Actions, Norms, Values: Discussions with Georg Henrik von Wright 9783110802450, 9783110154849

Citation preview

Actions, Norms, Values

1749

1999

Perspektiven der Analytischen Philosophie Perspectives in Analytical Philosophy Herausgegeben von Georg Meggle und Julian Nida-Rümelin

Band 21

W DE

G Walter de Gruyter · Berlin · New York 1999

Actions, Norms, Values Discussions with Georg Henrik von Wright Edited by Georg Meggle assisted by Andreas Wojcik

W DE

G Walter de Gruyter · Berlin · New York 1999

Printed on acid-free paper which falls within the guidelines of the ANSI to ensure permanence and durability. IJbrary of Congress Cataloging-in-Publication Data Actions, norms, values : discussions with Georg Henrik von Wright / edited by Georg Meggle : assisted by Andreas Wojcik. p. cm. — (Perspektiven der analytischen Philosophie : Bd. 21 = Perspectives in analytical philosophy) Papers presented at a colloquium held April 1996 at the Center for Interdisciplinary Research. University of Bielefeld. Includes bibliographical references and index. ISBN 3-11-015484-6 (alk. paper) 1. Wright, G. H. von (Georg Henrik), 1916— —Congresses. I. Meggle, Georg. II. Wojcik, Andreas. III. Series: Perspectives in analytical philosophy ; Bd. 21. B4715.W74A37 1998 198'.8-dc21 98-40608 CIP

Die Deutsche Bibliothek — Catalogmg-in-Publication Data

Actions, norms, values : discussions with Georg Henrik von Wright / ed. by Georg Meggle. Assisted by Andreas Wojcik. — Berlin ; New York : de Gruyter, 1998 (Perspectives in analytical philosophy ; Bd. 21) ISBN 3-11-015484-6

© Copyright 1998 by Walter de Gruyter GmbH & Co., D-10785 Berlin All rights reserved, including those of translation into foreign languages. No 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: Ingolf Max, Leipzig Printing: Arthur Collignon GmbH, Berlin Binding: Luderitz & Bauer, Berlin Cover design: Rudolf Hübler, Berlin

Preface

Actions, Norms, Values - these are the main topics in the philosophical work of Georg Henrik von Wright. He sketched a logic of actions and by turning to the rules of practical syllogisms he changed our entire discourse about the understanding of actions; he rediscovered (renewing Leibniz's ideas) the logic of norms; and he led the conceptual foundations for a systematic treatment of values. These achievements among others make Georg Henrik von Wright one of the most important philosophers of the twentieth century. Actions, Norms, Values - it was also these three key areas around which the discussions with Georg Henrik von Wright at the end of April 1996 involving some 30 experts (philosophers, law theoreticians and political scientist?) from all over the world revolved. They were held at the Centre for Interdisciplinary Research (ZiF) at the University of Bielefeld, which hosted the authors colloquium "Norms, Values and Actions - Discussions with Georg Henrik von Wright". During the scholarly preparation of the colloquium I was assisted by Werner Krawietz and Ernesto Garzon Valdes. Gerhard Sprenger of the ZiF and his assistant, Marina Hoffmann, made the organisation and the course of the colloquium a pleasure. The title of this ZiF colloquium was based on the volume published in 1994 by Suhrkamp entitled Normen, Werte und Handlungen (Norms, Values and Actions), in which von Wright compiled some of his main papers related to these subjects. Actions, Norms, Values is also the name henceforth to be used for these Discussions with Georg Henrik von Wright emerging from this colloquium. These proceedings document the present state of the art of von Wright's themes Actions, Norms, and Values. They continue earlier debates and launch new ones.

When preparing the publication of these proceedings, Andreas Wojcik relieved me of the dullest work. Ingolf Max provided an attractive, uniform layout. Hans-Robert Cram (de Gruyter publishers) enthusiastically took up the idea of this volume even before the Bielefeld colloquium; Gertrud Grünkorn supervised the results and Grit Müller handled the technical aspects of production. And finally, a very big thank-you to the ZiF, its staff,

vi

Preface

my co-organisers, all the participants in the colloquium and all the aforementioned.

Actions, Norms, Values was not just a scholarly event. Astuteness and cordiality usually live apart. But this was not the case during our Bielefeld discussions. And it is to the two most important reasons for this that this volume is dedicated (including on behalf of everyone else involved): Elisabeth and Georg Henrik von Wright. Georg Meggle

Contents

Abstracts

xi

Von Wright on Actions, Norms, and Values

GEORG HENRIK VON WRIGHT Ought to Be - Ought to Do GEORG HENRIK VON WRIGHT Value, Norm, and Action in My Philosophical Writings

3 11

Part I: Actions AULIS AARNIO Law and Action. Reflections on Collective Legal Actions

37

ROSARIA EGIDI Grounds for Acting, Grounds for Knowing

55

Ο LAV GJELSVIK On Mind and Matter

65

FRIEDRICH KAMBARTEL Remarks on Psycho-Physical Parallelism

79

FRANZ VON KUTSCHERA Explanation and Understanding of Actions

85

REX MARTIN Action Explanations as 'Understanding' Explanations

89

GEORG MEGGLE Understanding of Actions. Some Problems

103

viii

Contents

JEAN-LUC PETIT The Neurological Correlates of Action Types

Ill

HANS JULIUS SCHNEIDER Mind, Matter, and our Longing for the One World'

123

JOACHIM SCHULTE Willing and Acting

139

KRISTER SEGERBERG Results, Consequences, Intentions

147

MATTI SINTONEN On Freedom and Determinism

159

PIRMIN STEKELER-WEITHOFER What Is the Matter of Mind?

171

FREDERICK STOUTLAND Intentionalists and Davidson on Rational Explanation

191

Part II: Norms LILLI ALANEN Logical Modality and Attitudes to Propositions

211

NUEL BELNAP Concrete Transitions

227

EUGENIO BULYGIN

Existence of Norms

237

ERNESTO GARZON VALUES Reasonableness as a Criterion of Moral Correctness?

245

JÖRG HANSEN Paradoxes of Commitment

255

ERIC HlLGENDORF

Causality in Penal Law - Explanation or Understanding? A Sketch .. 265

WERNER KRAWIETZ Are There 'Collective Agents' in Modern Legal Systems?

273

FRANCISCO J. LAPORTA Legal Principles

279

Contents

ix

MARIJAN PAVCNIK The Importance of Legal Principles

285

OTA WEINBERGER Logical Analysis in the Realm of the Law

291

Part III: Values MANUEL ATIENZA AND JUAN Ruiz MANERO The Varieties of Legal Values

307

SVEN DANIELSSON The Norm-Value Distinction in 1963

325

BERNT ÖSTERMAN The Many Uses of "Good"

331

RUTH ZIMMERLING Ends: Rational or Reasonable?

345 365

Contributors/Addresses

359

Name Index

361

Subject Index

365

Abstracts AULIS AARNIO Law and Action. Reflections on Collective Legal Actions (page 37) This paper deals with collegia! judicial decisions as a form of human action. The scope is limited to three questions: (i) What are the structure and status of the general theory of action? (ii) Is this theory applicable to such performative acts as judicial decisions? (iii) Are the 'actions' of collective agents actions in the proper sense? The author defends the thesis that general theory of action is applicable not only to individual but also to collective actions; difficulties are not due to the structure of that theory or to its "individualistic character", but to the notion of a "collective will". This kind of "will" is epistemologically always a result of political procedures, and to speak of the "collective will" presupposes the analysis of these procedures.

LILLI ALANEN Logical Modality and Attitudes to Propositions (page 211) In discussing the nature and foundation of logical necessity Georg Henrik von Wright fights against a tendency to mystify necessity which Wittgenstein was fighting in criticizing the prejudice of the "crystalline purity of logic" and the idea of the "hardness of the logical must". The necessity attributed to the principles or laws of logic is not founded on any preformed logical structure of the world but stems, von Wright argues, from an attitude we take to some propositions. This paper examines the view of logic and logical necessity that emerges from his paper on "Logical Modality" and some of its implications. It outlines some traditional conceptions of modality and compares von Wright's view more particularly to Descartes's radical view of modality as dependent on the divine will and also to some contemporary views Descartes has been seen as anticipating. It purports to show that von Wright's way of detranscendentalizing modality by relating necessity to our attitudes or ways of treating sentences does not commit him to conventionalism or subjectivism.

MANUEL ATIENZA AND JUAN Ruiz MANERO The Varieties of Legal Values (page 307) In the following contribution, we, first, summarize the main aspects of our theory of legal sentences. We then present our conception of legal values. Such values are understood

xii

Abstracts

as the justificatory aspect of norms, seen as reasons for action. Concerning legal values, in our view, the most important classification is the one that distinguishes between ultimate, utilitarian, and merely instrumental values. Finally, that conception of legal values is compared with the theses on the relationship between values and norms, and on the distinction between intrinsic and extrinsic values, as presented by Georg Henrik von Wright in The Varieties of Goodness.

NUEL BELNAP Concrete Transitions (page 227) Following von Wright, 'transitions' are needed for understanding agency. I indicate how von Wright's account of transitions should be adapted to take account of objective indeterminism, using the idea of branching space-time. The essential point is the need to locate transitions not merely in space-time, but concretely amid the indeterministic, causally structured possibilities of our (only) world.

EUGENIC BULYGIN

Existence of Norms (page 237) In this paper several concepts of existence of norms are distinguished. Factual existence is related to certain empirical facts, which are characterized in different terms by different authors. The proposals of G.H. von Wright, Kelsen, Alf Ross and Hart are briefly discussed in this connection. Another concept is that of systemic existence or membership, widely used in legal discourse. This concept is also multivocal: a norm may be a member of a system (1) if it has been issued and not derogated by a competent authority, (2) if it is a primitive norm and (3) if it a logical consequence of some norms of the system. Another often used concept of existence is that of normative validity. In contradistinction to the first two it is a normative and not a descriptive concept. Finally, a very broad concept of existence as formulation is discussed. The different concepts are not incompatible, but only some of them are logically independent. As all of them are frequently used, confusions due to a lack of terminological distinctions are rather frequent.

SVEN DANIELSSON The Norm-Value Distinction in 1963 (page 325) In 1963 von Wright published three books about norms and values. It is evident that he there wants to distinguish more definitely than many other philosophers between normative and evaluative language, but it is less clear why. This question is the object of the present note.

Abstracts

xiii

ROSARIA EGIDI Grounds for Acting, Grounds for Knowing (page 55) The aim of this paper is to highlight some changes in von Wright's "dualistic" perspective after the publication of Explanation and Understanding. My focus will be especially on two theses characterizing the general theme "reasons for actions". I will make an attempt to show that, despite the background Wittgensteinian inspiration, both theses reflect a theoretical and systematic purpose, partially alien to the author of Philosophical Investigations.

ERNESTO GARZON VALDES Reasonableness as a Criterion of Moral Correctness? (page 245) Starting from some observations by Georg Henrik von Wright about the role "reasonableness" can play in the evaluation of ends pursued through norms, and about the desirability of overcoming what could be called "scientific fundamentalism", in this paper I formulate some doubts about von Wright's alleged moral skepticism and attempt to draw a somewhat more Optimistic' conclusion with respect to the possibility of an intersubjective foundation of moral norms. Added to the principles of symmetry and impartiality, "reasonableness" could perhaps be understood as a criterion of correctness for moral norms. This does, of course, presuppose that one concedes - as does von Wright - that there is a "naturalistic basis of all evaluation", namely, the universally shared desire of survival.

OLAV GJELSVIK On Mind and Matter (page 65) Georg Henrik von Wright's paper "On Mind and Matter" develops an interesting and important view on the grand philosophical issue of the relationship between the mental and the physical. It also professes an affinity with epiphenomenalism. I worry about epiphenomenalism, and I find it necessary to avoid it to keep some basic and indispensable ideas in the way we view ourselves. Von Wright is moved towards epiphenomenalism partly because he sees grave objections to rival views like identity-theories. I share many of his objections to rival views. In this paper I try to explain why we should worry about epiphenomenalism, and also how we can avoid the dilemma of choosing between epiphenomenalism on the one side and mind-brain identity-theories on the other. I outline briefly a view which is a weaker "Aristotelian" version of philosophical materialism, to which von Wright's arguments against identity-theories seem not to apply, and which is not epiphenomenalist. I offer this for as a suggestion for where to go, and not as a fully worked out view.

xiv

Abstracts JÖRG HANSEN Paradoxes of Commitment (page 255)

Professor von Wright's logic of norms as developed by him since 1981 is examined for paradoxes of commitment. In this logic the basic form of conditional norms is O(p -> q)\ It is shown that Chisholm's Paradox does not arise in this logic and that some Roman arguments concerning conditional promises can be reconstructed by it. Finally an example is presented which suggests that conditional obligation cannot generally be represented by O(p -4 9)'. ERIC HlLGENDORF

Causality in Penal Law - Explanation or Understanding? A Sketch (page 265) One can distinguish four types of causality that have relevance for penal law: causality between (natural) events, causality between (human) actions and events, causality between events and actions and causality between actions. All these forms can be sufficiently analysed with the help of the covering law-model. It may be that this model is reductionist; it certainly fails to take the emotional and intentional background of human behaviour into consideration the knowledge of which is indispensable for an understanding of actor and action. But an understanding in this sense is not necessary for causal analysis in penal law.

FRIEDRICH KAMBARTEL Remarks on Psycho-Physical Parallelism (page 79) The first part of the following remarks distinguishes two varieties of parallelism between physical and practical events. A thought experiment is described, which demonstrates the impossibility of a parallelism that is based on necessary and sufficient conditions and therefore constitutes causal dependencies. Subsequently the conceivability of a weaker form of parallelism is vindicated in the light of a particular understanding of strictly specific conditions. The second part focuses on the physical causes of sensations and perceptions. It is argued that the conceptual boundaries between the physical world and the world of sensations preclude the idea of a continuous causal chain, but nevertheless permit the application of an instrumentalist concept of causation; in the case of perceptions the pragmatic aspects of the involved judgmental structure seem to rule out the possibility of any causal determination.

WERNER KRAWIETZ Are There 'Collective Agents' in Modern Legal Systems? (page 273) Contemporary theories of norms and action are subject to rapid changes. Today individual action appears to be largely determined and informed by socially established

Abstracts

xv

institutions and social systems and by the modern communication systems of the law. These social institutions and systems do not, however, consist of individual actors and their intentional actions, to which the organized actions undertaken by 'collective agents' or 'collective subjects' also tend to be attributed. Social systems constitute, i.e. produce and reproduce themselves, in the legal system of modern society by self-referentially linking communicative acts (directives, norms) to former directives and norms which are interwoven with each other in legal communication.

FRANZ VON KUTSCHERA Explanation and Understanding of Actions (page 85) This statement was to serve as a short introduction to the papers on explanation and understanding of actions. Therefore it only briefly mentions two key-topics in this area and makes a few statements on them intended to trigger the discussion. The topics are: The nature of rational explanations and what they explain - as I see it, they explain not the occurrence, but the rationality of an action -, and the compatibility of rational and causal explanations of an action - referring to a rather exclusive notion of causality I maintain that what admits of rational explanation does not admit of causal explanation.

FRANCISCO J. LAPORTA Legal Principles (page 279) The paper is a proposal to put under discussion the possibility of including the notion of 'principle' or 'legal principle' in the framework of von Wright's thought. The author suggests that a change from a logic of the Ttinsollen type to a logic of the Seinsollen type might be a first step to clear the way for the incorporation of the elusive concept of 'principle' into a system of deontic logic.

REX MARTIN Action Explanations as 'Understanding' Explanations (page 89) My paper is concerned with setting out the views of Collingwood and von Wright on the explanation of action. Here I will identify a single main model or schema for the explanation of actions (that is, for explanations of actions by reference to reasons - to certain thoughts and motivations of the agent). This model, in my view, provides the root of both von Wright's notion of practical inference and Collingwood's idea of reenactment. In this paper I will turn as well to a critique of their two theories, by taking up and contrasting the role of understanding or intelligibility, often called Verstehen, in each of their accounts.

xvi

Abstracts GEORG MEGGLE Understanding of Actions: Some Problems (page 103)

An action is understood by us iff we see the action as being the conclusion of an appropriate practical syllogism. With this starting point of von Wright's Explanation and Understanding (1971) several other proposals are compared and more or less identified with, namely understanding as (i) knowing of the intention with which the action was done, as (ii) knowing the reasons for which it is or was rational to perform the action, and as (iii) knowing the subjective meaning of the action. Relative to these different versions of "understanding of an action", I summarise some differences between von Wright and myself.

BERNT ÖSTERMAN The Many Uses of "Good" (page 331) The paper discusses the multiplicity of uses of "good", a phenomenon which has been named the Varieties of Goodness by von Wright. With Mackie's definition of "good" in terms of requirements as a point of departure it is argued that von Wright's categories of goodness may be regrouped in a way that offers a more unified conceptual picture than he has presented.

MARIJAN PAVCNIK The Importance of Legal Principles (page 285) A legal rule expresses a type of conduct, whereas a legal principle "only" expresses a standard (e.g. reasonability, honest person) of how to act in legal relationships (e.g. in civil law), in which we are representatives of certain types of conduct. Thus the legal principle is neither an absolute measure nor one depending exclusively on the competent user, but a value basis that calls for acting "proportionately" and striking the "right balance". According to this distinction legal principles are measures directing the understanding of legal rules and the manner of their application. JEAN-LUC PETIT

The Neurological Correlates of Action Types (page 111) Taking seriously the recent cognitive development in science forces us to clash head-on with Wittgenstein's autonomy of intentional language argument in order to recognize the right of neuroscience to deal not only with bodily movements, but also with intentional actions. We discuss the chances of finding a way out of this dilemma in von Wright's action theory.

Abstracts

xvii

HANS JULIUS SCHNEIDER Mind, Matter, and our Longing for the One World' (page 123) In a first step I will summarize those accomplishments of von Wright's (1971) Explanation and Understanding that in my eyes are most important for the mind-body problem. In part two I will interpret our desire to live in One world' as one of the attractions of monism. I will characterize materialistic monism as an unjustified ideology and will (following ideas in EaU) offer an alternative (and non-idealistic) way for satisfying our monistic inclinations. Part three will take up von Wright's discussion in his paper On Mind and Matter (1994), interpret it, and by doing so will possibly push its significance into a direction the author himself did not intend. In part four, finally, I will ask how far the discussion has brought us to a solution or dissolution of the traditional problem, how the mind-body distinction relates to the mind-matter distinction, and what new problems the dissolution of the traditional problem has brought.

JOACHIM SCHULTE Willing and Acting (page 139) In spite of forceful counter-arguments known from the writings of Wittgenstein and von Wright does it seem possible to defend a certain view held by advocates of the notion of pure and/or isolable acts of will. However, in the light of further considerations based on insights also to be found in Wittgenstein and von Wright the claims of those advocates of pure volitions or acts of will can be seen to lose most of their initial plausibility.

KRISTER SEGERBERG Results, Consequences, Intentions (page 147) Von Wright's distinction between results and consequences of actions is fundamental. As soon as it has been made, it seems obvious; nevertheless it is not easy to account for the distinction within the framework of a rigorous logical theory. In this paper we study von Wright's theory, as it is presented in Chapter III § 5 of Norm and Action, in something like a laboratory setting - a modelling that is highly abstract but, thanks to its very abstractness, helps to highlight some interesting questions about von Wright's theory.

MATTI SINTONEN On Freedom and Determinism (page 159) Von Wright's Freedom and Determination (1980) widens the horizons of Explanation and Understanding (1971) by examining human action and its determination in a social

xviii

Abstracts

context. He distinguishes between the metaphysical problem of freedom and the problem of freedom relevant for social, moral, political and legal concerns. What a man can do depends on his abilities and on opportunities. The more immediate determinants divide into internal determinants, such as wants or desires, and external determinants, such as symbolic challenges and institutional practices. These determinants are not Humean causes of action, but rather reasons or grounds for their performance. The socially relevant issue? of freedom arise from the dialectics of internal and external determinants. Causation as such cannot threaten freedom to act, but internalization of institutional practices can be experienced as a loss of freedom: e.g., if "compelling" duties leave no choice, or if the agent lacks the required opportunities or abilities. This paper suggests that von Wright's separation between the causal world of bodily behaviour and the intentional world of action explains why the ontology and epistemology of the sciences of man differ from those of the natural sciences. The ultimate determinants of action, wants and duties, seem to provide a rock-bottom level for action-explanations. Wanting health, happiness and well-being neither can nor need have deeper (e.g. "naturalistic") explanations. These is no such rock-bottom level in the natural sciences. The main problem in their view is how these two realms are related.

PIRMIN STEKELER-WEITHOFER What Is the Matter of Mind? (page 171) In explaining actions or behaviour we distinguish reasons and causes (1.1) and, correspondingly, different forms and levels of explanation (1.2). Parts of these distinctions get blurred by the introduction of a rather vague concept of mental events (1.3). Von Wright's analysis of the relations between the behavioural, the mental and the neural (2.1-2.3) is designed to lead to a solution of the mind-body-problem similar to Davidson's Anomalous Monism (3.1), but the basic problem of an a priori belief in the causal connectedness of physical events is not addressed (3.2).

FREDERICK STOUTLAND Intentionalists and Davidson on Rational Explanation (page 191) It has become standard to take von Wright and Davidson's points of view about the nature of rational explanation of human action as polar opposites, von Wright representing the intentionalist approach and Davidson the causalist. This paper argues that this is a mistake, stemming largely from a misreading of Davidson's claim that rational explanations are causal. A more careful look at that claim shows that Davidson does not hold that the force of rational explanations derives either from empirical generalizations of any kind or from causal relations between reasons and actions. Their force stems rather from the conceptual and normative principles used to interpret rational agents, their attitudes,

Abstracts

xix

and their action. This is essentially the point of view of intentionalists like von Wright, which means that the two most important philosophers of action of our time differ not so much substantively as verbally.

OTA WEINBERGER Logical Analysis in the Realm of the Law (page 291) The principle of tolerance leads to the consequence that the appropriateness of proposed logical systems has to be tested in the field of their application. Such a test proves that classical deontic logic is not appropriate for legal argumentation. A genuine logic of norm sentences must be worked out for this purpose. The main philosophical presuppositions for this task are discussed. Structure theory of the legal system, legal dynamics and legal argumentation are the main fields of logical analysis in jurisprudence. Logical enquiry shows that logic cannot provide principles of natural law. Not only norm logic, but also other systems of practical reasoning are relevant for legal argumentation, namely: formal teleology, formal axiology, logic of preferences. Two examples of the application of logical analysis in the theory of justice are discussed: equality and universalization.

RUTH ZIMMERLING Ends: Rational or Reasonable? (page 345) In the following contribution, I take von Wright's essay "Rationality: Means and Ends" as the point of departure for a presentation of his conception of the difference between rationality and reasonableness. The presentation is introduced by some remarks on his late 'discovery' of the distinction, and on his impression that there is something peculiar to it in English. I then attempt a step-by-step reconstruction of von Wright's conception, pointing out what seem to me to be some implications not mentioned expressly by von Wright himself. The presentation is followed by a discussion of a few points in von Wright's conception that, in my view, do need clarification. The comment closes with the suggestion to 'take reasonableness seriously'.

Von Wright on Actions, Norms, and Values

GEORG HENRIK VON WRIGHT Ought to Be - Ought to Do

1 There is a familiar distinction between two uses of "ought" and also of "may". It is usually referred to with the phrases "ought to (may) be" and "ought to (may) do". Corresponding phrases are familiar from other languages known to me. Awareness of the relevance of the distinction to a theory of norms has perhaps been more manifest in German than in English philosophical writing. In German Sein-SollenfDürfen and TunSollen/Dürfen are established terminology for referring to the two uses of "ought" (so//) and "may" (darf). That which ought to or may be can generally be spoken of as a state of affairs - and that which ought to or may be done as an action (or type of action). 2 My first attempt to develop a formal logic of norms (a "deontic logic") was of the Tun- Sollen /Dürfen rather than the S'ein- Sollen /Dürfen type. Its basic concept, moreover, was that of permission. The ought, obligation or obligatoriness, was defined as the negation of the permittedness of the opposite. The things pronounced obligatory or permitted by norms were taken to be categories or types of action, such as e.g. murder, theft or smoking. This approach to a logic of norms is quite natural. More so, perhaps, than the conception of such a logic as one of the S'ein- Sollen /Dürfen type. It was, however, connected with some difficulties and limitations. One was caused by the application of the truth-connectives, particularly negation, to names of actions. What action is, for example, "not-murder" ? Is it to abstain from, omit murdering somebody or is it simply to do something else? Another apparent shortcoming of this logic was that it did not allow for an iteration of the normative characters, for example for saying that it is permitted that something is forbidden. The forbidden thing can be an action (e.g. smoking), but its being forbidden is a norm and not an action; therefore it makes no sense (in this system) to speak of the permittedness of a prohibition. The first ones to shift to a Sein-Sollen conception in deontic logic were, if I remember correctly, Arthur Prior and Allan Anderson. And I soon followed their example. The same generally holds true of the later development of

4

Georg Henrik von Wright

the subject. Sporadic returns to a Tun-Sollen conception, by myself and others, occurred, but remained outside the main stream. On a Sein-Sollen view the things pronounced obligatory, permitted or forbidden are states of affairs - for example that an agent is smoking or commits theft or murder. Whereas in my first system the variables p, q, etc. were schematic names of actions, they are in a Sein-Sollen logic schematic sentences expressing propositions (about action). 3 Most deontic logics use two operators: Ο for obligation ("ought") and Ρ for permission ("may"). On what may be called a received view, the operators are interdefinable with the aid of the symbol of negation ~, Ο = ~ Ρ ~ and Ρ — ~ Ο ~. This view I do no longer find acceptable. A special symbol for prohibition, however, is not needed, since a prohibition may be defined as an obligation to the contrary (O~) and an obligation as a prohibition to the contrary (0~~p = Op). This I find acceptable. The operators are prefixed to variables p, q,... or to molecular compounds of variables. In a Sein-Sollen logic Op is read "it ought to be the case that p"; here "p" stands for a sentence. In a Tun-Sollen logic the expression is read "one ought to p" where "p" stands for a verb of action. Every sentence of the schematic form Op or Pp can be used in two characteristically different ways. It can be used prescriptively for imposing an obligation or giving a permission. Or it can be used descriptively for stating that a certain obligation has been imposed or permission given. In the first case the sentence expresses a norm] in the second it expresses something which I shall call a norm-proposition. The sentences themselves will be called norm-formulations. 4 To affirm a norm-proposition is to state the existence of a norm. Normpropositions are true or false. The existence of norms I shall regard as empirical facts which may be established by ascertaining (social) reality. Norm-propositions "obey" logical laws which are valid for propositions generally. Whether there are logical principles which are peculiar to normpropositions, whether there is a Logic of Norm-Propositions as distinct from a Logic of Norms is a question in its own right. Few writers have bothered about the distinction. An exception are Alchourron and Bulygin. I shall not pursue the question here - and I do not pretend to know how to answer it. Norms as prescriptions of what ought to or may be or be done are neither true nor false. This is a commonly entertained view which I find acceptable. But if norms lack truth-value, how can logical relations obtain between norms? Can a norm logically contradict another norm, or one norm follow logically from another? Is a Logic of Norms possible? The great promise which early efforts, by me and by others, of building a deontic logic held forth tended to begin with to obscure the problematic nature of the enterprise. But the question soon began to torment me. I think it has an affirmative

Ought to Be - Ought to Do

5

answer. Before I briefly restate my view of what a deontic logic "looks like" some preliminary observations will be necessary. 5 That which a norm pronounces obligatory or allowed I shall call the content of the norm. A norm-content is doable if, and only if, it can come about as a result of human action. The notion of doability is relative to agents: what one agent can do another one may be unable to achieve. An existing norm emanates from a source. The source can be a sovereign king, a legislature, a worshiped God, etc. A norm can also have an impersonal source, originate ("grow up") within an ethnic or religious community. Norms emanating from the same source are said to constitute a normative order or system. Those to whom a norm applies, i.e. who are expected to do what the norm pronounces obligatory or perhaps avail themselves of that which the norm allows, are said to constitute the class of addressees of the norm. A norm will be called genuine if its content is doable by all agents to which it applies. A norm with a not-doable content will be said to be spurious. An example of a spurious norm would be one which pronounces a contradictory or tautologous state of affairs obligatory. The idea that only norms with a doable content are genuine, i. e. genuinely apply to the agents to whom they are addressed, is a variant to the wellknown idea, usually associated with the name of Kant, that what one ought to do one can do. "Ought implies Can". 6 In the course of years my opinions in deontic logic have undergone great changes. They have gradually moved away from what can perhaps be called the mainstream of developments of the subject. They have become increasingly what many people would regard as "nihilistic". Logical relations can obtain between norms only in an oblique sense which must be explained and justified separately. Given a set of genuine norms, stemming from the same source and addressed to the same subjects (agents). It can be divided into two sub-sets, one consisting of the O-norms, the other consisting of the P-norms of the original set. A set of O-norms is consistent if, and only if, the conjunction of the contents of all members of the set is doable. (One could also say: if all the member-norms are jointly satisfiable.) The notion of consistency does not apply to sets of P-norms. A joint set of O- and P-norms is consistent if, and only if, the conjunction of all the contents of all O-norms with the content of any one of the P-norms individually is doable. By the negation-norm of a O-norm I understand a P-norm, the content of which is the negation of the content of the O-norm. Similarly, the negation-

6

Georg Henrik von Wright

norm of a P-norm is a O-norm, the content of which is the negation of the content of the P-norm. Differently put: the negation of an obligation is a permission "to the contrary"; and the negation of a permission is an obligation to the contrary. We can now define the notion of normative entailment. A consistent set of norms will be said to entail a given norm if, and only if, adding the negation of the given norm to the set makes the set inconsistent. Thus, for example, the conjunctive permission P(p/\q) entails the simple permission Pp, because the set consisting of P(p Λ q) and O~ ρ is inconsistent. It is inconsistent because the conjunction p/\q Λ ~ρ is a contradiction, and thus something not doable. A case which has much puzzled deontic logicians is known as Ross's Paradox. Op entails O(p V q). This is so also on our definition. But this means, simply, that if it is obligatory that (to) p, then it cannot be permitted that (to) ~p Λ ~ς. This is obvious and in no way "paradoxical". The appearance of paradox arises from a confusion between norms and norm-propositions. The entailment in question does not mean that if there is (exists) a norm to the effect that p, then there also is a norm to the effect that p V q. The latter norm may, but need not coexist with the former. 7 A great many norms - perhaps a majority - are conditional. They prescribe what ought to or may be done if certain other things are the case (actions are performed). A standard example is the norm enjoining us to keep our promises. The correct logical form of conditional norms has been a matter of much dispute. There are two alternative conceptions. The one can, in the traditional symbolism, be represented by O(p —> q}; the other by p -> Oq. The material implication —>· may here be regarded as an adequate representation of the "if-then" or conditional relation. The difference between the two conceptions is the difference between "It ought to be the case that if ρ then q" and "If ρ then it ought to be the case that g". On the second interpretation it looks as though the truth (satisfaction) of the antecedent (p) engenders, gives rise to, an obligation. This view, I think, is not correct. On the first conception, the obligation is there independently of whether the antecedent of the conditional relation is true or not. This, I think, is correct. "Promises ought to be kept". This norm (obligation) was there since times immemorial. (We need not here inquire what is its "mode of existence", and whether the institution of promising is transcultural or not.)

Ought to Be - Ought to Do

7

The norm does not enjoin us to give (make) promises. Nor to do just any thing which might be the object of a promise. It only tells us that if we promise to do something, we ought also to do this thing. Now an agent promises to do something. Does this not mean that a new obligation is thereby created, viz. the self-imposed obligation for him to do this thing? I would answer as follows: Before he gave the promise, the agent was under an obligation either not to promise to do the thing or, if he promises, to do it. If he gives the promise, he will have to do something which, before he promised, he was not necessitated to do. That is: in order to satisfy the norm, to fulfil his obligation to keep given promises, he will now have to do what he promised. I shall call this a practical necessity incumbent on the promisor. It was not there before he had given the promise. In this sense it is a "new" thing. But it is not a new obligation, but something the agent has to do in order to satisfy an already existing one, viz. the obligation either not to promise or to do the promised. The statement of practical necessity is not a new norm but a true or false statement about what an agent has to do in order to fulfil his obligations. One may call the practical necessity too an "obligation", something one ought to do. But then one must remember that this "obligation" has a very different logical status from the primary one, the satisfaction of which requires us to act in such and such a way. It is natural and also common to speak of the practical necessities as things which must or have to be done (for the sake of some end) rather than things which are obligatory and therefore ought to be observed. But we need not be pedantic about the words "must", "ought", "obligation" and the rest - once we keep the meanings distinct. A useful phrase in this connections is "to see to it that". A conditional obligation, one could then say, enjoins (obliges) us to see to it that either the antecedent of the conditional is not or the consequent is satisfied (true). 8 The observations we made on conditional norms will turn out useful also when discussing the relation between an ought to (may) be and an ought to (may) do conception of a logic of normative discourse. I shall put forward, tentatively, the following thesis: Norms primarily prescribe what ought to or may be; propositions of practical necessity state what the addressees of the norm must or may do in order to satisfy their obligations or avail themselves of their permissions. In defence of the view that the idea of ought to (may) be holds a primary position in relation to the notion of ought to (may) do one can advance the following argument: The norms valid within a community of people generally express the idea (will, wish, desire) of a law-giver (norm-authority) about the state which

8

Georg Henrik von Wright

ideally should obtain within this community. (The authority can also be an impersonal "voice of the people".) The members of the community have to see to it that the ideal states envisaged by the norms actually are realized. To that end they have to consider what they have to (must) or are free to do. This again depends, not only on the contents of the norms, but also on variations in the circumstances under which they live. The following illustration may be helpful: Let the norm be Op. An addressee of the norm ought to see to it that the state of affairs that p obtains. Assume the prevailing state is that not-p and that it will remain so unless the agent interferes. Then he ought to (must) produce the state that p in order to satisfy his obligation. Assume that the state is that not-p but will change to that p also if the agent remains passive. Then he must not suppress the state, prevent it from originating. Next assume that the state is already there and does not change unless interfered with. Then the agent must not destroy it. Assume, finally, that the state is there but changes to its opposite unless there is interference. The agent has now to prevent this, i.e. to sustain the state in question. Producing, destroying, suppressing, and sustaining (a state) I shall call the four basic forms of action, of doing something. (Not all states are such that they logically permit all forms of "manipulation"; maybe they can be produced or destroyed, but not sustained or suppressed.) In the illustration we have been using the obligation which the agent has to observe is unchangingly to see to it that p, but what he must do, to which mode of action resort in order to satisfy his obligation, depends on and varies with the prevailing situation. The statements that the agent must or must not do this or that in order to satisfy his obligation are true (or maybe false) in relation to the norm imposing the obligation. These statements of practical necessity should therefore better not be called themselves norms. The Tun-Sollen/'Dürfen is descriptive, the S ein-So lien/Dürfen alone is prescriptive, one could say. 9 The fact that an agent performs a certain action is also a state of affairs of a sort. The variable in the expression Op and other formulas of deontic logic could represent a sentence describing an action. This was sometimes used as an argument against my original efforts to build a logic of the Tun-Sollen type and in favour of a logic of the Sein- Sollen type where the variables were schematic representations of sentences. A logic of the second type was thought to be able to "incorporate" everything expressible in a logic of the first type. But this was a superficial argument since it overlooked the complex structure of action-sentences. An example. A "law-giver" is anxious that an agent, over whom he has authority, should produce the state that p if it does not obtain - but not anxious that he should not destroy it if it is already there. He wants

Ought to Be - Ought to Do

9

perhaps to test the agents "productive" but not his "destructive" (nor his "sustaining" nor his "suppressive") capacity with regard to this state. To produce the state that p is only one of the four cases which are covered by seeing to it that p. To distinguish the four cases within the formalism of a deontic logic requires a developed formalism of a logic of action. What has the addressee of a norm making obligatory the production of the state that p to do in order to satisfy his obligation? The answer depends upon the prevailing situation. If the state is not already there nor originates independently of the agent, then he must (ought to, has to) change the state that not-p to its opposite. This is, for him, a practical necessity (under the norm). But if the state is already there, there is nothing he can do then to satisfy his obligation. If the norm is not cancelled he will just have to wait for an opportunity to satisfy it. Thus also in the case that we have been discussing. The content of the norm is a certain state of affairs, which ought to be, viz. that an agent performs a certain action, if there is an opportunity. What the agent has to do for the sake of complying with the norm depends upon the prevailing circumstances. The statement that he has to (ought to) do this or that is then true or false given the norm and the circumstances.

GEORG HENRIK VON WRIGHT Value, Norm, and Action in My Philosophical Writings With a Cartesian Epilogue

I am deeply indebted to the management of ZIP for arranging an Autorenkolloquium in my honour in Bielefeld in April 1996. The theme of the colloquium was Actions, Norms, and Values. Nine papers - three for each of the main topics - were read and discussed at the meeting. I responded to them all. Subsequently, no less than 30 essays were submitted for publication in the Proceedings. Rather than commenting on the individual contributions, I decided to write an essay on the place and role of the three topics of the colloquium in my thinking as a whole. The essay is "autobiographical" in that it follows the development of my thoughts in chronological order. This is why I have reversed the order of the themes to Value, Norm, and Action. In order to make the essay maximally self-contained I have also summarized my principal views and arguments. The summaries are often greatly oversimplified and may therefore strike a reader who is unfamiliar with the sources as uncritical or even dogmatic. The chronological perspective made it natural for me to deviate sometimes from the tripartite division value-norm-action and make a brief excursion into something which had been an off-shoot of my occupation with one or the other of the three topics. Most of these off-shoots have stemmed from action-theory. Since the mid-eighties they have increasingly deviated in the direction of the philosophy of mind and the metaphysics of the Cartesian divide between matter and mind. As such they signal a return to the problems of philosophy which were the first to attract me in my early adolescence. Some of the contributions in the book follow my thoughts in this direction. I therefore thought it appropriate to conclude my comments with a brief "Cartesian epilogue". Special thanks go to the chairmen of the three sessions of the colloquium, Professors Ernesto Garzon Valdes, Werner Krawietz, and Georg Meggle. The efficiency and unfailing patience of Professor Meggle, both in planning the colloquium and editing its proceedings is a gesture of friendship for which I shall be for ever grateful.

