions; i. e., the name "Aristotle" is not to be understood as an abbreviation of a description such as "the last great philosopher of antiquity." Rather, proper names have no intension at all, only extension; they designate something without themselves meaning anything at all. Their extension is, for Kripke, the same in all possible worlds. If an even greater ancient philosopher had lived after Aristotle, then the name "Aristotle" would not designate this later philosopher, but the man we know as Aristotle. This would be true even if Aristotle had not possessed any of the features and characteristics by which we identify him. In this sense, all proper names are rigid designators, i. e., they designate the same object in all possible worlds. 3 0 Kripke's position is further characterized by a particular use of the terms "a priori" and "necessary." "A priori" is, according to Kripke, an epistemological feature and applies to truths that can be recognized 29 30

Cf. Kripke 1971, 136; Kripke 1972, 3. Cf. Kripke 1972, 3 - 9 .

94

T h e Contingent Nature of Mind-Body Identity

independently of experience. Necessity, on the other hand, is a metaphysical property which expresses the idea that this aspect of the world could not have been otherwise. 3 1 Both notions not only do not have the same intension, they also do not have the same extension. True statements of the identity of proper names are, namely, necessary truths a posteriori. True, it must first be discovered that the morning star is identical with the evening star, but once this discovery has been made, it can never be otherwise. 3 2 Kripke's theory of the necessity of nominal identifications is merely a consequence of the terminology he chooses. Proper names are said to be rigid designators, i. e., they designate the same thing in all possible worlds. Necessary truths are true statements that can never be false in any possible world. 3 3 It follows from this that true identity relations between proper names are necessarily true. § 3 Kripke's theory of proper names is plausible — but this is also true of the opposite position. 3 4 In our context, however, proper names are less interesting than theoretical predicates. Kripke's central thesis here is that theoretical identifications, too, have the status of necessary truths a posteriori. T h a t is, it is necessary and not merely contingent that heat be identical with average molecular kinetic energy. 35 " H e a t " is a rigid designator and, in all possible worlds, refers to average molecular kinetic energy. This is not to say, on the other hand, that average molecular kinetic energy is responsible for the sensation of heat in all possible worlds. In fact, one could easily imagine a situation in which there are no creatures capable of perceiving heat. This would not be described as a situation in which there is no heat, but rather as a situation in which no one perceives heat. 3 6 This argument is based on the idea that the phenomenal description of heat (as the cause of the perception of heat) does not establish the 31 32 33 34

35

36

Cf. Kripke 1972, 3 4 - 3 7 . Cf. Kripke 1972, 1 0 7 - 1 0 8 . Cf. Kripke 1972, 109. T h e latter being an advanced variant of the Frege-Russell theory of proper names. An exemplary elucidation of it is found in Searle 1958, 166 — 173. It would be more accurate in this context to speak of "temperature" and not "heat." Kripke rejects this, however, because he anticipates complications because of the possibility of alternative temperature scales (Kripke 1972, 129). We adopt Kripke's terminology in this section. Cf. Kripke 1972, 1 2 9 - 1 3 0 .

Necessary Identity and Contingent N a t u r e

95

meaning of " h e a t , " but only the reference of this term. O n e d o e s not yet k n o w what the nature of heat is when one merely s p e a k s of the perception of heat; the latter is a contingent property of heat. By contrast, molecular motion is an essential property of heat. A phenomenon that lacks this essential property w o u l d not be called " h e a t : " We have discovered a phenomenon which in all possible worlds will be molecular motion — which could not have failed to be molecular motion, because that's what the phenomenon is. On the other hand, the property by which we identify it originally, that of producing such and such a sensation in us, is not a necessary property but a contingent one. This very phenomenon could have existed, but due to differences in our neural structures and so on, have failed to be felt as heat. 37 Hence, scientific characterizations are essential; they uncover the essence of the phenomenon. O n c e this true nature is recognized, no counterfactual situation is conceivable in which we would deprive the p h e n o m e n o n of its essential properties. For then it would be a different phenomenon. Hence, this p h e n o m e n o n possesses its essential properties in all possible worlds, i. e., necessarily. T h e apparent contingency of the identification of heat and molecular motion results f r o m the fact that the reference of " h e a t " is determined by a contingent feature (namely, the perception of heat). T h i s explains why a misleading impression of contingency arises despite the actual necessity of this identification. 3 8 Kripke's position can be reconstructed as embracing the following principles: (1) Realism thesis: T h e r e are natural, essential properties of things and of p h e n o m e n a . T h e essence of things and of phenomena is given by nature. (2) Essentialism thesis: T h e essential properties of things and of phenomena are discovered by science. Science brings a b o u t a constant accumulation of knowledge about the essence of things.

37

38

Kripke 1972, 133 (emphasis in the original). For the full argument, cf. Kripke 1972, 131-133. T h i s argument reveals that Kripke applies Leibniz's principle only to essential properties. T h e trans-world identity of an entity, i. e., its identity in all possible worlds, merely requires that the essential properties agree, while the inessential properties can be different. T h i s type of argument resembles our discussion of the deception argument (cf. III.l.S 5) and demonstrates once again that Leibniz's principle should not be applied indiscriminately.

96

The Contingent Nature of Mind-Body Identity

(3) Contingency thesis: Alongside the essential properties, there are features which are not attributes of things and phenomena by nature but are simply ascribed to them. Among these are the phenomenal features that describe the effect of these entities on our senses. (4) Semantic thesis: The essential properties determine the intension: the contingent features fix, at most, the extension. (5) Conclusion: All essential properties, and only these, are necessary. Alteration of essential properties is precluded, since we would then be talking ipso facto about another phenomenon. But this is not true of contingent features. Thus, it is indeed necessary that heat is identical with molecular motion, but it is not necessary that heat warms us. Kripke's view appears us not very plausible. The realism thesis is problematic, and there seems to be no way to support convincingly the essentialism thesis. But it is not our goal to criticize Kripke's model of theoretical identifications. The only question that we want to deal with here is why, according to Kripke, psychophysical identification cannot be of this type. § 4 Kripke discusses the problem of mind-body identity using the example of the connection of the psychological phenomenon of pain with its neural correlate, which is said to be the stimulation of "Cfibers." If the identity theory were correct, then stimulation of C-fibers would have to be the essential property of pain (just as molecular motion is the essential property of heat). The essence of pain would be determined by C-fiber stimulation and both would necessarily be identical to one another. This is, however, at least apparently not the case. Pain and C-fiber stimulation seem to be conceivable independently of one another; thus, their identity is, at best, contingent. All of this is obviously also true in the case of thermokinetic identification, since, there too, molecular kinetic energy and heat at least seem to be different. The decisive difference, for Kripke, is that the appearance of contingency can be explained away in the thermokinetic case and shown to be a deception. It can be shown that heat and molecular kinetic energy only seem to be different because the referent of "heat" is determined by the contingent property of heat perception. In the case of pain, this possibility does not exist. The perception of pain is the essential property of pain; the essence of pain is determined by its phenomenal quality. Unlike the case of heat,

Necessary Identity and Contingent Nature

97

whatever is felt as pain, is pain. The contingency of psychophysical identity can no longer be explained away as a deception; it exists in fact. But since all true identifications are necessary, the physical and the psychological are not identical with one another. 3 9 The pivotal point of this argument is that Kripke abandons the essentialism thesis for psychological phenomena: The essential properties of mental phenomena are determined not by science but by introspection. In other words, Kripke changes from realism for the physical domain into a phenomenalism for the domain of psychological entities. The soundness of his argument depends crucially on whether this shift is justified. First of all, it is evident that Kripke's argument cannot be generalized for all psychological states. In particular for intentional states (such as believing or knowing), reliable introspective access is not at all above suspicion (cf. III.1.§ 3). "For we seem to have lots of beliefs we don't know we have," 4 0 as Rorty correctly remarks. But then one might well think that sometimes only the judgment of others or the judgment of science teaches us what we really believe. T h a t introspection, as Kripke claims, reveals the essence of our inner life, would be a sound assumption, at most, for sensory qualities and not for intentional states. In fact, however, even in the case of sensory qualities, it is doubtful that Kripke is right about his psychological phenomenalism and the resulting claim that statements about one's own mental events are incorrigible. Doubts of that kind were already expressed by advocates of the identity theory even before Kripke conceived his argument. The central objection is that the incorrigibility thesis assumes that one's own sensory qualities are always correctly designated. However, it is possible to imagine situations in which this assumption would be abandoned in order to rely more upon neural indicators than on one's own perceptions. Thus, Rorty discusses the case where the neural correlate of perceiving pain (using Feigl's autocerebroscope, for example) could be found in a person who thinks he does not feel pain. The most obvious suspicion would be that this person does not know what pain is, that he uses the word "pain" differently than his fellow-creatures. 41

39 40 41

For this argument, cf. Kripke 1972, 1 4 8 - 1 5 3 . Rorty 1982, 333 (emphasis in the original). Cf. Rorty 1965, 43; also Aune 1966, 22.

98

T h e C o n t i n g e n t N a t u r e o f M i n d - B o d y Identity

Meehl employs the same argument in more detail. Meehl discusses a fictitious autocerebroscopic experiment with the perception of color, where incorrect identifications of colors occur in statistically significant numbers. In this case, one's own subjective classification of a color does not agree with neural indicators, nor with the color perceptions of others. Although one is firmly convinced that one is seeing green, for example, occasionally it happens that the physiological state of the vision center of one's brain indicates that one is seeing red. Moreover, it is assumed that a physiological examination reveals that the connection between the vision and speech centers are disturbed in such a way that a statistically significant number of perceptions are communicated incorrectly. T h a t is, although the vision center is, in fact, in the state "red," the speech center receives the message "green;" thus, one incorrectly believes that one is seeing green: It is therefore literally correct to say what many philosophers have considered nonsense, namely, sometimes my raw feels seem to be green when they are in fact red. 42 What Meehl describes here represents precisely the possibility that Kripke excluded outright, namely, the rejection of phenomenalism in the domain of sensory qualities. Correspondingly, it is in no way absurd to draw a parallel between this situation and Kripke's presentation of the thermokinetic identification, i. e., to assume that the criterion for the existence of sensory qualities is the neurophysiologies state, while the subjective perception determines the referent merely contingently. Even for psychological entities, one could defend the view that their essence is revealed by science, i. e., by neurophysiology. 43 This entails that psychophysical identity is necessary in the Kripkean sense. We must emphasize here that we ourselves are not arguing for this sort of necessity; on the contrary, we are arguing vigorously against it. T h e purpose of this discussion is merely to show that if Kripke's model of theoretical identification is accepted for the sake of argument, psychophysical and thermokinetic identity can be accommodated to the same extent. This implies that Kripke's reasoning fails to demonstrate their fundamental difference. 42

M e e h l 1966, 119; for the c o m p l e t e argument, cf. M e e h l 1966, 1 0 6 - 1 0 7 , 1 1 6 - 1 1 9 . T h e same argument is also found in A r m s t r o n g 1 9 6 8 , 1 0 8 - 1 1 0 . F o r a further discussion o f this p r o b l e m , cf. V I I . 1 .

43

T h i s is, in fact, precisely the position o f eliminative materialism (cf. II.6).

99

Intentionality of the Mental and Anomalous Monism

3. Intentionality of the Mental and Anomalous

Monism

§ 1 All the views discussed up to this point have tried to ban the identity theory to Hades by means of semantics and logic. By contrast, Davidson's influential argument uses the same means to raise the identity theory to the Olympus of philosophy. T h e object of Davidson's a priori proof is the token- and not the type-identity theory: It is supposed to show the identity of particular psychological states or events with particular physical states or events. Davidson does not aim at a theory reduction — and thereby an explanation — of mental phenomena by means of physiological laws, rather he explicitly rules out such a reduction. 4 4 He focuses solely on the identification of individual mental events with individual physical events. Davidson bases his argument on the following principles: (1) There states (2) Every (3) There

is a causal interaction between psychological and physical or events. causal connection must be based on a strict law. are no strict psychophysical laws. 4 5

From these principles, the token-identity theory is derived by the following reasoning: Assuming that a particular mental event caused a particular physical event, e. g., that the desire for strawberry ice cream led to the fulfillment of this desire. Both events, i. e., the desire and the corresponding trip to the ice cream parlor, must be connected — following (2) — by a strict law under some description. This law, however, cannot — following (3) — be a psychophysical connection and must therefore represent either a psychological or a physical law. Since only physics, and not psychology, can claim to afford a complete explanation — since it is more plausible to assume that the desire is a physical state than to presume that the trip to the ice cream parlor is a mental event —, one concludes that it is, in fact, a physical event. 4 6 Davidson's surprising and original argument seeks to base this ver— sion of the identity theory precisely on nomological independence instead of, as is usual, on nomological dependence — of mental and

44 45 46

Cf. Davidson 1970, 207. Davidson 1970, 2 0 7 - 2 0 9 . Davidson 1970, 223 — 225. Cf. the reconstruction of Davidson's argument in McGinn 1980, 1 8 5 - 1 8 6 .

100

T h e Contingent Nature of Mind-Body Identity

material events. Thus, Davidson characterizes his position as anomalous monism and does indeed consider all mental events to be physical ones, although not physically explainable.47 The core of this view is obviously principle (3), which asserts the impossibility of psychophysical laws. Davidson takes this impossibility to be based on the fact that physical and mental events can only be understood by means of essentially different theoretical structures. This difference precludes any reduction of psychology to a physically oriented theory such as neurophysiology. Before we reconstruct Davidson's argument in detail, we should first point to the various ways of applying a priori arguments of this type. Thus, for example, McGinn replaces Davidson's methodological argument against psychophysical laws with Kripke's ontological argument against such laws. In the context of Davidson's reasoning, Kripke's argument is no longer evidence for dualism, but rather for the tokenidentity theory.48 If one denies the premise of the completeness of physics in Davidson's proof, then the result is a strict demonstration of panpsychism. Moreover, a parallelistic dualism which rejects any psychophysical causation (cf. I.2.§ 3) remains completely unaffected by Davidson's argument. The parallelist could take the difference that Davidson claims exists between psychological and physical theoretical structures as evidence for his own dualistic convictions. This makes it clear how unspecific and ambiguous, how inconclusive such a priori proofs are even if there were no doubt about their validity. § 2 Returning to Davidson's argument against psychophysical laws, the first thing to note is that mental phenomena are characterized, for Davidson, by intentionality.49 The argument can be applied to psychology only insofar as it deals with entities such as beliefs and goals. 50 Furthermore, it must be emphasized that the argument is only intended to preclude the existence of lawlike connections. 51 Accidental psychophysical generalizations, on the other hand, are entirely possible. The question whether the latter actually exist cannot be the subject of philosophical argument, but solely of empirical research. Davidson's claim is, then, that linking a psychological predicate to a neurophysiolog47 48 49 50 51

Cf. Cf. On Cf. On

Davidson 1970, 214. McGinn 1980, 1 8 6 - 1 8 7 . the concept of intentionality, cf. II.5.§ 1 — 2. Davidson 1970, 2 1 0 - 2 1 1 ; Davidson 1973, 246. the concept of "lawlike," cf. I1.5.§ 3.

Intentionality of the Mental and A n o m a l o u s M o n i s m

101

ical predicate can only be accidental, i. e., it can at most have the epistemic status of the extensional equivalence of rational animals and featherless bipeds. For Davidson, the impossibility of psychophysical laws rests on the disparate commitments of the mental and physical schemes. It is a feature of physical reality that physical change can be explained by laws that connect it with other changes and conditions physically described. It is a feature of the mental that the attribution of mental phenomena must be responsible to the background of reasons, beliefs, and intentions of the individual. There cannot be tight connections between the realms if each is to retain allegiance to its proper source of evidence. [...] when we use the concepts of belief, desire, and the rest, we must stand prepared, as the evidence accumulates, to adjust our theory in the light of considerations of overall cogency: the constitutive ideal of rationality partly controls each phase in the evolution of what must be an evolving theory. [...] We must conclude, I think, that nomological slack between the mental and the physical is essential as long as we conceive of man as a rational animal. 5 2

The pivotal aspect of this argument lies in the assumption of a "constitutive ideal of rationality" which is supposed to determine the theoretical structure of psychology but not that of physics. When we attribute to a person intentional states, i. e., attitudes and goals, then we must base this attribution on the principle of rationality, insofar as we conceive of this person as "a rational animal." On the other hand, if lawlike psychophysical correlations existed, then the knowledge of the neurophysiological state would be sufficient to attribute the corresponding psychological state. In that case, then, a mental state could be attributed to a person without reference to the principle of rationality, thereby missing the realm of the mental. One would no longer be referring to intentional states and, thus, the topic would have been changed. In general, this means: Both psychology and physics are characterized by definite, constitutive principles that are in conflict with one another. A lawlike correlation between the two would lead to the transfer of these principles of each into the realm of the other, thereby leading to incoherence. 53 If one seeks to reconstruct a system of beliefs and motives from a person's behavior, then this can only be done adequately by

52 53

D a v i d s o n 1970, 222 - 223. We are following the reconstruction of Davidson's argument in K i m 1985, 376 — 381.

102

The Contingent Nature of Mind-Body Identity

recourse to reasons, i. e. to rationality, not simply by blindly applying physiological indicators: in inferring this system from the evidence, we necessarily impose conditions of coherence, rationality, and consistency. These conditions have no echo in physical theory, which is why we can look for no more than rough correlations between psychological and physical phenomena. 5 4

A reduction of psychology to neurophysiology is impossible, since psychological laws would have to be explained by physiological laws. This fails, however, because the conditions for attributing states are fundamentally different in the two realms, and the psychophysical abyss cannot, therefore, be bridged by any laws. Apparently, Davidson's concept of reduction is similar to Fodor's concept (as discussed in II.5.§ 4) in that it requires the connecting principles to be lawlike, while the model we developed in II.2.§ 2 considers them merely as definitions. Our model would, therefore, permit an explanation of psychology by neurophysiology even without using psychophysical laws. We shall not discuss the appropriateness of Davidson's concept of reduction, nor shall we examine more closely the adequacy of the claim that the mental is characterized by intentionality in every psychological theory and in all psychological laws at every level of generality (cf. VII.3). Instead, we shall merely look into the adequacy of Davidson's principle of rationality. § 3 Examining the adequacy of Davidson's concept of rationality is made more difficult by the necessity of first reconstructing its content. Davidson's own efforts to elucidate that concept are always sketchy and sometimes even empty. This regrettable fact has led a number of his interpreters to try to fill in the gaps in his arguments. Hence, we must start, by examining some of the suggested interpretations in greater detail. One of these suggested interpretations conceives of the rationality principle as a norm: Attributing propositional attitudes is done by considering which attitudes a person should have or which actions he

54

Davidson 1974, 231.

Intentionality o f the M e n t a l and A n o m a l o u s M o n i s m

103

should perform, given a particular system of beliefs. 55 This interpretation is supported by Davidson's comment that: One way

r a t i o n a l i t y is b u i l t in is t r a n s p a r e n t :

the cause must be a

b e l i e f a n d a d e s i r e in t h e l i g h t o f w h i c h t h e a c t i o n is r e a s o n a b l e . 5 6

According to this interpretation, "rational" is meant in the normative sense of "reasonable." One problem with this approach — apart from the difficulty of saying what is "reasonable" — is revealed by the well-known fact that there are always two reasons for an action: a good reason and the real reason. A normatively interpreted principle of rationality obviously provides the good reason, not the real reason. This would mean a turn from empirical science to practical philosophy. Psychology would be understood as applied ethics. Moreover, as Fodor pointed out, the attribution of propositional attitudes is occasionally based on precisely the opposite assumption, namely, that the corresponding person does not think entirely rationally — in the sense of "reasonably" and "adequately." Such a procedure is always applied when we want to trip someone up or set a trap for him (even if it is only in ethically unobjectionable contexts such as a chess game). In such cases, one hopes that the other does not notice or does not understand the trap we have set for him, or believes that there is no escape. In cases of this sort, intentional discourse is based on the expectation of a violation of the rationality norm. 57 Another reconstruction of the principle of rationality is to demand the transitivity of preferences — if a person prefers b to a and c to b, then he also prefers c to a — or by requiring that decisions suit preferences, or that the system of beliefs is deductively closed, i. e., that the obvious consequences of a belief are themselves beliefs. 58 It is certainly questionable whether these principles really hold true of actual (and not just of ideal) persons. But even if their validity is 55 56 57

58

C f . M c L a u g h l i n 1985, 3 5 5 ; K i m 1985, 3 8 3 . Davidson 1974, 2 3 3 . C f . F o d o r 1981, 106. F o d o r concludes this reflection with the r e m a r k : " W e do not play chess on the assumption that our o p p o n e n t will m a k e the o p t i m a l move; indeed (a small point) in any g a m e much m o r e c o m p l i c a t e d than, say, t i c k - t a c k - t o e , o n e is unlikely to have a clue as to w h a t the optimal move is, or even, indeed, whether o n e exists. T h e intentionality o f one's o p p o n e n t ' s propositional attitudes is not thereby i m p u g n e d " (ibid.). Cf. K i m 1 9 8 5 , 3 7 6 , 3 8 0 .

104

T h e Contingent N a t u r e of Mind-Body Identity

assumed, these principles have "no echo in physical theory" (to use Davidson's phrase) only because physics does not deal with beliefs. In a more general sense, however, this difference is not beyond doubt. After all, physical theories are more likely to be deductively closed than systems of beliefs, length measurements are more likely to be transitive than preferences. And insofar as beliefs and goals which cannot be directly observed must be attributed in such a way that they correspond to observable actions, this merely means that the states which are assumed for theoretical reasons must be specified in such a way that they correspond to the known data. And this trivial demand should also be satisfied in physics. Unobservable potentials, for example, are specified such that they lead to the correct values of observable field strengths. In this procedure we can hardly see a reason for giving special status to the mental. Another reconstruction sees the core of Davidson's principle of rationality in the idea that actions are intuitively intelligible. The condition of intelligibility determines the psychological interpretation of behavior, not however, the neurophysiological description of the brain. Accordingly, psychology seeks to understand meanings and, in this sense, represents a hermeneutical enterprise. 59 This interpretation, too, can be directly supported by Davidson's statements. In interpreting behavior, namely, we would have to assume a pattern of beliefs and motives which agreed with our own to a degree sufficient to build a base for understanding and interpreting disagreements. 6 0

While physical states are connected by blind regularities, the sequence of mental events must always satisfy the condition of rational intelligibility. It must be said here that the fundamental distinction between explaining nature and understanding humans does justice neither to physics nor to psychology. Einstein, for example, expresses his belief in the notion of "the reason that manifests itself in nature." 6 1 Not only men, but nature, too, is thought to be rational. It is this belief in the rationality of everything real which, for example, determines the high esteem of the scientific community for variational principles. Such variational 59 60 61

Cf. M c G i n n 1980, 187; Kim 1985, 383. Davidson 1973, 259. Einstein 1934, 5.

Intentionality of the Mental and Anomalous Monism

105

principles specify entities that become extremal for processes that actually occur (extremal principles). Examples are Fermat's principle (which says that light travels by the path which takes the shortest time) that deeply impressed Leibniz, 6 2 and the principles of Lagrange, Hamilton and Jacobi according to which certain functions of energy take on extremal values in actually realized processes (as compared to alternative courses of the same processes). Variational principles formulate a kind of economic principle for natural processes and this is often taken as an indication that nature has a rational structure and order which is intelligible to man. Conversely, it is quite problematic to regard the attribution of psychological states as being necessarily determined by "empathetic understanding." This view is based on Dilthey's conception of the humanities, whereby the humanities are distinguished from the natural sciences in that they obtain a direct knowledge of their subject matter by empathetically comprehending mental processes, while the natural sciences merely determine connections between states and events "from the outside." 6 3 Such notions, which are methodologically problematic and burden modern epistemological research more than they assist it, no longer adequately reflect the methods of modern psychology (of which we shall give a sketch in V.2). Current psychology aims at causal laws and predictions of behavior and, in that sense, equals the natural sciences. Though intelligibility and understandability do, indeed, play an important heuristic role in formulating laws, this is also true of the natural sciences. Moreover, principles such as intelligibility and understandability cannot be a substitute for an independent test of validity. 64 § 4 Up to this point, we have presented different reconstructions of Davidson's argument by various interpreters and examined the extent to which they are viable. However, in our opinion they are all substantially too limited or even flawed to count as adequate reconstructions of Davidson's position. Therefore, we shall suggest another reconstruction and discuss its viability. In our view, Davidson deems psychology a hermeneutic enterprise, to be sure, but he considers the term "hermeneutics" to be inadequately

62 63 64

Disc. met. § 2 2 ( = Leibniz 1686, 56/57); cf. Leibniz 1 8 7 5 - 1 8 9 0 , IV, 318. Cf. the detailed presentation in Patzig 1973, 45 — 75. Cf. Patzig 1973, 59-61.

106

T h e Contingent N a t u r e of Mind-Body Identity

defined by explanations such as "intelligibility" or "comprehensibility." T h e crux of hermeneutics is rather that it is only possible to interpret behavior, if one a s s u m e s that beliefs, goals and actions agree, i. e., if one assumes a coherent system of psychological states, and, in addition, supposes that the beliefs are to a large extent accurate, i. e., that they are true. T h a t such assumptions are necessary can be seen as follows: Someone's speech behavior (e. g., A's statement: "I think that the apple is red"), can justify the attribution of a definite belief (i. e., that A believes that the apple is red), only if the meanings that A attaches to his terms are k n o w n . Conversely, these meanings can only be derived f r o m knowledge a b o u t his beliefs. T h e result is an underdetermination of the belief system by the linguistic utterances, and this underdetermination can only be reduced or eliminated by applying additional m a x i m s which govern the interpretation. T h e principle of rationality is such a m a x i m of interpretation and, as such, has no counterpart in physical theory. Accordingly, D a v i d s o n says: Beliefs cannot be ascertained in general without command of a man's language; and we cannot master a man's language without knowing much of what he believes. [...] The reason we cannot understand what a man means by what he says without knowing a good deal about his beliefs is this. In order to interpret verbal behaviour, we must be able to tell when a speaker holds a sentence he speaks to be true. But sentences are held to be true partly because of what is believed, and partly because of what the speaker means by his words. The problem of interpretation therefore is the problem of abstracting simultaneously the roles of belief and meaning from the pattern of sentences to which a speaker subscribes over time. [...] In the case of language, the basic strategy must be to assume that by and large a speaker we do not yet understand is consistent and correct in his beliefs — according to our own standards, of course. [...] The interpretation of verbal behaviour thus shows the salient features of explanation of behaviour generally: we cannot profitably take the parts one by one (the words and sentences), for it is only in the context of the system (language) that their role can be specified. When we turn to the task of interpreting the pattern, we notice the need to find it in accord, within limits, with standards of rationality. In the case of language, this is apparent, because understanding it is translating it into our own system of concepts. But in fact the case is no different with beliefs, desires, and actions. 65 65

D a v i d s o n 1974, 238 — 239 (emphasis in the original).

Intentionality of the Mental and Anomalous Monism

107

This passage reveals the close connection between Davidson's argument against psychophysical laws and Quine's thesis of the indeterminacy of translation.66 Quine develops this thesis using the example of the problem of radical translation, i. e., of translating a previously unknown language, as it was spoken, for example, by a native tribe in a distant land. In this case, translation can only be based on observable behavior, i. e., it must be based on observing which sentences a native speaker accepts and which he rejects in given constellations of external stimuli. One of the results of Quine's discussion of this scenario is that, using this limited database, no unequivocal distinctions can be made between finding out the meaning of terms and examining the truth of statements. Sentences only obtain meaning by being linked to a network of other sentences, viz. only by being part of a terminological system or a theory, while conversely this network or system can only be derived from single sentences. The dilemma hinted at here is mitigated — if not avoided — by the possibility of deriving the meaning of certain single sentences directly from observed behavior without having to refer to the entire terminological system. This method does not, however, suffice to specify the meanings of all sentences without knowledge of their relations to other sentences. Thus, one cannot first explain the meaning and then examine the content; rather the meaning is often first derived from the content. From this mutual dependence arises a certain ambiguity in interpreting a foreign terminological system, which Quine proposes to eliminate by adopting a principle of charity. This principle expresses the maxim that, in doubtful cases, meaning should be attributed such that the greatest number of true statements results. 67 Quine's model can furnish a basis for a general theory of hermeneutics: Principles are developed which are supposed to govern the attribution of attitudes. The characteristics encountered in the phenomenon of radical translation are also found (even if in a weakened form) in the interpretation of language and behavior in general. Thus, in interpreting texts — especially apophantic texts, i. e., texts in which assertions are formulated with systematic intent, claims of validity are advanced, and arguments for and against such claims are made — the question typically

66 67

Davidson, in fact, explicitly refers to such a connection. Cf. Davidson 1970, 2 2 2 - 2 2 3 . Cf. Quine 1960, 9 - 17, 31 - 40, 58 - 60. Quine only formulates his thesis for reference, since he is skeptical about meaning.

108

The Contingent Nature of Mind-Body Identity

arises as to what meaning an author attributes to his terms. These meanings can only be derived within the context of the text, i. e., from the content of its assertions, while, conversely, this content can only be obtained from the meaning of the terms. If an author develops theses which appear outlandish or even outright false, one can either accuse him of simply being mistaken or one can assume that he has been misunderstood and consequently reinterpret his terms. Here, too, one sees the mutual dependence of meaning and content — or at least the tendency toward it — that is characteristic of the situation of radical translation. This makes it clear that the data (whether it is a text or linguistic behavior in general) do not, in general, unambiguously determine the attribution of beliefs, and that it requires additional principles to make it unambiguous. In addition to the abovementioned principle of charity, a principle of coherence is commonly invoked, which says that the attributed system of attitudes is supposed to be consistent and organized deductively. It is assumed that the beliefs are compatible and logically linked to one another, i. e., they are, as much as possible, not isolated from one another. This presentation reveals the similarity to Davidson's position as recounted above. In the quoted passage, Davidson emphasizes the analogy between the problem of translation (here: the problem of interpreting texts) and the interpretation of behavior. He points to the mutual dependence of meaning and content and concludes from this interdependence that the two aspects cannot be considered separately and consecutively. This mutual dependence produces an ambiguity which can only be reduced by limiting the range of possible interpretations. The strategy of interpretation must, therefore, be to introduce such limitations in the form of assumptions of the greatest possible consistency and correctness. This means that Davidson wants to invoke both the principle of charity and the principle of coherence to single out a definite interpretation and, to all appearances, these are the principles of rationality which are constitutive for the psychological. § 5 The status of this principle of rationality will now be examined. To this end, a look at the natural sciences will be helpful. Even in the natural sciences, the problem arises that a theory cannot be judged unambiguously using the facts at hand. Since false premises may have true consequences, theories can exist which are conceptually incompatible, but empirically indiscernible. Hipparchus already mentioned that an

Intentionality of the Mental and Anomalous Monism

109

eccentric theory of the sun's orbit (in which the sun moves in a uniform, circular orbit, but the earth is not at its center) leads to the same astronomical results as an epicyclic theory (in which the earth is at the center of the sun's orbit, but which attributes epicyclic motion to the sun).68 Similarly, Lorentz's electron theory and Einstein's special theory of relativity were, at first, empirically equivalent. In such cases, the facts are insufficient to decide about a theory's degree of confirmation or its measure of acceptability and the latter are to be determined by methodological criteria. Acceptable theories should not alone conform to the facts, but satisfy in addition methodological criteria such as simplicity, precision, explanatory power and heuristic fruitfulness. This means that the appropriateness of theories is, in part, determined by transempirical aspects. 69 This is also precisely the function of the principle of charity and the principle of coherence. Both principles make it possible to establish the superiority of certain interpretations. Only by means of these principles it is possible to make a choice from among a group of interpretations of a text or explanations of behavior which otherwise seem equally justified in the light of the available data. This demonstrates that principles of rationality and methodological criteria serve the same ends and are functionally analogous in this sense. Principles of rationality are, therefore, not unique and restricted to a hermeneutically oriented psychology; rather, they occur in the natural sciences as well. § 6 Let us now pass on to the adequacy of the two principles of rationality. While the principle of coherence seems to us to be a thoroughly plausible and natural maxim, we consider the principle of charity to be only of limited use. Particularly in the historiography of science, adherence to the principle of charity has often had an unfortunate impact. For example, Thorpe in 1921 believed the decisive aspect of Cavendish's essay of 1766 — in which the latter identified phlogiston with hydrogen — was the rejection of the theory of a universal and 68

69

Cf. Theon of Smyrna 1878, 166, 188. Also Duhem 1908, 8 ("Hipparchus seems to have been very much struck by the agreement between the results of two such very different hypotheses"); Mittelstrass 1962,170 — 171,176. A corresponding allusion to the geometric equivalence of the geocentric and the heliocentric systems is found in Leibniz's Tentamen de motuum coelestium causis (1689) (cf. Leibniz 1849—1863, VI, 146 footnote). For this presentation, cf. Carrier 1986a, 42 - 44.

110

T h e Contingent N a t u r e of M i n d - B o d y Identity

elementary air. In fact, few indications of such a rejection are found in the essay. Rather, Cavendish merely describes the different properties of several gases more precisely than was previously the case without, however, openly conceiving of this difference among properties as an argument against the interpretation that all gases are merely modifications of universal air. Thorpe's interpretation is apparently based on the principle of charity: The author being interpreted should be credited with as many correct views as possible. Cavendish's work is not evaluated on the basis of the phlogiston theory because this theory is wrong. Rather, the interpretation tries to reveal — at least in embryonic form — an aspect of the work that is still considered valid. The historiography of science under the patronage of the principle of charity interprets the past one-sidedly from the perspective of the present and judges the progress of science solely on the basis of whether or not contributions were made to our present knowledge; it views science as oriented finalistically to the present state of scientific knowledge. In short, the principle of charity, if it alone determines the presentation, privileges interpretations which for other — e. g., historical — reasons might be inappropriate. Thus, the principle is not fit to be recommended unreservedly as a guarantor of rationality. 70 This critique of the principle of charity in scientific contexts does not affect the maxim that — especially in philosophical contexts — the arguments of an author or a text should be made as strong as possible when developing an argument against them. In this way, one avoids the danger of the critique's not reaching the actual level of the text or author criticized. Thus, a correctly understood principle of charity is also tied to the concept of reconstruction which is so fundamental to the analysis of the mind-body problem and the attempts to solve it. Reconstruction means neither (mere) descriptive analysis nor arbitrary abandonment of the object to be reconstructed. Rather, a reconstruction is considered successful or adequate, when a construction C' is substituted for a given terminological context C, where C' not only correctly reproduces C in all its essential aspects but also better fulfills all the intentions that C seeks to fulfill — or, at least, does no worse. Thus, a reconstruction also differs from an interpretation in the usual sense in that it consciously alters the object, although retaining its essential properties. Recon70

For the e x a m p l e used here, cf. Carrier 1986b, 207 - 208. For additional arguments against the principle of charity, cf. McAllister 1986, 324 - 326.

Intentionality of the Mental and Anomalous Monism

111

structions, in contrast to interpretations, eliminate the distance between the production and reception. That is, they substitute themselves for the reconstructed objects — in the sense of the definition we have given — claiming to be those objects, only in a more perspicuous form. 7 1 Therefore, the effort to understand an author "better than he understands himself" makes perfect sense. It is essential to formulate a view in such a way that it remains plausible in the context of a changing history (be it the history of ideas, of philosophy, or even of science), without its being concluded that only current standards could be applied to older views. Understood in this way, the principle of charity is also a philosophical principle which can establish a "hermeneutics of truth," without somehow being able to guarantee this truth, especially in historical contexts. On the contrary. As the example given above shows, the principle of charity can falsify the historical truth if — with reference to current knowledge — it is taken as the sole guarantor of rationality and of the historical analysis of scientific developments. Moreover, the principle of charity — together with the concept of reconstruction as just explicated — serves another important purpose in that it sometimes gives rise to hermeneutic problems. If an author advances a thesis that, at first, sounds absurd, it is useful to ask whether he might have been misunderstood. Namely, if Davidson — in one of the passages quoted above — speaks of the conditions of coherence, rationality and consistency having no echo in physical theory, this seems at first to be clearly nonsense. After all, these conditions are often thought to be fulfilled in exemplary manner in the natural sciences. At this point, the principle of charity comes in, since, as Quine correctly comments: T h e c o m m o n sense behind the m a x i m [the principle of charity] is that one's interlocutor's silliness, beyond a certain point, is less likely than bad translation — or, in the domestic case, linguistic divergence. 7 2

The application of the principle of charity is fruitful in that it reveals a problem of understanding, where a rejection could otherwise be caused by a misunderstanding. In our case, this means: It is likely that Davidson is using his terms in an unusual way in the passages quoted; the task is to understand their meaning in the proper way. Of course, the choice of a particular interpretive hypothesis depends less on which view we ourselves 71 72

For this concept of reconstruction, cf. Mittelstrass 1984e, 89 — 93. Quine I960, 59.

112

The Contingent Nature of Mind-Body Identity

advocate, than on which interpretation is "harmonious," which best fits into the system of Davidson's theory as a whole, i. e., which is the most coherent. The principle of coherence is a reasonable hermeneutic principle. § 7 Finally, the question arises if the use of the principle of coherence allows one to give special status to disciplines which operate hermeneutically. Up to this point, we have only shown that the methodological criteria in the natural sciences are analogous to hermeneutic principles in a formal sense. Nonetheless, the content of the principle of coherence might still be an exclusive characteristic of the hermeneutic disciplines, or for that matter, of psychology. A number of things indicate that this is not the case, but rather that the natural sciences, too, apply a principle of coherence as a methodological rule. In order to demonstrate this, it will help to present Hempel's discussion of the evolution of methods for measuring time. In principle, according to Hempel, time could be measured by the pulse-rate of the Dalai Lama. This choice, however, would not lead to laws with the same simplicity and same degree of confirmation as those of classical mechanics. Basing the measurement of time on the earth's rotation is desirable, since it allows for developing a coherent corpus of laws. On the other hand, precisely on the basis of these laws, it becomes clear that tidal friction is gradually slowing the earth's daily rotation. If one wanted to retain rotation as the standard of time, it would lead to a substantially more complicated system of laws which would lack the original deductive and systematic coherence. Therefore, the transition to another standard of time seems justified — as has been achieved today in the choice of a certain frequency of light. 73 This description reveals the mutual dependence between the concept of the "same period of time" and the system of mechanical laws — which can be paralleled with the mutual dependence between the interpretation of single words and the interpretation of the whole system of beliefs and goals. Mere observation of empirical processes cannot determine which processes represent the same period of time — no more than one can directly infer attitudes merely by observing linguistic behavior. Rather, an adequate attribution of meaning to single concepts can only be achieved against the background of the coherence of the 73

Cf. Hempel 1952, 74. Hempel's discussion is based on the treatment of this problem in Schlick 1925, 7 1 - 7 2 .

Intentionality of the Mental and Anomalous Monism

113

whole system — just as the use of terms, and thereby the content of a single attitude, is only specified by the context of the system of beliefs. Conversely, the system of laws can only be formulated by specifying adequate single concepts — just as the system of beliefs can only be inferred from knowledge of the single attitudes. This mutual dependence between the meaning of the concepts and the truth of the laws — between the meaning of the terms and the attribution of a system of beliefs — produces an ambiguity which can be reduced by adopting the principle of coherence as a boundary condition for the whole system of laws — or the system of beliefs. This scenario of the co-evolution of terminology and theory demonstrates that one must always be prepared for further theoretical modifications. The principle of coherence controls each phase of the theory's evolution — and this holds literally for the case of hermeneutics. This presentation suggests that in the natural sciences the relation between concept and theory formation can exhibit precisely those features which, according to Davidson, are characteristic of the connection between the meaning of a term, single attitudes and the system of beliefs. Davidson would perhaps respond that this analogy ignores the difference that the natural sciences deal with general laws, while the hermeneutic situation refers to the beliefs and goals of an individual.74 But the natural sciences also consider certain individual cases. For example, using an astrophysical model of the sun and plausible assumptions about microscopic processes, one attempts to obtain an estimate of macroscopic parameters (such as the temperature of the sun's core) and derives from these definite values of observable quantities (e. g., neutrino flux). While it is true that the theoretical assumptions are not just valid for the sun, the model as a whole is set up to explain this single case. If there is still a difference here, it is certainly not important enough to justify the conclusion that the "constitutive principles" of the natural sciences and hermeneutic psychology differ. Thus, we conclude that even if one does, in fact, conceive of psychology as a hermeneutical discipline (as Davidson does), the essential divergence from the methods of the natural sciences, whose existence Davidson asserts, cannot be reconstructed. Davidson's argumentation is, therefore, not convincing.

74

This difference is emphasized in Davidson 1976, 267.

IV. The Self as Pilot § 1 After having described and discussed in the previous chapter the various attempts to prove or disprove the identity theory by means of a priori demonstrations, we will now examine a modern dualistic version, namely, the interactionist position of Popper and Eccles. More precisely, it is an expanded dualistic interpretation, in that it distinguishes among three worlds: World 1 of physical bodies, World 2 of mental states and events, and World 3 of "objective thought content," which already appears in Frege in the same way (as a "third realm"). 1 According to Frege, thoughts a r e neither things in the external world nor ideas. A third realm must be recognized. Anything belonging t o this realm has it in c o m m o n with ideas that it c a n n o t be perceived by the senses, but has it in c o m m o n with things that it does not need an o w n e r so as to belong to the contents of his consciousness. T h u s for e x a m p l e the thought we have expressed in the P y t h a g o r e a n theorem is timelessly true, true independently of whether anyone takes it t o be true. It needs no owner. It is not true only f r o m the time when it is discovered; just as a planet, even before anyone s a w it, was in interaction with other planets. 2

Elsewhere Frege speaks of "sense" 3 or of a "domain of what is objective, which is distinct from that of what is actual." 4 In any case, he means that "thought," i. e., the content or sense of statements, must not be confused with the subjective manner of thinking. "Objectivity" is a property of thought, not of the physical and psychological world. Popper takes up this idea and connects it, on the one hand, with his conception of the growth of knowledge and, on the other, in his treatment of the mind-body problem, with the conception of an interactionist, anti-reductionist dualism. In its epistemological characterization, World

1 2 3 4

Cf. Thiel 1980, 673; Ganslandt 1980, 499. Frege 1918/1919, 1 7 - 1 8 . Frege 1 9 2 3 - 1 9 2 6 , 61. Frege 1893, 1 5 - 1 6 .

T h e Self as Pilot

115

3 is an independent world, although it is brought about by man; 5 its objects are real objects which are not reducible to objects of World 2 and World 1. World 3 is constantly extended via solutions to problems which take the form Pi —> T T —> EE -> P2 (where Pi is the initial problem, T T the tentative theory as a creative conjecture, EE the elimination of error by rational critique, and P2 the new problems that result from the solution achieved). 6 At the same time, World 3 influences World 2 in the sense of the just stated form of problem solutions: Every problem solved not only creates new facts (Frege's "true thoughts"), but also additional problems which are treated in World 2, with consequences even for World 1. The relation of the three worlds is characterized by a feedback process. Moreover, Popper connects his conception of World 3 — which Eccles endorses — with ideas of evolution: the aim-structure of animals or men is not "given," but it develops, with the help of some kind of feedback mechanism, out of earlier aims, and out of results which were or were not aimed at. In this way, a whole new universe of possibilities or potentialities may arise: a world which is to a large extent autonomous.7

This World 3, which exhibits an evolutionary development insofar as it is subject to interaction with Worlds 1 and 2, has at the same time a phylogenetic and ontogenetic correlate in the brain's evolution: an interaction between World 2 and World 3, i. e., their linguistic structure, also leads to a change in World 1, in this case in the brain's structure. Eccles shares this idea, too, and also explains it using language as an example. Only man has a propositional language and this language can be employed only by subjects who have conceptual thought, which is essentially thought related to components of World 3. This thought transcends the perceptual present. It concerns the uniquely human development, utilizing concepts and symbols and rational arguments. By contrast the behaviour of animals is derived from their perceptual present and their background conditioning. When confronted by a situation, they rely on trial and error rather than on the attempt to understand and to act rationally. They also

5

C f . Popper 1972, 118; Popper/Eccles 1977, 5 6 3 ( " W o r l d 3 transcends not only W o r l d 1 but also W o r l d 2 . It really does exist; and not only does it exist, but it is active; it acts upon us (only, o f course, by way o f interaction). I conceive the relationship between W o r l d 1 and World 2 as being s i m i l a r " ) .

6

Popper 1972, 119. Popper 1972, 1 1 7 - 1 1 8 (emphasis in the original).

7

116

The Self as Pilot

rely on imitative behaviour. [...] This ideal world will be recognized as the world of the objective spirit — the third World of Popper. 8

World 3 has, according to Popper and Eccles, a history — in the history of man. It is reflected, however, not only in the history of the species, but also (and especially) in its culture, to which science belongs along with language and philosophy. Popper explains: I agree with Brouwer that the sequence of natural numbers is a human construction. But although we create this sequence, it creates its own autonomous problems in its turn. The distinction between odd and even numbers is not created by us: it is an unintended and unavoidable consequence of our creation. Prime numbers, of course, are similarly unintended autonomous and objective facts; and in their case it is obvious that there are many facts here for us to discover: there are conjectures like Goldbach's. And these conjectures, though they refer indirectly to objects of our creation, refer directly to problems and facts which have somehow emerged from our creation and which we cannot control or influence: they are hard facts, and the truth about them is often hard to discover. 9

Frege's position, which deals primarily with questions of validity, is supplemented with evolutionary considerations here. Evolution, the growth of knowledge and World 3 are connected — not only epistemologically, but also genetically. Frege's logical investigations are transformed into ontological and (insofar as they deal with the biology of evolution) empirical investigations. This is also the case with the application of Popper's three-world theory to the mind-body problem. § 2 Popper and Eccles advocate the dualistic independence of psychological and physical states or events and processes against the monistic point of view and emphasize, moreover, the autonomy and identity of the self (or consciousness) as opposed to its representations, even its physical ones — pointedly expressed in the phrase "the self and its brain." According to Popper, this self — in the sense of a "process" that generates the unity of a person's mental life, not in the sense of a thinking substance (as in Descartes) — is primarily anchored in World 3. 10 To be a self "is partly the result of inborn dispositions and partly the result 8 9

10

Eccles 1970, 170. Popper 1972, 118 (emphasis in the original). Goldbach's conjecture is the (still neither proved nor disproved) assumption that every even number except 2 can be represented as the sum of two prime numbers. Cf. Wolters 1980, 790 — 791. Cf. Popper/Eccles 1977, 1 0 8 - 1 0 9 .

117

T h e Self as Pilot

o f e x p e r i e n c e . " 1 1 T h e self o r s e l f - c o n s c i o u s n e s s is, t h u s , a p r o d u c t o f e v o l u t i o n . M o r e o v e r , a c c o r d i n g t o this v i e w , t h e b r a i n b e l o n g s t o t h e self, n o t t h e self t o the b r a i n . T h e self is the p r o g r a m m e r o f t h e c o m p u t e r " b r a i n " — the pilot, not the piloted: T h e active, psycho-physical self is the active programmer to the brain (which is the computer), it is the executant whose instrument is the brain. T h e mind is, as Plato said, the pilot. It is not, as David H u m e and William J a m e s suggested, the sum total, or the bundle, or the stream of its experiences. 1 2 T h i s n o t o n l y e m p h a s i z e s the i n d e p e n d e n c e o f t h e self in a n i n d e p e n d e n t w o r l d ( W o r l d 3 ) , b u t a l s o i n d i c a t e s h o w w e a r e t o v i e w a n intera c t i o n b e t w e e n W o r l d 1, t o w h i c h t h e b r a i n b e l o n g s , a n d W o r l d 2: we have seen that one kind of interaction between Worlds 2 and 3 ("grasping") can be interpreted as a making of World 3 objects and as a matching of them by critical selection; and something similar seems to be true for the visual perception of a World 1 object. T h i s suggests that we should look upon World 2 as active — as productive and critical (making and matching). 1 3 F o r E c c l e s , w h o s e c o n c e p t i o n d i f f e r s f r o m P o p p e r ' s h e r e , t h e selfc o n s c i o u s self h a s its p l a c e in W o r l d 2 . 1 4 A t the s a m e t i m e ,

Eccles

t r a n s l a t e s P o p p e r ' s i d e a o f t h e self a s t h e p r o g r a m m e r o f t h e c o m p u t e r " b r a i n " into a neurobiological language. H i s hypothesis, which seeks to c o m b i n e P o p p e r ' s p h i l o s o p h y o f t h e self w i t h n e u r o b i o l o g i c a l t h e o r i e s , is t h e f o l l o w i n g : Hitherto it has been impossible to develop any neurophysiological theory that explains how a diversity of brain events comes to be synthesized so that there is a unified conscious experience of a global or gestalt character. T h e brain events remain disparate, being essentially the individual actions of countless neurones that are built into complex circuits and so participate in the spatiotemporal patterns of activity. This is the case even for the most specialized neurones so far detected, the feature detection neurones of the inferotemporal lobe of primates [...]. Our present hypothesis regards the neuronal machinery as a multiplex of radiating and receiving structures: the experienced Popper/Eccles 1977, 111. ' 2 Popper/Eccles 1977, 120. 13 Popper/Eccles 1977, 48. For the (epistemological) problematic of the concept of " m a t c h i n g , " cf. Wolters 1988b, 1 3 3 - 1 3 6 . 14 Cf. Popper/Eccles 1977, 359 — 360. According to Eccles, memory belongs to World 3, which is also incompatible with Popper's conception. 11

118

The Self as Pilot

unity comes, not from a neurophysiological synthesis, but from the proposed integrating character of the self-conscious mind. We conjecture that in the first place the self-conscious mind is developed in order to give this unity of the self in all of its conscious experiences and actions. 1 5

According to this hypothesis, the self controls and interprets the neuronal processes "by virtue of a two-way interaction across the interface between World 1 and World 2." 1 6 It actively seeks brain events which lie in its domain of interest and integrates them into a unified and conscious experience. It constantly scans collective interactions of large numbers of neurons ("cortical modules"), which are open to an interaction with World 2 ("liaison brain"). Like a scanning device, it seeks among the 106 modules for the ones that are open to it and influences their behavior. This truly busy self, or this self-conscious mind, dwells primarily in the left hemisphere of the brain, which explains the latter's dominance. Further properties are attributed to the self, such as the capacity of building up a readiness potential for voluntary motions (by coordinating the modular activities) and concentrating it in the appropriate place. The unity of conscious experience, according to Eccles' central thesis, is provided by the self-conscious mind and not by the neural machinery of the liaison areas of the cerebral hemisphere. 1 7

In other words, the mind uses the liaison brain to bridge the psychophysical abyss to influence the brain mechanism that resides in World 1. § 3 Let's now turn to the reasons Popper and Eccles give in support of their dualistic theory. Popper relies on an anti-reductionist thesis, i. e., on the assertion that different domains of reality (such as physics, chemistry, ecology/sociology) 18 cannot be reduced to one another. This thesis is founded on (1) an argument by analogy from the history of science, (2) a reference to emergent properties, and (3) a Darwinistic or evolutionary argument.

15 16 17 18

Popper/Eccles 1977, 362 (emphasis in the original). Popper/Eccles 1977, 3 5 5 - 3 5 6 . Popper/Eccles 1977, 362. Cf. Popper/Eccles 1977, 1 6 - 1 7 , 20.

The Self as Pilot

119

With regard to the first argument, Popper contends that there are virtually no successful reductions in the history of science. The only known example is said to be the reduction of Fresnel's optics to Maxwell's electrodynamics, while the reduction of phenomenological thermodynamics to statistical thermodynamics, for example, is said to be "incomplete." 19 This argument makes it clear that, for Popper, the correction of the laws of the reduced theory by the reducing theory, i. e., the fulfillment of Nagel's "informal conditions" (cf. II.2.§ 1), precludes the existence of a reductive relation outright. In fact, a reduction generally goes along with a correction of the reduced theory. If this peculiarity precludes reducibility for Popper, then he is correct in stating that virtually no successful reductions exist. However, this argument has a number of shortcomings. First, it leads to the consequence that, even within the natural sciences, the majority of theoretical identifications are unjustified. As was presented in II.2.§ 3, the ontological reduction of temperature to average molecular kinetic energy, for example, succeeds by way of a corrective theory reduction. If the reduction claim is rejected under such circumstances, one is led to the very implausible conclusion that statistical mechanics does not treat thermal phenomena. Adopting Popper's concept of reduction would have truly revolutionary consequences for physical ontology. Second, even if we were prepared to pay this price, the advantages would be limited. After all, Popper's proviso against reductions does not exclude the possibility that neurophysiology could force psychology out of its role as a theory of behavior. Popper's argument admits the possibility that neurophysiology improves on psychology and, for this reason, replaces it. Only corrective reductions fall victim to Popper's proviso, not corrective theory replacements. Popper's argument against the identity theory could, thus, easily be taken as an argument in favor of eliminative materialism (cf. II.6.§ 1). Accordingly, it does not establish the point in question. Moreover, in another context, Popper praises the particular value of corrective reductions: I suggest that whenever in the empirical sciences a new theory of a higher level of universality successfully explains some older theory by correcting it, then this is a sure sign that the new theory has penetrated

19

Cf. Popper/Eccles 1977, 1 8 - 1 9 .

120

The Self as Pilot

deeper than the older ones. The demand that a new theory should contain the old one approximately [...] may be called (following Bohr) the "principle of correspondence".20 P o p p e r ' s principle of c o r r e s p o n d e n c e a p p a r e n t l y expresses N a g e l ' s r e l a t i o n of f r u i t f u l r e d u c t i o n (which is itself a special case of o u r relative interpretability (cf. II.2.§ 1—2)). W h i l e P o p p e r praises t h e " d e e p e r p e n e t r a t i o n " here, h e s p e a k s merely of " i n c o m p l e t e n e s s " in t h e a n t i - r e d u c t i o n ist c o n t e x t . In s h o r t , P o p p e r ' s p r o v i s o a g a i n s t r e d u c t i o n n o t only d o e s n o t achieve the i n t e n d e d g o a l , b u t also c o n t r a d i c t s P o p p e r ' s o w n ideas. § 4 P o p p e r ' s second a r g u m e n t is b a s e d o n the p h e n o m e n o n of emergence. T h e p r o p e r t i e s of systems o r w h o l e s t h a t are e m e r g e n t in the logical sense are t h o s e w h i c h c a n n o t be p r e d i c t e d f r o m t h o s e l a w s t h a t h o l d f o r their c o m p o n e n t p a r t s a n d t h e i n t e r a c t i o n b e t w e e n t h e m . " E m e r g e n t " is also used in a temporal sense. In this sense, e m e r g e n t p r o p e r t i e s are t h o s e w h i c h a p p e a r a f t e r a certain p o i n t in an e v o l u t i o n a r y process. F o r e x a m p l e , in the f r a m e w o r k of an e v o l u t i o n a r y c o s m o g o n y , e m e r g e n c e refers t o t h e a p p e a r a n c e of genuinely n e w qualities (such as t h e origin of life o u t of an i n o r g a n i c w o r l d ) . T h e antithesis of a n e m e r g e n t t h e o r y in t h e t e m p o r a l sense is n o t r e d u c t i o n i s m b u t r a t h e r a p r e f o r m a t i o n i s t view t h a t w o u l d , f o r i n s t a n c e , r e g a r d all p h e n o m e n a as spatial r e a r r a n g e m e n t s of particles. 2 1 B o t h c o n c e p t s of e m e r g e n c e are logically i n d e p e n d e n t of o n e a n o t h e r . T h u s , a p r e d i c t a b l e p r o p e r t y m i g h t only a p p e a r at a p a r t i c u l a r p o i n t in time (namely, w h e n all t h e c o n d i t i o n s f o r its realization a r e satisfied); a n d a logically u n p r e d i c t a b l e p r o p e r t y m i g h t h a v e existed all a l o n g . 2 2 P o p p e r , t o o , distinguishes b e t w e e n t h e t w o c o n c e p t s of e m e r g e n c e , 2 3 b u t he includes t h e a s s e r t i o n of a "creative e v o l u t i o n , " i. e., t h e existence

20 21

22 23

Popper 1972, 202 (emphasis in the original). Cf. E. Nagel 1961, 3 6 7 - 3 6 8 , 3 7 4 - 3 7 5 . Feigl already distinguishes between both concepts of emergence. Cf. Feigl 1958, 415. The recent development of irreversible thermodynamics shows precisely how the evolution of new kinds of structures can be explained from short-range interactions between the constituents and from global constraints of the system in question. Even "becoming" (i. e., temporal emergence) is accessible to physical theory if the latter is sufficiently complex. Cf. Prigogine 1980, 103 — 130. This is also true of biological evolution. As recent research (such as that of Eigen) shows, the occurrence of evolutionary processes and the development of complex molecular structures from these processes can, in principle, be predicted. Cf. E. Nagel 1961, 375. Feigl mentions this as well. Cf. Feigl 1958, 4 1 5 - 4 1 6 . Cf. Popper/Eccles 1977, 16.

T h e Self as Pilot

121

of temporal emergence, among the reasons for his anti-reductionist thesis, which leaves us with the impression that he considers this "creativity" a valid argument against reductionism, i. e., in favor of logical emergence. 24 An additional distinction between two types of logical emergence is appropriate. It concerns the difference between merely factual nonderivability of the properties of systems from the laws of their components and their non-derivability in principle.15 Factual non-derivability of the properties of systems is due to our ignorance of the prevailing regularities, non-derivability in principle is due to their non-existence. An effective anti-reductionist argument would have to give reasons why the emergent properties could not have resulted from interactions between the elements of the next lowest level. If a coil or a capacitor are alternately introduced into a circuit, no oscillation in the current is detected. If, however, both coil and capacitor are introduced at the same time, oscillations occur. 26 Emergence or non-predictability is obviously present only for someone who does not know the laws governing the interaction of the constituents. Invoking emergence without additional explanation merely feeds the suspicion that what is done here simply amounts to fencing in an asylum ignorantiae. Popper does nothing to dispel this suspicion. 27 This is, however, urgently needed, since the concept of emergence is also used within monistic approaches. Both Bunge 28 and Feigl emphasize the existence of emergent properties. Feigl's argumentative strategy tries to 24 25 26 27

28

Cf. Popper/Eccles 1977, 1 4 - 1 8 . This distinction is already made by Feigl. Cf. Feigl 1958, 424. For this example cf. Ko. Lorenz 1973, 31. Popper gives the diffraction by a lattice or crystal as an example of the effect of an irreducible structure. Cf. Popper/Eccles 1977, 1 8 - 1 9 . In doing so, he overlooks the fact that the diffraction can be entirely explained from the arrangements of the components. A second argument is based on supposedly permanent inaccuracies in the explanation of the half-lives of atomic nuclei, although he admits that an approximate explanation is possible (Cf. Popper/Eccles 1977, 27 — 28). If, however, the necessity of using approximations is considered sufficient to characterize a property as emergent, then this concept is completely trivialized. In quantum mechanics, even the helium atom can only be treated by methods of approximation and, hence, w o u l d be emergent relative to quantum mechanics. Such a weak concept of emergence would no longer serve as a basis for dualism. In that case it w o u l d be admissible to draw a parallel between the relation of mind and body and the relation of a hydrogen atom and a helium atom; and this w o u l d be compatible with any form of materialism. Cf. Bunge 1980, passim. Bunge makes no more of an effort to give a more precise explanation of his concept of emergence than Popper does.

122

T h e Self as Pilot

demonstrate that emergence is a property c o m m o n even in anorganic systems, and thus t o defuse it. 2 9 So emergence is only effective in support of dualism if it is handled circumspectly. O n the whole, this means that emergence in the factual sense refers to a " n e t w o r k p r o p e r t y " and it implies the derivability of the feature in question from a comprehensive theory o f the relevant system. E m e r g e n c e in the principle sense, on the other hand, means (in C h u r c h l a n d ' s m o r e precise form) that a property does, indeed, appear together with a feature described by a theory T n , but c a n n o t be explained within the theory's framework: a property F will be said to be an emergent property (relative to TN) just in case (1) F is definitely real and instantiated; (2) F is co-occurrent with some feature or complex circumstance recognized in TN; but (3) F cannot be reduced to any property postulated by or definable within TN. 3 0 T h e property F is subject to laws that are n o t derivable T n . Emergence in the principled sense is, therefore, a relative

from

property;

it always refers to a particular theory and entails non-reducibility. In the first place, only emergence o f this type can be used in support of dualism. Factual emergence only serves this purpose if it can be shown that our ignorance o f the pertinent laws is of a fundamental nature. T h o u g h the property in question may in fact be determined by certain laws, the corresponding derivation may still be ruled out for reasons o f principle. T h i s would be characterized as emergence in the principlefactual sense. In VIII.2, we will offer an argument in favor o f this option in order to support dualism. § 5

Popper's third argument in support o f his interpretation

of

the mind-body relationship aims to s h o w that the identity theory is irreconcilable with a Darwinistic point o f view. According t o Popper, it is inconceivable f r o m a D a r w i n i s t i c perspective that psychological processes could have developed at all if the identity theory were correct. Popper argues that the identity theory must explain all causal effects o f

29

30

Cf. Feigl 1958, 4 1 4 , 4 1 5 . It is clear that Popper's argument (cf. note 27) falls victim to this strategy. P.M. Churchland 1985, 12 (emphasis in the original).

T h e Self as Pilot

123

World 2 on World 1 in physical terms, i. e., in terms of World 1. The identity theorist cannot consider psychological processes a legitimate basis of explanation. This would imply that psychological processes are causally redundant, that the world would manage well enough without them. Thus, the development of psychological processes would not have had any advantage in natural selection, so that the origin of consciousness could not be explained on a Darwinistic basis. 31 In evaluating this argument, it must first be pointed out that it is inadequate as an objection to the token-identity theory. As presented in II.5.§ 3, functional materialism emphasizes precisely the autonomy of psychological concepts and points out that they are indispensable for explaining behavior. Thus, Popper's argument can, at most, be applied to a *Pone — ^one model of the psychophysical relationship, i. e., to a theory of type-identity. But there, too, the significance of Popper's objection seems limited. It must be emphasized that the identity theory (in contrast to Popper's formulation) does not assume a causal influence of World 1 on World 2 at all, but rather assumes the identity of World 1 with World 2. Psychological processes are physical events, they are merely observed from a different perspective. Thus, psychological processes are not causally redundant; after all, they are identical with causally effective physical processes. Properly understood, the heart of the difficulty that Popper describes resides in the question: how is it possible that for certain physiological processes another description — namely, a psychological one — can be given? But this question no longer has any recognizable connection to Darwinism. Moreover, Popper's argument can be used against certain theoretical identifications in the natural sciences. It would be possible, for example, to view the identification of colors with particular frequencies of the electromagnetic spectrum as a violation of the principles of Darwinism. Since, under the hypothesis of such an identity, only electromagnetic waves (i. e., entities of World 1), but not colors (i. e., entities of World 2) seem to be causally effective, it is completely incomprehensible how color perceptions could have evolved at all. Popper's objection would lead one to see a contradiction between Darwinism and the reduction of optics to electrodynamics. In the face of such consequences, one would certainly take Popper's argument, and not this reduction, to be inadequate. 31

Cf. Popper/Eccles 1977, 96.

124

The Self as Pilot

§ 6 In contrast to Popper, Eccles seeks to support dualism using neurophysiological arguments. 3 2 These arguments have already been shown to be inadequate by M a c K a y . Eccles' physiological indications for the intervention of an independent self can all be equally well understood from a monistic position. 3 3 Eccles, moreover, does not make his position anymore plausible by operating with such adventurous inventions as the "liaison brain." This reliance on pseudo-physiological hypotheses — which in a different fashion is also occasionally encountered on the monistic side — is one of the more annoying aspects of the contemporary debate on the mind-body relation. It is rightly pointed out in this context that within this conception it would make as much sense "if one substituted — by way of experiment — G o d , the World Spirit, the unconscious, an entelechy or the transcendental ego for the self-conscious mind." 3 4 Only, in this case, it is no longer classical metaphysics but rather a strange form of science fiction (the mind and its liaison brain). Whoever argues like Eccles has in substance not advanced far beyond Descartes' pineal gland hypothesis or the notion of a ghost in the machine. Despite this criticism, it should be said that we agree with many aspects of Popper's interpretation of the mind-body relation within the framework of the Popper/Eccles conception — even if we remain very skeptical about a hypostatization of World 3. We concur with the methodological idea of introducing psychological states and events as explanatory hypotheses that are open to empirical scrutiny: we can introduce, in psychology, c o n s c i o u s a n d unconscious mental events and processes, if these are helpful in explaining h u m a n behaviour, such as verbal behaviour. In this case, the attribution of a mind and of subjective c o n s c i o u s experiences to every normal h u m a n person is an e x p l a n a t o r y theory of p s y c h o l o g y of a b o u t the s a m e character as the existence of relatively stable material bodies in physics. In b o t h

32 33 34

Cf. Popper/Eccles 1977, 3 6 2 - 4 0 6 . Cf. M a c K a y 1978, 6 0 4 - 6 0 5 . Vollmer 1981, 63; cf. Cohen 1979, 303. Also Dennett 1979, 95 ("Eccles simply has no overarching psychological theory at all. His interactionism, which he touts as a 'radical hypothesis [...] that is essentially a new philosophy of perception' (271), is an unrivaled example of the invocation of a deus ex machina. Whenever a truly difficult theoretical issue arises, such as the 'integration' or 'interpretation' of the multifarious neural 'signals,' Eccles passes the buck to 'the self-conscious mind,' about whose apparently wonderful powers he is conveniently silent" (emphasis in the original)).

The Self as Pilot

125

cases [...] our theories are well testable: in physics, by the experiments of mechanics; in psychology, by certain experiments which lead to reproducible verbal reports. 3 5

We disagree primarily with the reasons that Popper gives for his view. But, here too, it must be emphasized that we agree with the basic lines of Popper's argumentative strategy, insofar as Popper also considers the identity theory to be a sensible, contingent, but empirically implausible claim. Popper sensibly does not look for a priori arguments against mind-body identity and in support of interactionism. Rather, he questions the plausibility of mind-body identity in the light of scientific theory and the philosophy of science. This line of questioning is also characteristic of the approach advocated here. Both of the following chapters accordingly seek to develop a different sort of justification for dualism.

35

Popper/Eccles 1977, 6 2 - 6 3 .

V. Science and Mind: Explanations of Behavior and Mentalistic Language The relation between mental states and brain processes cannot be clarified solely by analyses of a logical, semantic or terminological nature. In addition, the evaluation and philosophical interpretation of empirical science or empirical research strategies is required. Every position in the mind-body debate makes claims that can be backed, at least in part, only by empirical means. All of them make assertions about matters of fact, and none of them simply moves within the vast area of logical possibility and consistent thought. T h u s , it is unjustified to reject as does Bieri, for example, all questions about the empirical plausibility of mind-body identification and to see the guiding idea of the discussion solely in the problem of how to formulate coherently such an identification. 1 Here, too, the identity theory has been defended against objections which seek to deny that it is a meaningful and coherent position. However, one does not immediately become an adherent of a view simply because one believes it to be free of contradictions. After all, one could acknowledge the logical possibility of phlogiston chemistry or the caloric theory of heat without thereby becoming an advocate of these theories. Precisely because the identity theory is a sensible, contingent thesis, it can be — and, in fact, needs to be — supported by empirical evidence. Without empirical plausibility, the identity theory would merely have the status of such assertions as that the fairies and elves meet in glades to dance on mild moonlit nights. In fact, the identity theory's claim also aims at empirical confirmation. T h u s , for example, Feigl explains that the point is to show that monism is "very plausible on scientific g r o u n d s , " 2 and he asserts: N o r m a l inductive extrapolation from the successes of psychophysiology to date makes it plausible that an adequate theory of animal and human behavior can be provided on a neurophysiological basis. 3 1

Bieri 1 9 8 1 , 3 7 .

2

Feigl 1 9 5 8 , 4 8 3 . Feigl 1958, 3 8 2 .

3

Reduction Programs in Psychology

127

In the following, we shall attempt to show that such a claim is illfounded. At this point, the task consists primarily in sketching the reasons for introducing mental terms into the theory of behavior. In chapter VI we will then discuss the philosophical consequences of mentalism. The first question is whether recent developments in psychology make the success of reduction programs like the identity theory more probable. In the first section, we develop an argument by analogy from the history of science. This argument has an anti-reductionist intent that is specifically aimed at psychology and not at reductions in general as was Popper's argument which we rejected above. In the second section of this chapter, the structure of several psychological theories will be explained in order to clarify how mental terms enter into explanations of behavior, i. e., in what sense mentalism is scientifically justified (and why the corresponding reduction programs failed). In addition, we attempt to furnish a basis for the epistemological discussion of psychology in the following chapters.

1. Reduction Programs in Psychology: Behaviorism and the James-Lange Theory § 1 A reduction of psychology to neurophysiology is not in sight. Though the present state of research does, indeed, enable one to correlate specific psychological occurrences with events in an EEG (so-called "evoked potentials"), there is no psychophysical law in the sense that specific mental events can be reliably connected to particular neurophysiological processes. At the present time, no one can reliably specify the neural mechanisms that correspond to single psychological states. Thus, the neurophysiologist Schmidt explains: Almost nothing is known about the origins of neuronal impulse patterns, which lead from impulses for action to plans for motion. However, if thoughts lead to actions, the neurophysiologist is compelled to assume that the neuronal activity of the brain can be changed by thought. [...] This transformation of thinking and willing into cortical impulse patterns is however at the moment far beyond our comprehension. 4 4

Schmidt 1987, 200.

128

Science and M i n d

One of the disheartening accompaniments of this situation is that supporters of the identity theory occasionally aspire to make up for this shortcoming by means of a little science fiction (similar to that encountered in Eccles' dualistic theory). Thus, Bunge, for example, whose "emergentist materialism" is, in fact, a variant of the identity theory, 5 bridges the gap between psychology and neurophysiology solely on the strength of his imagination and places a visionary scenario before our eyes, in which a wide assortment of "psychons" and similar creatures of an imaginary cognitive neurophysiology is unfolded before the amazed spectator. 6 The motto seems to be: If the theory does not fit the facts, then furnish the corroborating empirical data yourself. But when philosophers of science, who lay claim to a certain philosophical and scientific seriousness, ignore all the pertinent facts when outlining the details and increasingly base their contentions on fictions, when they simply tell fairy-tales and give their "scientific" imagination free rein, then it is they themselves who have spoken judgment on the viability of their ideas. At present, we simply have no idea what form an eventual reduction of psychology to neurophysiology could take. An advocate of a monistic theory of identity could agree with that, too, and still remain a monist, namely, in the sense of asserting the feasibility of a reduction program. And this is precisely what Feigl does when he speaks of "extrapolating" from the present state of knowledge (cf. V. Introduction). Hence, Feigl's claim could be reconstructed as the assertion that the identity theory is a promising reduction program. At the moment, there are no criteria for promising reduction programs; but the fate of other reduction programs in psychology should admonish us to proceed with caution. Just because a number of reduction programs have had spectacular success in physics, we cannot forget that the history of psychology is filled with disappointments in this area. Blinded by the successful reductions in physics, one overlooks the rockiness of one's own turf. Wittgenstein's diagnosis is valid here: A main cause of philosophical desease — a one-sided diet: one nourishes one's thinking with only one kind of examples. 7

In the following, we present in brief the fate of two reduction programs in modern psychology: the failure of Skinner's behaviorism and the inglorious decline of the James-Lange theory of emotion. 5 6 7

Cf. Bunge 1980, X . Cf. Bunge 1980, 1 5 5 - 1 5 9 . Wittgenstein 1958, § 593.

Reduction Programs in Psychology

129

§ 2 Behaviorism aims to remove mentalistic terminology from psychology and seeks to explain human behavior entirely by recourse to the external conditions and consequences of behavior. Psychological states are superfluous to explanations of behavior: T h e objection to inner states is not that they d o not exist, but that they are not relevant in a functional analysis. 8

Of course, according to Skinner, one can always refer to inner states as the effects of external conditions or as causes of behavior; but this is merely a laborious detour which can be avoided if behavior (the dependent variable) is connected directly to the external conditions (the independent variable). 9 Theories of inner states are, if they function correctly, redundant and, hence, can be dispensed with; and, if they don't function correctly, they are wrong and, hence, can be dispensed with all the more. 10 Skinner's approach is a program to reduce mentalistic psychology to a theory that attempts to explain behavior solely as a function of its external antecedents and consequents. In this sense, the behavioristic reduction program is entirely analogous to that of the identity theory. It is no longer disputed that the behavioristic reduction program was a failure. Apart from all its empirical failings, it was primarily the following two objections that led to this negative evaluation: (1) Behavioristic stimulus-response correlations are only valid if they contain a ceteris-paribus clause. This supplementary condition must, however, necessarily have a mental character. 11 (2) Putnam's argument of the "super Spartan" calls attention to the fact that it is possible to imagine a community of people who do not, for example, exhibit the behavior associated with pain, although they do feel pain. 12 This schematic presentation should suffice for now, especially since we will soon (in V.2) return to the relation of behavioristic and cognitive 8 9

10

11

12

Skinner 1953, 35. Cf. Skinner 1953, 35 — 36. On the tendency against hypothetical-deductive theories in Skinner's conception, cf. also Hilgard/Bower 1966, 142—143. Skinner's argumentation here takes precisely the form which Hempel (with explicit reference to Skinner) calls the "theoretician's dilemma." Cf. Hempel 1958, 186. Bieri 1981, 33; Davidson 1970, 50. Both these authors refer this argument back to a study by Chisholm in 1957. In fact, it can be found already in Feigl. Cf. Feigl 1951, 197-199. Cf. Putnam 1963, 3 3 2 - 3 3 4 .

130

Science and Mind

psychology in more detail. Since our central arguments in support of a psychophysical dualism are based on the structure of modern cognitive psychology, we will have to discuss it in more detail anyway. § 3 The second failed reduction program in psychology is the JamesLange theory of emotion, developed near the end of the 19th century. This theory identifies so-called visceral processes, i. e., somatic, not brain processes, with emotions. In contrast to the everyday view, which says, for example, that someone is physically stimulated because he is emotionally stimulated, the reverse is true in the James-Lange theory. According to the latter, the emotional stimulation is controlled by the corresponding physiological process. This means, the emotion is (identical to) the perception of somatic processes. This theory, too, obviously embodies a reduction program that matches the identity theory to a tee. The James-Lange theory has since been rejected by experiment. The theory implies that there is a one-to-one correlation between emotions and physiological states. But empirical findings from the 1920s already contradict this claim. Artificial stimulation of internal organs with adrenalin injections merely produces "as if" emotions (such as, "it is as if I were angry"), not genuine emotions. This fact is confirmed by more recent investigations. In an experiment by Schachter and Singer (1962), different groups of experimental subjects were again placed in a state of physiological stimulation with adrenalin injections, and the subjects were all given different information about what to expect from the injection they had received. Having thus been given different information, the subjects had entirely different emotional reactions. This clearly contradicts the James-Lange theory and supports the assumption "that the quality of emotional states is determined by cognitive factors." 1 3 Schachter and Singer's experiment demonstrates that it is possible to alter emotions by altering cognitions, independently of the physiological stimulation. Just as in the case of behaviorism, the independence of mentalistic terminology has also prevailed in the case of the theory of emotion. 1 4

» R u c h / Z i m b a r d o 1975, 302. 14 For this presentation of the James-Lange theory, cf. R u c h / Z i m b a r d o 1975, 2 9 9 - 3 0 2 ; Lefrancois 1972, 49 - 52. The results of modern psychosomatic research point in the same direction. It was previously assumed, for example, that there was a connection

The Example of Cognitive Psychology

131

§ 4 This discussion of failed psychophysical reduction programs might convey the impression that such enterprises are only doubtful in psychology, but more successful in physics. But the path of physics is also strewn with the ruins of failed reduction programs. We need only recall the Descartes-Boyle program of reducing interactions of all kinds to particle collisions; a program that, in its time, was expounded with as much hope of success as the identity theory is today. In addition, the mechanistic program to reduce electrodynamics to mechanics and Planck's program to reduce quantum mechanics to Boltzmann's statistical mechanics should be mentioned at this point. All these positions can be seen (with hindsight) as not having grasped the true complexity of the situation.

2. The Example of Cognitive Psychology: Mental Events and Behavior § 1 One of the peculiarities of the current discussion of the mindbody problem is that the developments of modern psychology are only rarely taken into account. The argumentation is primarily based on neurobiology, on the one hand, and on everyday psychological experiences, on the other. The contributions of current psychological research are frequently overlooked. In this section, therefore, we will attempt to give some concrete content, based on current research, to the arguments for the necessity of mental terms which until now have been rather general. A cognitive psychology 15 that insists upon the irreducibility of mentalistic terminology cannot limit itself to expressing general reservations about reductionistic programs and citing its vague discontent as evidence; rather, it

15

between depression and dysfunction of the thyroid gland. This is not the case, however. Cf. Glatzel 1973, 70. In the following, we use the expression "cognitive psychology" as a synonym for "mentalistically-oriented psychology," i. e., for a psychology that introduces inner states as intermediary links for stimulus-response correlations. By contrast, the term "cognitive psychology" is often used in a narrower sense, in which case it characterizes certain abstract models for describing perception and thought. In addition, "cognitive psychology" is occasionally used to refer to cognitive science (cf. VII.2). Cognitive science is a philosophical program for interpreting cognitive psychology in the larger sense.

132

Science and M i n d

has to provide better solutions than a reductionistic approach. The following reflections will focus on a discussion of the behavioristic reduction program, with the goal of showing the extent to which cognitive terms stand up to a behavioristic reduction claim and the sense in which they are fruitful or even unavoidable. It is, of course, unnecessary to defend the adequacy of mentalism given the present state of affairs. That much is generally accepted; it doesn't make much sense to carry water to Lake Constance. The aims of this section are more specific. We intend to show: (1) that there are empirical arguments in favor of mentalism, and (2) that the relations between psychological theories are not fundamentally different from those between physical theories. It should become clear how cognitive theories actually function, and in what way (precisely) they cause behavioristic reduction programs to fail. The analogies that can be seen here between the patterns of theory dynamics in the social sciences and those in the natural sciences complement the thesis which will be developed in VI. 1 that there are notable similarities between explanations of human behavior and explanations of natural processes as regards the principles of concept formation and theory structure, too. Finally, (3) an overview of the content of some psychological theories is given, to enable us to use them as illustrations in the course of our argumentation. § 2 Not only is it impossible to give an exhaustive survey of the state of current psychological research but it would make little sense given the scope of our project. We must inevitably focus on just a few contributions that we hope are representative. We will be primarily concerned with learning theory and social psychology, since the behavioristic approach in the form of Skinner's theory undoubtedly had its greatest successes in this field. So we shall start with a short summary of Skinner's learning theory and then show how this approach was superseded by Bandura's cognitive learning theory and the motivation theory of Atkinson, Weiner and Heckhausen. In learning theory before Skinner, as a rule, behavior was taken to be a reaction to a previously or simultaneously occurring stimulus, and the effects of such stimuli on behavior were examined. Pavlov's theory of classical conditioning and Thorndike's theory of connectionism, for example, are part of this tradition. By contrast, Skinner asserts that the vast majority of human behavior occurs spontaneously, i. e., not as a response as previously assumed, but in a so-called operant manner.

The Example of Cognitive Psychology

133

Skinner sees such behavior as determined by its consequences, by the reactions that ensue from the behavior. The central concept of this theory is that of reinforcement. Reinforcements are those external conditions which increase the probability that certain behavior will occur. The theory's central claim is that reinforcements follow the action in time. A change in behavior that is caused by reinforcements is called an operant conditioning. Skinner's empirical experiments were essentially concerned with examining the effects of various reinforcement schedules, i. e., with the question of which reinforcement patterns resulted in the highest learning speed and greatest stability of learned behavior. 16 As just mentioned, cognitive factors have no place in Skinner's model. H u m a n behavioral patterns are, as it were, blindly produced and subsequently shaped by their consequences. This approach becomes problematic, however, if one examines how children acquire languages. The child's linguistic ability develops largely independently of parental reinforcement patterns. 1 7 Moreover, the classic experiment by Bandura, Ross and Ross (1963) on imitative learning revealed that (1) learning can take place even with merely vicarious reinforcement, or even (2) without any reinforcement. 1 8 Bandura and his colleagues showed a group of children movies which showed a child behaving aggressively. The results indicated that the rate of spontaneous imitation of the child's aggressive behavior by the viewing children significantly depended on whether the child in the movie was praised or punished for his behavior. In the case of the children watching the movie, behavior was, accordingly, not influenced by experiencing the consequences of their own behavior, but merely by observing the consequences of another child's behavior. A change in behavior resulted from information about the possible consequences of behavior. 19 In the second part of the experiment, all the children were encouraged to imitate the behavior of the aggressive hero of the movie and given rewards for doing so. The rate of induced imitation proved to be independent of the version of the movie that the children had seen, i. e., whether or not the child in the movie had been praised or punished for 16

17 18

19

For a more in-depth presentation of Skinner's theory, cf. Hilgard/Bower 1966, 107-145. Cf. R u c h / Z i m b a r d o 1975, 1 8 2 - 1 8 4 ; Bandura 1977, 1 8 - 2 0 . For the following presentation, cf. Bandura/Ross/Ross 1963, 601 — 607; Heckhausen 1974a, 1 1 2 - 1 1 3 . Cf. also Bandura 1977, 22.

134

Science and Mind

his behavior. This makes it clear that even those children who — because the vicarious reinforcement was lacking — had not spontaneously imitated the observed behavior had indeed learned it. Otherwise, they could not have reproduced it on demand. Accordingly, one must distinguish between the acquisition and the performance of behavior and note that acquisition, i. e., learning in the true sense, cannot be explained by recourse to reinforcement. 20 Imitative learning or modeling is learning without external reinforcements. It cannot be denied that reinforcements can effect behavior in various ways. But the question is: how do these effects come about? Several experiments on cognitive influences on the effect of reinforcement patterns can help to explain this. A group of experimental subjects were told that a shock would occasionally follow a certain stimulus. Except for a demonstration, however, no shock took place. Nevertheless, the previously neutral stimuli increasingly produced effects of anxiety. 21 Obviously, this is a case of classical conditioning by exclusively cognitive means. An analogous conclusion can be drawn from an experiment in which different groups of subjects were rewarded for performing certain actions according to an objectively equivalent reinforcement schedule, but they were given different information about this reinforcement schedule. It turned out that the influence of the supposed reinforcement pattern was greater than the consequences of the action that were actually experienced. 22 So Skinner's reinforcement does not work by shaping habits mechanically. Rather, reinforcements serve as information about the consequences of behavior, i. e., they have a predictive function; and they work as motivations, i.e, they create an incentive for certain behavior: In the social learning analysis, so-called conditioned reactions are considered to be largely self-activated on the basis of learned expectations rather than evoked automatically. T h e critical factor, therefore, is not that events occur together in time, but that people learn to foresee them from predictive stimuli and to summon up appropriate anticipatory reactions. 2 3

Skinner's theory — insofar as it produced valid results — actually functions cognitively. Our presentation of it here indicates that it might 20 21 22 23

Cf. Cf. Cf. Cf.

Bandura Bandura Bandura Bandura

1977, 1977, 1977, 1977,

28, 35. 70. 155-156. 68-69.

The Example of Cognitive Psychology

135

be better — instead of restricting psychology to a behavioristic approach — to say the reverse, namely, that cognitive psychology reproduces the correct results of behavioristic theory and improves the incorrect ones. § 3 In addition to learning by model and to external reinforcement, a third behavior-modifying influence can be detected that is also of a cognitive nature. That is, self-activated reinforcement. In the latter case, one establishes definite standards of behavior for oneself and gives oneself rewards or punishments according to the degree one complies with these standards. The details of this process are described in motivation theory. In this theory, motives have the conceptual status of hypothetical constructions and they are viewed as general dispositions of evaluating basic types of situations and anticipating their outcome. That is, motives express persistent idiosyncracies. 24 The achievement motive or the affiliation motive (i. e., the tendency toward social contact) are examples of such motives. Motivational structures are the result of experience over a long period of time. According to the kind of past individual experience, the motivational system is either of a more approaching or more avoiding nature. That is, it is either dominated by the hope of gratification or by fear of non-gratification. So each motive is divided into two tendencies; the relative strength of these parts differs according to the individual. 25 Motives do not lead directly to actions; rather, they must be stimulated by situations. So situations can be created experimentally in which two motives are activated such that a conflict results. Walker and Heyns used this fact to test motivation theory experimentally. 26 They measured the relative strength of the affiliation and achievement motives of individuals in a group of test subjects; 27 then they constructed situations in which both motives were stimulated equally and predicted individual behavior on the basis of the relative strengths of the two motives. These predictions were confirmed. 24 25 26 27

Cf. Heckhausen 1974b, 1 4 2 - 1 4 3 . Cf. Heckhausen 1974b, 1 4 6 - 1 4 7 . Cf. Walker/Heyns 1962, 5 4 - 6 8 . The Standard method for measuring achievement motive is the so-called thematic apperception test (TAT) (for a description, cf. Weiner 1972, 1 7 5 - 1 7 9 , 1 8 4 - 1 9 2 ) . To measure the affiliation motive, Walker and Heyns used a method analogous to TAT. Cf. Walker/Heyns 1962, 5 7 - 5 9 .

136

Science and Mind

In addition to specifying motivational structures, motivation theory attempts to reconstruct the precise influence of motivations on behavior. The theory of cognitive dissonance, which was developed primarily by Festinger, should be mentioned first. It seeks to explain a particular motivational process and its effects on behavior. The theory assumes that cognitive dissonance is a motivational state that always occurs if there is a psychological conflict, for example, between two incompatible beliefs or between attitude and behavior. This motivational state induces behavior appropriate to reduce dissonance. In a conflict between attitude and behavior, for example, either the attitude or the behavior is changed in such a way that harmony between them is attained again. The extent of cognitive dissonance does not depend solely on the degree of conflict between the belief and the action, but is also determined by the availability of justifications. For example, if good external reasons are found for behavior that contradicts one's own beliefs, then the tendency to change the beliefs to suit the behavior is reduced. This central prediction of dissonance theory was confirmed experimentally by Brehm and Cohen (1962), for example. Brehm and Cohen asked students to write essays which contradicted their true opinions (which had been determined in advance) and rewarded them for their efforts with money in varying amounts. Finally, the test subjects filled out questionnaires, which they were told would be anonymous, to collect their true opinions again. According to dissonance theory, there is a conflict here between opinion and behavior, and behavior is fixed by the conditions of the experiment. So the dissonance could be reduced in this case only by changing the attitudes. The theory predicts in that case that the less persuasive the justifications for the conflict-inducing behavior, the greater the difference between the opinions expressed in the questionnaires (which were influenced by the experiment) and the original opinions. The justifications, in this case, were the financial rewards. The prediction was that the smaller the reward, the greater the change in opinion induced by the experiment. This prediction was confirmed. 28 The experiment obviously leads to the interpretation that the less the reward, i. e., the Skinnerian reinforcement, the greater the induced effects. Dissonance theory points out the existence of anti-Skinnerian regularities.

28

For this presentation of the theory, cf. Lefrancois 1972, 57 - 62.

T h e E x a m p l e of Cognitive Psychology

137

§ 4 Cognitive dissonance represents only one type of motivating influences. Other motives and motivational processes are described by motivation theory as well. T h e achievement motive will be used as an example here, since it has been the most exhaustively researched. Heckhausen's process model was pioneering in this area. 2 9 According to his model, the following two phases are to be distinguished: (1) T h e incentives contained in a situation activate the persistent achievement motive; the consequences of an action are anticipated and evaluated {motivation). (2) T h e performance of the action is followed by the selfevaluation phase. Motivation theory aims to explain these intermediate cognitive processes, i. e., motivation and self-evaluation. Like all motives, the achievement motive is also made up of two tendencies — the success motive and failure motive. These two partial motives express the general hope for success and the general fear of failure. Success-motivated people actively seek situations with achievement opportunities hoping to succeed; failure-motivated people avoid such situations. T h e tendency to seek achievement situations in the hope of success is expressed as the product of the intensity of the success motive and the motivational strength. T h e latter is itself expressed as the product of the subjective incentive value and the subjective probability of the outcome sought. 3 0 T h e incentive value is, in its turn, determined by anticipating and assessing the achievement action and its outcome, as well as by the perceived difficulty of the task. At a moderate subjective difficulty, the incentive value is highest among success-motivated people and lowest among failure-motivated people. Self-evaluation of an achievement-related action is a function of two factors. A positive self-evaluation requires, on the one hand, that the person in question has a definite standard of quality, a level of aspiration, in the area in question, and that this level of aspiration is met or surpassed. On the other hand, the positive result must be attributed by the person to his or her own action. T h e procedure of determining the result's cause is called causal attribution. In general, four factors are brought into play, namely: talent, effort, the difficulty of the task and chance. Success-motivated people and failure-motivated people are characterized by different patterns of attribution. Success-motivated

29

F o r the following, cf. Heckhausen 1974b, 1 4 9 — 1 5 8 . Parts of this model are presented in English in Heckhausen 1967.

30

Cf. also Weiner 1972, 195.

138

Science and Mind

people attribute success to talent and/or effort (they attribute internally) and failure to insufficient effort or to chance. Failure-motivated people display no definite attribution strategy for success, but they always attribute failure to lack of talent. Both the details of the motivational process presented here and the above mentioned distinction between success-motivated people and failure-motivated people can be confirmed experimentally.31 Motivational theory's reinterpretation of the so-called Pygmalion effect is particularly impressive. It goes back to an experiment by Rosenthal and Jacobson (1968), in which teachers' attitudes to certain students were systematically manipulated and the effects on the students' progress were examined. The teachers were told that certain students (who had actually been chosen randomly) had done particularly well on an intelligence test and that their schoolwork was expected to improve significantly in the near future. Neither the students nor the parents were told of this prediction. Eight months later, these students did, in fact, improve both their performance in school and on intelligence tests. Rosenthal and Jacobson attribute this surprising result to the induced formation of favorable reinforcement patterns on the part of the teacher: The teacher pays more attention to the favored student and thus notices his achievements more readily and reinforces them more consistently.32 However, it must be said that attempts to repeat the results of the experiment showed only limited increases in scholastic performance and did not reproduce the improved results on intelligence tests at all. 33 In contrast to Rosenthal and Jacobson's interpretation, which is based on Skinner, motivation theory interprets the Pygmalion effect as a change in the attribution pattern on the student's part, induced by a corresponding change on the part of the teacher. The teacher now rates the students abilities more highly and no longer attributes failure to a lack of talent but rather to a lack of effort. To the extent that the student takes the teacher's new perspective to heart, his own motivation to learn improves and, therefore, his scholastic performance does, too. This interpretation leads to the prediction that a Pygmalion effect is only to be expected in the case of students whose intelligence test results would lead one to expect greater scholastic performance than they actually 31 32 33

Cf. Schneider 1976, 3 3 - 5 9 . Cf. Weinert 1974, 3 3 - 3 5 . Cf. Heckhausen 1974c, 571.

T h e Example of Cognitive Psychology

139

show (so-called "underachievers"), and only if this discrepancy is based on a lack of motivation. Hence, by empirically ascertaining the motivational structure and learning abilities, one can predict which students will show a Pygmalion effect and, especially, which students will not — in contrast to Rosenthal and Jacobson's explanation. This prediction has been confirmed. 34 Thus, motivation theory both explains the failure of the behavioristic interpretation (namely, the failure to reproduce the Pygmalion effect without qualification) and predicts a new kind of regularity (namely, the appearance of the Pygmalion effect only under certain circumstances). § 5 Our excursion through some areas of modern psychology makes clear — despite its brevity — that behaviorism is outdated for empirical reasons. This judgment can be supported and made more precise by applying Lakatos' methodology of research programs to the evolution of psychological theory as presented above. Lakatos regards science as competition among research programs. A research program is a series of theories (of versions of the program) that share the fundamental postulates (the "hard core" of the research program) and the guidelines for further theoretical development ("positive heuristics"). A methodologically qualified ("progressive") research program has the following features: (1) The program evolves in accordance with the requirements of positive heuristics. (2) Each variant of the program reproduces the successful explanations of the previous version and (3) predicts new and unexpected regularities. (4) At least a few of these predictions are experimentally verified. 35 Whenever a program's progressive (in this sense) explanatory achievements exceed that of a less successful alternative program, then the latter is superseded by the former. 36 If these criteria are applied to the competition between behaviorism and cognitive psychology, then it becomes clear that cognitive psychology can explain the effectiveness of Skinner's reinforcement mechanisms (by interpreting reinforcements as information and motivation) and, insofar as that is the case, reproduces the explanatory achievements of behavior34 35

36

Cf. Heckhausen 1974c, 5 7 1 - 5 7 3 . T h o u g h the demand for predicting novel phenomena is mitigated later, in the light of these weaker criteria as well such predictions increase the empirical support for a research program to a considerable degree. For a discussion of the various criteria entertained, cf. Carrier 1988, 2 0 5 - 2 1 6 . Cf. Lakatos 1970, 1 1 6 - 1 2 5 , 1 3 2 - 1 3 8 .

140

Science and Mind

ism. Moreover, it predicts new phenomena that were either unexpected or precluded by Skinner's theory and which can, in fact, be verified experimentally. Among these phenomena are (1) the detection of a classical conditioning that is brought about exclusively by cognitive means, (2) the prediction — derived from the theory of cognitive dissonance — that the induced change in behavior increases with decreasing intensity of reinforcement, (3) the prediction of a corrected version of the Pygmalion effect which, at the same time, reveals the reasons for the failure of the behavioristic interpretation of this effect. Thus, cognitive psychology explains everything that behaviorism could explain; moreover, it explains why behaviorism failed in particular cases, and, finally, it leads to empirically verified predictions of laws. Thus, in the light of Lakatos' criteria for scientific progress, behaviorism is superseded by cognitive psychology. § 6 It remains to be clarified whether the replacement of behaviorism by cognitive psychology constitutes a reduction of the former to the latter or rather a non-reductive theory substitution. In the former case, it would be an ontological reduction. The conclusion of the previous discussion would be that Skinner's reinforcements function in a different manner that Skinner himself assumed. In the case of a non-reductive theory substitution, however, one would conclude that Skinner's reinforcement does not exist at all. To answer the question of reducibility, we must examine the possibility of relative interpretability. In doing so, it should be recognized that cognitive psychology in no way denies that reinforcements are effective. Thus, Skinner's laws of behavior do, indeed, find their place in modern psychology. On the other hand, they don't hold without restrictions, but only under certain conditions. First, there are additional reinforcements of a kind different from that envisaged by Skinner (vicarious reinforcement, for example); second, all reinforcements are interpreted in a different fashion than that supposed by Skinner, namely, as information and motivation. The latter point means that reinforcements must be noticed in order to be effective; they cannot shape behavior blindly, as it were. This implies that the relativizing conditions relevant here are roughly as follows: Particular constellations of stimuli - such as those typical of modeling, for example — must be absent. Moreover, particular cognitive factors — such as conscious perception of reinforcement patterns — must be present. If these conditions are met, reinforce-

The Example of Cognitive Psychology

141

ments do, indeed, function in the way Skinner described. According to this (admittedly rough) analysis, Skinner's theory of behavior is, in fact, relatively interpretable in cognitive learning theory and motivational psychology. 37 Finally, we must ask whether the relativizing conditions are satisfied within the domain of Skinner's theory. According to our discussion of the concept of reduction in II.2.§ 6, there is no single strict answer to a question of this sort. Nevertheless, it is plausible to assume that nonSkinnerian reinforcements do not dominate human behavior, but are only significant in a minority of cases. 38 The same is certainly true of the assumed cognitive factors. We may assume that our information about the prevailing patterns of reinforcement is by and large correct, so that cognitively anticipated and actually realized patterns of reinforcement coincide. If this assumption is correct, then it would be proper to speak of a reduction of behaviorism to cognitive psychology. Thus, Skinner's reinforcements would exist, but they would be of a different nature than Skinner assumed — just as in the case of psychophysical reduction where mental events do, in fact, exist but are of a different nature than is often assumed. This analysis shows that the sciences of human behavior do not need special methodological standards. After all, psychology has produced precise, experimentally verifiable statements, and even predictions of new and unexpected laws. We regard this as an indication of the methodological unity of all the empirical sciences and as support for our argument against Davidson's proclamation of a special status for psychological discourse. We hope this has made it clear that (1) general methodological standards prevail in psychology as well, and that (2) the best available theories of human behavior, seen from the perspective of these standards, operate with mental terms.

37

38

It could be objected that we have not shown that Skinner's laws of behavior — such as: minimal extinction of behavior (i. e., "forgetting") can be achieved by a variable ratio reinforcement schedule — can be derived cognitively. In fact, such a derivation is not even possible on the basis of Skinner's theory itself. Rather, these laws are regularities which were disovered experimentally using Skinner's conceptual framework. Moreover, one might point out that the theory claims to be valid primarily for animal behavior. In any case, the model was developed on the basis of experiments with pigeons and rats.

VI. Philosophy and Mind: Psychophysical Dualism 1. The Logical Status of Psychological

Terms

§ 1 The foregoing short sketch of some theoretical developments in modern psychology reveals the type and structure of mentalistic terminology with which the best available theories of human behavior operate. We shall now discuss several characteristics of this terminology and of the structure of psychological theory as a whole. In doing so, one finds that, in all of the cases discussed, psychological terms are of an intentional nature. For example, motives are representations of expectations and goals and as such express particular contents. The same is true of causal attributions which express beliefs about causal origin. To this extent, the models sketched here employ a vocabulary which satisfies the traditional criteria for mental concepts. Thus, where conceptual structure is concerned, psychological theories (insofar as they were discussed here) are more conservative than behaviorism or a possible neuropsychology whose fundamental terms are or are supposed to be non-intentional. In this context, one may ask whether or not a non-intentional mentalism makes sense and whether cognitive psychology is necessarily committed to an intentional vocabulary. We will return to this question in VII.3. In addition, this outline of some theoretical models in psychology reveals that psychological laws have a different structure than is sometimes assumed. Thus, one encounters the assessment now and again that human individuality, i. e., the diversity and variability of human behavior, precludes the possibility of psychological regularities of the sort found in the natural sciences. At best, it is said, one could provide average values, i. e., rules for average behavior, from which significant variations would always be possible for particular individuals. This objection presupposes that psychological laws are formulated in classificatory concepts. Classificatory concepts-express whether or not an object belongs to a particular class. A psychological regularity of this type

The Logical Status of Psychological Terms

143

would take the form: "Extroverted people like strawberry ice cream." By contrast, comparative concepts establish an ordering among objects and, therefore, are capable of expressing more sophisticated and differentiated distinctions than classificatory concepts. Such a law framed in comparative concepts would take the form: "The more extroverted a person is, the stronger his preference for strawberry ice cream." In this more differentiated version, individual differences come out more clearly than in the classificatory version which does not recognize degrees, but only alternatives. The laws discussed above are precisely of the latter type. Thus, for example, in the Walker and Heyns experiment the behavior predicted depends on the relative strength of the relevant motives; and the more the Pygmalion effect is marked, the stronger the failure motive is present in the test subjects before the experiment. Thus, human individuality is compatible with the strict validity of psychological laws. § 2 The status of mental terms, as it follows from an analysis of cognitive psychology, will be reconstructed in the following; and in doing so, we shall make more precise and give some support to the contention already made several times, namely, that they are theoretical terms in the sense used in the philosophy of the natural sciences. Our thesis is that this is precisely where we find the kind of theoretical terms for psychology to which Feigl rightly aspired but which he wrongly sought in neurophysiology (cf. II.1.§ 1). Cognitive — not neurophysiological — concepts are the adequate hypothetical constructs for psychology. As was roughly sketched in II.1.§ 1, theoretical terms are terms whose meaning derives primarily from their integration in a theoretical structure, not from a connection to observables. This means that theoretical terms are not introduced by linking them one by one to specific observations or measurements: Rather, the constructs used in a theory are introduced jointly, as it were, by setting up a theoretical system formulated in terms of them and by giving this system an experiential interpretation, which in turn confers empirical meaning on the theoretical c o n s t r u c t s . 1

This means that theoretical terms are introduced by the basic postulates of the pertinent theory; they do not directly designate observable entities. The theory obtains an empirical interpretation when some of the theoretical terms it introduces are assigned empirical indicators. • Hempel 1952, 32.

144

Philosophy and Mind: Psychophysical Dualism

This assignment is carried out by means of so-called correspondence rules, each of which connects at least one theoretical term to one observation term. It is also possible for several correspondence rules, i. e., several empirical indicators, to be assigned to one theoretical term; and this assignment can change in the course of scientific progress. In other words, scientific progress might cause certain empirical indicators of a theoretical state to be eliminated and others to be introduced. 2 The application of theoretical terms is, therefore, determined by two factors: by their role in the network of the theory and by their connection to the empirical basis; or, in Feigl's words, their meaning is fixed by "«¡angulation in logical space." 3 Thus, theoretical terms contain a "surplus meaning" which results from the functional context of the pertinent theory, i. e., a meaning above and beyond their operational content. They are "open" terms, whose observational indicators can change in the course of scientific progress. The introduction of theoretical terms is based on the insight that such a flexible terminology is better suited to give the simplest possible systematization of empirical regularities than terms rigidly fixed to empirical indicators. It is primarily methodological and not epistemological reasons which suggest the use of theoretical terms. The particular character of theoretical terms and their role in formulating theories will now be discussed using two examples, one psychological and one physical. In particular, we try to show that theoretical terms are not synonymous with the indicators assigned to them, that they cannot be translated into these indicators. The presence of indicators is not logically sufficient to conclude that the corresponding theoretical term applies. 4 More precisely, this means that the conditions for applying theoretical terms are not specified exclusively in observation terms, but must in turn have recourse, at least partially, to theoretical terms. We hope to show by analyzing these examples, first of all, that the state of affairs in the physical case is very similar to that in the psychological case. This supports our thesis that mentalistic vocabulary in psychology should be considered legitimate theoretical terminology. Second, this discussion of scientific language should make it clear that mental terms represent an independent linguistic level, i. e., that they cannot be translated into the physical or physiological indicators that 2 3 4

For the concept of correspondence rules, cf. Wolters 1984b, 4 8 0 - 4 8 1 . Cf. Feigl 1963, 321. Carnap already called attention to this. Cf. Carnap 1956, 69.