12

Georg Henrik von Wright

I dedicate these autobiographical musings of mine to the contributors to the volume as a token of gratitude for the interest they have taken in my philosophical endeavours.

1 As an academic study, philosophy in Finland has traditionally been divided into two independent disciplines - theoretical and practical philosophy. In the first were counted logic, epistemology, and metaphysics; in the second ethics and social and political philosophy. In my student days theoretical philosophy also included psychology. I chose to study theoretical philosophy, my main motive being, as far as I remember, that I had heard about the Professor, Eino Kaila, who was married to a cousin of my mother. As a consequence of my choice I never studied moral, legal or social philosophy. Nor philosophy of history or religion. This is mildly ironic, considering that a main part of my work in philosophy, and certainly that part of it which is best known to the world, has been in the subject's "practical" dimensions - and thus falls within a discipline in which 1 have never had any professional training at all! 2 In the intellectual environment in which I grew up, however, the influence of Edward Westermarck was still strongly felt. He had been Professor of Practical Philosophy in Helsinki in the beginning of the century and was later Professor of Philosophy at Abo Academy and of Sociology in the University of London. He is one of the greatest scholars my country has fostered, and his influence on secular opinions in matters of morals and religion had been profound. It was therefore more or less "a matter of course" that in my early, non-professional reflections on questions of value I implicitly embraced the ethical subjectivism and relativism of which Westermarck had been a champion. In some sense - perhaps not easy to make explicit - I have remained faithful to these early views of mine. What I would call "genuine" valuations express a subject's approval or disapproval of an evaluated object. As such expressions, valuations are neither true nor false. The valuation of an object by one subject may differ from the valuation of the same object by another subject. But the statement "I approve this", unlike my approval "itself, is either true or false. In this regard value-sentences have a similar "systematic ambiguity" as normative (deontic) sentences. (Cf. below, II, 3) Westermarck would have said that valuations express feelings of approval or disapproval - and somebody may wish to say that they are expressive of approving or disapproving attitudes. But these expansions of the shorter formulation given above seem to me either otiose or else controversial.

Value, Norm, and Action in My Philosophical Writings

13

3 Westermarck also raised the question how to distinguish moral approval and disapproval from other kinds of valuation. It is to Westermarck's credit that he saw a problem here. Other supporters of an "emotive theory" of value have tended to bypass it. Thus, for example, Stevenson in his, at the time, influential book Ethics and Language. Westermarck's answer1 to his question was that moral approval is impartial (disinterested) and, in a certain sense, universalizable. These ideas are reminiscent of the ethics of Kant and also of the Golden Rule of Christian morality - two authorities in moral matters which Westermarck otherwise heavily censured. The two traits mentioned do not conflict with the subjectivity and relativity of moral valuations. But they challenge a question which Westermarck never raised: What is the logical nature of the statement that disinterestedness and universalizability make approval and disapproval moratf Is it a descriptive statement about the use of moral language? Then it can hardly claim to hold good for all societies throughout their history. Or is it an evaluative or normative statement urging us to view morality in a certain light? If so, then Westermarck's characterization is a contribution to what is also called normative ethics. It was certainly not meant to be so. But I think it is thus we have to understand it. The first contribution I personally made to ethical theory was a paper from the early 1950s "Om moraliska föreställningars sanning" (On the Truth of Moral Ideas}.2 The title was a conscious allusion to a famous lecture by Axel Hägerström, Sweden's "opposite number" to Finland's Westermarck. The paper contains what still seems to me a moderately good interpretation of the core of the two "giants'" moral philosophy but also an attempt to "rescue" the value-subjectivism of Westermarck and the value-nihilism of Hägerström from the swamps of moral anarchy and egoistic subjectivism. It also made use of ideas which I had learnt from Wittgenstein. I argued that a great many, perhaps a majority of judgements in the moral sphere subsume individual cases under accepted standards (criteria) of goodness or badness and thus are in a relative sense true or false. The standards themselves cannot be assessed from the point of view of truth, however. 4 It took another ten years before I presented to the public a major contribution to the philosophy of value. This was my book The Varieties of Goodness, based on the Gifford Lectures which I had given at St. Andrews.3 'Westermarck (1924). The moral emotions, Westermarck says, are "uninfluenced by the particular relationship in which they stand, both to those who are immediately affected by the acts in question, and to those who perform those acts". And the subject who makes a moral judgement "feels that it would be shared if other people knew the act and all its attendant circumstances as well as he does himself. 2 Von Wright (1954a), pp. 48-74. 3 Von Wright (1963a).

14

Georg Henrik von Wright

It was published in 1963. I had initially wanted to call it "Prolegomena to Ethics" but, as is well known, this title had already been used (Thomas H. Green: Prolegomena to Ethics, posthumously published by A. C. Bradley). An idea in Varieties which still strongly appeals to me is that notions central to the moral life of man such as good and evil, virtue and vice, justice and injustice, are concepts in search of a meaning. Although familiar from daily life they are at the same time obscure and vacillating. There may exist wide consensus about how to use them - but there is also much disagreement and controversy about their application to individual cases. In the "intellectualized" later stages in the history of a civilization it therefore becomes the task of philosophers to reflect on the meaning of moral ideas. The task is not descriptive. But neither is it a matter of mere stipulation. The philosopher's meaning-giving activity has a practical aim. It aims at being a guide for how to judge and evaluate people's conduct (including the philosopher's own). This is how I understand normative ethics. "Normative" does not here mean "legislating". Ethics cannot impose obligations or prohibitions. But it can propose standards of valuation in conformity with which obligations and prohibitions may then be imposed (or lifted) by an enlightened legal or moral authority. In Varieties I made a modest contribution to satisfy my own craving for the "meaning of morality". It can be said to have been in line with Westermarck's in that it tried to combine some Kantian elements with the Christian idea of agapeistic care for one's neighbour's good. I regarded a Principle of Justice as the cornerstone of morality. The moral duty is to act in accordance with this principle from what I called a moral motive, viz a will to secure for all a good which similar action on the part of our neighbours would secure for us. The moral will is beyond egoism and altruism - a disinterested and impartial will to justice. It treats your neighbour as though his welfare were yours and your welfare his. To have this attitude is to love your neighbour as yourself. - This still remains, in outline, my position. 5 After Varieties I made a few attempts to connect the idea of morality with that of rationality or a life in accordance with reason. A first attempt in this direction was a paper "Rationality: Means and Ends" of 1986.4 In it I examined the two forms of rationality which Max Weber distinguished by the names Zweck- and Wertrationalität. Aristotle had said that only means can be objects of deliberation, and that one cannot deliberate about ends.5 I made an effort to show in what sense deliberation about ends is possible and did this by relating the problems to another idea of Aristotle's, viz. the idea of phronesis or "practical wisdom". I disagreed with 4

Von Wright (1986), pp. 57-71. Aristotle: Nicomachean Ethics, III, iii, 11.

5

Value, Norm, and Action in My Philosophical Writings

15

Sir David Ross's definition of phronesis6 as "knowledge of how to secure the ends of human life". This is not what Aristotle meant when he characterized it as a "true and reasoned state of capacity to act with regard to the things which are good or bad for man" 7 In an unpublished recent paper I tried to connect the idea of phronesis with that of personal identity. The identity of a person consists in his knowing who he is, where he belongs, what he can do and what is beyond his power, what he is entitled to demand and expect of others, and others of him. A person's identity or self-knowledge, I argued, is incomplete, deficient unless he realizes the symmetry in the relationship between humans with regard to how their actions mutually affect their good (wellbeing). Inbuilt in the possession of self-knowledge is a moral attitude to our neighbour. It may be true to say that my efforts to connect, on the conceptual level, morality with practical wisdom (phronesis), life in accordance with reason and enlightened self-knowledge (identity), reflect a certain tension between subjectivist and objectivist tendencies in my thinking about value. It is also true that the "objectivist pull" has become stronger with age. But I continue to regard this as fully compatible with the subjectivism and relativism in ethics which, in my formative years, I learned from Westermarck. It is therefore with some dismay that I notice signs of a return to objectivist and absolutist value theories in the last few decades. They are in some quarters hailed as a remedy for the value chaos or vacuum which has followed in the wake of secularization of society and decay of traditional morality. But this, I think, is a mistake. 6 In 1952 I published a paper with the title "On the Logic of Some Axiological and Epistemological Concepts" .8 It was my first attempt to develop a formal logical theory of value. The inspiration to write it had little, if anything to do with interest in ethics or general theory of value. Its source was rather my "invention", a little earlier, of deontic logic. This was a formal theory of normative discourse - and I was now looking for a counterpart to it in evaluative discourse. But the attempt turned out to be abortive. I doubt whether anybody in addition to the author of the paper has ever gone through its rather tough formal arguments. And I am not sure that I any longer share the linguistic intuitions which the paper tries to formalize. In the years following, my interest was caught by another evaluative notion, preference. To prefer something to something else is to value or rate the first thing higher than the second. The preference relation is irreflexive and asymmetrical. Is it also transitive? And is it connected, i.e. is it so that if 6

Nicomachean Ethics of Aristotle, Translated and Introduced by Sir David Ross. The World's Classics, Oxford University Press, 1954, p. 142. 7 Nicomachean Ethics, VI, v, 4. "Von Wright (1952a), pp. 213-234.

16

Georg Henrik von Wright

a first thing is preferred to a second, then any third thing is either preferred to the second or the first to the third? Perhaps the questions had more than one answer. Thinking, "on and off', about them, eventually resulted in The Logic of Preference? published in the same year, 1963, as The Varieties of Goodness and Norm and Action. This attempt, unlike that of 1952, to create a formal logic of value concepts was not abortive - even though it could claim to cover only a part of a vast field. At long last I had found a sought-for counterpart in value theory to deontic logic in norm theory. I also proposed a "parallel" name for the new creation, Prohairetic Logic. Unlike its counterpart in the theory of norms, this label has not gained wide acceptance. In the early 1970s I returned once more to the topic of preference and wrote a paper "The Logic of Preference Reconsidered".l° It is a difficult paper, the arguments being very compressed and tense. It contains, however, some new ideas and a further elaboration of old ones - and may still have something valuable to offer. One of its main thoughts is a distinction between two notions of indifference between options - one meaning simply "no preference either way" the other, a stronger one, meaning "rated equal" or "of equal value". Before leaving this topic, I must mention that my ideas about a logic of the preference relation had been anticipated by Sören Hallden in his book The Logic of 'Better' which appeared in 1957.u I knew of its existence but was unfamiliar with its content when, in haste, I wrote The Logic of Preference. It had not even occured to me that "better" is (almost) another word for "preferred". This is an example among others of my failure to note and duly acknowledge related work done by my contemporaries. I shall soon have more to say about this. (Below II, 2) Hallden's thoughts fell on fertile soil. His pupil Bengt Hansson and, following him, Peter Gärdenfors made important contributions to the further development of formal preference theory and not least to the debate stirred by what is known as Arrow's impossibility Theorem.

1 My interest in norms has a very different origin from my interest in questions of value. The latter can be said to have been part and parcel of my early attraction to philosophy. In my pre-student days I had not only implicitly been imbued with the spirit of Westermarck's ethical subjectivism 9

Von Wright (1963b). Von Wright (1972), pp. 140-169. n Hallden (1957). 10

Value, Norm, and Action in My Philosophical Writings

17

and relativism, but had also been an enthusiastic reader of texts by Schopenhauer, Nietzsche, and Guyau, It took, however, a long time for this trait in my philosophic character to mature. This was no doubt partly due to lack of early academic training in the field. (Above I, 1) I came straight to norms from a professional occupation with logic. After having finished my doctoral work on induction and probability, I became interested in the Entscheidungsproblem of the predicate calculus. I wanted to show that Wittgenstein's idea in the Tractatus of logical truth as tautology could be extended beyond propositional logic. This demanded a new theory of quantifiers. And one day it occurred to me, "almost out of nowhere", that the modal words "possible", "impossible", "necessary" are mutually related to each other in the same way as the quantifiers "some", "no", and "all". I knew nothing about modal logic then - but boldly set myself the task of constructing one. I was aware of Lukasiewicz's use of modal ideas for his polyvalent logic, and I soon also discovered the subject's chief modern classic, C. I. Lewis. But I remained ignorant of what Carnap and Ruth Barcan had been doing just a few years earlier. It was in the course of this work that - in conversation with friends about ethics - I was struck by the observation that the normative ideas of the permitted, the forbidden, and the obligatory seemed to obey the same patterns of interdefinability as the basic modalities and quantifiers. I quickly wrote a paper about this observation and sent it to Mind, where it soon appeared. Shortly after, Oskar Becker's Untersuchungen über den Modalkalkül appeared in Germany12 and Jerzy Kalinowski's "Theorie des propositions normatives" in Poland.13 These three independent publications mark the birth of the study now commonly known as Deontic Logic. Later I criticized the analogy between modal and deontic ideas. Its sore point is the presumed identity "not-permitted" = "forbidden" and the consequent interdefinability of the deontic operators and P. I think the analogy is untenable and that the identities which are extracted from it depend on a confusion between genuine norms and norm-propositions or statements about the existence or non-existence of norms. From Norm and Action (1963)14 on I have treated obligation and permission, and P, as the two basic normative categories, - prohibition being defined as an obligation to omit. But I think it is right to say that a majority of deontic logicians continue to stick to the interdefinability thesis. This is not the only position in deontic logic where I have with time come to be in the minority - in some cases perhaps a minority of one. 12

Becker (1952). Kalinowski (1953), pp. 17-53. 14 Von Wright (1963c). 13

18

Georg Henrik von Wright

It is indeed an ironic thought that the idea from which deontic logic took its start was in fact a mistake! 2 Much of my creative work in logic I have done in ignorance of predecessors and contemporary workers. When embarking on modal logic, I thought I was entering a more or less virgin land. This was true, at most, of the off-shoot of it represented by deontic logic. But even this has an ancestry which later research has brought to light. Similar ignorance had also been characteristic of my occupations with the Entscheidungsproblem. When in 1948 I presented a solution to the problem for monadic predicates,15 I did not know that Quine a few years earlier had published a substantially similar solution.16 (We were presumably both ignorant of the fact that decades earlier Behmann had solved our problem.) And when I later solved the decision problem for what I called "double quantification",17 the reviewer (Ackermann) in the Journal of Symbolic Logic pointed out that my solution could be obtained from earlier results of Gödel, Kalmar, and Schütte.18 What should one think about such ignorance? When one enters a new field of research and believes one has an "original" solution to a problem in that field, would not the right thing be first to familiarize oneself with the literature and ascertain whether the problem had perchance already been solved? Had I followed this recipe, I would have avoided many blind alleys and, possibly, some mistakes. But it has also seemed to me that a new solution to an already solved problem can be of value both because it may have been reached through a new type of approach and because it may direct future research towards a new direction. It is also a fact about me personally that whenever a new thought dawns on me, working it out absorbs for a time all my energies and closes my mind to outside influences. (I simply shut them out.) And when I am finished with a problem, I often lose interest in it and am therefore not anxious to dig into its prehistory. This, however, is not always the case. Sometimes my curiosity is raised. What have others had to say about the same topic? Then I have often done quite extensive reading and this may have resulted in adding assenting or dissenting comments to a manuscript still in statu nascendi. Reviewers have sometimes said that I have been "influenced" by one or another of the writers thus referred to. This has usually not been the case, and when it has, I have said so explicitly. I have acknowledged the influence of Elizabeth Anscombe and Charles Taylor on my thought in Explanation and Understanding. On the other hand, there was hardly any influence from my readings in the 15

Von Wright (1947). Quine (1945). 17 Von Wright (1952b). 18 Journal of Symbolic Logic 17 (1952), pp. 201 f.

16

Value, Norm, and Action in My Philosophical Writings

19

German tradition in writing this book. Only afterwards did I discover that E&U could be seen as an attempt to "bridge the gap" - still very much open at the time when the book was published - between an Anglo-American analytic and a Continental-European hermeneutic tradition. 3 The Swedish philosopher Ingemar Hedenius was, if not the first to note, at least the first critically to exploit the distinction between norms and statements to the effect that such and such norms exist.19 The distinction can be said to be hidden in ordinary language. A deontic sentence, for example "you may not park your car here", can either be used for issuing a prohibition to someone or for informing him of existing parking regulations. If the sentence is used in the first way, it is used (interpreted, understood) prescriptively, if used in the second way it is used descriptively. It is a time-honoured opinion that norms proper have no truth-value. This opinion seems to me indisputable, if understood to mean that deontic sentences, prescriptively interpreted, do not express true or false propositions. Views which attribute truth-value to norms, as for example some "natural law" doctrines do, seem to me guilty of a confusion between norms and normpropositions. The plausibility of such views then depends upon the criteria which they employ for the existence of norms. If norms have no truth-value, how can logical relations such as contradiction and entailment (consequence) obtain between them? Is a deontic logic possible? To begin with I assumed, naively, that the mere fact that one can construct a formal calculus with intuitively plausible sounding axioms and rules of inference was enough to show that the answer to the above question was affirmative. Or even to show - as I said in the Preface to my Logical Studies (1957) - that logic has a wider reach than truth. 20 And this seemd to me to be an achievement of considerable philosophical interest. But my judgement then was certainly too hasty. On the other hand, a syntactic structure of deontic sentences and their molecular compounds, such as my 1951 system, might, on a descriptive interpretation of the sentences, be a Logic of Norm-propositions. Was this what deontic logic really amounted to? In Norm and Action (1963) I nearly subscribed to this opinion. But I also felt confused. 4 In the 1951 system one can prove the formula Pp V O~p. It says that any state of affairs has a deontic status, is either permitted or forbidden. A normative order or system for which this holds true is said to be complete or 19 20

Hedenius (1941). Von Wright (1957), p. vii.

20

Georg Henrik von Wright

"gapless". One can also prove the formulas ~(ΟρΛθ~ρ) and ~(ΟρΛΡ~ρ). The first says that no state of affairs is such that it and also its contradictory are obligatory; the second that it cannot be the case that the contradictory of an obligatory state is permitted. A normative order for which this holds true is said to be consistent or contradiction-free. Interpreted as a logic of norm-propositions, the 1951 system would thus be valid only for such normative orders which possess the "perfection-properties" of gaplessness and freedom from contradiction. But it seems a notorious fact that, for example, many legal orders both contain gaps and mutually contradictory norms. Kelsen had thought that legal orders "of norm-logical necessity" are complete and consistent. Therefore he was, for a short time, enthusiastic about deontic logic, since it seemed to him to prove two of the meta-juristic ideas characteristic of his thinking.21 But he was mistaken. The "logical moral" which follows from the de facto existence of gaps and contradictions in some normative orders is that the 1951 system could not pretend to be the uniquely correct Logic of Normpropositions. And in view of the fact that the norms themselves are neither true nor false, it was at least doubtful whether this system could claim to be valid as a Logic of Norms either. 5 The first ones to see clearly the necessity of distinguishing between a logic of norm-propositions and a logic of norms (a "real" deontic logic) were, I think, Carlos Alchourron and Eugenio Bulygin, co-authors of their classic Normative Systems,22 A few years prior to the book's appearance, Alchourron had made his first attempt to construct a Logic of Normpropositions - or as he called it "Normative logic" - in contradistinction to a Logic of Norms.23 It is a great merit of the two authors to have understood this double task of a logical study of norms. It has taken me nearly thirty years to see its full significance. And not all logicians and philosophers have seen it even now. 6 After Norm and Action the question "Is a Logic of Norms possible?" continued to torment me. But now I think that at long last I am able to justify an affirmative answer. I shall try briefly to summarize my final position.24 On the descriptive interpretation of deontic sentences, their molecular compounds are themselves sentences expressing true or false propositions. 21

On Kelsen's position and my meeting with Kelsen in 1952 see my paper "Is and Ought". Von Wright (1985a), pp. 263-281. 22 Alchourron/Bulygin (1971). 23 Alchourron (1969). 24 The account given here essentially follows the one given in my paper "Is there a Logic of Norms?" (1991) and its revised version included in my collection of papers Normen, Werte und Handlungen (1994). See also the introductory lecture "Ought to Be - Ought to Do", this volume.

Value, Norm, and Action in My Philosophical Writings

21

The compounds are, in other words, truth-functional, and have descriptive meaning. But according to the prescriptive interpretation of deontic sentences, molecular compounds formed by them with the aid of sentential connectives do not express norms, and have no prescriptive meaning. The connectives, in other words, do not apply to norms, i.e. to prescriptively interpreted deontic sentences. (There are a few apparent exceptions to this. I have elsewhere called them "semantic accidents". I do not need to consider their case further here.) It follows from what has been said that it is not possible to construct a Logic of Norms in the form of a syntactic structure of sentential variables, connectives, and deontic operators. The deontic sentences in such a structure can only be given a descriptive interpretation. This means that the structure itself is, at most, a Logic of Normpropositions and not a Logic of Norms. Does this, if accepted, not wreck the whole enterprise of building a "real" deontic logic? I do not think it does. There is another approach to the problem beside the syntactic one. I shall call it the semantic approach. It is as follows: I shall say that a set of norms is consistent if, and only if, the conjunction of all things (norm-contents) which the norms in the set pronounce obligatory with any one of the things which they permit is a doable state of affairs, i. e. is something which can be achieved as a result or consequence of human action. A set which is not consistent is inconsistent or self-contradictory. According to this definition, to take a very simple example, the two norms Op and O~p constitute an inconsistent set. This is so because the conjunction of their contents, ρ Λ ~ρ is not a doable state of affairs. That ρ and ~ ρ are inconsistent with each other is a fact of logic. But why should we call Op and Ο ~ ρ mutually inconsistent? I do not think we can answer the question satisfactorily, unless we consider the purpose or rationale of norms and norm-giving activity. A norm-authority (law-giver) who enjoined his norm-subjects to see to it that ρ and also, at the same time, that ~p, would be "crying for the moon", asking for something which cannot be, and therefore would be behaving irrationally. But there is nothing irrational about one authority enjoining that ρ and another authority that ~ p. In calling Op and Ο ~ ρ mutually inconsistent, the sameness of the norm-giver and the occasion for acting is presupposed. Next we introduce the notion of a negation-norm of a given norm. The negation of an obligation Op is a permission Ρ ~ p. And the negation of a permission Pp is an obligation O~p, i.e. the prohibition of p. This answers to a common prescriptive understanding of the phrases "not-obligatory" and "not-permitted". We can now also define the notion of normative entailment. I shall say that a consistent set of norms entails a further norm when, and only when,

22

Georg Henrik von Wright

adding the negation norm of this further norm to the set makes the set inconsistent. To take a very simple example: Op entails Pp. The negation of Pp is O~p. The set consisting of Op and O~p is inconsistent. Hence obligation ("ought") entails permission ("may"). Deontic logic, as I see it now, rests on the possibility of defining the notions of normative consistency and entailment in ways which seem to be in good accord with our intuitions. It took me some thirty years to reach this position. It is unlikely that I shall change it again, except perhaps in details. The seeds of my view are already in Norm and Action. In the 1980s my views radicalized and became more and more like the "nihilism" of Hägerström and of Kelsen in his later years. Like them I came to hold that in the traditional sense "genuine" logical relations cannot obtain between norms. But unlike them I thought it possible to "transcend" logic and apply the notions of consistency and logical consequence beyond the borders of the true and the false, thus vindicating the original intuitions behind my essay of 1951. 6 In my introductory paper at the Bielefeld symposium I dealt with some other central problems which had intrigued me over the years. They have to do with the controversial subject of hypothetical norms and the important distinction between what I have called "the ought of obligation" and "the must of practical necessity". I shall not here repeat what I had to say about these topics but refer the reader to the slightly abbreviated version of the paper which is included in the present volume.25

1 My interest in action was an immediate consequence of the invention of deontic logic. The "things" which norms pronounce obligatory, permitted or forbidden are, broadly speaking, either actions or "doable states of affairs", i.e. states which can be achieved, brought about as the results or consequences of action. An action of the achievement type has what I called an "inner" and an "outer" aspect. The latter consists in changes (or not-changes) in the physical world - immediately in behavioural reactions in the agent's body. The former is something "mental", the intention or the will of the agent to achieve a change or ensure that something remains unchanged. My concern with action focused to begin with on the external aspect, a study of change. It resulted in my first attempt, in Norm and Action, to 25

The paper was subsequently published under the title "Ought to Be - Ought to Do" in Garzon Valdes et al. (1997), pp. 427-435.

Value, Norm, and Action in My Philosophical Writings

23

create a Logic of Action. But I soon came to take an interest also in the mental or "volitional" aspect of action. This resulted in what later became the main theme of my action-theory, viz. the explanation of action. I dealt with it in Explanation and Understanding (1971)26 and later work. Let "p" represent the state that a certain window is open. This state can be there as a result of two different actions. One is the action of opening the window, i.e. changing its state from closed, ~p, to open, p. The other is the action of preventing the window from closing, for example because someone else is trying to shut it. The two actions may differ in "deontic status". The first may be permitted or even obligatory, the second strictly prohibited. But this difference could not be expressed in "classic" deontic logic. Its expression required a symbolism for change. Let "~p Tp" stand for the change from window closed to window open, and "p Tp" for the window staying open. If these two events are thought to result from action, then "O(~p Tp)" stands for the obligation to open the window, and "O ~(p Tp)" for the prohibition to prevent it from closing. (This is only a first step to a full symbolism, but will suffice for present purposes.) The upshot of such observations as these was that a fully developed Logic of Norms has "to stand on the shoulders" of a Logic of Action, which in its turn must be based on a Logic of Change. But these missing foundations of deontic logic had yet to be created. In Norm and Action I embarked upon the task. I thought I had to start from scratch. As far as Logic of Action was concerned I was presumably right. I do not know of any systematic earlier efforts in the same direction. And thirty years and more later I still have the impression that the formal logical study of action remains relatively little developed and is sometimes unduly neglected. My work in action logic after Norm and Action aimed at constructing an axiomatic theory for the four atomic or elementry types of action which have the nature of achievements, viz. productive, preventive, sustaining, and destructive action and the corresponding omissions of their performance. My most recent contribution is the essay "Action Logic as a Basis of Deontic Logic" from 1988.27 I think it will also be my last. Others who have worked in this field have, on the whole, not built on the "tetradic" foundation of elementary modes of action. An interesting approach of a somewhat different nature is Krister Segerberg's dynamic logic. Another is Nuel Belnap's See-To-It-That combination of actionist and normative ideas. I was glad to find both writers among the contributors to this volume. 26 27

Von Wright (1971). Von Wright (1988).

24

Georg Henrik von Wright

With the bottom level of my new construction, change logic, things stood somewhat differently. Its origin is another example of my frequent ignorance of parallel work done by other logicians. My Logic of Change turned out to be but a small fragment of the work in Tense-Logic, which had been initiated some years earlier by Arthur Prior.28 I had started with a theory of an asymmetric conjunction "and next". It captures transitions (progressions, changes) from one state to another in a discrete time-medium. Shortly after, Segerberg produced an improved version of it under the title "On the Logic of 'to-morrow'" ,29 My next target was another asymmetric conjunction "and then". It concerns progressions of unspecified length in a time-medium which need not be discrete. Its axiomatization turned out to be much trickier than that of "and next". I was intensely occupied by it during a trimester at the University of Pittsburgh in the mid-sixties. This was an exceptionally fertile period in my creative life. It resulted, among other writings, in An Essay in Deontic Logic and the General Theory of Action (1968).30 Prior visited Pittsburgh and gave a lecture. Thanks to him I realized that my work with the two asymmetrical "ands" had been my lonely and maybe unnecessarily cumbersome way to tense-logic. But I did not regard this as a waste of time and energy - far from it. Thus the study of change, originally a preliminary to the study of norms and actions, awakened my interest in what I would myself call the Logic of Time rather than use the language-oriented term Tense-Logic. For some years I continued work in this new direction. If I contributed anything of original value it was, perhaps, the beginning of a study of "time-division" - the "chopping up" of a stretch of time into smaller and smaller parts in contrast to "adding up" such stretches in progressions towards the future and the past.31 I also ventured into the spatial dimension, into something 1 called the Logic of Place.32 Ultimately, these preoccupations stimulated my interest in various kinds of non-classical, deviant logics.33 They signal, I believe, the end to my itinerary in logic. 2 Closely parallel in time with my efforts to base deontic logic on a logic of action, was my work on Practical Inference. The inspiration, if I remember correctly, came from my study of Aristotle when I was preparing The Varieties of Goodness - and it was greatly supported by Elizabeth Anscombe's masterpiece Intention (1957). 28

0n it and its relations to some work done by me see Prior (1967). Segerberg (1967). 30 Von Wright (1968). 31 Von Wright (1969). 32 Von Wright (1983a). 33 See my paper "lYuth-Logics" (1987), pp. 311-334. 29

Value, Norm, and Action in My Philosophical Writings

25

A practical argument (inference, syllogism) has two premises. One is evaluative; it refers to an end which is coveted (wanted) by an agent. The other is epistemic; it mentions what the agent thinks about means to this end. The conclusion, finally, consists in the agent's proceeding to action, i.e. to using the means in order to attain the end. - This is only a sketchy initial description; there are alternative and different ways of describing basically the same structure. It took me, for example, some time before I realized that Anscombe's understanding of the Aristotelian doctrine was rather different from my own.34 Is practical inference a logically valid argument? The answer, of course, depends upon how one states the premises and conclusion in more precise terms. The critical question is whether there is some kind of necessary connection (entailment) between, on the one hand an agent's ends and estimated means and, on the other hand, his actual efforts to attain his ends. In countless discussions with Norman Malcolm35 and Fred Stoutland the three of us tried to formulate a plausible version of what Stoutland had called "the logical connection argument" .36 A practical argument is forward-looking in the sense that it proceeds (in thought) from the background of an action in ends and means to the projected performance of the action itself. It is related to prediction. One can, however, reverse the order and argue from the performance of the action to its background in ends-means considerations. Then the schema becomes related to the explanation of action. This does not mean that one could infer from the conclusion back to the premises. But it means that one can refer to the premises as reasons for the action. Explanation of action has been one of the main themes in my thinking in the last thirty years, or so. I am afraid that it retarded the maturation of my thoughts that I long tended to view the problems in the prospective light of action-prediction rather than the retrospective light of action-explanation. 3 Are the reasons for an action causes of the action? The question is not about linguistic usage. Reasons for actions are commonly (also) called "causes". This is in good order. What I have been anxious to do when answering the above question negatively is to emphasize a conceptual difference between explanation of intentional action in terms of reasons (and motives) and the explanation of natural events in terms of cause and effect. Therefore, I have distinguished between rational explanation of action and causal explanation of natural events. The general schema of a causal explanation is the deductive-nomological or Covering Law schema. It subsumes a token pair of events under a general 34

See Anscombe (1989), and my reply to Anscombe. See Malcolm (1989), and my reply to Malcolm. 36 See Stoutland (1970). 35

26

Georg Henrik von Wright

rule or law correlating types of the events with the tokens of their instantiation. The schema for a rational explanation is a reversed practical inference, one could say. It understands an individual action against the background of its reasons, i.e. the agent's wanted ends and estimated means for their attainment. In Explanation and Understanding and later I defended the conceptual autonomy of rational explanations of actions as against causal explanations of events. For some time I argued that what the covering law schema was for the natural sciences, the practical inference schema is for the human sciences.37 This was an exaggeration - though I think it contained a germ of truth in it. My position was a defence of the autonomy of the human sciences (the humanities or Geisteswissenschaften) in relation to the natural sciences (Naturwissenschaften). It thus stood in opposition to the "unity of science" ideas of the logical positivists and later empiricists. My "dualist" view can be related to the Erklären-Verstehen distinction and the debate familiar from German philosophy round the turn of the century, revived in more recent "hermeneutic" philosophy. (See above II, 2) The distinction between explanation and understanding, however, is far from always clear and sometimes it is confusing. In more recent years I have therefore focused on another dichotomy which seems to me to cut deeper. It is the division between knowing truth and understanding meaning or significance. Knowing and understanding, I would say, are the two basic categories or dimensions of the cognitive life of man - and it has been to the detriment of our intellectual culture that is has overemphasized knowledge at the expense of understanding. It was in "Münsteraner Disputationen" with Georg Meggle in the late 1980s that there took place a decisive turn in my thinking from the explanation-understanding-thematics to the relation between knowing and understanding. I tried to defend in our talks - against Meggle - the view that understanding is not a form of knowledge (Wissen) - although it presupposes knowledge (for example of what will count as an intelligible explanation of an action in terms of its reasons). Understanding so to speak "makes its own truth" in the concrete act of connecting the understanding mind to its object. I hope that I shall still be able to give to my insights in these relationships a clearer and more convincing expression.38 4 In Explanation and Understanding I also defended what is sometimes called an actionist or interventionist view of causal or nomic connections in nature. The question was: Can one distinguish between mere regular sequence or concomitance of events - the Humean view of causation - and some stronger bond between them which deserves perhaps to be called nat37

Von Wright (1971), p. 27. Von Wright/Meggle (1989); reprinted in von Wright (1994).

38

Value, Norm, and Action in My Philosophical Writings

27

ural necessity? In E&U I argue for this distinction. In a much simplified form, my argument can be stated as follows: It is the hallmark of a nomic connection between two events, E\ and EI, that it supports contrafactual statements of the following type: If E\ had happened on an occasion when in fact it did not happen, then E? would have happened too. But how can one make it plausible that this is so? My answer was: By producing E\ in a situation when we are reasonably sure that neither it nor EI will not occur, as we say, "by themselves" or because of some already known cause, and then noticing that EI too makes its appearance. This presupposes that E\ is a "doable state of affairs" and that we master the technique of producing it. The presupposition is normally fulfilled in the important type of human action we call a (scientific) experiment. My theory could therefore also be called an experimentalist theory of causation. It is easily misunderstood. It makes causation conceptually dependent on action, - but it does not mean that every time when something causes something else the cause is produced by some agent. (This only seldom happens.) An objection has been that the theory holds, at most, for terrestrial phenomena, accessible to manipulation, but cannot claim validity for the physical universe as a whole. To this one can say two things. First, that lawlike connections in cosmic dimensions - for example relating to the origin of the universe - have a speculative character which may make the distinction between the accidental and the nomic difficult or impossible to apply to them. And second, that many of the laws which we do not hesitate to hold true for phenomena beyond the reach of action - for example in geology or in theories of evolution - are in the last resort extrapolations from findings under conditions when experimental manipulation was possible. A theory of natural necessity similar to my actionist view is not unknown from the literature. It is reminiscent of the view held by Collingwood and, in more recent times, by Gasking.39 It was gratifying for me to notice that some of the contributors to the present volume, Egidi, von Kutschera and Schneider, seem sympathetic to it - at least "mildly". 5 Of all books I have written, Explanation and Understanding is perhaps the most challenging or even controversial. For me it constituted an urge to return to its main themes - action and causality - trying to state my positions clearer and to modify and develop them further. These endeavours resulted in two new books. One is called Causality and Determinism and is based on the Woodbridge Lectures which I gave at Columbia in 1972.40 The other has the title Freedom and Determination (1980) and remains, I should say, the fullest and best-argued statement of my position in the philosophy 39 40

Gasking (1954). Von Wright (1974).

28

Georg Henrik von Wright

of action.41 It represents an advance from my earlier views mainly thanks to a distinction between internal and external determinants of action. It takes into account the dependence of the agent on conditions, institutions, and practices in the society of which he is a member. With this the problem of human freedom alters and becomes a major problem of social philosophy. I regret that I never followed up this line much further. It would have taken me to the neighbourhood of the Hegelian-Marxist tradition and the "critical theory" of the Frankfurt School and of Habermas. When I returned to the problem of freedom in my Tanner Lectures of 198442 my thoughts had already started to move into a different field of inquiry which no longer falls under the rubric "Actions, Norms, and Values". But since some very interesting contributions - Gjelsvik, Kambartel, Schneider - to the Bielefelder Proceedings follow my thoughts in this new direction, I shall conclude my paper by a short and much simplified comment on it.

IV 1 I distinguished (above III, 1) between the outer and the inner aspect of an action. The first is a change (or a not-change) in the outer world. The second is the mental background of this change in the volitions, intentions, motives, and reasons of an agent, the "operations" of which, as we say, brought about the change. The inner is here, in some sense, the cause of the outer and the outer the "output" of the inner. Thus: mental (psychical) cause - physical effect. There is also the reverse case. An event in the physical world, for example a sound, affects a sense-organ and "calls forth" a sensation. Here the outer thing is the input and the inner the output. Physical cause - mental (psychic) effect. This description of psycho-physical interaction slurs over a number of grave problems. Since the days of Descartes they have been of main concern to Western philosophy. Descartes made a sharp distinction between the mental and the physical, mind and matter. But he also thought that the two causally interact. Is this really the case? And if so, how shall we understand it? It is no exaggeration to say that the major metaphysical positions defended by philosophers since Descartes have been attempts to answer the above questions. In this sense one can say that Descartes overshadows the philosophic landscape from his own day up to the present. Some of Descartes' immediate successors denied interaction between mind and matter and thought that their "matching" was something occasioned 41 42

Von Wright (1980). Von Wright (1985), pp. 107-170.

Value, Norm, and Action in My Philosophical Writings

29

by a divine will; Leibniz spoke of it as a "preestablished harmony". Others suggested a reductionist solution - either a materialist reduction of mind to matter as with Hobbes, or an idealist one of matter to mind as with Berkeley. Others still, the monists, thought of the two substances as one. Thus Spinoza in the 17th century and the empirio-criticists in the late 19th. Not long ago interactionist dualism experienced a revival taking developments full circle back to the neighbourhood of Descartes.43 2 Serious thinking about these questions must take into account the role played by the nervous system (the brain). Our initial description above said nothing about this. The volitional (motivational) forces behind an action do not immediately bring about the behavioural (bodily) changes in the world which constitute the first stage of its outer aspect. What they call forth are neural changes which in their turn are causally responsible for the behavioural aspect of the action. Similarly, the affectation of a sense-organ by a stimulus does not immediately result in a sensation but in neural processes which are then "transformed" from a central state in the brain to a sensation. The neural processes - with or without additional physical stimulation - may also causally produce behavioural reactions which would show ("mean") to an outside observer that the subject has the sensation. (For example, his getting up from a chair in response to a knock on the door.) The "mysterious" nature of the transition from neural process to mental sensation is responsible, I think, for a temptation to say that the sensation and the neural state which is there when the subject has the sensation are, somehow, the same or identical. And there is a similar temptation to identify the volitional mental background of an action with the neural states and processes which are causally responsible for the behavioural part of its outer aspect. To yield to these temptations is to embrace an identity theory about the mind-body relationship. But the identification results from confusion. 3 A way to clarify these matters involves a detailed mapping of the interplay of the three main factors of the case: the mental, the neural, and the behavioural. Characteristic of their interplay, as I see it, are three different relations of conceptual priority. I shall call them epistemic, causal and semantic. The mental, I shall say, is epistemically prior to the neural. By this I mean the following: In order to get to know the neural counterparts of mental states (phenomena) we must be able to identify those states - for example hearing a sound, feeling pain, being afraid, believing or expecting something 43

Cf. Popper/Eccles (1977).

30

Georg Henrik von Wright

- using criteria which are independent of their neural basis. Because otherwise one could not locate the neural events among which one has to look for counterparts or equivalents to the already known mental ones. - We would, for example, not know anything about a "pain-centre" in the brain, unless we had experimented with subjects known to feel pain as a consequence of certain physical stimuli being applied to them. Further, there is a causal priority of the neural in relation to the behavioural. This is a straightforward matter of scientific truth. Bodily reactions of the kind we call behaviour - not for example unconditional reflexes - are caused by efferent impulses from motor centres in the brain. Finally, there is what I propose to call the semantic priority of the behavioural in relation to the mental. It consists, roughly speaking, in this: that the behavioural reactions caused by the neural equivalents of the various mental states constitute what it means to say that the subject experiences ("suffers") those states. This semantic involvement of the behavioural with the mental is a more complex, and therefore also more controversial, relationship than the one between the mental and the neural and the neural and the behavioural. 4 Do body and mind interact then? In some, philosophically noncommittal, sense they certainly do. Of course, physical stimuli of sense-organs call forth sensations, and reasons move agents to act. Descartes, who believed in body-mind interaction, was also worried by the question whether this was compatible with strict determinism in nature. One way to rescue compatibility is to subscribe to the following Principle of the Causal Closedness of the Physical World Order: Whenever an event in nature E can be rightly said to cause a mental event M, there exists another physical event F such that E causes F and M temporally coincides with (the whole or a part of) the duration of F. And when, conversely, some mental event M can be rightly said to cause something in the physical world E, then there exists another physical event F such that F causes E and the temporal duration of M is included in that of F. Accepting this idea, which seems to me plausible, one can replace talk of body-mind and mind-body interaction by talk of causal connections between physical events only. Using terminology now in fashion, one can say that the Causal Closedness of the Physical World Order means that causal relations between mental and physical phenomena are supervenient in time on causal relations the terms of which are physical phenomena only. The suggested position comes as "near" the psycho-physical identity thesis as is, in my opinion, "logically permissible". It is also reminiscent of versions of the position known as epiphenomenalism.

Value, Norm, and Action in My Philosophical Writings

31

5 The body-mind distinction, as I see things, is not a sharp divide. The epistemic priority of the mental in relation to the neural, and of the behavioural in relation to the mental, mean that mind and body are conceptually intertwined in the groundwork of reality. Attaining this insight has meant easing myself of what I felt to be a Cartesian load.44 Also the three major efforts in the philosophy of this century, those of Husserl, Heidegger, and Wittgenstein can, in their different ways, be seen as efforts to come out from under the shadow of Descartes. Perhaps I may be allowed to say about myself that I have tried to follow an independent track in the same direction.

References ALCHOURRON (1969). Carlos E. Alchourron: Logic of Norms and Logic of Normative Propositions. Logique et Analyse 12 (1969). ALCHOURRON/BULYGIN (1971). Carlos E. Alchourron and Eugenio Bulygin: Normative Systems. Wien 1971. ANSCOMBE (1989). Elizabeth Anscombe: Von Wright About "Practical Inference". In Schilpp/Hahn (1989). BECKER (1952). Oskar Becker: Untersuchungen über den Modalkalkül. Meisenheim 1952. BERNARDO (1988). Giuliano di Bernardo (ed.): Normative Structures of the Social World. Amsterdam 1988. BULYGIN ET AL. (1985). Eugenio Bulygin, Jean-Louis Gardies and Ilkka Niiniluoto (eds.): Man, Law, and Modern Forms of Life. Dordrecht 1985. GARZON VALUES ET AL. (1997). Ernesto Garzon Valdes, Werner Krawietz, Georg Henrik von Wright and Ruth Zimmerling (eds.): Normative Systems in Legal and Moral Theory, Festschrift for Carlos Alchourron and Eugenio Bulygin. Berlin 1997. GASKING (1954). D. Gasking: Causation and Recipes. Mind 63 (1954). HALLDEN (1957). Sören Hallden: On the Logic of "Better". Lund 1957. HEDENIUS (1941). Ingemar Hedenius: Om rätt och moral. Stockholm 1941. KAHNOWSKI (1953). Jerzy Kalinowski: Theorie des propositions normatives. Studia Logica 1 (1953). MALCOLM (1989). Norman Malcolm: Intention and Behavior. In Schilpp/Hahn (1989). McMURRlN (1985). Sterling M. McMurrin (ed.): Tanner Lectures on Human Values, Vol. VI. Salt Lake City 1985. POPPER/ECCLES (1977). Karl Raimund Popper and John C. Eccles: The Self and Its Brain. London 1977. PRIOR (1967). Arthur Norman Prior: Past, Present, and Future. Oxford 1967. 44

On this, see my book In the Shadow of Descartes (1998b).

32

Georg Henrik von Wright

QuiNE (1945). Willard Van Orman Quine: On the Logic of Quantification. The Journal of Symbolic Logic 10 (1945). SCHILPP/HAHN (1989). Paul Arthur Schilpp and Lewis Edwin Hahn (eds.): The Philosophy of Georg Henrik von Wright. La Salle, 111. 1989. SEGERBERG (1967). Krister Segerberg: On the Logic of "To-morrow". Theoria 33 (1967). STOUTLAND (1970). Frederick Stoutland: The Logical Connection Argument. American Philosophical Quarterly 7 (1970). WESTERMARCK (1924). Edward Westermarck: The Origin and Development of the Moral Ideas, Vol. I, Second edition. London 1924. VON WRIGHT (1947). Georg Henrik von Wright: On the Idea of Logical Truth I. Helsingfors: Societas Scientiarum Fennica, Commentationes Physico-Mathematicae, Vol. IV, no. 4, 1947. VON WRIGHT (1952a). Georg Henrik von Wright: On the Logic of Some Axiological and Epistemological Concepts. Ajatus 17 (1952). VON WRIGHT (1952b). Georg Henrik von Wright: On Double Quantification. Helsingfors: Societas Scientiarum Fennica, Commentationes Physico-Mathematicae, Vol. XVI, no. 3, 1952. VON WRIGHT (1954a). Georg Henrik von Wright: Om moraliska föreställningars sanning. In von Wright (1954b). VON WRIGHT (1954b). Georg Henrik von Wright: Vetenskapens funktion i samhället. Copenhagen 1954. VON WRIGHT (1957). Georg Henrik von Wright: Logical Studies. London 1957. VON WRIGHT (1963a). Georg Henrik von Wright: The Varieties of Goodness, London 1963. VON WRIGHT (1963b). Georg Henrik von Wright: The Logic of Preference. Edinburgh 1963. VON WRIGHT (1963c). Georg Henrik von Wright: Norm and Action. London 1963. VON WRIGHT (1968). Georg Henrik von Wright: An Essay in Deontic Logic and the General Theory of Action. Amsterdam 1968. VON WRIGHT (1969). Georg Henrik von Wright: Time, Change, and Contradiction. London 1969. VON WRIGHT (1971). Georg Henrik von Wright: Explanation and Understanding. London 1971. VON WRIGHT (1972). Georg Henrik von Wright: The Logic of Preference Reconsidered. Theory and Decision 3 (1972). VON WRIGHT (1974). Georg Henrik von Wright: Causality and Determinism. New York 1974. VON WRIGHT (1980). Georg Henrik von Wright: Freedom and Determination Amsterdam 1980. VON WRIGHT (1983a). Georg Henrik von Wright: A Modal Logic of Place. In von Wright (1983b). VON WRIGHT (1983b). Georg Henrik von Wright: Philosophical Papers, vol. II, Philosophical Logic. Oxford 1983.

Value, Norm, and Action in My Philosophical Writings

33

VON WRIGHT (1985a). Georg Henrik von Wright: Is and Ought. In Bulygin et al. (1985). VON WRIGHT (1985b). Georg Henrik von Wright: Of Human Freedom. In McMurrin (1985). VON WRIGHT (1986). Georg Henrik von Wright: Rationality: Means and Ends. Epistemologia 9 (1986). VON WRIGHT (1987). Georg Henrik von Wright: Truth-Logics. Logique et Analyse 30 (1987). VON WRIGHT (1988). Georg Henrik von Wright: Action Logic as a Basis of Deontic Logic. In Bernardo (1988). VON WRIGHT (1991). Georg Henrik von Wright: Is there a Logic of Norms?. Ratio Juris 4 (1991). VON WRIGHT (1994). Georg Henrik von Wright: Normen, Werte und Handlungen. Frankfurt 1994. VON WRIGHT (1997). Georg Henrik von Wright: Ought to Be - Ought to Do. In Garzon Valdes et al. (1997). VON WRIGHT (1998a). Georg Henrik von Wright: Ought to Be - Ought to Do. In this volume. VON WRIGHT (1998b). Georg Henrik von Wright: In the Shadow of Descartes. Dordrecht 1998. VON WRIGHT/MEGGLE (1989). Georg Henrik von Wright und Georg Meggle: Das Verstehen von Handlungen (Münsteraner Disputation). Rechtstheorie 20, 1989; reprinted in von Wright (1994).

Part I: Actions

AULIS AARNIO Law and Action Reflections on Collective Legal Actions

Abstract: This paper deals with collegial judicial decisions as a form of human action. The scope is limited to three questions: (i) What are the structure and status of the general theory of action? (ii) Is this theory applicable to such performative acts as judicial decisions? (iii) Are the 'actions' of collective agents actions in the proper sense? The author defends the thesis that general theory of action is applicable not only to individual but also to collective actions; difficulties are not due to the structure of that theory, or to its "individualistic character", but to the notion of a "collective will". This kind of "will" is epistemologically always a result of political procedures, and to speak of the "collective will" presupposes the analysis of these procedures. 1 2 3

Conceptual Specifications Understanding an Individual Act Legal Actions

1

Conceptual Specifications 1.1

The Problem

My presentation focuses on collegial legal actions as a form of human action. The scope is, however, limited to three questions: (i)

What is the structure and the status of the general theory of action introduced by Georg Henrik von Wright;

(ii)

Is this theory applicable to such performative acts as judicial decisions or passing a statute; and finally,

(iii)

Is it possible at all to speak about action in connection with such highly institutionalized collective agents as collegial courts or legislator?

To elaborate these problems, some conceptual clarifications are necessary.

38

Aulis Aarnio

1.2

Three Perspectives

Legal activities like judicial decisions can be analysed from at least three angels, namely from a descriptive, justificatory and teleological point of view. The Descriptive Point of View. In the (heuristic) description, either the context of discovery or the procedure leading from the reasons to the action is presented. An example of a scientific discovery which is often mentioned is the well-known story of Newton's discovery of the law of gravity, whereas the "question and answer-method" used by Agatha Christie elucidates heuristically the detective's thinking procedure in finding the murderer. This kind of logic of discovery has in recent years been developed on a general level e.g. by Jaakko Hintikka. The "logic of discovery" in the latter case may be either structural (legal-logical) as in Hintikka's theory or psychological, i.e. it may describe either the "moves" from one argument to another and finally to the conclusion, or be interested in the internal psychological mechanisms causally guiding the decision-maker to utilise the reasons. Depending on perspective, the matter can be seen from the point of view of either individual or social psychology. A distinction can also be drawn between normative and descriptive approaches. Normally, there is simply a description at issue. A normative theory, nonetheless, aims to recommend what a person should do when, for example, solving a criminal case. On Justification. As far as justification is concerned, the starting point is not the action itself, but the conclusion of the problem-solving or the decision-making procedure. Thus, the aim of legal justification is to convince the legal community of the "rightness" of the decision. For this, the justification formulates a set of arguments (reasons R) legitimising C. Using Jerzy Wroblewski's terminology, there are two different aspects with regard to the justification, i.e., the internal (deductive) and external (discursive) ones. The internal justification proves the proposition to be valid by means of the rules of alethic logic. Let us call this logical justification sensu stricto. The core of this justificatory procedure is the Aristotelian theoretical syllogism. The reasons used in syllogistic justification are norms and facts, and the conclusion is reached by deduction only. The function of the internal justification is, however, quite restricted: The syllogism can only give an ex post legitimation to the decision. Therefore, the internal justification is nothing but one side of legal justification. The other is intertwined with setting the premises of the syllogism where the question crops up of why these and not those premises? This problem cannot be answered without going outside the internal justification, i.e. without weighing up the justificatory material. This is why

Law and Action. Reflections on Collective Legal Actions

39

Wroblewski spoke about external justification or logical justification sensu largo.

The rationality of the external justification is not only deductive but essentially discursive1. In modern argumentation theory, this discursive approach to law has usually been understood as weakly normative in the sense that the theory is a rational reconstruction of the justificatory procedure2. In this regard, the justification of legal decision is a special case of practical reasoning. It is practical basically due to the subject-matter: Legal reasoning concerns practical sentences, i.e. norms. However, it is also practical as far as the logical structure of the reasoning is concerned, which is why the procedure is often called new rhetoric or topical argumentation. It produces a complex set of reasons, like a chain of syllogisms, and the aim of this procedure is acceptability and the final criterion is coherence3. Explanation and Understanding. In the following, the focus of the presentation is on neither the heuristic description nor the justification of legal decisions but on the explanation of the legal behaviour. In this regard, a special problem is encountered. The reasons (R) in the justificatory procedure are not necessarily reasons for action at all, but norms, facts or values supporting ex post the normative consequence, and therefore a theoretical syllogism is, as will be seen later on, a useless tool when one tries to understand, for instance, judicial decisionmaking procedure or the passing of laws in the legislative procedure. As is well-known, according the so-called Humean guillotine it is logically impossible to infer from the deontic world something relating to the empirical one. If the major premise refers to a deontic state of affairs, the conclusion should also belong to the deontic world. This is fatal for the explanation of legal activities because the conclusion of the explanation is not a norm but an act. When explaining a legal activity like a judicial act or legislation, the focus lies on an act. Somebody has acted in a certain way, and our task is to understand this activity. For this purpose, one needs a theory formulated in other than justificatory terms. This theory can be called, using Georg Henrik von Wright's terminology, the theory of action.4 As in the field of justification, the notion of practical reasoning is encountered here too, but in a different sense. The key concept is purposive (volitive) behaviour. The person acting in society (the agent) has a certain motivational basis for his behaviour. The motives make the behaviour understandable, giving a meaning to the activity. Seen from this perspective, understanding (and teleological explanation) Wroblewski (1974), pp. 42 f.; Alexy (1983), pp. 273 f.; and Aarnio (1987), pp. 119 f. Aarnio (1997), pp. 174 f. 3 See especially Alexy/Peczenik (1990), pp. 272 f. 4 Von Wright (1968), pp. 37 f. 2

40

Aulis Aarnio

of the behaviour means roughly the same as Wittgenstein's "looking close up". By viewing already familiar matters in a new context, one learns to understand the connections between different facts, events and procedures. Gradually, the behaviour turns out to be a meaningful part of a larger whole. Keeping this in mind, one encounters, for instance, the following types of questions: (i)

What is the theoretical structure of the explanative approach as far as human behaviour is concerned?

(ii)

Is it possible to interpret judicial and/or legislative activities as intentional behaviour?

(iii) Is the general theory of action applicable to individual behaviour only, or does it also suit collective agents? The first question deals with the basic notions of action theory itself and their use in the explanation of human behaviour. The second refers to the limits of intentional behaviour in judicial life, and the third presupposes an answer to the classical problem of ontological, epistemological and methodological individualism.

2 2.1

Understanding an Individual Act

Intentional versus Institutional Behaviour

According to Peter Winch, understanding social behaviour necessarily presupposes established rules or rule-like standards. This is the very element that creates interactive ties between different events and modes of behaviour5. If a person's behaviour cannot be said to follow any socially established practice, there is no basis for understanding this behaviour. However, at least the following of a rule seems to exclude any purposiveness to human behaviour. To follow a rule is like going along the rails, as Wittgenstein put it. Is it therefore impossible to govern "normatively guided" behaviour by means of the concept of action? Are the following of a rule and purposive activity opposite phenomena? In one sense, yes. If one has a disposition to follow given rules, one may behave intentionally without making any purposive choices between alternative options. This is even more clear in a case when the agent follows the rules "blindly", i.e. with no volitive concern as regards the norms. This is linked with the distinction between institutionalised and intentional behaviour. According to von Wright, behaviour is institutionalised if, 5

Winch (1958), pp. 45 f.; and Sandbacka (1987), pp. 47 f.

Law and Action. Reflections on Collective Legal Actions

41

and only if, an individual has the disposition to follow a rule "mechanically". The agent has internalised the rules of the game and even in applying them he acts without questioning their interpretation. Not only social habits but also many other forms of social behaviour, for instance obeying traffic rules, are internalized. In the field of law, I have called these routine cases. With regard to them, the following of the rule and purposiveness are mutually exclusive. When a person has internalised a rule which forbids him to cross the street when the traffic lights are red, he does so "automatically" and not due to his purpose to follow those rules. Internalised behaviour is not accompanied by purposive deliberation. Here, one can speak, so I believe, about a strictly norm-oriented behaviour, or from a slightly different perspective and in keeping with the Weberian terminology, about norm-rationality (sensu stricto). However, the reflective "following the rule" can normally be considered intentional. This is even more clear in the case of "following a principle". Principles are only standards open to weighing and balancing. The reflective following a norm (either rule or principle) is goal-rational, or in the term suggested by Ronald Dworkin, policy-oriented6. The agent is forced to weigh up the consequences of his action and to make a choice between certain alternatives available in each case. In doing so, the agent intentionally directs his action according to the goals using the framework given by the norms as his means. This is also the case in the interpretation of ambiguous legal statutes, in filling normative gaps7 as well as in legislation. In the following, this type of intentional (legal) behaviour will be the topic of this contribution, and then analysis will only focus on the legislative procedures, i.e. on the behaviour of Parliament. 2.2

On the Concept of Action

Action versus Activity. According to Georg Henrik von Wright, an action can be individual or generic. Here only the first type of action will be dealt with because the idea of my paper is to defend the thesis that the general theory of action is applicable not only to individual actions but to collective ones, too. An individual action is at issue, for instance, when a person moves his hand to greet, as well as in a series of actions such as opening the window, or running to the railway station. Playing football, in its turn, is an example of an activity consisting of several individual actions. 6

Dworkin (1957) passim. See about gaps Alchourron/Bulygin (1971), pp. 31 f.

7

42

Aulis Aarnio

The Logic of Actions: A Short Introduction. The term "action" means here, as von Wright puts it, effecting a change in the world or preventing it. The description of an action can thus be given describing all the changes in the world consisting of generic states of affairs. Their existence is not relative to a certain time or place (for example, the sun shines). At a given moment, every generic state of affairs either prevails or does not. A description of a state of the world (mi) at the moment (ti) thus depends on whether a state of affairs (a) 'constituting' the world is present or absent. Reality does not actually consist of separate parts. There exist numerous dependencies between the 'parts' of the world. The atomistic world-picture leads, in its extreme, to a distribution of the world. The simplified model is, however, of great theoretical value, as von Wright emphasises. Therefore, the atomistic picture serves well the purposes of the theory of action. An event is to be understood as a change in the state of the world: πΐί —>· rrij. The different states of the world can be temporally arranged into a successive series, i.e. into the history of (this) world. According to von Wright, the transition can be logically described with connective T. Thus, a(Ta) means no change. Consequently, a(T non-a) is "o disappears" non-a Ta means "a appears" and non-a Τ non-a the same as "a fails to appear". As von Wright suggests, the logic of change can actually be seen as a formal theory of connective Τ which co-ordinates states of the world over a time span, provided that inquiry is confined to finite changes in the system and that it happens within a discrete period of time. A mere theory of change as such is an inadequate tool in describing individual actions in the world. Action is an active interference with the course of world history. A certain world history results from an action if and only if the world would have been different without the agent's activity. We must know not only how the world actually changes, but also how it would have been changed from one occasion to the next if there had been no interference by an agent with the course of history. In this regard, an action can be characterised as a change (from the initial state Ft to end state Fe) taking place with regard to the counterfactual state Using the connectives "T" and "/" the notion of action can now be logically expressed in the formula:

"/" is here "instead of and the formula can thus be read: % and then Oj due to the agent, instead of having ac, owing to nature." Yet, the impact of an action on world history can be demonstrated in the form of a specific path of activity, i.e. in the form of a topological lifetree starting from a certain initial state (Fj). The life-tree indicates the

Law and Action. Reflections on Collective Legal Actions

43

possibilities at the agent's disposal in the world8, actually limited on many different grounds9. The Why-Question. The description of the life-tree is not enough for the notion of action. We also have to answer the question: Why did the person act as he did? And this question presupposes information about the agent's goals, generally about his intentions. Understood in this way, the concepts of "goal" and "consequence" are referentially identical only in a special sense. The aim to achieve a certain state of affairs makes the consequence a goal, but not vice versa; not all consequences are the goals of an agent. Nonetheless, setting a goal does not necessarily lead to its successful realisation. An act may also fail. The achievement of a goal usually only fully succeeds in routine actions. By taking the goal into account, the theory of action is supplemented by a volitive addition, the aspect of human will. Unless the agent interferes in the course of events, the state of world 5Ί will come about. The agent himself wants to bring about state £2, which differs from state 5i. To bring about 82 the agent has to take measures which, in his understanding, will ensure success. If the world still remains in the state Si, the agent has failed. 2.3

On the Teleological Understanding

Intentional and Epistemic Presuppositions. besides the goals, two kinds of beliefs:

The action thus presupposes,

(i)

the assumption concerning the counter-factual state Fc, and

(ii)

evaluations of the alternative end-states (E).

Every current strategy is deficient. The agent does not know for certain the initial state, nor is he sure of the consequences. Therefore, the theory of action necessarily has stochastic features. The action leads from one stochastic state to another. In analysing the notion "doing", the aspect of "interfering in the world" is important. This human aspect presupposes the agent being aware not only of his goals but also of the state of world. Let us call these two intentional and epistemic presuppositions. The latter covers the initial state at the moment of the action as well as a forecast of how the world would look without any interference. In this regard, the analysis of "doing" may concern either a prospective or a retrospective point of view10. Prospectively, the focus lies on planning: "Von Wright (1971), pp. 49 f. 9 Aarnio (1976), pp. 152 f. 10 Aarnio (1976), pp. 154 f.

44

Aulis Aarnio

The agent has fixed certain goals and he then plans the "moves inside the life-tree" choosing proper means to reach the goal or at least to maximise the chances of reaching it. The prospective aspect is thus not only futureoriented but also internal, the agent's own point of view. The problems of disagreement vs. agreement on goals, ambiguity of beliefs, and other difficulties in forming a homogeneous basis for planning are especially connected to the prospective point of view. The retrospective aspect, in turn, looks at the past. The subject matter of the analysis is an existing behaviour, something which has already happened. The task is to understand this behaviour as an action, to give it an intentional (teleological) explanation. In the following, only the retrospective aspect will be dealt with. The retrospective aspect also clearly points out that the analysis is neither descriptive nor prescriptive. Using the tools defined by the general theory of action, one not only describes what has happened nor gives norms for how one should act but reconstructs the conceptual structure. The idea is to make explicit the implicit preconditions of using, for instance, the notion of action in the common language. In the same sense, the analysis can reveal the implicit ways of using such terms as "Parliamentary will" "the intentions of the judge" or the "purpose and intent of the law giver". On the Practical Syllogism. Schematically, this kind of an understanding/ explanation is a form of practical reasoning, and here "practical reasoning" does not, as in the case of the justification of judicial decisions, concern norms or values but an action: A did T because he sought to bring about E and thought that T is necessary as regards reaching E. It is practically necessary for A to perform T under the condition that the premises at issue (goal and epistemic assumption) are valid. Practical reasoning can thus be formulated in the form of a so-called practical inference (or syllogism; cf. the theoretical one): PI:

A aims at E

P2:

A considers T to be necessary for E

C:

A undertakes to do T

Following von Wright, the practical inference can be read: "The starting point, that is, its major premiss, refers to a matter that is desired, aimed at; the minor premiss associates with this matter an individual act, as a means leading approximately to the end; the conclusion consists of adopting the means to achieve the end"11. 11

Von Wright (1971), pp. 96 f.

Law and Action. Reflections on Collective Legal Actions

45

In other words, the minor premise is based on a technical norm12. It expresses what has to be done if a certain state of affairs prevails. Practical syllogism and the notion of the technical norm are thus closely connected to each other. From the major and minor premises follow not only a normative consequence (A ought to do T), but a fact: A undertakes T. This is the most radical difference between the theoretical and practical syllogisms. A practical syllogism, formulated in this way, does not violate the distinction between Is and Ought (Sein and Sollen). The premises PI and PI both refer to facts, not to norms or evaluative standpoints. Hence, the factual consequence results from facts. On the other hand, the practical syllogism gives an answer to an old philosophical problem: How is knowledge transformed into human action. Bindingness of the Practical Inference. However, there are problems as far the bindingness of the inference is concerned. In other words, does the practical inference actually proceed from intentions to action, or is it logically conclusive, as von Wright asks13. According to von Wright, "A intends to do T" can be translated into an expression "A does T intentionally". This is certainly so taking into consideration the time dimension: A now intends to do T. As von Wright says, the expression is here rendered in the language of doing. The conclusion can easily be put in the form "A does T". A cannot intend "to do now" and, at the same time, not to do T. A has an intention to do something at the present moment. This formulation does not, of course, take into account interfering factors, namely circumstances preventing A from carrying out his intention. A may, for instance, (i) die; (ii) modify his intention; or (iii) become aware of his inability to achieve the goal. Furthermore, it is not necessary to take into account here either nonexpected or hidden consequences (dysfunctions). The agent's behaviour can be understood only on the basis of the material which was available for him at the moment of action. Hence A, at any rate, undertakes to perform T. Although "undertaking" would seem to save the conceptual framework, A may still be prevented while he undertakes to do T. One more reservation is thus essential. Still, a certain criticism has to be remembered here: the syllogism never actually leads to an action, but at most, to a decision to bring about the intended state of affairs. This point has been emphasised by Risto Hilpinen14. 12

See Aarnio (1997), pp. 188 f. Von Wright (1971), pp. 96 f. 14 Hilpinen (1972), pp. 200 f. 13

46

Aulis Aarnio

Hilpinen maintains that the inference is valid, moving only on the level of intentions, because doing does not necessarily result from the premises. Instead, the alternative, "to do now" leads to an action. Yet what about "a decision to do"? It can be seen equivalent to undertaking an act, because it does not make much sense to claim "A now (really) decides to do T" and, at the same time, to say "A will not do T". The case is analogous with the analysis of "intending"15.

3 3.1

Legal Actions

A Performative Point of View

As was mentioned above, the focus of this presentation lies on the behaviour of Parliament. Are there theoretical reasons to define this kind of behaviour as an action in the sense of the general theory of action? The legislative procedure as such can certainly not be understood as a proper action in terms of von Wright's theory. It is more an activity consisting of numerous individual actions. However, at least in some cases the "final step" of the procedure, passing the statute, seems to be intentional, but not exactly similar to, for instance, running to the railway station or opening a window or raising one's hand in a classroom. Legislation is a performative act in the true Austinian sense, and in this it is similar to, for instance, sentencing a person to prison. Both of these, a statute and the sentence, are performed by the decision-maker himself. Giving a penalty or passing a statute can be compared to a donation: A gift of a watch is performed by the act of donation. In a similar way, the statute is constituted by the decision. Judicial decision-making is merely norm-oriented only in such routine cases as, for instance, fining a person for illegal parking. In so-called difficult cases, the judge always has only a normative framework available. The core of the decision-making process then focuses on uncertainty either about the content of a norm or about the evidence. In these, not uncommon cases, the judge's behaviour is as intentional as that of the person running to the railway station. Especially in cases of so-called consequential reasoning, the judge orientates himself according to certain systemic or social goals16. The same holds true as far as the legislation is concerned. The performative character of passing a statute is thus not an obstacle to applying the general theory of action to this type of human activity. It 15

See Aarnio (1976), pp. 157 f., where Hilpinen's analysis is examined more closely. Aarnio (1997), pp. 214 f.; and Aarnio (1987), pp. 131 f.

16

Law and Action. Reflections on Collective Legal Actions

47

differs, of course, from simple forms of behaviour like raising one's hand in a classroom, but it is still intentional activity. Something is, literally, done by Parliament. 3.2

Collective Actions

In General. It is possible to speak about collective action if and only if every individual belonging to the collective agent shares at least to a certain degree the same intention with the others. As Raimo Tuomela puts it, there must be a shared "we-intention" prevailing in the group, i.e. every individual is conscious about the intentions of the others and also knows that every single member of the group presupposes the others to be committed to that very intention.17 A mere (ad hoc) number of people at the market or on the street does not constitute a collective agent. There are, however, different types of collective agents (and actions) depending on the degree of institutionalisation of the agents. At the one end of the scale are so-called group actions. Group Actions. In the cases of individual actions, there is only one agent: A is running to the railway station and the outsider is curious to know why he is doing so. However, it is also possible for several persons together to perform the same action. This can be called a group action if and only if the above presupposition of a shared intention prevails. The description of a group action is of a similar type to the description of individual actions: The group X changed the state of the world in the way Y. The understanding of a group action is, however, more complicated. Let us take an example: A, B and C are pushing A's car on the street. All of the persons may have the same intention to act and the same beliefs on the facts of the case. All of them - A, B and C - want to get the engine to start in order to visit their friend, and they believe that pushing the car is, on that occasion, the necessary condition to start the engine. However, the co-agents' intentions may, in details, also differ from or be even partially contradictory to each others. A as the owner of the car wants to visit his friend, B intends to help A, and C is only there for fun, although all of them believe that pushing the car is necessary to start the engine. This means that there is not necessarily a perfectly homogeneous intentional basis to understand the group behaviour. In some cases this may have consequences as far as the division of labour inside the group is concerned. Group actions are often also analysed referring to the agent's retrospective responsibility with respect to actions. Then, a group action can respectively be thought of in terms of a composite or conjunction of the actions of two 17

Tuomela (1977), pp. 36 f.

48

Aulis Aarnio

or more different agents. As Gregory Mellema points out, it is perfectly conceivable, however, that the group action includes the action of an individual agent who fails to share responsibility for the outcome of the group action18. A person whose actions fail to constitute a causal contribution to the outcome cannot be said to be responsible for it, either. In the present contribution, the group actions and the problem of responsibility due to this type of action are left out of account. The Notion of Collegial Action. Depending on the degree of institutionalisation, the collective agents may also be more structured than mere groups. This is the case for example as far as the General Assembly of the IVR is concerned. There is a valid Constitution of the IVR which defines the structure of the Assembly as well as individual actions and activities within it. In this case it is possible to distinguish between different levels of agents on the basis of their roles inside the collective agent. (i)

Executive agents have shared intentions concerning their activities, and they have certain shared mutual beliefs on, for instance, the goals and tasks of the organisations. According to the Constitution of the collective agent, the participation of the executive agents is also jointly necessary for the functions of that organization. Without such joint actions there will not be any General Assembly of the IVR.

(ii)

Some agents are only contributory but in this role jointly necessary for the collective agent. Their actions are controlled by the executive agents who also formally direct their intentions and beliefs concerning the actions.

(iii) In addition to these two groups of agents there may be different other types of agents belonging to the collective agents but not being necessary for it in the same way as the executive and contributory agents are. They may be passiv3 members of the association, paying only the fee but not taking part in the activities, or absolute free-riders such as some individuals using the social resources without taking any societal responsibility. Some collective agents are even more structured than the General Assembly of the IVR. Let us mention as examples the (Finnish) Parliament and the (Finnish) Supreme Court. The Parliament is Parliament due to the Finnish Constitution in the same way that the General Assembly is the highest organ in the IVR on the basis of the IVR Constitution. Yet the Supreme Court is not only thus based on the Constitution but is also collegial due to the strictly defined procedures to make proposals and decisions. The behaviour 18

Mellema (1988), pp. 12 f.; especially p. 19.

Law and Action. Reflections on Collective Legal Actions

49

of every single judge in the Supreme Court is norm-based so that his actions are necessary to a valid decision. In both cases, there are executive and contributory agents but no perfectly neutral members, much less free-riders. In the following, only these kinds of highly institutionalised actions, especially those made by Parliament, are the subject-matter of this contribution. According to the legal language, for instance, Parliament "behaves" in certain way, it makes decisions and gives valid laws, it has "goals", "purposes" and "intentions" and it is said to "believe" certain matters to be necessary or sufficient with respect to the decisions etc. What does all this mean from the point of view of the theory of action? Is it at all possible to apply the basic notions of von Wright's general theory of action to this kind of collective agent, and if it is possible, in what sense is the collective agent "doing" something and doing it "intentionally"? An Ontological Point of View. As far as the ontological status of a collective agent, say the Finnish Parliament, is concerned, one can qualify it referring to Popper's World 3. Collective agents are man-made entities relatively independent of their members, i.e. of the individuals. I dare to think that Ilkka Niiniluoto for example would be inclined to define the ontological status of collective agents in this way.19 For my purposes this is not a satisfactory answer, or is, at least, only a tentative one. According to my ontological assumptions a collective agent like the Finnish Parliament is not an independent entity. It consists of individuals functioning in certain social roles. In this regard, the Finnish Parliament is a collective constituted by individual members of Parliament (200 persons) all functioning in a certain social role denned by Finnish statutory law (especially by the Constitution) and by the political tradition of the country. The social role is thus an intermediate link between an individual A and the collective called Parliament, because only such an A who has the role of a Member of Parliament can be an element of the institutional entity "Parliament". However, this is not my point as far as the ontology of collective agents is concerned. One can ask at least one more question: What combines the social roles so that one can speak about "giving law", "having a goal" "intending something" etc., or what makes, for instance, two hundred people sitting in the City Hall of La Plata participants of the XVIII World Congress of the IVR? Here, I suppose, Eerik Lagerspetz gives decisive help to me20. Such an entity as "The Finnish Parliament" can be compared to phenomena like "money", "faculty of law", "marriage" etc. Ota Weinberger and Neil MacCormick would speak about institutional facts. They are neither facts in the sense of the Popperian World 3, nor facts of a simple ontological type. As 19

Niiniluoto (1980), pp. 125 f. Lagerspetz (1989), pp. 15 f.

20

50

Aulis Aarnio

Lagerspetz puts it, institutional facts do exist due to certain shared (mutual) beliefs, and because those who believe in the existence of the institutional facts also act according to their beliefs. Lagerspetz writes: "Neither it is enough that members of a community just happen to believe that some objects are money: they must also know that this belief is generally shared by other members [...]. And lastly, these attitudes must be related to actions. They must appear as (at least partial) reasons for the members of the respective societies to do certain things" .21 The notion of mutual belief can, in this context, be defined as follows: A belief p is mutual in a population S if and only if: (i)

everyone in S believes that p;

(ii)

everyone in S believes that everyone else in S believes that p; and so, when i is the number of reiterations needed to describe the beliefs of the members in 5;

(iii) everyone in S believes that no one in S has any such beliefs of higher order (> i) about the beliefs of the members of S which would have an effect on the behaviour of any member. The mutual belief is thus reflexive, or if one likes, circular. The beliefs referred to in the analysans refer, in their turn, back to these beliefs. The reflexivity is, however, not a weakness of the definition. As Lagerspetz points out, every individual sees the behaviour of all others as a part of his own environment with which he or she has to scope. Referring to money as a conventional fact Lagerspetz says: "From the individual's point of view, the fact is just there: others use certain objects as money". He also continues that by giving a conventional analysis of rules, it is possible to build up a theory of such basic institutions as law, money and language. What about a situation when there prevails in the (Finnish) Parliament not only an agreement about certain decisions but a significant disagreement among its members? Can one speak in this very situation about shared mutual beliefs at all? To my mind, yes. For the constitution of a collective agent the shared mutual beliefs are the necessary condition. The members of the collective agent have to share beliefs concerning what that collective agent, say the Finnish Parliament, or the General Assembly of the IVR, is. Without this kind of "agreement" or "convention" there a collective agent does not exist because in the case of full disagreement the situation would be similar to everybody warring against everybody. There has to be at least "agreement" concerning the "hard core" of the agent and that "agreement" must cover at least the majority of the members. 21

Lagerspetz (1989), pp. 15 f.