The Logical Status of Psychological Terms

145

are assigned to them. The analysis of the psychological example starts with Feigl's argument against a reductionistic behaviorism (cf. II.1.§ 1, V.l.§2); in the physical example, we examine the concept-indicator relation in electrodynamics. § 3 In the attempt to connect a psychological state, e. g. a particular state of motivation, to a particular behavioral indicator the difficulty lies in evaluating the influence of other psychological states that could also influence behavior. For example, a particular state of motivation *Pm may actually exist, but the performance of the associated behavior might be inhibited by the simultaneous presence of the state of fear Y f . In order to determine if 'PM is present, therefore, it must be known whether *Pf is realized. Generally speaking, this means that in order adequately to attribute a psychological state to a person, its interaction with other psychological states must be known. Any inference from the observed behavior BM to the mental state I'M presumes other psychological states that could influence the connection of I'M and BM do not fluctuate uncontrolledly. Psychological concept-indicator relations hold only given a ceteris paribus clause which must itself use mental concepts. The correspondence rules are formulated with provisos which are in turn couched in mental terms. This mentalistic ceteris paribus clause can only be tested if reliable indicators can be assigned to the possible distorting influence ^ i , and this assignment must itself take into account the influence of other distortions (possibly including Tm). These are the peculiarities in the application of theoretical terminology that we referred to when speaking of the theory-specific application conditions of such terms. The measurement of theoretically determined psychological states is decisively influenced by the theory in which they appear. This means that laws enter into the relation of concept and indicator, i. e., into the correspondence rules. On the one hand, these laws first establish certain connections by permitting one to infer a psychological state *FX (e.g., thè presence of cognitive dissonance) from a behavior Bx (e. g., a change in opinion in the dissonance experiment). On the other hand, the laws limit the applicability of these connections by tying them to certain mentalistic ceteris paribus clauses (such as the lack of a strategy of justification). In this sense, the meaning of theoretical terms is influenced by their role, by their "functioning" in a theory; the formulation of correspondence rules for a theoretical term is dependent on the nomological relations of this term to other theoretical terms.

146

Philosophy and Mind: Psychophysical Dualism

All the essential features of the psychological example are encountered in physical theory as well. For example, the electrical field strength E can be measured by the acceleration a of a particle with a charge q and mass m, using the equation E = m a / q . T h u s , the acceleration (or the force ma) represents the empirical indicator of the theoretical entity "electrical field." W h a t is significant about this is that it makes little sense to consider this theoretical term to be defined by the indicator. T h e electrical field strength can only be obtained from the acceleration if there are no other accelerations — such as those caused by the magnetic field — which distort the results. This shows h o w an additional theoretical entity enters into the concept-indicator relation, with the result that this relation is only valid under a ceteris paribus clause, which employs other theoretical terms. In order to determine field strength f r o m acceleration, the relevant accelerations have to be distinguished; and what is relevant can only be determined by a theory of electrical fields. This example shows, moreover, that only an empirical measurement of the product of charge and field strength can be obtained in this way; an approach for obtaining field strength separately requires the assumption of a constant charge (which is also to be considered a theoretical entity). If one wants to measure electrical field strength, one needs nomological assumptions about other electromagnetic entities (namely, electrical charge and magnetic field strength), which effect or are effected by the entity being sought. Here, too, laws are essentially involved in the concept-indicator relations. T h e result of our analysis of these examples is that mental terms play the same role and show the same characteristics in psychology as theoretical terms in physics. Both the fact that correspondence rules contain certain provisos which are themselves expressed in theoretical terms as well as the peculiarity that laws even enter into the correspondence rules themselves can be seen in both cases. We conclude f r o m this that mental terms in psychology are theoretical terms. § 4 T h e introduction of theoretical terms is not necessary for epistemological reasons; a strict assignment, i. e., an identification of concept and indicator, could be achieved by a conventional fiat. Feigl, too, emphasizes this possibility, but hints at the methodological cost to our explanatory and predictive powers that such a strategy would imply: A strict o n e - t o - o n e correspondence of indicator variables and intervening variables [i. e., theoretical terms] can be achieved by arbitrary definition — but the severely limited success of the prediction and

T h e Logical Status of Psychological Terms

147

explanation of behavior on this basis shows convincingly that we must search by whatever means available for a more complete frame of concepts and laws. 5

T h a t the damage that comes from rejecting theoretical terms would be of an essentially methodological nature, is also one of the central insights of Hempel's pioneering study of how theoretical terms can be defined by observational terms. 6 Hempel also emphasizes the loss of explanatory and predictive power that would go along with insisting on all terms being strictly tied to the observation basis. Moreover, he points out the limitations to the heuristic fruitfulness of theories that such a strategy would introduce. 7 Thus, the admission of theoretical terminology is, in the end, determined by the commitment to certain methodological criteria (such as explanatory power). Cognitive concepts in psychology are to be seen as explanatory constructions to which behavior, reports about inner states and neurophysiological phenomena are attributed as indicators. Cognitive concepts and their indicators are not synonymous. Behavior or neurophysiological states indicate the applicability of cognitive concepts, they refer to the presence of the corresponding psychological phenomena, but are not identical to them. T h u s , while Feigl assumes a one-to-one correlation of the phenomenal predicate P, the psychological predicate *P and the neurophysiological predicate ® (P «-»• ) (cf. II.1.§ 4), we represent the situation as follows: Theoretical Level

Psychological State

Correspondence Rules

Empirical Level

Neurophysiological State

Introspective Experience

Behavior

Fig. 3 According to this reconstruction, the neurophysiological state is merely an indicator of a corresponding psychological state, not its criterion (as it is for the identity view). 8 5 6 7 8

Feigl 1 9 5 1 , 1 9 9 - 2 0 0 . Hempel 1958. Cf. Hempel 1958, 2 1 3 - 2 1 9 . F o d o r seeks to base an argument for materialism on viewing psychological states as inferred entities (i. e., as theoretically determined states). Since behavior is not a

148

Philosophy and Mind: Psychophysical Dualism

2. The Ontological Status of Mental States § 1 After having reconstructed the logical status of mental terms and having demonstrated that such terms can legitimately be considered theoretical terms, we will now discuss the ontological consequences of this result. The criteria for establishing an ontology have changed fundamentally in the course of the past few centuries. It had been assumed previously that the ontology must be established by general metaphysical or epistemological principles which in turn determine the entities that science can legitimately introduce. In more recent developments, by contrast, the idea that the answers to ontological questions follow upon scientific theories has gained acceptance. This change in the philosophical approach will be briefly outlined using the example of the discussion of the nature of space. Then the consequences of this new orientation for psychophysical ontology will be expounded. § 2 The two basic theories of space can be thought of as a relationalist and a substantivalist view. The relationalist theory accepts only spatial relations between phenomena as existent and considers the idea of space as an entity independent of objects to be an inadmissible hypostatization. The substantivalist or container view, on the other hand, assumes a spatial order which is independent of all phenomena and ontologically prior to them. For the relationalist, objects constitute space, for the container theoretician, objects are in space. Leibniz is one of the classical advocates of the relationalist view. Leibniz's defense of relationism as against the concept of absolute or container space, which in Newtonian mechanics serves to define absolute motion and, hence, inertial motion, 9 is based on the two principles of sufficient condition for the application of such terms, only neurophysiological processes can still be considered. Due to the lack o f other possibilities, the latter are the only remaining, plausible, sufficient observational criteria for attributing mental events (cf. F o d o r 1968, 99). F o d o r ' s argument assumes that at least one sufficient condition of application must be given in observational terms for all sensible theoretical terms. However, as was shown above in detail — and already emphasized by C a r n a p with

9

explicit reference to theoretical terms in psychology (cf. C a r n a p 1956, 71 — 72) —, this is not correct. O n e of the characteristics of theoretical terms is precisely that their applicability can never be conclusively determined solely by means of observables. N e w t o n 1687, 5 (Scholium). N e w t o n attempts to show that the centrifugal forces that o c c u r during rotational motions c a n n o t be attributed to relative rotation and must, therefore, be understood as " t r u e " rotations, i. e., as rotations with respect to an

The Ontological Status of Mental States

149

the sufficient reason and the identity of the indiscernibles. If space were substantial, Leibniz argues, then a situation could come about for which it would be impossible to give a sufficient reason: All points in space are of equal status. If they were taken to be independent entities besides representing the order of objects, then two states in which the position of all objects remained unchanged but where the directions (east and west, for example) were switched, would have to be considered different states. But there is no reason whatsoever why either the one or the other should be preferred. 1 0 On the other hand, if both indiscernible states were taken to be identical (as a relationalist theory maintains), then the principle of sufficient reason remains unaffected. Relational space, in Leibniz's analysis, becomes the "abstract space" of the "order of all possible positions" (meaning the spatial relationships between physical bodies). 1 1 While Leibniz bases relationalism on metaphysical principles, M a c h adduces epistemological principles in its support. A body's motion in an ontologically prior, substantial space is completely imperceptible, and absolute or container space is therefore empirically inaccessible. When we speak of a body's motion, it only becomes a definite statement with respect to reference bodies: "In reality, therefore, we are simply cognizant of a relation of the body K to [the reference bodies] A, B, C . . . " 1 2 T h a t we can change these reference bodies creates the illusion that we could do without them. In fact, however, every means would be wanting of forming a judgment of the behavior of K and of putting to the test what w e had predicted — which latter therefore w o u l d be bereft of all scientific significance. 1 3

T h u s , Mach's relationalism rests on epistemological considerations about the verifiability of statements.

10

11

12

13

absolutely immobile space. The concept of absolute space would, thus, also have experimental content. For more on this and for a history of the concept of space, cf. Janich/Mittelstrass 1973, 1160-1161. 3rd Letter to Clarke, Leibniz/Clarke 1715/1716, 53 ( = Leibniz 1875-1890, VII, 363-364). 5th Letter to Clarke, Leibniz/Clarke 1715/1716, 171 ( = Leibniz 1875 - 1890, VII, 415). Cf. also the famous definition of relational space in the letter to des Bosses dated June 16th, 1712: "space is the order of simultaneously existing things, as time is the order of consecutive ones" (Leibniz 1875-1890, II, 450). Mach 1883, 281. Ibid. For Mach's critique of the Newtonian conception and for the formulation of his principle of inertia, cf. Wolters 1987, 3 7 - 4 9 .

150

Philosophy and Mind: Psychophysical Dualism

A common feature of both approaches is that they seek to support relationalism with arguments that precede any particular theory. The terminology and ontology of science are preemptively restricted. What science may reasonably assume as existent is determined without reference to scientific results. The modern alternative to this strategy is to discuss ontological questions with reference to the best available theories. In doing so, one does not demand that single, isolated concepts be adequate in the light of philosophical postulates, but rather assesses theories as a whole according to methodological standards, in order to examine the ontological consequences of the theories that satisfy these standards. Thus, the adequacy of relationalism, for example, would be judged on the basis of an interpretation of Einstein's gravitational field equations. It could be argued, for example, that the fact that these equations contain oscillations of the metrical field (gravitational waves) even in a vacuum shows that there could be space (i. e., a metrical field) even where there is no matter. This would then imply that the relationalist position is inadequate. The same result is obtained if one bears in mind that acceleration is independent of the frame of reference in question, in the General Theory of Relativity no less than in Newtonian mechanics. Accelerations, therefore, should not be considered (or at least not without additional reflection) relative accelerations, as the relationalist theory of space would have it. 1 4 § 3 This twist in the ontological discussion was introduced primarily by Quine. According to Quine, a theory considers those objects existent which it must presuppose for its own validity, that is, those objects to which at least one predicate of the theory must be truthfully applicable in order for the theory to hold. In Quine's words: T o be assumed as an entity is, purely and simply, t o be reckoned as the value of a [quantified] variable. [ . . . ] T h e variables of quantification, " s o m e t h i n g , " " n o t h i n g , " " e v e r y t h i n g , " range over o u r whole ontology, whatever it may be; and we are convinced of a particular ontological supposition if, and only if, the alleged presuppositum has t o be reckoned a m o n g the entities over which o u r variables r a n g e in order to render one of our affirmations t r u e . 1 5

Of course, such a discussion would have to be much more subtle. This argument is only intended as an example of the type of argument, is Quine 1953, 13.

14

T h e Ontological Status of Mental States

151

When we speak of the cross-breeding of zoological species, for example, we must recognize species as existing entities, at least if we do not succeed in paraphrasing the corresponding statements such that there is no longer any mention of species. In general, it is true that: a theory is committed to those and only those entities to which the bound variables of the theory must be capable of referring in order that the affirmations made in the theory be true. 1 6

The turn in the ontological debate in the philosophy of science that Quine introduced has meanwhile developed into so-called scientific realism. According to this position, the usefulness of a theoretical apparatus is the criterion for judging whether its theoretical terms genuinely refer, i. e., whether they designate objects that actually exist. This means that the theoretical terms of our most successful theories are genuinely referential. Leplin mentions the following as two of the fundamental claims of scientific realism: The central terms of the best current theories are genuinely referential. [...] Scientific theories make genuine, existential claims. 1 7

For example, since our best theories of matter operate with electrons, electrons exist. The crucial point for our context is that mental entities also stand up to this criterion. Our best available theories of behavior use cognitive concepts as theoretical terms. Hence, the psychological states characterized in this way exist. Scientific realism is not incontrovertible. The question whether theoretical terms can be meaningfully said to be genuinely referential still awaits an answer. But it is not our intention to intercede for this position. On the contrary, we prefer to speak of explanatory constructions rather than the discovery of theoretical entities. In fact, an argument like ours is in no way committed to scientific realism. Although it also assumes that empirically and methodologically qualified theories can supply us with a well-supported picture of the phenomenal domain they describe, still it should not be forgotten that this is a picture that we have drawn and that does not simply reproduce the order of things. Our argument does not commit us to the interpretation that science reveals the essence of nature that exists independently of all human effort. Rather, we believe the scientist is as much an 16 17

Quine 1953, 1 3 - 1 4 . Cf. also Quine 1960, 2 4 0 - 2 4 3 and Quine 1969, 9 3 - 9 8 . Leplin 1984, 1 - 2 .

152

Philosophy and Mind: Psychophysical Dualism

inventor or constructor as a discoverer: He constructs the conceptual means for describing the world, he produces them, he does not simply find them in the world or in nature; so the human image is reflected in these conceptual means. Even in the framework of a constructivist (in this sense) or instrumentalist orientation, the distinction can be upheld between theories that influence our picture of nature and those that do not. "Ontologically relevant" means in this case: "relevant to our picture of nature." Accordingly, this does not mean that this picture corresponds to a true world or a true nature detached from all human activity. Here we do not share scientific realism's more extensive claims. Hence, we must distinguish between two levels of reflection. On the one level, the point is that ontological questions are to be discussed on the basis of the best available scientific theories — and here we agree with scientific realism. On the other level, we maintain that ontological statements cannot be understood to be claiming the existence of a world that lies behind the phenomenal world and is independent of all human activity — and here we part ways with scientific realism. This disagreement with the view of scientific realism, however, has no relevance to the question at hand. The only crucial point here is that the existence of psychological entities follows from the same criteria as the existence of theoretically determined physical entities. Thus, if electrons exist, then failure motives also exist. Or, to put it a bit more crudely: if there is matter, then there is also mind. § 4 Our argument for introducing talk about psychological states and events treats these states and events as theoretically inferred. Recourse to them is justified, because they help us to achieve the best explanations of behavior and introspective reports. The empirical basis for attributing psychological states and events to other people is thus essentially intersubjective. This approach must be kept strictly separate from the procedure that uses induction and analogy to infer the existence of mental phenomena in others from one's own mental states and events. Such an approach by analogy as advocated by Feigl, 18 for example, is methodically very dubious, since its empirical basis is merely subjective and, moreover, extremely narrow since it only refers to a single person. 1 9 18 19

Cf. Feigl 1958, 380; also II.1.§ 4. For the distinction between an "inferred-entity approach" and an "analogy-approach," cf. Fodor 1968, 9 3 - 9 4 ; P.M. Churchland 1981, 69.

The Ontological Status of Mental States

153

Our argument in favor of psychophysical dualism essentially takes the form of an inference from mentalism to dualism. This inference is based on (1) the assumption that it is science that illuminates the nature of what there is, and (2) the consideration that, given the current state of research (and that of the foreseeable future), human behavior can best be explained by reference to psychological states and events. Finally, it is based (3) on the rejection of the "double talk, which would repudiate an ontology while enjoying its benefits" 2 0 that Quine lamented. Any discussion of the mind-body problem and any position in this debate, wherever it is located between interactionist dualism and eliminative materialism, essentially concerns existence and ontology. Our argument is that such an ontological debate should recognize the consequences of scientific research and not simply ignore them. For clarity's sake, it should be added that the sort of dualism we are advocating here should in no way be understood as embracing the doctrine of two substances. In our conception, unlike that of Descartes, properties are not inseparably tied to substances. To Descartes it seemed certain that a thought cannot exist without a thing that is thinking; and in general no act or accident can exist without a substance for it to belong to. 2 1

We do not envisage a substantial opposition of res cogitans and res extensa (i. e., of a thinking and an extended substance) and are not arguing in the tradition of the fluidum theories of the 18th century, in which phenomena such as heat, electricity and magnetism were explained by introducing special fluids, i. e., substances. We are not proposing a non-spatial mind-substance; rather, what is at issue here is the recognition of the independence or autonomy of mental states and events. Not substances, but processes are the focus of our argument. In this we follow Kant who criticized Descartes' hypostatization of processes and advocated an empirical, non-metaphysical dualism (cf. 1.2.§ 5). Moreover, it should be emphasized that there are in turn qualitative differences between psychological processes (analogous to those between physical ones). Thus, hoping for warmer weather has as little in common with thinking about Einstein's field equations as entropy has with electrical field strength. Thus, the position advocated here would be

20 21

Quine 1960, 242. Descartes 1641, 124 ( = Descartes 1 8 9 7 - 1 9 1 0 , VII, 1 7 5 - 1 7 6 ) . Cf. I.2.§ 1.

154

Philosophy and Mind: Psychophysical Dualism

more appropriately characterized as psychophysical process-pluralism. Schlick already mentioned this multiplicity of qualities: every sensible and philosophically honest world view must be pluralistic. For the universe is variegated and manifold, a fabric woven of many qualities no two of which are exactly alike. 22

What Schlick characterizes here as a "pluralistic world-view," and that is also expressed in Whitehead's formula "the universe is more various, more Hegelian," 2 3 has not only ontological meaning, but also — and, in our context, primarily — methodological meaning: Phenomena that appear different and are experienced differently should be treated as such, until their relatedness or identity is proved by experimental or theoretical means. This kind of pluralism is surely better suited to the manifold color and form of the world than is Parmenides' profession of faith in the One.

3. Dualism and its Problems § 1 Such an inference to dualism is, of course, only justified if this interpretation of the psychophysical relationship is not discredited for other reasons. We indicated earlier that a sensible distinction can be made between ontologically relevant and ontologically irrelevant theories even in the framework of an overall instrumentalist approach (cf.VI.2.$3). If it is found that there are important, independent arguments against a dualistic interpretation, then the ontological significance of cognitive psychology would have to be rejected. In that case, mental terms would merely be seen as useful instruments for explaining and predicting behavior. They would not refer to anything real. This kind of situation is not unknown in physics. In hydrodynamics (e. g. in the examination of the flow of fluids) or in technical mechanics (e. g. in the analysis of the warping of beams) arbitrary volume elements are considered and the forces acting on them are calculated. Thus, both approaches regard matter as a continuum; they disregard the atomic structure of the material. Despite the empirical success of these the22 23

Schlick 1925, 333 (emphasis in the original). Whitehead 1947, 118. On the Hegelianism that finds expression in this phrase of Whitehead's work, cf. Kambartel 1965, 7 2 - 9 8 .

Dualism and its Problems

155

ories — which is documented by the fact that our bridges do not generally collapse —, one would not conclude that matter has a continuous structure. Consequently, it is not adequate to take the fundamental entities of all successful theories to be ontologically relevant. In fact, precisely this kind of view is the basis of the functionalist interpretation of psychological regularities (cf. II.4.§ 2). Psychological natural kind terms have the status of terms like "valve lifter" and "mouse trap," which also do not designate physical kinds. It does not make much sense to grant ontological relevance to theories that use this kind of terminology. Only kinds of valve lifters or mouse traps realized in a particular manner can sensibly be assumed to exist and be considered physical (or technical) natural kinds. The functional category itself, on the other hand, is an abstractum and, hence, has no ontological significance. If functionalism correctly interpreted psychological laws, then our inference from mentalism to dualism would, in fact, be invalidated. But it would have to be shown that the functionalist interpretation is appropriate, which in turn presupposes that a multiple reduction claim is met. As the discussion in II.4.§ 7 made clear, such a claim is, on the one hand, implausible (at least at first sight) and, on the other, in fact not realized. Hence, this consideration does not constitute a sound objection to the ideas we have developed here. The same is true of a more concrete version of this objection: In describing computer processes, we use arithmetic or logical terms. If one wanted to transfer our interpretation of psychological terms to this case, one would have to say that there are, in fact, numbers or logical implications in computers. This interpretation would be obviously inadequate. What, then, is the difference between interpreting psychological terms and interpreting computer states? In the case of a computer, in contrast to the interpretation of psychological terms, the functionalist interpretation of internal states is completely justified. The same internal state (i. e., the same state of the program) can be realized technically in different ways. That the demand for multiple reductions is satisfied justifies the functionalist interpretation of internal computer states. We understand how a program state can be transformed into a physical process. For this reason, the argument we present here in support of a dualistic understanding of psychological states and events does not require that we attribute a mental life to computers.

156

Philosophy and Mind: Psychophysical Dualism

Of course, there could be other reasons for granting merely instrumental significance to mentalistic explanations. If dualism can be criticized with independent arguments that are based, for example, on possible internal inconsistencies or methodological inadequacies, or, conversely, monism could be based on considerations about methodological or conceptual plausibility, then any ontological interpretation of psychological entities would be inappropriate, even if the empirical and methodological qualifications of the corresponding psychological theories were beyond question and their reduction to neurophysiology were not in sight. Hence, in the following, we will attempt to determine whether any indications of the appropriateness of this sort of (as it were, non-Quinean) approach are present in the psychophysical case. First, we will examine possible reasons for disqualifying dualism and, in VI.4, take a look at possible independent arguments in support of a monistic interpretation. § 2 T h e emphasis of objections to dualism lies in the field of methodologgy. A dualistic treatment of the psychophysical relationship — thus runs the argument in various versions — lacks essential characteristics of a methodologically appropriate, scientifically acceptable approach. T h e soundness of such an objection depends crucially on which variant of dualism it is intended to attack. If mental states are seen as essentially subjective, qualitative, undetermined by laws and are contrasted to the material world, then we are certainly justified in wondering whether this strategy is methodologically irreproachable. Insofar as psychological states are introduced as capricious and fleeting entities, we are justified in mistrusting them. T h e same is true if one believes (with Bunge) that dualism is committed to an everyday description of psychological phenomena. Given these assumptions, it is not surprising that a dualistic approach is seen as being alien to the conceptual f r a m e w o r k of science. 24 Although criticisms of this kind may be sound when levelled against other dualistic positions, they still d o not constitute a valid objection to the version we have developed. O u r argument is not based on the anomaly of the mental, but rather on the assumption that psychological entities are subject to special laws. It does not emphasize the ability of the mental to escape all measuring procedures, but rather the special 24

Cf. Bunge 1980, 4, 33.

Dualism and its Problems

157

methods that can detect mental phenomena. The psychological entities that we are assuming here are the states of cognitive psychology. The psychologist who tests anxiety, spatial perception or the performance of the achievement motive, is measuring psychological states; psychological laws deal with such states. The methods for measuring them can be applied intersubjectively, their results can be reproduced, the laws can be tested. In view of these circumstances, it may be premature to deny that such approaches are scientific and to expect all salvation from neuropsychology. 2 5 The objection that dualism lacks simplicity is equally unconvincing. Bunge contends that dualism would require two kinds of states: one physical and one psychological, and this duplication "would go against the grain of contemporary science." 2 6 Of course, it is methodologically sound to keep the number of fundamental entities to a minimum. But it is also methodologically sound to use these fundamental entities to make statements with the greatest possible precision in the widest possible domain of application. A neurophysiological psychology has the advantage of greater terminological parsimony; on the level of empirical content, however, this turns into stinginess. While it is true that less is added at the top, nothing comes out at the bottom as a result. T h e theory is simplified before it really exists. This circumstance makes the heuristic argument in favor of monism less plausible. Bunge says that dualism hampers the progress of psychology, while monism encourages it, since it promotes cooperation between psychology and biology. 2 7 In fact, however, in recent decades, mentalistic heuristics has proved to be significantly more conducive to psychology than materialistic heuristics. After all, in the framework of neurophysiology, a cognitive mechanism is only acceptable if a neural counterpart

25

Bunge's defense of such a neuropsychology has, moreover, unfortunate consequences if one considers that psychology also has a psychotherapeutic branch. Bunge explicitly declares that all psychological illnesses are neuronal illnesses (Bunge 1980, 75, 79), so that effective psychotherapy must be a medicinal therapy: "sick neurons call for a biochemical (or psychopharmacological) approach instead of logotherapy: cells do not listen" (Bunge 1980, 79). Keeping in mind that medicinal therapy has, as a rule, proved to be less successful than psychotherapeutic treatment (combined with the treatment of the s y m p t o m s with drugs) (for the case of bronchial asthma, cf. Brautigam/Christian 1981, 166), and that medicinal f o r m s of therapy in general are accompanied by sideeffects, it becomes clear that Bunge's recommendation is inappropriate.

26

Bunge 1980, 33. Bunge 1980, 19, 24.

27

158

Philosophy and Mind: Psychophysical Dualism

can be associated with it. Thus, if materialistic heuristics had been followed consistently, it would have prevented the evolution of cognitive psychology entirely. To all appearances, Feigl's "extrapolation" of the evolution of science (cf. V. Introduction) leads into a quite different direction than Feigl assumed: it leads one to expect cognitive psychology to be more successful than a neurophysiological psychology. Hence, contrary to the views of Bunge and Feigl, it appears a mentalistic approach is heuristically fruitful. § 3 The traditional difficulty associated with a dualistic interpretation is to provide a plausible reconstruction of the status of psychophysical laws. As a rule, a causal relation exists between a theoretical state and its indicators. For example, an electrical field accelerates charged particles; a bound electron's quantum jump causes light to be emitted. Of course, a non-causal common variation of two entities is also possible, as is the case with kinematic laws such as the law of falling bodies. In that case, there is a nomological relation between the displacement and the time interval, without its being possible to interpret sensibly one of the entities as the cause of the other. Obviously, these two possibilities correspond to the interactionist and the parallelist positions in the mind-body debate. We ourselves are inclined to an interactionist interpretation. At first glance, this is certainly the most obvious assumption: Mental events like goals and beliefs influence our actions, or so it seems; conversely, the dependence of the psychological condition on neurophysiological and somatic states also suggests a causal relationship. A parallelist interpretation would have to insist that psychological processes can be explained solely by recourse to other psychological processes (cf. 1.2.§ 3). It would have to advocate the view that the appearance of psychophysical interaction is based on an illusion and that closer observation reveals the inner-psychological or inner-physical causes for the apparent interaction. For example, all (so-called) psychosomatic diseases would have to be explained exclusively by reference to physiological phenomena. The assumption of this kind of autonomy of the physical and the psychological domain is, however, not justified at present and must be considered implausible. The difficulty of a parallelist interpretation is that the common variation of entities without mutual causation only seems to occur in natural processes if these entities are effects of a common cause (as in the case of falling bodies). But we have no idea what sort of common

Dualism and its Problems

159

cause there could be in the case of the psychophysical relation; after all, Spinoza's divine substance could hardly claim to be plausible anymore. And the assumption of a parallelism without recourse to any causal relation is without parallel in nature. In that case, the mind-body relation would, in fact, be unique, incomparable and, hence, not analyzable (as Shaffer, for example, assumes 28 ). But this resignation can only be a last refuge after all other alternatives have failed. Therefore, our next question will be: how viable is psychophysical interactionism? In the traditional discussion, the fundamental difficulty in the assumption of mind-body interaction is thought to be the deep-rooted difference of mind and body; there can be no connection which bridges the ontological abyss between the psychological and the physical. This idea is based on the view that only similar things can effect one another; causality requires that cause and effect be of a similar nature. The historical origin of this view is in Empedocles, who traces everything back to the interaction of related materials: So sweet seized on sweet, bitter rushed to bitter, sharp came to sharp, and hot coupled with hot. 2 9

This idea of causality, whose influence can be seen reaching from antiquity to modern times, and which says that like only effects like, is also found in Democritus. According to Aristotle, he asserts that agent and patient are identical, i. e. "like." It is not possible (he says) that "others," i. e. "differents," should suffer action from one another: on the contrary, even if two things, being "others," do act in some way on one another, this happens to them not qua "others" but qua possessing an identical property. 3 0

This theory of causality is also found in Leibniz's view of the equality of cause and effect and also — as evidence of terminological continuity — in the formula "causa qualis, talis esse semper effectus solet" (The nature of the cause always coincides with the nature of the effect). 31 The chemistry of principles of the 17th and 18th century also assumed a 28 29

30

31

Shaffer 1967, 345 (cf. Introduction above). VS 31 B 90; cf. B 62, [Empedocles 231] for an epistemological application: B 109 ("With earth we perceive earth, with water water, with air divine air, with fire destructive fire, with love love, and strife with baneful strife" [Empedocles 233]). De gen. et corr. A 7 . 3 2 3 b l 0 - 1 5 ( = VS 68 A 63) [engl. 323]. Cf. Democritus, VS 68 B 164, A 37, A 38, A 99a; Leucippus, VS 67 A 1 (za 5(xoia Ttpoi; xa S^oia). Gruter 1625, 142. For Leibniz, cf. Leibniz 1676. Also Breger 1984, 118.

160

Philosophy and Mind: Psychophysical Dualism

particular attraction of the like and used it to explain the cohesion of materials. "Gleich und gleich schickt sich zusammen" (The like appropriately combines with the like), assured Stahl in 1703. 3 2 It is no longer disputed today that a valid argument against psychophysical interaction cannot be based on the dissimilarity of mind and body. The similarity or dissimilarity of cause and effect no longer plays a role in the modern understanding of causality. Rather, the constant conjunction of events (in Hume) or their functional dependence (in Mach) stand at the center of modern theory of causality. The objection of dissimilarity collapses with the decline of the outdated doctrines of causality. Schlick already noted: But even if the physical and the mental w e r e in fact t w o different domains of the real, n o difference in kind, however great, could constitute a serious obstacle to the existence o f a causal relation between them. F o r we k n o w of n o law stating that things must be of the same kind in order t o act on one another. O n the contrary, experience everywhere s h o w s that the m o s t disparate things stand in a relation of dependency t o one another and thus interact with one another.33

It is implausible to demand that causally interacting entities be ontologically similar. This does not, however, permit one to conclude conversely that psychophysical interaction would not require any additional explanation. Thus, Ducasse argues that only causal connections between spatially distant events must be explained by intermediate processes. But to inquire after intermediate steps in the case of contiguous causal effects, as in the case of psychophysical interaction, would amount to a category mistake. 34 In contrast to such an idea, we would rather agree with Feigl who sees the central puzzle of the mind-body problem in the question

32

33 34

Stahl 1703, 38. For a short summary of the theory of bonding advocated by the chemistry of principles, cf. Carrier 1986c, 3 3 3 - 3 3 7 . Schlick 1925, 301. A similar argument appears in Shaffer 1967, 341; Puccetti 1985, 49. Ducasse 1960, 88. In fact, this is precisely the thesis expressed by Newton/Clarke against Leibniz. For Newton/Clarke, the psychophysical gap in perception processes is bridged by the presence of the mind: "the Mind of Man, by its immediate presence to the pictures or Images of things, form'd in the Brain by the means of the Organs of Sensation, sees those Pictures as if they were the Things themselves" (Leibniz/Clarke 1715/1716, 2 9 - 3 0 ( = Leibniz 1 8 7 5 - 1 8 9 0 , VII, 353), cf. 48, 70, 113 ( = Leibniz 1 8 7 5 - 1 8 9 0 , VII, 360, 369, 386)). For Leibniz's response: Leibniz/Clarke 1715/1716, 37, 55, 95 ( = Leibniz 1 8 7 5 - 1 8 9 0 , VII, 3 5 6 - 3 5 7 , 365, 3 7 5 - 3 7 6 ) .

D u a l i s m a n d its P r o b l e m s

161

of the logical nature of psychophysical correlation l a w s . 3 5 O n this point a philosophical explanation is indeed called for; reservations a b o u t whether such laws are permissible should be taken seriously and evaluated carefully. § 4 O n e such reservation a b o u t interactionism is based on the lack of specificity of the laws governing psychophysical interaction. T h u s , Feigl maintains that interactionism contains the claim that several psychological states are connected to just one neurophysiological state (i. e., a psychophysical relation of the type * F m a n y — ( C i , . . . , Cn-i), Ch => ( C i , . . . , Cn) or, in the simpler case, (Ci) => (Ci), Ci => (Ci,C 2 ). The "=>" signifies the relation of the normative-genetic foundation. Cf. Lorenzen 1969, 85 — 89; Lorenzen 1974, 4 1 - 4 3 . — Sellars' argumentative strategy is clearly based on this sort of reconstruction of historical developments as well. Though Sellars is not, of course, concerned with the justification of statements but rather with the meaning of terms. This meaning of mental terms is determined using a systematic history of the development of mental descriptions. 12

P.M. Churchland 1985, 1 5 - 1 6 .

T h e Empirical Basis of Psychology

187

introspective certainty that one's mind is the seat of beliefs and desires may be as badly misplaced as was the classical man's visual certainty that the star-flecked sphere of the heavens turns daily. 13

Hence, in the f r a m e w o r k of this interpretation, a correction of the conceptual system of folk psychology is possible, at least insofar as it applies to observation statements. Thus, it is equally possible to imagine that introspective reports will one day be framed in a neurophysiological language. 1 4 § 5 This model of the empirical basis of psychology is of f u n d a m e n tal importance. It is based on the most sophisticated conception of observation language in the philosophy of natural science, which, for its part, has grown out of a mutual fertilization of philosophy and psychology. T h e essential aspect of this conception is that the possibility of systematic errors on the level of the observation language is taken into consideration. In the f r a m e w o r k of the first of the three views presented here, the possibility of introspective observation errors is admitted, too. But this only means that the possibility of accidental errors is considered, i. e., of mistakes in w o r d usage, that are due to improperly applying the system of rules of the language in question. Only errors in speaking and writing or confusions regarding w o r d s or phenomena are taken into account. T h e possibilities for error admitted in the third of the views presented here, on the other hand, are of a more f u n d a m e n t a l nature; they are due to shortcomings in the system of linguistic rules as a whole. Since conceptual description always involves a pre-structuring of the corresponding domain, a situation can arise in which all the terms are used correctly and in accordance with the rules, but the description is, nonetheless, completely inadequate. Put more precisely, this means that statements such as "I am seeing blue n o w " or "I am feeling pain n o w " might have the same status as observation statements such as " T h e sulfur is n o w being dephlogisticated to sulfuric acid" or " M a r s is n o w moving in retrograde motion with respect to its deferent." This approach is surely counterintuitive at first glance, but it owes its plausibility to the development of analogous ideas in the philosophy of natural science. In the latter, it is n o w generally accepted that relying 13 14

P.M. Churchland 1981, 70 (emphasis in the original). Cf. Feyerabend 1963, 55; P.M. Churchland 1985, 16.

188

Psychology and Consciousness

on the given means skating on rather thin ice. This becomes evident, since general theories enter into both these examples of observation statements — phlogiston chemistry in the former case and Ptolemaean astronomy in the latter. In this sense, observation statements express data that are already interpreted, whereby the process of interpretation is often fused with the process of perception and cannot be neatly separated from it. The experienced phlogiston chemist perceived dephlogistication in the same way that the particle physicist of today immediately identifies protons and electrons in cloud-chamber photographs. The objection arises that this feature may — at most, if at all — find its psychological parallel in statements like "A is experiencing a cognitive dissonance" or "This stimulus stimulates B's achievement motive," of which one could reasonably say that theoretical terms enter into observation statements. Usual perception statements, however, are of a different kind, or so it is said. But the thesis is that all perception statements are of this kind. According to this interpretation, which derives primarily from the psychologist Bruner and the philosopher of science Hanson, all perception has the structure of a process of judgment and of framing hypotheses, insofar as certain physiological stimuli are attributed to certain classes of things or phenomena, i. e., they are organized according to conceptual categories. An object is perceived as something and, hence, is placed into a categorial scheme. All the relevant processes occur unconsciously, i. e., the conscious perception appears only at the end of this process of categorization. Hence, perceptions are already inferences from experience.15 According to this view, certain physical or psychological stimuli are assigned to certain categories; the stimuli indicate the applicability of the category in question. The essential aspect of this process is that the categorial system in general and the relevance of individual categories in particular are determined using general assumptions. The structure of the accepted theory determines which categories are used at all and which stimuli indicate the applicability of which categories. This model of the theory-ladenness of physical facts underlies this interpretation of the problem of the empirical basis of psychology. Both theoretical and methodological reasons favor such a view; therefore, we consider it to be adequate. The unity of scientific method — 15

For the development of this view, cf. Bruner 1957, 6 8 6 - 6 9 1 ; Hanson 1958, 4 - 3 0 . An excellent presentation of this view is given in Papineau 1979, 1 8 - 3 3 . Fodor also latches on to this theory of perception. Cf. Fodor 1975, 42 — 48.

189

Intentionality and Cognitive Science

i. e., the principle of physicalism of form (cf. VI.4.§ 3) — suggests that the problem of the empirical basis should be treated as uniformly as possible. The interpretation of the empirical basis of psychology sketched here, however, is based on the most sophisticated of the available theories of the empirical basis of physics and, hence, is preferable. Introspective reports, thus, have the same structure as physical observation statements. The difference is simply that the relevant indicators are internal and not external phenomena. Hence, this methodological perspective leads to the same result that was already formulated in III.1.§ 3: The peculiarities of the mental are not given by direct acquaintance, but rather are inferred from theoretical description.

2. Intentionality

and Cognitive

Science

§ 1 The treatment of the problem of the empirical basis of psychology leads to the more general question of the relation of consciously experienced mental states and events or those identified by folk psychology, on the one hand, to the cognitive states and events specified by scientific psychology, on the other. Intentionality is widely thought to be a characteristic, or even the defining feature, of folk psychology's understanding of the mental, since it is supposed to document the special status and the peculiarity of psychological as opposed to purely physical phenomena (cf. II.5.§ 1). Folk psychology (as discussed in II.6.§ 2) is said to include all theories that (1) use intentional terminology, and (2) whose laws refer to the propositional attitudes that occur in these intentional statements, and especially to the contents of these attitudes. Hence, if one asks whether folk psychology and scientific psychology refer to the same entities, then the issue is really to determine what justifies our attribution of beliefs, desires and goals, as well as whether explanations of behavior can be founded on laws of the following type: If a person A wants to achieve the goal a and believes that action b is suited to that goal, then A puts b into action (cf. II.6.§ 2). An identification of the states of folk psychology with those of cognitive psychology is by no means obvious; after all, we do indeed attribute to other people certain beliefs and desires in folk psychology but not failure motives and cognitive dissonances. Only closer analysis

190

Psychology and Consciousness

can reveal whether there is an identity of reference here and, if so, in what sense. Hence, we have to complete our reconstruction of the status of psychological terms. Fig. 3 (VI.1.§ 4) lacks any reference to folk psychological concepts, which would clearly have to find their place between the terms of cognitive psychology and those of introspective description. So we need to take a closer look at the middle column of Fig. 3. Following our discussion of the status of introspective reports in II.1.§ 2, III.1.§5 —6 and VII.1, we start with the assumption that the mental states and events attributed to others and those known by introspective reports are identical. In other words, folk psychology treats those states that we usually call "conscious." The mind-mind, problem is then characterized by the additional question whether these conscious states are identical with the psychological states assumed by cognitive psychology or whether the former are merely indications of the latter. The question is: which of the two following possibilities is realized (whereby "—• " once again indicates connection by correspondence rules and "•" indicates identity): State of Cognitive Psychology I

or

I

State of Folk Psychology I Qualities of Perception Fig. 4 If a concept-indicator relationship is realized here, then it can be assumed that the mental states described by folk psychology, i. e., consciously perceived mental states, do not reveal the whole architecture of our internal psychodynamics. In the following, we want to discuss two approaches to interpreting cognitive psychology and to illuminate the status of intentional speech within their respective frameworks. In doing so, we shall start with a discussion of cognitive science, which can be regarded as a further development of functionalism. In discussing it, we will concentrate primarily on the form that Fodor and Pylyshyn gave to it. It is representative of the spectrum of approaches in cognitive science, so that such a

Intentionality a n d Cognitive Science

191

restriction is suited to the matter at hand. In the next section, we shall examine the syntactic theory of the mind, which can be reconstructed as an implementation of the program of eliminative materialism. In the framework of this conception, as it is advocated by Stich, intentional speech is said to be methodologically questionable and inadequate to the facts. This is followed by investigations into psychosemantics and additional reflections on the mind-mind problem. § 2 Cognitive science starts from the assumption that folk psychology's scheme of laws is at least approximately true. It is essentially correct to explain human behavior by recourse to beliefs, expectations, goals or motivations. Mental states of this type are characterized by certain contents, i. e., they are described by intentional concepts. These contents are, moreover, essential for explaining behavior, i. e., mental states of different content lead, as a rule, to different behavior. If someone believes that it is raining, then she will behave differently when leaving her home than she would if she were convinced that the sun was shining. If somebody reaches for her umbrella upon leaving her home, then folk psychology would explain this by attributing to her the belief that it is raining or that it will rain. The explanation is given by invoking mental states of a particular content, i. e., it makes use of intentional terms. It is one of the central claims of cognitive science that this type of explanation represents the most successful of the available models for explaining behavior. This means that intentional terms are vastly superior to physical terms in the domain of psychology, and the reason for this superiority is that laws can be expressed in intentional terms that cannot adequately be cast in physical terms. Assuming that a person A witnesses a serious accident and hurries to the nearest telephone booth to call the ambulance. This behavior falls under the following law of folk psychology: When someone sees that someone else is injured, the desire to help the injured person ensues. Hence, if we know that A is witnessing an accident, then we can predict his behavior on the basis of this intentionally formulated law: A will try to get help. Such a prediction would not be possible, if we could only base it on the physical circumstances of the situation. First, because there can be a physically identical situation (such as the acting out of such an accident while filming a television program), that would not induce an observer to get help. Second, a physically different situation

192

Psychology and Consciousness

can be imagined (A does not witness the accident himself, he is simply told about it) that would result in the same behavior. 1 6 Hence, the intentionally formulated psychological generalization cannot be understood simply by reference to the external circumstances of an action: Physical and neurophysiological terms taxonomize the w o r l d in w a y s that do not permit us to express such generalizations. T h e y often distinguish aspects of the world and of behavior that are equivalent with respect to their psychological import, and sometimes fail to make distinctions that are psychologically relevant. For this reason, descriptions cast in such terms typically fail to capture important psychological generalizations concerning human behavior. 1 7

If behavior cannot be adequately understood as a simple reaction to a physically described situation, one could still try to paraphrase intentional sentences in a more refined fashion by using physical statements in order finally to avoid the former altogether. Cognitive science does not believe that this is a reasonable option. In this context, Fodor discusses the "fusion theory" of intentionality that seeks to avoid attributing contents. According to this theory, the intentional statement "A believes that it is raining" should be reconstructed as "A believes-thatit-is-raining." The sentence as it is reconstructed contains a predicate that cannot be further divided; thus, it corresponds to a statement about A's physical activity (like "A is drinking"). The fusion theory fuses a psychological verb with its object (the content sentence) to an amorphous collective expression (whereby the content sentence disappears). Fodor objects that such a reconstruction turns peculiarities that appear to be plausible and consistent in the intentional approach into mere coincidences. For example, it would be a mere coincidence that a phrase is common to two different propositional attitudes with the same content sentence (such as "A hopes that it is raining" and "A fears that it is raining"). Both attitudes are, in fact, completely independent of one another. Similarly, it would be purely coincidence that statement sentences occur as grammatical objects after psychological verbs. The sentences "A believes-that-it-is-raining" and "it is raining" have no connection to one another. But, so Fodor concludes from these (and other) peculiarities: "It's the mark of a bad theory

16 17

For this example, cf. Pylyshyn 1984, 7 - 8 . Pylyshyn 1984, 17.

Intentionality and Cognitive Science

193

that it makes the data look fortuitous." 1 8 Relinquishing intentional terminology is indeed possible, but it is not methodologically expedient. For cognitive science, this observation implies that propositional attitudes like beliefs and goals should be taken seriously, and that it is methodologically sound to attribute such propositional attitudes to others. It is justified to assume that the contents of propositional attitudes are somehow represented in a person's cognitive system. Hence, cognitive systems contain mental representations, which represent possible circumstances or situations and organize preferences among these circumstances. If A believes that a, then a is represented in the cognitive structure of A; and if B prefers situation b to circumstance c, then this is expressed by an internally represented order of preferences. According to Fodor, this representation is cast in a private "language of thought," that is built into, as it were, the cognitive system in question. This means that a mental representation has the structure of a sentence; it is a sentence in the mental language. This idea of an internal language owes its plausibility to an analogy to computers: Computers also use at least two languages, one for communicating with their environment and another for carrying out their operations. In this model, propositional attitudes are directed toward mental representations — toward ideas that are not themselves formulated in the public language. Rather, they are formulated in a language that is internal to the system. Propositional attitudes are relations between people and sentences. 19 § 3 According to what we have said, the central problem that cognitive science seeks to solve lies in the fact that, on the one hand, there is no doubt that the content of mental states effects behavior and, on the other hand, it is certainly not clear at the outset how they could effect behavior at all. After all, it is anything but obvious how an idea's content can have physically tangible effects. To resolve this difficulty, cognitive science resorts to a computer analogy, which is a further development of the condition of Turing translatability (cf. II.4.§3). The model of internal language mentioned above is elaborated by the computer analogy. The reason for resorting to computers here is that they do, in fact, perform operations with semantically interpreted entities 18 19

Fodor 1981, 1 8 0 - 1 8 1 . For the argument as a whole, cf. Fodor 1981, 1 7 9 - 1 8 1 . Cf. Fodor 1975, 2 8 - 3 3 , 5 5 - 5 6 , 6 5 - 6 6 ; Pylyshyn 1984, 1 - 2 1 .

194

Psychology and Consciousness

like numbers; and the way computers perform these can be taken as a model for mental operations. The computer analogy is the only successfully elaborated hypothesis about how differences in meaning could result in physically tangible differences. 20 Cognitive science distinguishes among three levels in computers: the physical realization, the syntactic structure and the semantic interpretation. On the level of the physical realization, computer states are described by a pattern of electronic states in integrated circuits. On the syntactic level, these states are represented as strings of formal symbols, and a sequence of such states is characterized using transformation rules for the strings of symbols. For example, these transformation rules take the form: If memory register 1 is in a state that is described by the string xox, and memory register 2 assumes the state characterized by xxo, then memory register 3 changes into the state that corresponds to xoxx. This is merely a syntactic description, since no meaning is ascribed to the individual symbols. They are discrete, atomic symbols that are introduced as parts of a formal language. Hence, at this level of description, strings of symbols are merely characterized by syntactic correctness (or a lack of syntactic correctness), i. e., they are (or are not) produced by applying the transformation rules of the formal language. Correspondingly, the same holds for transformations of strings of symbols. A syntactically described state corresponds to a variety of physical realizations, i. e., the same formal string of symbols can be realized technically in a variety of ways. Finally, on the level of semantic interpretation, meaning is associated with such a string of syntactic symbols. Thus, the formal transformation of the strings given above can be understood as the binary encoding of the addition 5 + 6 = 11. By interpreting the strings as numbers and the change in state of the third memory register as addition, these states or processes which were initially only formally characterized, are given a semantic interpretation. 21 This explains how computers can perform semantically interpreted operations with semantically interpreted entities. In these machines, contents are syntactically encoded in such a way that the code reflects the relevant semantic differences. A syntactic state of this kind is, moreover, to be understood as an equivalence class of physical states, 20 21

C f . Pylyshyn 1984, 7 , 2 6 - 2 7 . C f . Pylyshyn 1984, 5 9 - 6 2 .

Intentionality and Cognitive Science

195

so that sequences of syntactic states can also be described by physical causal laws. 22 Hence, causality is not located on the semantic level. Rather, semantically interpreted entities are transferred into a syntactic code that, for its part, is also realized physically. This code's transformation rules are formed such that, on the one hand, they take the semantic features of the original expressions into account and, on the other hand, they can be represented by physical relations between machine states: The physics of the machine thus guarantees that the sequences of states and operations it runs through in the course of its computations respect the semantic constraints on formulae in its internal language. 23

Thus, in automated theorem proving, for example, the machine states M i , . . . , M n are connected to semantically interpreted statements S i , . . . , S n in such a way that M i , . . . , M n are run through precisely if Si, . . . , S n - i is a proof for S n . The proof is not derived from the semantic properties of the statements Si, but rather from the physical properties of the states Mi, with which syntactic characteristics are associated in turn. It is this association between types of features (in this case, based on the machine's structure and the program's conception), that enables one to interpret the sequence of physical states as a proof. 2 4 The essential move of cognitive science lies in applying this model to the brain's functioning. As with the computer, a neurophysiological chain of states or events can be described from a physical or syntactic standpoint, as well as from a psychological or semantic standpoint. In the physical description or from the physical perspective, these states or events stand in a causal relationship with one another, while in the psychological description or from the psychological perspective, these same events are connected to one another by logical relations: In short, the organic events which we accept as implicated in the etiology of behavior will turn out to have two theoretically relevant descriptions if things turn out right: a physical description by virtue of which they fall under causal laws and a psychological description by virtue of which they constitute steps in the computation from the stimulus to the response. 25

22 23 24 25

Cf. Pylyshyn 1984, 39. Fodor 1975, 66. Cf. Fodor 1975, 6 6 - 6 7 , 73. Fodor 1975, 74.

196

Psychology and Consciousness

Hence, cognitive science assumes a kind of co-variation between causal-syntactic and logical-semantic relations in the same chain of states or events. For psychology, the situation can be represented as follows. Mental states are themselves semantically interpreted entities; they express a certain content and, in that sense, they have semantic attributes such as truth values or referents. In addition to semantic features, mental states also have syntactic or formal properties, which include all their properties that are independent of their meaning. Cognitive science's central idea is expressed in the so-called formality condition: All mental operations proceed exclusively by virtue of the formal properties of mental states. If differences in the semantic content of mental states are supposed to be psychologically relevant, then these differences must be reflected by the syntactic differences between these states. 26 § 4 It is obvious that cognitive science in general and the formality condition in particular are inspired by a commitment to materialism. The problem of how semantic contents can have physically tangible effects is primarily a problem for those who do not believe that propositional attitudes in particular and mental states in general are serious candidates for an ontological interpretation. Cognitive science's solution to this problem lies in taking a token mental state to be real and causally effective (precisely because this state has a physical realization), but in attributing mere instrumental status to psychological types of states. Hence, cognitive science advocates a realism for single propositional attitudes, but an instrumentalism for types of propositional attitudes. This interpretation of psychological types of states arises from a commitment to functional materialism that, for its part, represents a variant of the token-identity theory (cf. II.4.§4). But this monistic orientation is by no means imperative. On the contrary, like functionalism in general, cognitive science in particular is ontologically neutral and, hence, compatible with dualism (cf. II.4.§ 4). It is in no way inevitable to view neurophysiological processes as the exclusive medium for causally relevant syntactic operations. Rather, precisely because computer programs are viewed as a model of the brain's mode of operation, neurophysiological processes cannot be the sole instantiation of cognitive activities. This 16

Cf. Fodor 1975, 32; Fodor 1981, 2 2 6 - 2 2 7 , 2 4 0 - 2 4 1 . For a more detailed presentation of this key concept, cf. Sayre 1987, 247 — 251.

Intentionality and Cognitive Science

197

claim to the multiple realizability of cognitive activities lies at the heart of cognitive science. Accordingly, dualism may well assent to cognitive science if the latter's accent is shifted. The point is not to make intentional modes of description acceptable even to those who believe that the world can be exhaustively analyzed by physics, but rather to understand the mechanism of intentionally characterized processes. In other words, cognitive science should be seen as a program for explaining the laws of cognitive psychology. As can be gathered from the process model of motivational psychology, for example (cf. V.2.§ 4), the mental processes referred to in that context (such as judging situations or comparing actions with goals) are characterized by certain kinds of content and are of an intentional nature in this sense (the reason being that situations that differ in content in the relevant aspect are represented differently and, as the case may be, judged differently). But the effects of intentional mental states on behavior are not yet understood, if one adheres to the view (as we do) that these states demand serious consideration, not only from a methodological but also from an ontological perspective. A dualistic interpretation is not free of the need to provide an explanation for how mental states of different content can lead to different actions. Cognitive science's approach appears to be the only one that is currently in a position to offer a (more or less) detailed theory about how one can operate with semantically interpreted entities, without in turn referring to intentionally characterized processes. In this sense, it is a plausible approach to explaining how we manage to act according to reasons and follow goals. § 5 In the framework of cognitive science's interpretation of mental activity, the status of intentionally characterized states and aspects deserves particular interest. First, it is noteworthy that cognitive science understands psychological processes as formal operations on semantically interpreted entities. This means that every propositional attitude is supposed to correspond to a formal relation between an organism and a formula in its internal system of representation. It is true that this formula is itself semantically interpreted, but the relation does not hook into its semantic aspects but rather into its syntactic ones. This peculiarity is precisely what characterizes the formality of this relation. The crucial question is: what is the status of the semantic interpretation of

198

Psychology and Consciousness

the mental state, i. e., of the content of the propositional attitude, in the framework of this approach? Fodor is sometimes accused of (and attacked for) attempting a syntactic definition of these semantic aspects. 27 But this is an inaccurate reconstruction. Fodor has in mind a syntactic theory of mental operations, not, however, a syntactic theory of the objects of these operations (i. e., not a theory of the content of propositional attitudes). Thus, he mentions explicitly that describing a sequence of computer states by means of its internal machine language (i. e., its syntax) does not tell anything about the content of the program being run. That is, a single, formally described sequence of commands can be interpreted semantically in a variety of ways. A single sequence of states expressed in the machine language can, for example, represent both a simulation of the Six Days' War as well as a game of chess. 28 Moreover, each of these situations can be simulated by different programs, i. e., by different sequences of formally characterized machine states. This implies that there exists a many-many relation between semantic interpretations and syntactic structures. Fodor's observation leads to the result, that the connections between formally characterized mental states do not yet provide a semantic interpretation of these states. This result is a corollary of an insight dating from around the turn of the century. This insight was gained in the context of the discussion of Hilbert's conception of definition by axioms. In this conception, mathematical terms are supposed to be definable by the condition that they satisfy the basic assumptions, i. e., the axioms, of the corresponding mathematical theory. And these axioms are themselves given in a formal, uninterpreted language. The problem with this approach is that such a system of uninterpreted symbols can be interpreted semantically in a variety of ways. The formal system represents an abstract structure that, in general, possesses a number of 'models that satisfy it. Thus, for example, the formal system of Peano axioms has the natural numbers as one model alongside others. 2 9 The same is true of Hilbert's formalized version of the axioms of Euclidean geometry, which is satisfied not only by the entities usually associated with it (such as 27

Sayre 1987, 2 5 4 - 2 5 5 ; M e n d o n ? a 1987, 1 9 6 - 1 9 7 . Cf. Fodor 1981, 207. 2 * Cf. Wolters 1984c, 9 1 1 - 9 1 2 .

28

Intentionality and Cognitive Science

199

points, straight lines and planes), but also by a model developed by Wellstein, which operates with spherical surfaces (arranged in a special way) instead of plans and great circles instead of straight lines. In plane geometry, moreover, one can speak of pairs of numbers instead of points, for example, and of linear functions instead of straight lines. All these substantively different entities satisfy the formalized version of the Euclidean axioms. 3 0 The possibility of ascribing content to the same formal structure in different ways makes it impossible to determine, using only this structure, the semantic properties of the terms occurring therein. Formal structures do indeed impose constraints on a semantic interpretation, since not all situations can serve as adequate models, but they do not single out any particular interpretation. They determine only an equivalence class of interpretations that exhibit the same structure; that is, they determine meaning only up to isomorphism. For cognitive science, this means that it is impossible to infer the content of mental states from an analysis of the syntactic relations between states. A psychology operating on the basis of the formality condition, i. e., one that explicitly disregards the semantic features of mental states, cannot contribute anything to an explanation of the semantic features of mental states. The claim is precisely that the internal psychodynamics is not influenced by semantic properties. The course of the (causally relevant) syntactic operations is assumed not to depend on whether (and, if so, in what way) the strings of symbols being processed are interpreted with respect to content. Mental states do, indeed, possess semantic characteristics, but mental operations do not have access to them. In this sense, cognitive science explains the nature of cognitive processes by tracing it back to syntactic or algorithmic procedures; this is how it explains the nature of intelligence. But in no way does it present a theory of the representational character of mental states; it does not deal with the contents of mental states. It explains mental operations, not mental representations. 31 Psychology interpreted in terms of cognitive science does, indeed, make use of intentional speech, but it does not explain intentionality. Cognitive science does not provide a theory of psychosemantics. 32 30 31 32

On multiple interpretability, cf. Schlick 1925, 3 1 - 3 8 . Cf. Fodor 1981, 20 - 24, 232 - 233. For a more detailed elaboration of the relation between cognitive science and the theory of psychosemantics cf. Carrier 1991b. For more on the problem of psychosemantics proper cf. VII.4.

200

Psychology and Consciousness

It can be gathered from this presentation that Searle's famous (and often misunderstood) "Chinese room argument" draws upon just this peculiarity. Searle's argument is directed against the thesis that computer programs can provide a model of mental activity and exhibits a strong anti-functionalist bent. Searle argues that it is impossible to trace human, content-based thought-processes back to formal symbol manipulations as they occur in computers. M o r e specifically, he addresses the functioning of question answering programs. Such programs have the capacity to give sensible answers to questions about some story presented to them. T h u s their internal mechanism appears to provide a plausible means for illuminating what might happen in human story understanding. This assumption, however, Searle argues, is far off the mark. To make his point clear, Searle invents a thought experiment that is supposed to elucidate the functioning of such a program. He tries to imitate its internal procedure by putting himself into the role of a program designed to answer questions posed in Chinese about a story told in Chinese. He imagines that he is locked in a closet (the Chinese room) where he is given a large batch of Chinese symbols. He is unable to understand a single one of these signs; they are all meaningless squiggles to him. However, he has some instructions formulated in English at his disposal, and these instructions correlate these symbols to other Chinese symbols in a second batch. H e is now in a position to assign to each of the symbols of the first batch ("the questions") a symbol of the second batch ("the answers") just by recognizing the shape of these symbols. This means, without any understanding of the meaning of the writing, he can simulate understanding by applying rules that are insensitive to content and only refer to the shape, i. e., the formal aspects of the symbols. T h e upshot of this thought experiment is that the capacity o f manipulating formal symbols according to formal rules has nothing to do with human understanding. T h e Chinese room inmate simply does not understand Chinese — though he can fool an outside observer about his deficiency. 3 3 Searle's argument is correct and quite straightforward. T h a t is, it does not rely on a "mysticism of understanding" or on dubious, everyday intuitions about understanding as some critics have objected. W h a t this argument employs is simply the multiple interpretability of formal structures. If a formal structure can determine meaning only up to 33

Cf. Searle 1980, 4 1 7 - 4 1 8 .

Intentionality and Cognitive Science

201

isomorphism (i. e., if it can determine only an equivalence class of isomorphic semantic interpretations), then meaning cannot be inferred from the formal structure. And if, moreover, human mental processes are content-based and thus have recourse to meaning, then they cannot be completely or exhaustively characterized as manipulations of strings of symbols. The point is, however, that this result perfectly agrees with the tenets of cognitive science. Searle's argument affects only a position that attempts to derive contents from formal structures, and that is precisely what cognitive science does not aim at. As just stated, cognitive science in no way regards a theory of psychosemantics as a mere consequence or corollary of a theory of mental processes. It is the latter, and not the former, that the computational model of the mind is supposed to provide. As regards cognitive science, Searle's argument — though valid — is thus rather pointless. Cognitive science does, on the other hand, indeed assume an association between formal-syntactic and content-semantic features. Mental states that differ in content are supposed to be distinguished by their formal characteristics, so that purely syntactic operations can proceed in a way that respects the semantic properties of the corresponding states. It is this peculiarity on which Searle's thought experiment is based. Hence, cognitive science claims that it is possible to recognize or identify semantic content using syntactic features; it does not claim to define semantic content using formal characteristics. It postulates a onesided correlation between semantics and syntax: Semantic differences must be reflected in syntactic differences, not necessarily the other way around. 34 Thus, semantics is supervenient with respect to syntax. 35 This subtle aspect of cognitive science has not yet been sufficiently appreciated; nonetheless, it has significant effects on the status of intentional speech in the framework of this theory. There is, however, one important qualification — hitherto unnoticed — to be added to the formality condition. As just mentioned, the formality condition requires that every semantic difference be reflected in a syntactic difference among the corresponding symbols. The multiple interpretability of formal structures makes it clear, on the other hand, that this condition — in the general form just stated — can never be 34 35

Cf. Pylyshyn 1984, 31, 74; Marras 1987, 1 2 1 - 1 2 2 . For the concept of supervenience, cf. II.3.§ 4.

202

Psychology and Consciousness

met. After all, the different semantic interpretations of the same formal structure are syntactically completely alike. It is this very feature that qualifies them as different semantic interpretations of the same formal structure. This implies that the differences in content between different interpretations of the same structure cannot be captured by syntactic differences; only semantic differences within an interpretation can, if at all, be grasped formally. This is to be expressed such that the formality condition holds at most up to isomorphism or, in other words, it is confined to a given interpretation. It should be emphasized at this juncture that the existence of syntactic features sufficient to identify semantic content has until now only been postulated, but not demonstrated. The formality condition has the status of an empirical hypothesis. T h e current state of research in this area can be summarized by saying that (1) we do not presently have an adequate syntactic theory at hand; i. e., we are not, in fact, able to identify mental states using formal criteria. On the other hand, (2) there are fruitful starts toward a solution to this problem. Consequently, it should not be ruled out that cognitive science's program might be accomplished in the foreseeable future. § 6 This assessment remains unimpaired if one takes more recent developments in computer science into account. It has been objected to cognitive science that its use of the computer analogy applies only to today's digital computers. If there is at all a computer model that can claim to be a plausible means for elucidating cognitive processes, so the objections runs, it is "connectionist processing" or the model of "neuronlike circuits." However, relying on that approach implies rejecting the idea — characteristic of cognitive science — that the mind performs rule-governed operations on formal symbols. 3 6 In our view there is some truth to this objection but its impact is far less significant than is usually assumed. 3 7 Neuronlike circuits consist of an assembly of a large number of simple electronic elements ("nodes") each of which is in a definite state of activation (characterized by a certain voltage). Each of these nodes is connected to a large number of other nodes and can transmit electric impulses to the other nodes which change the latters' state of activation. 36 37

Vgl. Bechtel 1988, 64. T h e following argumentation roughly follows Carrier 1991a, section 4.

Intentionality and Cognitive Science

203

The impulses can be considered as excitatory or inhibitory stimuli. The state of the whole circuit or network at a certain time is characterized by a point in a configuration space. A configuration space is an abstract space whose dimension is equal to the number of elements represented; the activation state of each element is represented by a point on a coordinate axis in that space. The coordinate values of all axes taken together (i. e., the set of voltage values of all elements) characterize the overall activation state. To each of these overall states an energy value can be ascribed that expresses the total energy of the network in that particular state. In this approach the possible states of the network can be described by an energy surface in the configuration space, and the temporal evolution of the network can be represented by a trajectory on the pertinent energy surface. In general, the energy surface has a "mountainous" shape, i. e., it possesses lots of local maximum values (peaks) and local minimum values (valleys). Computing or problem-solving amounts to starting off with certain initial data, processing these data, and ending up with certain output values. Data processing can, accordingly, be represented by a trajectory on the energy surface that connects the initial state (representing the parameters to be processed) with the final state (representing the solution). In order to make sure that the system arrives at the right solution it must be so constructed that this solution corresponds to a local energy minimum. That is, the initial state must possess a comparatively high energy and thus be unstable, and it must be transferred by means of the internal excitations and inhibitions into an energetically low und thus stable state. Solving a problem by means of a neural network requires the invention of a nodal structure that has the rightly shaped energy surface. The shape of the energy surface thus represents the constraints of the problem at hand, the state of activation of some selected nodes represents the input values that the parameters in some concrete instance of the problem assume, and the state of activation corresponding to the energy minimum represents the solution. 38 In our context the relevant aspect of neuronlike circuits is that they appear to solve some problems such as pattern recognition in a more human-like fashion than usual; sequentially operating digital computers. Hence, they provide a much more promising model for the brain's 38

For this sketch of neural networks cf. Tank/Hopfield 1987.

204

Psychology and Consciousness

functioning than the latter. But in that case one must ask if the moral that cognitive science draws from the computer analogy can be transferred to neuronlike circuits. And the problem is, that on the face of it the answer patently seems to be no. After all, unlike a digital system, a neural network does not follow a set of rules for manipulating formal symbols. There are neither strings of formal symbols explicitly encoded in the circuit, nor are there any rules programmed into it. Computing or problem-solving does not have the structure of a sequence of discrete steps following one another according to certain rules of operation that are explicitly represented in the circuit. Rather, it has the structure of a continuous process that is governed directly by the laws of physics. The objection to cognitive science can now be stated thus: Mental activity does not consist of performing operations on formal symbols. In neural networks the syntactic level has simply disappeared from the scheme. There is no intermediate link between the semantic interpretation and the physical realization. It has to be admitted, first, that there is a point to this argument. It is clear that the content of the portrait that cognitive science draws of human mental activity has to be changed. It surely makes a difference whether or not mental operations are discrete or continuous and whether or not they follow explicitly encoded rules. And if connectionism is right, as it seems to be in light of the available evidence, cognitive science has to be revised substantially. Our claim is, however, that whereas the content of the portrait cognitive science gives must indeed be modified, its structure remains intact. The colors have to be changed, as it were, but the lines remain in their place. More specifically, the triade of physics, syntax, and semantics — central to the view of cognitive science — can be identified on closer inspection in neural networks, too. In order to back that claim we must trace the apparently missing syntax. The crucial aspect in that context is that a neuronlike circuit does not model precisely one problem; it rather models a whole class of substantively different but formally equivalent problems. One and the same network may represent a specific brain activity, a combinatorialoptimization problem (such as the traveling salesman problem of finding the shortest path through a large number of cities), and the electronic spin state of so-called spin glasses. 39 In all these cases the problem as well as the manner of its solution can be described in a formal fashion 39

Cf. Stein 1989, 4 0 - 4 2 .

Intentionality and Cognitive Science

205

and without resorting to the details of a particular application. In neuronlike networks, too, it is an abstract or formal quantity, namely the shape of the energy surface, that brings about the desired effect. Computing and problem-solving still depend essentially upon the formal aspects of the system and not on the particulars of their interpretation. T h e shape of the energy surface " m e a n s " different things in different applications. As we just emphasized, the interpretation of all formal systems is only determined up to isomorphism, and this peculiarity leaves room for a multiple interpretation of one and the same formal structure. T h e traits characteristic of the relation between semantic interpretation, syntactic structure, and physical realization — as specified in cognitive science — can easily be inferred from this description. As just stated, the same (formally described) network has the capacity to solve problems with a different content (though the same structure). Moreover, there is no need for detailed argument to realize that the same (substantively described) problem can be solved by networks structured differently. All that matters for problem-solving is the right connection between input and output; the precise trajectory the system moves along from the former to the latter is irrelevant. From this it can be concluded, first, that there is a many-many relation between semantic interpretations and syntactic structures. And it is clear, second, that in order for the formally described network to represent adequately a certain problem and its solution, all relevant differences in content must be reflected in formal differences. T h a t is, if two sets of initial parameters are to lead to distinct output values (which is sufficient for a difference in content between them), then they must be associated with differently shaped regions of the energy surface (which indicates a difference in their formal aspects). But this is nothing other than the formality condition translated into network language. T h e relation between syntax and semantics is thus faithfully preserved in connectionism. T h e same holds for the relation between syntax and physics. It is obvious that the same formal structure — i. e., the same shape of the energy surface — can be realized by means of several distinct physical systems; and any such system can by means of an appropriate choice of the constraints represent distinct formal structures — i. e., they can assume distinct shapes of the energy surface. So here a many-many relation between syntax and physics is present as well. Moreover, it goes without saying that — just as in digital computers — causality is located on the physical level, not on the semantic one. It is not the content of

206

Psychology and Consciousness

a question that puts the electrons in motion. Here again there is no deviation from the claims of cognitive science. The upshot of this discussion is that connectionism furnishes the same sort of relations between semantics, syntax, and physics as do the more traditionally oriented versions of cognitive science. Both entail a many-many relation on either level. It is obvious, on the other hand, that there is in fact one substantial difference between the two approaches. Cognitive science assumes that a cognitive system acts according to explicitly encoded rules, and this is precisely what connectionism denies. The point is, however, that the behavior of connectionist systems can still be described by a set of formal rules. It is strongly suggested by the present state of knowledge that the performance of any such system can be modeled by a rule-based system. 40 Connectionist systems behave as if the rules were encoded in them. This means that it is not alone the sort of relations between semantics, syntax, and physics that coincides in the two approaches; moreover, the relations themselves agree. The particular input-output correlations of connectionist systems can be reproduced by rule-based systems. Accordingly, connectionism leaves the structure of the cognitive science triade unaffected. These similarities allow for the conclusion that the mental architectonic as specified by the two positions virtually coincides; their differences only concern the nature of the building blocks by means of which this architectonic is realized. § 7 In the following, we assume that the claims of cognitive science are justified and inquire into the consequences for the philosophy of consciousness. It is useful to start by recalling (1) that cognitive science emphasizes the significance of intentionally described mental states, i. e., of ideas with a specific content. In addition, it denies (2) that intentional categories can be defined by syntactic features. Semantic properties (such as truth value or reference) cannot be accounted for in this way. Nevertheless, (3) the application or operation of intentional statements can be sensibly controlled and regulated using the syntactic properties of sentences of the mental language that these statements characterize. For this reason, mechanisms that are blind to semantic interpretations of formulas, can still operate with these formulas as if these interpretations were available to them. 40

Cf. Bechtel 1987, 2 4 - 2 5 , 30.

Intentionality a n d Cognitive Science

207

One aspect of cognitive science's interpretation coincides with that of folk psychology: Folk psychological states such as beliefs and goals in particular and introspectively available mental states (i. e., those available to consciousness) in general are, in fact, causally effective; they cause human behavior and interact with other psychological states. This causal efficacy does not, however, result from those features that are attributed to them by folk psychology or introspection. This means, the belief that a is the case does, in fact, effect behavior, but not as a result of the content a, rather, because the brain translates it into a semantically uninterpreted string of symbols, which is in turn physiologically implemented. The subtle point in this interpretation is that states of consciousness are, in fact, causally relevant, but not those properties of these states (namely, the content) that are accessible to consciousness. Since, on the other hand, the causally irrelevant, semantic features cannot be defined formally, the intentional terminology cannot be abandoned in favor of an exclusively syntactic way of speaking. Semantic or intentional characteristics are to be understood as emergent properties in the sense described in IV.§ 4: they occur together with formal features, but cannot be understood by means of the latter. Thus, in the framework of cognitive science, there is no primacy of syntax. Rather, one can speak of (following cognitive science's selfassessment) a parallel between syntactic processes and semantic operations. 41 To be surp, this parallelism should not be understood as a oneto-one correspondence of syntactic and semantic properties, but rather in the weaker sense of an association of syntactic differences with differences in content. If, moreover, one sees the syntactic specification of mental states as a characteristic of scientific psychology and sees folk psychology's types of states as determined by the content of mental states, then cognitive science's interpretation leads to a parallelistic view of the mind-mind problem, as it were. Syntactic differences correspond to semantic differences; and, in this sense, there is a psychopsychological parallelism of properties. As a result, intentional terminology (or states of consciousness as defined by folk psychology) remains an essential means for describing mental states (or an important characteristic of psychological laws), although it loses its central position. Though Brentano and Chisholm are, indeed, correct in their thesis that an adequate physical definition 41

Cf. Pylyshyn 1984, 40.

208

Psychology and Consciousness

of intentional concepts is not possible, this fact is much less significant than Brentano and Chisholm assume. Mental phenomena are only incompletely characterized by intentionality. Recourse to intentionality does not provide access to causal relations between psychological states. Intentionality does determine psychokinematics, as it were, but not psychodynamics. In the end, the viability of cognitive science's approach must be measured by whether it succeeds in providing a reconstruction of the intentional processes that cognitive psychology assumes. There haven't yet been any convincing successes in this field. Thus, cognitive science is a program for explaining mental operations which still lacks substantial empirical support. Nevertheless, this program has the advantage of basing its explanation of mental operations on processes that can in principle be realized technically or constructively. Naturally, the lack of empirical support for cognitive science's interpretation also affects the reliability of the consequences of such an interpretation for a philosophy of consciousness as elaborated here. In that sense, we have merely discussed an option for a philosophy of consciousness, whose adequacy to the facts has yet to be shown. Apart from that, it is, in any case, problematic to restrict scientific psychology to the laws of folk psychology, i. e., to expect that the progress of psychology would merely refine that which we knew all along (cf. VII.3 — VII.4). Moreover, we doubt that Fodor's emphasis on sentence-like structures as the bearers of psychological operations is appropriate. Several authors have warned against underestimating prelingual mental skills; and it seems to us as well that associative, imagistic thinking has been unjustifiably neglected in this program. The idea of modeling all thought processes on discursive argumentation is surely one that could only occur to philosophers.

3. Intentionality

and Syntactic

Psychology

§ 1 The syntactic theory of the mind, as advocated above all by Stich, claims that cognitive science is inadequate. Stich denies the parallelism of properties that is claimed by cognitive science and deems it impossible to reproduce syntactically differences in content. Hence, the formality principle is considered incorrect. The surprising aspect of all

Intentionality and Syntactic Psychology

209

this is that syntactic theory blames this defect on semantics and not on syntax. This means that the semantic interpretation of mental states is not intended to reappear in its familiar role as the principle of mental causation; on the contrary, the semantic interpretation of mental states is said to be unfit for this role and conceptually outdated. Rather, mental representations are completely characterized by their syntactic features. What is at issue here is the role of intentional terms, and Stich's arguments focus primarily on problems with the justification for the attribution of beliefs. Stich is primarily concerned with the status of sentences like "A believes that a." He contends (1) that such intentional sentences should not, for methodological reasons, play a role in any psychology deemed scientific; and (2) that, as regards cognitive psychology, they do not, in fact, play such a role. According to Stich, the folk psychological strategy of expressing a mental state or event by a content sentence (i. e., a) is fundamentally inadequate; cognitive psychology and intentional terms do not match up. Stich concludes from this that the mental states or events assumed by folk psychology do not, in fact, exist. In reality, there are no such things as beliefs. Stich's central argument against the formality condition lies in his claim that folk psychology's criteria for individuating belief states, i. e., its criteria for establishing which beliefs have the same content or belong to the same class of beliefs, deviate significantly from those for respectable theoretical states. Regular theoretical terms get their meaning by being integrated into the pertinent theoretical network (cf. VI.1.§ 2). If beliefs are supposed to be regular psychological states, then they must be individuated by their relation to other psychological states, to sensory stimuli and behavioral responses. The individuality of a belief state should be determined by its inferential profile. T h a t is to say, two mental states are of the same type or are instances of the same kind, if their respective relations to other mental states (including stimuli and behavior) exhibit the same structure. In the case of the belief states determined by folk psychology, however, individuation is achieved differently. Stich defends this contention with several arguments, the central one of which will be examined more closely here. § 2 If beliefs are to be adequately characterized as theoretically determined states, then it must be possible to identify them by the role they play in the mental structure. So the first step must be to reconstruct the criteria for identifying belief content given this assumption. It should

210

Psychology and Consciousness

be kept in mind that not only the interaction with the person's other, actually present beliefs ought to be taken into account when attributing content, but that possible beliefs should also be considered here. Let's assume that two people A and B express their belief that all men are mortal. A believes, moreover, that Socrates is a man, which leads him to conclude that Socrates is also mortal. B, on the other hand, has never even heard of Socrates. Clearly, the actual causal links of "All men are mortal" are different in each case. Nevertheless, we would attribute the same content to the beliefs of A and B. Hence, sameness of belief content does not require sameness of the actual interactions, but rather of the possible ones. If B knew who Socrates was, then his belief in the mortality of all men would also lead him to conclude that Socrates is mortal: The essential point is that [...] the type identity of a mental state is determined by its potential causal interactions with other mental states, with stimuli, and with behavior. Its type identity does not depend on the other mental states the subject happens to be in at the moment in question.42

Taken together, this means: If beliefs are theoretically determined states, then their identity must be determined by their possible causal links, by their causal potential, and must be independent of the interactions with other beliefs that are actually realized in a person. For contrast, let's examine the case of Mrs. T who, as a youth, was deeply affected by the assassination of President McKinley in 1901. In old age, she suffered from increasing memory loss and, though she still knew quite well that McKinley had been the victim of an assassination, she could no longer explain what an assassination was, or what death was; she didn't even know, whether she herself was still alive or whether she had died. Intuitively evaluated, i. e., from the point of view of folk psychology, we would surely say that Mrs. T did not hold the belief that McKinley had been assassinated. She didn't know what she was saying. Nevertheless, the corresponding sentence has lost nothing of its causal potential. If Mrs. T's memory could be restored, the belief statement which is isolated here would regain its earlier relations to other beliefs. By the causal criterion of individuation, one would have to conclude that Mrs. T believed in McKinley's assassination before and after her memory loss. Examples of this kind lead Stich to the conclusion 42

Stich 1983, 5 4 (emphasis in the original).

Intentionality and Syntactic Psychology

211

that folk psychology's taxonomy deviates significantly from the taxonomy that is derived from causal potential. If belief states are understood as theoretical states, then their intuitive criteria for individuation cannot be reproduced. 4 3 § 3 In fact, folk psychology's criteria for individuation are of an entirely different sort. In Stich's reconstruction of folk psychology, attributing a belief of a given content to another person follows by analogy with ourselves. When we ascribe a content sentence to someone, then we are characterizing this person's belief by means of a possible belief of our own: When A says, "S believes that p," he is saying that S has a mental token of a sentence stored in the way characteristic of beliefs, and this token is content-identical to the one which he (A) expresses by uttering "p." 4 4

Thus, according to Stich, folk psychology's essential means of characterizing belief contents is the similarity of these contents with our own belief contents. Such a strategy reaches its limits, however, if we are to identify the beliefs of people whose views and structures of thought are significantly different from ours. One could hardly claim to give an adequate characterization of the beliefs of children, the mentally ill and members of foreign cultures by relying on such a procedure. A sort of Protagorean parochialism is thus built into intentional folk psychological concepts, since when it comes to ascribing content, men and women like us are the measure of all things. 45

While the first step of the argument is supposed to demonstrate the significant difference between the folk psychological and the causal individuation of beliefs, the second step claims the folk psychological approach is methodologically inferior. In essence, Stich's methodological argument says that folk psychology suffers from a parochial approach that is fixed on ourselves and misses the peculiarities of other, different types of mental structures. 46 § 4 Stich sees the root of all evil in the fact that semantically interpreted, meaningful representations are used at all in scientific psych43 44 45 46

For this argument and this example, cf. Stich 1983, 51 —56. Stich 1983, 76. Stich 1983, 7 (emphasis in the original). Cf. Stich 1983, 1 0 0 - 1 0 4 , 1 3 5 - 1 4 8 .

212

Psychology and Consciousness

ology. Psychology should undertake to identify its terms using the causal profile, but reject every semantic interpretation of these terms (such as beliefs of a certain content). The core of the syntactic theory of the mind is, therefore, the idea that the causal links between mental states or events should be reflected in the formal relations in a structure that is defined exclusively by syntactical means: T h e basic idea of the S T M [Syntactic Theory of the Mind] is that the cognitive states whose interaction is (in part) responsible for behavior can be systematically mapped to abstract syntactic objects in such a way that causal interactions a m o n g cognitive states, as well as causal links with stimuli and behavioral events, can be described in terms of the syntactic properties and relations of the abstract objects to which the cognitive states are mapped. M o r e briefly, the idea is that causal relations a m o n g cognitive states mirror formal relations a m o n g syntactic objects. 4 7

This makes it clear that syntactic theory and cognitive science sketch the same picture of mental operations. Psychological theory associates formal entities with cognitive states in such a way that the causal relations between those states correspond to the formal relations between those entities. Causality is reflected in the syntactic structure. This view of the structure of psychological theories is modeled on explanations in the natural sciences, in which likewise (1) the correspondence of causal and logical connections occurs 48 and, moreover, (2) the latter are conceived of as formal, mathematical operations. So there is unanimous approval for the goal of expressing effects by means of formal consequences. The differences only appear when one asks which features can be appropriately ascribed to the objects of mental operations, i. e., which properties mental representations possess. Cognitive science demands, in sum, a double correspondence of levels of description: a correspondence of causal and syntactic relations (and this correspondence determines the picture of mental operations), as well as a correspondence of syntactic and semantic features (through which intentional terminology receives its place in psychology). Syntactic

47 48

Stich 1983, 149. Hertz already regards this correspondence of causality and logic as the true center of scientific explanation: "We form for ourselves images or symbols of external objects; and the f o r m which we give them is such that the necessary consequents of the images in thought are always the images of the necessary consequents in nature of the things pictured." (Hertz 1894, 1).

Intentionality and Syntactic Psychology

213

theory only denies the appropriateness of the second postulate, not of the first. Hence, its picture of psychodynamics remains within the framework of cognitive science. Only intentionality, which is so often thought to be the defining characteristic of mental language, is barred from psychology. Stich's argument for an exclusively syntactic interpretation of mental states is conceived following Fodor's rejection of a formal definition of semantic content. As presented in VII.2.§ 5, Fodor rightly points out that the content of a syntactically characterized string of symbols can be interpreted in a variety of ways. Similarly, a psychological theory that specifies types of mental states using only the latter's interaction profiles, is not able to provide a semantic determination of the mental states that it assumes. 4 9 This is the reason for Stich's surprising maneuver not to defend a semantics that deviates from the usual one, but rather to reject any psychosemantics. § 5 We shall next test the viability of Stich's arguments by first examining the claim that a scientific psychology cannot reproduce the belief states as individuated by folk psychology. Stich's justification for his thesis, that there are divergences between the folk psychological and the causal taxonomy that cannot be bridged, rests on an inadequate characterization of the causal criteria for identifying beliefs. Namely, Stich's argument is based on an ambiguity in the concept of possibility. His insistence on considering possible interactions with other beliefs — and his complementary neglect of how beliefs are imbedded in someone's actual system of beliefs — derives its plausibility from the example of the mortality of all men, in which one of the partners (B) happened not to have heard of Socrates, but could, of course, be told about him at any time. In this case one is in fact justified to disregard B's ignorance of this fact when individuating beliefs and to take into account the possibility that he could be informed of it. In the case of Mrs. T. by contrast, it simply would not be possible for someone to come along and explain to Mrs. T what an assassination is and what it entails. T h e pertinent conceptual context cannot be established. So the relevant belief states are, in fact, not available to Mrs. T and not just coincidentally unrealized.

49

C f . Stich 1 9 8 3 , 108.

214

Psychology and Consciousness

We wish to express this difference such that, in the case of B, interaction with the belief states addressed is technically possible (meaning it can, in fact, be produced), in the case of Mrs. T, on the other hand, it is at best physically possible (namely, if her loss of memory could be cured without violating the laws of nature) or only logically possible (namely, without self-contradiction). 50 Lack of actual interaction can only plausibly be taken as irrelevant to the identification of beliefs, if the interaction would technically be possible at any time. Hence, if beliefs are individuated on the basis of causal potential, then this should be specified as technically possible causal potential. Using this explication, Stich's examples can be reconstructed without any trouble. Hence, Stich's argument does not suffice to show that cognitive science's formality condition cannot be satisfied, i. e., to preclude the possibility that belief states could be adequately individuated by recourse to their integration into a network of other beliefs. 51 This conclusion is supported by the investigations carried out in VI.1, whose essential goal was precisely to show that the intentional concepts of cognitive psychology can legitimately be understood as theoretical terms. Thus, it cannot be ruled out by means of general considerations that beliefs turn out to be respectable theoretical states. This is true even if one believes (with Stich) that folk psychology is characterized by an approach which proceeds from analogies to ourselves. Even if intuitive explanations of behavior are supposed to be governed by a hermeneutical method in this sense, the mere use of intentional terms need not oblige us to use the methods of hermeneutics. This can be seen more clearly by recalling the presentation of cognitive psychology in V.2. As is apparent from the examples discussed there, the use of intentional terms does not require analogies based on empathetic understanding; rather, it is governed by criteria that can be intersubjectively determined. Apart from that, it should be emphasized that even the existence of these sorts of methodological distinctions need

50

51

The distinction between technical, physical and logical possibility was developed in the framework of Logical Empiricism to specify the principles of verificationist semantics. According to verificationist semantics, only verifiable statements have meaning. But "verifiability" means precisely the "possibility of verification"; and the distinction used here was introduced in this context. This conclusion is also reached (even if for other reasons) by Marras 1987, 1 1 8 - 1 1 9 . However, there are other difficulties that stand in the way of such an individuation; these will be addressed in VII.4.

Intentionality and Syntactic Psychology

215

not imply actual distinctions among the things that are characterized differently. The same entities can be arrived at by different paths; an epistemological difference does not necessarily lead to an ontological difference. § 6 After Stich's argument for his thesis of fundamental divergences between a taxonomy according to content and a taxonomy according to inferential relations has been examined and rejected, we will now discuss the plausibility of Stich's arguments in favor of an exclusively syntactic use of psychological terms. To begin with, Stich's claim (developed with reference to Fodor's presentation) that the consideration of relations between formally characterized states does not yet establish a semantic interpretation of these states is correct. N o r is this surprising, since it is simply (as we said) the reformulation of an old insight (cf. VII.2.§ 5). On the other hand, a semantic interpretation can be achieved, as a rule, by ascribing certain observable properties to certain expressions in the syntactic structure. For example, after the observable properties of a straight line had been ascribed as indicators to a merely formally characterized entity in Hilbert's axioms, Euclidean geometry is singled out as a model. In this way, the formal system of axioms has attained a semantic interpretation. In other words, the specification of the inferential relations (i. e., of the causal potential) can, together with the attachment to an observation basis, indeed lead to a semantic interpretation of a structure. Precisely this strategy is characteristic of the approach described in VI.1.§ 2 for interpreting the content of theoretical terms: A theory is endowed with a semantic interpretation by means of both the formal network of terms contained in it and the link of several of its theoretical terms to observational indicators. Hence, an empirically applied structure is, by the same token, a semantically interpreted structure. Stich also notes that there is a difference between a psychological theory and Fodor's syntactic machine states, insofar as there is in the former but not the latter an attachment to sensory stimuli and behavioral responses (and not just to other syntactic entities). Nevertheless, he counters the relevance of this fact by means of the following thought experiment: If all children immediately after birth had optical instruments that systematically altered colors installed in front of their eyes, then, say, a green stimulus would lead to the belief that one was seeing,

216

Psychology and Consciousness

say, a red object. This means that the causal profile does not determine the semantics, even if stimuli and responses are also taken into account. 5 2 But Stich's example in no way permits this inference. It merely shows that a different connection to the empirical basis (namely, in one case with and in one case without the color inversion) leads to a different semantic interpretation of the descriptive terms. In both cases, in fact, every color predicate has a well-defined referent, 53 and the terminological system as a whole has a semantic interpretation (even if it differs in each case). The uncertainty expressed in Fodor's example of the ambiguity of syntactic machine states can be resolved analogously, namely, by taking into account the empirical connection of the abstract states (and that means, in this case: the respective contents of the screens). At this level, there are clearly differences between a simulation of the Six Days' War and a game of chess. In sum, this consideration leads to the following result: At least some of the terms of a psychological theory must necessarily be connected to empirical features; otherwise, one merely has a mathematical algorithm, not a psychological theory. But in light of our foregoing discussion, it is apparent that this empirical connection already contains a semantic interpretation of the descriptive terms. Hence, the idea of a purely syntactical psychology is self-defeating or incoherent in this sense. It is true, on the other hand, that there is a significant peculiarity here in that an interpretation of the descriptive vocabulary does not by itself provide an exhaustive interpretation of the contents of mental states. 54 § 7 As mentioned in VII.3.§ 1, Stich seeks to support syntactic options not only by using general methodological considerations, but also by claiming that intentional terminology does not, in fact, play any role in cognitive psychology. It would follow from this that intentional terms are questionable not only for empirical reasons but also for methodological ones. Stich tries to justify this assessment by showing that folk psychology — of which intentional language is characteristic — conflicts with cognitive psychology in a fundamental way.

52 53

54

Cf. Stich 1983, 1 0 8 - 1 0 9 . This referent is, moreover, also accessible to the unlucky inhabitants of the colorinverted world, since they can perceive the instrument causing the inversion and can correctly analyze its effects (just as we can do with our system of sensory perception). On this important fact, cf. VII.4.

Intentionality and Syntactic Psychology

217

Stich puts forward two arguments to back this claim. The first appeals to cognitive dissonance (cf. V.2.§ 3). Stich points out that none of the subjects in the relevant experiments answered the question about the reasons for their behavior by talking about the processes of dissonance reduction that the theory postulated — instead they gave other reasons. Nevertheless, the behavior of the subjects could still very well be explained by recourse to these processes. This suggests the interpretation that we are endowed with two cognitive systems: one responsible for our verbal behavior and the other for our non-verbal behavior; and both systems are largely isolated from one another. Such a hypothesis of dual control is, for Stich, incompatible with folk psychology. According to the latter, namely, the beliefs identified by their similarity have certain functions. They are both the basis for the utterances as well as the basis for actions. For this reason, one can assume that there are no such things as beliefs: What is striking about the results [that support the dual hypothesis] [...] is that they strongly suggest that our cognitive system keeps two sets of books [subsystems] in this way. And this is a finding for which folk psychology is radically unprepared. If it is true, then states similar to the one underlying our own ordinary utterance of "p" do not also participate in the production of our nonverbal behavior. In those cases when our verbal subsystem leads us to say "p" and our nonverbal subsystem leads us to behave as though we believed some incompatible proposition, there will simply be no saying which we believe. Even in the (presumably more common) case where the two subsystems agree, there is no saying which state is the belief that p. If we really do have separate verbal and nonverbal cognitive storage systems, then the functional economy of the mind postulated by folk theory is quite radically mistaken. And under those circumstances I am strongly inclined to think that the right thing to say is that there are no such things as beliefs.55

Stich's second empirical argument appeals to theories of memory. To satisfy the demands of folk psychology, one would have to assume modularity in a theory of memory, i. e., it must be possible to identify the contents as assigned by folk psychology as the contents of memory. That is, the only theories of memory that are compatible with folk psychology are those that assume memory stores sentences. But this is

55

Stich 1983, 231 (emphasis in the original). For the complete argument, cf. Stich 1983, 230-238.

218

Psychology and Consciousness

not the case in several of the models currently under discussion. One of these suggests instead that none of the distinct units or parts of the mental model " h a v e m e a n i n g s in themselves" and thus none can be identified with individual beliefs, desires, etc. M o d u l a r i t y [...] is violated in a radical way since m e a n i n g or content emerges only f r o m " g r e a t webs of s t r u c t u r e " a n d no natural part of the system can be correlated with " e x p l i c i t " or verbally expressible beliefs. 5 6

Horgan and Woodward discuss Stich's empirical arguments. They point out that, in the case of dissonance theory, there is, in fact, no problem with ascribing beliefs, since one can appeal to unconscious beliefs. Thus, the usual psychological interpretation of these findings is in no way characterized by a fundamental rejection of intentional terms. Rather, it is assumed that behavior is at least occasionally determined by unconscious psychological states, while introspective reports on the reasons for this behavior, on the other hand, depend on conscious psychological states. While such an interpretation is, indeed, a modification of folk psychology, in no way does it imply a wholesale rejection of folk psychology. Moreover, for a number of psychological explanations in the framework of cognitive psychology (such as the attribution theory as well as dissonance theory), the appeal to beliefs is essential. On the whole, it cannot be said that psychology counsels us to abandon intentional terms. 57 In addition, according to Horgan and Woodward, the modularity principle is an overly strong condition for matching folk psychology and cognitive psychology. It is not necessary that the fundamental states assumed or the fundamental concepts introduced within the frameworks of both theories coincide in order for both theories to refer to the same entities. After all, the complex states of cognitive psychology could well be identical to the simple states of folk psychology. 58 In that case, there would be no correspondence on the fundamental level, but there would, nonetheless, be agreement of reference. § 8 Horgan and Woodward's argumentation clearly proceeds along the same lines as our discussion of Fodor's reduction condition (cf. 56 57 58

Stich 1983, 241. Horgan/Woodward 1985, 2 0 7 - 2 0 9 . Cf. Horgan/Woodward 1985, 211-217..

Intentionality and Syntactic Psychology

219

II.5.§5). In fact, Stich's modularity is nothing other than an application of Fodor's d e m a n d for one-by-one assignment of natural kind terms. In light of this reconstruction, Stich's condition for identifying the objects described by folk psychology and scientific psychology a m o u n t s to a demand for an overly strict reduction relation between the t w o theories. We agree with H o r g a n and W o o d w a r d that Stich's examples in no way suggest that all types of intentional speech should be rejected. O n the other hand, Stich rightly calls attention to the possibility that the content of scientific psychology's terminology may deviate significantly f r o m the familiar vocabulary of folk psychology, so that an identification of the states and processes referred to in the t w o approaches is problematic and in need of defense, even if both approaches employ intentional concepts. Even independently of Stich's arguments, there are prima facie doubts regarding such an identification. We just mentioned that many of the terms of scientific psychology seem to lack a counterpart in folk psychology. Moreover, the functional role that the familiar folk psychological terms play in scientific psychology, is not always familiar. In dissonance theory, for example, the relationship of belief and action is often surprising. Following folk psychology, we take the introspectively accessible belief to be the cause of the action, but, as in the case of dissonance, it is sometimes the action and its effects that are the cause of the beliefs. This can be gathered f r o m the example of the changes in attitude induced by dissonance that was discussed in V.2.§ 3. Even where there is a greater conceptual similarity, the natural kind terms of folk psychology and cognitive psychology often do not coincide completely. For example, medical psychology divides the single folk psychological term "headache" into — a m o n g others — the terms "muscle contraction, or tension, headache" and "migraine" and attributes different causes and effects to the entities so characterized. 5 9 But if psychological terms are regarded as theoretical terms (as we d o here), a drastic change in the theory can be expected to lead to a change in the meaning of the relevant terms. To be sure, this does not entirely rule out the possibility that psychology refers to our familiar mental states (among other things); but it is anything but certain that such an identity of reference exists.

59

This is, at least, the received view. In more recent approaches both types of headache are traced back to a common metabolic dysfunction. We shall leave that aspect out of consideration.

220

Psychology and Consciousness

§ 9 Horgan and Woodward have presented an interesting argument in favor of the existence of such an identity. They also emphasize that one can only sensibly speak of folk psychological mental states, if these states exhibit the structure of causal links that folk psychology attributes to them. 6 0 This means that the adequacy of assigning singular properties (such as "S believes that p" or "S has pains") crucially depends on the correctness of the general principles that connect the belief or perception states to one another or with behavior. Thus, it doesn't make sense to say that the tokens may exist, but not the relevant types. Horgan and Woodward point out that the natural kind terms of cognitive psychology and those of folk psychology do not, in fact, coincide, but that every natural kind term of the latter can be expressed by means of several natural kind terms of the former. In Horgan and Woodward's argument, this implies that cognitive psychology does not refute folk psychology, but rather that the mental states specified by folk psychology are actually causally effective. To this end, they introduce the concept of minimal causation-. An event e is a minimal cause of an event f , if e causes f and no part of e causes f . This leads to the following two assertions: (1) e can be a genuine cause of f , without being a minimal cause of f . (2) If e causes f as well as g, then the part of e that minimally causes f can be different from the part of e that minimally causes g. The pertinent thesis is then: If folk psychology refers both f and g back to e, and cognitive psychology introduces two partial events e\ and £2, so that e\ minimally causes f and ei minimally causes g, then this does not refute folk psychology. 61 The intuition behind Horgan and Woodward's argument is obviously that cognitive psychology is, indeed, more precise than folk psychology, but that the former leaves the causal architecture of the latter intact to such an extent that one cannot really speak of a true dislodgement and substitution. In fact, the argument can be reconstructed as the claim that there is a relationship of corrective reducibility between cognitive and folk psychology. T h a t is, an identification of the relevant entities is supposed to be established by means of a reduction of the pertinent theories. Translated into our terminology, this means that Horgan and 60 61

Cf. H o r g a n / W o o d w a r d 1985, 217. Cf. H o r g a n / W o o d w a r d 1985, 2 1 8 - 2 1 9 .

Intentionality and Syntactic Psychology

221

Woodward claim that folk psychology is relatively interpretable in cognitive psychology. If "—>" expresses the causal relation, then the folk psychological laws e —>/" and e —» g must be relativized to the conditions e\ (in the first case) or ei (in the second case), in order to become valid laws in cognitive psychology. The above-mentioned subdivision of headaches into different types (such as migraine and tension headache) can serve as an example of such a relation. One of the effects of migraine is increased sensitivity to light, one of the effects of tension headaches is a numbness in the neck and limbs. One could imagine two folk psychological laws of the following type here: (1) headaches (sometimes) cause increased sensitivity to light (corresponding to e —• f ) , (2) headaches (sometimes) cause numbness of limbs (corresponding to e - > g ) . Medical psychology's relativization would then consist of relativizing the truth of (1) to the satisfaction of the condition "migraine" and the truth of (2) to the satisfaction of the condition "tension headache." A relative interpretation of folk psychology in cognitive psychology could be achieved in this way. To be sure, relative interpretability is only a necessary, not a sufficient condition for reducibility (cf. II.2.§ 4). In addition, the relativizing condition must be approximately fulfilled in the interpreted law's domain of application. If this is not the case, then one would have to speak of a non-reductive theory replacement and not of a reduction. We are not in a position to decide which of these two possibilities is actually realized. Nevertheless, even a superficial look reveals that a reduction certainly does not exist, if the folk psychological event e can be subdivided alternatively and exhaustively into e\ and ei (which, to all appearances, Horgan and Woodward have in mind), and the situation is otherwise as described. In that case, either the relativizing condition e\ is approximately fulfilled (namely, if most of e are also ei), or ei is approximately realized (namely, if most of e are also ei), but not both. And this implies that one of the two interpreted laws (i. e., either e -> f or e —> g) cannot be reduced. This should cause us to be skeptical about Horgan and Woodward's thesis of a smooth and unproblematic adjustment of folk psychology into scientific psychology. § 10 This judgment should not yet be considered a reason to favor the syntactic option. On the contrary, we believe this option is poorly supported and insufficiently elaborated. In any case, it is not clear what kind of psychology Stich's approach amounts to. For example, Stich

222

Psychology and Consciousness

discusses the case we also mentioned of the belief a building was on fire. This case is intended to show the existence of functionally equivalent, but neurophysiologically different states (cf. II.5.§ 2). In this context, Stich emphasizes that syntactic theory exactly reproduces the functionalist explanation, since it also advocates the view that neurally different paths result in the same "B-state of the type F " (which reminds one of the belief that there is a fire, but in no way expresses this belief). 62 Clearly, the old, familiar semantically interpreted mental representations simply appear in new syntactic clothes. In a similarly conciliatory mood, Stich explicitly raises the possibility that a syntactic psychology approximately reproduces the folk psychological architecture of desires and beliefs, so that the latter could be reduced to the former. 63 Since, just a few chapters before, he was talking about fundamental and insuperable divergences between the two approaches, this reversal makes Stich's position somewhat less clear. Occasionally, syntactic theory seems to be a mere verbal radicalism, whose main achievement is providing well-known phenomena with a new label. The precise form of syntactic terminology also remains in the dark. Marras supposes here (despite all Stich's protests to the contrary) that it is a revival of behavioristic vocabulary: in describing behavior, "arm-raisings" should be replaced by "arm-risings." 6 4 It seems more likely to us that Stich's model essentially favors a means of expression that corresponds to the fusion theory of intentionality (cf. VII.2.§ 2). As we said, Stich cannot entirely avoid a semantic interpretation of psychological entities, but he attributes no content to mental states or events. But this is analogous to fusion theory's model, which says the statement "A-believes-that-tf" is to be interpreted as a semantic whole, in which no content sentence can be isolated. Stich's theory would then have to be understood as the claim that Fodor's methodological argument against such an approach was insufficient, since the possible disadvantages are outweighed by other sorts of methodological virtues. It follows from our discussion of the syntactic model, however, that there are no such virtues. Stich's theory represents an attempt to implement the program of eliminative materialism, 65 which rejects intentional speech (cf. II.6.§ 1). Cf. Stich 1983, 1 7 2 - 1 7 3 . « Cf. Stich 1983, 222 - 223, 227. 6 4 Cf. Marras 1987, 1 2 6 - 1 2 7 . 6 5 Accordingly, Stich's approach is praised by Churchland. Cf. P. S. Churchland 1986, 382-383. 62

Psychosemantics and Folk Psychology

223

But his argument also calls attention to an important fact: namely, even if Brentano and Chisholm are correct in their thesis that intentional terms cannot be defined by physical terms, this thesis represents a valid objection to an exclusively physical description of the mental only if intentionality is an essential feature of the mental. Every form of eliminative materialism is compatible with this kind of non-definability; the latter would merely be another indication that intentional terms should be replaced by more suitable ones. If it is thought to militate against neurophysiology or physics that intentional terms cannot be expressed in its framework, this would be comparable to objecting that thermodynamics did not permit a definition of the matter of heat. Accordingly, intentionality is not to be defined, but to be gotten rid of. 66 Cognitive science does not aim at a physical definition of intentional terms anymore than eliminative materialism does, but it does try to use non-intentional indicators in order to understand and identify differences that can be expressed intentionally. Moreover, in both approaches the causality of psychodynamics is not located on the level of intentional description. What can be gathered from all this is that, even given a consensus about the irreducibility of intentionality, the status of intentionally described mental states or events remains to be explained. They could be significant or insignificant, causally effective or ineffective, accessible or inaccessible to consciousness. This means that philosophy of consciousness has a lot of free play that needs to be examined more carefully. The first step is to reconstruct the status of intentional speech in the framework of the positions in the philosophy of mind that have been discussed so far; then we will discuss the different theories about the content of mental states and, finally, develop our own view.