Law and Action. Reflections on Collective Legal Actions

51

It is quite another problem whether the (Parliamentary) members agree as far as a certain decision (Dj) in a certain problem (Pi) is concerned. This disagreement does not concern the constitution of the agent but the content of the decision concerning for example the commercial law. On Epistemological and Methodological Individualism. What is said now about the ontology has an important consequence as far as the epistemological and methodological perspectives are concerned. The "will of Parliament" is only in a very vulgar sense the brute sum of the "wills" of individual Members (200) of Parliament. This could be the case if and only if all the Members of Parliament (i)

actually have that special "will" (intent to achieve E);

(ii)

express this "will" in a non-contradictory way, and

(iii) there are no interfering factors preventing an individual member from realising his "will" In the majority of Parliamentary decisions this is not the case. Therefore this type of vulgar epistemological or methodological individualism does not hold true. On the contrary, there exists a scale of situations, and these situations are not at all like the above simple example. In an extreme case, a Member of Parliament may even sleep throughout the whole procedure. In this case, an individual intention or will ofthat member does not actually exist. Thus, the principle of moral symmetry, discussed by Georg Henrik von Wright in his "The Varieties of Goodness", cannot be applied here either. Von Wright describes this man's sense of symmetry in the following words: "If my wants are satisfied at the expense of another man's, then why not his wants at my expense"22. Parliamentary decision-making is not comparable to moral decisions. Normally, members of Parliament are divided into (political) groups having asymmetric, even contradictory intentions and, concerning the subjectmatter, different epistemic beliefs. In such a system an individual "will" is not significant because in Parliamentary voting individuals normally follow the orders and rules of their group instead of their own intentions. Using von Wright's terminology, the action is rather parasitic than symmetric. Von Wright defines a parasitic action as follows: "If X adds to the greater advantage of getting his share also the smaller advantage of skipping his due, he will necessarily deprive some of his neighbours of their share in the greater good"23. 22 23

Von Wright (1963), p. 210. Von Wright (1963), pp. 203 f.

52

Aulis Aarnio

This parasitic nature of political relations is typical both of the formation of the "will of the group" and that of the Parliamentary "will". Neither of these "group wills" is normally the same as an individual will or a number of such wills. Taking all this into account, the above conventionalist view touches the epistemological and methodological core of our problem. It is difficult, even impossible, to identify Parliamentary "will", "goal" or "purpose" by reducing it simply to individual volitive acts. Parliamentary "will" is epistemologically always only a result of a political procedure, and speaking about the "will" presupposes the analysis of that procedure from a systemic point of view as for example Werner Krawietz in many contexts has done24. Norm-Oriented and Intention-Oriented Understanding. Methodologically it is thus not so important to try to grasp the individual "wills" of Members of Parliament but to analyse the procedure formulating the "collegial will of the Parliament". As far as the explanation (understanding) of this procedure is concerned, it may, so to speak, be either norm-oriented or intention-oriented. A norm-oriented Parliamentary procedure can be understood by referring merely to the norms regulating it. For instance, in order to understand why Parliament followed a certain order in passing the statute, one may only refer to the wording of the Constitution, according to which no other procedure was possible, and this very fact provides a sufficiently exact basis for our understanding. However, there are still numerous cases (perhaps even the majority) where it is perfectly inadequate only to refer to certain norms. In these cases Parliament did not only intentionally follow certain norms but intentionally took the material decision A and not B. The alternative A was chosen deliberately. Behaviour is in these situations intention-oriented (purposive) in the sense of deliberation, and there seem to be no problems at all to connect such notions as "goal" and "belief to a collective agent, too, and say, for instance: "The Finnish Parliament behaved in such-and-such a way, because it wanted to achieve the goal E and believed the statute S to be a necessary means to realise this goal." The "behaviour" of a collective agent, like the Finnish Parliament, also consists in this case of a series of activities, which in their turn are constituted by a number of individual actions. Despite this complexity, one has good reasons to speak about "the action of Parliament" if the notion of "action'' is understood as a general description of the "total" behaviour of the agent, for instance, of "passing a certain statute". 24

See Aarnio (1987), pp. 107 f.

Law and Action. Reflections on Collective Legal Actions

53

Conclusion. In cases of deliberation, the lawgiver's behaviour can thus be understood on the motivational and epistemic basis, too, i.e. on the basis of shared mutual beliefs, and this, again, makes it possible to interpret statutes referring to the "intention" or "purpose" of the lawgiver. Hence, the general theory of action is applicable to collective action, although with certain reservations. The difficulty of the application is not in the structure of the theory as such, or in its so-called "individual character" . As far as, for example, the theory of legislation is concerned, the future theoretical and empirical challenge will be in the analysis of notions like "agent", "motive" and "belief. This type of approach presupposes intensive interdisciplinary research in the fields of sociology, political science and in legal theory. Let us take up this challenge.

References AARNIO (1976). Aulis Aarnio: On Legal Reasoning, Turku 1976. AARNIO (1983). Aulis Aarnio: Man and the Changing Society. Some Thoughts on Leo Tolstoy's Conception of History. Essays in Legal Theory in Honor on Kaarle Makkonen. Oikeustiede-Jurisprudentia 16 (1983). AARNIO (1987). Aulis Aarnio: The Rational and Reasonable. Dordrecht 1987. AARNIO (1997). Aulis Aarnio: Reason and Authority. Aldershot 1997. ALCHOURRON/BULYGIN (1971). Carlos Alchourron and Eugenio Bulygin: Normative Systems. Wien 1971. ALEXY(1983). Robert Alexy: Theorie der juristischen Argumentation. Frankfurt am Main 1983. ALEXY/PECZENIK (1990). Robert Alexy and Aleksander Peczenik: The Concept of Coherence and its Significance for Discoursive Rationality. Ration Juris 3 (1990). DwORKIN (1957). Ronald Dworkin: Taking Rights Seriously. New York 1957. HILPINEN (1972). Risto Hilpinen: "Praktinen päättely ja oikeuskysymyksen ratkaiseminen (Practical Inference and Solving the Legal Question)". In: Lakimies (Lawyer) (1972). KRAWIETZ (1984). Werner Krawietz: Recht und Rationalität in der modernen Systemtheorie. In: Krawietz et al. (1984). KRAWIETZ ET AL. (1984). Werner Krawietz, Theo Mayer-Maly, Ota Weinberger (eds.): Objektivierung des Rechtsdenkens. Gedächtnisschrift für Ilmar Tammelo, 1984. LAGERSPETZ (1989). Eerik Lagerspetz: The Opposite Mirrors. Dordrecht 1989. MELLEMA (1988). Gregory Mellema: Individuals, Groups and Shared Moral Responsibility. New York 1988. NHNILUOTO (1980). Ilkka Niiniluoto: Johdatus tieteenfilosofiaan. Keuruu 1980.

54

Aulis Aarnio

SANDBACKA (1987). Carola Sandbacka: Understanding Other Cultures. Helsinki 1987. TuOMELA (1977). Raimo Tuomela: Human Action and its Explanation. Dordrecht 1977. WlNCH (1958). Peter Winch: The Idea of a Social Science and its Relation to Philosophy. Norwich 1958. VON WRIGHT (1963). Georg Henrik von Wright: The Varieties of Goodness. London 1963. VON WRIGHT (1968). Georg Henrik von Wright: An Essay in Deontic Logic and the General Theory of Action. Amsterdam 1968. VON WRIGHT (1971). Georg Henrik von Wright: Explanation and Understanding. London 1971. VON WRIGHT (1974). Georg Henrik von Wright: Causality and Determinism. New York 1974. WROBLEWSKI (1974). Jerzy Wroblewski: Legal Syllogism and Rationality of Judicial Decision. Rechtstheorie 5 (1974).

ROSARIA EGIDI Grounds for Acting, Grounds for Knowing Abstract: The aim of this paper is to highlight some changes in von Wright's "dualistic" perspective after the publication of Explanation and Understanding. My focus will be especially on two theses characterizing the general theme "reasons for actions". I will make an attempt to show that, despite the background Wittgensteinian inspiration, both theses reflect a theoretical and systematic purpose, partially alien to the author of Philosophical Investigations. 1 2 3

Changes in the Dualistic Outlook of Action Explanation The Conceptual Priority of Agency on Causation The Puzzle of the Groundless Knowledge

1

Changes in the Dualistic Outlook of Action Explanation

The points I will discuss here concern a "classic" topic of von Wright's thought: the relation of intention to action or intentional behaviour. And yet, I shall limit myself to emphasizing some changes in perspective that occur in the philosophy of action formulated in Norm and Action,1 and reflected in the fundamental theses elaborated in Explanation and Understanding,2 and especially in later writings focused on the theme "reasons for actions".3 1.1 In fact, in many passages of his recent works, von Wright does maintain that his views have considerably changed since the publication of Explanation and Understanding. But what exactly is the change as to the central theses of this work? And, most important of all, what is the change as to the approach to the theory of action known as the "dualistic" outlook, according to which there is a conceptual distinction between the model of causal explanation and that of intentional explanation? Surely, as von Wright says in his intellectual autobiography, the view of the "sui generis character of action explanation in conformity with patterns of practical inference" and in contrast "with explanations which rely on lawlike connections",4 is the Won Wright (1963). Von Wright (1971). 3 See von Wright (1974); (1976); (1980); (1981); (1982). 4 Von Wright (1989), p. 39. 2

56

Rosaria Egidi

same as ever. Moreover, he maintains the view that intentionality "is not anything 'behind' or Outside' the behaviour. It is not a mental act or characteristic experience accompanying it" ,5 and thus it cannot be reduced to an expression of individual thoughts, accessible by introspection. Finally, the author does not deny his view that the explanation of intentional behaviour draws meanings out of the rules governing the social game, and rests on the network of reasons and motivations which has "its place in a story about the agent", as opposed to a physical explanation concerning mere behaviour and the determinant causes. In fact, what changes in von Wright's later perspective, and will help to correct the very dualistic starting point - or perhaps to better specify the scope of its validity - is the non univocal view of the concept of causation: the discovery of its intrinsic intermingling with the concept of action, and the dependence of the former on the latter. 1.2 There are at least two sources for von Wright's practical philosophy, which justify the significant role he ascribes to action in scientific explanation: the first source can be traced to his interest in normative discourse, which resulted in the creation of a new branch of logic - deontic logic; the other one is connected to Wittgenstein's influence, or more precisely to his view of the pragmatic nature of intentionality, and to the non causal theory of action put forward in Philosophical Investigations? I wish to focus mostly on the second source of inspiration in order to establish whether the changes in the original dualistic perspective also show traces of Wittgenstein's influence and, in such a case, whether they represent an independent development of his perspective. I will sketch two theses characteristic of von Wright's thought about practical philosophy topics: the first thesis claims the priority of agency on causation; the second one is concerned with the puzzle of the groundless knowledge and its impact on the problem of free action. I will make an attempt to show that, despite the background Wittgensteinian inspiration, both theses reflect a theoretical and systematic purpose, partially alien to the author of Philosophical Investigations.

2

The Conceptual Priority of Agency on Causation

In a series of lectures published in 1974, entitled Causality and Determinism, von Wright questions the universal validity which is traditionally ascribed to the paradigmatic concept of causation, viewed as intrinsic to any "scientific" explanation; instead, he introduces a very specialized use of causation and causal determinism, which "can claim validity only for limited portions of the world, and not for the world as totality". It is, as it were, a dynamic concept 5

Von Wright (1971), p. 115. See Wittgenstein (1953), §§ 466-490.

6

Grounds for Acting, Grounds for Knowing

57

of causation "rooted in the idea that there are agents who can interfere with the natural course of events", a concept which "is therefore secondary to the concept of human action" J This concept and its implications in the field of practical philosophy, theory of action and ethics have been developed in the recent works, mostly in Freedom and Determination, which, outlining the most comprehensive analysis of the intentional action, is a sort of pendant of the treatment of causation included in Causality and Determinism. It is my purpose here to make a few remarks just on this treatment. 2.1 Von Wright's change in perspective took place in Explanation and Understanding: its general aim might be summed up as the rehabilitation of the agency in the thematic area of the scientific explanation. The final outcome of this conception - which he called the "actionist" or "experimentalist" version of causation - is a reversal of the received views, i.e. of the primary and exclusive role ascribed to causality in natural sciences. Without presupposing the concept of action as well as the active role of human action on natural events - as von Wright affirms - we could not recognize the external world as a set of events connected by causal laws. Thus, the thesis of the cause-reason dichotomy, which inspired the view of the so-called new dualism, has been replaced by the conception according to which the idea of a causal nexus (causation) somehow incorporates the notion of human agency. In this sense, therefore, the two notions seem to be intrinsically connected. In the large variety of meanings which the concept of causation takes on in the context of natural facts and in that of historical and social facts, von Wright picks out a peculiar concept of causation, dependent on that of action and, more precisely, on the crucial scientific concept of "experiment": "The reason why I nevertheless want to give a basic priority to this 'actionist' or 'experimentalist' notion of cause, is that, in addition to holding an important place in the experimental natural sciences, it seems largely to figure as a prototype for the idea of cause in the discussions of philosophers about universal causation, determinism versus freedom, interaction of body and mind, etc."6

2.2 In the actionist use of the concept of causation, there is a further element, worked out in Causality and Determinism, specifying the logical peculiarities of causal determination in the domain of action: it is the element of counterfactual conditionality informing the dependence of causal relation on the concept of action, i.e. on factual conditions making action logically possible. In a synoptic passage in Causality and Determination, von Wright makes it clear about the specific use of "counterfactual" involved in the actionist view of causality: 7 8

Von Wright (1980), p. 2. Von Wright (1971), pp. 36 f.

58

Rosaria Egidi "In the concept of action is thus implicit a comparison or contrast between a state of affairs resulting from the action and another state which would otherwise, i. e. had it not been for the performance of the action, have obtained. But ofthat which 'would otherwise have obtained' we cannot possess 'strict knowledge', if it is thought that such knowledge is possible only of that which we witness (observe, verify) as having occurred. For that which 'would otherwise have obtained' never comes true (occurs). It is 'contrary to fact'."9

Von Wright warns against the difference between counterfactuals involved in actions and causal counterfactuals involved in nomic connections. While the latter kind of counterfactuality presupposes a causality related to laws, to lawlike connections, the first kind of counterfactuals presuppose agency. This is, therefore, a more "primitive" concept, basically connected to our knowledge of an ability to do certain things or of "a power to interfere with 'nature' to make the course of the world (a little bit) different from what it otherwise would be".10 As von Wright says: "If man throughout stood quite 'passive' against nature, i.e. if he did not possess the notion that he can do things, make a difference to the world, then there would be no way of distinguishing the accidental regularity from the causal one. Nor would there be any way of distinguishing the case when p has the 'power' of producing the sequence of q upon p. Man would simply not be familiar with the notion of counterfactuality, with the idea of how it would have been, if ... This is the ground for saying that the concept of causal connection rests on the concept of action."11

2.3 The idea of a form of causation dependent on agency definitely shows the overcoming of the dualistic approach involved in the rigid Wittgensteinian distinction between causes and reasons, while the recognition of a specific actionist version of causality foreshadows von Wright's more mature attempt to recover a role for determinism in the domain of human action, thus - as it will be stated in later works - reconciling it with the concept of free will, traditionally regarded as irreducible to determinism. The rejection of the univocal use of the concept of causality and the resulting formulation of the actionist theory are a clear instance of the Aufhebung and Aufbewahrung strategy, characteristic of von Wright's philosophical style. In fact, the choice of such a strategy shows a need, on the one hand, to preserve the Wittgensteinian lesson but, on the other, to reject it. The thesis of the basic priority of the concept of action is definitely an elaboration of the typical view of Philosophical Investigations about the pragmatic ground 9

Von Wright (1974), p. 41. Von Wright (1974), p. 45. u Von Wright (1974), pp. 52 f. 10

Grounds for Acting, Grounds for Knowing

59

of any language game, a view which we find summed up in an emblematic passage in Vermischte Bemerkungen: "Der Ursprung und die primitive Form des Sprachspiels ist eine Reaktion; erst auf dieser können die komplizierteren Formen wachsen. Die Sprache will ich sagen - ist eine Verfeinerung, 'im Anfang war die Tat'."12

Von Wright shares with Wittgenstein the idea of the priority of agency on any language game; however, his application of this idea in the domain of causal determinism is an autonomous development of his own thought. In the Wittgensteinian theses on intentionality, in fact, it is difficult to track down arguments relating to the dependence of causation on agency and the role played by the counterfactual element in the specification of the conceptual nature of such relation. If we wish to find a forerunner of the actionist interpretation of causality, perhaps we should go back to a tradition different from the Wittgensteinian one, namely the Aristotelian tradition, for which von Wright has always had a strong attraction. One of the "dynamic" meanings ascribed by Aristotle to the term "cause" in Metaphysics is in fact that of "efficient cause". 2.4 Von Wright's thesis on the dependence of the concept of causation on that of agency really brings about a crisis in all the series of postulates traditionally connected to the univocal concept of causal determinism, exposing the underlying misunderstandings and conceptual illusions. I would like to mention some of these misconceptions, but without discussing them. The misunderstanding which has prevented the logical peculiarities of the concept of causation in the domain of actions from being singled out is the misleading analogy between causal laws and reasons or determinants of actions, with the result that an explicative power has been ascribed only to the "reasons" which are also "causes". Another misunderstanding arising from not distinguishing between causality in the context of lawlike connections and causality in the context of reasons for actions, is what von Wright called the methodological misunderstanding of "internalization": the relationship between actions and their determinants like wants, desires, feelings is patterned on the issue of the relationship between neurophysiological processes and macroscopic reactions in a body. Lastly, the "deterministic illusion," which treats determinism in the study of man - i.e. the search for laws governing individual and social life - in the same way as determinism in natural sciences is, on von Wright's view, what impedes giving an appropriate solution to the issues concerning free action. I will not comment on these topics here, but I rather mean to dwell upon some analogies between causal determinism and knowledge problems. 12

Wittgenstein (1980), p. 31.

60

Rosaria Egidi

3

The Puzzle of the Groundless Knowledge

In "A Lecture on Freedom of the Will"13 Wittgenstein had put forward a point of view very similar to that which von Wright argues in many writings following Causality and Determinism: after exposing the misleading analogies blurring the differences between determinism in the context of natural laws and determinism in the realm of reasons for actions, the old philosophical problem of free will appear in a new light. Clarifying the different uses of causal determinism, the illegitimate transfer from one area to another is a way of resolving the puzzle which Wittgenstein states at the beginning of his lecture: "Could one say that the decision of a person was not free because it was determined by natural laws? - There seemed to be a point in saying that if it determined by natural laws, if the history of people can be determined if we know their anatomy etc., then a decision can't be said to be free."14

According to von Wright, the incompatibility between "determinism" and "freedom" disappears if one considers intentional behaviour under two aspects, which are not incompatible but peculiarly connected with each other: we could call them the mental and the bodily or physical aspects. In Freedom and Determination, in the analysis of the "determinants" of intentional actions, von Wright had already made it clear that all human actions also present a physical aspect: they are in fact embedded in a "causal history". In An Essay on Door-Knocking, he specifies that a sort of parallelism obtains between the agent's action grounded on reasons and the brain processes which determine the physical behaviour or the bodily movement, i.e. between intentionality and mere behaviour. If this relation between the mental aspect and the physical aspect were contingent, the parallelistic perspective he proposes would be associated to some form of reductionistic or monistic approach or to the identity theory. But, according to von Wright's parallelism or compatibility thesis, the connection between these two aspects is not contingent but conceptual.15 Hence, the notion of "parallelism" he refers to is deeply different from that of psycho-physical parallelism typical of scientific psychology; its use here aims at increasing the distance from any form of reductionism and identitism, and confirming that behaviour understood under the aspect of intentionality and mere behaviour do not exclude each other, since they belong to different but parallel conceptual levels. However, the resort to "parallelism" is rather instrumental after all. As von Wright 13

Wittgenstein (1989). This text originates from the notes by Y. Smythies at the lectures delivered by Wittgenstein in Cambridge in 1945-46 or 1946-47 or more probably earlier in 1939. 14 Wittgenstein (1993), p. 429. 15 Von Wright (1988), p. 286.

Grounds for Acting, Grounds for Knowing

61

reaffirms in An Essay on Door-Knocking, "the philosophical interest" of this term vanishes as soon as one gets rid of the "reductionist illusion". But - as we know - von Wright is pursuing another goal, associated with the parallelistic thesis, that of dispelling the "deterministic illusion". 3.1 We can view the essay Determinism and Knowledge of the Future as a lucid corollary of the attempt to distinguish the meanings of "determinism", thus avoiding to make the category errors pointed out in Freedom and Determination. The emancipation from the deterministic illusion goes through the acknowledgement that causal determination is not of the same logical type as the reasons determining actions: i.e. the grounds for acting are not of a kind, human actions can be described in a variety of aspects. According to von Wright, the deterministic illusion does not relate only to the domain of acting, but also to the domain of knowing: for it has its root in the idea that the knowledge of the future, the foreknowledge, entails the predetermination of the truth of the objects to be known. But, as it happens in the domain of intentionality, even in the field of knowledge we don't have just a single way of taking for granted our claim to know something or replying to the question concerning how something is known. As it is the case with the grounds for acting, when talking about the grounds for knowing we are not thinking of a single meaning of "ground", and sometimes we use "knowledge" improperly. The problems concerning the determinism of action are therefore similar to those concerning foreknowledge. On the basis of these background analogies, in Determinism and Knowledge of the Future, von Wright showed that the deterministic illusion "has its root in a misconception of the notion of knowledge".16 The arguments supporting such an analogy and their application to the issues concerning the freedom of action are among von Wright's significant contributions to the logical grammar of "determinism"; however, the indication of the misconception surrounding the notion of knowledge has its inspiration source in some remarks in Wittgenstein's On Certainty. 3.2 The indication of the different role played by the grounds for knowing when it is a matter of genuine knowledge, which is based on experience and empirical evidence, and when it is a matter of knowledge called "certainty", i.e. which we take on trust, that "we unquestioningly take for granted", is put forward in the first part of On Certainty (§§ 1-65). Here Wittgenstein singles out a "highly specialised" use of the verb "to know",17 which is not to be confused with the ordinary use we can notice in propositions which instead express genuine knowledge, based on experience or derived from it. According to Wittgenstein, the category error of epistemology from Descartes to Moore lies in not distinguishing these two uses or, more precisely, in taking the propositions we call "certainties" as if they were propo16

Von Wright (1982), pp. 52 f. Wittgenstein (1969), §§ 11 f.

17

62

Rosaria Egidi

sitions concerning contingent facts. Wittgenstein explicitly rebukes G. E. Moore for this error, denying that his truisms, identified with commonsensical propositions, really express "knowledge". In spite of what Descartes Moore claim, the propositions which are undoubtedly certain have no epistemological ground. "I know that I have two hands"; "I know that earth has been in existence for millennia"; "I know that an amputated arm cannot grow again"; "I know that every human being has parents" are propositions which we take for granted, without asking for their grounds. In this sense, they represent cases of groundless knowledge, and therefore they do not deserve to be called "knowledge". The groundless knowledge pertains to systems of beliefs shared by all human beings, institutions, consolidated practices and techniques, in a word to the Weltbild, which is "the inherited background against which I distinguish between true and false".18 Like Wittgenstein, von Wright calls the groundless knowledge "certainty". Thus, the propositions expressing certainties have a peculiar epistemological status: being independent on any experience and empirical evidence, they are neither true nor false, and are free of spatio-temporal determinations. In so far as they are about the conditions that make it possible to determine (justify, ground) the truth or falsity of empirical propositions, the propositions which are undoubtedly certain lie, as Wittgenstein says, "beyond being justified or unjustified",19 and in this sense they are without grounds. Thus, an entire order (class, domain) of propositions, which traditionally has been labelled as "knowledge", no longer has this epistemological status. 3.3 In Determinism and Knowledge of the Future, von Wright includes among these propositions the laws of nature, coming to expose the category error of ascribing to them the function of justifying knowledge. The latter can be given only through propositions which are derived from experience or based on contingent facts. Pieces of empirically grounded knowledge belong to a logical level which is different from that of the certain knowledge and the natural laws, and in this sense they cannot interfere or contradict each other. However, this categorial difference does not relegate the groundless knowledge to a sort of sub-level of arbitrariness and subjectivism, but simply means that, like laws of nature, the certainties are epistemically groundless, but just because are pragmatically grounded. The application of the analysis of the grounds for knowing at the level of the grounds for acting allows to solve the "conceptual misconception" of viewing the causal determination of events as a "menace" to the freedom of the human action. Once the level difference is established, the deterministic categories stop playing any role in the area of intentional behaviour: it remains among them a sort of principled parallelism, which refers to their compatibility and non interference. That 18

Ibid., § 94. Ibid., § 359.

19

Grounds for Acting, Grounds for Knowing

63

man is subject to natural laws does not challenge the freedom of his actions. A paradigmatic von Wright's passage is perhaps the best description of the way to escape from the puzzle: "There is no circularity of a logically vicious kind in this fact that mankind is both slave and master of his own destiny" ,20

References SCHILPP/HAHN (1989). Paul Arthur Schilpp and Lewis Edwin Hahn (eds.): The Philosophy of G. H. von Wright. La Salle 1989. VON WRIGHT (1963). Georg Henrik von Wright: Norm and Action. London 1963. VON WRIGHT (1971). Georg Henrik von Wright: Explanation and Understanding. London 1971. VON WRIGHT (1974). Georg Henrik von Wright: Causality and Determinism. New York 1974. VON WRIGHT (1976). Georg Henrik von Wright: Determinism and the Study of Man. In von Wright (1983); first publ. in 1976. VON WRIGHT (1980). Georg Henrik von Wright: Freedom and Determination. Amsterdam 1980. VON WRIGHT (1981). Georg Henrik von Wright: Explanation and Understanding of Action. Revue Internationale de Philosophie 35, 1981; also in von Wright (1983). VON WRIGHT (1982). Georg Henrik von Wright: Determinism and Knowledge of the Future. In von Wright (1984); first publ. in 1982. VON WRIGHT (1983). Georg Henrik von Wright: Philosophical Papers, I. Oxford 1983. VON WRIGHT (1984). Georg Henrik von Wright: Philosophical Papers, III. Oxford 1984. VON WRIGHT (1988). Georg Henrik von Wright: An Essay on Door-Knocking. Rechtstheorie 19 (1988). VON WRIGHT (1989). Georg Henrik von Wright: Intellectual Autobiography. In Schilpp/Hahn (1989). WITTGENSTEIN (1953). Ludwig Wittgenstein: Philosophische Untersuchungen. Philosophical Investigations. Oxford 1953. WITTGENSTEIN (1969). Ludwig Wittgenstein: Über Gewißheit. On Certainty. Oxford 1969. WITTGENSTEIN (1980). Ludwig Wittgenstein: Vermischte Bemerkungen. Culture and Value. Oxford 1980. WITTGENSTEIN (1989). Ludwig Wittgenstein: A Lecture on Freedom of the Will. Philosophical Investigations 12 (1989), also in Wittgenstein (1993). WITTGENSTEIN (1993). Ludwig Wittgenstein: Philosophical Occasions 19121951, Indianapolis 1993. 20

Von Wright (1976), p. 52.

OLAV GJELSVIK On Mind and Matter

Abstract: Georg Henrik von Wright's paper "On Mind and Matter" develops an interesting and important view on the grand philosophical issue of the relationship between the mental and the physical. It also professes an affinity with epiphenomenalism. I worry about epiphenomenalism, and I find it necessary to avoid it to keep some basic and indispensable ideas in the way we view ourselves. Von Wright is moved towards epiphenomenalism partly because he sees grave objections to rival views like identity-theories. I share many of his objections to rival views. In this paper I try to explain why we should worry about epiphenomenalism, and also how we can avoid the dilemma of choosing between epiphenomenalism on the one side and mind-brain identity-theories on the other. I outline briefly a view which is a weaker "Aristotelian" version of philosophical materialism, to which von Wright's arguments against identity-theories seem not to apply, and which is not epiphenomenalist. I offer this for as a suggestion for where to go, and not as a fully worked out view. 1 2 3 4 5 6 7 8

What is Epiphenomenalism? Why is There a Problem in Seeing Reasons as Epiphenomena? Von Wright's Affinity to Epiphenomenalism Psycho-Physical Parallelism, the Identity-Theory and Epiphenomenalism The Trouble with the Identity-Theory A Dilemma A Way Out of the Dilemma? Where this Way Leads

Georg Henrik von Wright's paper "On Mind and Matter" develops an interesting and important view on the grand philosophical issue of the relationship between the mental and the physical. I admire his way of doing philosophy very much. When I disagree with him, I sympathize deeply with the concerns which move him. I shall focus on an issue which I believe puzzles all of us. The issue is epiphenomenalism of the mental in the philosophy of mind. Von Wright's professes an affinity with such epiphenomenalism. Here are some views of von Wright's I think are absolutely right: Interactionism of the Cartesian sort is unacceptable, and so is "reductionism". There are at least these differences between us: I believe I am a philosophical

66

Olav Gjelsvik

materialist, but of a weak or minimalist type. I try not to be an epiphenomenalist. Von Wright voices worries about mind-brain identity-theories of all sorts. I share these worries, but I believe that they should not result in abandoning weak materialism and as a result embracing epiphenomenalism. That price is too high to pay. As von Wright I am basically a dualist when it comes to explanation of why events occur. Von Wright voices a basic distinction between rational explanation and causal explanation, while my dualism is between explanation within the logical space of reason and explanation in the realm of natural law. I see the notion of cause as a basic and unreducible notion which we need, among other things, in order to account for why explanations of both types are good explanations. In a sense I see all explanations of why events occur as causal explanations, but then "causal explanation" has a different sense from what it has in von Wright's writings. I think we ought to be pluralists about causal explanation, where von Wright perhaps prefers pluralism about the meaning of "cause". Why is there an affinity to epiphenomenalism on von Wright's view? Let us ask:

1

What is Epiphenomenalism?

C.D. Broad formulated it thus: "Epiphenomenalism may be taken to assert one of two things, (a) That certain events which have physiological characteristics have also mental characteristics, and that no events which lack physiological characteristics have mental characteristics. That many events which have physiological characteristics are not known to have mental characteristics. And that an event which has mental characteristics never causes another event in virtue of its mental characteristics, but only in virtue of its physiological characteristics. Or (b) that no event has both mental and physiological characteristics; but that the complete cause of any event which has mental characteristics is an event or set of events which has physiological characteristics. And that no event which has mental characteristics is a cause-factor in the causation of any other event whatsoever, whether mental or physical."1 The basic point of epiphenomenalism can be formulated thus: Mental properties have no real causal powers. If events are seen as having both mental and physical properties, and we are monists, then they cause other events in virtue of their physical properties only. If events have only one kind of property, either mental or physical, and there is a dualism of mental and physical events, then mental events cause no other events. Causation hooks up the (1925), p. 472.

On Mind and Matter

67

physical events. If you are a dualist, mental events become epiphenomena, if you are a monist, mental properties seem eventually to be without causal significance. Consider this case from Wesley Salmon2 as an illustration: There is a rotating spotlight at the centre of a circular room casting a spot of light on the wall. A light ray from the spotlight to the wall is a causal process, the motion of the light on the wall is not a causal process in the way ordinary motion is. A moving spot of light at place A is not the cause of a moving spot at the adjacent place B a short time later. The two spots are rather related to each other as two successive mirror images are related. They mimic causal processes and may be mistaken for causal processes. What we see on the wall is not a process involving a real causal chain. The spots on the wall are reminiscent of epiphenomena in the sense that they fail to cause what they seem to cause. (The analogy is not complete; the spots do of course have some causal powers. There can hardly be a perfect analogy to epiphenomena among physical phenomena.)

2

Why is There a Problem in Seeing Reasons as Epiphenomena?

If mental phenomena are epiphenomena, then they are quite like the spots on the wall in the sense of having no real power to cause the things we take them to cause. They do not cause the events which follow them in time and seem to be caused by them. Causal powers are restricted to the physical. The scientific picture of our bodies as a physiological machine which we have inherited from Descartes not only excludes interactionism, it effectively excludes the mental from the realm of things which make a causal difference. The problem with epiphenomenalism is in my judgement this: If our sensations, reasons and actions lack the power to cause things, this is a threat to the conception we have of ourselves. As I see the conception we have of ourselves, our reasons matter causally for what we do and our actions matter causally for what happens in the world, and thereby, for instance, for how we come to have knowledge of the world. The world out there matters causally for what we perceive and believe, and neurophysiological events in us matter causally for what we desire. Our reasons and our actions must make a difference to what happens in the material world. If the difference they make is not a causal difference, then we do not know how they can make a difference. All difference made, everything which happens in the material world, seems attributable to the causal powers of happenings. It 2

Salmon (1977).

68

Olav Gjelsvik

is hard to see how we can really make anything happen, if nothing we do is among those things which make a difference.

3

Von Wright's Affinity

to Epiphenomenalism

Von Wright asks whether our reasons for an action are not causes of the action, and answers that it is unobjectionable to see reasons as causes as long as we see them as causing actions (and actions only). Therefore he seems, at first glance, to have no affinity with epiphenomenalism about reasons. Problems turn up if we ask about what actions cause. We might think that actions cause bodily movements and events in the world. Let us look at the causes of a bodily movement, "(the result of) an action of an agent."3 Beware of the nice parenthesis around "the result of! It opens for the possibility of an identification of the action and the bodily movement. We are then, von Wright says, "stepping into a conceptual morass."4 Here is the trouble we are in: What happens in the world outside our bodies seems caused by the bodily movement which again is a result of the action which is caused by the reasons. We seem to need to see it like that in order to be able to maintain the conception we have of ourselves as agents who make things in the world happen. The bodily movement is undoubtedly an event which causes events in the material world outside our bodies. In fact we need nothing more than the occurrence of this bodily movement at the present location to account causally for all those further material events outside the body. The bodily movement was a result of the action. It is not, however, caused by the action. The bodily movement is caused, von Wright stresses, by prior physiological events. These physiological events make up the full cause of the bodily movement - perhaps in a chain leading all the way back to sensory input. If this view on the causes of the bodily movement is right, and it surely seems to be right, then the action does not cause anything; there is nothing for it to cause. The action is epiphenomenal. (Recall Broad's definition.)5 If the action is epiphenomenal, the reasons have nothing but an epiphenomenon to cause, and they become epiphenomenal too, and so it goes, all the way back to experiential input. Von Wright does, however, think of the action somewhat differently, or so it seems: "A, the action, is M, the bodily movement, viewed (conceived, 3

Von Wright (1995), p. 3. Von Wright (1995), p. 4. 5 Von Wright (1995, p. 18) seems to identify the action and the bodily movement, when he says that the action, A, is the bodily movement viewed (conceived, understood) under the aspect of intentionality. I turn to this next. 4

On Mind and Matter

69

understood) under the aspect of intentionality. Viewing M under this aspect means relating it to the mental things we call reason for an action. This relation is not causal."6 Here von Wright seems to commit himself to a limited identity-theory, where he identifies the action and the bodily movement. I discuss this view in the next section. Note that the case of action then differs from the cases of sensation and reason, where an identity-theory leads to absurdities, according to von Wright.7 Let us look closely at the proposal at hand as regards epiphenomenalism. If the action is the bodily movement seen from a certain perspective, then it has, I assume, the same causal powers as the bodily movement. Can we then avoid the epiphenomenalist threat? Consider this: Whether or not the bodily movement can be seen from the intentional perspective at all, depends on whether we can relate the bodily movement to reasons. Those relations to reasons, if present, are not and cannot be causal according to von Wright. Therefore the relationships whose occurrence determine whether a bodily movement is an action or not, are not causal relationships. Then the natural thing to say is that the properties in virtue of which the bodily movement is an action are not among its causal properties. But then the action as action is without causal powers, since whether or not the bodily movement can be seen as an action does not depend on any of its causal properties. We are then left with epiphenomenalism of the monist type in the case of action (remember Broads two types). Since the action again seems to be epiphenomenal, the reasons, which we intuitively see as causing the action, are under the threat of epiphenomenalism as well. If the action causes no real event, and also is the only event caused by the reasons, there seems to be nothing in the causal order for the reasons to cause, and then the reasons become epiphenomenal too.8 And so it goes. If the causal picture of our sensations, reasons and actions is analogous to the case of the moving spots on the wall, there seems to be a threat to the conception we have of ourselves as agents which as agents make things happen in the material world. On von Wright's view, I believe, it is no part of the conception we have of ourselves that our actions themselves cause material events. He differs from classical epiphenomenalism because he does not see sensations, reasons and actions as caused by material things in the first place. But a basic similarity with epiphenomenalism remains, and I 6

Von Wright (1995), p. 18. I turn to this argument of von Wright's later. 8 Alternatively, if you become an identity-theorist, and identify the bodily movement and the action, reasons become epiphenomenal since the physiological events exhaust the set of causes of the bodily movement, and the claim that the reasons cause the action must be deemed false. The reasons must be identical with some of those prior events, and how can that be? 7

70

Olav Gjelsvik

am unsure about how much the difference to it matters for the problems. My worry is the similarity. To repeat myself: which reasons move us, which action we choose to perform must make a difference to what happens in the material world. If the way these things make a difference is not causal, we want an account of how they make a difference. I fail to see a workable alternative to causation at this point. Let us for a moment leave this issue, and approach things from a different angle.