4. Psychosemantics

and Folk

Psychology

§ 1 Cognitive science and the syntactic theory of the mind can be reconstructed as explanatory programs for cognitive psychology (a sample of which was presented in V.2) or folk psychology. T h a t is to 66

This makes it clear again in hindsight how it is premature of Davidson to accept intentionality as the defining feature of the mental without any further argument (cf. III.3.§ 2). The analysis presented here makes it clear that Davidson must show that Stich cannot be right, a n d that for reasons a priori.

224

Psychology and Consciousness

say, both are meta-theoretical approaches. Cognitive science aims at a reproduction of folk psychology by means of a reconstruction of cognitive psychology that is based on the formality condition; syntactic theory is a program for replacing folk psychology while simultaneously reproducing cognitive psychology. A significant characteristic is that semantically psychological states in cognitive psychology are meaningful, interpreted entities; this characteristic is accepted by cognitive science, while syntactic theory attempts to abolish it. Another significant aspect is that only in cognitive psychology is the semantic interpretation causally significant as well. Cognitive science and syntactic theory seek, by means of a theory of formal operations, to reproduce or, respectively, to overcome generalizations about the content of mental representations. As presented in VII.2.§ 7, in the case of cognitive science this leads to a remarkably parallelistic view of the relationship between those properties that are accessible to conscious experience and those on which a scientific description is based. The syntactic interpretation of phenomena of consciousness, on the other hand, can be characterized as epipbenomenalistic. For epiphenomenalism, conscious phenomena are merely by-products of physiological mechanisms that do not themselves interfere with these mechanisms (cf. 1.3.$ 3). Though epiphenomenalism is an interpretation of the mind-body relation, it can easily be transferred to the mind-mind relation. An epiphenomenon is generally a state (or event) which occurs in the context of a particular complex of states (or events), but is not itself causally effective nor can it reasonably be considered an indicator of a causally effective state (or event). Furthermore, not only can certain states be epiphenomena, but also certain aspects or properties of such states. Using this terminology, it is clear that, in the framework of syntactic theory, the conscious aspects of mental states become mere epiphenomena of the syntactic characteristics of these states; it is the latter that have complete command over the mode of operation of the psychological machinery. Expressed in terms of Huxley's analogy, only the sound of the whistle can be experienced introspectively, not the piston strokes. A third significant aspect of a philosophy of consciousness is the individuation of states. Here the issue is the relationship of the mental types, as assumed within the three approaches we discussed, to the natural kinds, as assumed within folk psychology. Cognitive science understands cognitive psychology as an organic continuation of folk

Psychosemantics and Folk Psychology

225

psychology and, hence, contends that the types of states in scientific psychology coincide (extensionally) approximately with those of folk psychology. Syntactic theory, on the other hand, claims there are significant divergences between the types of phenomena in syntactic and folk psychology. We ourselves hold the view (in contrast to the perspective of cognitive science) that considerable differences exist between the taxonomy of folk psychology and that of cognitive science; this view will be substantiated in the following. But first, the results of our discussion up to this point should be summarized in a table. The following claims are characteristic of the approaches discussed:

Cognitive Psychology Cognitive Science Syntactic Theory

Semantic Interpretation of Mental States

Causal Efficacy of Interpretation

Individuation of States like that of Folk Psychology

Yes

Yes

No

Yes

No

Yes

No

No

No

Table 2 § 2 For a philosophy of consciousness, it should be noted that — as already mentioned several times — although cognitive science is framed in intentional terms (i. e., the type of the concepts it employs coincides with that of folk psychology), the psychological states thus described are completely new (i. e., their content differs from that of the mental phenomena assumed by folk psychology). Although motives, for example, are introduced as dispositions for expecting and evaluating the possible outcomes of a situation (cf. V.2.§ 3) and, hence, fall under categories familiar from folk psychology, the function that motives take over in information processing and in regulating behavior transcend the limits of folk psychological regularities. The Pygmalion effect has no place in folk psychology. By using folk psychological (i. e., intentional) concepts, one can formulate regularities other than the folk psychological ones. This observation suggests that we should distinguish between folk psychology in the technical sense and in the intuitive sense. 67 Folk 67

On folk psychology in the practical sense, cf. IX.§ 5.

226

Psychology and Consciousness

psychology in the technical sense is characterized by reference to prepositional attitudes and, correspondingly, by its use of intentional terminology. Folk psychology in the intuitive sense, on the other hand, means the model for explaining actions that is familiar to us from our daily lives; it includes the pearls of wisdom grandmother's jewelry box supplies to explain human behavior. Our claim is that folk psychology in the technical sense goes beyond the realm of folk psychology in the intuitive sense. While it is true that cognitive psychology is oriented towards folk psychology in the technical sense, since it makes use of intentional terminology, it nevertheless contains laws that are not found in grandmother's jewelry box and, accordingly, is not based on folk psychology in the intuitive sense. The same picture emerges if unconscious psychological processes are included in the discussion. As was mentioned, Horgan and Woodward bring the possibility of introducing unconscious processes to bear against Stich's empirical arguments, in order to show that folk psychological concepts are still applicable. This succeeds, but only at the price of the applicability of folk psychological regularities. The acceptance of unconscious states implies that one of folk psychology's essential observational indicators of mental states or events — namely, introspective reports — must be limited in its domain of application. Unconscious states or events cannot be reported at will; the subjective privileged character of the mental is lost. If one considers that many of the intermediate cognitive processes assumed by psychology are of this kind, i. e., they possess no introspectively accessible indicators, then one is led to conclude that our conscious experience provides an, at best, incomplete picture of our mental life and the factors that affect it. This situation can be expressed in two different ways. A considerable part of folk psychology's explanations of actions is based on the principle introduced in II.6.§ 2: (A wishes that b) A (A believes that a b) {A sets about to produce a) (if A is not distracted or confused, no other goals take precedence and A believes that he can bring about a, etc.)

(1)

This law can be retained if it is stipulated that the person does not always know his own beliefs and goals. This must be expressed in such a manner that the following folk psychological law is false: (A believes that a) A (A is sincere) —> (A reports on demand that he believes a)

(2)

Psychosemantics and Folk Psychology

227

The alternative is to retain (2), which would, however, require abandoning (1). In that case, the person always knows his desires and beliefs; but they are not always linked to his actions according to the model of (1). In both cases, there is unquestionably a considerable deviation from the model of explanation familiar from intuitive folk psychology. This conclusion is also supported by the psychological interpretation of experiments with split-brain patients. The neural connections between the two hemispheres of the brains of these patients have been severed (for medical reasons). One of these patients is asked to perform a certain action and this request is made using optical information provided only to the left side of the patient's field of vision and, hence, to the right (mute) cerebral hemisphere. When the patient is then asked why he performed this action, the left hemisphere of the brain (where the linguistic capability resides) knows nothing of the request the right side obtained; the patient gives other, entirely sensible reasons. Gazzaniga and LeDoux have generalized this trait to physiologically intact people. They don't regard the ego as a unified psychological entity and assume, instead, a variety of distinct mental systems in the brain. All of these systems can produce behavior (if certain conditions are realized). These mental systems are often disconnected from one another; and this has the consequence that the conscious ego which is capable of speech is not always aware of the actual reasons for an action (which may be rooted in the activity of an isolated mental system). The ego perceives these actions and invents plausible and sensible reasons for them. Hence, in this model, consciousness is not a unified entity, but rather contains a variety of isolated, individual entities. The ego has a "compartmentalized" structure. 6 8 As with cognitive dissonance, here too, the beliefs are rather the consequences of actions than their causes; and such a result is certainly incompatible with intuitive folk psychology. This discussion does not yet outright preclude the possibility that folk psychology (in the intuitive sense) is reducible to scientific psychology. This depends on the degree to which the folk psychological regularities are distinct from those of cognitive psychology, i. e., on how well the corresponding relativizing conditions are met. The answer to this question demands a comprehensive and detailed examination of the relationship of folk psychology to cognitive psychology, which would 68

Cf. Gazzaniga/LeDoux 1978, 1 4 6 - 1 5 1 .

228

Psychology and Consciousness

exceed the scope of this study. 69 In any case, the cases we have looked at thus far make a reduction seem unlikely. In conclusion, this means that use of intentional vocabulary is, indeed, a property characteristic of scientific psychology, but that this peculiarity is much less significant than is often assumed. Despite this peculiarity, the contents of the intentionally described states can deviate widely from the states known from folk psychology or from conscious experience. This can be expressed such that the cognitive states, but not the folk psychological states, are the bearers of psychodynamic causality, i. e., such that (folk psychologically described) consciousness is merely a by-product of psychological processes. 70 On the other hand, both causally effective and causally irrelevant mental states can be adequately described in intentional language; hence, both are of the same type. It is a common mistake in today's philosophy of mind to identify the folk psychological approach with the use of intentional terms, without considering that another sort of picture of our conscious existence could also arise within the boundaries of intentional description. "What should be made of folk psychology in light of all this? It is hardly justified to suppose that intuitive folk psychology presents an adequate picture of our mental existence. This means, it does not even correctly reproduce psychokinematics (as was still assumed in VII.2.§ 7). It does, however, fulfill another important function. Namely, it serves as an observation theory for the ascription of mental contents. In order to make this claim plausible, we will first discuss the problem of psycho semantics and then (intuitive) folk psychology's contribution to the solution of this problem. § 3 Psychosemantics seeks an answer to the question: what determines the content of mental states, or in what way can mental states represent something? This question is especially important to cognitive science, since it claims that mental states are specified syntactically in the first place. In doing so, it sets itself the problem of determining what fixes the semantic interpretation of formally characterized mental states. In light of the analysis in VII.3.§ 6, the most obvious strategy for solving 69

70

Such an examination is, moreover, hardly possible at present, because there is not always a neatly characterized state of research for all fields of cognitive psychology. For this reason, the (potential) reducing theory often cannot be clearly recognized. As a consequence, the formality condition refers to cognitive states, not folk psychological ones.

Psychosemantics and Folk Psychology

229

this problem lies in seeing how a mental state is integrated into a network made up of the sensory perceptions that precede it, the actions that follow it, and other mental states. If mental terms can be legitimately characterized as theoretical terms, then the theoretical context in which they appear and their connection to the observational basis determine their meaning. For types of psychological states (like beliefs and goals), this condition is satisfied. For example, the psychological type "belief" can be roughly characterized by saying that beliefs are possible effects of sensory perceptions and that they interact with goals and actions in the way described by principle (1) above. 71 Hence, the characterization of types of mental states corresponds to the model for determining the content of other theoretical terms and does not, therefore, produce any particular difficulties. This is not the case in the characterization of the content of mental states. It does not suffice here to appeal to the interaction profiles of the corresponding states. Such a content attribution based on the interaction profile could be oriented, on the one hand, on those interactions of a mental state that are actually realized in a particular person. But, as was discussed in VII.3.§ 2, this would lead to an overly specific individuation of beliefs. Obviously, it makes no difference to the content of the belief in the mortality of man, whether someone knows who Socrates is, even though the actual profile of interaction of this belief might very well be influenced by this knowledge. On the other hand, the possible interactions of a mental state could be used to determine its content. In fact, our reply to Stich's objections is based precisely on a more specific version of this option (cf. VII.3.§ 5). Although this reply invalidates Stich's argument, it has other weaknesses (as we must admit now) that disqualify it as a general contribution to the problem of psychosemantics. The core of the difficulty is that actual interactions may sometimes very well influence the content of mental states. Let's consider the tragic case of Oedipus, for example, who unwittingly marries his own mother. The dramatic aspect of this story lies in the fact that Oedipus does not know at first the true identity of his wife Jocasta; he learns this only later. The revelation of Jocasta's identity clearly leads to a dramatic change in the content of Oedipus' beliefs with respect to his marriage. But this change can only be reconstructed 71

Cf. Fodor 1987, 6 9 - 7 0 .

230

Psychology and Consciousness

by recourse to the actual interaction profile of Oedipus's beliefs. After all, the new information "Jocasta is my mother" crucially alters the profile of interactions of the previously existing belief "my marriage is morally sound and it pleases the gods." Correspondingly, the content of this belief has suddenly turned into its opposite. If one is relying on the technically possible interactions, this change is incomprehensible. For an interaction between new information and old beliefs was always technically possible. Oedipus could have been informed about the true identity of his wife at any time. In this version, the interaction theory results in an overly coarse individuation of the content of beliefs. This example makes it clear that the identity of the content of beliefs cannot be adequately determined on the basis of the interaction theory, since essential and inessential interactions must be distinguished. In some cases, a change in the actual connections of beliefs is relevant to the content of these beliefs, but not in other cases. And determining the relevant connections apparently requires in turn knowledge of the contents of the pertinent beliefs. For this reason, the interaction theory is unfit for identifying the content of beliefs. 72 § 4 In addition to the interaction theory, there are essentially two other theories of psychosemantics currently being advocated, namely, correlational the teleological or Darwinistic model and the causal or model. The teleological model, in the form that derives from Papineau, for example, sees the content of beliefs as being determined by their truth conditions, and these in turn are determined by the success of the actions that are based on these beliefs. The success of an action is decided for its part by criteria from the theory of evolution: Successful actions contribute to the dissemination of one's own genes. In sum, the teleological theory asserts: That the belief "tree" represents a tree can be accounted for by showing that it is the biological purpose — as it resulted from natural selection — of the belief "tree" to be evoked by trees. 73 But this (rather outlandish) approach, too, has a series of weak points. The most obvious objection is that the assumption that the 72

73

T h e Churchlands overlook this aspect when they claim that intentionality is sufficiently accounted for by the context theory or network theory of meaning. Cf. P. M . Churchland 1981, 6 8 - 6 9 ; P. S. Churchland 1986, 3 4 4 - 3 4 6 . Cf. Papineau 1987, 63, 6 9 - 7 0 .

Psychosemantics and Folk Psychology

231

content of beliefs is determined by natural selection implies that the identity and individuality of all belief contents must be innate. Papineau rejects this objection and would have it that the concept of natural selection also applies to learning processes in general: T h e r e is no need to restrict ourselves to inter-generational selection here. N a t u r a l selection occurs within generations, by learning, as well as between generations, by genetic changes. We can think of learning as selecting c o m p o n e n t s for our cognitive m e c h a n i s m s , a n a l o g o u s l y to the w a y that inter-generational evolution selects genes. 7 4

When extended in this way, the teleological model is, however, Darwinistically oriented only in a metaphorical sense. Apparently, the point is merely that the content of a person's beliefs is fixed by external rewards (or Skinnerian reinforcements). 75 But then it is difficult to understand why a person would maintain beliefs that have unfavorable consequences or are even (e. g., in the case of a martyr) responsible for his untimely death (which would significantly decrease his chances for disseminating his genes). Papineau tries to take this type of objection into account by determining the value of beliefs in natural selection in the first place by seeing whether they satisfy a person's needs or serve the goals he is trying to achieve. And all beliefs and goals, according to Papineau, can finally be reduced to and explained by the goal of the dissemination of genes (even if the person in question does not realize it). 76 Hence, the teleological model is based on the assumption that all human goals and needs (and, hence, indirectly all of human behavior) can be explained by evolutionary-biological considerations. This kind of universalism of biological evolution is, however, implausible and is not supported by the facts. Papineau gives no indication whatsoever how beliefs that appear to hinder the dissemination of genes could, in fact, support it in the end. This means that Papineau's Darwinistic imperialism is essentially a mere postulate (and a rather adventurous one, at that). Thus the teleological model does not contribute to an explanation of the problem of psychosemantics. § 5 The causal or correlational theory represents a third approach to psychosemantics. According to this theory, a mental state acquires 74 75 76

Papineau 1987, 66. Cf. Papineau 1987, 66-67. Cf. Papineau 1987, 6 8 - 6 9 .

232

Psychology and Consciousness

content by being correlated in a special way to an external state of affairs. More specifically, the content of a belief is the class of external causes that produce this belief under optimal conditions; it is determined by a covariation between this content and an external state. 77 The correlational theory has been advocated in two variants. The first is Fodor's psychophysical version, the second is the information-theoretic account that was primarily developed by Dretske. We discuss Fodor's version in this section and address Dretske's account in the next one. Fodor's psychophysical brand of the causal theory is the psychological version of referential semantics, which identifies the meaning of terms with the reference or extension of these terms. Psychosemantics is not interested in the meaning of terms in general, but in the more limited problem of interpreting the content of mental states. Put more precisely, this means that the content of the belief "it is red there" (i. e., not necessarily of the term "red" itself) should be defined such that single cases of this belief are instantiated in a person A, if A observes a red object under optimal conditions. By this recourse to the reference of a belief, the content of this belief should be able to be reconstructed in non-semantic terms. This is not the place to analyze Fodor's carefully constructed and brilliantly defended theory of psychosemantics in detail. For our purpose, it suffices to sketch two of its main features. On the one hand, its nonsemantic specification of the optimal conditions and, on the other hand, the notorious problem of all referential semantics, namely, the differences in meaning of co-referential terms. The necessity of demanding optimal conditions in the causal relation between external circumstances and the corresponding belief derives from the possibility of erroneous beliefs. One could, in fact, look at a cat and believe that one was seeing a dog — due to unfavorable circumstances (such as excessive distances, inadequate light, etc.). The problem lies in characterizing these optimal conditions in a way that doesn't in turn make use of the conceptual system of the observer or of his other beliefs. Fodor's claim is that there is a special class of beliefs (namely, beliefs in the presence of simple sensory qualities) for which this is, indeed, possible. Beliefs in the qualities of colors, for example,

77

Cf. McLaughlin 1987, 272.

Psychosemantics and Folk Psychology

233

force themselves upon an observer independently of how he is otherwise cognitively equipped: For example: Paint the wall red, turn the lights up, point your face toward the wall, and open your eyes. The thought "red there" will occur to you; just try it and see if it doesn't.78 T h e optimal conditions f o r such beliefs can be completely characterized by physical terms. One need only describe an observer with open eyes looking at a wall with sufficient paint and sufficient light, and one already has a characterization of the external circumstances that are sufficient for the corresponding belief to result. T h e precise description of the optimal circumstances can be based on psychophysics, which explains precisely how large a retinal image must be, what the optimal distance is f o r seeing, etc. 7 9 Such a direct definition is, however, only possible with terms for designating direct perceptions. It no longer works with beliefs that such properties as "is a horse" or "is a proton" are given. Psychophysics cannot guarantee that a horse will be seen as a horse; for that, one would have to have the concept "horse," and this stipulation is of a semantic nature. 8 0 Fodor suggests the following solution to this difficulty. For all beliefs that do not involve immediate sensory perceptions of the object, a connection with sensory perceptions is established by other theories. Physics, for example, specifies the conditions under which the presence of protons leads to certain perceptible indications (such as the density of photographic plates, trails in a cloud chamber, etc.), and these indications are, in turn, accessible to psychophysics. Psychophysics thus establishes a connection between the sensory indications and the corresponding beliefs. Hence, the belief "there is a proton" is indirectly connected to a state in the external world by means of sensory qualities as well as by the laws of psychophysics and physics. And this external state is precisely the content of the corresponding belief: physics — including the physics of the experimental environment — guarantees a reliable causal covariation between instantiations of proton and the psychophysical properties of the photographic plate, or the cloud chamber, or the voltmeter, or whatever apparatus you're 78 79 80

Fodor 1987, 112. Cf. Fodor 1987, 1 1 2 - 1 1 3 . Cf. Fodor 1987, 1 1 6 - 1 1 7 .

234

Psychology and Consciousness

using to detect protons (if physics didn't guarantee this correlation, what you're using wouldn't count as a proton detector). And psychophysics guarantees a reliable causal covariation between the observable properties of the apparatus and the tokening of concepts in the belief box (if psychophysics didn't guarantee this correlation, these properties of the apparatus wouldn't count as observable). 81

It is clear, on the other hand, that the causal chain from the external state to the corresponding belief is mediated by the knowledge of the corresponding physical theory; we have to know that certain trails in cloud chambers reveal the presence of protons. The mechanisms that connect the external state to the corresponding belief are, therefore, of an intentional nature. Fodor asserts, however, that this fact is irrelevant to the problem of psychosemantics. For the question as to what a given belief represents, it is only essential that a reliable connection exists between this belief and an external state; it is not essential how this connection was brought about. Psychosemantics merely requires that the causal control should actually obtain, however it is mediated. The claim, to put it roughly but relatively intuitively, is that it's sufficient for "proton" to express proton if there's a reliable correlation between protons and "protons", effected by a mechanism whose response is specific to psychophysical traces for which protons are in fact causally responsible. And that claim can be made in nonintentional, nonsemantic vocabulary. It just was. 8 2

This means that someone has the concept of proton at (or has the belief "there is a proton"), if his concept of proton is connected to physical indicators in the proper way; beyond that, he can think whatever he likes about protons. Thus, someone might take protons to be alive; as long as he maintains that protons are detected by the usual cloud chamber trails, he has the correct concept of the proton. 8 3 The observable features that are assigned to a belief determine its content. To begin our discussion of Fodor's theory, it should be noted that it could not provide a complete explanation of the content of mental states, even if it were completely correct. A referential psychosemantics works, at best, for beliefs — and Fodor does limit his discussion to beliefs - but it fails to work in the case of goals. After all, goals express

81 82 83

F o d o r 1987, 119 (emphasis in the original). F o d o r 1987, 121 (emphasis in the original). Cf. F o d o r 1987, 1 2 4 - 1 2 5 .

Psychosemantics and Folk Psychology

235

(yet) unrealized states, to which no objects in the real world correspond. 84 But even when limited to beliefs, Fodor's approach is not sufficient to determine content. First, Fodor's model of a perceptual basis is inadequate; second, the content of beliefs cannot be traced back to this perceptual basis with sufficient precision. McLaughlin has already pointed out that sensory states and sensory experiences, too, can be influenced by beliefs. Namely, if one looks at a red wall and believes that the wall is being lit with a red light, then one will also believe that the wall is white. 85 But this demonstrates that the optimal conditions specified by psychophysics are in no way sufficient to establish that someone will have a certain belief. It must furthermore be assumed that no other beliefs keep one from having the belief in question; and, once again, this requirement clearly appeals to the content of beliefs. In addition, someone who has the conceptualized belief "it is red there" must at least know that red is a color (although someone can have the sensory impression "red" without having the corresponding belief). 86 To this should be added that shades of colors are understood conceptually in different ways in different cultures. Hence, two different shades might lead one person to belief states of the same content, but lead another person to different belief states. Such differences in the individuation of the belief contents are clearly beyond the scope of any psychophysical characterization. Moreover, someone looking at a red wall, which is optimally lit by psychophysical criteria, may not come to believe "it is red there," because he is distracted or confused. Hence, the characterization of what is optimal must also contain psychological conditions. The result of this discussion, then, is that no level exists on which the content of beliefs could be determined by taking only external states into account or by considering the "given" exclusively. The content of beliefs is always dependent on the context of other beliefs. 87 Even if one accepts Fodor's idea of basic sensory perceptions, his model still results in an inadequate individuation of the content of beliefs. Fodor's referential psychosemantics implies that empirically equivalent mental terms (and, correspondingly, beliefs with the same observable 84 85 86 87

Cf. Pylyshyn 1984, 48. Cf. McLaughlin 1987, 280. Cf. McLaughlin 1987, 281. This result in turn supports our analysis of the problem of the empirical basis of psychology in VII.1.

236

Psychology and Consciousness

counterparts) are semantically equivalent in any case. Apparently, Fodor explicitly sanctions this consequence, as shown by his example of the belief in living protons. Such exclusive recourse to the observational content of a belief is, however, inadequate. If we assume that a medieval sorcerer and a modern psychiatrist identify a particular illness, e. g., epilepsy, using the same observable symptoms, such as loss of consciousness, muscle cramps and foaming at the mouth. The sorcerer would attribute this illness to possession by, say, the demon Epilepsia, while the psychiatrist would assume improper neuronal function. Despite the supposed empirical equivalence of both terms for the illness, it seems extremely implausible to assign the same content to the belief "this is a case of epilepsy" in both cases. The conclusion is the same, if one considers the following (slightly idealized) case from the history of science: the relation of the terms "oxygen" and "dephlogisticated air" in the early stages of the Chemical Revolution. The empirical tests (i. e., the perceptual indicators) for the use of both terms were equivalent in this case (e. g., causing candles to flare brightly, production of this gas by heating mercuric oxide/mercury calx). If we imagine that Priestley and Lavoisier, let's say in 1776, stood in front of a container of gas produced by mercury calx/mercuric oxide and said "there is dephlogisticated air" and "there is oxygen," respectively, on the basis of Fodor's model, we would have to ascribe the same content to their beliefs. In view of the different kinds of theoretical ideas that are related to their respective statements, however, such an identification of content would be inappropriate. While Priestley's belief refers to air from which the phlogiston (the carrier of the property of inflammability, which escapes from the burning object as it burns) has been removed, Lavoisier means the elementary gas "oxygen" which combines with the burning object and thereby gives off heat. Oxygen as an elementary gas no more exists for Priestley than phlogiston does for Lavoisier. To attribute nevertheless the same content to their beliefs would hardly do justice to the significant differences in their views. It becomes even clearer that Fodor's causal theory of the content of beliefs is inadequate, if one considers the later development of Lavoisier's theory of oxygen. Lavoisier supplements his original conception with a theory of acidity that regards oxygen as the cause of acid properties. In doing so, he introduced a new empirical indicator of a substance's oxygen content, namely, its acidity. Hence, Lavoisier's early and late

Psychosemantics and Folk Psychology

237

concepts of oxygen no longer share the same perceptual indicators, and, therefore, are no longer semantically equivalent according to Fodor's model. But to locate the gap between pre- and post-revolutionary semantics in the formulation of a theory of acids, rather than in the theory of oxygen, appears to be the least plausible of all imaginable alternatives. All this leads to the conclusion that Fodor's causal psychosemantics cannot adequately reconstruct the individuation of the content of beliefs. § 6 Dretske's information-theoretic account of psychosemantics starts with the same basic assumption as Fodor's psychophysical model, namely, that all representation is correlation. The key notion here is that the specific aspects of representations can be clarified by having recourse to the concept of a nomological relation between two physical systems. More concretely, representations are based on the existence of a lawful dependence between the states of two systems Si and S2. Suppose that there is a causal connection between these two states and that the state of S2 is deterministically and nomologically linked to the state of Si; the state of S2 is caused by the state of Si. Suppose furthermore that there exists no causal intervention from a third system, i. e., the relation between Si and S2 is undistorted. Then we can infer with certainty from a given state of S2 that the corresponding state of Si is also present. This is to be expressed such that in this case the state of S2 reliably gives information about the state of Si. Because the latter is the cause of the former there is a constant connection between them, and this enables the state of S2 to refer, or to stand in the relation of "aboutness," to the state of Si. The former indicates the presence of its cause. This relation of reference or aboutness, however, constitutes the most basic feature of intentionality. This model is used in the first place for elucidating the nature of perceptual knowledge, i. e., of true beliefs about observational traits of the external world. In that case the knowledge is caused by a sensory input which in turn indicates the presence of an observable state of affairs. Just as in Fodor's version of the causal theory, it is concluded that a mental state can represent a state of affairs by virtue of being caused by (and therefore correlated to) that state of affairs. Reference is not sufficient for representation or intentionality. Whereas all representation is correlation, the reverse does not hold. The first thing to be noted is that mental representations or intentional states are intensional, i. e., non-extensional. Intensionality means that

238

Psychology and Consciousness

coextensional expressions may possibly not be substituted for one another salva veritate in a sentence. T h a t is, replacing an expression in a sentence by a coextensional one may affect the truth value of that sentence. This is typically the case for mental states characterized by propositional attitudes. If somebody knows that a, and a is extensionally equivalent to b, it is not generally true that she also knows that b. While someone may know that the man down at the corner is upset she may not know that Fred Dretske is upset although the two (let's suppose) are in fact identical (cf. II.5.§ 3, III.1.§ 3). This peculiarity of mental states can be accomodated in the information-theoretic account by making use of the condition that the correlation between the two states be lawful. This means that the correlation must be based on a true, lawlike connection; mere accidental generalizations won't do. 8 8 This implies that property G of system S2 may represent or give information about property F belonging to Si, whereas G does not represent the property H of system S3 if F and H are accidentally coextensional. That is, G is caused by F but not by H. For this reason G carries information about F but not about H — although F and H always happen to occur together. To make things more perspicuous think of the following example. Let's assume that all male members of the church choir of St. Ambrosius happen to be bald and are also stricken with hoarseness and that no other male in the community under consideration exhibits these regrettable properties. In such a case learning about a man's hoarseness also carries information about his choir membership while learning about his baldness does not. The reason is that the former connection, but not the latter, is based on a causal law and is thus lawlike in character. The relevant causal law is that frequent rehearsals strain the voice and lead to hoarseness. Adding the boundary condition that the church choir of St. Ambrosius — due to the limited skills of its members — practices frequently, allows us to derive the prediction that all singers (and especially the male ones) are in fact hoarse. N o such derivability exists in the case of baldness. Although (by hypothesis) all male singers and only the male singers are bald, this property is only coextensional to membership and not related to it in a lawlike fashion. Hence, it provides no information about the membership. This trait makes is clear that information has the same intensional nature as intentional states; there 88

On the distinction between lawlike regularities and accidental generalizations cf. II.5.§ 3.

Psychosemantics and Folk Psychology

239

is no substitutivity of coextensionals in either case. Statements describing information are thus intensional sentences and informational content indeed exhibits intentional properties. It is obvious that this claim can't yet be the end of the story. For on that basis intentionality is no longer a peculiar trait of the mental but rather pervades the whole of nature. After all, it follows that thermometers, galvanometers and the like may possess intentional states. They are connected in a lawful fashion to some outside state of affairs and thus represent it. As can be gathered easily, the intensional characteristics are also present in that case. If a thermometer is immersed in a bath tub it represents the temperature of that tub and not the temperature of a tub placed next to it, though as a matter of fact the two may coincide. Hence, if no further qualifications are introduced one might conclude from Dretske's model that the mind is all-pervasive or that the world is made of mind-stuff. Dretske attempts to avoid panpsychism by adding two further conditions an intentional system has to meet in order to qualify as a truly representational or cognitive one. Such a system must have the capacity of (1) aspect separation (our term) and (2) "digitalization." The first condition is based on the strategy of accepting outright the seemingly unwelcome conclusion that non-cognitive states can possess content. They can indeed possess content. The catch is, however, that they possess too much content. That is, non-cognitive, intentional states such as the readings of a measuring instrument don't represent a single external state but rather a whole bunch of such states that are all nomologically related to its pointer readings. Consider, for example, a galvanometer measuring current intensity. In addition to representing current intensity, the gadget cannot help representing voltage states (by virtue of Ohm's law). Moreover, when attached to a photomultiplier placed at the end of a telescope, the registered current intensity pattern may indicate an astronomical object in outer space, whereas when attached to an ionization chamber — practical awkwardness aside — it may record the presence of an alpha-particle. The point is that in all these circumstances the galvanometer does not represent precisely one external state; it represents a number of states which appear to be of a quite different nature from a cognitive point of view. One and the same reading of the instrument contains information about a large number of intuitively distinct physical quantities. Accordingly, the pointer state has no definite content; it carries no unique piece of information.

240

Psychology and Consciousness

This description of the malaise suggests the following remedy. Cognitive systems are set apart from non-cognitive, intentional systems by their capacity to distiguish the different pieces of information contained in a given signal. In contrast to intentional systems in general, cognitive systems may represent the same token in different fashions; it may mean different things in different contexts. That is, a cognitive system can respond to one and the same physical state in a variety of ways. Whereas in nature a given state of current intensity always goes along with a voltage state, a cognitive system can represent one of them without at the same time representing the other. That's what's distinctive about cognitive systems: they can unpack the information batch. The second characteristic of cognitive systems is their capacity of digitalization. That is, their capacity of digesting, as it were, the incoming signal (i. e., the sensory experience) and transforming it into a conceptualized belief. A cognitive system can convert experience to knowledge. More specifically, digitalized representation means that not all states that are different in the system Si are represented as different states in the system S2. The representing system S2 collects distinct states of Si into equivalence classes; i. e., it represents them as the same states. Take the cognitive representation of a certain flower as an example. In the cognitive representation of, say, a rose as a rose, a vast amount of information is disregarded. We abstract from the blossom's color, the flower's length, and from lots of other accidental features like the presence of some spots on the leaves. That is, the peculiarities of a token's sensory image are systematically neglected; a whole class of physically distinct signals is subsumed under the common label "rose." This means on the whole that, in Dretske's view, it is these two features, namely aspect separation (i. e., sorting out components of information) and digitalization (i. e., stripping off such components), that are constitutive of cognitive systems.89 There has been a lot of telling criticism of Dretske's model, and we shall spare ourselves the repetition of these objections. We just want briefly to call attention to four problems involved here and afterwards spell out our own doubts about that approach. The first difficulty we would like to mention is the claim inherent in Dretske's model that there 89

This reconstruction of Dretske's position is based on Dretske 1980; Dretske 1983. A brief sketch of it is given in Bechtel 1988, 65 — 67. A somewhat similar approach is pursued in Sayre 1987, 260 — 266.

Psychosemantics and Folk Psychology

241

exists a level of pure perceptual knowledge that is uninfluenced by what an observer already knows. 9 0 This proposal obviously flies in the face of the theory-ladenness of perception that we stressed in VII. 1. Second, Drestke's treatment of false beliefs is generally considered to be unconvincing. Dretske regards false beliefs as being produced by wrong token typing. T h a t is, a given signal is put into an improper conceptual basket (e. g., a rose is conceptualized as a violet). False beliefs are thus the result of a misapplied digitalization procedure. This model has aroused various misgivings that we don't wish to rehearse. 91 Third, it is not clear to us how to accomodate the capacity of mental states to refer to non-existent objects in the framework of Dretske's model. H o w are we to account for the belief that Pegasus has wings in terms of some sensory input? In such cases there is obviously no physical system whose causal influence might give rise to a corresponding belief state. And it is precisely this possible reference of mental states to nonexistent objects that Brentano considered to be a central aspect of intentionality ("intentional inexistence" (cf. II.5.§ 1)). A viable theory of intentionality has somehow to come to terms with this important peculiarity. Dretske, however, remains silent on that problem. True, the information-theoretic model addresses primarily perceptual beliefs. It is clear, on the other hand, that the model must offer some guidelines for a generalization to belief states in general if it is to qualify as an acceptable account of psychosemantics. Dretske might perhaps resort to Hume's doctrine of "multiple, partial causation" in order to solve the problem of intentional inexistence. H u m e tried to reconcile his conception that all human ideas are caused by sensory impressions with the capacity of the mind to imagine objects it has never experienced by the following construction: In thinking of such an object, e. g., Pegasus, the mind combines some simple ideas, i. e., horse and wings, each of which can be traced back to sensory impressions. 92 Applying this construction to the problem of intentional inexistence leads us to the interpretation that the notion of a nonexistent object is generated by joining together several causal influences of existing objects. Each such influence gives rise to a part or aspect of

90

91 92

In fact, Dretske takes preliminary knowledge into account but tries to get rid of it in the end by tracing it back to perceptual knowledge; cf. Dretske 1983, 5 7 - 5 8 . Cf. the "Peer Commentary" in Dretske 1980, 63 - 83 and Bechtel 1988, 66. Cf. H u m e 1748, 2 7 - 2 8 .

242

Psychology and Consciousness

the resulting notion. It is obvious, however, that this approach transcends the framework of Dretske's model. For it is impossible to describe this mechanism of multiple, partial causation in terms of Dretske's two cognitive procedures, namely, aspect separation and digitalization. Adopting the Humean doctrine would thus require in any event a considerable alteration of Dretske's model. Fourth, it is doubtful whether Dretske's causal explanation of the non-substitutivity of coextensionals really hits the mark. Dretske's explanation relies on the distinction between causal and accidental relations. But it is hard to see how this distinction is of any help in the following standard example of intentional non-substitutivity: John knows that aspirin relieves headache. Aspirin is acetylsalicylic acid. But John does not know that acetylsalicylic acid relieves headache. After all, aspirin is as causally efficient in relieving headache as is acetylsalicylic acid. The causal relation is of exactly the same strength and quality in either case. Accordingly, Dretske does not offer a general explanation of intentional non-substitutivity. Our central difficulty with Dretske's approach concerns the nature of the principles that guide the procedures of aspect separation and digitalization. At the start, the approach of interpreting representation as correlation is straightforward and promising, but second thoughts arise when these auxiliary procedures are looked at more closely. The need to introduce such auxiliaries stems from the fact that physical natural kinds do not at all coincide with mental natural kinds. Signals that are physically alike may belong to distinct cognitive types, they may indicate or mean different features; and, conversely, physically different signals may be subsumed under the same cognitive label. This crossclassification or many-many relation between physical natural kinds and cognitive natural kinds is the basis of the explanatory value of cognitive kinds; it is the basis of the capacity of propositional attitudes to capture generalizations that cannot be captured in purely physical terms (cf. II.5.§ 2; VII.2.§ 2). Dretske is, of course, aware of this peculiarity; this insight underlies his introduction of the two auxiliary procedures. Our complaint is, however, that these procedures don't work in accordance with the fundamentals of his project to specify a naturalized psychosemantics. What both procedures are essentially supposed to accomplish is a reshuffling of the incoming physical signals according to their cognitive significance. Signals that are physically but not cognitively alike are

Psychosemantics and Folk Psychology

243

dissociated and put into their appropriate cognitive places; they are assigned to their pertinent cognitive equivalence classes. The same holds analogously for physically distinct signals. The question is now: H o w is this feat to be performed? H o w are cognitive equivalence classes to be extracted from the physical ones? Apparently, there is only one answer available: by having recourse to meaning. After all, the whole reshuffling procedure comes down to distinguishing between relevant and irrelevant aspects and accordingly latches on to the content of the incoming information. If the semantic foundation of the reshuffling is not taken into account, the whole procedure has the air of an arbitrary combinatorial exercise. A viable theory of psychosemantics has to specify in nonsemantic and non-intentional terms what separates physically similar signals and what ties together physically dissimilar ones. And since Dretske gives no non-semantic criterion for cognitive type identity, his whole project of formulating a naturalized psychosemantics fails. § 7 This discussion of some psychosemantic theories which are representative of approaches in this field allows for the conclusion that the program of naturalizing psychosemantics cannot be successfully implemented at present. There exists no worked-out method for adequately ascribing content to mental states by exclusive recourse to nonintentional and non-semantic indicators. The interaction theory and the causal-correlational theory mark in a sense two extremal, or even contrary, approaches to tackling the problem of psychosemantics. Interaction theory regards the inferential profile of mental states as the criterion for mental type identity and accordingly stresses the importance of the cognitive architecture for content ascription. By contrast, causal-correlational theory disregards the inferential role and instead places the connection of a belief system to observable traits of the external world at center stage. It is noteworthy that both approaches — despite their different structures and orientations — stumble into the same pitfall: They fail to individuate correctly the contents of beliefs. They don't succeed in distilling cognitive states exhibiting the same content from physical states by purely physical (i. e., non-intentional and non-semantic) means. It is this peculiarity that makes Searle's argument (cf. VII.2.§ 5) nontrivial. It is a general trait of formal structures that they do not possess content in and by themselves and that they need some external source of content. What is special about the mental is, however, that content

244

Psychology and Consciousness

cannot be ascribed to cognitive states according to the principles that govern content ascription in other branches of the sciences. After all, the interaction theory is a descendant of the theoretical context account of meaning (as described in VI.1.§2 — 4), and the correlational theory is a relative of the verificationist theory of meaning (as sketched in I.l.§ 3). Whereas both approaches can be successfully employed to supply a theory with meaning, they fare worse when used for attributing meaning to a network of beliefs. Therefore, the point in Searle's argument is not that a theory of mental operations must be supplemented with a theory of mental content. The point rather is that the usual methods for ascribing content are insufficient for ascribing content to mental states. This shows that Brentano and Chisholm rightly attribute a special place to intentional predicates. It also shows that Stich was right in asserting that the functional definition of mental states involves special difficulties (even if Stich's reasons for his assertion are unconvincing). Nevertheless, it would be inappropriate to conclude from this fact (as Stich does) that we must give up an intentional description of psychological states. After all, the discussion in VII.2.§ 2 made it clear that such intentional descriptions succeed empirically, and it wouldn't be proper to reject an empirically successful approach, because it can't be reduced to other theoretical approaches. 9 3 Intentional interpretation of mental states is emergent relative to the semantic interpretation of both nonintentional psychological terms as well as of physical terms. But this fact should not be considered a sufficient reason for abandoning intentional speech. § 8 But if the program of a naturalistic psychosemantics fails, and the contents of mental states are not accessible solely by recourse to non-intentional descriptions, then how can we gain empirical and theoretical access to these contents at all? H o w do we determine the content of mental states? In our view, intuitive folk psychology plays an essential role here, and, in this sense, folk psychology functions as an observation theory. With its help, we attribute to others mental states of a definite content, whose interactions and further properties can then be explained by recourse to the states of a theoretical psychology. In Bandura's experiment with learning by imitation, for example, even 93

Cf. our discussion of physicalism in VI.4.§ 4.

Psychosemantics and Folk Psychology

245

those children who did not show imitative behavior spontaneously did so after being explicitly told to imitate. From this, it was concluded that these children, too, had learned the corresponding behavior before (cf. V.2.§ 2). Clearly, such an inference is based on the following kind of folk psychological law: (A shows a behavior a on demand) -*• (A knows a) —• (A learned a) That is to say, that a learning process had taken place is established by means of intuitive folk psychology; the task of scientific psychology is to explain the preconditions and peculiarities of such a learning process (as this is done in the framework of social learning theory). Folk psychological laws often go unnoticed in this context, since they seem to be trivial and obvious. This impression is, however, deceptive; this is shown by the fact that they can be wrong (as we emphasized). While, in this case, the existence of a propositional attitude (A learned that a) is directly inferred from the application of folk psychological laws to overt behavior, such attitudes are usually determined less directly. In the case of Oedipus, for example, we say that the content of Oedipus' belief system has been changed by the information about the true identity of Jocasta, because Oedipus' subsequent self-destructive behavior would otherwise be incomprehensible from the perspective of intuitive folk psychology. The latter assumes that the goal of protecting one's own body from damage is only abandoned in the most extreme state of despair; hence, the content of Oedipus' beliefs must be reconstructed in a way that makes this despair plausible. The same is true of the example given in VII.2.§ 2 of a person who witnesses an accident. In order to make this person's helpful behavior understandable from a folk psychological perspective, one must attribute to him, among other things, the belief that another person was just injured and needs help. On the other hand, two beliefs have the same content, if possible differences have no effect on behavior (from a folk psychological perspective). The example given in VII.3.§ 2 is such a case, in which the belief "all men are mortal" is connected with the name "Socrates" in person A, while person B lacks this connection. The content of both beliefs is the same in this case, because the difference is irrelevant to a folk psychological description of the behavior of A and B. This leads us to the general conjecture that the content of a belief can be defined as an equivalence class of the actual causal interactions

246

Psychology and Consciousness

of this belief with its surrounding cognitive field. Though a person's actually realized interactions within his cognitive system, as well as with the outside world, fix the content of this person's mental states, differences in the causal profile determined in this way are not necessarily reflected in differences in content. This is why an interaction theory of psychosemantics fails. The actual causal profile does not give the correct types of states for the content of beliefs; certain actual connections are irrelevant to an identification of the content. Intuitive folk psychology provides just this distinction between relevant and irrelevant connections; or to put it differently: folk psychology combines several actual causal profiles of a singular mental phenomenon into an equivalence class. And this equivalence class can be reasonably identified with the type of the corresponding belief, i. e., with its content. This shows, first, how intuitive folk psychology identifies the content of mental states, and, second, makes it clear why this approach does not contribute to a naturalistic psychosemantics. The conditions for the equivalence of causal profiles are themselves framed in intentional terms. After all, to put it bluntly, this approach aims at determining the content of mental states in such a way that they harmonize with a person's other mental states, as well as his interactions with the outside world. It is apparent that this procedure is a refinement of the method discussed in III.3.§ 7 for interpreting language and behavior. Mental content is specified in such a way that it forms the most coherent system possible in light of folk-psychological laws. In the case of Oedipus, attributing the state of despair to Oedipus establishes a reasonable link between the knowledge of Jocasta's identity and Oedipus' selfmutilation. To attribute to him a state of joy upon rediscovering his mother would create an incoherence in Oedipus' belief system. His subsequent actions would be incomprehensible (in light of principle (1) (VII.4.§ 2) and the folk-psychological assumption of the high value of protecting one's own body). In this way, the application of folk psychological laws creates the empirical basis with which the theoretical explanation can operate. Consequently, the proper place for intuitive folk-psychological laws is the explanandum (that which is to be explained) and not the explanans (that which explains). Intuitive folk psychology is an observation theory of scientific psychology, i. e., folk-psychological states are indicators for theoretical states (even if they are sometimes vague or flawed); the former only receive theoretical import and explanatory power by being

Psychosemantics and Folk Psychology

247

connected to the latter. Because of their role as indicators, folk-psychological states should not be thought of as epiphenomena. True, they themselves do not enter into the causal mechanisms, but they at least indicate part of the causally effective processes. In order for a theory to be considered an adequate observation theory, two conditions are necessary, even if both together are not sufficient. First, the terms of the observation theory T o must correspond to the terms of the explanatory theory TE in such a way that all the empirical indicators assigned to the terms of T o also appear among the indicators of TE, but not the other way around. Let's look, for example, at the case discussed in VII.3.§ 8 of dividing headaches into migraine and tension headaches. The only observational indicator for applying the everyday expression "headache" is the appearance of the corresponding pain, or of its behavioral manifestations. This indicator is likewise associated with both of the medical terms, though the latter are also connected to indicators like sensitivity to light and numbness of the limbs. Clearly, the full observational content of the everyday expression appears in the scientific terms; the difference is simply that the latter are further specified by additional indicators. In such a case, it seems reasonable to view the term "headache" as part of the observation language and the two psychological terms as its theoretical differentiation. Headache is an indicator for both migraine and tension headaches. On the other hand, the term "unconscious belief," for example, lacks an essential intuitive folk-psychological observational indicator, namely, accessibility to introspection. Hence, this term cannot be understood as a part of folk psychology, if the latter is to be considered an observation theory. The headache example makes it clear that the folk-psychological indicators can be connected to theoretical states in an unspecific way. One and the same empirical event apparently indicates two different theoretical states here. Such an imprecise connection nonetheless does not preclude the possibility that folk-psychological states could be precisely explained by recourse to the theoretical states. Such an explanation does not require an inference from indicator to state, but rather from state to indicator; and ambiguity in the one direction is, indeed, compatible with an unequivocal relation in the other direction. If one limited oneself to the demand that such a proper-subclass relation exist among the empirical indicators, this would make it possible to prevent grossly inadequate theories from being rejected by interpreting

248

Psychology and Consciousness

them as observation theories. Phlogiston chemistry, for example, considers sulfuric acid an elementary substance that turns into the various sulfurous oxides (to use the modern name) by increasing the supply of phlogiston. Sulfur is considered the maximum degree of phlogistication of sulfuric acid. This means that the terms "sulfur oxidized to a certain degree" and "sulfuric acid phlogisticated to a certain degree" may designate the same empirical substance and, hence, have the same observational indicators. This fact could be used to transform phlogiston chemistry into an observation theory with oxygen chemistry as the pertinent explanatory theory. But it is clearly inadequate to first describe chemical phenomena in phlogistic language, only to explain them subsequently in terms of modern chemistry. Therefore, we must also demand of an observation theory that it satisfy certain requirements for its empirical and methodological quality. This second condition would rule out phlogiston chemistry as an observation theory, while folk psychology doubtlessly satisfies this requirement in light of the current state of research. From this, it is also clear that the conditions for an adequate observation theory are weaker then those appropriate to a reduction. Reducibility requires that the corresponding terms can be defined, whereas all that is necessary here is to assign the observational indicators of the corresponding terms. Possible differences in the causal profile are not significant. This discussion shows that it is possible to use folk psychology as an observation theory; it does not show that its use is compelling. It is possible that observation theories will be advanced whose terms provide more accurate indicators for theoretical states. Precisely for this reason, the satisfaction of both conditions is not sufficient to consider folk psychology an adequate observation theory. Though, at present, there are no alternative observation theories; folk psychology has thus far had no serious competition in the field of describing mental facts. In addition, it can be assumed that folk psychology would be retained even if it had significant shortcomings. The reason is that folk psychology is of considerable value in non-descriptive contexts, i. e., it is used in interpreting practical human behavior. 94 Thus, folk psychology does not just serve theoretical purposes. This discussion permits us to conclude that folk psychology supplies us with (part of) the facts that are subsequently explained by scientific 94

For a discussion of this aspect, cf. IX.§ 5.

Psychosemantics and Folk Psychology

249

psychology. Folk psychology is, in fact, of dubious value as an explanatory theory, but as an observation theory it is still useful and significant. On the other hand, it should be pointed out that (in light of the discussion of the problem of the empirical basis of psychology in VII.l.§ 5) there is no guarantee that it will remain so in the foreseeable future, and that, moreover (in light of the discussion of Einstein-Feigl completeness in VI.3.§ 5), it is not even desirable that folk psychology play the role of an observation theory in the long run. § 9 The results of our discussion of the mind-mind relation can be summarized as follows: (1) the content of mental states is essential to any explanation of behavior. To that extent, folk psychology in the technical sense will remain an important part of scientific psychology. Nonetheless, (2) this content cannot be grasped naturalistically. In addition, (3) the reducibility of folk psychology in the intuitive sense to scientific psychology is doubtful. Nevertheless, (4) intuitive folk psychology serves as an observation theory for scientific psychology. These results have an important impact on the interpretation of the mind-body relation, which we shall discuss in conclusion. In the original form given to it by Schlick and Feigl, the identity thesis assumed an identification of immediately experienced psychological phenomena with physical processes (cf. II. Introduction, II.1.§ 4). This position is to be understood as the claim of a double identification. It requires not only the reducibility of psychology to neurophysiology, but also the identification of folk psychological states and events with the states and events of cognitive psychology. Our presentation, on the other hand, has made it clear that the latter relation (the mind-mind relation) is more likely to be understood as a concept-indicator relation than as identity. If this is correct, and intuitive folk psychology is not, therefore, reducible to cognitive psychology, then this means that one cannot reasonably speak here of a theoretical identification by way of an ontological reduction. To be sure, this does not preclude the possibility that there is an identity of reference in individual cases, i. e., that both approaches designate the same singular mental event in a particular application of a term. But this would be merely coincidental and comparable to the case discussed in II.5.§ 3, where the two predicates "is a lion" and "located within 3 km of the University of Constance" apply to the same entities.

250

Psychology and Consciousness

Hence, a token mind-mind identity is possible, but (unlike token mind-body identity) cannot reasonably be understood as a theoretical identification. 95 The reason for this difference is that the conceptual similarity of neurophysiology and cognitive psychology (as conceived by Fodor (cf. II.5.§4)) is greater than the conceptual affinity between cognitive psychology and folk psychology (in the form we take to be plausible). In the first case, the relationship of the corresponding theories can be reconstructed as a corrective reduction, but not in the second case. Thus, it is certain that the identity thesis, in the form which found its classical expression in the works of Schlick and Feigl, is quite problematic: Experienced states of consciousness can probably not be identified theoretically with physically specified states. This conclusion is based on arguments of which Schlick and Feigl were not even faintly aware, since they are the results of recent developments in cognitive psychology. And this shows once again that it would be unwise to anticipate the path of science beyond the visible horizon for the sake of a philosophical interpretation.

95

That such a coincidental token identity exists, is precisely Fodor's claim applied to the mind-body relationship. Our presentation, on the other hand, has shown that this self-assessment is inadequate (cf. II.5.§5).

Vili. Epistemic Limitations of the Mind-Body Relation Up to this point, the mind-body problem has been discussed as the problem of the actual nature of the mind-body relation. We were concerned with providing an adequate reconstruction and with showing the logical nature of the connections between mental states or events, brain processes and behavior. In the following, we shall address possible epistemological limitations, i. e., possible limits on our knowledge of the mind-body relation. That such limits actually exist is asserted by Gierer. Gierer claims that brain states and states of consciousness are identical, but that the details of the relation between them cannot in principle be settled. We shall first test this claim and then develop a different kind of argument for the possible existence of epistemic limitations.

1. Epistemic Limitations? Principles of Finitism and Godelian Incompleteness § 1 Gierer's ontological monism essentially rests on two arguments. First, every nerve cell can, as a first approximation, be thought of as a digital switch; and such a network of digital building blocks can, in fact, perform any logical operation that can be described formally. This suggests that all the brain's functions are based on the physical mechanism of information processing. 1 Second, Gierer relies on the principle of the causal closedness of the physical world: All results of the sciences point to the conclusion that the laws of physics completely cover the realm of the objectively observable; this applies to the human brain as well. 2

1 2

Cf. Gierer 1985, 1 8 4 - 1 8 8 . Gierer 1985, 225.

252

Epistemic Limitations of the Mind-Body Relation

The hope that physics is complete is the basis for ontological monism here. Gierer's central claim is that, although physical laws hold without restriction in the brain, it is not possible to describe psychological states by these laws: Although all evidence indicates that the state of consciousness has a definite relationship t o the state of the brain, the conclusion is n o t logically compelling, that the mental state of a person must be c o m p l e tely derivable from the physical state of his brain. 3

Rather, one supposes: that the mind-body relation c a n n o t be decoded with finite m e a n s in a finite world. Although physics applies without restriction to the brain, conscious experiencing c a n n o t be grasped completely by a purely objectivizing scientific analysis. 4

To justify this assumption, Gierer has recourse to (1) Godel's

in-

completeness theorem and (2) to a principle of finitism. According to Godel's theorem, formal systems of sufficient complexity contain true statements that cannot be proved by means of this system, i. e., they contain formally undecidable formulas. Such a system has more true consequences than can be derived. Also included among these undecidable statements is the assumption that the system is consistent; even if the system is, in fact, consistent, this cannot be proved using the means of the system. 5 The principle of finitism says that everything that cannot be decided in a limited number of steps is undecidable for us. If a question can only be answered by a process whose steps would require more time to carry out than the age of the universe would permit, then the question is unanswerable. Godel's incompleteness theorem leads Gierer to suppose that the brain cannot develop a complete picture of itself. In particular, one can expect to encounter undecidability in abstract statements about brain processes and statements about self-referential brain processes (such as about self-consciousness, for example). 6 Gierer seeks to support this general supposition with a concrete thought experiment. 7 Let's imagine a computer that is capable of imitat3 4 5 6 7

Gierer 1985, 15. Ibid. Cf. Gierer 1985, 41 - 42. An overview of Godel's discovery is given in Heijenoort 1967. Cf. Gierer 1985, 46, 237, 247. Gierer 1985, 2 4 1 - 2 4 8 .

Epistemic Limitations I

253

ing the brain, and whose internal structure is known. And then ask ourselves whether we could determine any disposition to behavior that is supposed to be anchored in the psychological structure of such a machine. It turns out that we quickly encounter the limits set by the principle of finitism. We could, indeed, decide whether any given disposition to behavior is present in any given circumstance. This is done by simply questioning the apparatus or by analyzing its physical state. But not every true disposition could be determined in this way: Behavioral dispositions refer to general characters ("If the economic situation improves, he will build a house") that hold for an unsurveyably large number of physically different consequences or environmental influences and physically different actions. [...] One would have to test all these possibilities by putting the appropriate sense impressions in all details into the machine in order to confirm the validity of the behavioral dispositions. 8

Even if this were supposed to be true of individual dispositions, it would not be possible for all dispositions imaginable. Since "the number of possibilities exceeds any finitistically testable amount." 9 Dispositions to behavior are encoded in the physical state of the machine (or the brain); and the thought experiment makes it clear that, in general, this code cannot be decoded by finite means. 1 0 It is apparent from this that the limits of derivability suggested by the incompleteness theorem do, in fact, result in restrictions on decoding the mind-body relation. 11 § 2 Combining an ontological monism with an epistemological thesis which denies reducibility is characteristic of Gierer's position. Gierer's ontology is conceived in a non-Quinean fashion: Our assumptions about the existence of things are not dependent on our best theories about these things. Hence, we will first investigate the viability of Gierer's argument for ontological monism. Gierer's first argument concerning the digital structure of the nervous system crucially depends on it being possible to describe all brain

8 9 10 11

Gierer 1985, 2 4 4 . Gierer 1985, 2 4 5 . Cf. Gierer 1985, 2 4 8 - 2 4 9 . Cf. Gierer 1985, 2 4 8 .

254

Epistemic Limitations of the Mind-Body Relation

processes in a formally exact manner. But this is far from being certain, as Gierer also points out: It remains open [ . . . ] whether really all interesting properties o f the h u m a n brain can be formalized. 1 2

But this calls into question the relevance of this argument to the mind-body problem. And, even if its relevance were given, the only consequence of this argument would be that the human brain could be entirely made up of a digital network. The gap between possibility and reality would have to be closed by other, independent arguments. But such arguments can hardly be expected, since numerous empirical indications suggest that the structure of neurophysiological processes deviates considerably from the structure of electronic processes in today's digital computers (cf. VII.2.§6). Hence, the digitalization argument cannot provide any convincing reasons to support an ontological monism. The second argument fares even worse. In the discussion of the mind-body problem, the completeness of physics is precisely what is at issue. And we are of the view that higher-level psychological processes cannot be described physically, i. e., that physics is not complete in this sense (cf. VI.3.§ 4, VI.4.§4). What Gierer assumes is precisely what is in question. But an ontological monism can hardly be grounded on mere assumptions. Therefore, we conclude that Gierer's ontological monism is not well-founded. Once Gierer's epistemological argument has been stripped of its premature attempt to tip the scales to monism, it has a strongly antiidentity bent. A reduction of psychology (i. e., the theory of behavioral dispositions) to the physical laws in the brain (or even in Gierer's supercomputer) is the only way to justify an identity theory (cf. II.3.§ 3 - 5 ) . If this route is blocked, then the identity theory cannot be justified. If there actually were epistemological limitations of the sort Gierer describes, then their existence would constitute an important argument in support of dualism. If for epistemological reasons we are forced to use mental terms to talk about our behavior, if an independent psychological terminology cannot be avoided, then this is one of the strongest indications imaginable for the adequacy of a dualistic conception. With this as a background, everything depends on the viability of Gierer's argument in support of such epistemological limits. 12

Gierer 1985, 191.

Epistemic Limitations I

255

§ 3 Gierer explains that no material structure (and that includes the brain) can contain a complete image of itself. After all, it would have to contain its own image, the image of that image and so on, ad infinitum, and this is impossible. That may be. But what follows from that is open to question. The argument merely says that a particular person A cannot have complete knowledge of his own, present state; he cannot be acquainted with all of his own, current mental states. But this is a very weak claim. Let's just see what would still be possible: (1) A can have complete knowledge of his psychological states at an earlier point in time. By referring to two different points in time, the self-referentiality vanishes and with it the essential reason for incompleteness in Gierer's argumentation. (2) Another person can know A's current psychological state. This also avoids the self-referentiality that Gierer emphasizes. (3) Finally, A can know his own, currently relevant mental states. It isn't necessary to have recourse to all the other psychological states, in order to give an explanation of one particular psychological state. When examining the structure of A's self-consciousness, it is probably safe to leave out all the information about the structure of A's olfactory center. Nothing in Gierer's argument excludes the possibility that one part of the brain can know another part of the same brain completely. These three possibilities permit substantially more than would be necessary to decode the mind-body relation. If they are taken as given, then one cannot speak of limits to knowledge in any relevant sense. Gierer is operating with an extraordinarily weak concept of incompleteness here. Such weak forms of incompleteness do not provide any serious obstacles to a reduction of psychology to neurophysiology, particularly not if the latter is conceived of in the moderated form expressed in II.3.§ 5. Thus, reference to Godelian incompleteness contributes little to a solution of the mind-body problem. The same is true of Gierer's computer analogy. In essence, this latter argument says that behavioral dispositions cannot be determined reliably. If this were really the case, then it would have consequences that reached far beyond the mind-body problem. After all, the argument amounts to the claim that a reduction of psychology to neurophysiology is impossible, because psychology is impossible. Behavioral laws cannot, in general, be established with finite

256

Epistemic L i m i t a t i o n s o f the M i n d - B o d y R e l a t i o n

means — that is what Gierer concludes from his thought experiment. This conclusion contains a fundamental attack on psychology as a science. In fact, Gierer's argument can easily be generalized even further. Let's imagine a computer in whose structure all true laws of nature are coded, and attempt to find out by testing individual cases what form the laws of nature take. The laws of nature are general rules that apply to an unmanageably large number of physically different entities. All these possibilities would have to be tested by entering the corresponding initial conditions into the machine and following the development of the physical states. Even if this could be done for individual laws of nature, it couldn't be done for all the possibilities imaginable. It would exceed the limits of any finite method. This argument is entirely analogous to that of Gierer and clearly proves that natural science is impossible. Hence, Gierer's limits to the decodability of the mind-body relation are only the limits to the decodability of nature. But one would rather suspect that these limits do not represent objective restrictions, but merely the limits of the methods that Gierer considers. This suspicion is confirmed when Gierer's arguments are inspected more closely. Gierer requires of a test of behavioral dispositions that it examine all of the individual cases that it comprises. This is, of course, impossible in all fields; the inductive risk can never be avoided. Gierer also requires an algorithm to judge truth claims. So he claims that deriving behavioral dispositions is impossible because they cannot be derived by an "automatic method." 13 But it is, of course, in all cases unrealistic to demand that the inductive risk be eliminated and that an algorithm be provided to decide which laws hold. So it is not surprising that the mind-body problem cannot be solved in this way. Gierer's conception of science essentially takes the following form: One constructs the logical space of all possibilities and seeks a method that can reliably locate the case that is actually realized in a finite number of steps. This idea is apparently modeled on combinatorial problems, as is also suggested by Gierer's frequent reference to codes of all kinds. In fact, Gierer describes the situation very well as it applies to code-breaking. First, a gradual approach to a solution is not possible in such cases. Either the code is broken or it isn't; it is, as it were, a »

Gierer 1 9 8 5 , 2 4 5 .

Epistemic Limitations II

257

case of all or nothing. Second, the task of breaking a new code is not made essentially easier by previous knowledge, i. e., by knowledge from other fields. Neither of these difficulties, as a rule, occur in scientific problems in the narrower sense. Of course, there can be approximate neurophysiological explanations of behavior (just as there are approximate explanations of the precession of equinoxes or of the dispersion of light). Furthermore, in most cases, previous knowledge allows us to eliminate a large number of the logical possibilities and makes the situation manageable. The error expressed in this combinatorial approach is that it takes the problem of deciphering unknown cuneiform characters as the model for the mind-body problem.

2. Epistemic Limitations? The Brain and the Theory of Self-Organization § 1 Although the details of Gierer's argument do not hold up and the epistemological limits that he points out cannot be justified, Gierer's point of view does prove to be fruitful in a certain way. This is meant in the following sense: Even if all processes of consciousness are, in fact, physical processes, then it may, indeed, be impossible for us to recognize this. Even if we assume that the identity thesis correctly describes the situation, a reduction of psychology to neurophysiology, which is the epistemic criterion for the appropriateness of mind-body identification, could be unfeasible. But, since a successful reduction is the only theoretical reason for adopting the identity thesis, it might not be possible to substantiate a monistic interpretation, even if it were true. To demonstrate the possibility of such epistemic limits, we draw on recent results of synergetics, i. e., the theory of self-organization, and develop arguments for the impossibility of a psychophysical reduction, using a plausible extrapolation from current developments in neurophysiology. Our argumentation proceeds from the phenomenon of deterministic chaos in dissipative systems. Dissipative systems are open, i. e., they exchange energy with their surroundings and are far from thermal equilibrium. When non-linear effects (i. e., effects that cannot be described using linear equations) occur, this can result in the formation of

258

Epistemic Limitations of the Mind-Body Relation

coherent structures. In chemistry, such non-linear effects are found in autocatalytic reactions (such as those of the form X + Y ->• 2X); 1 4 in physics, one encounters them in forced, damped oscillations or in convection currents, 15 in ecology, they appear in the so-called logistic map that represents the growth of a population in a closed area. 1 6 In all these cases, large non-equilibrium structures are formed whose regularities can be described by non-linear dynamics and thermodynamics. These kinds of processes are thought of today as the basis of selforganization. Thus, biological systems, too, are considered to be nonequilibrium structures whose form (unlike crystalline order) is maintained by constant exchanges of energy with their surroundings. It is generally assumed that the brain in particular can be described as a dissipative system. Another characteristic feature of such systems is: not only does order arise out of chaos in them, but they also display chaotic behavior themselves under certain conditions. This was already noted by Poincaré for non-linear, conservative systems, i. e., systems that do not exchange energy with their surroundings. Since the origin of such effects is more easily accessible in conservative systems than in dissipative systems, we will sketch Poincaré's discussion of these effects as a first step, then we will examine the behavior of dissipative systems. § 2 Poincaré observes that, even in a deterministic world, where every state is connected to every other by strict laws, the future state cannot always be derived from the present one and chance cannot be eliminated. This becomes clear, if one looks at a cone balancing on its tip and asks oneself, if, when and to which side it will fall. Clearly, its unstable balance is destroyed by the slightest distortion of its symmetry, e. g., by an imperceptible breeze. In this case, therefore, an insignificant cause has a considerable effect. The problem here is that we always know the initial conditions only approximately, i. e., we do not know the cone's position and the air's motions with absolute precision. Of course, if this approximate knowledge of the causes permits the description of the effect with the same precision, then one could (despite the uncertainty) speak of predictability. 14 15 16

Cf. Prigogine 1980, 9 5 - 9 6 . Cf. Schuster 1984, 8 - 9 . Cf. Schuster 1984, 3 1 - 3 2 .

Epistemic Limitations II

259

But that isn't the case in this example; a slight disturbance, a tiny inaccuracy in our knowledge of the initial conditions may have important consequences. This means that, although the situation is governed by strict laws, it is not possible to predict to which side the cone will fall: If we knew exactly the laws of nature and the situation of the universe at the initial moment, we could predict exactly the situation of that same universe at a succeeding moment. But, even if it were the case that the natural laws had lo nonger any secret for us, we could still only know the initial situation approximately. If that enabled us to predict the succeeding situation with the same approximation, that is all we require, and we should say that the phenomenon had been predicted, that it is governed by laws. But it is not always so; it may happen that small differences in the initial conditions produce very great ones in the final phenomena. A small error in the former will produce an enormous error in the latter. Prediction becomes impossible, and we have the fortuitous phenomenon. 17

In general, the state of a system can be expressed by a position in an abstract configuration space in which all its relevant parameters are represented. In configuration space, a system's state can be described completely. Thus, the state of mechanical systems is characterized by giving the components of positions and momentum of all the degrees of freedom. Consequently, such a state can be expressed formally by a position in position-momentum space, i. e., in phase space. Hence, a change in state would be understood as a trajectory in phase space (cf. VII.2.§ 5). In chemical systems, concentration space, whose axes represent the concentrations of the reacting substances, is chosen for practical reasons as the appropriate configuration space. In these terms, the effects that Poincaré pointed out can also be described in such a way that neighboring points in phase space may belong to entirely different trajectories in phase space. This means: States that are initially indistinguishable can develop differently over the course of time. Trajectories in phase space that start off close to one another will drift apart over a period of time. Of particular interest are those cases, in which there are divergent trajectories in every area around a point in phase space. In these systems, despite precise knowledge of the relevant laws and despite the availability of observational data with arbitrary, finite precision, no prediction of the system's evolution is possible. Only on the basis of infinitely accurate knowledge about the 17

Poincaré 1908, 68 (emphasis in the original).

260

Epistemic Limitations of the Mind-Body Relation

initial conditions, could a prediction be made. 1 8 A case in point is the many-body problem, i. e., the problem of calculating the temporal evolution of a system of more than two bodies that interact gravitationally with one another. In this case, a strict prediction of future developments is not possible. This result implies, incidently, that the stability of the solar system cannot be proved. § 3 The aspect introduced here using the example of conservative systems appears analogously in dissipative systems. The long-term behavior of these systems is described by the concept of attractors. Attractors are those states of the system toward which the system is evolving. If we imagine a bar, for example, that has different temperatures at either end, then the bar (as a consequence of the second law of thermodynamics) will with time assume the same temperature throughout. This can be expressed by saying that thermal equilibrium is an attractor for non-equilibrial states in isolated systems. 19 To put it more precisely, an attractor can be expressed as a bounded region in configuration space toward which all (sufficiently close) trajectories approach asymptotically in the course of the system's development in time. In the simplest case, the attractor takes the form of a fixed point; i. e., from whatever point in configuration space a system starts, it always moves toward a fixed point. Then it remains there in a stable equilibrium. Another form attractors take is the limit cycle: Whatever its initial position, the system moves toward a particular periodic trajectory and remains there. Limit cycles can be found in chemical systems, for example, where certain stable oscillations occur in the concentrations of the reacting substances. 20 In our context, the so-called strange attractors are of particular interest. These attractors are also bounded regions in phase space toward which all system trajectories move in the long run. But the trajectories are essentially dependent on the initial conditions in these cases. Points in the attractor, however close together they may be initially, will be macroscopically separated after a sufficiently long period of time. There is an exponentially growing divergence of neighboring trajectories. 2 1 The 18 19 20

21

Cf. Prigogine 1980, 4 3 - 4 4 . Cf. Prigogine 1980, 8. Cf. Prigogine 1980, 9 8 - 1 0 0 . On forms of attractors, cf. also E.N. Lorenz 1963, 1 3 2 - 1 3 3 ; Schuster 1984, 111; Crutchfield et al. 1986, 4 2 - 4 6 . For a precise explanation, cf. Schuster 1984, 93 — 94.

Epistemic Limitations II

261

first example of a strange attractor was the simplified mathematical model of Benard convection developed by the meteorologist Lorenz in 1963. Benard convection is the phenomenon of the formation of regular patterns of convection currents in a horizontal cross-section of a fluid in which a temperature gradient is present. Beyond a certain critical point in this temperature gradient, the heat flow begins to oscillate around its average value and, as the temperature gradient increases further, it finally exhibits chaotic oscillations. Though these phenomena cannot be completely described using the simplified Lorenz model, the characteristic chaotic behavior occurs there as well. Since these sorts of convection processes have a crucial effect on weather patterns, similar chaotic effects are to be expected there as well. 22 The phenomena described here lead to a self-induced noise of considerable amplitude that must be distinguished from random thermal noise (which is found as well), and that also has nothing to do with quantum-mechanical indeterminacy. 23 Rather, its origin can be traced back to the system's extraordinary sensitivity to perturbations, or to the fact that the system's temporal evolution is heavily dependent on the initial conditions. While in conservative systems like Poincare's cone this sort of sensitivity only occurs at single, special points, in dissipative systems, it exists within the entire domain of the strange attractor, i. e., in an extended area of phase space. In dissipative systems, therefore, chaotic behavior is not limited to singular and rarely occurring situations, but is found on a significant scale. This sensitivity to small fluctuations in the boundary conditions or in external control parameters is expressed as the exponentially growing divergence of neighboring trajectories in phase space. More precisely, this means that, in the region of the attractor, trajectories diverge within any given neighborhood of any given point. If two neighboring points in phase space are selected such that their trajectories remain close to one another, then there is always a third point between them whose trajectory differs widely from the other two. 2 4 For this reason, minute, microscopic fluctuations can determine a system's macroscopic state; the perturbations do not average out, instead they drive the system in

22

23 24

E.N. Lorenz 1963, 137, 141. Cf. Schuster 1984, 9. For an experimental and theoretical analysis of this phenomenon, cf. also Grossman 1983. Cf. Grossmann 1983, 139; Schuster 1984, 2. Cf. Hunt 1987, 131.

262

Epistemic Limitations of the Mind-Body Relation

a certain direction. With reference to meteorology (which was the original impulse for studying chaotic effects in dissipative systems), Lorenz spoke of the butterfly effect: Even the flapping of a butterfly's wings can crucially effect the convection currents in the earth's atmosphere and, hence, meteorological developments. 25 This leads to the problem that it is possible to know the exact equations of motion for a system, without being able to predict the evolution in time of this system. Although meteorological developments (as it is generally understood) can be completely described by thermodynamic equations, this is of little help. Because all observations are always finitely accurate, the future behavior cannot be predicted using these equations. Though weather can be predicted in the short run; the chaotic effects described here will still appear in the middle to long run. It is important to see that it is not the system itself that behaves chaotically; its development is, quite the contrary, strictly deterministic. The chaos exists only for us, not for the thing being studied; it results from the imprecision of our knowledge of the initial conditions. 26 But this means that there is an epistemological limit that occurs in the phenomenon of deterministic chaos. Although a system is, in fact, strictly deterministic and can be completely understood according to certain laws, we are not in a position to describe the behavior of this system, despite our precise knowledge of these laws. § 4 The essential point in the context of our discussion is that many systems, including biological ones in particular, exhibit this property of deterministic chaos. For example, the phenomenon is found in the electrical potentials of nerve membranes. 27 Chaotic behavior is also exhibited in individual neurons and (as a collective phenomenon) in networks of neurons. 28 Judging from current neurophysiological research, it can be assumed that deterministic chaos occurs in the brain. It is supposed, moreover, that strange attractors and deterministic chaos are important to the brain's ability to process information. 29 25

26

27 28 29

Cf. Schuster 1984, 2. Poincaré also applied his model to the weather. Cf. Poincaré 1908, 69. This shows, incidentally, that the term "deterministic chaos" really makes sense, though it seems at first glance to be a contradiction in terms. Cf. Schuster 1984, 7 9 - 8 0 ; Crutchfield et al. 1986, 48. Cf. Albano et al. 1986, 235. Cf. Farmer 1982, 243; Nicolis 1983, 336; Babloyantz 1986, 245.

Epistemic L i m i t a t i o n s II

263

We shall treat this supposition as a premise and take it to be a n a t u r a l e x t r a p o l a t i o n f r o m current research t h a t the brain's f u n c t i o n is essentially dependent on deterministic-chaotic processes. T h e question is, w h a t effects w o u l d it have on the scope of neurophysiological theories, if this hypothesis turned o u t to be correct, a n d , in particular, h o w the identity theory's reduction claim fares in light of this. Neurophysiological research reveals t h a t the neural states or changes in state t h a t could influence psychological processes or h u m a n behavior are considerably complex. A great many individual elements (i. e., synapses, neurons, etc.) take p a r t in such changes in state. T h i s m e a n s t h a t every change in state t h a t effects behavior must be u n d e r s t o o d as a long chain of changes in the states of these individual neural elements. Hence, in neurophysiological configuration space (which is supposed to describe unambiguously the states of all the various elements), t w o states t h a t are different to such a degree t h a t they are associated with different behavior are widely separated f r o m one another. O u r hypothesis t h a t deterministic chaos is significant to the functioning of the brain implies precisely that all functionally different n e u r o p h y siological states are always separated by a large number of intermediate neural steps. Although t w o of these neurophysiological states may, in fact, be b o u n d t o one another by strict neurophysiological regularities, that is to say, although each step determines the next, it may still n o t be possible to derive the w h o l e of the system's evolution in time f r o m these regularities. T h e reason being t h a t those chains of neurophysiological states t h a t effect behavior are of a complex nature. This means t h a t every relevant change in state includes a large n u m b e r of individual steps and, hence, contains a long-term change in the state of the system (though it may, in fact, only require a short a m o u n t of time). But these long-term changes can n o longer be predicted, if deterministic chaos is present. If we c o m p a r e this model of the structure of neurophysiological theories with t h a t contained in Fig. 2 (II.5.§ 4), which represents F o d o r ' s f u n c t i o n a l materialism, then we see that the latter presentation needs an essential modification. In Fig. 2, namely, the individual neural states ®ii were directly connected to the states 2j. T h e scenario developed here, on the other h a n d , implies that (1) the connections contain a large n u m b e r of intermediate steps, which (2) are extremely sensitive to changes in the b o u n d a r y conditions. Hence, whether the state 3>n produces O22 (as assumed in Fig. 2), depends o n a series of other

264

Epistemic Limitations of the Mind-Body Relation

entities O m n . Depending on the values the latter take, entirely different neurophysiological states follow O n . Some of which may be other 2j's (which are associated with the same psychological state as O22), but some of which may be states Oki, which correspond to entirely different psychological entities. What is essential here is that these processes cannot be tracked and predicted in detail by the theory. To be sure, the laws that connect every neurophysiological state with the next are known; but, because of the presumed sensitivity to small fluctuations in the boundary conditions, every such chain of events drowns in the chaos of the self-induced noise after just a few steps. Given these assumptions, therefore, the chains of the neurophysiological states that effect behavior cannot be traced and reconstructed theoretically. § 5 This creates the impression that predictions of behavior are impossible under the circumstances described here, and, hence, no regularities of human behavior can be specified. If this were, in fact, a consequence of the ideas developed here, then it would clearly be a reductio ad absurdum. After all, we are in general very capable of predicting the behavior of people we know. But chaos and predictability do not rule each other out entirely. Rather, in particular cases, it is possible to find regularities in chaotic systems by changing the level of description. "We will show how this is possible on the example of statistical mechanics. On the level of molecular description, a gas is made up of a chaotic confusion of colliding particles. Though the system is governed by the deterministic laws of classical mechanics, small variations in the boundary conditions cause it to change completely after just a few collisions and make it impossible to predict the situation on the level of molecular description. The processes of collisions in a container filled with a gas cannot be traced in detail by the theory. But despite this chaos on a microscopic level, we can still specify regularities which permit us to provide a reasonable approximation of certain aspects of this system's evolution. These regularities are formulated on a macroscopic level of description and include the laws of phenomenological thermodynamics. The complexity on a microscopic level can be treated using macroscopic terms that express the collective properties of the system (in this case, pressure and temperature, for example). This is possible because (as already mentioned in II.2.§ 3) states that are different from a microscopic perspective may be equal from a macroscopic point of view. In a gas at

Epistemic Limitations II

265

constant temperature, the kinetic energy of individual molecules changes continually; but the energy distribution and, in particular, its average value (and, hence, the temperature) remain unchanged. Therefore, it isn't necessary to track the temporal development of individual microstates, in order to be able to specify the nomological relations between these sorts of collective properties. Rather, statistical hypotheses of equiprobability are entirely sufficient for this purpose. For a comparison of the thermodynamic and the neurophysiological case, we can look at the derivation of the second law of thermodynamics in the framework of statistical mechanics, that was sketched in II.2.§ 3. According to the second law, entropy never decreases in a closed system, but rather it either remains constant or it increases. This law thus expresses an entity's development in time and in the context of the present discussion resembles in this way the psychological law *Fi -* ¥ 2 (i. e., a psychological state ¥ 1 is followed by a psychological state ¥2). To derive the second law, one first specifies the velocity distribution of the molecules in a gas. As previously mentioned, the same distribution of velocity (or energy) can be realized by a large number of microscopically different states. The number of possible microscopic realizations determines the probability of the corresponding distribution. This shows that, on the microscopic level, the second law corresponds to the statement that improbable distributions tend to develop toward more probable distributions. It is essential to our argument to look at the assumptions that enter into this derivation. Namely, in addition to the theorems of classical mechanics, statistical hypotheses play an important role here. The latter formulate expectations about what should be seen as equally probable, and express, for example, the assumption that the distribution of molecular velocity is isotropic, and that, on the average, there is no privileged direction in collision processes. Similarly, it is assumed that equal volumes in phase space always correspond to equal probabilities of the states thus described. With the help of such assumptions, the mean evolution in time of distribution functions, each of which comprises a large number of individual microscopic distributions, can be described with sufficient precision. Despite the fundamental chaos, a macroscopic description is successful in this case, because we can identify a class of equiprobable or, as it were, functionally equivalent micro-states. Hence, the precise structure of the microscopic chaos doesn't have any macroscopic conse-

266

Epistemic Limitations of the Mind-Body Relation

quences, and for that reason, it is unnecessary to follow the microfluctuations in detail. Since only statistical assumptions of a global nature are necessary, it is possible to detect statistical regularities on the microscopic level that can be translated into relations between the macroscopic variables. Thus, phenomenological thermodynamics is reducible to statistical mechanics. Accordingly, the conclusion is: Deterministic chaos does not in principle result in the complete unpredictability of evolution in time. Rather, this can be avoided by choosing an appropriate level of description. Chaos does not preclude reducibility in general. § 6 One essential aspect of this situation can be seen in the neurophysiological case. In Fodor's characterization of functional materialism (Fig. 2), several neurophysiologically different states are assigned to the same psychological state. Different neural processes are functionally equivalent. This means: Though several neurophysiologically different states (namely, the diverse members of the class 02j, but also other types of states such as 3k or C>4i) follow a particular state On, most of them are not functionally different. This shows how the chaos on the physiological level can be fenced in by changing the terminology. Analogously to the thermodynamic case, the complexity can be reduced by introducing appropriate macroscopic variables that characterize the classes of functionally equivalent micro-states. In this sense, the role of psychology is comparable to the status of phenomenological thermodynamics. In another sense, on the other hand, there are significant differences between the thermodynamic and the neurophysiological example. Up to this point, our argumentation can be understood as the demand that a statistical neuropsychology (analogous, as it were, to statistical mechanics) should be developed from which psychological laws can be derived. But this is impossible. To make this clear, let's consider Fodor's neurophysiological scenario again. According to this scenario, there is a statistical correlation between two psychological states and ¥ 2 , and a disjunction of neural states Oii's and 02j's is assigned to each of them. In principle, this assignment can be established by experimental means, i. e., by observing that always appears together with a member of the class of ii's and that ¥ 2 is analogously coupled to a state 2j. This could, for example, be accomplished with the help of Feigl's autocerebroscope. This means that a statistical correlation between the

Epistemic Limitations II

267

members of the state-classes of ®ii's and ® 2 j ' s can be demonstrated experimentally. Hence, the psychological law -> ¥ 2 can be translated into neurophysiological terms and the correctness of the translation can be tested experimentally. But the difficulty remains that this neurophysiological correlation cannot adequately be derived from the laws of neurophysiology. T h e reason for this difficulty is that the necessary statistical assumptions have a different status. As we have mentioned, the thermodynamic case only required very basic postulates about the equiprobability of all directions or velocity intervals. But the neurophysiological example would require assumptions about the equiprobability of particular neural states. O f course, one could seek to explain the correlation between Oii's and ®2j's by declaring all ® 2 j ' s (under the assumption that a ® i i is realized) to be equally probable, in order to maintain the analogy to the thermodynamic case. But such an assumption concerning the conditioned equiprobablity of neural states would no longer be acceptable as an explanatory hypothesis, because it would not refer to fundamental, but rather to theoretically derived processes. Hypotheses about the equiprobability of complex neurophysiological processes are not explanations, but are themselves in need of explanation. T o trace the regularity ®ii —> ®2j back to statistical assumptions about the equiprobability of all ® 2 j ' s (for a given ® i i ) would not explain the regularity, it would merely reformulate it. But tracing psychological laws back to neurophysiological assumptions of equiprobability is not just theoretically inappropriate, it is also inadequate to the facts. Up to this point, it has been assumed that microscopic chaos did not have any macroscopic effects, i. e., it did not effect behavior. T h i s assumption is unrealistic. In Fodor's scenario, neurophysiological correlates o f are also connected to neurophysiological correlates of psychological states that differ from ¥ 2 . Our premise that deterministic chaos is essential to the brain's processing of information suggests that at least some of the microscopic differences in states also have macroscopic consequences. For example, it certainly makes a psychologically tangible difference whether the state IN or the state ®IM is realized in Fig. 2. Generalizing from this fact suggests that an equiprobability of the relevant neurophysiological states does not actually exist. Therefore an approach that is based on mere equiprobability assumptions is not only formally insufficient but also materially inappropriate.

268

Epistemic Limitations of the Mind-Body Relation

The neurophysiological description cannot be accomplished exclusively by recourse to concepts like the equiprobability of neural states; rather, specificity — the particular type and manner of neural organization — is essential here. This means that one really does have to descend to the level of the temporal evolution of micro-structures, in order to achieve a psychologically relevant theory of brain processes. This requirement is based precisely on the fact (only too familiar to meteorologists) that microscopic chaos often becomes macroscopically relevant in a variety of ways. § 7 Hence, the argument developed here takes the following form: Deterministic chaos permits predictions in principle, if one succeeds in finding an adequate level of description. This is possible because small fluctuations in the initial conditions can lead to states that are, to be sure, microscopically different but are, nonetheless, macroscopically equivalent. This is precisely the situation that one encounters in statistical thermodynamics; and, in this respect, there is an analogy to the neurophysiological scenario discussed here. In other respects, the thermodynamic model and the neurophysiological model diverge significantly. In the neurophysiological model, the equiprobable alternatives cannot be derived from fundamental assumptions (such as the equiprobability of all directions). In addition, only a few, and certainly not all, of the neurophysiological states Oki that follow a particular neurophysiological state ij are functionally equivalent to one another. Hence, the strategy borrowed from thermodynamics of combining functionally equivalent micro-states into an equivalence class and associating this class with one macroscopic entity and thereby limiting the uncertainty by changing the level of description is only partially applicable. In no way are all ®kl's functionally equivalent to one another; some of them are connected to mental states of different kinds. This peculiarity is what distinguishes the neurophysiological scenario from the thermodynamic example; and precisely for this reason, it is not only conceptually but also empirically inadequate to employ an equiprobability assumption. Unlike the gas container (described by means of statistical mechanics), which micro-state is actually realized does, indeed, often make a difference to the brain. A neuropsychological theory has to be able to trace the succession of micro-states, and this is impossible according to our premise.

Epistemic L i m i t a t i o n s II

269

The transition to a macroscopic description in psychological terms allows us to formulate statistical regularities, even if they cannot be traced back theoretically to the micro-level of neurophysiological states; that is, even if they cannot be proved to be consequences of neurophysiological principles. Though these psychological laws can be translated into neurophysiological terminology, it cannot be demonstrated that these regularities expressed in neurophysiological terms are actually neurophysiological laws. There are two analogous correlations, namely * F i —y 4*2 and O i j - » < D k i , both of which are in need of explanation. But there are reasons to presume that such an explanation cannot succeed in the latter case, whereas, to all appearances, nothing prevents this in the former case. Hence, it is possible to assign neurophysiological concepts to psychological concepts, but it is not possible to derive psychological theorems from neurophysiological laws. In that case, psychology could not be reduced to neurophysiology. Though neural correlations can be specified that correspond to psychological laws, it cannot be proved that these correlations are neurophysiological theorems. But this is precisely what is required to be able to distinguish between the two alternatives represented in Fig. 1 (II.3.§4). Without this evidence of neurophysiological derivability, the experimentally demonstrable correlations Ojj —> Ok! could be precisely the effect of the psychological relation, i. e., of the lawful connection of and ¥ 2 (cf. VI.3.§ 4). To eliminate this possibility, we must show that there is causal closedness on the neurophysiological level, i. e., that neurophysiology is complete; and this can only be done by proving that the corresponding neurophysiological correlations can be derived theoretically. In our scenario, however, this is impossible precisely because of the presumed deterministic chaos. This means, from whatever else psychological laws can be derived, they cannot be derived from neurophysiological theories. Neurophysiology is not, then, the pertinent fundamental science for psychology, even if psychological states or events are, in fact, of a neurophysiological nature. We wish to distinguish between two levels of description here. The first one is the level of neurophysiological concepts. It is concerned with the structure of individual neurophysiological systems and describes the way they work by means of deterministic laws. On this first level, therefore, structural concepts and deterministic theories are employed. This is to be distinguished from the psychological level of description. Here psychological states are identified by their function as explanations

270

Epistemic Limitations of the Mind-Body Relation

of behavior, and these explanations are (like the psychological regularities on which they are based) themselves of a statistical nature. Thus, the second level is characterized by functional concepts and statistical theories. The crucial result is, then, that there is a gap between the two approaches that might rule out tracing the second level of description back to the first. Even if there were an identity of psychological and physical states, in the light of the state of neurophysiological research that is extrapolated here, a reduction of psychology to neurophysiology would not be possible. Hence, our thesis has the status of an emergence claim-, psychological laws are emergent relative to neurophysiology. By which we mean the principle-factual variant of logical emergence mentioned in IV.§ 4: The properties and the evolution of deterministic-chaotic systems cannot, in fact, be derived from the laws that apply to their components. That we are dealing with principle-factual emergence is suggested by the fact that limitations to reduction are of a principle nature and are based on fundamental limits. That is, our only claim is that the behavior of the whole cannot be derived from the behavior of the parts. It has not been shown that the behavior of the whole is not determined by the behavior of the parts (and their interactions). Rather, the structure of our argument presupposes this. In summary, we have shown that, given that the identity theory in functional materialism's version is true and assuming what we consider to be a plausible hypothesis about the future state of neurophysiological research a reduction of psychology to neurophysiology is not possible. With that, the only reason for supporting the identity theory collapses. If the identity theory is correct, then it cannot justifiably be advocated. If there is, in fact, mind-body identity, we still cannot know it. Apart from that, we do not associate any actual truth claims with our scenario. Neurophysiology cannot be pursued from the philosopher's armchair. Our only point was to show that a plausible projection for the evolution of neurophysiology could have significant philosophical consequences.

IX. The Mind-Body Relation and Human Self-Understanding § 1 T h e foregoing analyses and discussions may have created the impression that all that philosophy can contribute to the explanation of consciousness is a mere reflection about science. T h a t is, that whoever has understood that mental terms are theoretical terms which represent, for their part, explanatory constructions would also know what consciousness is. But this would be an extreme misunderstanding. Consciousness, seen philosophically, is not only a term in scientific theory — and, seen historically, it has been so only in extremely rare cases. This does not mean that philosophy over and above its jurisdiction as philosophy of science has a special source of knowledge of its own that is completely closed to science. This, too, would be a misunderstanding that would run counter to the inner unity (notwithstanding all reciprocal frictions) of philosophy and science, namely as expressions of a rational view of things. What is meant, rather, is that the concept of consciousness is richer in meaning than its explication by philosophy of science reveals, and that philosophy has other ways of approaching the concept of consciousness than the approach by way of philosophy of science that has been emphasized thus far. Of course, this does not mean that philosophy could put itself in the place of science: Philosophy should not (and cannot) explain the consciousness that science is trying to explain. N o r would anyone, in the words of the neurophysiologist Creutzfeldt, deny that our experiences and reactions are based on the ability of the nervous system to pick up certain physical stimuli, to code them as sequences of action potentials and to transform this information into appropriate sequences of the motoric systems, which gives a sufficient explanation to reductionism. And finally no one will deny that motoric, complex and social behavior can be found in humans that differs from that of lower animals at most in quality, as ethology has taught us. 1

1

C r e u t z f e l d t 1981, 3 3 .

272

T h e Mind-Body Relation and H u m a n Self-Understanding

But this is not the issue here, as Creutzfeldt himself emphasizes immediately after these remarks with reference to inappropriate applications of properly described mechanisms precisely to areas "which they cannot explain and describe." 2 Among these areas is the philosophical concept of consciousness. "Consciousness," "self-consciousness," and "ego" — these are primarily philosophical terms, not neurophysiological ones, and only secondarily general scientific terms that have an empirical background. Thus, the neurophysiologist's comment, itself philosophical in kind, that reflection about consciousness is not primarily a scientific question. It is a special peculiarity of human consciousness, that it cannot give any definitive answer about itself. 3

"Definitive" is to be understood here in the sense of a scientific answer. By contrast, a philosophical answer leads into epistemological and (philosophico-)anthropological domains. Once again: Philosophy should not hope to solve problems that science can solve better — and, hence, that are defined as scientific, not philosophical problems. Its task is to produce clarity in all domains of our self-understanding, and understanding of our situation including our scientific understanding. It serves the latter as philosophy of science. What is peculiar to philosophy here is that it undertakes this task by thinking, not by research. This, of course, does not mean that research does not think, but that philosophical reflection is not research in the same way that science — in the light of theories and empirical methods — does research. "Consciousness," "self-consciousness," "ego," but also "self-understanding" are titles of this specifically philosophical way of orienting oneself in thought and through thought. And this orientation is not a cultural luxury that a rational society performs alongside its everyday life of work and formation of scientific knowledge; nor can it be performed by science itself. Pointedly formulated: Socrates' question about man's proper understanding of himself and his situation is not answered, or even made easier to answer, by increasing our scientific knowledge of the brain. This would, again, only mean that we believed (scientistically) that we had understood everything if we had understood ourselves from a neurophysiological perspective. In this sense, the dual2 3

¡bid. Creutzfeldt 1983, 9.

The Mind-Body Relation and Human Self-Understanding

ism advocated here is not only a dualism in theoretical in practical intent.

273

intent, but also

§ 2 Of course, this should not be taken as a invitation to all philosophers gayly to continue on with talk about consciousness, etc., as if neurophysiology and scientific psychology had never been invented. Scientific knowledge helps one to avoid the naïveté with which philosophical opinions tend to put themselves in the place of scientific research. The Socratic question of man's proper situational and selfunderstanding does not permit us to ignore scientific knowledge. Rather, properly understood, it includes the latter and is simply opposed to naïve scientism of the sort mentioned. Wittgenstein, too, opposed this sort of scientism, i. e., the unreflected transfer of the methodological and epistemological ideals of the empirical (and the exact) sciences to other, quite differently structured realms of knowledge. But he does so in a way that throws the (scientific) baby out with the (philosophical) bathwater. Wittgenstein starts with a critique of the notions of the non-objectifiability and non-experienciability of a transcendental subject, 4 as well as the transcendental philosophy's idea that the "philosophical" ego is not the subject matter of psychology. The philosophical ego shrinks to a point without extension, and there remains the reality coordinated with it. T h u s there really is a sense in which philosophy can talk a b o u t the self in a non-psychological way. W h a t brings the self into philosophy is the fact that " t h e w o r l d is my world". T h e philosophical self is not the h u m a n being, not the h u m a n body, or the h u m a n soul, with which p s y c h o l o g y deals, but rather the metaphysical subject, the limit of the w o r l d — not a part of it. 5

Later, in the Philosophical Investigations, Wittgenstein criticized (in the form of the Cartesian two-substances dualism) a dualistic understanding of the mind-body relation ("Where our language suggests a body and there is none: there, we should like to say, is a spirit"6). He gives this relation an analytical form, that not only presents the classical mind-body problem as pseudo-problem, but also presents psychology as 4 5 6

Tract. 5.631-5.633 ( = Wittgenstein 1922, 116). Tract. 5.64-5.641 ( = Wittgenstein 1922, 116-118). Wittgenstein 1958, §36 (emphasis in the original).

274

T h e Mind-Body Relation and H u m a n Self-Understanding

a whole, and, hence, cognitive psychology as well, as a fundamentally flawed enterprise: The confusion and barrenness of psychology is not to be explained by calling it a "young science"; its state is not comparable with that of physics, for instance, in its beginnings. [...] For in psychology there are experimental methods and conceptual confusion. [...] The existence of the experimental method makes us think we have the means of solving the problems which trouble us; though problem and method pass one another by.7 Not only is such a judgment hardly in accord with the development and with the present state of cognitive psychology; it also no longer meets the standard of a philosophical reconstruction of scientific psychology in general. Moreover, we are merely mystifying ourselves, if we define our insight without regard to the present state of knowledge and, in addition, as Wittgenstein does, take cognitive psychology to be possible only as non-scientific knowledge. Ryle's critique of philosophical and scientific mentalism is different, although it is close to Wittgenstein's position in many respects. Ryle criticizes both a dualistic conception of the mind-body problem oriented towards a Cartesian dualism of two substances, as well as a reductionist behaviorism. 8 Ryle exposes category mistakes in the two-substances dualism (and in its monistic alternative), 9 and accuses reductionist behaviorism of confusing action with behavior. At the center of his reflections is the reconstruction of psychological statements, or talk about psychological states or events, as statements about dispositions, i. e., (intersubjectively accessible) possibilities of action. 10 An opposition of (private) "internal world" and (public) "external world," as Cartesian dualism states, is eliminated by representing the "internal world" by the organizational forms of "external" action, or by dispositions towards action. 7

Wittgenstein 1958, X I V (232e) (emphasis in the original). Adherents to Wittgenstein's radical assessment can still be found today, w h o advance linguistic arguments against the possibility o f a cognitive psychology. Cf. M a l c o l m 1 9 7 1 , 3 8 5 - 3 9 2 ; B a k e r / H a c k e r 1984a, 11; also B a k e r / H a c k e r 1984b, VII - XIII. An attempt to, as it were, reconcile Wittgenstein with modern forms of cognitive psychology is made by M e n d o n j a 1987, though without succeeding in transforming cognitive psychology's orientations into a Wittgensteinian p r o g r a m .

8

Ryle 1949, 1 1 - 2 4 , 3 2 7 - 3 3 0 . A category mistake results from introducing expressions o f a particular logical type into the variables of schematic sentences that only produce meaningful sentences for expressions o f other logical types. Cf. K. Lorenz 1984c, 3 7 0 .

9

10

Cf. Ryle 1949, 1 1 6 - 1 5 3 .

T h e Mind-Body Relation and Human Self-Understanding

275

Problems that result from this reconstruction of the distinction between things and persons, were made the subject of analyses by Strawson 1 1 and Sellars. 1 2 Ryle's critique of a "philosophy of mind" is much more precise than Wittgenstein's exaggerated critique, and it is more

systematic.

Nevertheless, Ryle has no plan for laying the foundations of a (nonCartesian) philosophical psychology or even of a methodology for psychology; rather, he is merely content with destroying the Cartesian myth that is hidden behind the traditional "philosophy of mind." N o r does he intend a philosophy of psychology: It has not been a part of the object of this book to advance the methodology of psychology or to canvass the special hypotheses of this or that science. Its object has been to show that the two-worlds story is a philosophers' myth, though not a fable, and, by showing this, to begin to repair the damage that this myth has for some time been doing inside philosophy. I have tried to establish this point, not by adducing evidence from the troubles of psychologists, but by arguing that the cardinal mental concepts have been credited by philosophers themselves with the wrong sorts of logical behaviour. 13 § 3

However, Ryle's declaration that the mind-body problem is a

pseudo-problem is not the end of the matter. If philosophy were to be satisfied with this information, then the only route left would be to choose between mere opinions and scientific orientations. But science all t o o often gives a description of the life world that we can hardly recognize: The world presents itself to the brain by way of the sense organs. Already here, its unity is divided into a multiplicity of phenomenal forms, in as much as every sense organ responds only to a limited spectrum of energy transfer: the eye to the limited wave length sector of "visible" light, the ear to the narrow area of mechanical vibrations, the sense of touch to long-wave heat radiation and to lower frequency mechanical vibrations, and the senses of smell and taste to a narrowly restricted area of concentrations of certain molecules. The world, as it presents itself to us is thus limited to a narrow sector of physical and chemical phenomena. This is our life world. 14 11 12

13 14

Strawson 1959. Sellars 1963. For a more detailed treatment of the mind-body problem in Wittgenstein and Ryle, cf. Rentsch 1980, 201 — 204; on the general conception of a critique of language in Ryle, cf. K. Lorenz 1970, 131 - 1 3 9 . Ryle 1949, 329. Creutzfeldt 1981, 34.