4

Psycho-Physical Parallelism, the Identity-Theory and Epiphenomenalism

Since von Wright sees Rational Explanation and Causal Explanation as two distinct types of explanation, his way of understanding the psycho-physical parallelism is flavoured by that. At each end we have physical events, at the early end an incoming acoustic signal and at the late end an outgoing bodily movement. In between the ends we have a neural route of causes, and a rational route from experience through thought to action. First a question is about the first end. What is the relationship between the incoming sound and the acoustic sensation? It is, von Wright says, natural to think of it as cause and effect. How can that be? "A supporter of psycho-physical identity theory would say that the neural effect of the sound and the hearing of the sound, the acoustic sensation, are the same (identical). [... ] A 'conceptual gain' of identifying the sensation with something neural would be that then the sound uncontroversially is a cause of the sensation."9 The thought I want to latch on to, is that the identity-theory would, if accepted, make the sound the cause of the sensation, and thereby make the sensation part of a causal chain. There are arguments against the identitytheory, whereof von Wright gives two. The point I want to make here is the minimal point that the identity-theory seems to undo some of the affinity to epiphenomenalism in the case of sensation. If the action at the end of the mental process were to be thought of as identical with the bodily movement, it might by analogy be clear and uncontroversial how the action makes a difference to what happens, namely by causing further events, the very same events as the bodily movement causes. What is required for the gain to occur, is that what makes it possible to see the bodily movement as an action is that it is caused by reasons. If the properties in virtue of which the bodily movement is correctly seen as an action, does not include causal relations to reasons, I think that the epiphenomenalist threat which can be reduced when you are an identity-theorist, looms large again, as argued above. I think that von Wright might agree with this limited point, that the 9

Von Wright (1995), p. 15.

On Mind and Matter

71

"conceptual gain" he speaks about is much more clearly there if we treat perception, reason and action the same way, and as causally related to each other and thereby to their causes and effects as well.

5

The Trouble with the Identity-Theory

Let us look carefully at the arguments against identity-theories, and concentrate on the one argument von Wright sees as most important. The argument is this: A brain state exists in space, open to inspection from the outside by the neurophysiologists. The sensation, however, is registered only by the hearing subject. Only that person has it, and no one else. It seems absurd to identify the "private" sensation with a public, bodily, "material" state. I find von Wright's argument against the mental-physical identity-theories important. I see the argument as stating that there are certain requirements the referents of the terms in true identity-statement have to meet, and those requirements are not met by the two candidates here considered. Whether the identity-claim then becomes absurd or obviously plain false is not an issue of great concern to me. The point is rather that it would be absurd to cling to the thought that the identity-statement might be true. The crucial differences in this case are the difference in spatial location, and differences in availability to outside inspection. These are clear differences. It is right that the inspectability from the outside is a property of the neural event the other event seems not to have. This does not mean, however, that the event which is a person having the sensation in question is not located in space. It is located in space to where that person is. Persons are located in space, and they have minds. Any person's mind is epistemologically available in the peculiar way only minds are available, and which von Wright describes so well. We might say "I saw that he heard the sound". This is not, however, inspecting the sensation. The epistemological access to a mental event and that to a neural event are very different indeed. This is a conceptual point. Are these differences sufficient for ruling out an identity-theory? Von Wright seems to think so. In my judgement he seems to make too much out of the problem that mental events are not in space. If mental events have no spatial location, then an identity-theory is really absurd. But as noted above, it is not true that mental events have no location in space. Thoughts themselves have no location in space, but a person's entertaining of a thought has such a location: it is located to where that person is. There might be another Cartesian argument to the effect that thinking, and mental predicates which presuppose thinking, do not presuppose (con-

72

Olav Gjelsvik

ceptually speaking) spatial extension for the subjects which get ascribed the predicate. All material predicates presuppose extension for their subjects. By general metaphysical principles one can, perhaps, from this starting point reach the conclusion that res extensa and res cogitans are different modes of being and cannot be different properties exemplified by a single substance.10 But I do not think that we in the end can make sense of a subject without a spatial location and extension enjoying what we think of as objective thought and experience. I think Strawson (and Kant and others), and also Wittgenstein in his special way, have done much to help us appreciate that the Cartesian premise is wrong. Von Wright's reasons do however make a strong case against thinking of the relation between the mental and the neural as an identity-relation. Mental events are located in space in a way which is clearly different from the way a neural event is located in space. The neural event has a specific location mental events fail to have. Perhaps this in itself should lead us to give up the identity-theory.11 Furthermore, if we think of events as having constitutive objects undergoing changes, as on Kim's conception, we see that the mental events have the person as a constitutive object, while the neural event has neural entities as constitutive objects. If we think of events this way, the identity theory again seems impossible. The issue about whether we can see the mental event as having material properties is still open. I do not fully subscribe to Kim's view on events, but I see here an objection to the identity-theory which might be independent of specific views on events.

6

A Dilemma

We therefore seem left with a dilemma. As I see it, the dilemma von Wright sees is this: We seem forced to choose between the comfort of the prospect of avoiding epiphenomenalism and the causal views of the identity theory with its resulting absurdities on one side, and the affinity with epiphenomenalism we seem committed to if we refuse to accept these absurdities. Von Wright chooses the latter. It matters much for me whether von Wright recognises a dilemma a bit like this. There is a real dilemma for him if he agrees with me that epiphenomenalism is worrying: That it is a threat to our concept of agency which is so central, philosophically and otherwise. I would like him to respond to this last point, and I would like our interchange to start with this focus. 10

See Descartes (1974), pp. 32 and 254. See also Donagan (1978). This is Jennifer Hornsby's view, as I interpret her. See Hornsby (1980a).

11

On Mind and Matter 7

73

A Way Out of the Dilemma?

I believe there is a way between the Scylla of absurdity and Charybdis of epiphenomenalism. I believe we have to grapple with ontology and identities to find a solution, and that we cannot rest content with the epistemological turn von Wright recommends. We must take the further step of seeing epistemological practices as having ontological implications: The epistemological practice where the intentional has its home, should be seen as an explanatory practice with ontological commitments of its own. These entities we are ontologically committed to as long as we commit ourselves to this practice, and commit ourselves to the causal relations as they are depicted by the practice and in the practice. And at the last point we have no choice: we are so deeply committed to this practice that we cannot live the lives we live without this commitment.12 The pressures against seeing these entities as having causal powers, should be seen as stemming from the tendency to identify the realm of entities with causal powers as the realm of entities identified by the vocabulary suited for entering natural laws. But the latter identification must be resisted. To put it positively: There is a possible "conceptual gain" in an identitytheory which makes proper room for seeing perceptions, reasons and actions as having causal powers. But it seems impossible to enjoy this gain since the view results in difficulties if not absurdities of its own, difficulties we cannot live with. My suggestion is that an ("absurd") identity-theory might be sufficient for this conceptual gain, but it is not necessary. Happily for us, we can have the gain without the difficulties. I want to point to the parallel case of substances, and make use of the fact that identities between substances are much rarer than we might believe at first. In my view a statue is not identical to the lump of clay it is made of. Their relationship is one of constitution, not of identity. Both objects might cease to exist while the other does not; there are conceivable changes which only one of them can survive. Think of them a Statue and Lump. Lump constitutes Statue, and they are not identical. Furthermore, neither of them should be thought of as a mereological sum of atoms or molecules. I think the latter is very important for our purposes. It has been well argued by Mark Johnston.13 To single something out as a statue, we have to think of it as having been made with something like an artistic intention. Since this is so, we can see that particular origin as essential for something to be a statue. It is not 12

This I see as the moral of Strawson's celebrated essay on Freedom and Resentment (Strawson 1962). 13 See Johnston (1992). See also Price (1992), and many other writings critical of mereology.

74

Olav Gjelsvik

essential for being a lump, or for being Lump. Presented with Lump, we might wonder whether Lump is a statue. Presented with Statue, we might wonder whether it is a lump of clay or a lump of something else. Statue and Lump occupy the same place at the same time, and their actual causal powers are the same: They share a set of actual causal powers. If I knock you in the head with Statue, I knock you with Lump, and the effect on you is one and the same.14 Attributions of causal powers make up extensional contexts, causal explanations make up intensional contexts.15 The important point for my purposes is that Statue and Lump are not identical. Two distinct practices come together here, and result in distinct entities from an ontological point of view; entities with very much in common, among other things their actual causal powers. The way I think I see out of the dilemma is that we view the relationship between some mental and physical events on the model of the relationship between Statue and Lump. In particular I want to suggest that we view the relationship between the action and the bodily movement on this model. Von Wright seems to want to think of this relationship as one of identity, but I hesitate. There are general reasons for hesitating, but also more specific ones. I want to dwell on the specific ones. A bodily movement is something which happens to the body. When I move the body, that is something I do. There is a transitive verb "move", (mover) and an intransitive verb "move", move/.16 I can mover the chair, and when I do, I act. Then the chair moves/. The movement/ of the chair is not an action on the chair's part. The chair is not even capable of agency. The phrase "the movement of the body" is in itself ambiguous between the movement^ of the body which is an action, and the movement/ of the body which is something which happens to the body as it happens to the chair in the other case. As in the case of the moving/ of the chair, no action or intention is necessary for there to truly be a moving/ of the body. In the case of the moving/ of the chair we see the action, the moving^, as its cause. We might wonder whether there also is a causal relationship between the moving^ of the arm and the moving/ of the arm. I have argued against this move elsewhere, their relationship should not be seen as causal.17 So what is it, if it is not causal? There has to be a movement/ of the arm for there to be an action. On the other hand such a movement/ of the arm is only necessary, not sufficient for there to be an action. For there to be an action, there has to something I do, and not merely something which happens to my arm. We can say, roughly, that there being a movement^ is 14

I argue this point in a brief paper in Gjelsvik (1988). See Davidson (1980a). 16 For this distinction, see Hornsby (1980b). 17 In Gjelsvik (1990). l5

On Mind and Matter

75

parallel to there being an action when there is an intention, a special origin in me is required. (Remember that I can mover my arm when I am asleep and have no intention whatsoever.) There has to be an intention for there to be an action; an action is a movement^ with a particular causal origin.18 So my suggestion is this: the relationship between the action which is the bodily movement^ and the bodily movement/ is not one of cause and effect and not one of identity. It is one of constitution; it is like the relationship between Statue and Lump. The origin which is essential for the action to be an action is not essential for the movement/ to be a movement/ of the body. Still they may happen at the same place at the same time and have the same actual causal powers.19

8

Where this Way Leads

We have so far just looked at the case of the action and the movement of the body, where von Wright is tempted by an identity-view. I want to subscribe to the sentence "the action is the bodily movement". It is true on both readings. If "bodily movement" means movementr, things are easy, but when "bodily movement" means movement/, then the "is" is the "is" of constitution, the sense in which the statue "is" a lump of clay. This is a weaker view than the view von Wright's suggests at this point, since it upholds an ontological distinction between the action and the bodily movement (when the latter is understood a particular way, i.e. the intransitive reading "movement/"). This distinction is based on the noted ambiguity in the verb move, and the fact that an action requires a special origin. The origin required is that it is caused by reasons in the right sort of way. For that to be the case, reasons have to be causes. What is the most significant difference between the case of action and movements of the body and the case of sensation and the case of reasons? The most significant difference is that the case of action and bodily movement is like the case of Statue and Lump in the respect that there are natural candidates from common sense ontology and our epistemological practices for being identical. Lump stands to Statue as bodily movement M stands to action A in von Wright's discussion. In the case of sensation S and reason R, there is no reasonable natural candidate with the role of M. But the absence of any such candidate should not make us deprive S and R of their 18

I do not want to enter the difficult debate about whether there has to be an intention for here to be a movement χ of the body. I think not, and this view of mine complicates the picture: For there to be a movement?· there has to be an internal origin for the movement. But also in a reflex movement I mover my hand, but without an intention. 19 I owe much to David Wiggins's many works when I speak of constitution-relations.

76

Olav Gjelsvik

causal powers. The fact that neither of them can be identical to a sum of micro-events in the brain, should be seen as a nice companion to the fact that neither the action nor the bodily movement is such a sum. Such mereology will not work anyway. The fact that any possible candidate along the lines of construction out of microevents, will have to be secondary to the intentional event when it comes to identification, a point heavily stressed by von Wright, is just the mirror image of the point that there is no event with the function of Lump here. The other point, that the systems of identification we employ in our explanatory and classificatory schemes suffice for ontological commitment, as it surely does in the Statue and Lump case, therefore extends to this case, and we can speak of both the perception and the having of a reason without having anything like a candidate for an event which constitutes them. The explanatory scheme we employ commits us to seeing these items, perceptions and the entertaining of thought, as having causal powers as the action does. This does not mean that the we do not study the neurophysiological correlate of these things when we study what happens in the brain, as we do when we study the correlate of the action. I believe we have to see the explanatory practices and schemes we use and trust as ontologically committing. Perceptions, reasons and intentions, the inhabitants of the logical space of reasons, also enter causal relations as long as we speak of a person's perceptions, a person's having of reasons, and a person's actions. Neurophysiological events enter the realm of natural law and enter causal relations seen that way. Our understanding of both realms presupposes the concept of cause. A perception is a causal effect of the sensory input which is answerable to rational norms, and an intentional action is a causal effect of our beliefs, desires and reasons which is similarly answerable to rational norms. Perceptions, reasons and actions do not enter the realm of natural law as neurophysiological events do. The realm of natural law is filled by causes, and the notion of nomic connection presupposes the notion of cause. Nomic connections make up one out of many causal mechanisms by which we, epistemologically, identify causes in this explanatory practice which also, by and large, contains science. In science, explanation by generalities, and in particular by (nomic) laws, is to be preferred when it can be had. Still the basic philosophical point is that there is more to the realm of causes than what fills the realm of natural law. The notion of cause is primitive, as primitive as truth, and cannot be defined by notion which do not presuppose cause. At best it can be elucidated as truth also can. (Tarski showed the way). The notion of cause has a basic role to play in all issues about individuation of substances, events and their properties. So I agree with von Wright that there is an epistemological priority of the mental over the neural, in the specific sense that the mental has to be

On Mind and Matter

77

identified the way we always identify mental phenomena in order to study correlates of the mental. There is a semantic priority of the behavioural in relation to the mental, but here behaviour must be understood as an expression of agency within the logical space of reason and not as a product of neurophysiology. There is a causal priority of the neural over the behavioural, but only as long as the behavioural is not understood as an expression of agency. If the behavioural is understood as an expression of agency, however, reasons are causally prior in explanations of why events (actions) occur. The relationship between the neural and the mental in general is one where we also might want to raise the issue of (causal) priority. Clearly mental capacities depend heavily on the functioning of the neural and the physical. They are not necessarily caused by the neural even if they depend on it, but they might be. A person's loss of his eyesight or his getting forgetful are obviously correlated with neural or physical changes. Because that is so, these mental changes can be said to be caused by whatever causes those neural changes. The dependencies mentioned are at the heart of my (weak) materialism (without mental-physical identities): On the whole the mental depends on the physical; there is supervenience. This view requires seeing all things and events as having physical correlates, otherwise we do not understand rightly how there could be such dependencies. I do not here want to go into the task of specifying this supervenience or dependency. Causal explanations of mental changes which presuppose (and are based upon) existing dependencies of mental properties on physical properties are quite different explanations from explanations of actions or beliefs: They are cases where one exploits known dependencies to extend explanation within the realm of natural law into the mental. That can be done in many cases without amounting to any sort of threat to the epistemological practices within the logical space of reason: they are explanations of phenomena which cannot be properly explained within the logical space of reason. Therefore they do not in any way compete with rational explanations; they simply have different explananda. To sum up: People can be seen as biologically functioning bodies and also as persons. These perspectives are very different even if there are many meeting points and dependencies. The difference is not that one perspective is permeated by causality and the other is not. Both are permeated by causality. Apart from having causality in common the perspectives are indeed different. Functioning bodies are studied in neurophysiology, persons have experiences, entertain thoughts and act for reasons.20 20

I thank Eyjolfur Emilsson for helpful comments, and the participants in Bielefeld, especially G. H. von Wright and F. Stoutland.

78

Olav Gjelsvik

References BROAD (1925). Charlie Dunbar Broad: Mind and Its Place in Nature. London 1925. DAVIDSON (1980a). Donald Davidson: Causal Relations. In Davidson (1980b). DAVIDSON (1980b). Donald Davidson: Essays on Actions and Events. Oxford 1980. DoNAGAN (1978). Alan Donagan: Descartes' 'Synthetic' Treatment of the Real Distinction between Mind and Body. In Hooker (1978). DESCARTES (1974). Rene Descartes: The Philosophical Works of Descartes. Cambridge 1974. GJELSVIK (1988). Olav Gjelsvik: A Note on Objects and Events. Analysis 48 (1988). GJELSVIK (1990). Olav Gjelsvik: On the Location of Actions and Tryings. Erkenntnis 33 (1990). HOOKER (1978). Michael Hooker: Descartes, Critical and Interpretative Essays. Baltimore 1978. HORNSBY (1980a). Jennifer Hornsby: Which Mental Events are Physical Events. Proceedings of the Aristotelian Society 81 (1980). HORNSBY (1980b). Jennifer Hornsby: Actions. London 1980. JOHNSTON (1992). Mark Johnston: Constitution is not Identity. Mind 101 (1992). PRICE (1992). Huw Price: Metaphysical Pluralism. Journal of Philosophy 89 (1992). SALMON (1977). Wesley C. Salmon: An 'At-At' Theory of Causal Influence. Philosophy of Science 44 (1977). STRAWSON (1962). Peter F. Strawson: Freedom and Resentment. Proceedings of the British Academy 48 (1962). VON WRIGHT (1998a). Georg Henrik von Wright: On Mind and Matter. In Wright (1998b). VON WRIGHT (1998b). Georg Henrik von Wright: In the Shadow of Descartes. Essays in the Philosophy of Mind. Dordrecht 1998.

FRIEDRICH KAMBARTEL Remarks on Psycho-Physical Parallelism Abstract: The first part of the following remarks distinguishes two varieties of parallelism between physical and practical events. A thought experiment is described, which demonstrates the impossibility of a parallelism that is based on necessary and sufficient conditions and therefore constitutes causal dependencies. Subsequently the conceivability of a weaker form of parallelism is vindicated in the light of a particular understanding of strictly specific conditions. The second part focuses on the physical causes of sensations and perceptions. It is argued that the conceptual boundaries between the physical world and the world of sensations preclude the idea of a continuous causal chain, but nevertheless permit the application of an instrumentalist concept of causation; in the case of perceptions the pragmatic aspects of the involved judgemental structure seem to rule out the possibility of any causal determination. 1 2

Some Remarks on So-Called Parallelisms between Physical and Practical Events On the Conceptual Problems with Sensations and Perceptions

1

Some Remarks on So-called Parallelisms between Physical and Practical Events

In a weak sense, we could speak of specific (e.g. physiological) conditions C for an act (type) ο in a situation where an actualization (performance) of ο is only possible if C are fulfilled. A stricter form of specificity would then in addition include that the conditions C can only hold when accompanying an α-token. The stricter form of specific conditions, it seems, roughly reconstructs our understanding of a psychophysical parallelism. Namely: It would allow us to "translate" physiological occurrences into parallel (corresponding) acts and, in the reverse direction, to derive the validity of the necessary physiological basis from the performance of certain acts. With acts, though, both directions of this parallelism lead to conceptual difficulties. In both directions, namely, we could use it for causal strategies. But bringing about a token of α by realizing C can, in fact, only mean causing compulsory behavior, contradicting our assumption that the parallelism holds between acts and physiological events. The contradiction

80

Friedrich Kambartel

becomes clearer when we try to suppress a in cases where C is valid. This not being possible would then make us experience or feel the compulsory situation mentally. We are familiar with such cases from psychopathology, compulsory neurosis e.g. Causing C by actualizing a, on the other hand, while not inducing this sort of problems, may create other ones. In cases where C must be realized before a takes place, we would then have to consider the possibility of causally influencing the past. We also run into conceptual problems in cases of simultaneous causation, namely, when C is brought about by a causal history independent of the performance of a (as we normally assume to be the case with physiological conditions). Is there a way out? - We could say that strictly specific conditions are in some sense unthinkable. There cannot be experience or experiments which contradict this. The physiological basis of action is and can only be related to a (general) competence as a whole, not to the special acts which form part of this competence. Searching for specific physiological brain correlates to, let us say, the (semantically) well-determined utterances of a sentence would then be in vain. The results of physiological research thus far do not contain any contradicting data. They mainly correlate e.g. brain areas and levels of stimulation with a generally defined competence. A great deal of the contributions in this area come from the investigation of brain injuries and their connection to the loss of certain abilities. Just think of neurophysiological aphasia research. But naturally, a grammatical problem cannot be solved empirically. We have to find conceptual reasons which make the "empirical" results inevitable. Such reasons are available if we change the concept of strictly specific conditions (and thus that of parallelism) a bit: A strictly specific condition C in the new sense is still a necessary condition for the actualization of an act a. But it may be fulfilled without the corresponding occurrence of a. Although, if we wanted to, we could then produce an actualization of a. Being strictly specific, however, C should not be the necessary physiological basis for another (different) act a*. If those specific conditions C can arise independently of an α-token, there is no reason why they should be caused by an actualization of a, or an actualization of ο be caused by C. In this framework, the question as to whether C will indeed occur independently of a appears to be of an empirical nature: We may work out experiments in which we produce C with a person Ρ - or this (that C holds) may just happen. In such situations, Ρ should now try to forebear enacting a. If Ρ succeeds, the independence under investigation has been proven. If

Remarks on Psycho-Physical Parallelism

81

P should regularly fail to forbear enacting a, the resulting α-behavior would be compulsory. That the actualization of α be an aci-token (and thus not compulsory) was, however, essential to the whole setting of our argument. This construction of alternatives uses experimental fictions. The distinction of cases used to support our argument, however, is complete. It therefore shows on purely conceptual grounds that a parallelism allowing a causation of act-tokens is not possible. Let me further illustrate these considerations by a simple model: Imagine a stab or a suitably bandaged arm which has a solid connection with a kind of propulsion shaft which is itself turning around a wheel. Movements of the stab or arm thus translate into movements of the wheel (which is the physiological counterpart in the model). If the wheel is blocked, acts of moving the stab or the arm cannot be performed. A certain technical or physical condition of the wheel - its being able to revolve freely - here becomes a necessary and sufficient condition (and therefore a strictly specific condition in the sense of p. 75 f.) for the execution of certain acts. On the other hand, movements of the stab or arm by a causal mechanism induce corresponding movements of the wheel, and conversely, movements of the wheel cannot occur without the corresponding movements of the stab or arm. This is obviously a parallelism causally connecting movements of the wheel and certain motor behavior, and this in both directions. Assuming a suitably constructed wheel, its movements will coerce the corresponding movements of the stab or arm, which - should we try to suppress them could be experienced as compulsory and not as action.

2

On the Conceptual Problems with Sensations and Perceptions

Concerning so-called sense perception (in German: Sinneswahrnehmung) three things should be carefully separated. Taking acoustic sensations as an example, these are: (a)

the corresponding physical events or states, e.g. the spreading of sound waves or physiological events in the ear or the brain;

(b)

sounds and noises, sometimes described as "pure" sensations, in the narrow sense of the word (e.g. a long, high, shrill sound);

(c)

hearing a sound as a well-determined, interpreted perception (e.g. hearing "the whistle of an engine").

Between the languages in which we talk about physical events on the one hand, about sensations and perceptions on the other, there are conceptual boundaries. To describe sensations or perceptions in terms of physics would

82

Friedrich Kambartel

be a category mistake. Therefore, it does not make sense to say - as materialists might do - that a certain oscillation of the air is a sound. We can talk about the occurrence of sound waves in the absence of somebody having heard them, whereas a sound is only possible as a heard sound. Imagining a sort of transformation of the physical into the mental disguises the unbridgeable conceptual gap between the two spheres. Relations between what we hear and certain neurophysiological events or oscillations of the air or water, all these relations are empirical relations. The semantics of talking about sensations and perceptions obviously does not stipulate these relations. Think of Wittgenstein's empty head argument. Naturally, there may come a time in which we invent a new (common) language enabling us to talk about the physical and the mental in one and the same conceptual frame. Then the old puzzles and mysteries might vanish! At the moment, however, I cannot conceive of such a situation. Can there be a causal influence between events on different sides of this conceptual boundary? A simple instrumentalist understanding of causality seems to have no problems with this: In the elementary case that events e cause corresponding events e* means that we can produce the e*-events by establishing the corresponding e-events. A physical, e.g. mechanical continuity of causal steps (a chain of close connections) is not conceptually necessary. It is a fact that sound waves or physiological events cause certain acoustic sensations. We know that by producing sound waves or physiological occurrences of a certain kind, at the same time we cause sounds with certain attributes. If, however, we think of a continuity of influence across conceptually defined borderlines, we get a picture which makes no sense. Therefore, the causal impact of the physical upon the mental, though a fact, remains a riddle. How do things change if we consider perceptions instead of sensations? We can, it seems, readily imagine that physiological conditions specific to a certain perception cause this perception. We may think of brain states or events in our sense organs as such causes. And yet the situation here is not quite as simple as it may seem: What we perceive obviously depends not only upon physiological states or events but also on our understanding of the relevant situation. This again cannot simply be conceived of as an effect of yet other specific physiological structures which interfere with the immediate physiological causes of our perception. This remark holds because our interpretation of the situation has a pragmatic component. We do not acquire such interpretations by mere induction from private experience, but essentially in interaction, meaning especially dialogical practice and linguistic behavior.

Remarks on Psycho-Physical Parallelism

83

In general, the hermeneutic activities of our biographical and social past provide a familiar, self-evident background ("horizon") B for the classification and integration of our perceptions. We may think that certain brain physiological conditions are built up parallel to the forming of B. These "engrammata" could then work jointly with the immediate physiological causes of a certain perception and thus causally fix the hermeneutic part of our perception. Paranoid perception is perhaps of this form. In the non-pathological case, the pragmatic aspect of perception has consequences. Perceptions can obviously undergo subsequent hermeneutic deliberation, leading us to distinguish between what we thought we perceived and what we really perceived. Perceptions are thus located between events which happen to us and those produced by our action. Even so-called immediate perception is generally accompanied by an element of judgement. It is in this sense that normal perceptions differ from compulsory cases; and therefore what I said about acts in a slightly modified way should also be valid for perceptions.

FRANZ VON KUTSCHERA Explanation and Understanding of Actions Abstract: This statement was to serve as a short introduction to the papers on explanation and understanding of actions. Therefore it only briefly mentions two key-topics in this area and makes a few statements on them intended to trigger the discussion. The topics are: The nature of rational explanations and what they explain - as I see it, they explain not the occurrence, but the rationality of an action -, and the compatibility of rational and causal explanations of an action - referring to a rather exclusive notion of causality I maintain that what admits of rational explanation does not admit of causal explanation. 1 2

What Is Explained by Rational Explanations? Are Rational Explanations Compatible with Causal Ones?

1

What Is Explained by Rational Explanations?

Since we understand something if and only if we can explain it I can confine my remarks to the explanation of actions. My main topic are rational explanations. The present discussion of them was triggered by William Dray with his book on Laws and Explanation in History (1957) and has gained much by Georg Henrik von Wrights analyses in (1971). Their ancestral line goes back to Aristotle's practical syllogism. Different versions of it appear in his works, but its point comes out most clearly in De anima III, 11 (434al7-21) and De motu 7 (701a7-25) where it is formulated thus: (I)

Every man in a situation of type S should do F. I am in a situation of type 5. Therefore: I should do F.

This is an argument by which an agent, in a situation where he has different options, decides on what to do. It is not an explanation of the fact that he actually does F or will do F. The validity of the argument in no way depends on whether he follows his own advice. The point of the syllogism is not explanatory but deliberative. It gives the agent a sufficient reason for doing something.

86

Franz von Kutschera

A practical syllogism more in the spirit of decision theory would be this (II)

In the present situation my preferences and expectations are such that the expected utility of doing F is higher than that of any other option open to me. Therefore: I should do F.

This deliberative argument immediately transforms into a rational explanation (III) The preferences and expectations of X in situation S are such that the expected utility of doing F is higher than that of any other option X has in S Therefore: It is rational for X to do F in S. A rational explanation of an action, generally speaking, is an explanation by the aims, intentions or interests of the agent and his beliefs or expectations. According to most modern authors, however, in order to explain the action, it has to be an argument with the conclusion that it was actually performed, an argument showing that, given the premises, its occurrence was necessary or at least highly probable. For that the premises of (III) are clearly not sufficient. This, also, is a frequent objection to Aristotle, who, especially in De motu (701al2-13 and 702al6-18), says that a deliberation of type (I) leads to immediate action if there are no external obstacles - in the Nicomachean Ethics VII, 3 sect. 6-9 (1147al-36) he also takes weakness of will into consideration. What is missing is a premise to the effect that the agent X is a rational person, i.e. that he, either generally or in the specific situation, does the rational thing. In both versions, however, such an assumption is problematic. Probably no one acts rationally all of the time, and a claim that the agent behaves rationally in the specific situation can only be ascertained by observing his actual behavior. For an explanation, a premise relying on the explanandum is of dubious value, however. Are rational explanations unsound, then? I think not. The first step towards a re-evaluation is the recognition that actions are events, and that an explanation of an event is not always an argument with the conclusion that it happened. There are different things to understand about an action, different properties to account for: that it was murder and not only manslaughter, for instance, why it was done in a certain way, at a certain time or place. All these explanations have different conclusions. There is, then, no objection to accept (III) as an explanation of X's doing F, provided X actually did it. The premises of the argument do not imply that X did F, but they conclusively show that this action had a certain property, namely that of being rational for X in the given situation. A rational explanation is an answer to a why-question. An action is something the agent could have refrained from doing, i.e. something resulting

Explanation and Understanding of Actions

87

from a choice. If we ask why he did it we want to know why he chose this specific option, what his reasons were. An answer is given by a rational explanation, and if he really had a choice in the manner and his behavior was not determined by outside circumstances, this is the only possible answer.

2

Are Rational Explanations Compatible with Causal Ones?

This brings me to the second point I want to raise, the compatibility of rational and causal explanations of actions. (Like the first point it has been elaborated in Kutschera 1993.) If an episode in the behavior of a person X is an action in the sense just specified, i.e. if X is its originator, it cannot also have a second origin, a determining cause. (The decision itself is no cause, since it does not necessitate the action, as we already have seen.) Actions, therefore, do not admit of causal explanations. This thesis, like any other one on the compatibility or incompatibility of rational and causal explanations of actions, has to be based on a specific conception of causality. In view of the divergent theories of causality-regularity theories, counterfactual analyses, probabilistic and modal theories - it makes no sense to discuss the question without saying what type of causal explanations one has in mind. My incompatibility thesis relies on the notion of causality I have described in (1993a). It is related to the one von Wright has developed in (1974). According to him, a causal law stating that A-events are necessarily followed by B-events, in contradistinction to a statement of a mere regularity in the normal course of events, presupposes that there are Α-events that normally do not happen but can be brought about by an agent. Agents and actions are, then, conceptually prior to causal laws, and therefore actions cannot, generally, be causally explained. This, I think, was an important step towards a revision of the official doctrine that causal explanations are the only respectable and scientifically sound ones. In an intellectual atmosphere comparable to our own today Platon, in the Phaidon, turned against the 'Physicists' of his time. Socrates has been condemned to death and now awaits his execution in prison, although he was given an opportunity for flight. Why did he stay? A rational explanation would run like this: "Socrates is now sitting in his cell because he rather wants to suffer the death penalty than break Athenian law". The answer of the Physicists, however is: "Socrates sits there because his legs consist of bones and sinews, and the sinews are loosened so that the bones are bent at the knees." "But", says Socrates in the dialogue, "to call that an explanation is simply absurd". And we may add: Even if the Physicists had spoken of neural synapses and excitations instead of bones and sinews their explanation would have lost nothing of its absurdity.

88

Franz von Kutschera

References DRAY (1957). William Dray: Laws and Explanation in History. Oxford 1957. KUTSCHERA (1993). Franz von Kutschera: Die falsche Objektivität. Berlin 1993. KUTSCHERA (1993a). Franz von Kutschera: Causation. Journal of Philosophical Logic 22 (1993). VON WRIGHT (1971). Georg Henrik von Wright: Explanation and Understanding. London 1971. VON WRIGHT (1974). Georg Henrik von Wright: Causality and Determinism. New York 1974.

REX MARTIN Action Explanations as 'Understanding' Explanations Collingwood and von Wright

Abstract: My paper is concerned with setting out the views of Collingwood and von Wright on the explanation of action. Here I will identify a single main model or schema for the explanation of actions (that is, for explanations of actions by reference to reasons - to certain thoughts and motivations of the agent). This model, in my view, provides the root of both von Wright's notion of practical inference and Collingwood's idea of reenactment. In this paper I will turn as well to a critique of their two theories, by taking up and contrasting the role of understanding or intelligibility, often called Verstehen, in each of their accounts. 1 An Explanation Schema for Actions 2 Von Wright on Understanding 3 Collingwood on Understanding 4 Failures of Intelligibility or Understanding 5 The Truth About Efficacious Reasons 6 Collingwood's Alternative View 7 A Summary and Conclusion Appendix to Footnote 14

In my judgement R. G. Collingwood and Georg Henrik von Wright have done some of the most interesting and creative work, in our time, on the theory of action explanation. The present paper sets out to explore and contrast their main contributions in this area.

1

An Explanation Schema for Actions

One of the standard kinds of explanation is that in which an action of an agent is accounted for by reference to certain thoughts and motivations that the agent has. Such explanations are often called intentionalist explanations. Characteristically, such an explanation would look something like the following one. The agent did A (the deed performed) because (i)

the agent perceived that he or she was in a certain situation and was disposed to act toward it in some definite way;

90

Rex Martin

(ii)

there were a number of alternative courses of action (designated A, B, C, .D, and so on) open to the agent who had the situational motivation described in (i);

(iii) the agent did want to achieve or accomplish such-and-so end, which (iv) the agent believed would satisfy his or her initial situational motivation; (v)

the agent believed that doing A was, in the circumstances already described, a means to accomplishing the stated purpose in (iii) or a part of achieving it;

(vi) there was no action other than A, which action was believed or seen by the agent to be a means to the goal, that the agent preferred or even regarded as about equal; (vii) the agent had no other purpose which overrode that of accomplishing such-and-so; (viii) the action to be taken was timely and, when the time was ripe, the agent had not forgotten the relevant purpose, overlooked the time, or what have you, and (ix) the agent knew how to do A, was (generically) able to do it, and physically able to do it in the situation as given and, at the timely moment, had the opportunity, and so on. We can take some such schema, in a suitably amplified version, as the standard one for the purposes of this paper.1 An explanation of an individual action (in the account I am giving) is, then, an exemplification of the schema just developed. Such an explanation is afforded by substituting, under each condition of the schema, statements of fact that satisfy - in one way other or another - the terms of that condition. Every such explanation breaks down, then, into two main parts: (a) the formal part as given in the schema itself and (b) a material one, represented by the statements of fact that satisfy the schema in a given case.

2

Von Wright on Understanding

I want, next, to indicate a second dimension to this account, the dimension of understanding (or Verstehen). In his more recent writings, von Wright has attempted to give Verstehen a constitutive role in his account of inten1

This schema, as will become clear as we proceed, draws upon both von Wright and Collingwood.

Action Explanations as 'Understanding' Explanations

91

tionalist explanations. This particular line of development reaches its fullest statement in his idea of an "understanding explanation."2 In this more recent work, von Wright has, in effect, bifurcated intentionalist explanations into two main parts (or sub-schemas, if you will). One sub-schema follows, roughly, the means/end analysis he originally seemed to lean to.3 Here the sense - the only important sense - in which we understand a person's action is that, once we have the agent's beliefs in hand, then we can "see" or interpret (even if we had not been able to beforehand) the agent's action as a means to some particular end or goal. In defense of this view von Wright has said, "[I]f we come to the conclusion that he really believes that he must do A in order to achieve B and that B is what he is after, then we also understand why he did A."* For, clearly, there is a sense in which we understand the agent's action in such a case; we understand it as a means to B. That is how we see it - as serving in that role. Accordingly, we can be said to understand why the agent did A (namely, to achieve B). Or, to put the point differently, we can explain the agent's doing A by referring to B (and to the agent's means/end belief). However, the sense in which we understand a person's action in the other sub-schema von Wright had in mind is quite different from this. Here the agent's situation cum motivation is envisioned as a sort of "external impulse" or "demand" to which various possible alternative courses of action (including the action actually performed) are responses. But one does not understand these various courses of action by reference to the agent's belief that some (or one) of them would be responses to demands. Rather the onlooker or investigator here more or less directly "sees" or interprets these courses of action as responses - and does so without any intermediary agent's belief as an element (or necessary element) in the resultant interpretation.5 2

Von Wright's account of "understanding explanations" ('verstehende Erklärungen') is developed in two papers in 1985 and further developed in a paper in 1989. The papers from 1985 will be cited as von Wright (1985b) and (1985a). It is the paper from 1989 that I will emphasize in my account of "understanding explanations." I should add that the term "understanding explanation" is also used in von Wright's Tanner Lecture (1985 TL, p. 136). For discussion see von Wright (1985 TL), esp. part I, sects. 7-16, pp. 128-147. Von Wright (1985 TL), shortened, with all but sect. 8 of those sections omitted, and von Wright (1985a) and (1989) are reprinted (in German) in von Wright's 1994 book Normen, Werte und Handlungen [Norms, Values and Actions], in sect. Ill, pp. 141-255. 3 In his book Explanation and Understanding, von Wright (1971). 4 1 quote from a letter he sent me, dated 29 August 1990 [hereafter: L1990]. See also von Wright (1985a), p. 5; (1989), pp. 14-15. 5 The idea of the agent's situation cum motivation as a "demand" or "challenge" to which an agent then responds by acting is developed most fully in von Wright (1985a); see sects. 2-4. Much the same ground is covered in von Wright (1988), at pp. 275-277,

92

Rex Martin

Von Wright appears to think that, of these two distinctive types of understanding, the response approach is the basic one.6 It gets closer to how most actions happen, how they get started or come about in the first place. For not only deeds performed but also goals themselves (that is, the end states that are envisioned, in a goal or purpose, as something to be brought about or established) are typically responses to situational demands. Now, let us quickly complete the picture here. We will have a full "understanding explanation," then, when we go back to the first of von Wright's sub-schemas and pick up a crucial detail. Here the investigator identifies and relies on the agent's belief that the particular action performed was in fact a means to the end in view and interpolates that belief into the explanatory account, thereby filling the remaining gap in it.