276

T h e Mind-Body Relation and H u m a n Self-Understanding

Is it really like this? The fact that scientific terms are to be seen as explanatory constructions already speaks against it. The monistic side also emphasizes that our picture of the world is always a construction,15 And what is true of this picture must also be true of our biological knowledge of the processes of consciousness. To put it differently: If everything that we normally know about the world is at least partially a construction of the brain, then what biology knows about the brain is also unable to exist without being touched by constructions. Neurophysiological research, in part, also sees it this way, as does Creutzfeldt, for example, when he speaks of the human brain's capability of symbolic representation emphasizing thereby that the symbols themselves are not a function of the nervous system, nor are they the nervous system, nor the world itself. 16 It is obvious from the start that a symbolic representation of the world "represents neither the world as it is nor the activities of the brain as the neurophysiologist measures it." 1 7 The symbolic representation of the world is not identical to the brain's representation of the world. But this means that our theoretical symbolizations are also constructions in this sense. Hence, Creutzfeldt draws a dualistic conclusion from these reflections, which he pointedly expresses as follows: "Dualism — i. e. standing opposite oneself — is the nature of consciousness." 18 It is clear that this is not a neurophysiological statement, but, rather, a philosophical one. With regard to what has been said about a practical dualism, it therefore need not be a meaningless or superfluous statement. This is also true, though with some restrictions, of the following statements: The symbols, i. e., "the brain's symbolic representation of reality, not only refer to the world but rather independently form their own world, to which our brain in turn constantly refers: the world of the mind." 1 9 Or: The unity of consciousness presents itself to the philosopher and to the neurophysiologist [...] in the same way as a problem: It can be derived neither from a self-analysis of reason nor from the scientific-

15

16 17 18 19

Cf. Pöppel 1985, 1 3 9 - 1 4 0 . Seen from this point of view, Pöppel's reductionistic critique of dualism, whereby the mind plays a dubious role as a "deus ex machina" (Pöppel 1985, 1 4 5 - 1 4 6 ) , is methodologically premature. Creutzfeldt 1981, 43. Creutzfeldt 1981, 41. Creutzfeldt 1981, 42, cf. 43. Creutzfeldt 1981, 42.

T h e Mind-Body Relation and H u m a n Self-Understanding

277

technical analysis of the anatomic-functional organization of the brain. On the contrary, both analyses — that of reason, whose nature (as Kant put it) it is to pose questions about itself that it cannot answer, and that of the neurophysiologist, who with technical measurements poses questions to the nervous system that are excluded by this scientific technology itself — lead to the same p a r a d o x . They must make a logically compelling but scientifically not further justifiable assumption about the ultimate ground of the unity of consciousness. 2 0

Or, with reference to the statement that dualism explains the nature of consciousness: N o reductionist theory does justice to this essence of consciousness. They all grasp only one side, namely, the brain mechanisms, but not the other, the world of symbols, with which these mechanisms, and through them the brain, confront themselves, and which are as real as the natural world. 2 1

" T h e world of the mind," "the unity of consciousness" — these formulas, like the statement (reminiscent of Popper) that the world of symbols is as real as the natural world, breathe the spirit of classical philosophy, not that of the relatively sober, modern philosophy of science. Nevertheless, they are also compatible with the dualism we advocate here, with respect to the emphasis on the constructive character of our orientations, including scientific orientations, and the rejection of the reductionist thesis of an identity of brain mechanisms and consciousness. What is really crucial in all of this is not that pure mind once again lifts over biology, but rather that the practice of biology is seen in its theoretical research programs from the perspective of methodological clarification. Whether the "unity of consciousness" or the asserted "reality" of the symbolic world has to be a part of this, is a question of secondary importance. § 4 The constructive character of our orientations, both the scientific and the non-scientific, means that we are living beings that construct not only the world in which we live — our world — but also ourselves, by living in and with our self-understandings (in Heidegger's terms: by living in and with our Entwürfen). There is a difference between developing scientific knowledge about oneself and self-understanding of oneself. The concept of consciousness stands for that, too, within the practical 20 21

Creutzfeldt 1981, 38. Creutzfeldt 1981, 43.

278

The Mind-Body Relation and H u m a n Self-Understanding

sense of dualism sketched here. Accordingly consciousness — and this seems tautological only at first glance — is always experienced or precisely "conscious" consciousness. To put it in scientific terms, in consciousness humans confront themselves, particularly in the form of self-consciousness: Consciousness is the experience of ourself and our world. This experience confronts us as an observer with ourselves and our perceptions. T h e experience of every instant must be related to earlier experiences, so that it presupposes memory. And it must be communicable to oneself and to others. The manifold of experiences finally stands in a unified relation to all our experiences and thus presupposes the experience of a self that understands itself as identical. 22

However, the observational terminology, which here assimilates experience and self-experience to scientific experience in the usual methodological sense, is misleading in many respects; it suggests "objective" conditions. Nor is the observer on a tower or in front of a cloud chamber meant, but the ego or self that knows itself in the process of forming knowledge, i. e., that lives, experiences, knows. These, in turn, are not scientific, but life world or philosophical concepts; and this less in the sense of a Lebensphilosophie, i. e., a movement aimed against the tendency toward an increasingly scientific appropriation of the world and the self (Kierkegaard, Nietzsche, Bergson, et al.), but rather in an epistemological sense, just as Kant meant in substituting the transcendental ego for the empirical ego (the "empirical consciousness") (cf. I.2.§ 5). As Kant put it: All representations have a necessary relation to a possible empirical consciousness. For if they did not have this, and if it were altogether impossible to become conscious of them, this would practically amount to the admission of their non-existence. But all empirical consciousness has a necessary relation to a transcendental consciousness which precedes all special experience, namely, the consciousness of myself as original apperception. [...] T h e synthetic proposition, that all the variety of empirical consciousness must be combined in one single self-consciousness, is the absolutely first and synthetic principle of our thought in general. But it must not be forgotten that the bare representation "I" in relation to all other representations [...] is transcendental consciousness. 23

22 23

Creutzfeldt 1981, 33. Critique of Pure Reason A 1 1 7 - 1 1 8 footnote (emphasis in the original).

The Mind-Body Relation and Human Self-Understanding

279

This is the Kantian "I think" which must be able "to accompany all my representations," 24 and it is the concept of consciousness that does not lead to a two-substances dualism in the Cartesian sense, but rather to the concept of "duality" of consciousness, 25 meaning the dual structure of consciousness (in the practical as well as the theoretical sense). To this context belongs the "undismissable first-person asymmetry," 26 emphasized by Dennett, as well as Popper's comment that the self does not belong to the brain, but rather the brain to the self. All of this makes good sense, particularly if it is viewed from the perspective of life-world practice, not from the perspective of scientific theory. Theoretical consequences are linked to this. Namely, if consciousness, especially in the sense of self-consciousness, is a philosophical term for which there is no corresponding neurophysiological term, then it does not make sense to speak of an "evolution" of consciousness, or to take as the basis of all investigations of the concept of consciousness 27 the claim "that in the human brain as a product of evolution only functions that can be traced back to evolutionary adaption processes are represented." 2 8 § 5 Hence, a philosophical analysis that cannot be reduced to a scientific analysis, in this case a neurophysiological analysis, also leads to a dualistic conception. "Consciousness," "self-consciousness," and "ego" are dualistic terms; they cannot be formed in a monistic conception, or — if we assume that they are philosophical terms — they cannot be reconstructed. And that itself is an argument that psychology cannot be reduced to neurophysiology. Furthermore, dualism in the practical sense emphasized here corresponds directly to experience of the life world. Let us take the case of someone sitting down at his desk, not exactly overjoyed, and telling himself — like it or not — to start thinking (because he has to finish a book like this one, for example). Does scientific knowledge change anything in such a sometimes consoling, 24 25 26 27

28

Critique of Pure Reason B 1 3 1 - 1 3 2 . Cf. I.2.J 5. Cf. Creutzfeldt 1981, 48. Dennett 1982, 351. Creutzfeldt also responds critically to this, Creutzfeldt 1981, 31 ("If there is [...] no possibility at all of determining the consciousness of our non-dominant hemisphere, to say nothing of other creatures that cannot communicate their life worlds to us — and apparently to themselves as well —, we must seriously ask whether it is meaningful to speak of an 'evolution of consciousness'."). Poppel 1986, 75.

280

T h e Mind-Body Relation and H u m a n Self-Understanding

sometimes desolate situation? Certainly not. We are also basically quite right when we conceive of ourselves as masters of our own home, to which our brain belongs as well. "Scientific" explanations are unnecessary for that. From a practical perspective, then, it should be emphasized that a scientific answer to the mind-body problem is at most an answer to the question of man's place in nature. But man is not just a creature of nature, but also a moral creature that sets goals. Even a complete description of his actual behavior could not (under penalty of the naturalistic fallacy) adequately take the realm of ethical values into account. In the same sense, the (by his own admission) vulgar materialist, Rorty, admits that we need many different descriptions of ourselves — some for some purposes and others for others, some for predicting and controlling ourselves and others for deciding what to do, what meaning our lives shall have. 29

The perspectival existence of man emphasized here suggests a way to extend the concept of folk psychology beyond its present framework. In the preceding presentation of psychosemantics and folk psychology, we made the distinction between folk psychology in the technical sense and in the intuitive sense (VII.4.§ 2). In the technical sense, folk psychology refers to propositional attitudes, as characterized by the use of an intentional terminology; in the intuitive sense, it displays the familiar pattern of explanations of action that make up life-world orientations. This distinction, in which cognitive psychology is placed on the side of technical folk psychology — it has an intentional terminology, but contains regularities that intuitive folk psychology lacks —, is according to what we have presented not yet complete. It should be supplemented by the concept of a folk psychology in the practical sense. The latter deals with interpretations of actions in non-descriptive contexts; these interpretations are not objects of scientific psychology, nor of technical folk psychology, nor of intuitive folk psychology in its function as an observational theory. 30 Folk psychology in the practical sense — or practical folk psychology — is not a scientific (explanatory) theory in the narrower sense, nor does it serve theoretical (scientific) ends. Rather, 29 30

R o r t y 1982, 3 4 5 (emphasis in the original). Von Wright's theoretical analyses, for example, could serve here as a p a r a d i g m a t i c basis for this sort of folk psychology. Cf. von Wright 1971.

The Mind-Body Relation and Human Self-Understanding

281

as a philosophical conception, it aims more at the philosophical problems that we have characterized under the heading "human self-understanding," the problem of consciousness, for example. Even if (contrary to expectations), in the distant future, neurophysiological theories could do justice to our psychodynamics, this would still not give us cause to view our practical relations with one another as interactions of neural mechanisms and, hence, to exile every form of practical folk psychology to the realm of "outmoded theories." This would be analogous to the fallacious inference from Darwinism to social Darwinism. Our practical (or moral) self-understanding remains unaffected by any empirical success of a monistic interpretation. It is not true here that: scientia locuta, causa finita. One could also speak of a complementarity of different descriptions in Bohr's sense. In its original sense, complementarity means that a causal description (i. e., the strict applicability of conservation laws) and a space-time description (i. e., the precise space-time representation of a system's evolution) preclude each other, but that they are both necessary to a complete understanding of the phenomena. Choosing one approach makes it impossible to determine the characteristic magnitudes of the complementary approach. 31 This does not mean that there are several, equally valid theoretical interpretations of the mind-body relation, but that there is an analogous complementarity between a scientific and a philosophical, life-world description of man. Both are based on different, mutually exclusive assumptions and both seek different ends. In science's intersubjective approach, for example, private, autonomous self-consciousness cannot be defined; anyone who insists on the indispensibly privileged character of the private, cannot pursue science. Though it would be superfluous to refer to self-consciousness to explain our behavior, it is indispensable for explaining our self-understanding. If we want to orient ourselves, we require different methods than if we want to control ourselves. In this sense, freedom and human dignity do not stand under the proviso of the evolution of science, and the declaration of human rights cannot be revoked by neurophysiology. But this need not mean that philosophical and scientific culture should settle on islands far apart from one another. This is especially the case if we are dealing with questions to which both have an answer — 31

Cf. Bohr 1928.

282

T h e Mind-Body Relation and H u m a n Self-Understanding

even if we usually find that they have answered different questions. Creutzfeldt, to whom we have referred several times, ends his talk by returning once again to the problem of the unity of consciousness, with the remark: "The neurophysiologist's answer is the beginning of the philosopher's questioning." 3 2 This is certainly true, at least in this case. But it is also true that the answer to the philosopher's question should be informed by science. Otherwise, what philosophy is so often (and, unfortunately, rightly) accused of would hold: that it seeks to escape reality in speculation, instead of grasping it in thought.

32

Creutzfeldt 1981, 43.

Conclusion At the end of our explorations into the realm of the philosophy of mind, it will be useful to recapitulate briefly the part of our argumentation that deals with philosophy of science and to draw a few conclusions. Our argument in support of dualism proceeds from the simple principle: What looks different, should be treated as different, until it is shown otherwise. From this it follows, first, that the burden of proof lies with monism and not dualism. It is not up to dualism to demonstrate that a physical explanation of mind is impossible, but rather up to monism to prove that such a physical explanation is adequate. Expressed in the language of statistics: Dualism is the null hypothesis; there must be special reasons to reject it, but not to accept it. 1 In principle, there are three possible lines of argument that could justify a monistic mind-body identification. The first possibility is an ontological reduction, which is itself based on a theory reduction of psychology to neurophysiology. Second, psychophysical identity might possibly be established experimentally, and, third and finally, it is possible to imagine it being based on principles of a general nature. It was argued that the experimental option does not, in fact, exist (cf. II.3.§ 3—4), and that there are no principles that are both logically strong enough to preclude dualism and logically weak or flexible enough to assure that they will not interfere with the progress of science (cf. VI.4). That leaves reduction as the only route to monism; and, in fact, as we have tried to make clear, all monistic approaches can be reconstructed (in some cases contrary to their own self-assessment) as reduction claims (cf. II.7). This implies that a sound and justified defense of monism requires that these reduction claims be met. But nothing can be found at present that could satisfy such a demand; in the present state of affairs, a reduction of psychology to neurophysiology is not available even in broad outline. Hence, the claim of reducibility is ill-supported. 1

Thus, the attempts to pass on some or all of the burden of proof to dualism — as is the case in Schlick 1925, 3 1 8 - 3 1 9 ; Rorty 1965, 41; and Fodor 1981, 2 - are to be rejected.

284

Conclusion

In addition, two reasons speak against the assumption that such a reduction awaits us in the foreseeable future: First, an argument by analogy taken from the history of science makes it clear that attempts to reduce psychological theories to physiologically or non-mentalistically structured theories have failed time and again (cf. V.l). True, despite these experiences, reduction claims can still be upheld, but these tend to become mere personal confessions. Second, extrapolating from current tendencies in neurophysiology suggests that a reduction claim cannot justifiably be advocated, even if it is in fact correct. Even if reducibility exists in fact, it is possible that we will never know (cf. VIII.2). Not only are monistic claims of identification insufficiently supported; positive reasons for a dualistic interpretation can also be given. In doing so, we first assume (with Rescher) that there is no hierarchy among the sciences. It is inappropriate to rank the natural sciences as unconditionally superior to the social sciences. 2 This judgment is backed by our investigations into the structure of psychological theories. It was shown that psychological concepts have the same logical features as physical concepts, and that psychological theories stand up to the same methodological criteria as physical theories (cf. V.2.§ 5 — 6; V I . 1 . § 2 — 4). Despite this conceptual and methodological symmetry between physics and psychology, the philosophical interpretations of each of these sciences are usually carried out in completely different ways. While one refers to present-day physics in discussing the philosophy of nature, the philosophy of mind operates with estimates of the state of future psychology. This sort of asymmetry in the respective philosophical approaches is ill-founded and inadequate. Long-term extrapolations about the state of scientific research are, moreover, not possible. The course of scientific progress is never immune to fundamental alterations; and such scientific revolutions can lead to essential changes in our image of nature. But this precludes predictions about the projected path of science, which leads one to conclude, as Rescher rightly remarks, that present-day science cannot speak for future science: The inherent unpredictability of future scientific developments [...] means that present-day science cannot speak for future science: it is in principle impossible to make any secure inferences from the substance of science at one time about its substance at a significantly different time. 3 2 3

Cf. Rescher 1979, 186. Rescher 1984, 102 (emphasis in the original).

Conclusion

285

This means that the answers to philosophical questions about science, i. e., about what science knows, should approach it as it is and not be based on fictions about its future. Precisely this approach (as has been emphasized many times) is the proper practice for the philosophical interpretation of physical theories. We can see no reason why the philosophy of mind should instead follow the model of the Naturphilosophie of the early 19th century, which took the state of the sciences as a mere starting point and then gave free rein to the imagination. The state of the sciences should not simply be taken as raw material, stimulation and source of inspiration for the development of far-reaching interpretations. This respect for science, which is expressed by taking into account what science knows and, conversely, what it does not know, suggests we should take the current state of psychological research seriously. And this, in turn, demands that we take the ontological commitments of psychological theories into consideration to the same degree as the ontological commitments inherent in physical theories. If we base our best available theories of behavior on mental entities, then this must be reflected in our judgment about what exists. In this way, mentalism leads to dualism (cf. VI.2), unless serious obstacles stand in the way of such an inference — and that, as we have argued, is not the case (cf. VI.3). In order to forestall a possible misunderstanding it should be stressed once again that the mere fact that neurophysiology is at present unable to account for higher level behavioral regularities is by no means sufficient to support a dualistic claim. What is required in addition is the existence of empirically and methodologically qualified psychological theories that allow us to derive these regularities. Our argument is based on the observation that there in fact exists a successful rival research program to neurophysiological psychology, namely, cognitive psychology. Accordingly, we do not simply recommend jumping from neurophysiological ignorance to philosophical claims. What we do recommend is to progress from psychological knowledge (taking into account all the qualifications and considerations mentioned) to philosophical claims. A dualistic interpretation alone does not teach us much about the nature of the mind; we have not yet understood the psychological mechanisms, if we emphasize the autonomy of mental phenomena. Dualism shares this fate with monism, which also needs to be sup-

286

Conclusion

plemented by a theory of psychological functions. In fact, the problems in this domain are largely independent of ontological commitments. Nevertheless, it is clear in this context that basing our studies of psychological processes on modern psychology has implications for the interpretation of the mind-body relation. This is because the classic brand of the identity view, namely, the identification of processes in consciousness (i. e., folk psychological natural kinds) with neurophysiological processes (i. e., physical natural kinds), is called into question by the current state of psychological research. The latter makes it unlikely that the mental phenomena that are accessible to consciousness correctly reflect the dynamics of mental life (cf. VII.4.§ 9). This proves once again that one is on shaky ground with overly hasty predictions about the evolution of science. After all, the identity thesis ran into difficulties here where its classical advocates never even sensed a problem. As a whole, the dualism advocated here is of a pragmatic nature in two respects. It is, first, a dualism suggested by a life-world perspective (cf. IX), and a dualism of this sort is almost universally agreed upon. Moreover, our brand of dualism embraces, second and more importantly, a theoretical position. The pragmatic aspects of that position concern only the nature of the reasons adduced in its support. In contrast to a strict or non-pragmatic dualism, we hold the view that psychophysical identity cannot be rejected out of hand; no logico-terminological obstacles stand in its way. Mind-body identity is a coherently formulated option for the future evolution of science. If we distinguish (following Frege) between the content of a proposition and the method we use for assessing its validity, the pragmatic aspect of our dualism is confined to the second part of that distinction; it refers to the methods of justification and not to the nature of the claim entertained. Our concern is primarily that the current state of research be treated with more respect and that the philosophical consequences of limiting ourself to that which we (halfway) know be explored. This should introduce a greater degree of sobriety into the mind-body debate, which has often been lost due to bold a priori constructions and imaginary neurophysiologies and neuropsychologies. Since we don't know what we don't know, it seems to us that sober restraint is the most appropriate philosophical stance. On this basis, a (pragmatic, interactionist) dualism appears to be the best supported option.

Bibliography References to texts from classical antiquity as well as from the philosophical tradition, unless listed differently in the bibliography, use the standard abbreviations: e. g. Aristotle De an. = De Anima; Pbys. = Physica; Aquinas, S. th. = Summa tbeologica; Descartes, Princ. Philos. = Principia Philosophiae-, Leibniz, Disc. met. = Discours de métaphysique. Kant's Critique of Pure Reason is cited in the usual manner according to the pagination of the original edition (A = 1st ed., 1781; B = 2nd ed., 1787). A. M. Albano et al. (1986) "Lasers and Brains: Complex Systems with Low-Dimensional Attractors," in: Mayer-Kress, 1986, 2 3 1 - 2 4 0 Alexander von Aphrodisias (1887) Alexandri Aphrodisiensis praeter commentaria scripta minora: De anima liber cum mantissa (ed. I . Bruns) Berlin, 1887 (Supplementum Aristotelicum II 1) M. Alonso/E. J. Finn (1967) Fundamental University Physics I (Mechanics), Reading Mass. etc.: AddisonWesley, 1967 Aristotle, De gen. et corr. The Works of Aristotle (ed. W. D. Ross), De Generatione et Corruptione, (trans. H. H. Joachim), London: Oxford University Press, 1966 D. M. Armstrong (1968) A Materialist Theory of the Mind, London, New York: Routledge, Humanities Press, 1968 D. M. Armstrong (1977) "The Causal Theory of the Mind," in: Philosophische Psychologie?, ed. R. Bubner et al., Göttingen: Vandenhoek u. Ruprecht, 1977 (Neue Hefte für Philosophie 11), 8 2 - 9 5 A. Augustinus (1841 - 1849) Opera omnia, / — X V (ed. J.-P. Migne), Paris 1841 —1849 (Patrologiae cursus completus. Series Latina 32 — 47) B. Aune (1966) "Feigl on the Mind-Body Problem," in: Feyerabend/Maxwell, 1966, 17 — 39 E. Averill/B. F. Keating (1981) "Does Interactionism Violate a Law of Classical Physics?" Mind 90 (1981), 102-107

288

Bibliography

A. J. Ayer (1959) Logical Positivism, ed. A. J. Ayer, Glencoe: The Free Press, 1959, 4 1963 (The Library of Philosophical Movements) A. Babloyantz (1986) "Evidence of Chaotic Dynamics of Brain Activity During the Sleep Cycle," in: Mayer-Kress, 1986, 2 4 1 - 2 4 5 G. P. Baker/P. M.S. Hacker (1984a) Language, Sense and Nonsense. A Critical Investigation into Modern Theories of Language, Oxford: Blackwell, 1984 G. P. Baker/P. M.S. Hacker (1984b) Scepticism, Rules and Language, Oxford: Blackwell, 1984 A. Bandura (1977) Social Learning Theory, Englewood Cliffs: Pretnia Hall, 1977 A. Bandura/D. Ross/S. A. Ross (1963) "Vicarious Reinforcement and Imitative Learning," Journal of Abnormal and Social Psychology 67 (1963), 6 0 1 - 6 0 7 W. Bechtel (1987) "Connectionism and the Philosophy of Mind: An Overview," The Southern Journal of Philosophy 26 (1987), Supplement, 1 7 - 4 1 W. Bechtel (1988) Philosophy of Mind. An Overview for Cognitive Science. Hillsdale NJ: Lawrence Erlbaum, 1988 G. Berkeley (1710) A Treatise Concerning the Principles of Human Knowledge, Dublin 1710, London 2 1734, in: The Works of George Berkeley, Bishop of Cloyne II, ed. A. A. Luce/T. E. Jessop, London: T. Nelson, 1949 W. Berkson (1974) Fields of Force. The Development of a World View from Faraday to Einstein, London: Routledge, 1974 W. Bernard (1988) Rezeptivität und Spontaneität der Wahrnehmung bei Aristoteles. Versuch einer Bestimmung der spontanen Erkenntnisleistung der Wahrnehmung bei Aristoteles in Abgrenzung gegen die rezeptive Auslegung der Sinnlichkeit bei Descartes und Kant, Baden-Baden: Körner, 1988 (Saecula Spiritalia 19) K. Bernath (1969) Anima forma corporis. Eine Untersuchung über die ontologischen Grundlagen der Anthropologie des Thomas von Aquin, Bonn: Bouvier, 1969 P. Bieri (1981) "Generelle Einführung/Einleitung," in: Analytische Philosophie des Geistes, Königstein: Hain, 1981 N. Bohr (1928) "The Q u a n t u m Postulate and the Recent Development of Atomic Theory," Nature 121 (1928) (Supplement), 5 8 0 - 5 9 0 D. A. Bonevac (1982) Reduction in the Abstract Sciences, Indianapolis/Cambridge: Hackett, 1982

Bibliography

289

W. Bräutigam/P. Christian (1981) Psychosomatische Medizin. Ein kurzgefaßtes Lehrbuch, Stuttgart: Thieme, 3 1981 H. Breger (1984) "Elastizität als Strukturprinzip der Materie bei Leibniz," in: Leibniz' Dynamica. Symposion der Leibniz-Gesellschaft in der Evangelischen Akademie Loccum, 2. bis 4. Juli 1982, ed. A. Heinekamp, Stuttgart: Steiner, 1984 (Studia Leibnitiana. Sonderheft 13), 112 — 121 J. W. Brehm/A. R. Cohen (1962) Explorations in Cognitive Dissonance, New York/London: Wiley, 1962 F. Brentano (1874) Psychology from an Empirical Standpoint, ed. O. Kraus, trans. A. C. Rancurello/D. B. Terrell/L. L. McAlister, London: Routledge & Kegan Paul, 1973 J. Brozek/H. Gundlach (1988) G. T. Fechner and Psychology (International Gustav Theodor Fechner Symposium Passau 12 to 14 June 1987), ed. J. Brozek/H. Gundlach, Passau: Passavia Universitätsverlag, 1988 (Passauer Schriften zur Psychologiegeschichte 6) B. Brundell (1987) Pierre Gassendi. From Aristotelianism to a New Natural Philosophy, Dordrecht etc.: Reidel, 1987 J. S. Bruner (1957) "On Perceptual Readiness," in: Readings in Perception, ed. D. C. Beardslee/ M . Wertheimer, Princeton etc.: Van Nostrand, 1958, 6 8 6 - 7 2 9 M . Bunge (1980) The Mind-Body Problem. A Psychobiological Approach, Oxford etc.: Pergamon Press, 1980 R. Carnap (1928) "Pseudoproblems in Philosophy," in: R. Carnap, The Logical Structure of the World and Pseudoproblems in Philosophy (trans. R. A. George), Berkeley: University of California Press, 1967, 301—343 R. Carnap (1931) "The Elimination of Metaphysics Through Logical Analysis of Language," (trans. A. Pap), in: Ayer, 1959, 6 0 - 8 1 R. Carnap (1931/1932) "Über Protokollsätze," Erkenntnis 3 (1932/1933), 2 1 5 - 2 2 8 (repr. in: Schleichen, 1975, 8 1 - 9 4 ) R. Carnap (1932/1933) "Psychology in Physical Language" (trans. G. Schick), in: Ayer, 1959, 165-198 R. Carnap (1956) "The Methodological Character of Theoretical Concepts," in: Feigl/Scriven, 1956, 3 8 - 7 6 M . Carrier (1986a) "Empirische Wissenschaft und methodologische Normen," in: Wahrheit, Richtigkeit und Exaktheit, ed. J. Klein/H. D. Erlinger, Essen: Die blaue Eule, 1986 (Siegener Studien 40), 4 1 - 6 3

290

Bibliography

M . Carrier (1986b) "Wissenschaftsgeschichte, rationale Rekonstruktion und die Begründung von Methodologien," Zeitschrift für allgemeine Wissenschaftstheorie 17 (1986), 2 0 1 - 2 2 8 M . Carrier (1986c) "Die begriffliche Entwicklung der Affinitätstheorie im 18. Jahrhundert. Newtons Traum — und was daraus wurde," Archive for History of Exact Sciences 36 (1986), 3 2 7 - 3 8 9 M . Carrier (1988) " O n Novel Facts. A Discussion of Criteria for Non-ad-hoc-ness in the Methodology of Scientific Research Programmes," Zeitschrift für allgemeine Wissenschaftstheorie 19 (1988), 2 0 5 - 2 3 1 M . Carrier (1989) "Circles without Circularity. Testing Theories by Theory-Laden Observations," in: An Intimate Relation, ed. J. Brown/J. Mittelstrass, Dordrecht/ Boston: Reidel, 1989, 4 0 5 - 4 2 8 M . Carrier (1991a) "Geist und Computer. Formale Operationen als Vorbild mentaler Prozesse," in: W. C. Zimmerli (ed.), Das Problem des künstlichen Gehirns (forthcoming) M. Carrier (1991b) "On the Disunity of Science or Why Psychology is not a Branch of Physics," in: Einheit der Wissenschaften (Internationales Kolloquium der Akademie der Wissenschaften zu Berlin, Bonn, 2 5 . - 2 7 . Juni 1990), ed. J. Mittelstrass, Berlin: de Gruyter (forthcoming) R. Chisholm (1955/1956) "Sentences about Believing," Proceedings of the Aristotelian Society 56 (1955/1956), 1 2 5 - 1 4 8 P. M. Churchland (1981) "Elimiriative Materialism and the Propositional Attitudes," The Journal of Philosophy 78 (1981), 67 - 90 P. M. Churchland (1985) "Reduction, Qualia, and the Direct Introspection of Brain States," The Journal of Philosophy 82 (1985), 8 - 2 8 P. S. Churchland (1986) Neurophilosophy. Toward a Unified Science of the Mind-Brain, Cambridge Mass./London: M I T Press, 1986 L. J. Cohen (1979) Review of Popper/Eccles, 1977, Mind 88 (1979), 3 0 1 - 3 0 4 J. W. Cornman (1962) "The Identity of Mind and Body," The Journal of Philosophy 59 (1962), 486-492 O. Creutzfeldt (1981) "Bewußtsein und Selbstbewußtsein als neurophysiologisches Problem der Philosophie," in: Reproduktion des Menschen. Beiträge zu einer interdisziplinären Anthropologie, Frankfurt/Berlin/Wien: Ullstein, 1981 (Schriften der Carl-Friedrich-von-Siemens-Stiftung 5), 29 — 44 (Discussion 44 — 54)

Bibliography

291

O. Creutzfeldt (1983) "Das Bewußtsein entzieht sich der Wissenschaft (Protocol of discussion)," MPG-Spiegel 6 (1983), 8 - 1 0 L. J. Cronbach/P. E. Meehl (1956) "Construct Validity in Psychological Tests," in: Feigl/Scriven, 1956, 174 — 204 J. P. Crutchfield et al. (1986) "Chaos," Scientific American 255 (1986), No. 6, 3 8 - 4 9 L. J. Daston (1982) "The Theory of Will versus the Science of Mind," in: The Problematic Science. Psychology in Nineteenth-Century Thought, ed. W. R. Woodward/ M . G. Ash, New York: Praeger, 1982, 8 8 - 1 1 5 D. Davidson (1970) "Mental Events," in: Davidson, 1980, 2 0 7 - 2 2 7 D. Davidson (1973) "The Material Mind," in: Davidson, 1980, 2 4 5 - 2 5 9 D. Davidson (1974) "Psychology as Philosophy," in: Davidson, 1980, 229 — 239 D. Davidson (1976) "Hempel on Explaining Action," in: Davidson, 1980, 261—275 D. Davidson (1980) Essays on Actions and Events, Oxford: Clarendon Press, 1980 D. C. Dennett (1979) Review of Popper/Eccles, 1977, The Journal of Philosophy 76 (1979), 9 1 - 9 7 D. C. Dennett (1982) "Comments on Rorty," Synthese 53 (1982), 3 4 9 - 3 5 6 R. Descartes (1641) The Philosophical Writings of Descartes II (trans. J. Cottingham/ R. Stroothoff/D. Murdoch), Cambridge: Cambridge University Press, 1984 R. Descartes (1897-1910) Oeuvres, I - X I I , ed. Ch. Adam/P. Tannery, Paris: Vrin, 1 8 9 7 - 1 9 1 0 Diog. Laert. Diogenis Laertii Vitae Philosophorum, I— II, ed. H. S. Long, Oxford: Clarendon Press, 1964 F. I. Dretske (1980) "The Intentionality of Cognitive States," Midwest Studies in Philosophy 5 (1980), 2 8 1 - 2 9 4 F. I. Dretske (1983) "Precis of Knowledge and the Flow of Information," The Behavioral and Brain Sciences 6 (1983), 5 5 - 9 0 E. Du Bois-Reymond (1880) Über die Grenzen des Naturerkennens [1872], Die sieben Welträtsel [1880], Leipzig: Veit, 1907 C. Ducasse (1960) "In Defense of Dualism," in: Dimensions of Mind. A Symposium, ed. S. Hook, New York, London: Collier Books, Collier-Macmillan, 1960, 1966, 85-89

292

Bibliography

P. Duhem (1908) To Save the Phenomena. An Essay on the Idea of Physical Theory from Plato to Galileo, trans. E. Doland/C. Maschler, Chicago/London: Chicago University Press, 1969 I. During (1966) Aristoteles. Darstellung und Interpretation seines Denkens, Heidelberg: Winter, 1966 R. A. Eberle (1971) "Replacing one Theory by Another under Preservation of a Given Feature," Philosophy of Science 38 (1971), 4 8 6 - 5 0 1 J. C. Eccles (1970) Facing Reality. Philosophical Adventures by a Brain Scientist, N e w York/ Heidelberg/Berlin: Springer, 1970 A. Einstein (1934) The World as I see it (trans. A. Harris), London: Watts 1935 (The Thinker's Library 79) A. Einstein (1949) "Autobiographisches/Autobiographical Notes," in: Albert Einstein Philosopher — Scientist, ed. P. A. Schilpp, Evanston 111.: Cambridge University Press, 1949, 3 1970, 2 - 9 4 Empedocles Empedocles. The Extant Fragments (ed. M . R. Wright), New Haven/London: Yale University Press 1981 J. D. Farmer (1982) "Dimension, Fractal Measures, and Chaotic Dynamics," in: Evolution of Order and Chaos in Physics, Chemistry, and Biology (Proceedings of the International Symposium on Synergetics at Schloss Elmau, Bavaria, April 26 —May 1, 1982), ed. H. Haken, Berlin/Heidelberg/New York: Springer, 1982, 2 2 8 - 2 4 6 G. T. Fechner (1860) Elemente der Psychophysik I, Leipzig 1860 (repr. Amsterdam: Bonset, 1964) H. Feigl (1950) "Existential Hypotheses. Realistic versus Phenomenalistic Interpretations," Philosophy of Science 17 (1950), 3 5 - 6 2 H. Feigl (1951) "Principles and Problems of Theory Construction in Psychology," in: Current Trends in Psychological Theory, ed. W. Dennis et al., Pittsburgh: University of Pittsburgh Press, 1951, 1 7 9 - 2 1 3 H. Feigl (1958) "The 'Mental' and the 'Physical'," in: Concepts, Theories, and the MindBody Problem, ed. H. Feigl/M. Scriven/G. Maxwell, Minneapolis: University of Minnesota Press, 1958 (Minnesota Studies in the Philosophy of Science II), 3 7 0 - 4 9 7 H. Feigl (1963) "Physicalism, Unity of Science and the Foundations of Psychology," in: Herbert Feigl. Inquiries and Provocations. Selected Writings 1929 — 1974, ed. R. S. Cohen, Dordrecht/Boston/London: Reidel, 1981, 3 0 2 - 3 4 1

Bibliography

293

H. Feigl/M. Scriven (1956) The Foundations of Science and the Concepts of Psychology and Psychoanalysis, ed. H. Feigl/M. Scriven, Minneapolis: University of Minnesota Press, 1956 (Minnesota Studies in the Philosophy of Science 1) P. K. Feyerabend (1962) "Explanation, Reduction, and Empiricism," in: Scientific Explanation, Space, and Time, ed. H. Feigl/G. Maxwell, Minneapolis: University of Minnesota Press, 1962 (Minnesota Studies in the Philosophy of Science 111), 28-97 P. K. Feyerabend (1963) "Materialism and the Mind-Body Problem," The Review of Metaphysics 17 (1963/1964), 4 9 - 6 6 P. K. Feyerabend/G. Maxwell (1966) Mind, Matter, and Method. Essays in Philosophy and Science in Honor of Herbert Feigl, ed. P. K. Feyerabend/G. Maxwell, Minneapolis: University of Minnesota Press, 1966 J. A. Fodor (1968) Psychological Explanation. An Introduction to the Philosophy of Psychology, New York: Random, 1968 J. A. Fodor (1975) The Language of Thought, New York: Crowell, 1975 J. A. Fodor (1981) Representations. Philosophical Essays on the Foundations of Cognitive Science, Cambridge Mass.: M I T Press, 1981 J. A. Fodor (1987) Psychosemantics. The Problem of Meaning in the Philosophy of Mind, Cambridge Mass.: M I T Press, 1987 G. Frege (1893) The Basic Laws of Arithmetic (trans. M . Furth), Berkeley/London: University of California Press, 1967 G. Frege (1918/1919) "Thoughts," in: G. Frege, Logical Investigations (trans. P. T. Geach/R. H. Stroothoff), Oxford: Blackwell, 1977, 1 - 3 0 G. Frege (1923 - 1926) "Compound Thoughts," in: G. Frege, Logical Investigations (trans. P. T. Geach/R. H. Stroothoff), Oxford: Blackwell, 1977, 5 5 - 7 7 H. R. Ganslandt (1980) Article: "Dritte Welt," in: Mittelstrass, 1980, 4 9 9 - 5 0 0 P. Gassendi (1658) Syntagma philosophicum, I — II, Lyon 1658 M . S. Gazzaniga/J. E. LeDoux (1978) The Integrated Mind, New York/London: Plenum Press, 1978 C. F. Gethmann (1984) Article: "Intentionalität," in: Mittelstrass, 1984, 2 5 9 - 2 6 4

294

Bibliography

A. Geulincx (1893) Opera philosophica 111, ed. J . P.M. Land, The Hague 1893 A. Gierer (1985) Die Physik, das Leben und die Seele, Munich/Zurich: Piper, 3 1986 J. Glatzel (1973) Endogene Depressionen. Zur Psychopathologie, Klinik und Therapie zyklothymer Werstimmungen, Stuttgart: Thieme, 1973 P. Gohlke (1958) Aristoteles. Über den Himmel. Vom Werden und Vergehen, Paderborn: Schöningh, 1958 S. Grossmann (1983) "Chaos — Unordnung und Ordnung in nichtlinearen Systemen," Physikalische Blätter 39 (1983), 1 3 9 - 1 4 5 A. Grünbaum (1987) Psychoanalyse in wissenschaftstheoretischer Sicht. Zum Werk Sigmund Freuds und seiner Rezeption, Konstanz: Universitätsverlag, 1987 J . Gruter (1625) Bibliotheca exulum seu Enchiridion divinae humanaeque prudentiae, Frankfurt 1625 N. R. Hanson (1958) Patterns of Discovery. An Inquiry into the Conceptual Foundations of Science, Cambridge: Cambridge University Press, 1965 (Cambridge 1958) W. F. R. Hardie (1964) "Aristotle's Treatment of the Relation between the Soul and the Body," The Philosophical Quarterly 14 (1964), 5 3 - 7 2 E. Hartman (1977) Substance, Body, and Soul. Aristotelian Investigations, Princeton: Princeton University Press, 1977 H. Heckhausen (1967) The Anatomy of Achievement Motivation (trans. K. F. Butler/ R. C. Birney/ D. C. McClelland), New York/London: Academic Press, 1967 H. Heckhausen (1974a) "Faktoren des Entwicklungsprozesses," in: Weinert, 1974a, 101 —132 H. Heckhausen (1974b) "Motive und ihre Entstehung," in: Weinert, 1974a, 133 - 1 7 1 H. Heckhausen (1974c) "Lehrer-Schüler-Interaktion," in: Weinert, 1974a, 547 — 573 M. Heidelberger (1988) "Fechners Leib-Seele-Theorie," in: Brozek/Gundlach, 1988, 6 1 - 7 7 J. van Heijenoort (1967) Article: "Gödel's Theorem," in: The Encyclopedia of Philosophy III, ed. P. Edwards, New York, London: Macmillan, Collier-Macmillan, 1967, 348-357 H. Heimsoeth (1966) Transzendentale Dialektik. Ein Kommentar zu Kants Kritik der reinen Vernunft. Erster Teil: Ideenlehre und Paralogismen, Berlin: de Gruyter, 1966

Bibliography

295

C. G. Hempel (1952) Fundamentals of Concept Formation in Empirical Science, Chicago: University of Chicago Press, 1952 (International Encyclopedia of Unified Science II, No. 7) C. G. Hempel (1958) "The Theoretician's Dilemma: A Study in the Logic of Theory Construction," in: Hempel, Aspects of Scientific Explanation And Other Essays in the Philosophy of Science, New York, London: Free Press, Collier-Macmillan, 1965, 173-226 C. G. Hempel/P. Oppenheim (1948) "Studies in the Logic of Explanation," Philosophy of Science 15 (1948), 135 - 1 7 5 H. Hertz (1894) The Principles of Mechanics. Presented in a New Form (trans. D. E. Jones/ J. T. Walley), New York: Dover, 1956 E. R. Hilgard/G. H. Bower (1966) Theories of Learning, New York: Appleton-Century Crofts, 31966 Th. Hobbes (1839-1845) The English Works of Thomas Hobbes, I —XI, ed. G. Molesworth, London 1839-1845 T. Horgan/J. Woodward (1985) "Folk Psychology is here to Stay," The Philosophical Review 94 (1985), 197-226 D. Hume (1748) An Inquiry Concerning Human Understanding, ed. C. W. Hendel, Indianapolis/New York: Bobbs-Merrill, 1955 G. M. K. Hunt (1987) "Determinism, Predictability and Chaos," Analysis 47 (1987), 129 — 133 Th. H. Huxley (1893) Collected Essays (1893-1894) 1 (Method and Results), London 1893 (repr. Hildesheim/New York: Olms, 1970) T. Irwin (1988) Aristotle's First Principles, Oxford: Clarendon Press, 1988 P. Janich/J. Mittelstrass (1973) "Raum," in: Handbuch philosophischer Grundbegriffe II, ed. H. Krings/ H. M. Baumgartner/C. Wild, München: Kösel, 1973, 1154-1168 P. Janich/K. Mainzer (1980) Article: "Gesetz (exakte Wissenschaften)," in: Mittelstrass 1980, 7 6 1 - 7 6 3 F. Kambartel (1965) "The universe is more various, more Hegelian. Zum Weltverständnis bei Hegel und Whitehead," in: Collegium Philosophicum. Studien Joachim Ritter zum 60. Geburtstag, Basel/Stuttgart: Schwabe, 1965, 72 — 98 I. Kant (1902ff.) Gesammelte Schriften, ed. Königlich Preussische (afterwards: Deutsche) Akademie der Wissenschaften (zu Berlin), Berlin: Reimer, 1902ff.

296

Bibliography

F. Kaulbach (1969) Immanuel Kant, Berlin: de Gruyter, 1969 (Sammlung Göschen 536/536a) A. Kenny (1980) Aquinas, Oxford: Oxford University Press, 1980 J. Kim (1985) "Psychophysical Laws," in: LePore/McLaughlin 1985, 3 6 9 - 3 8 6 A. Koyré (1939) Galileo Studies, Hassocks: The Harvester Press, 1978 (Original: Etudes Galiléennes, I —III, Paris: Hermann, 1939 (Actualités scientifiques et industrielles 852 — 854), in one volume: Paris: Hermann, 1966) S. A. Kripke (1971) "Identity and Necessity," in: Identity and Individuation, ed. M. K. Munitz, New York: New York University Press, 1971, 135 - 1 6 4 S. A. Kripke (1972) Naming and Necessity, Cambridge Mass.: Blackwell, 1980 H. Kuhlenbeck (1982) The Human Brain and Its Universe, I —III, ed. J. Gerlach, Basel etc.: Karger, 1982 (2., revised and expanded edition of: Brain and Consciousness, 1957) W. Kullmann (1974) Wissenschaft und Methode. Interpretationen zur Aristotelischen Theorie der Naturwissenschaft, Berlin/New York: de Gruyter, 1974 F. v. Kutschera (1982) Grundfragen der Erkenntnistheorie, Berlin/New York: de Gruyter, 1982 I. Lakatos (1970) "Falsification and the Methodology of Scientific Research Programmes," in: Criticism and the Growth of Knowledge (Proceedings of the International Colloquium of Philosophy of Science, London, 1965, 4), ed. I. Lakatos/ A. Musgrave, Cambridge: Cambridge University Press, 1970, 91 —195 I. Lakatos (1971) "History of Science and Its Rational Reconstructions," in: PSA 1970. In Memory of Rudolf Carnap (Proceedings of 1970 Biennial Meeting. Philosophy of Science Association), ed. R. C. Buck/R. S. Cohen, Dordrecht: Reidel, 1971 (Boston Studies in the Philosophy of Science VIII), 91 — 135 L. Laudan (1977) Progress and Its Problems. Toward a Theory of Scientific Growth, Berkeley/ Los Angeles/London: Routledge, 1977 G. R. Lefrancois (1972) Psychological Theories and Human Learning: Kongor's Report, Belmont Cai.: Brooks Cole, 1972 G. W. Leibniz (1676) "De Arcanis Motus et Mechanica ad puram Geometriam reducenda" (LH 35; XIII,3, Bl. 81 (Niedersächsische Landesbibliothek Hannover)). Published in: H.-J. Hess, Die unveröffentlichten naturwissenschaftlichen und technischen Arbeiten von G. W. Leibniz aus der Zeit seines Parisaufenthaltes. Eine Kurzcharaktistik, in: Leibniz à Paris (1672 — 1676). Symposion de la G. W. Leibniz-Gesellschaft (Hannover) et du Centre National de la Recherche

Bibliography

297

Scientifique (Paris) à Chantilly (France) du 14 au 18 Novembre 1976 I: Les Sciences, Wiesbaden: Steiner, 1978 (Studia Leibnitiana Supplementa XVII), 202 - 205 G. W. Leibniz (1686) Discours de Métaphysique/Metaphysische Abhandlung, ed./trans. H. Herring, Hamburg: Meiner, 1958 G. W. Leibniz (1714) Principes de la Nature et de la Grâce fondés en Raison. Monadologie/ Vernunftprinzipien der Natur und der Gnade. Monadologie, ed. H. Herring, Hamburg: Meiner, 1956 G. W. Leibniz (1849 - 1863) Mathematische Schriften, I-VII, ed. C. I. Gerhardt, Berlin/Halle 1 8 4 9 - 1 8 6 3 (repr. Hildesheim: Olms, 1962) G. W. Leibniz (1875-1890) Die philosophischen Schriften, I — VII, ed. C. I. Gerhardt, Berlin/Leipzig 1 8 7 5 - 1 8 9 0 (repr. Hildesheim: Olms, 1 9 6 0 - 1 9 6 1 , 1978) G. W. Leibniz (1969) Philosophical Papers and Letters, ed./trans. L. E. Loemker, Dordrecht: Reidei, 2 1969 G. W. Leibniz/S. Clarke (1715/1716) Correspondance Leibniz-Clarke présentée d'après les manuscrits originaux des bibliothèques de Hanovre et de Londres, ed. A. Robinet, Paris: Presses universitaires de France, 1957 J. Leplin (1984) "Introduction," in: Leplin (ed.), Scientific Realism, Berkeley/Los Angeles/ London: University of California Press, 1984, 1—7 E. LePore/B. P. McLaughlin (1985) Actions and Events. Perspectives on the Philosophy of Donald Davidson, ed. E. LePore/B. P. McLaughlin, O x f o r d / N é w York: Blackwell, 1985 G. A. Lienert (1969) Testaufbau und Testanalyse, Weinheim/Berlin/Basel: Beltz, 3 1969 E. N. Lorenz (1963) "Deterministic Nonperiodic Flow," Journal of the Atmospheric Sciences 20 (1963), 1 3 0 - 1 4 1 K. Lorenz (1970) Elemente der Sprachkritik. Eine Alternative zum Dogmatismus und Skeptizismus in der Analytischen Philosophie, Frankfurt: Suhrkamp, 1970 K. Lorenz (1975) "Die Monadologie als Entwurf einer Hermeneutik," in: Akten des II. Internationalen Leibniz-Kongresses Hannover, 17.-22. Juli 1972, III (Metaphysik — Ethik — Ästhetik — Monadenlehre), Wiesbaden: Steiner, 1975 (Studia Leibnitiana Supplementa XIV), 317 — 325 K. Lorenz (1980) Artide: "Ereignis," in: Mittelstrass, 1980, 568 K. Lorenz (1984a) Artide: "Identität," in: Mittelstrass, 1984, 189 - 1 9 2

298

Bibliography

K. Lorenz (1984b) Article: "intensional/Intension," in: Mittelstrass, 1984, 256 — 257 K. Lorenz (1984c) Article: "Kategorienfehler," in: Mittelstrass, 1984, 3 7 0 - 3 7 1 Ko. Lorenz (1973) Behind the Mirror. A Search for a Natural History of Human Knowledge, (trans. R. Taylor), London: Methuen, 1977 P. Lorenzen (1969) Normative Logic and Ethics, Mannheim/Zurich: Bibliographisches Institut, 1969 P. Lorenzen (1974) Konstruktive Wissenschaftstheorie, Frankfurt: Suhrkamp, 1974 E. Mach (1883) The Science of Mechanics. A Critical and Historical Account of Its Development, (introd. K. Menger, trans. T. J. McCormach), La Salle 111.: Open Court, 1960, 31974 D. M. MacKay (1978) "Selves and Brains," Neuroscience 3 (1978), 599 — 606 N. Malcolm (1971) "The Myth of Cognitive Processes and Structures," in: Cognitive Development and Epistemology, ed. T. Mischel, New York/London: Academic Press, 1971, 3 8 5 - 3 9 2 N. Malebranche (1959) Oeuvres X, ed. H. Gouhier/A. Robinet, Paris: Uria, 1959 Marcus Aurelius (1979) Ad se ipsum libri XII, ed. J. Dalfen, Leipzig: Teubner, 1979 A. Marras (1987) "Critical Notice: Stephen Stich's From Folk Psychology to Cognitive Science: The Case Against Belief;' Philosophy of Science 54 (1987), 1 1 5 - 1 2 7 G. Mayer-Kress (1986) Dimensions and Entropies in Chaotic Systems. Quantification of Complex Behavior (Proceedings of an International Workshop at the Pecos River Ranch, New Mexico, September 11 —16, 1985), ed. G. Mayer-Kress, Berlin/ Heidelberg/New York/Tokyo: Springer, 1986 J. W. McAllister (1986) "Theory-Assessment in the Historiography of Science," The British Journal for the Philosophy of Science 37 (1986), 3 1 5 - 3 3 3 C. McGinn (1980) "Philosophical Materialism," Synthese 44 (1980), 173-206 B. P. McLaughlin (1985) "Anomalous Monism and the Irreducibility of the Mental," in: LePore/ McLaughlin, 1985, 3 3 1 - 3 6 8 B. P. McLaughlin (1987) "What is Wrong with Correlational Psychosemantics?" Synthese 70 (1987), 271-286