3

Collingwood on Understanding

At this point it would prove useful to turn to Collingwood. A convenient way, indeed perhaps the best way, of understanding Collingwood's idea of the explanation of action is to regard it as building on the very same schema we used in the case of von Wright. Here Collingwood laid special stress on judgements of intelligibility or plausibility. If we were to put the matter in the way Collingwood put it, we would say that these judgements of intelligible connection allow us, once we have in mind a particular situational motivation and a particular purpose of the agent, to re-enact the agent's action. For we can see, with these points in mind and in the light of available evidence, that one of the courses of action - the deed actually performed - makes sense in the situation envisioned and its being done is plausible. Thus, we can successfully get to the deed performed, by citing thoughts and beliefs that the agent had, and in that sense re-enact it (in imagination). For Collingwood, then, what must underwrite the claim that a given action A is a response to a demand is the intelligibility of that action in that role; likewise, what must underwrite the similar claim that a given action Λ is a means to an end, to a purpose or end in view of the agent, is the intelligibility of that action in that particular role. Thus, what underwrites these interpretative claims, in his view, is the same sort of thing in each case. There's no bifurcation here at all. 280-283, 286-287 in particular; see also von Wright (1989), pp. 15-17, 24, and (1985 TL), p. 129. For the point about simply understanding or "seeing," see von Wright (1988), p. 277; (1989), pp. 25-26. 6 See von Wright (1989), p. 17.

Action Explanations as 'Understanding' Explanations

93

A fully successful (or fully satisfactory) intentionalist explanation of an individual action, in this Collingwoodian conception, is a special sort of exemplification of the standard schema mentioned earlier (and sketched out in sect. 1). Here the factual filler which provides the stuff of any given intentionalist explanation should not only instantiate one or another of the conditions of the schema (in an instantiation well supported by available evidence) but should do so in an intelligible or plausible way. This is provided for when three main points in the schema - that is, the agent's situation cum motivation, the agent's relevant purpose, and the deed performed - are satisfied by facts which are themselves intelligibly connected in the specific relationships they have with one another as, respectively, (a) a plausible thing to do in a situation, (b) a situationally responsive end in view, (c) an action that serves understandably as a means to that end (or as part of accomplishing it). 7 Thus, to give an example: Caesar was faced with a lot of trouble from the British tribes (for they were engaged in raids and were causing unsettlement in the Gaelic world) and he wanted to put an end to this trouble, in order to facilitate his conquest of Gaul, so he invaded Britain to carry out an expedition against the tribes, hoping thereby to pacify them. The points identified here are connected with each other in the ways just prescribed. The end in view (the goal of pacifying Britain) is connected with Caesar's original motivating perception of the situation (as a troublesome unsettlement in the Gaelic world, stemming from the British incursions) in that it represents a responsive way of resolving that particular situation. And the action (Caesar's invasion) is an understandable means to the end in view, or part of accomplishing it. Finally, invading Britain is a plausible thing to do in view of the British incursions, a main feature of the unsettlement which Caesar hoped to quell. In making coherence or intelligibility the central issue - as Collingwood has here, in the relations cited in points (a), (b), and (c) above - one is saying something over and beyond what the agent thinks (believes or intends) in the matter and, thus, over and beyond what the evidence might support as to the truth about agent beliefs and so on. For one is saying - to cite one possible case - not so much that the agent's action (e.g., Caesar's invasion) was intended to be part of, or a way of accomplishing, the agent's particular end but that it was intelligible to us, and presumably to any other serious inquirer, in that role. And the same could be said for the other intelligibility relations cited above. 7

I am here summarizing what I take to be Collingwood's account of understanding or intelligibility in action explanations. I have based my view of Collingwood here, principally, on what he says in his 1946 book The Idea of History, in part V, chs. 1-5.

94

Rex Martin

A useful parallel might be noted here. An attempt at explanation which used false statements might count as schematically sound, but it would be clearly unsatisfactory or inadequate as an explanation. By the same token, we would have an explanation of sorts (where the relevant conditions were factually filled) even if the facts cited were not intelligibly connected in the specific relationships they have with one another along the various lines cited in (a), (b), and (c) above. Imagine here an action that does not serve understandably as a means to such-and-so end (or as part of accomplishing it).8 Such an explanation would only be schematically sound; it would not be fully adequate or satisfying. For it would fall below a certain standard for "understanding explanations," that given by having intelligible connections at all the appropriate points.

4

Von Wright on Failures of Intelligibility or Understanding

Now, let us consider in greater detail the case where the called-for "fit" or "coherence" was not present. For example, we might look at the explanation offered by an ancient Aztec priest for their practice of human sacrifice: the explanation that such ritual slaughter helped slow the decline of the universe and thus kept the present age (or eon) in existence. This, I surmise, is a case where deed and reason do not match up, where their conjunction in a means/end relationship, for example, seems at least to us "crazy, unintelligible, irrational."9 Clearly, von Wright allows for such cases. I mean cases where no consensus may exist, between agents and investigators, as to why the agent did the act in question; more particularly, cases where agents and investigators differed, not on the facts of the matter, but on whether certain facts, juxtaposed in certain specific relationships, are plausible or even intelligible in those relationships.10 As best I can make out, von Wright proposes to resolve these interesting non-consensus cases by turning to the idea that what we're after (in 8

Consider, for example, a peasant who, having inflicted a knife wound (accidentally) on his own leg, proceeds to clean the knife meticulously while leaving the wound itself totally unattended. Here the facts, as described, are not intelligibly related to one another in the relationship of means (cleaning the knife)/end (getting the wound to heal) - even though we could affirm that they apparently stand in that relationship (on the basis, for example, of what the agent reports). For discussion, see my book Historical Explanation (1977), at pp. 88-89. Another example, of failure of intelligible connection, is offered in the next paragraph of the present paper. 9 The quoted phrase is found in L1990, p. 2. 10 Von Wright makes precisely such a point in the paragraph that continues from (1989), p. 35 onto p. 36. See also von Wright (1985b), p. 133, and (1985a), p. 10.

Action Explanations as 'Understanding' Explanations

95

an explanation) is the efficacious reason,11 the reason that actually makes the action occur. Several things are involved, I think, in this emphasis on efficacious reasons. (i)

That reason, whatever it might be, that was actually efficacious in the performance of an action is always to be chosen for purposes of explaining the action.

(ii)

Von Wright thinks efficacy can be modelled in two distinct ways (in accordance with his two sub-schemas). (a) In the one case, the efficacious reason is the agent's belief that a certain action is a way to accomplish such-and-so end; the belief here is efficacious, for, given the belief, the agent will do that thing to achieve the end intended, (b) In the other case, we try to have in view (on the basis of evidence) the agent's overall motivational background: the set of existing reasons for or against a given action. That reason (or balance of reasons) that actually moved the agent to perform the action is here the efficacious one.

(iii) In any event, it is the agent's actual way of seeing or comprehending (Verstehen) the action, and the reasons for it, that counts. For only this understanding allows "entry into the subjectivity involved with the act of understanding"12 and hence, through that particular understanding (and only through that one), with the efficacious reason for the agent's deed. (iv) Thus, in the end, von Wright's account (especially in the sub-schema concerned with means/end beliefs) is tied to the perspective of agent's understanding. (v)

Agent's understanding, as here characterized, is not the same as what the agent reports. For the agent can be lying or mistaken or even merely insufficiently self-conscious. The point, rather, is that to find the efficacious reason, the agent and the investigator always take one and the same perspective: they look to the agent's relevant means/end beliefs, in the one sub-schema; they look to the agent's existing motivational background, in the other.

(vi) Von Wright's principal rationale for taking up this particular perspective is that it allows us more or less successfully to duplicate, at certain1 crucial points, the features of what might be called the standpoint of agency - of performative understanding. "Von Wright (1989), p. 25. See also von Wright (1985a), pp. 9, 12, and (1985 TL), pp. 135, 137. 12 Von Wright (1989), pp. 29-30. See also von Wright (1985b), p. 134.

96

Rex Martin 5

The Truth About Efficacious Reasons

But how does one know that the reason selected as efficacious is the one that actually figured in the agent's deed? Von Wright's answer here involves two claims. The first is that we use evidence of a certain sort: consideration is given to what the agent has said and done in the past (or has recently said and done, more or less at the time of acting), and consideration can be given to relevant things the agent says or does in the future. It is on the basis of evidence of this sort that one selects, from the agent's existing motivational background, that one reason which actually moved the agent in the case at hand. To this von Wright adds, as a way of achieving explanatory closure, the agent's relevant means/end belief (as attested to by evidence, of course). But again one asks, how do we know this: how do we know that this is the reason that moved the agent? that this belief brought about the action in question? We come, then, to von Wright's second claim. There is, he says, no further fact that one can adduce to answer these questions. Rather, if the data used has been assembled as completely and carefully as possible on a basis of considering extensive relevant evidence and if, in so assembling, one can then achieve a continuity between deed and thought, a coherence or fit between them (of the sorts identified in the two sub-schemas), that's all there is to it. It's simply that we do understand the deed as done from these reasons. And this understanding (with the firm evidential base on which it rests kept in full view) is what warrants the claim that the agent actually acted for these reasons.13 It's not so much that efficacy claims are true (for this suggests some pre-existing matter of fact that we could find, a real connection out there somewhere) but, rather, that they are warranted - by understanding. Here it is not the case that understanding is itself certified for use by its attachment to efficacy (where efficacy is a matter that can be independently established - by neurology, for example). Rather, it's the very opposite: here claims to efficacy are themselves warranted, in von Wright's view, by the sheer fact 13

For the main argument here see von Wright (1985a), sect. 7. Von Wright describes his account of understanding explanations (in 1989, p. 24; 1985b, p. 135, and elsewhere) as a coherence theory of understanding (Kohärenztheorie des Verstehens). What he particularly had in mind, I surmise, is that there is a "fit" or "matching," a coherence, if you will, between deed and reason (von Wright 1989, p. 25). Thus, he says (using this time the notion of a "consensus in the understanding" to describe a coherence of agreed-upon facts): "The 'subjectivity of understanding' does not make the explanations arbitrary. Every attempted explanation ought to respect the facts of the case: that there existed such and such reasons for, and perhaps also against, the action and that an action which matches the proposed description actually took place. But when these facts are established, agreed upon, there is no further fact in addition to the consensus in the understanding which establishes that the action took place for that reason and not for that other one. Finis" (in a letter to me of 30 July 1993 [hereafter: L1993], pp. 1-2).

Action Explanations as 'Understanding' Explanations

97

of understanding (as spelled out in his two sub-schemas) - and the base of evidence on which it rests. This is the only support these claims have, or can have.14 We have no guarantee, in understanding so conceived, that investigator and agent will always agree in given cases on the reasons for deeds (and this brings us back to the non-consensus case from which we started). Most often they will agree, however, and that's an end on it. But sometimes they won't. Here another party attempts to overrule the agent's own account: the onlooker says, "No, you didn't act for the reason you profess but for another; you acted for this reason, as the evidence of your overall behavior suggests." Perhaps an impasse results. The one can then try to "convert" the other, at the time or later on (even in the history books), but this may not avail.15 The point is, neither the agent nor the investigator is the court of last appeal, or the highest authority; neither has nor can have a uniquely privileged and infallible access to the truth, to the correct understanding. Rather, in the end, we must rely simply on understanding (and the firm evidential base on which it rests), rely, that is, on an understanding of the sort that is typically but not always shared between agents and investigators, as the warrant for efficacy claims. These remarks do not solve the problem of non-consensus; they simply locate the problem, indicating how it might be resolved. They tell us where to look - to the agent's existing motivational background and stock of means/end beliefs - and they suggest the sort of resolution we seek.

6

Collingwood's Alternative View

At this point I think we've arrived at that which principally serves to separate the views von Wright and Collingwood have taken on understanding. Von Wright is concerned with efficacious reasons; and because he is, he is committed to the perspective of agent's understanding - to the perspective of the agent's existing motivational background and relevant means/end beliefs). If challenged, he would likely say: but it is the way the agent actually 14

For further discussion and for citations, see the Appendix. Von Wright regards such cases of disagreement as possible ones, but not as typical: "Usually the agent himself relates his action to the same reason for doing it as most outside observers of his conduct. But sometimes there is a disagreement - and neither the agent nor the observer can be 'converted' to see (understand, explain) the action in the same way. Sometimes the observers are unanimous and the agent is 'judged,' praised or blamed, sentenced or acquitted on the basis of their understanding of him" (L1993, p. 1). See also von Wright (1985b), pp. 133-134; (1985a), sects. 8-10; (1989), pp. 28-29; (1985 TL), pp. 134, 139-141. 15

98

Rex Martin

understands things, the agent's efficacious reason (as exhibited, for example, in the agent's means/end belief), that brings about the deed. Collingwood, on the other hand, was more concerned with explanatory reasons; and, because he was, he was committed to the view that the reasons used in an explanation must be intelligible to the investigators (to those who are in the explanatory mode, as givers or as receivers of explanations).16 Explanations that don't exhibit intelligible connections in this regard (that are not re-enactible by investigators) are not satisfactory as explanations of action. If challenged, Collingwood or, rather, the defender of Collingwood might say something like the following: performing an action and explaining it are two different things. A person can perform an action even where intelligibility of connection fails, for crazy people do act; and actions can be done that we don't understand (by people in another culture, say). What we can't do, however, is have a satisfactory explanation of action where intelligibility of connection fails for the explainers. For where re-enactibility is not exhibited or not present, an explanation is simply not intelligible. Though it may be an explanation of sorts (in fulfilling the terms of the schema, say) it nonetheless falls below a certain standard. For a proper explanation of an action should yield intelligibility. We want explanations which are satisfactory by that standard. One main difference between von Wright's position and Collingwood's, then, is the perspective each takes up. In von Wright's case (certainly in the means/end sub-schema, and in the other sub-schema as well), it is the perspective of the agent and the efficacious reason for the agent's action. In Collingwood's, it is the perspective of persons in an explanatory mode. Significantly, Collingwood described such persons as engaged in re-enactment. For re-enactment (note the name) is something investigators attempt to achieve; it is something they achieve when they use intelligibly connected reasons in an effective way, so as to afford plausibility in an explanation. Neither perspective here identified is the correct one (just as neither is wrong); they are compatible, in fact. It would be well if they could be brought together more completely, then. It may prove, though, that the various strategies we employ do not readily lead to an consensus between agents and investigators. What then? The situation is not irretrievable on the basis Collingwood has provided. Investigators, in Collingwood's view, are trying to provide an internal understanding, one that tracks and is ultimately faithful to the agent's own thoughts. Sometimes, though, the agent's thoughts may have to be redescribed (perhaps extensively so) to achieve re-enactibility. And often, 16

By intelligible I mean, of course, what I said in the paragraph in sect. 3 to which note superscript 6 is appended.

Action Explanations as 'Understanding' Explanations

99

even then, investigators may need to learn to think in new ways in order to track the thoughts of agents. But in no case does the investigator desert the perspective of investigator. To try to 'become' the agent would be pointless, for then one would simply be acting (and not even trying to explain). Nor does the investigator attempt to 'duplicate' the agent's thought; for if the agent's thought is already explanatorily opaque (as, by hypothesis, it is here, as in the Aztec or the knife/wound example), then reproducing it accurately will not result in something intelligible to the investigator, something the investigator can re-enact. In the end, the thought of the agent and that of the investigator are separate and may even be distinctive; what a successful re-enactment can achieve, in those cases that remain irreducibly disparate, is a "fusion of horizons" (in Gadamer's phrase). My point here is that Collingwood's theory is open to the sort of resolution I've just been describing.17 But von Wright's really is not - that is, it is not if we insist, as he has done, on the primacy of efficacious reasons, as determined from the agent's perspective, in non-consensus cases.

7

A Summary and Conclusion

My assessment of the main differences between Collingwood and von Wright can now be summarized briefly. First, von Wright bifurcated the notion of understanding, Collingwood (by contrast) attempted to give a unified account. Second, Collingwood's account of understanding was different from von Wright's (most clearly so from the account where von Wright emphasized the agent's means/end belief as the key to one of his two kinds of understanding). Collingwood would, I think, regard von Wright's emphasis here as simply inappropriate - in that using the agent's means/end belief might not yield intelligibility to anyone (not even to the agent, upon reflection). Finally, the perspective each occupied was different: in von Wright's case it is the perspective of the agent and the efficacious reason for the agent's performance of an action, as seen from that perspective; in Collingwood's, it is the perspective of persons in an explanatory mode who are thereby using re-enactible (that is, intelligibly connected) reasons. These differences are crucial, especially the latter two. And we have or may have good reasons to prefer Collingwood's stance on both these points. 17

What we would need, to supplement Collingwood's theory, then, is a further theory of how actions (in other times or other cultures) might be understood (or, better, come to be understood) even in non-consensus cases. See here my paper (1991c). The Aztec example is discussed there, at some length. For further elaboration of the point made in the first paragraph of this note, see esp. Martin (1991c), pp. 362-363 and n. 8 on p. 365.

100

Rex Martin

Thus, the first difference (the unified versus the bifurcated account) could be resolved as well, along the lines Collingwood has suggested. Von Wright's bifurcation of ways of understanding is really quite arbitrary, in my judgement. It's simply not necessary to think that the means/end relationship is always and necessarily mediated by an explicit belief of the agent. That relationship can be established in exactly the same way as was a response to a demand: we can straightforwardly see or interpret a given action as a means to an end, without the mediation of a distinctive means/end belief on the part of the agent. The problem runs deeper than this, though. The problem is that von Wright's account of understanding, as modelled in the two sub-schemas, is bivalent. Thus, each of his senses of "understanding" can work against the other and (in so doing) can threaten to take over the domain that the other sense ostensibly controls. It might be the case, for example, that the kind of intelligibility identified with the "response" sub-schema could become completely dominant, thereby suppressing the emphasis found in the other sub-schema on certain explicit beliefs of the agent, while at the same time undermining reliance on the sort of understanding such beliefs can yield. Conversely, the other or means/end belief schema might come to predominate totally, thus requiring reliance on mediating beliefs at other specific points as well (for example, such beliefs as the more or less selfconscious one, on the part of the agent, that a certain end in view is a suitable response to a given situational "demand"). And that change would underwrite a wholesale deployment of the notion of understanding characteristic of the means/end belief schema and, with it, the abandonment of the kind of intelligibility identified in the "response" sub-schema. The point is that the approach in each of these sub-schemas is equally eligible in von Wright's account, so the tension I've described can never be tamped down complete. The bifurcation von Wright suggests, then, is both arbitrary and unstable. But this particular tension, I've already noted, is not present in Collingwood's account and would not be a problem there. Von Wright's work, like Collingwood's, is not particularly easy to grasp in the area we have been examining: the area of understanding (or intelligibility) in action explanations. I have, accordingly, sought to locate the thought of each on this matter more precisely by putting it in close relation to that of the other. Letting their ideas play against one another, in this way, should make the views of each of these thinkers clearer to us.18 18

In the present paper I have drawn on earlier writings of my own: on Martin (1991a), on (1993), and on (1994). There is, of course, some overlap in content between the present paper and these other three. The present paper is a shortened version, drawing on sect. 1 only, of "Von Wright and Collingwood on Causation and the Explanation of Human Action," in Contemporary Action Theory, Vol. I: General Philosophy of Action, Raimo

Action Explanations as 'Understanding' Explanations

101

Appendix to Footnote 14 Thus, "That which links an action with one or more reasons [Gründen] is simply the fact that we see or understand [verstehen] the action as having arisen from these reasons. We see the action 'in the light of particular reasons - and it is in this [fact of] seeing or understanding that the 'effectiveness' [or 'efficacy'] of these reasons lies" (von Wright 1989, pp. 25-26). See also von Wright (1985d), p. 130; (1985a), p. 12. And again: "[T]he position was taken in the moment of recognition [from context: in the moment of the actor's acknowledgement of it], and it is not until this moment that the connection of reason and action is created" (von Wright 1989, p. 29). See also von Wright (1989), p. 27; (1985(1), pp. 132, 134; (1985a), pp. 12, 16; (1985 TL), pp. 143-144. "What I am saying amounts to this: The effectiveness/efficacy of these reasons or motives [Beweggründe] cannot be separated from the recognition and acknowledegment of them in the process of taking a comprehending position [verstehende Stellungnahme]. The truth of an explanation of action is created together with and is identical to the position taken to the reasons for acting. This position is not 'complete' or 'closed' [fertig'] until it is recognized [that is, until this sort of recognition has occurred]" (von Wright 1989, p. 29). See also von Wright (1985d), pp. 132-134; (1985a), pp. 18-19; (1985 TL), p. 137. But von Wright notes, "I'm afraid many would consider my position much too 'subjectivist' when it makes understanding itself the criterion of truth of explanations" (L1990, p. 2). See also von Wright (1985d), p. 130; (1985a), p. 12; (1989), p. 26; (1985 TL), p. 137. In a subsequent letter to me (L1993) von Wright returned to the same theme with the following observation: "There is an aspect of my idea about 'understanding explanations' . . . which to me seems crucial - though probably very hard to digest for most explanation theorists. It is this: One cannot separate the truth of the explanation from the very act of understanding itself. The 'efficacious' reasons for an action are those in the light of which Tuomela and Ghita Holmström-Hintikka (eds.), Dordrecht, The Netherlands: Kluwer, (1997). I want to thank the editors for allowing publication of the present papers in this collection of essays, ed. by Georg Meggle. Further citations to von Wright's writings and supporting arguments to points made in the present paper can be found in my article, Martin (1990). A discussion of one important necessary condition of intelligibility (an ontic condition) can be found in my paper, Martin (1991b). I want to thank audiences in Germany, England, Canada, and Wales for their comments on earlier versions of the present paper. And for providing me with serviceable translations from German and Italian texts, I want to thank James Gilkeson and Mirella Vaglio respectively.

102

Rex Martin

the action is understood or explained." See also von Wright (1985a), p. 12; (1985 TL), p. 143.

References

BIANCO (1985). Franco Bianco: Dilthey e il Pensiero del Novecento [Dilthey and the Thought of the Nineteenth Century]. Milan 1985. COLLINGWOOD (1946). R. G. Collingwood: The Idea of History. New York 1956; first publ. in 1946. McMURRlN (1985). Sterling M. McMurrin (ed). The Tanner Lectures in Human Values, vol. VI. Cambridge, England 1985. MARTIN (1977). Rex Martin: Historical Explanation. Ithaca, NY 1977. MARTIN (1990). Rex Martin: G. H. von Wright on Explanation and Understanding: An Appraisal. History and Theory 29 (1990). MARTIN (1991a). Rex Martin: Collingwood on Reasons, Causes, and the Explanation of Action. International Studies in Philosophy 33 (1991). MARTIN (1991b). Rex Martin: Intelligibility. Monist 74 (1991). MARTIN (1991c). Rex Martin: The Problem of Other Cultures and Other Periods in Action Explanations. Philosophy of the Social Sciences 21 (1991). MARTIN (1993). Rex Martin: On G. H. von Wright's Theory of Practical Inference. Archiv für Rechts- und Sozialphilosophie Beiheft 51 (1993). MARTIN (1994). Rex Martin: Collingwood and von Wright on 'Verstehen', Causation and the Explanation of Human Action. Collingwood Studies 1 (1994). VON WRIGHT (1971). Georg Henrik von Wright: Explanation and Understanding. Ithaca, NY 1971. VON WRIGHT (1985a). Georg Henrik von Wright: Probleme des Erklärens und Verstehens von Handlungen [Problems in the Explanation and Understanding of Actions]. Conceptus 19 (1985). VON WRIGHT (1985b). Georg Henrik von Wright: Sulla Veritä Delle 'spiegazioni' Comprendenti [On the Truth of 'Understanding' Explanations]. In Bianco (1985). VON WRIGHT (1985 TL). Georg Henrik von Wright: Of Human Freedom. In McMurrin (1985). VON WRIGHT (1988). Georg Henrik von Wright: An Essay on Door-Knocking. Rechtstheorie 19 (1988). VON WRIGHT (1989). Georg Henrik von Wright: Das Verstehen von Handlungen - Disputation mit Georg Meggle [The Understanding of Actions ...]. Rechtstheorie 20 (1989); reprinted in von Wright (1994). VON WRIGHT (1994). Georg Henrik von Wright: Normen, Werte und Handlungen [Norms, Values and Actions]. Frankfurt am Main 1994.

GEORG MEGGLE Understanding of Actions: Some Problems

Abstract: An action is understood by us iff we see the action as being the conclusion of an appropriate practical syllogism. With this starting point of von Wright's Explanation and Understanding (1971) several other proposals are compared and more or less identified with, namely understanding as (i) knowing of the intention with which the action was done, as (ii) knowing the reasons for which it is or was rational to perform the action, and as (iii) knowing the subjective meaning of the action. Relative to these different versions of "understanding of an action", I summarise some differences between von Wright and myself. 1 2 3 4 5

Which Practical Syllogism? Which Intention? Which Rationality? Which Acting? Is Understanding not Knowing?

In contrast to scientific explanation, understanding applied to actions and their agents (persons) and results (e.g. signs, expressions, languages) was initially regarded in analytic philosophy as an area best avoided in the academic world. The fact that this situation has radically altered is largely due to both Wittgenstein's Philosophical Investigations (as far as rule-based acting is concerned) and Georg Henrik von Wright (in the field of intentional acting). It was in von Wright's book Explanation and Understanding (1971) that he first posed the crucial question clearly and unambiguously: If scientific explanations take the form of deductive-nomological and probabilistic inferences, but this form is inappropriate for understanding actions - what form of inference does understanding have? Von Wright's reply was that, as a special type of teleological explanation, the understanding of actions takes the logical form of Practical Syllogisms. In other words: (U-PS)

An action is understood iff it is regarded as the conclusion of a (suitable) Practical Syllogism.

My own occasional reflections on understanding are mostly also based on this approach, and the days and weeks during which I was able to discuss

104

Georg Meggle

with Georg Henrik our common and yet different understanding of understanding are among my happiest times. However, it seems that even at the best of times a vestige of uncertainty is unavoidable in philosophical questions. There is only one way to further minimise this vestige - to try and make it explicit. And this is what I am doing here, by summing up the differences already developed in more detail elsewhere (see in particular von Wright/Meggle (1989)) along with their genesis.

1

Which Practical Syllogism?

1.1 There are many types of 'Practical Syllogisms'. Let us leave aside the nuances of such syllogisms and restrict ourselves to the following: (PS-I)

(1) (2) Therefore (3)

X wants A. X believes A will occur iff he himself does /. X does /.

Is this a valid syllogism? It ought to be, if Practical Syllogisms are to provide the same substantiation for (the performance of) actions as 'DM' (deductivenomological) explanations for (the occurrence of certain) events. For only such DN explanations are genuine explanations, in which the deduction of the explanans regarding the explanandum is a valid syllogism. However, if (PS-I) is valid, then ditto (PS-II) - and vice versa. (PS-II)

(1) (2) Therefore (3)

X wanted A. X believed A would occur iff he himself did /. X did /.

Let's ask the question again: Are (PS-I) and (PS-II) valid syllogisms? Von Wright's quick reply was No, as there are too many factors which still need to be considered before this question can even be put. To what times do the premises and conclusion specifically refer? Do X's wanting and believing really occur at the same time? Can/could X do / in the first place? Did X believe himself that he could do /? etc. These and other loopholes were discovered and plugged by von Wright in Explanation and Understanding (III.5 and III.6). But even if all the possibilities which could get in the way were to be considered, the conclusion-validity question would still remain unsettled. How did von Wright solve this question? 1.2 This was where my initial difficulties arose. I found von Wright's answer ambivalent. It seemed as if von Wright wanted to answer the validity question for (PS-I) vs. (PS-II) not just separately but even divergently: namely with Yes for (PS-II) - but only iff its conclusion (3) is already known

Understanding of Actions: Some Problems

105

to be correct - and a tendency towards No for (PS-I). Furthermore, he also (later) declared that the usability of a Practical Syllogism (of the above type) for the purpose of understanding an action ultimately does not depend on the validity of the relevant inference (cf. von Wright (1978), p. 279). 1.3 Naturally, this was and still is a mine of material for discussion, albeit probably no longer for us, for this part of the controversy at least ought now to be a thing of the past. Now (I believe) we share the opinion that in order to understand an action neither (PS-I) nor its past-tense form (PS-II) but rather a Practical Syllogism with a different conclusion type is relevant: (PS-III)

(1) (2) Therefore (3)

X wants A. X believes A will occur iff he himself does /. It is/would be rational for X to do /.

The Practical Syllogisms corresponding to the understanding of the action substantiate the respective actions in so far as the premises of these syllogisms (i.e. the 'voluntative-cognitive complex' (1) & (2)) provide sufficient reasons for the rationality of these actions. X's voluntative-cognitive complex results in (as long as, as tacitly assumed, we regard his believing and wanting to be strong believing and wanting) a decision under certainty. (From X's viewpoint, doing / definitely leads to A, i.e. to a result which is best for him, whereas not doing / will definitely not do so.) And for such a situation, the syllogism (PS-III) holds simply by definition. (Those who choose an action which will definitely provide an optimum result are acting rationally.) According to this new approach, whether X actually does or did what would be or would have been rational for him thus depends only on one thing: whether in the given situation X is or was actually rational. If this situation-relative rationality assumption is brought into play, all the other difficulties listed above dissolve into nothing; and (previous) disregard for this assumption also explains how these difficulties arose in the first place.

2

Which Intention ?

2.1 If I do /, I want A, and I believe that I can achieve A only and precisely by doing / (in other words, in addition to (3), conditions (1) & (2) of the above syllogisms also apply to me), it could also be said that I perform the action / with the intention of bringing about A. Von Wright's concept of understanding (U-PS) and the following concept of understanding thus amount to one and the same thing: (U-I)

Understanding an action := Knowing the intention behind it.

106

Georg Meggle

2.2 At any rate, that's what I thought. But von Wright put me right. How? Like this: Whereas from my point of view in both situations (SI) and (S2) I do something with a certain intention (51) I want to get home quickly. Shall I take a taxi? Shall I wait for the bus? "The bus will only come along in 20 minutes; I'll easily be home by that time if I take a taxi," is the thought which goes through my mind, and I hail the next taxi I see. (52) I'm driving close behind a lorry on the motorway. Suddenly the lorry's brake-lights come on. Automatically I put the brakes on too - after all, I don't want to cause an accident. and due to (U-I) I treat both cases as being the same, von Wright shrinks from such egalitarianism. His "intentions" are accompanied by something like conscious reflections; mine are not. Von Wright would not even talk of intentional acting in connection with (S2), as this would clash with the automaticity of my braking reaction. 2.3 Is this just a disagreement about the choice of words? If not, we're immediately in trouble. Is deliberate intention part of acting? And in which of the various senses of "deliberate intention"? (Cf. for example "intention with" vs. "intention in", and "intended" vs. "not unintentional".) What distinguishes acting from mere behaviour - or behaviour from mere reflexes? Luckily, we can keep out of trouble at this point. The central question which needs to be asked here is merely this: Should and do we really only want to discuss whether actions resulting from conscious reflections are understandable? And what argues in favour of and against Yes or No in each case? Naturally, more general systematic viewpoints come into play here. (We will examine one of these below.) Von Wright still seems to tend towards Yes, while I loudly declaim No.

3

Which Rationality?

3.1 The interim result of 1.3 above can also be expressed more generally as follows: An action is understood by those who know and know why it is or was rational for the agent to perform it. In short: (U-R)

Understanding an action := Knowing the reasons for its rationality.

Although the premises of a Practical Syllogism provide (as we can see in (PS-III)) such reasons, they only do so unrestrictedly for the special case of a decision under certainty, and perhaps also (albeit with some limitations) for a few other special cases of a decision under risk - in both cases, however,

Understanding of Actions: Some Problems

107

only because the decision situation can be formulated merely by means of the qualitative concepts X wants A and X believes that A. But what about the more general case in which we can only take more or less strong (i.e. comparatively or even metrically formulated) preferences or probability assumptions as our starting-point? In this case, Practical Syllogisms of the above type cannot be used for the actions characterised as rational (according to the Bayesian criterion of rationality, i.e. by providing a maximum expectation of benefit). Should these actions be regarded as incomprehensible merely for this reason? This would be absurd. The understanding approach referring back to Practical Syllogisms is thus only usable for one special area (namely that of means-end rationality). A more general approach ought to begin by determining the concept of understanding in Bayesian rationality in general instead of with Practical Syllogisation. And as (U-PS) and (U-I) mean the same or (if this is too strong) are at least closely linked, the same also holds for the understanding approach referring back to intentions. The more general approach is (U-R); like (U-PS), (U-I) is also at best a specification thereof. 3.2 What's the state of play now with our above (see 2) disagreement concerning situations (Si) and (S2)? If we let rationality take precedence, the old question of whether (U-I) is applicable to both situations or just (SI) appears less important. The more explosive question is now whether our rationality approach (U-R) is suitable for both. Our disagreement on this point is unfortunately stronger than our joint plea for rationality. Von Wright continues to deny the predicate of rationality for my behaviour in (S2), whereas I see no problem in explaining my fortunately quasi-automatic braking reaction as rational. And of course, I'm not alone: at this point too I'm merely acting as the spokesman of all those decision theorists who have no problem at all in for example explaining even unconscious consumer behaviour in terms of decision and thus rationality theory. Von Wright's minority judgement may nevertheless be the more convincing of the two. But why? Because rationality considerations are only relevant for actions - but (S2) doesn't describe an action? If so, while happily subscribing to this because-sentence, I would dispute the but-premise. I didn't have to put on the brakes; I could have refrained from doing so. And I would like to regard behaviour which I could have refrained from doing as an action. 3.3 Perhaps this disagreement harbours an even more general one: belief assumptions usually do not need (including - but not only - within the context of decision theory) to be explicit assumptions (staring the decider in the face as he makes his decision); they can also be implicit (and (S2) provides

108

Georg Meggle

a perfect example). There's no doubt that the vast majority of assumptions which determine our behaviour are merely implicit. (For example, I've spent the whole day working on this paper. And although I definitely wouldn't have done so if I'd not assumed that everything was okay at home, until just now the thought "Everything's okay at home" hadn't explicitly gone through my mind at any time today.) By contrast, von Wright appears (at least in the context of his minority judgement as far as rationality is concerned) to have exclusively aligned himself behind the first alternative. Is this the case? And if so, why?

4

Which Acting?

4-1 The rationality grounds which we need to know in order to understand an act by the person X are exclusively composed of probability assumptions and the preferences of X himself. If we know these reasons, we know the function and the sense which the action concerned has for the agent himself from his viewpoint. We thus know, as can also be said with Max Weber, the subjective sense of the action. It's our rationality approach again merely expressed in different words: (U-MW)

Understanding an action := Knowing the subjective sense of the action.

This formulation makes it especially clear that everything so far has revolved around the understanding of concrete actions. Behaviours and their sense and understanding are ignored. 4-2 Some forms of action have what we call an 'intersubjective meaning' relative to a group. And sometimes we have to know this kind of meaning in order to be able to recognise the subjective sense of the action being performed. To use von Wright's favourite example, you have to know that tipping your hat is regarded in some areas as a gesture of greeting if you are to correctly understand X's concrete hat movement as a gesture of greeting on his part. 4-3 Although von Wright drew attention early on to such a 'sense priority' of intersubjective sense, I believe that he somewhat exaggerated the gap between the corresponding cases of understanding. That I correctly understand someone's behaviour as a greeting may indeed imply that I know that behaviour of the type concerned (within the reference group) intersubjectively has the meaning of a greeting, but greeting still includes the desire to greet - and so understanding in the sense of (U-MW) continues to remain relevant for these actions too. Occasionally, von Wright's remarks appear to

Understanding of Actions: Some Problems

109

represent the thesis that for the (performance of) actions with an intersubjective meaning, the concept of understanding actually tailored to concrete actions becomes irrelevant. Am I mistaken? 4-4 By the way, the above sense priority of the intersubjective does not exclude the possibility that "intersubjective meaning" itself (at least for some actions) can be explicated by means of the concept of subjective sense. For example, elementary concepts of the conventional communicative meaning of actions and thus also of linguistic expressions (as the products of such acting) can be explained in this way. Does von Wright believe this special application of the programme of methodological individualism to be unsuccessful right from the start? Although I would of course be very reluctant to want to believe this, his own understanding approach has, I believe, precisely in this field proved to be extremely fruitful.

5

Is Understanding not Knowing?