Bibliography

299

P. E. Meehl (1966) "The Compleat Autocerebroscopist: A Thought-Experiment on Professor Feigl's Mind-Body Identity Thesis," in: Feyerabend/Maxwell, 1966, 103 - 1 8 0 P. Meehl/W. Sellars (1956) "The Concept of Emergence," in: Feigl/Scriven, 1956, 239 — 252 W. P. Mendon ? a (1987) "Zur Möglichkeit kognitiver Psychologie aus Wittgensteinscher Sicht," Zeitschrift für allgemeine Wissenschaftstheorie 18 (1987), 183—203 J. Mittelstrass (1962) Die Rettung der Phänomene. Ursprung und Geschichte eines antiken Forschungsprinzips, Berlin: de Gruyter, 1962 J. Mittelstrass (1970) Neuzeit und Aufklärung. Studien zur Entstehung der neuzeitlichen Wissenschaft und Philosophie, Berlin/New York: de Gruyter, 1970 J . Mittelstrass (1980/1984) Enzyklopädie Philosophie und Wissenschaftstheorie, I and II, ed. J. Mittelstrass, Mannheim/Wien/Zürich: Bibliographisches Institut, 1980, 1984 J . Mittelstrass (1984a) Article: "Impetustheorie," in: Mittelstrass, 1984, 2 1 2 - 2 1 3 J. Mittelstrass (1984b) Article: "Leib-Seele-Problem," in: Mittelstrass, 1984, 5 8 1 - 5 8 4 J . Mittelstrass (1984c) Article: "Nus," in: Mittelstrass, 1984, 1 0 4 4 - 1 0 4 5 J. Mittelstrass (1984d) Article: "Okkasionalismus," in: Mittelstrass, 1984, 1 0 6 7 - 1 0 6 8 J. Mittelstrass (1984e) "Scientific Rationality and Its Reconstruction," in: Reason and Rationality in Natural Science. A Group of Essays, ed. N. Rescher, Lanham/New York/ London: University Press of America, 1985, 83 — 102 J . Mittelstrass (1988) "Die Philosophie der Wissenschaftstheorie. Über das Verhältnis von Wissenschaftstheorie, Wissenschaftsforschung und Wissenschaftsethik," Zeitschrift für allgemeine Wissenschaftstheorie 19 (1988), 308 — 327 E. Nagel (1961) The Structure of Science. Problems in the Logic of Scientific Explanation, London: Routledge, 1961 T. Nagel (1974) "What is it like to be a Bat?," The Philosophical Review 83 (1974), 435 - 450 N. J . Nersessian (1984) Faraday to Einstein: Constructing Meaning in Scientific Theories, Dordrecht/Boston/Lancaster: Nijhoff, 1984 I. Newton (1687) Philosophiae naturalis principia mathematica, London 1687 (repr. London: W. Dowson, 1960, Brussels: Culture et civilisation, 1966)

300

Bibliography

J. S. Nicolis (1983) "The Role of Chaos in Reliable Information Processing," in: Synergetics of the Brain (Proceedings of the International Symposium on Synergetics at Schloss Elmau, Bavaria, May 2-7, 1983), ed. E. Ba§ar et al., Berlin/Heidelberg/New York/Tokyo: Springer, 1983, 3 3 0 - 3 4 4 F. Nietzsche (1889) Ecce homo. Wie man wird, was man ist [1889], Werke. Kritische Gesamtausgabe VI/3, ed. G. Colli/M. Montinari, Berlin: de Gruyter, 1969, 2 5 3 - 3 7 2 D. Papineau (1979) Theory and Meaning, Oxford: Clarendon Press, 1979 D. Papineau (1987) Reality and Representation, Oxford: Blackwell, 1987 G. Patzig (1973) "Erklären und Verstehen. Bemerkungen zum Verhältnis von Natur- und Geisteswissenschaften," in: Patzig, Tatsachen, Normen, Sätze. Aufsätze und Vorträge, Stuttgart: Reclam, 1980, 45 — 75 (Original in: Neue Rundschau 84/3 (1973), 3 9 2 - 4 1 3 ) Plotinus (1964-1982) Opera, I — 11I, ed. P. Henry/H.-R. Schwyzer, Oxford: Clarendon Press, 1964-1982 Plutarchus (1960) Moralia V 3, ed. C. H u b e r t / M . Pohlenz/H. Drescher, Leipzig: Teubner, 1960 E. Pöppel (1985) Grenzen des Bewusstseins. Über Wirklichkeit und Welterfahrung, Stuttgart: Deutsche Verlagsanstalt, 1985 E. Pöppel (1986) "Materialien des Bewusstseins," Physikalische Blätter 42 (1986), No. 3, 75-77 H. Poincaré (1908) Science and Method (trans. F. Maitland), New York: Dover, 1952 E. P. Polten (1973) Critique of the Psycho-Physical Identity Theory. A Refutation of Scientific Materialism and an Establishment of Mind-Matter Dualism by Means of Philosophy and Scientific Method, The Hague/Paris: M o u t o n , 1973 K. R. Popper (1972) Objective Knowledge. An Evolutionary Approach, Oxford: Clarendon Press, 1972 K. R. Popper/J. C. Eccles (1977) The Self and Its Brain, Berlin/London/New York: Springer, 1977, 1981 I. Prigogine (1980) From Being to Becoming. Time and Complexity in the Physical Sciences, San Francisco: Freeman, 1980 M . Prince (1904) "The Identification of Mind and Matter," The Philosophical Review 13 (1904), 4 4 4 - 4 5 1

Bibliography

301

R. Puccetti (1985) "Popper and the Mind-Body Problem," in: Popper and the Human Sciences, ed. G. Currie/A. Musgrave, Dordrecht/Boston/Lancaster: Nijhoff, 1985, 45-55 H. Putnam (1963) "Brains and Behavior," in: Putnam, 1975, 3 2 5 - 3 4 1 H. Putnam (1967a) "The Mental Life of Some Machines," in: Putnam, 1975, 4 0 8 - 4 2 8 H. Putnam (1967b) "The Nature of Mental States," in: Putnam, 1975, 429 - 440 H. Putnam (1975) Mind, Language and Reality (Philosophical Papers 2), Cambridge etc.: Cambridge University Press, 1975 Z. W. Pylyshyn (1984) Computation and Cognition. Toward a Foundation for Cognitive Science, Cambridge Mass.: M I T Press, 1984, 1986 W. V. O. Quine (1953) From a Logical Point of View. Logico-Philosophical Essays, Cambridge Mass.: Harvard University Press, 2 1964 W. V. O. Quine (1960) Word and Object, Cambridge Mass.: M I T Press, 1960 W. V. O. Quine (1969) "Existence and Quantification," in: Quine, Ontological Relativity and Other Essays, New York/London: Columbia University Press, 1969, 91 — 113 R. Reininger (1923) Kant. Seine Anhänger und seine Gegner, Munich: Reinhardt, 1923 B. Rensch (1979) Gesetzlichkeit, psychophysischer Zusammenhang, Willensfreiheit und Ethik, Berlin: Dunker, 1979 T. Rentsch (1980) Article: "Leib-Seele-Verhältnis III," in: Historisches Wörterbuch der Philosophie V, ed. J. Ritter/K. Gründer, Basel/Stuttgart: Schwabe, 1980, 2 0 1 - 2 0 6 N. Rescher (1970) Scientific Explanation, New York, London: Free Press, 1970 N. Rescher (1979) Cognitive Systematization. A Systems-Theoretic Approach to a Coherentist Theory of Knowledge, Oxford: Blackwell, 1979 N. Rescher (1984) The Limits of Science, Berkeley/Los Angeles/London: University of California Press, 1984 R. Rorty (1965) "Mind-Body Identity, Privacy, and Categories," The Review of Metaphysics 19 (1965), 2 4 - 5 4 R. Rorty (1982) "Contemporary Philosophy of Mind," Synthese 53 (1982), 3 2 3 - 3 4 8

302

Bibliography

R. Rosenthal/L. Jacobson (1968) Pygmalion in the Classroom. Teacher Expectation and Pupils' Intellectual Development, New York etc.: Holt, 1968 F. L. Ruch/P. G. Zimbardo (1975) Lehrbuch der Psychologie. Eine Einführung für Studenten der Psychologie, Medizin und Pädagogik, Berlin/Heidelberg/New York: Springer, 2 1975 G. Ryle (1949) The Concept of Mind, London: Hutchinson, 1949 K. Sayre (1987) "Cognitive Science and the Problem of Semantic Content," Synthese 70 (1987), 2 4 7 - 2 6 9 S. Schachter/J. E. Singer (1962) "Cognitive, Social, and Psychological Determinants of Emotional State," Psychological Review 69 (1962), 3 7 9 - 3 9 9 K. F. Schaffner (1970) "Outlines of a Logic of Comparative Theory Evaluation with Special Attention to Pre- and Post-Relativistic Electrodynamics," in: Historical and Philosophical Perspectives of Science, ed. R. H. Stuewer, Minneapolis: University of Minnesota Press, 1970 (Minnesota Studies in the Philosophy of Science V), 3 1 1 - 3 5 4 H. Schleichen (1975) Logischer Empirismus — Der Wiener Kreis. Ausgewählte Texte mit einer Einleitung, ed. H. Schleichen, Munich: Fink, 1975 M. Schlick (1925) General Theory of Knowledge (trans. A. E. Blumberg), Wien/ New York: Springer, 1974 M. Schlick (1934) "The Foundations of Knowledge" (trans. D. Rynin), in: Ayer, 1959, 2 0 9 - 2 2 7 R. F. Schmidt (1987) "Motorische Systeme," in: Grundriss der Neurophysiologie, ed. R. F. Schmidt, Berlin/Heidelberg/New York: Springer, 6 1987, 1 5 7 - 2 0 4 K. Schneider (1976) "Leistungsmotiviertes Verhalten als Funktion von Motiv, Anreiz und Erwartung," in: Leistungsmotivation und Verhalten, ed. H.-D. Schmalt/W.-U. Meyer, Stuttgart: Klett, 1976, 3 3 - 5 9 H. G. Schuster (1984) Deterministic Chaos. An Introduction, Weinheim: Physik-Verlag, 1984 O. Schwemmer (1980) Article: "Erklärung," in: Mittelstrass, 1980, 5 7 8 - 5 8 4 J. R. Searle (1958) "Proper Names," Mind 67 (1958), 1 6 6 - 1 7 3 J. R. Searle (1980) "Minds, Brains, and Programs," The Behavioral and Brain Sciences 3 (1980), 417-457 W. Sellars (1953) "A Semantical Solution of the Mind-Body Problem," Methodos 5 (1953), 45-82

Bibliography

303

W. Seilars (1956) "Empiricism and the Philosophy of Mind," in: Feigl/Scriven, 1956, 2 5 3 - 3 2 9 W. Sellars (1963) Science, Perception and Reality, London, New York: Routledge, Humanities Press, 1963 A. D. Sertillanges (1954) Der heilige Thomas von Aquin, Cologne/Olten: Hegner, 21954 J. A. Shaffer (1967) Article: "Mind-Body Problem," in: The Encyclopedia of Philosophy V, ed. P. Edwards, New York, London: Macmillan, Collier-Macmillan, 1967, 336-346 B. F. Skinner (1953) Science and Human Behavior, New York: Macmillan 1953, 1966 J. J.C. Smart (1959) "Sensations and Brain Processes," The Philosophical Review 68 (1959), 141 - 1 5 6 R. Specht (1966) Commercium mentis et corporis. Über Kausalvorstellungen im Cartesianismus, Stuttgart-Bad Cannstatt: Frommann, 1966 R. Specht (1976) Article: "Influxus physicus, Influxionismus," in: Historisches Wörterbuch der Philosophie IV, ed. J. Ritter/K. Gründer, Basel/Stuttgart: Schwabe, 1976, 354-356 R. Specht (1980) Article: "Leib-Seele-Verhältnis," in: Historisches Wörterbuch der Philosophie V, ed. J. Ritter/K. Gründer, Basel/ Stuttgart: Schwabe, 1980, 1 8 6 - 2 0 1 R. Specht (1984) Article: "Occasionalismus," in: Historisches Wörterbuch der Philosophie VI, ed. J. Ritter/K. Gründer, Basel/Stuttgart: Schwabe, 1984, 1090-1091 B. Spinoza (1967) Opera II (Tractatus de intellectus emendatione. Ethica), ed. K. Blumenstock, Darmstadt: Wissenschaftliche Buchgesellschaft, 1967 B. Spinoza (1985) The Collected Works of Spinoza, ed./trans. E. Curley, Princeton: Princeton University Press, 1985 G. E. Stahl (1703) Einleitung zur Grund-Mixtion Derer Unterirrdischen mineralischen und metallischen Cörper [lat. 1703], Leipzig 1744 D. L. Stein (1989) "Spin Glasses", Scientific American 261 (1989), No. 1, 3 6 - 4 2 S. P. Stich (1983) From Folk Psychology to Cognitive Science. The Case Against Belief, Cambridge Mass./London: M I T Press, 1983 P. F. Strawson (1959) Individuals. An Essay in Descriptive Metaphysics, London: Methuen, 1959 D. W. Tank/J. L. Hopfield (1987) "Collective Computation in Neuronlike Circuits", Scientific American (1987), No. 6, 6 2 - 7 0

257

304

Bibliography

A. Tarski (1953) "A General Method in Proofs of Undecidability," in: Tarski/ A. Mostowski/ R. M. Robinson, Undecidable Theories, Amsterdam: North-Holland, 3 1971, 1-35 Theon von Smyrna (1878) Expositio rerum mathematicarum ad legendum Platonem utilium, ed. E. Hiller, Leipzig 1878 C. Thiel (1980) Article: "Frege," in: Mittelstrass, 1980, 6 7 1 - 6 7 4 Thomas Aquinas (1970) Summa theologiae, lat./engl., London: Blackfriars, 1970 G. Vollmer (1981) "Altehrwürdig, aber unhandlich: Popper und Eccles zum Leib-Seele-Problem," Allgemeine Zeitschrift für Philosophie 6 (1981), 6 0 - 7 0 VS H. Diels, Die Fragmente der Vorsokratiker. Greek and German (Berlin 1903), I —III, 5th Ed., ed. W. Kranz, Berlin: Weidmann, 6 1951 (reprinted without changes, 171974) E. L. Walker/R. W. Heyns (1962) An Anatomy for Conformity, Englewood Cliffs: Prentice-Hall, 1962, 1967 (Chapter 6 ["Conformity and Conflict of Needs," 5 4 - 6 8 ] ) B. Weiner (1972) Theories of Motivation. From Mechanism to Cognition, Chicago: Markham, 1972 F. E. Weinert (1974) "Einführung in das Problemgebiet der Pädagogischen Psychologie," in: Weinert, 1974a, 2 9 - 6 3 F. E. Weinert et al. (1974a) Pädagogische Psychologie 1 (Funk-Kolleg), ed. F. E. Weinert et al., Frankfurt: Fischer, 1974 A. N. Whitehead (1947) Essays in Science and Philosophy, New York: Greenwood Press, 1947, 1968 William of Ockham (1981/1980) Opera theologica V (Quaestiones in librum secundum sententiarum (Reportatio, ed. G. Gál/R. Wood)/ IX (Quodlibeta Septem, ed. J. C. Wey), St. Bonaventure: Institutum Franciscanum, 1981/1980 K. V. Wilkes (1980) "Brain States," The British Journal for the Philosophy of Science 31 (1980), 111-129 K. V. Wilkes (1981) "Functionalism, Psychology, and the Philosophy of Mjnd," Philosophical Topics 12 (1981), 1 4 7 - 1 6 7 W. Windelband (1957) Lehrbuch der Geschichte der Philosophie, ed. H. Heimsoeth, Tübingen: Mohr, 151957

Bibliography

305

L. Wittgenstein (1922) Tractatus logico-philosophicus (Germ./Engl.), London: Routledge 1922 L. Wittgenstein (1958) Philosophical Investigations, Oxford: Blackwell, 1958 Ch. Wolff (1736) Philosophia prima sive ontologia, methodo scientifica pertractata, Frankfurt/ Leipzig 2 1736 (repr. Hildesheim/New York: Olms, 1977) Ch. Wolff (1740) Psychologia rationalis, methodo scientifica pertractata, Frankfurt/Leipzig 2 1740 (repr. Hildesheim/New York: Olms, 1972) M . Wolff (1978) Geschichte der Impetustheorie. Untersuchungen zum Ursprung der klassischen Mechanik, Frankfurt: Suhrkamp, 1978 G. Wolters (1980) Article: "Goldbachsche Vermutung," in: Mittelstrass, 1980, 7 9 0 - 7 9 1 G. Wolters (1984a) Article: "Kant (I)," in: Mittelstrass, 1984, 3 4 3 - 3 5 1 G. Wolters (1984b) Article: "Korrespondenzregel," in: Mittelstrass, 1984, 4 8 0 - 4 8 1 G. Wolters (1984c) Article: "Modell," in: Mittelstrass, 1984, 9 1 1 - 9 1 3 G. Wolters (1987) Mach I, Mach II, Einstein und die Relativitätstheorie. Eine Fälschung und ihre Folgen, Berlin/New York: de Gruyter, 1987 G. Wolters (1988a) "Verschmähte Liebe. Mach, Fechner und die Psychophysik," in: Brozek/ Gundlach 1988, 1 0 3 - 1 1 6 G. Wolters (1988b) "Evolutionäre Erkenntnistheorie — eine Polemik," Vierteljahrsschrift der Naturforschenden Gesellschaft in Zürich 133 (1988), 1 2 5 - 1 4 2 G. Wolters/P. Schroeder-Heister (1984) Article: "Logik, intensionale," in: Mittelstrass, 1984, 666 — 667 G. H. v. Wright (1971) Explanation and Understanding, London: Routledge, 1971

Index of Names Albano, A. M . 262 Alexander of Aphrodisias 14 Alonso, M . 48 Aristotle 4, 10 - 1 5 , 19, 22 - 24, 28 - 29, 47, 49, 77, 93, 159 Armstrong, D. M. 90, 98 Arnauld, A. 18 Atkinson, J. W. 132 Augustine, A. 14, 16 Aune, B. 97, 181, 183 Averiii, E. 162 Babloyantz, A. 262 Baker, G. P. 274 Bandura, A. 1 3 2 - 1 3 4 , 244 Bechtel, W. 202, 206, 240 - 241 Bénard, H. 261 Benedetti, G. B. 4 7 - 4 9 Bergson, H. L. 278 Berkeley, G. 28, 3 0 - 3 1 Berkson, W. 175 Bernard, W. 11, 13 Bernath, K. 15 Bieri, P. 68, 126, 129, 168 Bohr, N . 120, 281 Boltzmann, L. 4 5 - 4 6 , 64, 131 Bonevac, D. A. 44 Bosses, B. des 149 Bower, G. H. 129, 133 Boyle, R. 131 Bräutigam, W. 157 Breger, H. 159 Brehm, J. W. 136 Brentano, F. 67 - 68, 70, 207 - 208, 223, 244 Brouwer, L. E. J. 116 Brundell, B. 28 Bruner, J. S. 188 Büchner, L. 31 Bunge, M . 121, 128, 1 5 6 - 1 5 7 , 179 Cabanis, G. 29 Carnap, R. 3 9 - 4 0 , 144, 148, 171, 1 8 2 - 183 Carrier, M . 1 0 9 - 1 1 0 , 139, 160, 165, 199, 202

Cavendish, H. 1 0 9 - 1 1 0 Chisholm, R. 68, 129, 2 0 7 - 2 0 8 , 223, 244 Christian, P. 157 Churchland, P. M. 51, 61, 75 - 76, 78, 122, 152, 184, 1 8 6 - 1 8 7 , 230 Churchland, P. S. 51, 59, 6 3 - 6 4 , 69, 7 6 - 7 8 , 222, 230 Clarke, S. 149, 160 Clauberg, J. 20 Cohen, A. R. 136 Cohen, L. J. 124 Copernicus, N. 165 Cordemoy, G. de 19 Cornman, J. W. 81 - 82, 84 - 85 Creutzfeld, O. 2 7 1 - 2 7 2 , 275 - 279, 282 Cronbach, L. J. 166 Crutchfield, J. P. 260, 262 Darwin, Ch. 50 Daston, L. J. 32 Davidson, D. 5, 85, 99 - 108, 111 - 113,129, 141, 168, 223 Democritus 11, 159 Dennett, D. C. 124, 279 Descartes, R. 1, 4, 9, 1 6 - 2 0 , 24, 2 7 - 3 0 , 32, 47, 116, 124, 131, 153, 178, 2 7 3 - 2 7 5 , 279 Dilthey, W. 105 Diogenes Laertius 14 Dretske, F. I. 232, 2 3 7 - 2 4 3 Du Bois-Reymond, E. 1 Ducasse, C. 160 Düring, I. 1 2 - 1 3 Duhem, P. 109 Dykes, P. W. 55 Eberle, R. A. 44 Eccles, J. C. 5 - 6 , 1 1 4 - 1 1 9 , 1 2 0 - 1 2 5 , 128 Eigen, M. 120 Einstein, A. 104, 109, 150, 153, 165 - 1 6 6 , 175, 178 - 1 7 9 Empedocles 159 Epicurus 14, 28 Euclid 1 9 8 - 1 9 9 , 215

308

Index of Names

Farmer, J. D. 262 Fechner, G. Th. 4, 13, 3 1 - 3 2 , 34, 37, 52 Feigl, H. 3 5 - 4 2 , 5 1 - 5 3 , 55, 7 4 - 7 5 , 78, 8 7 - 8 9 , 92, 1 2 0 - 1 2 2 , 126, 128, 129, 143 - 1 4 5 , 146 - 1 4 7 , 152, 158, 160 - 1 6 5 , 1 7 1 - 1 7 2 , 1 8 0 - 1 8 3 , 2 4 9 - 2 5 0 , 266 Fermât, P. de 105 Festinger, L. 136 Feyerabend, P. K. 43, 49, 75, 84, 90, 187 Finn, E. J. 48 Fodor, J. A. 5 8 - 6 2 , 6 9 - 7 4 , 8 3 - 8 4 , 9 0 - 9 1 , 1 0 2 - 1 0 3 , 1 4 7 - 1 4 8 , 152, 1 6 8 - 1 6 9 , 174, 188, 190, 1 9 2 - 1 9 3 , 1 9 5 - 1 9 6 , 198, 199, 208, 213, 215 - 2 1 6 , 218 - 219, 222, 229, 232 - 237, 250, 263, 2 6 6 - 2 6 7 , 283 Frege, G. 82, 94, 1 1 4 - 1 1 6 Fresnel, A. J. 119 Galilei, G. 44, 47, 178 Ganslandt, H. R. 114 Gassendi, P. 28 - 29 Gazzaniga, M . S. 227 Gernhardt, R. 9 Gethmann, C. F. 68 Geulincx, A. 1 9 - 2 0 , 23 Gierer, A. 251 - 257 Glatzel, J. 131 Godei, K. 7, 252, 255 Goldbach, Ch. 116 Grossmann, S. 261 Grünbaum, A. 86 Gruter, J. 159 Hacker, P. M . S. 274 Hamilton, W. 105 Hanson, N. R. 188 Hardie, W. F. R. 11 Hartman, E. 11 Heckhausen, H. 1 3 2 - 1 3 3 , 135, 1 3 7 - 1 3 9 Heidegger, M . 277 Heidelberger, M . 32 Heijenoort, J. van 252 Heimsoeth, H. 24 Hempel, C. G. 43, 112, 129, 143, 147 Heraclitus 83 Herder, J. G. v. 19 Hertz, H. 212 Heyns, R. W. 135, 143 Hilbert, D. 198, 215 Hilgard, E. R. 129, 133 Hipparchus 108 - 1 0 9

Hobbes, Th. 16, 29, 175 - 1 7 6 Holbach, P. H. D. v. 29 Hopfield, J. L. 203 Horgan, T. 76, 2 1 8 - 2 2 1 , 226 Hume, D. 30, 117, 160, 2 4 1 - 2 4 2 Hunt, G. M . K. 261 Husserl, E. 68 Huxley, Th. H. 3 2 - 3 3 , 224 Irwin, T. 11 Jacobi, C. G. J. 105 Jacobson, L. 138 - 139 James, W. 117, 128, 130 Janich, P. 70, 149 Kambartel, F. 154 Kant, I. 19, 2 4 - 2 8 , 3 0 - 3 1 , 82, 153, 277-279 Kaulbach, F. 2 4 - 2 5 , 27 Keating, B. F. 162 Kehl, H. 34 Kenny, A. 15 Kierkegaard, S. 278 Kim, J. 101, 1 0 3 - 1 0 4 Knutzen, M . 19 Koyré, A. 47 Kripke, S. A. 5, 9 2 - 9 8 , 100, 183 Kuhlenbeck, H. 87 Kullmann, W. 11 Kutschera, F. v. 81 Lagrange, J. L. 105 Lakatos, I. 1 3 9 - 1 4 0 , 1 8 5 - 1 8 6 Lamarck, J.-B. P. A. de Monet de 50 La Mettrie, J. O. de 29, 32 Lange, C. G. 1 2 8 - 1 3 0 Laudan, L. 178 - 179 Lavoisier, A. L. de 50, 236 LeDoux, J. E. 227 Lefrancois, G. R. 130, 136 Leibniz, G. W. 5, 1 9 - 2 4 , 26, 31, 8 1 - 8 3 , 85, 88, 95, 105, 109, 1 4 8 - 1 4 9 , 1 5 9 - 1 6 0 Leplin, J. 151 Leucippus 159 Lewin, K. 38 Lienert, G. A. 166 Lorentz, H. A. 64, 109, 178 Lorenz, E. N. 2 6 0 - 2 6 2 Lorenz, K. 2, 22, 81, 83, 274 - 275 Lorenz, Ko. 121 Lorenzen, P. 181, 1 8 5 - 1 8 6

Index of Names Mach, E. 149, 160 MacKay, D. M . 124 Mainzer, K. 70 Malcolm, N . 274 Malebranche, N . 20 Marcus Aurelius 14 Marras, A. 201, 214, 222 Maxwell, J. C. 64, 175 McAllister, J. W. 110 McGinn, C. 9 9 - 1 0 0 , 104, 164 McKinley, W. 210 McLaughlin, B. P. 103, 232, 235 Meehl, P. E. 55, 83, 87, 88, 90, 98, 166, 1 7 0 - 1 7 3 , 183 Mendon?a, W. P. 198, 274 Mittelstrass, J. 3, 10, 12, 19, 22, 24, 47, 109, 111, 149 Nagel, E. 42 - 43, 54, 1 1 9 - 1 2 0 Nagel, T. 1 8 3 - 1 8 4 Nersessian, N. J. 175 Newton, I. 44, 4 7 - 4 9 , 1 4 8 - 1 5 0 , 160, 165, 167, 175, 178 Nicolis, J. S. 262 Nietzsche, F. 30, 278 Ockham, W. of 16, 42, 53, 54 Oppenheim, P. 43 Papineau, D. 188, 2 3 0 - 2 3 1 Parmenides 154 Patzig, G. 105 Pawlow, J. P. 132 Peano, G. 198 Pereira, G. 28 Planck, M. 131 Plato, 1 0 - 1 1 , 1 3 - 1 5 , 117 Plotinus, 14 Plutarchus, 14 Pöppel, E. 276, 279 Poincaré, H. 2 5 8 - 2 5 9 , 2 6 1 - 2 6 2 Polten, E. P. 8 2 - 8 3 Popper, K. R. 5, 1 1 4 - 1 2 5 , 127, 179, 277, 279 Priestley, J. 236 Prigogine, 1. 120, 258, 260 Prince, M . 34 Protagoras 211 Ptolemy, C. 188 Puccetti, R. 55, 160 Putnam, H . 5 8 - 6 1 , 6 5 - 6 6 , 129

309

Pylyshyn, Z . W. 190, 1 9 2 - 195, 201, 207, 235 Quine, W. V. O. 107, 111, 1 5 0 - 1 5 1 , 153, 156, 177, 253 Reininger, R. 27 Rensch, B. 34 Rentsch, T h . 275 Rescher, N. 70, 177, 284 Rorty, R. 75, 8 5 - 8 7 , 97, 1 8 3 - 1 8 4 , 280, 283 Rosenthal, R. 138 - 139 Ross, D. 133 Ross, S. A. 133 Ruch, F. L. 130, 133 Rüdiger, A. 19 Russell, B. 94 Ryle, G. 8, 2 7 4 - 2 7 5 Sayre, K. 196, 198, 240 Schachter, S. 130 Schaffner, K. F. 179 Schleichen, H . 39, 182 Schlick, M . 3 5 - 3 7 , 53, 8 5 - 8 9 , 91, 112, 154, 160, 162, 1 6 8 - 1 7 1 , 174, 184, 199, 2 4 9 - 2 5 0 , 283 Schmidt, R. F. 127 Schneider, K. 138 Schopenhauer, A. 1 Schroeder-Heister, P. 83 Schuster, H . G. 258, 2 6 0 - 2 6 2 Schwemmer, O. 43 Searle, J. R. 94, 2 0 0 - 2 0 1 , 2 4 3 - 2 4 4 Seilars, W. 68, 90, 170 - 1 7 3 , 185, 275 Sertillanges, A. D. 15 Shaffer, J. A. 1, 30, 1 5 9 - 1 6 0 Singer, J. E. 130 Skinner, B. F. 129, 1 3 2 - 1 3 4 , 136, 138 - 141 Smart, J. J. C. 81, 84 Socrates 210, 213, 229, 272 - 273 Sophie Charlotte 23 Sophocles, 76 Specht, R. 10, 1 6 - 1 9 Spinoza, B. 2 0 - 2 1 , 3 1 - 3 3 , 159 Stahl, G. E. 50, 160 Stein, D. L. 204 Stich, S. P. 76, 168, 191, 2 0 8 - 2 1 9 , 2 2 1 - 2 2 3 , 226, 229, 244 Strawson, P. F. 275 Suärez, F. 16

310

Index of Names

Tank, D. W. 203 Tarski, A. 43 - 44 Theon of Smyrna 109 Thiel, Ch. 114 T h o m a s Aquinas 14 — 15, 28 Thorndike, E. L. 132 Thorpe, E. 1 0 9 - 1 1 0 Turing, A. M . 6 0 - 6 1 , 63, 1 7 6 - 1 7 7 , 193 Voider, B. de 26 Vollmer, G. 124 Walker, E. L. 135, 143 Weiner, B. 132, 135, 137

Weinert, F. E. 138 Whitehead, A. N. 154 William of O c k h a m f Ockham, W. of Wilkes, K. V. 67, 77 Windelband, W. 29 Wittgenstein, L. 2, 8, 128, 2 7 3 - 2 7 5 Wolff, Ch. 24 Wolff, M . 47 Wolters, G. 2 6 - 2 7 , 32, 83, 1 1 6 - 1 1 7 , 144 Woodward, J. 76, 2 1 8 - 2 2 1 , 226 Wright, G. H . v. 280 Zimbardo, P. G. 130, 133

Index of Subjects Actuality (entelechy) 12, 23 Anima intellectiva 16 Animal spirits 17 — 18 Anima sensitiva 16 Anti-reductionism 5, 74, 118, 121 Apperception 25 — 26, 278 Aristotelianism 14, 29 Atomism 11, 28 Attitudes, propositional 68, 70, 76, 1 0 2 - 1 0 3 , 180, 189, 1 9 2 - 1 9 3 , 1 9 6 - 1 9 8 , 226, 238, 245, 280 Attractor 260 - 262 Autocerebroscope 42, 52, 97, 266 Behaviorism 3 3 - 3 4 , 1 2 9 - 1 3 2 , 1 3 9 - 1 4 2 , 145, 222, 274 Bodily a priori 26 — 27 Brain paradox 87 Cartesian dualism (Cartesianism) 5, 9, 19 Category mistake 8 1 - 8 2 , 8 4 - 8 5 , 88, 274 Causae occasionales 19 Causae procatarcticae 20 Causal attribution 137, 142 Chaos, deterministic 7, 257, 262 — 270 Charity, principle of 107 — 111 Chinese room argument 200 Cognitive 1 Cognitive science 7, 131, 1 9 0 - 2 0 8 , 202, 204 - 206, 212 - 214, 223 - 224, 228 Completeness of physics 100, 168, 172, 252, 254 Concepts, complete 21 - 23 Concepts, individual 21 Concepts, theoretical 6, 37, 1 4 3 - 1 4 8 , 151, 161, 164, 167, 1 8 6 - 1 8 7 , 209, 2 1 4 - 2 1 5 , 219, 229, 271 Concursus dei 19 Conditioning 133 Configuration space (phase space) 259 - 260, 261 Confirmability 70 Connecting principle (bridge law) 42, 71-72

Connectionism 202, 204 - 206 Consciousness 2 - 3 , 6 - 8 , 2 5 - 2 6 , 3 2 - 3 3 , 87, 117, 123, 1 8 0 - 1 8 4 , 1 8 9 - 1 9 0 , 2 0 7 - 2 0 8 , 218, 223, 225, 227, 2 7 1 - 2 7 3 , 2 7 6 - 2 7 9 , 2 8 1 - 2 8 2 , 286 Consciousness, empirical 25 - 26, 278 Consciousness, transcendental 26, 278 Constancy objection 88, 90 — 92 Construction 8, 276 Construct validity 166 — 167 Content, semantic 6 8 - 7 0 , 77, 142, 180, 189, 191 - 1 9 4 , 196 - 205, 207, 209 - 212, 215 - 216, 222, 224, 228 - 236, 244 - 246, 249 Contingency thesis (Kripke) 96 Contrary-to-fact sentences (counterfactuals) 70, 76 Correspondence rules 144—147, 190 Covering-law model j Explanation Darwinism 118, 1 2 2 - 1 2 3 , 230 - 231, 281 Deception argument 89 - 90, 91, 95 Deus sive natura 22 Dispositions 40, 6 0 - 6 1 , 274 Dissonance, cognitive 136 — 137, 140, 145, 1 7 3 - 1 7 4 , 189, 2 1 7 - 2 1 9 , 111 Double-aspect doctrine 31 - 3 3 , 34, 52 Dualism 2 - 3 , 6, 8, 16, 24, 55, 57, 61, 122, 125, 130, 1 5 5 - 1 5 8 , 167, 174, 179, 1 9 6 - 1 9 7 , 2 7 4 - 2 7 5 , 277, 279, 283, 285-286 Dualism, interactionist 1, 5, 8, 17—18, 5 5 - 5 6 , 1 1 4 - 1 1 5 , 117, 1 2 4 - 1 2 5 , 153, 158 - 166, 169, 172, 179, 286 Dualism, practical 8, 273, 2 7 6 - 2 7 8 , 279 Dualism, pragmatic 286 Dualism, theoretical 8 Ego (self) 7 - 8 , 8 5 - 8 6 , 1 1 6 - 1 1 8 , 124, 179, 227, 272, 2 7 8 - 2 7 9 , 281 Ego, empirical 25, 27, 278 Ego, transcendental 25, 27, 273, 278 Einstein-Feigl completeness 164—166, 249

312

Index of Subjects

Emergence 5, 54, 1 2 0 - 1 2 2 , 172, 207, 244, 270 Emergence thesis f Emergence Entropy 4 6 - 4 7 , 153, 265 Epiphenomenalism 32 — 33, 224 Equivalence, extensional 83, 101 Equivalence, functional 45, 66, 268 Esse est percipi 30 Essentialism thesis (Kripke) 95 — 97 Event 1 Everyday language 2, 41, 186 Explanation (covering-law model) 43 Explanatory constructions 147, 151, 271, 276 Extension (reference) 82 - 84, 93 - 95, 107, 232 Extremal principles 105 Folk psychology 41, 76 - 80, 180 - 181, 187, 1 8 9 - 1 9 2 , 2 0 2 - 2 1 1 , 214, 2 1 6 - 2 2 7 , 280-281 Folk psychology, intuitive 8, 225 — 228, 244 - 250, 280 Folk psychology, practical 8, 280 — 281 Folk psychology, technical 8, 225 — 226, 249, 280 Forma assistens 28 Forma informans 28 Formality condition 196, 201 - 2 0 2 , 205, 208, 214, 224, 228 Functionalism 4, 5 8 - 6 6 , 68, 7 7 - 7 9 , 123, 155, 1 7 6 - 1 7 7 , 190, 263 Genesis, normative 185 — 186 Genidentity 82 Gestalt psychology 89 - 90 Given, the 38, 40, 181, 185, 1 8 7 - 1 8 8 Godel's incompleteness theorem 7, 252 - 253, 255 Grain objection 88 — 90 Growth of knowledge 114, 116 Hermeneutics 104, 1 0 6 - 1 0 7 , 1 0 9 - 1 1 3 H-O-scheme of explanation f Explanation Hylomorphism 11 Hylozoism 10, 34 Idealism, dogmatic 31 Identity theory 2, 5, 7, 33 - 35, 46 - 47, 52, 55, 81, 83, 8 5 - 8 7 , 9 2 - 9 3 , 9 7 - 9 8 , 114, 1 2 2 - 1 2 3 , 1 2 4 - 1 2 8 , 131, 147, 254, 257, 270, 286

Identity theory, contingency of the 5, 81, 9 2 - 9 3 , 9 6 - 9 8 , 126 Identity theory, token- (particular) 36, 55, 58, 61, 66, 71, 7 4 - 7 5 , 7 8 - 7 9 , 91, 9 9 - 1 0 0 , 1 2 2 - 1 2 3 , 168, 1 9 6 - 1 9 7 Identity theory, type- (general) 4 — 5, 3 5 - 3 7 , 52, 5 6 - 5 8 , 6 1 - 6 2 , 66, 75, 77-80, 90-91, 122-123 Identity view f Identity theory Immaterialism 28, 30 Impetus theory 4 7 - 4 9 Individuation of states 224 Influxionism 19 - 20, 24 - 25, 28 Instrumentalism 152, 196 Intension 8 2 - 8 4 , 9 3 - 9 4 Intentional inexistence 70, 241 Intentionality 7, 6 7 - 6 9 , 71, 7 5 - 7 6 , 84, 1 0 0 - 1 0 3 , 142, 1 8 0 - 1 8 1 , 1 8 9 - 1 9 3 , 196-199, 206-209, 212-216, 2 1 8 - 2 1 9 , 2 2 2 - 2 2 3 , 2 2 5 - 2 2 6 , 228, 234, 237 - 239, 2 4 1 - 2 4 4 Inter-species functionalism 58, 65 Interactionism f Dualism, interactionist Interpretability 43 - 44, 46 - 47, 49, 73 - 74, 140-141, 220-221 Intra-species functionalism 58, 66 — 67, 71 Introspection 6 - 7 , 75, 86, 89, 91, 97, 147, 152, 169, 1 8 0 - 1 8 3 , 1 8 5 - 1 8 7 , 1 8 9 - 1 9 0 , 207, 2 1 8 - 2 1 9 , 2 2 4 - 2 2 6 , 247 James-Lange theory 128 — 130 Konstatierung

184

Language of science 3 Lawlike 70, 1 0 0 - 1 0 2 , 238 Leibniz-identity 5, 8 1 - 8 5 , 88, 9 2 - 9 3 , 95, 148 Leibniz's principle f Leibniz-identity Liaison-brain 5, 118, 124 Life world 7, 275, 279 Logical Empiricism 39, 42, 214 Materialism 3 1 - 3 4 , 61, 8 4 - 8 5 , 8 9 - 9 0 , 121 Materialism, eliminative 4, 7 5 - 8 0 , 98, 119, 153, 191, 2 2 2 - 2 2 3 Materialism, emergentist 128 Materialism, functional 61, 123, 196, 2 6 3 - 2 6 4 , 266, 2 6 9 - 2 7 0 Materia prima 15 Materia secunda 15

Index of Subjects Mechanics, statistical 3 7 - 3 8 , 4 4 - 4 6 , 57, 6 2 - 6 4 , 8 4 - 8 5 , 9 1 - 9 2 , 119, 131, 1 6 4 - 1 6 5 , 223, 2 6 4 - 2 6 6 , 2 6 8 - 2 6 9 Mental 1 Mentalism 6, 127, 131 - 1 3 2 , 142, 153, 155, 274, 285 Mind-body identity 5, 7, 40, 57, 77, 79 - 80, 9 2 - 9 3 , 96, 1 2 4 - 1 2 5 , 249, 270, 286 Mind-mind problem 41, 180, 190, 207, 224, 249-250 Minimal causation 220 Modularity principle 217 — 218 Modules, cortical 118 Monad 2 1 - 2 3 , 26 Monism, anomalous 9 9 - 100 Monistic position (monism) 28, 31, 57, 78, 99, 1 2 1 - 1 2 2 , 126, 128, 1 5 6 - 1 5 8 , 168, 174, 196 - 1 9 7 , 251 - 252, 253 - 254, 283, 286 Motivation 132, 1 3 4 - 1 4 2 , 145, 189, 197, 225-226 Multiple interpretability of formal structures 1 9 8 - 2 0 2 , 205, 213, 215 Natural kind terms 70 - 74, 220 Neuronlike circuits 202 — 205 Neurophysiology 3, 37, 4 1 - 4 2 , 5 1 - 5 2 , 56 - 59, 66 - 67,- 69, 71 - 72, 74, 77 - 80, 8 9 - 9 1 , 9 8 - 1 0 0 , 102, 119, 1 2 7 - 1 2 8 , 143, 1 6 4 - 1 6 5 , 167, 1 7 9 - 1 8 0 , 223, 249 - 250, 255, 257, 262 - 263, 266 - 2 6 7 , 269 - 2 7 2 , 2 7 6 - 2 7 7 , 2 7 9 - 282,283 - 284 Neuropsychology 156 - 158, 167, 266, 268 Nous (receptive and constructive) 12 - 15 Observation terms 3 7 - 3 8 , 144, 147 Occasionalism 1 9 - 2 0 , 2 4 - 2 5 , 28 Ockham's razor 42, 5 3 - 5 4 , 169 One's own mental events 97 Other minds 3 9 - 4 0 Panpsychism 34, 100 Parallelism, psychophysical 19 — 25, 28, 32, 5 2 - 5 3 , 100, 1 5 8 - 1 5 9 , 1 6 8 - 1 6 9 Parallelism, psychopsychological 207 — 208, 228-229 Phenomenalism 38, 4 0 - 4 1 , 9 6 - 9 8 , 181, 183 Phlogiston theory 50, 110, 126, 188, 236, 248 Physicali/physicah 1 7 0 - 1 7 3

313

Physicalism 57, 1 6 5 - 1 6 6 , 1 6 9 - 1 7 7 , 188-189 Physicalism, confirmatory ] Physicalism Physicalism, explanatory | Physicalism Physicalism, of content | Physicalism Physicalism, of form | Physicalism Possibility (technically, Physically, logically) 213-214 Potential, causal 54, 210, 2 1 4 - 2 1 5 Pre-established harmony 1 9 - 2 1 , 2 3 - 2 4 Principle of causal closedness 168 —169, 251 Principle of coherence 1 0 8 - 1 0 9 , 1 1 1 - 1 1 3 Principle of correspondence 120 Principle of finitism 7, 252 - 253 Principle of indiscernibility | Leibnizidentity Principle of rationality 101 - 104, 106, 108-111 Principle of sufficient reason 24, 148 — 149 Private language 38, 182 Problem of the empirical basis 181 — 183, 2 3 5 - 2 3 6 , 249 Process 1 Process-pluralism 154 Proper name 93 — 94 Propositional attitudes Î Attitudes, propositional Protocol statements 182 Psychodynamics 7, 199, 208, 213, 223, 228, 281 Psychokinematics 208, 228 Psychological 1 Psychology, Aristotelian 10 - 13, 19, 23 - 24 Psychology, cognitive 2, 3, 4, 6, 38, 1 2 9 - 1 3 0 , 131, 135, 1 3 9 - 1 4 2 , 145, 157 - 158, 189 - 1 9 0 , 197, 207 - 208, 209, 214, 2 1 6 - 2 1 7 , 2 1 8 - 2 2 1 , 2 2 3 - 2 2 6 , 2 2 7 - 2 2 8 , 2 4 9 - 2 5 0 , 274, 280 Psychology, syntactic 7, 191, 1 9 7 - 1 9 8 , 2 0 8 - 2 0 9 , 2 1 2 - 2 1 3 , 216, 222, 2 2 3 - 2 2 5 Psychoneurology 167 Psychons 128 Psychophysical law 100, 102, 107, 127, 158 Psychophysics 32, 3 6 - 3 7 , 74, 233, 235 Psychosemantics 7, 69, 191, 199, 201, 213, 228-229 Psychosemantics (causal (correlational) theory) 231, 2 4 3 - 2 4 4 information theoretic version 232, 237 - 243 psychophysical version 232 — 237

314

Index of Subjects

Psychosemantics (interaction theory) 229-230, 243-244 Psychosemantics (teleological (Darwinistic) theory) 2 3 0 - 2 3 1 Pygmalion effect 1 3 8 - 1 4 0 , 143, 2 2 5 - 2 2 6 Q u a n t u m mechanics 3, 64, 121, 131, 167-168 Realism, scientific 151 — 152 Realism thesis (Kripke) 95 — 97 Reciprocal reducibility 79 Reconstruction 1 1 0 - 1 1 1 , 1 8 5 - 1 8 6 Reduction 36, 4 2 - 4 7 , 4 9 - 5 1 , 5 4 - 5 5 , 7 1 - 7 4 , 77, 82, 85, 1 1 9 - 1 2 2 , 248, 266 Reduction, functionalist 62 — 66, 71—74, 7 8 - 8 0 , 155-156, 218-219 Reduction, ontological 36, 42, 44 - 45, 55, 6 4 - 6 5 , 66, 77, 8 1 - 8 2 , 85, 92, 119, 140, 2 4 9 - 2 5 0 , 283 cf. reduction Reduction, physicalist 28, 170, 1 7 2 - 1 7 3 , 174-176 Reductionism 71, 74, 1 2 0 - 1 2 1 , 271 Reduction of behaviorism to cognitive psychology 1 4 0 - 1 4 1 Reduction of folk psychology to scientific psychology 76 - 80, 218 - 219, 220 - 221, 2 2 7 - 2 2 8 , 248, 2 4 9 - 2 5 0 Reduction of psychology to neurophysiology 4 - 5 , 7, 4 1 - 4 2 , 5 2 - 5 3 , 5 6 - 5 7 , 73 - 74, 78 - 80, 92, 99 - 100, 102, 249, 2 5 4 - 2 5 6 , 257, 2 6 9 - 2 7 0 , 2 8 3 - 2 8 4 , cf. neurophysiology Reduction programs in psychology 6, 127-128, 130-132 Reference 39, 70, 82, 151, 1 8 9 - 1 9 0 , 196, 206, 216, 232, 249 Reinforcement, Skinnerian 1 3 2 - 1 3 3 , 134, 136, 1 3 9 - 1 4 0 , 141, 231 Relations, intertheoretical 64, 78 Relativity Theory, General 3, 150 Relativity Theory, Special 64, 109, 167, 175, 178 - 1 7 9 Representation, mental 193, 199, 212, 224, 230, 237, 239 - 240 Representation, symbolic 276 Res cogitans 16, 18, 24, 29, 153 Res extensa 16, 19, 24, 153 Retrospection 86 Self-consciousness 7 - 8 , 2 5 - 2 6 , 116 — 117, 252, 255, 272, 2 7 8 - 2 7 9

Self-movement 10, 12 — 13 Self-organization 257 - 258 Semantic content f Content, semantic Semantics of science 4 Semantic thesis (Kripke) 96 Sensory qualities 75, 8 7 - 8 8 , 89, 9 7 - 9 8 , 176, 1 8 0 - 1 8 1 , 1 8 3 - 1 8 4 , 1 9 0 - 1 9 1 , 2 3 2 - 2 3 3 , 235 Space 3, 87, 148 - 150 Space-time 3 State 1 Stoa 11 Subject, transcendental ] Ego, transcendental Supervenience 55, 201 Synergetics 257 Systems, dissipative 257 — 258, 260 — 262 Tarski-interpretability 44, 73 - 74 Theory dynamics 132 Theory-ladenness of perception 185 — 189 Theory of substantial forms 14, 16, 23, 28 Theory reduction 3 6 - 3 7 , 42, 4 9 - 5 1 , 6 4 - 6 5 , 66, 77, 92, 99, 119, 283, f Reduction Theory replacement 50, 221 Thermodynamics, phenomenological 37-38, 44-46, 50-51, 63-64, 91-92, 119, 260, 2 6 4 - 2 6 6 Thermodynamics, statistical f Mechanics, statistical Thoughts 1 1 4 - 1 1 5 Three-world theory 5, 114, 116 Token-token-identity j Identity theory, tokenTranslation, indeterminacy of 107 Translation rules 43 — 44, 73 Truth, necessary 93 - 94 Turing machine 6 0 - 6 1 , 63, 1 7 6 - 1 7 7 , 193 Two-substances dualism 24 - 25, 27 — 28, 3 0 - 3 1 , 153, 2 7 3 - 2 7 4 , 279 Type-identity j Identity theory, typeType-type-identity f Identity theory, typeUltimate foundation 4 Variational principles 104 — 105 Verificationist semantics 38 — 39, 214 Worlds, possible 93 - 95

w DE

G

Walter de Gruyter Berlin New York

GRUNDLAGEN DER KOMMUNIKATION UND KOGNITION FOUNDATIONS OF COMMUNICATION AND COGNITION

Meaning-Scepticism Edited by Klaus Puhl Large-octavo. IX, 258 pages. 1991. Cloth DM 142,ISBN 3 11 011833 5 (Library Edition) Quine, Putnam and, more recently, Kripke have brought forward sceptical arguments which in different ways attack the traditional ideas of linguistic meaning, rule and intention: Quine through his indeterminacy thesis, Putnam via model-theoretic doubts and Kripke by means of the rule-sceptical arguments he finds in the later Wittgenstein. The philosophical essays presented here have been specially commissioned for this volume to explore the extent to which these arguments pose a real challenge to meaning and meaningtheory. At the same time Meaning-Scepticism contributes to current discussion of Wittgenstein's work. Most of the papers are concerned in one way or another with the question of whether Wittgenstein endorsed meaning-scepticism, and with the problem whether, if he did not, his later philosophy contains arguments which refute such scepticism.

Price is subject to change

w DE

G

Walter de Gruyter Berlin - New York

GRUNDLAGEN DER KOMMUNIKATION UND KOGNITION FOUNDATIONS OF COMMUNICATION AND COGNITION

Thinking and the Structure of the World Das Denken und die Struktur der Welt Hector-Neri Castañeda's epistemic Ontology presented and criticised Hector-Neri Castañedas epistemische Ontologie in Darstellung und Kritik Edited by Klaus Jacobi and Helmut Pape Large-Octavo. XIV, 573 pages. 1990. Cloth DM 238,ISBN 3 11 011302 3 (Library Edition) The American Philosopher H.- N. Castañeda is well-known for his work on practical philosophy and on the context - and persondependency of language. This volume presents and discusses his metaphysics and philosophy of language based on a theory of experience. In four studies, Castañeda develops his theory of the indicators of proper names and the ontology of forming from a semantics of thinking Pnce is subject to change