The area which gives me the most problems is the coherentistic change which has emerged with increasing clarity over the past few years, or to be more accurate with the consequence of this change which undermines the foundations of the following distinctions. We have to distinguish between understanding and believing to understand in the same manner as we generally do between knowing and merely believing to know (i.e. strong belief). Just as knowing implies the correctness of what's believed, understanding also implies the correctness of what's believed to be understood. And this is exactly what distinguishes a case of understanding (as knowing) from the more general case of understanding something as something (i.e. believing, regarding, considering something as something). Something can be correctly or incorrectly understood as something; by contrast understanding is always correct - or it's not understanding. ('Correctly understanding' is thus just saying the same thing twice over.) Von Wright evades these distinctions with his thesis that understanding need not be knowing. I simply cannot understand this, or his reasons. Or is it more to the point that I simply don't want to understand it? If that's the case, I need someone to give me the motives to change my attitude accordingly.

This was just a brief rundown of our differences. Naturally, as already stressed at the start, our common ground is much larger and more significant.

110

Georg Meggle

And of course behind this summary is something like another dream of new, happy days. References APELETAL. (1978). Karl Otto Apel, JuhaManninen and RaimoTuomela(eds.), Neue Versuche über Erklären und Verstehen. Frankfurt a.M. 1978. MEGGLE (1978). Georg Meggle: Eine Handlung verstehen. In Apel et al. (1978). MEGGLE (1990). Georg Meggle: Eine konventionelle Handlung verstehen. Rechtstheorie 21/4 (1990). VON WRIGHT (1971). Georg Henrik von Wright: Explanation and Understanding. Ithaca 1971. VON WRIGHT (1978). Georg Henrik von Wright: Erwiderungen. In Apel et al. (1978). VON WRIGHT/MEGGLE (1989). Georg Henrik von Wright und Georg Meggle: Das Verstehen von Handlungen (Münsteraner Disputation). Rechtstheorie 20, 1989; reprinted in von Wright (1994). VON WRIGHT (1994). Georg Henrik von Wright: Normen, Werte und Handlungen. Frankfurt a.M. 1994.

JEAN-LUC PETIT

The Neurological Correlates of Action Types Abstract: Taking seriously the recent cognitive development in science forces us to clash head-on with Wittgenstein's autonomy of intentional language argument in order to recognize the right of neuroscience to deal not only with bodily movements, but also with intentional actions. We discuss the chances of finding a way out of this dilemma in von Wright's action theory. 1 2 3

The Cognitive Trend in Neuroscience, a Challenge for Philosophy of Mind Wittgenstein's Argument in Favour of the Autonomy of Action in Relation to Neurophysiology Does von Wright's Two-Level Stance Provide an Alternative?

1

The Cognitive Trend in Neuroscience, a Challenge for Philosophy of Mind

What seems largely admitted nowadays is that neuroscience has become cognitive. By qualifying it as "cognitive", one does not mean, contrary to a regrettable misunderstanding, that it has occurred recently to this neuroscience to take any particular interest in the internal mental states representative of human cognition, or intellectual knowledge, as if it were returning to the straight and narrow after decades of behaviourist distraction. By qualifying recent neuroscience as "cognitive", it is meant primarily that, without leaving its proprietary ground as an experimental, natural science, it has acquired the self-confidence in its epistemological right, not only the technological, instrumental capability to deal with the so called high level, secondary, associative, supra-modal or mental functions of the brain. If one prefers, it is no less than those long-lasting mysteries of the functioning of the human mind that stubbornly defied all traditional philosophers, that progress in neuroscientific investigations seems to be promised to successfully see through by casting its powerful new light on them. Or, discarding unnecessarily inflated language, among the psychological functions of the human mind which philosophers from Plato and Aristotle on (with exceptions, but very few and late) have made their special professional task of describing and understanding, none could be henceforward considered as being out of the range of neuroscientific research.

112

Jean-Luc Petit

At least, there are two facts that even a sweeping view at the chapter titles of any authoritative collection of articles in neuroscience1 tend to enhance. That contemporary neuroscience has decisively freed itself from its former confinement in the short-sighted study of anatomical structures as localizations of the brain functions and of low level motor reflex responses to the simple stimuli of artificially impoverished environments. And that contemporary neuroscience is now considering its right to encroach upon the human mind, without letting itself be intimidated by the latter being a traditional philosopher's territory. Alternatively, if one goes through the popular scientific literature, one can't help but acquire the conviction that "looking inside the human brain" (the title of a recent issue of La Recherche2), however inadequate may be this crude metaphor to the specific dealings of neuroscientists, seems to be mainly considered in the laboratories as in laymen's opinion as a fully legitimate, if not the only possible entry to the mind. Surely, what could appear, from a certain viewpoint, as an encroachment of neuroscience on philosophy's domain would be revealed as not actually a seriously offending one, if a changing of this point of view could make it appear more challenging than threatening. But, realizing that the mere attempt at philosophical thinking cannot but inscribe us into a tradition, it almost would be contradictory to imagine that we could have an unprejudiced view on the human mind and decide uncompromisingly of the best way of approaching it. By viewing mental acts, states or attitudes as we view them, we are under the influence, explicit or implicit, of at least some thinkers whose style of thought turned itself into a paradigm for every attempt of ours at "starting afresh" to think about that kind of subject matter. These background influences command the openness or closedness of new avenues of thinking in front of us. So, we'd better look straightforwardly and unswervingly at those conceptual constraints or obstacles that cannot be so easily bypassed, and that we have as responsible philosophers to confront honestly in the perspective of our own tradition of thought, in view of a meeting of philosophy and neuroscience on the human mind, as a possible common ground of their respective enquiries. Thus, if what our present query is about is the possibility of throwing a bridge between a genuine philosophy of mind and the new cognitive neuroscience, we cannot be held but at bay by the great figure of Wittgenstein. He founded its denial of the very possibility of such a bridge on the fact that the language of mind, that it is up to philosophy to analyse, is grammatically J

Cf. "Vision", "Representation", "Audition", "Attention", "Memory", "Reasoning", "Language", in Kosslyn/Andersen (1992); "Strategies and planning", "Attention", "Memory", "Language", "Thought and imagery", "Emotion", "Consciousness", in Gazzaniga (1995). 2 n° 289 Juillet-Aoüt 1996.

The Neurological Correlates of Action Types

113

autonomous in relation to the cerebral or mental events and causal processes objects of the empirical research of neuroscience. As an alternative to this stance, I would like to oppose the one taken up by von Wright, who tried to be equally fair to the demands of scientific explanation and to the demands of humanistic understanding.

2

Wittgenstein's Argument in Favour of the Autonomy of Action in Relation to Neurophysiology

Roughly, we could distinguish two modalities of language use, one that could be dubbed "extensional" and one that could be dubbed "intensional", or "non-extensional". We use language extensionally when we have to spot and describe some object of interest into some extra-linguistic domain, mainly an area of physical space. We use language intensionally when we ourselves move, as it is the case in everyday conversation, amongst expressions of attitudes, dispositions, or intentions that relate us essentially to an interlocutor, actual or virtual, personal or collective, and only contingently to the objects, facts or situations of extra-linguistic reality. The grammatical devices of the extensional use of language are those of referring to an object by its name and stating a proposition by giving its truth-grounds (i.e. the objective conditions that obtain if, and only if, it is true). Alternatively, the grammatical devices of the intensional language use are those linking the contents of an enunciation to its pragmatic context (the speaker, the place, the date, etc.) and enclosing into the context of some propositional attitude locution any proposition stating the facts, in order to turn the latter statement dependent upon the subjective or cultural point of view of the one who makes it. For example, keeping to extensionality, I can call anybody in the room by his/her name, I can describe in whatever detail the scene in front of me in terms of the objects that fill the room, their properties and relations; but I leave extensionality and enter the intensional language mode as soon as I express any psychological attitude, be it mine or others's, about the latter scene or any of its components. The transition may not seem dramatic, but it actually is considerable. To realize that, suffice it that whatever you say in the intensional mode straightaway ceases to depend on objective reality to be confirmed as true or refuted as false, as if it were screened all around from the conditions that reign outside the perspective on the world of the speaker's linguistic community, since its private point of view escapes anyway from language. What urged Wittgenstein to enter into the philosophy of psychology was that the psychologists and psychophysiologists of his time, in their attempt to start a new discipline modelled on physics, seemed to him to be unaware

114

Jean-Luc Petit

of this distinction between the extensional and the intensional. Illuded by a tendency to treat uniformly all language uses on the model of the extensional physical language of natural science idealized by logicians crippled by their own formalisms, these psychologists and psychophysiologists unconsciously were led to enquire into the supposed events, states or processes inside the speaker's organism behind each expressions of psychological attitudes in ordinary conversation. Misled by grammatical forms that are mere conventions between users of signs of a certain human community, they construed such conceptual entities as if they were substances accessible to objective methods of observation, manipulation, measurement, law induction, etc. In their scientistic enthusiasm, they had no regard for our human condition as members of communities of sign users. They did not notice the fact that as language users, we normally include the objective situations or state of things into the context of the interlocutors' attitudes. Nor that, as the result of multiple overlapping of these attitudes threatens to obscure and abort communication, we enforce amongst us certain common rules that we observe with constancy. So, that our sole warrant of the accessibility in a common world of any other language user, whatever his α priori unknown dispositions, is the fact that we were all trained by our parents or teachers to use the devices of language precisely as I said earlier. By distinguishing two modalities of use, one extensional, one intensional, and abstaining from using the devices that are conventionally intensional in the extensional mode, which threatens to spoil our precious and unique mutual understanding. Nobody could be so ingenuous as to imagine that such mannerisms be anything but purely conventional, and that, allowing for exceptions, they imply any internal mental state or process. Thus, generalizing boldly along the same line, Wittgenstein has suspected all psychology (or psychophysiology). When this psychology was not happy about relating bodily reflex movements to sensory stimuli, he accused it of erroneously taking such conventional forms of the social game as evidence of the occurrence of so many events or processes, either in the ethereal realm of mind, or in the wet material inside our skulls. What Wittgenstein's editors - but not he, himself - have emphatically called his "philosophy of psychology" is nothing but a careful sifting of the predicates of the language of mind, with a view to depriving them of their alleged mental and cerebral support, to tearing them away from the web of the causal chains of nature, and restore them to the web of the rules of social life. So, putting under scrutiny the expression "remembering what one was going to say at the moment one was interrupted", he throws doubts onto that which we are tempted to project into the past in this context: the image of a definite intention to say something that was already completed "in one's mind", but whose expression was repressed for some reason. Whatever the persuasiveness of this image, he counterbalanced it by making us

The Neurological Correlates of Action Types

115

pay attention to the function in actual conversation of such an admission, showing that it has most of the time nothing to do with an avowal referring to any episode in one's mental life. Critically examining the expression "I made it because I wanted to", he dissuades us to trust in the image of an agent inexorably lead to whatever he does by the causal influence of prior intentions, wants or desires in his mind. The sole fact of evoking the many alternative contexts of this expression, including the one in which it serves as a mere stopping of the conversation, suffices to divert us from psychological interpretations of mental causalities by focusing under our attention on the far more pressing social constraints that urge us to justify our conduct. All these analyses converge on the conclusion that language, as considered in its effective occasions of use in everyday life, couldn't be less concerned by the eventual states, events or processes occurring in our minds or our brains. That is not to deny the existence of any correlation between the predicates of the language of mind and such items, but to limit their importance to that of mere contingencies which do not help us understand the meaning of these predicates, a meaning that each particular context of use bestows on them. Wittgenstein offers us in his last writings a comparison that I found rather shocking at first, but that is, after all, not unfit to illustrate his point. He invites us to imagine some plants of different species (verschiedene Pflanzenarten), but whose seeds look the same, or are such that we could discover no difference between them, at least no difference that could be relevantly correlated with their difference in species. So, from viewing the seeds without knowing the plants they come from, one could not predict the plants they will produce. That, one could only say if one knew the plants they come from. In this case, Wittgenstein affirms very reasonably that such prediction, that has nothing objectionable to it, would yet not be a causal judgement. One would not say of such and such seed: "it will give a poppy because it came from a poppy", a causal judgement of course, but a false one based on something like action at a distance. The prediction we would make would rather be an explanation of our knowledge of the outcome of the seed alluding to the knowledge one have of its former history. Wittgenstein's point is precisely to plead for the recognition of this type of explanatory judgements that are not based on a search for some hidden difference to which the difference in the manifest outcome could be traced back as being its cause. He was well aware of having to fight for the recognition of this kind of judgements against "the powerful urge we have to see everything in terms of cause and effect" (Wittgenstein 1976, p. 375). Of this type are the judgements we express daily about each other on the sole basis of our knowledge of our respective publicly imputable actions and understandable reasons, while we couldn't care less about what's going on

116

Jean-Luc Petit

in each other's minds. In order to decipher the seed metaphor in this sense, let us substitute for the plants, the actions of some human agent; for the seeds what's going on in his mind; and for the prediction of the outcome of the seeds from our knowing which plants they come from, substituting the normal mutual understanding of each other's actions. Here Wittgenstein puts his finger on a common truth that is no less shocking for the scientifically biased mentality. For it seems to shake the belief in the universality of causality, or at least to threaten the hegemony of the inductive knowledge of natural science since it carves out a place for a type of practical knowledge without observation nor induction, that such science ignores. The reader may not be impressed by the trite truth that we can reliably anticipate each other's actions by taking only into account our reasons for action. But did Wittgenstein really claim more? Of course, I cannot claim that he did not go so far as to flirt with a bold prediction that was to suffer a flagrant contradiction provided by recent data obtained from cerebral imagery technology3. This prediction was that one would find no constant differences between the cerebral processes of agents observed while doing actions of different kinds. Anyhow, it must be said in Wittgenstein's defence, that his point was not a question of fact, but of language. We say "there must be a difference" and look for one where none was up to now discovered; we do not content ourselves by saying "there is a difference" while there is one in front of us. Viewing this grammatical usage as a specially interesting one, possibly characteristic of our modern scientific culture, the philosopher asked himself about our motivations for maintaining it as we do, come what may. A question that can only be answered, precisely, by putting such usage in the larger context of the other characteristics of our present way of life with its proprietary prejudices, Schemas of thought or world-views. Lastly, it remains that one can safely agree with Wittgenstein when he limits himself in opposing, on the one hand, to the cerebral processes that respond to general laws of the living, whatever the environment, and, on the other hand, to human actions that cannot but answer to the particular social pressures of each context in daily life, whatever each organism's inner condition. For, surely, whether in the good citizen or in the hardened lawbreaker, 500 msec, before the onset of muscular movement, the pre-motor cortex changes its activation level in the same way.

3

By measuring changes in cerebral blood flow (CBF) with positron emission tomography (PET) during the performance of predetermined tasks.

The Neurological Correlates of Action Types

3

117

Does von Wright's Two-Level Stance Provide an Alternative?

Wittgenstein, as every unprejudiced reader will agree, is more elusive, ironical and thought-provoking than systematic - von Wright's is a more architectonic style of thought. Wittgenstein contented himself to lowering our threshold of logical sensibility to the many absurdities that result from the fact that we tend to accommodate our conception of psychological life to a model of classrooms physics. Von Wright was inspired by Wittgenstein's paradoxes and profound but scattered remarks to articulate a large-scale model of understanding that could counterbalance as a rational alternative the current explanatory model of empirical science. In Wittgenstein, the pervading theme of "language games" is more a Leitmotiv than a working part of a system. Von Wright sets it at work, freeing it from excessive fixation to language, and integrating it into a generalized theory of intentionality including every object-oriented intentional behaviour, be it answering to an internal motivation or to a symbolic challenge for participation in an institutional practice. Wittgenstein tirelessly derided the naive generalization of the extensionalist treatment of descriptive discourse to our verbs and nouns of intentional attitudes. But he did not disclose the true reason for his insistence, apart from warning us of the dangers of our seemingly chronic tendency to reify imaginary entities on the basis of a misinterpretation of some non referential language uses. Von Wright, for his part, has experimented with a systematic mode of determining the meanings of these expressions. Stressing their exclusivity to human agents that normally intend to do whatever they do - by tracing the links that weave them within a closed conceptual web through which framework we, the agents, view everything; by remembering the rule-governed internalized practices of social life that regulate their use - he established firmly the fact that our understanding of such expressions stand on its own footing. This is not normally derived from a prior understanding of expressions referring to non human facts and their causal relationships. He even went so far as to hypothesize that our prior understanding of human actions might be necessary for the recognition of such causal relationships in the non human world of facts. Taking all that into account, von Wright's philosophy of mind would better be characterized as purely intentionalistic, as he clings to the fact that everything mental (or cultural) is screened from objective reality by the frontiers of an intentional context centered on the bearer of attitudes (eventually a collective). And he teaches us to reconsider this very same condition of intentionality, not as a deplorable deficit in semantic force that mars some expressions that we should avoid, but as an indispensable - so to say, α priori - condition of the intelligibility of human conduct. In spite of not being, then, under the threat of a denial by objective facts of the

118

Jean-Luc Petit

external world that hovers around non-intentional concepts, our intentional expressions of "action", "intention", "will", "desire", "belief, etc. are not for all that, deprived of any truth-grounds, nor of any meaning content. But we must be aware of the fact, von Wright contends, that what bestows on them truth-value and meaning is just so much and no more as the way human agents understand themselves and the things they do; the way human agents explain themselves to each other in their daily transactions; and the generally expected and reassuring adequacy of the fitness of their actions and expressions in a certain pattern of common understanding of the world that is in use in their social communities. Thus, von Wright agrees with Wittgenstein in presupposing an insurmountable conceptual (or "grammatical") gap between the intentional language and the modern scientific one. Now, in scientific or common parlance, we not infrequently hear or read such categorically impure modes of expression as "the class of neurons F represents (or computes, decides, simulates, or what not, choosing what you like in our intentional vocabulary) the intentional content V". Is any logical principle offended by this shifted application of the intentional attributes of inter-personal ordinary discourse to the typical infra-personal entities (brains, nervous cells, molecules) of natural science? Not necessarily. Those who maintain that the language of science could not be anything but a rigorously causal and physical reduction of our everyday intentional vernacular that refers to personal agents and their goal-seeking conducts are dreaming of an ideal science that does not exist actually and may never exist at all. Meanwhile, let us not prejudge the logico-semantic homogeneity of such a motley array of disciplines as that which, today, creatively interlace neurophysiology with molecular biology on the one hand, ethology, psychology and psychiatry on the other. Pending the final outcome of the controversy over the unity of science, one cannot help but wonder if the socalled "cognitive neurosciences" are not weaving precisely the kind of Φ-Φ systematic correlations on a generic level that the intentionalism of linguistic analysis seems not to admit without reservations as being accessible to experimental research. While classically, the pertinent correlations were between external stimuli of low level reflexes and mere bodily movements, not to be confused with human actions so that no overlapping of the empirical and the conceptual was then to be feared, we have recently been confronted by a dramatically altered situation. A mature neuroscience seeks correlations between the high level supra-modal functioning of associative, secondary cortical areas of the brain and, if not exactly the human actions and perceptions, intentions and motivations experienced from within by the subject, at least by the corresponding traits of their empirical analysis. And this empirical analysis sometimes expands and carries on, sometimes straightens, interpolates

The Neurological Correlates of Action Types

119

unforeseen new details and threatens superseding their (common sense, or) logical analysis. Would the establishment of such "cognitive" correlations on an ever growing scale not constitute a serious challenge to the thesis of the autonomy of the intentional? Perhaps, we can answer that "understanding" will forever remain safe from the vicissitudes of "explanation". But, beforehand, nobody can disregard the possibility that his understanding - even his selfunderstanding - would some day end benefit from the empirical findings, to say nothing of the most tough-minded mechanistic hypotheses of neuroscientists. This eventuality, whether or not we like it, is one we must nevertheless take account of. But how? We would be glad to answer: "Along the lines of von Wright's intentionality theory", only that it cannot be without some reservations. For, in order to cope with this problem in the perspective of a logical analysis of intentional language, this theory would see itself as being torn between two claims. The one is to graft intentionality onto the preintentional quasi-teleology, quasi-causality of the biological functions of the organism. The second is Wittgenstein's refusal to let his sharp distinction between the intentional and the non intentional be blurred. Would such theory bear the strain of taking up both requirements at the same time, supposing that these are not logically contradictory? At first blush, von Wright's (1980) generous interpretation4 of the thesis of the "equality" of the reasons and the causes of our actions, upheld by the causal theorists, has the double advantage of preserving something of the autonomy of the reasons in relation to the causes, without excluding (as boldly as Wittgenstein 1967, § 608) the very possibility of empirical research finding neurological causes for our movements, if not our actions: 4

This is to be contrasted to the rather ungenerous appreciation of it by Wittgenstein, after von Wright himself (1995, p. 15): "The metaphysics which Wittgenstein is fighting is thus not one rooted in theology but one rooted in science [... ] Of this he gives clear warning in the Blue Book where he wrote: 'Philosophers constantly see the methods of science before their eyes, and are irresistibly tempted to answer questions in the way science does. This tendency is the real source of metaphysics, and leads the philosopher into complete darkness (Wittgenstein 1958).' He immediately gives examples: The craving for general theories, contrasted with what he calls 'the contemptuous attitude to the particular case'; the tendency to [...] reduce [...] intentional behaviour to bodily movement. The most vulgar examples of these tendencies we find, it seems to me, in contemporary philosophy of mind, be it in the form of the physicalists' 'identification' of so-called mental states with brain processes or the eliminative materialists' rejection of our common sense psychological concepts - what they call 'folk-psychology' - as a 'radically false theory' eventually to be replaced by a perfected neuroscience."

120

Jean-Luc Petit "This is, I suppose, the point where my position crucially differs from the one taken by supporters of a causal theory of action explanation. A 'causalist' would argue that in order to explain - and not merely justify - an action the reason must, somehow, equal a cause. The difficulty for this position, as I see it, is with the 'equal'. According to one influential opinion the reason must, in order to be explicative, have a neuro-physiological causal equivalent. This need not mean that generically the same brain event corresponds to the same generic reason, e.g. the receiving of a certain order. The equality between reason and cause is an equality at the level of tokens, not of types. But then to insist that the reason causes the behaviour seems to me merely to be to insist upon the existence of sufficient neuro-physiological causes of the bodily movements in which an action manifests itself. Whether there are such causes or not is a question (of physiology) in its own right. Even if the answer to it turned out to be that there are not, or not always, such causes operating, the explanation of the action remains unaffected. Its truth-grounds are not physiological" (von Wright 1980, p. 61, n. 2).

Thus, if a reason explanation is true, and really explains an action, von Wright concedes that the reason alluded to must also be the cause of that action. But he understands such an assertion as referring to the action in two distinct senses, the one, of the reason explanation, the other, of the causal explanation: (1)

As a generic act, an action is answering to a reason, equally generic: one "opens a window" because one "wants to get some fresh air"; one "passes the salt" because somebody "asked one to", etc.

(2)

As it manifests itself through a well determined bodily movement, an action has to have also an well-determined neurological cause - as such it can't possibly be other than an action token, not, as in the first case, an action type.

So that, if von Wright admits the identity of reasons and causes, it is only at the level of the tokens and in the context of a causal explanation. At the same time, he can maintain that an explanation of the action as such remains a reason explanation that has nothing to do with causal explanations, and this so much so, that it relates the action as a type to its reasons, that are considered also as types. However, what this solution seems to presuppose is the definitive inability of a neurological enquiry into the causes of our movements (and/or actions) to access the level of action types at which we habitually understand and "explain" our actions. So that, every pretended action explanation in neurophysiology seems to be interpreted by von Wright, as it was traditionally interpreted, as referring only to bodily movements and low level reflexes that link local stimulations from the environment upon specific surfaces of

The Neurological Correlates of Action Types

121

the body to the pre-wired motor reactions of one definite member. Such an interpretation looks like a mere mitigation of Wittgenstein's denial of the possibility of "brain reading" under the pretext that we talk about what we do, not about the (for him) questionable neurological processes inside our brain that physicians affirm to occur when we are doing it. Nevertheless, some recent progress in neuroscience, for example, the discovery of "mirror neurons" by Rizzolatti et al. (1992)5 evidenced the fact that if the neuroscientists can't exactly look inside our brain and say what we are doing, they can literally do that with monkeys (the accepted model of man in this area). Having surgically implanted electrodes in the premotor cortex of a monkey, they literally6 can say, if not predict, while looking at the "raster" and listening to the loud speaker retrieving the electrobiological discharge of certain neurons, what is the precise "action" that the monkey is presently doing (or that it is observing his experimentator do), as the curve profile correlated to the action of "picking up a raisin with the tips of the fingers" is distinctively recognisable from the one correlated to "grasping a banana with full hand", or to "tearing something by twisting it in a contrary sense with the hands", etc. And the same researchers insist that what the neurons encode is not such low level and narrowly specified characteristics of the bodily movement as certain innervation paths, muscle chains or individual motor members, but the actions as such, with their fully intersubjective, practical meaning, as distinctively opposed to one another inside the ecological repertoire of the animal ("picking" vs "grasping", etc.). How are such phenomena to be interpreted? Are they a refutation of the autonomy of our language of actions and their reasons in relation to their neurological causal explanation, or are they urging us to re-evaluate the deep rootedness of the practical categories of this language in the neurological structure of the organism. Or are they both? To this question, von Wright's kind response in Bielefeld was: "To some extent, it seems that the idea of what is sometimes called a 'token materialism' has been a kind of - not misuse - but something which has been proposed just, so to say, faute de mieur. that one can't, one has not means even to observe correlations of the type (level). Whether that holds altogether, whether false or true, on the grounds of principle, I don't know; but it seems to me that, although one can have qualms about materialist readings, this functions more like a kind of programme, or just a postulate: there must be something in the brain that corresponds to the types which correspond to the mental states. So, that, for these findings about the monkeys and their reactions to what (a man does), they are very, very 5

At the Istituto di Fisiologia Umana, University of Parma, Italy. And astonishingly, I must confess, for having had a video-tape projection of this experiment in a 1995 workshop on philosophy and neuroscience in Strasbourg. 6

122

Jean-Luc Petit interesting, and it's an advance towards not only speculative ideas about correlations, but also findings about generalities."7

Although such a response very aptly reminds us not to be too dogmatic about his own distinction between a type, and a token materialism, it does not seem to me to remove entirely the risk of inconsistency I tried to call attention to in his two-level stance. In fact, if the neurosciences are recognized - as, it seems to me, they should be - to have a fully legitimate access to the level of actions as generic entities of human understanding, then the Wittgensteinian move of defining this very same level as proprietary to ordinary language and prohibiting scientific empirical enquiry from entering it has to be abandoned. Even if this is at the cost of putting the theory of action in the challenging situation of a dialogue with neuroscience, that its founder (Wittgenstein) wanted so badly to keep separate from philosophy.

References EGIDI (1995). Rosaria Egidi (ed.): Wittgenstein: Mind and Language. Dordrecht 1995. GAZZANIGA (1995). M. S. Gazzaniga (ed.): The Cognitive Neurosciences. Cambridge, Mass. 1995. KOSSLYN/ANDERSEN (1992). S. M. Kosslyn and R. A. Andersen (eds.): Frontiers in Cognitive Neuroscience. Cambridge, Mass., 1992. RIZZOLATTI ET AL. (1992). Giorgdo di Pellegrino, Luciano Fadiga, Leonardo Fogassi, Vittorio Gallese, and Giacomo Rizzolatti: Understanding Motor Events: A Neurophysiological Study. Experimental Brain Research 91 (1992). VON WRIGHT (1980). Georg Henrik von Wright: Freedom and Determination. Acta Philosophica Fennica Vol. 31 (1980). VON WRIGHT (1995). Georg Henrik von Wright: Wittgenstein and the Twentieth Century. In Egidi (1995). WITTGENSTEIN (1958). Ludwig Wittgenstein: The Blue and Brown Books. Oxford 1958. WITTGENSTEIN (1967). Ludwig Wittgenstein: Zettel Oxford 1967. WITTGENSTEIN (1976). Ludwig Wittgenstein: Philosophical Occasions 19121951. Indianapolis 1993; first ed. first publ. in 1976.

7

Bielefeld, 27 April 1996.

HANS JULIUS SCHNEIDER Mind, Matter, and our Longing for the One World' Abstract: In a first step I will summarize those accomplishments of von Wright's (1971) Explanation and Understanding that in my eyes are most important for the mind-body problem. In part two I will interpret our desire to live in One world' as one of the attractions of monism. I will characterize materialistic monism as an unjustified ideology and will (following ideas in EaU) offer an alternative (and non-idealistic) way for satisfying our monistic inclinations. Part three will take up von Wright's discussion in his paper On Mind and Matter (1994), interpret it, and by doing so will possibly push its significance into a direction the author himself did not intend. In part four, finally, I will ask how far the discussion has brought us to a solution or dissolution of the traditional problem, how the mind-body distinction relates to the mind-matter distinction, and what new problems the dissolution of the traditional problem has brought. 1 2 3 4

"Explanation and Understanding" and the Mind-Matter Problem Materialistic Monism and Our Longing for the One World' Von Wright on Mind and Matter The Dissolution of the Old Problem and the Next Steps to be Considered

1

"Explanation and Understanding" and the Mind-Matter Problem

Many years ago I had the pleasant experience that an attempt of mine to carry a thought of Georg Henrik von Wright just one small step further1 was greeted by him with acceptance and encouragement. With the present paper I boldly try to repeat this experience: For a long time now I have had the impression that EaU2 contains all the ingredients of what might be called a solution or 'dissolution' of the traditional mind-body problem, and I would like to use the publication of this volume as an opportunity to spell out my reasons, especially since he has recently taken up the problem himself.3 This time, of course, my proposals might have a different fate. But since I was fortunate enough to get to know his most friendly, generous, and modest Schneider (1978). Von Wright (1971); quoted in the following text as EaU. - I would like to thank John Granrose for taking the trouble of correcting my English. 3 Von Wright (1994). With his kind permission I am also using an unpublished typescript with the same title, prepared by him for lectures in Leipzig. 2

124

Hans Julius Schneider

personality, and, unavoidably, have grown older and become less vulnerable to criticism, I am prepared to accept even a complete rejection from him. What then are the accomplishments of EaU that are most pertinent for the mind-body problem? The book succeeded in showing, firstly, that the realm of understanding actions has epistemological priority over the realm of explaining events in the material world of physics and, secondly, that the world of our actions is broader, more encompassing, than the world as pictured (actually or even potentially) by the natural sciences. Von Wright established the first point, the epistemological priority, by demonstrating that the concept of a 'law of nature' (and in particular, the concept of causality in a Humean sense) depends on the concept of action (in particular, on the concept of handling and watching what he has described as 'systems' of events in the material world). To put it in a nutshell: Without the experience of myself freely experimenting with chains of states of affairs, of acting and refraining from action, the very concepts of causality and lawlikeness could not exist. Turning to the second point, the broader scope of the world of action as compared to the world of causality, one might distinguish a 'horizontal' from a 'vertical' aspect of this scope, although both dimensions are closely related. Horizontally, the broader scope of the world of action is established by the same argument that shows the epistemological priority: If indeed the existence of a world of science (i.e. of a totality of reasonably defensible claims about laws of nature) presupposes an active being, a person conducting experiments, it follows that for every claim about a particular lawful connection there must be a person who has freely designed and conducted a related experiment4 and can judge about the evaluation of its result in a causally unconstrained manner (i.e. her words are arguments, not just causally produced noises). So the experimenting person necessarily is situated outside the particular experiment in question, outside the 'system' under investigation. Its initial state, the first state of the causal chain, is situated at a moment after the scientist has acted, and for this reason the investigated system is narrower than the total situation of designing and conducting an experiment. It follows that every gain in the effort to enlarge the scope of our scientific control of systems of events, every gain in our manipulative capacities, necessarily is produced by actions outside this enlarged scope. Metaphorically speaking: Regardless of how much the table will be enlarged on which our experiments are set up, there will always be a scientist standing (in any particular experiment) beside the table. She has prepared it and 'sets it in motion', so to speak, but is herself situated outside the experiment. Even 4

The expression 'related' is important here, since we obviously cannot handle e.g. astronomical objects.

Mind, Matter, and our Longing for the One World'

125

if one would accept a success in manipulating certain 'mental events' as a sufficient reason to call them 'causally produced' by the (result of the) manipulating action (I shall return to this crucial point), the unmanipulated actions that are necessary conditions for the scientific experiment demonstrating this success could themselves not be called 'causally produced' in that particular instance. Using the same simile once more: Even if the experimenting scientist himself (Prof. A) is put upon the table by force and is in that way (together with the subject of his own research, person X) made a part of an enlarged 'system', this step necessarily involves an investigating colleague (Prof. B) who conducts the experiment on Ά experimenting on X\ So the enlarged experiment necessarily is conducted by some acting person (Prof. B), and so on, for every enlargement, ad infinitum. This is why von Wright can claim that the race between action and causality will for conceptual reasons always be won by action: Without genuine actions there can be no claims to causality. Here we see already the other, the 'vertical' aspect of the fact that the world of action has a broader scope (or 'depth', in this case) than the world of causality. It is worked out by von Wright by demonstrating that reasons, i.e. the entities we refer to when we describe actions, cannot be construed as (a special kind of) causes. Not only does any specific causal claim, regardless of whether it concerns billiard balls or brain states, presuppose a scientist setting up his experiments in a reasonable way, but it can also be shown that for conceptual reasons it is impossible to redescribe what has happened by treating her reference to her reasons for the particular setup as just a special case of a reference to causes. The efforts we make in explaining our actions in the usual way of referring to goals and 'states of mind' (like convictions, preferences, etc.) cannot be likened to explanations we give of an event by referring to its causes, even if we would know a great deal more about neural states and the causal relations between them. So not only are actions necessary to establish claims about causal connections, but it is also the case that the particular vocabulary we have to use to discuss actions cannot be transformed or translated into a vocabulary for the description of causal processes. It is a logical point von Wright is making here, and for this reason it is independent of the state of knowledge in the cognitive sciences. It is known as his version of the 'logical connection argument' and I will return to it at a later point. The conclusion it allows is the following: On both counts 'mental language games' are unavoidable. Horizontally, as far as the scope of 'things talked about' in causal language is concerned, there must always be a place left for the scientist who explains the rationality of his experiments, and whose very words cannot in this moment be taken as causal events. And vertically, as far as the means for talking are concerned, the language

126

Hans Julius Schneider

of giving reasons for actions (like setting up a particular experiment) cannot be replaced by a language citing causes. The prima facie difference between the manipulating action on the one hand and the manipulated system on the other can be maintained and shown to be crucial by philosophical analysis. So the following overall picture emerges: We see that the scientific experiment is a very special kind of action, yielding a very special kind of experience. As technology shows, this kind of experience brings great power with it, and this aspect might lead us to think that there is no limit to it: Ecological damage, for instance, can be detected in the most remote areas of the earth. On the other hand we must not forget that our knowledge of the 'world at large' rests on many different kinds of experience, of which the experience produced by science is only one. The world as seen by scientific experiments is the result of a very specialized view; necessarily it leaves out many things we know from other sources, from the experience we have with our own bodies, for example, or with our social activities, including the special case of linguistic interactions. This broader realm of non-scientific experience ('die Lebenswelf) is by no means restricted to 'the mental' in a narrow sense; it includes kisses and toothaches. So we should not forget that the world of science is very much 'smaller' indeed than the world as perceived and articulated in those areas of experience that are free of the methodological constraints that make science possible. This observation about the partial character of the world of science does not deny that at the same time it is larger in so far as it reaches out to invisible stars and 'down' to subatomic particles.

2

Materialistic Monism and Our Longing for the One World'

At this point some readers might have in mind the following objection: Even if it is admitted that in order to establish (i.e. to get to know about) a causal relation it is necessary to freely conduct experiments, why should it not be admissible to generalize after the experiments? Why should it not be possible to claim retrospectively that at the time when the experiments were conducted by Professors A and B, laws of neurophysiology etc. must have been effective which causally also forced the experimenting scientist B (watching A's brain during his research on X) to behave in the way she did? Would not honesty and sobriety demand the conjecture that her sense of freely setting up her experiments on A must have been an illusion? Von Wright here rightly asks: How could such a claim be established? And his answer is: Only by acting once again, i.e. by freely conducting experiments on B, and so again the realm of causality cannot be total, the person con-

Mind, Matter, and our Longing for the One World'

127

ducting the new experiment (Prof. C) is outside of it. This much I have tried to express above by the simile of the enlarged table. In addition I would like to stress here that to deny this epistemological restriction of our claims about the existence of 'laws of nature' would be to transgress the border from science to ideology. One would yield to the feeling that somehow there must be universal causal chains, even if we will never be able to completely establish them. In order to avoid this transgression I equated the 'world of science' not with something Out there', something 'given', but with the totality of reasonably defensible claims about laws of nature. If science is to remain science and not to become a new faith, this modesty is mandatory. There is no rational source where the cited feeling of a 'must' could come from. It is one thing to say that to do natural science means to look for 'laws of nature' as hard as one can, i.e. to try to push our ability to manipulate systems as far as possible. But it is something completely different to say that the 'laws of nature' must be out there, fixed and ready, waiting to be discovered by us. To my mind claims of this kind are an expression of misunderstood theology. On the other hand, if it is not completely misguided to speak of a feeling that there should be universal laws, one might well ask what the sources of such a feeling could be and whether it could not be expressed by or at least related to another, more reasonable claim that avoids ideology. It seems to me that what is at stake here is a longing for the One world', for the unity of the world as we (sometimes) experience it. This desire is what makes all forms of monism so attractive as compared to dualism. And it is worthwhile to pause and see what might be behind this longing for (or even experience of) unity. When we have answered this question we will find that von Wright has to offer an alternative understanding of the desired unity. I would like to propose that it is our actual experience of having lived in One world' before our encounter with higher learning and the sciences, and the experience that the complex reality of an adult person in an industrialized society has emerged via a process of differentiation, - that it is this double experience that stands behind the need to also project some kind of unity into the future, as the desired result of systematic inquiry. The way in which we experience ourselves in unreflected moments, for example, is (in the non-pathological case) not one of duality in the sense of 'mind' standing against 'matter'. So when we come to learn more and more about ourselves by reading about medicine, psychology, etc., the plausibility of the picture we are offered depends, among other things, on the degree of unity it can offer. In this respect, any form of monism has a clear advantage over dualistic positions. The physicalist picture, the monistic claim that 'basically' or 'strictly speaking' all reality is physical reality, surely offers us the unity of One world', not to mention its additional advantage of making life in

128

Hans Julius Schneider

Scottish castles much easier because it discourages the belief in disembodied ghosts. And as far as science and technology are still dominating the Zeitgeist, the physicalist picture is more attractive than an idealistic monism, especially since some authors of popular science books have assured us that modern physics is all we need to get seriously acquainted with the realm of the mental, including even Eastern Mysticism. What von Wright offers in EaU might for some readers be less entertaining and exciting, but it is a much more plausible picture: The unity is and remains that of our experience as acting persons in a social context. To understand the realm of the mental means to be capable of language games like giving reasons, expressing wishes, fears, or opinions.5 These language games are tightly interwoven with normal, visible activities that involve our bodies; no need for ghosts or modern physics at this point. This understanding of meaningful actions in a social context is well established long before an understanding of the physical (in the sense of physics, not of eating something or hurting one's fingers) can enter the scene, because, as we have seen, physics is made possible by a special form of reasoned action, namely by scientific experiment. When we understand and accept this unity in our experience as acting persons, when we see that scientific thinking does not threaten it as long as it does not transgress the border to ideology, it is no longer necessary to hope for a unity as a goal that science should strive at, as its projected end, located far ahead in the distant future. And in so far as science is made possible by abstraction it is even paradoxical to expect it to regain the territory the leaving out of which has constituted it in the first place. Using a phrase of Nelson Goodman6, some of our intellectual, artistic, and other cultural endeavors can be described as 'ways of worldmaking'. The scientific part of these endeavors need not lead us to a grand unified theory. And there is nothing to regret about this, as long as we do not loose sight of our primary awareness of unity out of which all intellectual and artistic endeavors spring. I think that it is one of the tasks of philosophy (brilliantly exemplified in the work of von Wright) to help us stay in touch with this unity. It would be a grave mistake to trade it for the price of making philosophy 'more scientific'.

3

Von Wright on Mind and Matter

I shall now summarize and interpret some main points of von Wright's recent discussion of the mind-matter problem, without trying to draw a line 5

'Language games' must be understood here in a very broad sense that includes forms of non-verbal communication. 6 Goodman (1978).

Mind, Matter, and our Longing for the One World'

129

between my interpretation of his thought and my attempts to carry it further. He himself formulates as the problem he sets out to overcome "the apparent irreconcilability of interactionism with what we think of as a scientific understanding of natural phenomena".7 And he announces right at the beginning that he will reject the idea of body-mind interactionism.8 Von Wright formulates his main conclusion in the following words: "Simplifying a little, one might say that science investigates facts about the world and their causal interrelations. In the nexus of these facts and their connections the mental has no place. No 'ghost in the machine' must be allowed to disturb the scientist's peace of mind."9

In the traditional idea of a body-mind interactionism two directions of influence can be distinguished which are both contained in our ordinary ways of speaking. On the one hand something mental is taken to influence the body, as when what even in non-scientific talk by now is called a 'motive' sets a person in 'motion'. We say, for instance, 'the desire to take revenge motivated him to pick apples from his neighbour's tree'. In popular thought, the motive is conceived of as an inner person, a ghost, an invisible force acting on the limbs of the motivated person. On the other hand something happening to the body is taken to influence the mind, as when, for instance, alcohol (or being hit on one's head) disturbs (or definitely ends) a person's mental activities. As already noted, it was one of the main results of EaU that mental entities like goals and beliefs cannot, for logical reasons, be construed as causes for actions. The (quite complicated) argument that shows this impossibility is an epistemological one (I have hinted at it already above): The attribution of certain goals and beliefs (mental states M) to a person P in order to explain his or her action A is not possible independently of the performance or not-performance of the very action A on the given occasion. In this way the explanans (P wants to accomplish a goal G and thinks that doing A will under the given circumstances lead to G) is not logically independent of the explanandum (P does ^4). And since for causal explanations explanans and explanandum have to be logically independent, the explanation of an action with reference to mental states like goals and beliefs must be of a different, non-causal type of explanation. Hence mental states are not causes for actions; there is no causal interaction from mental states to actions. This conclusion in no way discredits our usual way of making actions understood by citing goals and beliefs, it only insists that this type of explanation differs from causal explanations. And it does undermine the popular idea of 7

Von Wright (1994), p. 101, abstract. "Ibid. 9 Von Wright (1994), p. 109.

130

Hans Julius Schneider

'the will' being a cause for an action. In his book (following a well known philosophical tradition) von Wright speaks of 'understanding' actions versus 'explaining' natural events. 3.1

Criteria and the Existence of the Mental

In his more recent paper On Mind and Matter™ he mainly discusses the other direction of influence associated with interactionism, the influence of the material on the mental. Can certain physical happenings in the nervous system (including the brain) be construed as causes (in the strict sense elaborated in Eall) for 'mental events'? Can the stimulus of a certain nerve connected to a person's ear, for example, be construed as the cause for his or her hearing a certain noise? Can certain happenings in the history of his or her brain be causes for a person's having certain goals and beliefs? Von Wright's answer is clearly negative: The relation in question is not a causal one. Interestingly, he points out that our dim realization of the non-causal character of this relation might be one reason for the prima facie plausibility of the identity theory.11 If we would assume that A and B are identical, it would follow that they cannot stand in a causal relation to each other. So the claim that brain state and mental state are identical would (whatever else we think of it) have the advantage to ascertain the truth that there is no causal relation between them. But this alone, of course, is no sufficient reason for subscribing to the identity theory. And as von Wright points out, there are strong reasons against it: a nerve-state is something spatial, the sensation of hearing a noise is not; the first can be observed by different people, the second is only 'had' by (or attributed to) one subject; the first had to be discovered by science, the second is known without science; etc. So why can physical happenings in the nervous system not be construed as causes for 'mental states and events' ? In order to answer this question we should be able to say what mental events and states are. Without an answer to this question we can hardly judge whether it makes sense to construe them as standing at the end of a causal chain. Von Wright's first move to answer this question consists in taking an epistemic approach. He asks: What are the circumstances that permit us to correctly ascribe a certain mental state to a person? To take this question as an epistemic one means to ask for criteria for mental states, not for symptoms. This difference is very important for the argument; it can be explained in the following way: Asking for criteria means to ask for the meaning of an expression; it means to ask how our first encounter with the Object' in question must be understood. 10 n

Von Wright (1994). l have proposed above that our longing for One world' might be another reason.

Mind, Matter, and our Longing for the One World'

131

To ask for symptoms, on the other hand, presupposes that the object in question is already known; it is familiar but difficult to detect. A person can know, for instance, what it means to tell a lie, but since cases of lying are often hard to detect, it makes sense to look for clues, for signs that can be taken as indicating that probably on the given occasion a suspect does lie. In this sense a red face and nervous behaviour may, for example, be symptoms of lying. Von Wright's second move (and here I might interpret him in a more radical way than he intended to be read) is the answer he gives to the epistemic question. It is that all ascriptions of even the most primitive of mental states to a living being have as their criteria interpreted actions of that being, i.e. behaviour described in terms of its meaning. In order to contrast this criteriological view once more with the symptomatic we can formulate: Mental states are not hidden entities the presence of which we infer from certain symptoms in the behaviour of living beings. Especially, they are no states of the nervous system, watched from 'inside'. Ί can cite Poe's poem The Raven by heart' does not mean 'looking into my brain from inside I find that it is in such and such a state'. In order to assess the correctness of an ascription of a mental state to a person whose actions we have seen and interpreted, we typically ask the person. And to understand the answers we get we once again have to take them as meaningful actions, this time as actions in the medium of a shared language, i.e. a shared practice. And to take them in this way does not mean to take them as symptoms. On the level of animals we cannot resort to speech acts but have to watch their behaviour and consider it, for instance, as exhibiting the desire to eat, to hide, to avoid pain, etc. So the actions of persons (and the behaviour of animals interpreted in analogy to our actions) are the criteria for ascribing mental states. As we have noted, this is to say that to ascribe one of them to a person (or another living being) means to make certain claims about him or her in terms of a language of actions, i.e. in the domain of understanding meanings, in contradistinction to the domain of manipulating systems of events. The very existence of the entities we have learned to talk about (via the discussed criteria) as 'mental states' depends on the step that we decide to use a language with the help of which we attribute actions to agents and discuss their meaning. It is the possibility to establish this practice that gives rise to the possibility of 'referring to mental states' in the first place; outside of it, expressions of the form 'the mental state M' have no meaning. This, I understand, is the view to which we are led by von Wright's claim that the relation between mental states and certain characteristic actions is a semantic one; that the activities of a living being are the criteria, not the symptoms for its mental states.

132

Hans Julius Schneider

If this is correct, we can say that mental states 'exist' only in the context of a discussion about the proper understanding of an action. Another formulation for the same claim would be that the speech act 'to refer to a mental state' is only possible in the context of discussing actions. In contrast, it seems correct and fair to say that the language of physics, when it is confronted with a living being, abstracts from the meanings of its actions, i.e. it abstracts from those characteristics that make an action an action, in contradistinction to a mere series of states of affairs. To look at bodily processes from the point of view of physics or chemistry means to methodically exclude purposes, goals, beliefs, etc. from the description of the behaviour. And this means that for science (in so far as its defining characteristics include the search for causality and laws of nature) mental states 'do not exist'. And it is clear that to say this is nothing but to pronounce a conceptual truth, a tautology. 3.2

No Causal Chain

But if indeed mental states simply have no place in causal science, this entails that they have no place as possible effects or end-states of causal chains. If we stick to the methodical restrictions that define causal inquiry, we by definition exclude all reference to actions as meaningful activities of beings with certain interests, etc. And this exclusion is once and for all: What is methodically kept out of the picture can and will never come in again. Mental states will not suddenly reappear, neither in the role of causes for material events, nor in the role of their effects. Once we have settled for the language of causality (for the 'machine-view'), no 'ghosts' are going to 'emerge' as long as we avoid conceptual mistakes, they will appear neither as causes nor as effects. For what would we have to show to establish the contrary? If we wanted to claim that a certain physical happening P in the nervous system of a being is the cause for a certain mental state M, we would have to show that the willful production of P in an experiment will invariably be followed by those actions the occurrence of which justifies the claim that the being is in the mental state M. But the last sentence contains a category mistake the repair of which will show that the causal chain we are looking for cannot be constructed. For what does the word 'actions' mean in the above sentence ('will invariably be followed by those actions')? When we conduct causal experiments, we can only speak about chains of observable material events, considered strictly as material events. We are not allowed to view the animal behaviour in question as a case, for instance, of fleeing a stimulus, or expressing pain, only as exhibiting such and such changes. So what we can willfully produce can only be bodily changes, not actions. And since ac-

Mind, Matter, and our Longing for the One World'

133

tions and actions alone are the criteria for mental states, we cannot causally produce mental states. The goal of including mental states as possible results into our experimental setup can only be reached by adopting a description of the behaviour in question in terms of actions. But with such a move the realm of causal investigations would already have been left behind; we could no longer interpret our investigation as an attempt to show that causal language can reach out to mental states. Only when from the very beginning we adopt the richer language of action can we talk about mental states and their relation to the 'situation' in which they occur. This situation will be the special kind of 'situatedness' in which living beings find themselves. This 'environment' will have to be described in a language that is richer than the language of physics. And it is no contradiction to admit that it will treat of 'physical' things like eatable food or hospitable climate; they are 'physical' in the sense of 'related to the body', which is not the same as 'a subject matter of the science of physics'. Admittedly, our common way of speaking about causality does not agree with these results. Although everybody will admit that certain mental effects (like understanding Kant) are impossible to unfailingly produce by physical means, we tend to think that there are also cases in which physical changes do cause mental changes. One might think of the following extreme case as a plausible candidate: Destroying the brain of a deer by shooting it in the head certainly means to produce a physical happening in its nervous system. Also we will have an invariable 'physical' result here, the unfailing production of which justifies us in speaking of causality in von Wright's action-oriented sense, namely, the death of the animal. Is not this result of the envisaged action of shooting also something mental (when we allow action-terminology for the description of the behaviour of deer in the first place)? The result we are dealing with is not a set of characteristic actions that could be criteria for a mental state, but it is the ending of all actions. To say of the animal that it is dead, does not mean to ascribe a mental state to it. On the contrary, it seems that an adequate definition of death should entail that all 'mental states' in the sense discussed here have come to an end. So the case that more than others seems to deserve the description that by a manipulation of a physical system a mental result has been produced and hence seems to be the strongest candidate for the desired kind of causality, is at the same time most remote from the domain of actions and mental states. If we change the example from shooting the animal to inflicting severe pain, our observations from above immediately become relevant: To describe the animal as 'being in pain' (instead of being in such and such a state as far as its nerves and limbs are concerned) means to describe it in terms of

134

Hans Julius Schneider

meaningful actions. And although severe pain can be considered as a mental state as close to the living body as possible, it still remains a mental state, in contradistinction to a physical state. But this means that the expression meant to describe a cause, namely 'to inflict pain', itself belongs to the mental vocabulary, so again we have not crossed the line from the physical to the mental. To put it simply: an expression like 'to slap somebody in his face' is a part of meaning-related, i.e. 'mental' language; it does not belong to physics, to the production of an initial state of a system, although in court the act will be classified as one of 'physical violence'. This observation shows that the vocabulary we use in everyday language to describe what we do to each other ('you are stepping on my foot', 'you hurt me', 'please hold my hand') belongs to the realm of meaningful actions, not to the realm of causality in von Wright's sense. The fact that what is done involves the 'physical' bodies of the participants (in German: 'der Leib') in no way justifies the conclusion that the subject matter we are talking about is 'the physical' in the sense of physics (Lder Körper1}. For this reason Descartes' choice of the terms res extensa is highly misleading: By no means everything extended in space must for that reason be seen as an object of physics, as the briefest reflection on works of art, persons or philosophy books shows.

4

The Dissolution of the Old Problem and the Next Steps to be Considered

How close did the above considerations bring us to a solution of the mindbody problem? Certainly it has not been explained how ghosts (res cogitans) can operate machines (res extensa) and how machines can have causal effects on ghosts. This is why the problem as traditionally formulated has not been solved, but rather dissolved. Von Wright says about his paper that it "argues that the problem of interactionism may be overcome by the bodymind distinction being viewed as a division between two ways of looking at living beings: one consisting in relating overt behavioural reactions to intra-bodily causes and effects, the other consisting in understanding what these reactions mean."12

The overcoming or dissolution of the problem consists in realizing that what seemed to be a distinction between two kinds of things, two kinds of 'stuff', can more adequately be described as a distinction between two ways of 'looking at' or 'talking about' living beings. To me this result is quite convincing, and it is an important step. But I would like to point out some consequences 12

Von Wright (1994), p. 101, abstract.

Mind, Matter, and our Longing for the One World'

135

of the strategy that has led to this dissolution of the problem, consequences that open up new domains in need for clarification. How does the traditional opposition between mind and 'body' relate to the distinction between mind and 'matter' that von Wright is using (rightly, I think) in the title of his paper? In the sentence just quoted from his abstract, strictly speaking he should have said that the maiier-mind distinction can be viewed as a distinction between two ways of looking, not the body-mind distinction. Equally, when he goes on to say that the material way of looking at living beings relates their "overt behavioural reactions to intra-bodily causes and effects" his readers must be clear that the behavioural reactions as well as the intrabodily processes must not be described in the vocabulary we would normally use to describe the behaviour even of animals, because these terms carry with them an understanding of its meaning (the animal is hungry, it sleeps, sweats, etc., the immune system is activated against intruders). As has been noted above, what has emerged in our discussion as the 'material' part of our means for describing living beings is by no means identical either with our vocabulary for 'extended' things or with our vocabulary for 'bodily' (leibliche) things and happenings. Indeed, if one would try to consistently stick to a purely material, purely causal terminology in the description of living beings, one would find it quite a difficult thing to do. I myself suppose that it will turn out to be impossible for a biologist. It would be worthwhile to investigate in this respect the terminology actually used in the life sciences, including cognitive science. My impression is that this language is full of 'anthropomorphisms', i.e. terms for actions that, as we have seen, entail an understanding of the meanings of those actions. I suppose that the use of this terminology is unobjectionable, even mandatory. What I object to is only that it is often unintentional and is very seldom a subject of explicit reflection.13 In a parallel fashion, our discussion has made clear that what Descartes thought of as the separate domain or 'stuff' of res cogitans can no longer neatly be separated from the world of extended bodies. It is not a world of pure, unembodied meanings, of chains of 'thoughts' or 'mental events', because the expressions we need to speak about the meaning of actions are either expressions for actions themselves or have actions as their criteria. Talking about the living body entails much more than talking about matter and its changing states (namely: meaningful actions), and talking about mental states entails much more than talking about meanings (again: meaningful 'embodied' actions). So the terms 'behavioural reaction' and its contrast to 'meaning of a reaction' in the quoted sentence must not be mis13

I do not want to defend vitalism here. My point is not to introduce a quasi-physical 'life-force' as an additional ingredient of causal explanations in biology, but to decide for a descriptively adequate terminology. Cf. Schneider (1989) and Schneider (1995).

136

Hans Julius Schneider

read to express more Cartesianism than is actually left after von Wright's clarifications. Once more we find that the domain of the 'material' or 'physical' language of causality is very small indeed. That the natural scientist can retain his intellectual peace of mind and does not have to fear the encounter with a 'ghost in the machine' remains true. But we can now see that to speak about the 'absence of ghosts' has two meanings: It can mean the absence of mental states in a strictly causal, material kind of inquiry that is so restricted that its usefulness for biology and cognitive science can be questioned. But this 'absence of ghosts' can also point to the necessity to re-interpret the domain of res cogitans as not constituting a separate territory, distinct from the body, but as being concerned with the meaning-aspect of 'embodied' actions and animal behaviour. In this respect also, since Descartes' special kind of mental stuff has disappeared, no ghosts have to be feared. But if we want to have a science of living nature that is able to treat meaningful behaviour of animals and will for that reason include mental states, we will have to consciously and reflectively use an 'anthropomorphic' language. It might as a matter of fact be used already, but mostly, I suppose, without methodological reflection. It will be richer than that of causal, material investigation, but at the same time it must not suggest consciously planned, intentional actions at places where there are none.14 So I think that the next step on our agenda is not a progressive refinement of the causal language of physics, chemistry, etc. until some day the mental will 'emerge'. A program of this kind would be an illusion. From the arguments developed above it appears to be much more promising to try to 'downgrade' our language of actions and mental states in such a way that the result (a consciously and deliberately modified sense of the relevant terms like e.g. 'to desire', 'to believe', etc.) allows us to speak of the meaning of the behaviour even of primitive animals. Von Wright's discussion of the clear cases of causally manipulating material systems versus consciously acting according to goals and beliefs has immensely cleared the ground and has dissolved the mind-matter problem in its Cartesian form. It is very helpful indeed to be able to clearly distinguish these two ways of speaking about living beings. But now we need to bring the extremes together again, not by refining our talk about brains to such a degree that we miraculously arrive at the mind, but by consciously 'loosening' our talk about actions 14

I agree with von Wright that what he calls 'quasi-teleological' explanations in biology are a kind of causal explanations. On the other hand I think he has given no reasons (and indeed may not demand) that all of biology must be causal. I myself think that at least for the description of higher animals a vocabulary for meaningful actions is unavoidable and quite in order, as long as it is kept in mind that it rests on a restricted form of communication. Cf. Schneider (1989).

Mind, Matter, and our Longing for the One World'

137

in such a way that it can consciously and perspicuously be used to cover the meaningful behaviour of animals or even of (parts of) our own bodies, for instance in cases of so called 'psychosomatic reactions'.15 These steps, I believe, will also strengthen our awareness that in many respects it is One world' that all living beings inhabit and that we should help to preserve.

References GATZEMEIER (1989). Matthias Gatzemeier (ed.): Verantwortung in Wissenschaft und Technik. Mannheim 1989. GOODMAN (1978). Nelson Goodman: Ways of Worldmaking. Hassocks 1978. MITTELSTRASS/RIEDEL (1978). Jürgen Mittelstraß and Manfred Riedel (eds.): Vernünftiges Denken. Berlin 1978. SCHNEIDER (1978). Hans J. Schneider: Die Asymmetrie der Kausalrelation. Überlegungen zur interventionistischen Theorie Georg Henrik von Wrights. In Mittelstraß/Riedel (1978). SCHNEIDER (1989). Hans J. Schneider: Anthropomorphes versus anthropozentrisches Denken. Zur ethischen und wissenschaftstheoretischen Bedeutung einer Unterscheidung. In Gatzemeier (1989). SCHNEIDER (1995). Hans J. Schneider: Wie kommt Geistiges zur Sprache? Ethik und Sozialwissenschaften 6 (1995). SCHNEIDER (1996). Hans J. Schneider: "Den Zustand meiner Seele beschreiben" - Bericht oder Diskurs? Deutsche Zeitschrift für Philosophie 44 (1996). SCHNEIDER (forthcoming). Hans J. Schneider: Handlung - Verhalten - Prozeß: Skizze eines integrierten Ansatzes. To appear in: Straub/Werbik (forthcoming)· STRAUB/WERBIK (forthcoming). Jürgen Sträub and Hans Werbik (eds.): Handlungsbegriff und Handlungserklärung. VON WRIGHT (1971). Georg Henrik von Wright: Explanation and Understanding. London 1971. VON WRIGHT (1994). Georg Henrik von Wright: On Mind and Matter. Journal of Theoretical Biology 171 (1994).

15

Cf. Schneider (forthcoming).

JOACHIM SCHULTE Willing and Acting

Abstract: In spite of forceful counter-arguments known from the writings of Wittgenstein and von Wright does it seem possible to defend a certain view held by advocates of the notion of pure and/or isolable acts of will. However, in the light of further considerations based on insights also to be found in Wittgenstein and von Wright the claims of those advocates of pure volitions or acts of will can be seen to lose most of their initial plausibility. 1 2 3

Forces behind Actions Irrational Motives, Passions Struggles

1

Forces behind Actions

When beginning to reflect on problems of human action, and especially on problems concerning free human action, we want to become clearer about the processes we feel to be behind such action. Wanting to understand these assumed processes we try to abstract from our own cases, analyse what we find into separable types of occurrences and operations, and use our findings to draw a picture of the mental mechanism regarded as responsible for our actions. After further reflection and reading up on what other people have found out about these matters we may arrive at a picture of the following kind: "We are now in a position to describe what happens in deliberate action, or when the mind is the seat of many ideas related to each other in antagonistic or in favorable ways. One of the ideas is that of an act. By itself this idea would prompt a movement; some of the additional considerations, however, [... ] block the motor discharge, whilst others [... ] solicit it to take place. The result is that peculiar feeling of inward unrest known as indecision. Fortunately it is too familiar to need description, for to describe it would be impossible. As long as it lasts, with the various objects before the attention, we are said to deliberate; and when finally the original suggestion either prevails and makes the movement take place, or gets definitively quenched by its antagonists, we are said to decide, or to utter our voluntary fiat in favor of one or the other course. The reinforcing and

140

Joachim Schulte

inhibiting ideas meanwhile are termed the reasons or motives by which the decision is brought about." This splendid characterization of what happens before we get down to doing something is by William James.1 Several elements of this account have been discussed by Wittgenstein, and von Wright refers to a picture of this kind when, in his lectures "On Human Freedom" he sketches what he calls "the traditional Freedom-of-the-Will problem". At the beginning of these lectures von Wright observes that even the most permanent types of philosophical problems cannot help being modified by changes of the context in which they are posed. Thus, while one may claim that the problem of our freedom to act comes as near to being a perennial problem of philosophy as one can imagine, the way relevant questions are raised becomes different as our scientific and philosophic outlook and interests change. "For a long time it was customary to think that human action as overt manifestations of behaviour are caused by something called volition or acts of the will. Human freedom, it was then often said, just consists in this: that an agent's actions are determined by his will and not by external forces over which he has no control or power"2. This conception, von Wright points out, will promptly run into serious difficulties for the reason that we shall sooner or later have to ask ourselves: What about "the will itself? Are we free to will what we will? Or is the will determined by something else? If the will is not free, action determined by the will can be free at most in some relative sense" and so forth3. But it is not only the problems deriving from the uncertain, perhaps constantly misleading, nature of the will understood as a denizen of our minds and the real decision-maker behind our actions which induce von Wright to formulate his questions about human freedom in a non-traditional way. Even more weight is probably carried by von Wright's conviction that "The 'classical' way of posing the problem of freedom can be said to obscure the factors which are normally said to determine our actions, viz., the reasons we have for performing them" 4 . When wondering whether we are free to will what we will von Wright implicitly raises a question brought to the fore by Wittgenstein in his discussion of the will, which can usefully be read as a critical comment on the picture drawn by William James and briefly quoted at the beginning of this paper. In this discussion Wittgenstein emphasizes that our attempts at understanding the nature of the will are time and again frustrated by the fact that the will keeps behaving like a will-o'-the-wisp: we seem to see it quite Barnes (1890), Vol. II, p. 528. Von Wright (1985), I, § 1. 3 Von Wright (1985), ibid. 4 Von Wright (1985), ibid. 2

Willing and Acting

141

clearly but never manage to get hold of it. Two of the main reasons for our frustration and repeated running up against unexpected obstacles are, as Wittgenstein explains, first, our mistaken belief that willing itself is a kind of action and, second, the traditional picture behind our considerations which represents the antecedents of action as quasi-causal goings-on within a more or less mechanically understood mind-machine. Wittgenstein's criticism is memorably summarized in § 613 of his Philosophical Investigations: "In the sense in which I can ever bring anything about (such as stomach-ache through over-eating), I can also bring about an act of willing. In this sense I bring about the act of willing to swim by jumping into the water. Doubtless I was trying to say: I can't will willing; that is, it makes no sense to speak of willing willing. 'Willing' is not the name of an action; and so not the name of any voluntary action either. And my use of a wrong expression came from our wanting to think of willing as an immediate non-causal bringing-about. A misleading analogy lies at the root of this idea; the causal nexus seems to be established by a mechanism connecting two parts of a machine. The connexion may be broken if the mechanism is disturbed." To this one may of course want to object that one knows from one's own experience what the workings of the will are like. Even if it is correct to say that our descriptions of our volitions tend to be misleading, we still feel their force; we intensely feel the conflict between our acts of will and various obstacles and counter-tendencies; and above all we remember willing, wanting, desiring something. And what could possibly render my conviction of the existence of a given kind of mental act more certain than a series of clear memories of occurrences of such acts? This notion, however, is not sufficiently thought out. No doubt I remember willing, wanting, desiring this and that, but whatever I may be able to recall and vividly re-present to my inner eye is not an act of willing or wanting or desiring. I may see all sorts of pictures and feel all sorts of rekindled emotions, but I shall not see pictures of volitions, nor shall I feel pure acts of willing or desiring. It is with this sort of situation in mind that Wittgenstein writes: "Ί remember that I should have been glad then to stay still longer.' - What picture of this wish came before my mind? None at all. What I see in my memory allows no conclusion as to my feelings. And yet I remember quite clearly that they were there." (PI, § 651) Wittgenstein does not deny the reality of certain memory-phenomena; nor has he any wish to dispute the former presence of feelings which are now remembered. What he does deny is that these feelings are feelings of pure or somehow isolable volitions and that it is such pure or isolable acts which are now remembered. Of course, Wittgenstein does not stop short at mere denial. Reasons for regarding his denial as convincing are supplied by his numerous further remarks on our use of psychological concepts and, in particular, by his com-

142

Joachim Schulte

plex discussion of the achievements of introspection. The refusal to regard willing as an independently characterizable kind of action is based on a vast context of relevant considerations which may certainly be thought sufficient and persuasive. Here, however, I shall confine myself to vaguely reminding you of the existence of all these further considerations. Instead of going into them, I want to have a look at the possibility of continuing the argument between denier and advocate of volitions or acts of will.

2

Irrational Motives, Passions

Have we really seen the advocate of volitions at the end of his tether, as both Wittgenstein and von Wright seem to suggest, or is there a further move which he might make? It appears that such a further move is still open to him and, as it happens, he might take his inspiration from something von Wright says in the same lectures we already have quoted from. In these lectures von Wright makes a fair number of helpful distinctions while reminding us that these distinctions are not clearly marked in ordinary language and that, even as philosophers, "One must not be pedantic about the use of words". When he reaches the pair of concepts mentioned in the last sentence of my quotation from William James - namely, the concepts of reason and motive -, von Wright notes that a distinction between them might be made by means of referring to their different relations with rationality. As von Wright writes: "Motives have not the same link with the rational faculties of man that reasons have. Motives can be irrational. And irrational motives can prompt a man to act perfectly rationally for reasons"5. As examples of such motives (in contrast with reasons) von Wright mentions what he calls '"passions' such as jealousy, hatred, greed. They tend to 'move' people to action; under their influence people do various things. That a man, for example, hates another man will usually manifest itself in various 'ends' of action which he then pursues. He may want to inflict harm on the object of his hatred"6. Passions of this kind are, I take it, something whose force an agent cannot help feeling. If you are under influence of hatred or greed, you will have to combat against it if you wish to avoid acting out of hatred or greed. That is, passions like hatred or greed are forces which make themselves felt. There may be reasons or causes elsewhere - outside yourself - that could be blamed for your having these passions, but this is an ulterior question which at this point must be left out of account. But what can you do if you do not want to act in accordance with these passions? After all, people have been known to act virtuously. They have 5

Von Wright (1985), I, § 8. Von Wright (1985), I, § 8.

6

Willing and Acting

143

withstood greed and hatred, they have abstained from steeling their neighbour's cattle or gold and they have overcome their hatred and refrained from taking revenge. Notoriously these abstainings, overcomings, and refrainings do not happen without the taking place of a previous struggle. So, independently of ontological and other questions about the status of those "passions" we may - within the picture drawn by von Wright - take them to be forces against which we feel to have to battle if, instead of behaving thus "passionately", we want to act virtuously, rationally, or simply like gentlemen.

3

Struggles

Now, what are these struggles? What are they if not struggles of the will against passion, even if sometimes they figure as struggles of duty against inclination or good against evil? And these are, nota bene, questions, not about the real essence or the true nature of things, but about the way they appear to the person in question. It is the struggler himself who feels that he has to fight a power which is at work in his own bosom. And is that not a real feeling, a feeling which simply is the will combating against passion, evil, or what have you, just as a headache simply is a headache, even if the doctor can give you a long jaw about what the headache may, or actually is, due to? If a man wants to live up to certain standards, he will occasionally, and perhaps frequently, have to fight against something within himself. And these will certainly be occasions where you literally feel the will taking on an enemy on his own ground, that is, on the battlefield of the mind. To generalize a bit: wherever one wishes to live up to certain standards - no matter if it's Oscar Wilde's "blue china" which he said he found it difficult to live up to7, or the most heroic ends of Greek or Christian ethics - there will always be purely inner struggles which are experienced as the will fighting against some equally internal obstacle. The point I want to bring out may become even clearer if you consider entirely irrational cases, as is suggested by von Wright's remark about motives in contrast with reasons. This does not mean cases of evidently stupid mistakes or foolish miscalculations. What it means is cases of known absurdity, gross foolhardiness, even contradictoriness. In such cases, I suppose, your mind really is, as William James puts it, quite clearly the "seat" of a variety of ideas in conflict and you witness their coming to grips with one another. We all know, if not from our own experience, then from what the poets tell us, that on certain occasions one makes the present moment stand still, that one may want to fly; that one may want to make the dead live. 7

Ellmann (1987), p. 44.

144

Joachim Schulte

And what could be more absurd, in some cases even self-contradictory, than wishes of this kind? And still, we understand what is being said; and at least some of us know from their own experience what it feels like. And this is what I want to stress: here there seems to be a feeling which one can experience and remember, and we seem to understand that it is this feeling which is being reported when someone tells us that a struggle between the will and a passion or some obstacle has taken place. But here we have to be careful. No doubt, there are typical kinds of feeling generally associated with descriptions of a struggle between the will and another force, which may be represented as extremely strong ("My hatred/greed/jealousy etc. is stronger than myself. And this typical sort of feeling appears the more clear-cut the stronger the obstacle or the more the desired object borders on the impossible, in which case "the limits of possibility" from the obstacle one wishes to overcome. But does that truly mean that this typical sort of feeling, however real it may prove to be, is a pure and isolable feeling of an act of will in the way a feeling of headache simply is a headache? No, I do not think that it means that. But in order to see this, "the axis of reference of our examination must be rotated" as Wittgenstein puts it at one point (PI, § 108). Instead of looking into our minds as if we were observing struggling heroes and villains coming to grips with each other, we should have a closer look at what we are saying and see our words as "tools" as Wittgenstein says in another passage (PI, § 360). We do speak of conquering one's weaker self, of victories of the mind over the weakness of the flesh, and of wanting what is impossible; and we know perfectly well what is meant by these expressions, but we do not really expect a well-defined or definable inner process to correspond to these descriptions. By rotating the axis of reference we shall notice that, even in examining our own attitudes and propensities, we use expressions of that kind, not to identify a certain feeling, but to characterize a certain type of situation in which we find ourselves. This situation may be marked by more or less typical kinds of feelings, but the struggles mentioned are not struggles between inner processes but between objectives, reasons, principles and the agent's disposition, character, morals, morale, or what have you. We shall soon notice that what we regarded as the graphic character of a certain experience - an assumed pure and isolable feeling - is nothing but our graphic way of describing a certain kind of experience. But this "nothing but" has to be understood in the right way. It has two sides. On the one hand, it means that however much I stare at what William James calls the "seat of many ideas" I shall not be able to identify a pure volition. That is, by employing the introspective method usually recommended by advocates of pure acts of will I shall get nowhere near such acts.

Willing and Acting

145

If, on the other hand, I pay attention to the way we typically talk about the relevant situations, and especially about those situations which we, following von Wright, have graphically described as struggles between ourselves, or our better selves, and irrational motives or overwhelming passions, I have taken another step in the direction of seeing that, even though introspection may be inevitable, it rarely is a useful guide towards philosophical clarity.

References ELLMANN (1987). Richard Ellmann: Oscar Wilde. London 1987. JAMES (1890). William James: The Principles of Psychology. New York 1890. WITTGENSTEIN (1953). Ludwig Wittgenstein: Philosophical Investigations. Oxford 1953 (PI). VON WRIGHT (1985). Georg Henrik von Wright: On Human Freedom. In The Tanner Lectures on Human Values, Vol. VI, Salt Lake City 1985.

KRISTER SEGERBERG Results, Consequences, Intentions Abstract. Von Wright's distinction between results and consequences of actions is fundamental. As soon as it has been made, it seems obvious; nevertheless it is not easy to account for the distinction within the framework of a rigorous logical theory. In this paper we study von Wright's theory, as it is presented in Chapter III § 5 of Norm and Action, in something like a laboratory setting - a modelling that is highly abstract but, thanks to its very abstractness, helps to highlight some interesting questions about von Wright's theory. 1 2 3 4 5 6

Modelling Systems Agentless Systems One-Agent Systems: Results One-Agent Systems: Consequences One-Agent Systems: Intentions Envoi

1

Modelling Systems

Assume a room with one door and one window. There is a single agent in the room, but to begin with we will assume that he is asleep and so does not interfere with what goes on in the room. Suppose that, for some reason, all that interests us is whether the door and the windows are open or closed and what the indoor and outdoor temperature is. Then, abstracting from an indefinite mass of information, such as the supply of radiators, the position of the sun, and the agent resting on a couch, we may represent the stateof-the-world by a quadruple (d, w,iin>*out), where 0 for d or w represents Open' and 1 represents 'closed' and where fj n and tmt are integers giving the indoor and outdoor temperature in centigrade. Thus on a hot summer's day, even in Scandinavia, the representation might be (0,0,30,32), whereas (1,1, —8,20} would not be unusual in winter. Let us wax somewhat technical. If s is a quadruple, then we write s(z) for the (i + l)st coordinate. For example, (0,0,30,32)(2) = 30. A possible history or simply path of length ξ, where ξ is an ordinal, is a function from ζ into the set of possible states-of-the-world. Thus in effect a path is a sequence σ = ($η : η < ξ) of such quadruples. In human affairs we are normally interested in paths of finite length, although it would be possible to

148

Krister Segerberg

contemplate paths of length ω (the order type of the set of natural numbers). Our understanding is that a path σ = (SQ, . . . , s n _i) represents a possible development in the room with SQ, ..., s n _i representing η successive stages of that development; if η = 0, then σ is the path φ = ( ) of length zero, called the null path or the empty sequence.1 If σ is of finite, positive length we may write σ(#) for the last member of the sequence; if σ is empty or infinite the expression σ(#) is meaningless. If σ and τ are paths we write στ for their concatenation. We define an event as a set of paths. Notice that the null event 0 is different from the event {φ} - the latter has an element, viz., the null path, whereas the former has no element at all. With this terminology we can model, at least to some extent, a number of events in our system; some examples are listed below. Remember that we are talking about a causal system without any agent to infer with it. Hence events must be described in a way that does not imply agency. Say that the door is now open, next closed. Then the event that the door closes or, equivalently, the event consisting of the door closing (or the door's closing), has been instantiated. The event of the door closing: the set of paths (SQ, Si) such that So(0) = 0 and si (0) = 1. The event of the window opening: the set of paths (SQ, Si) such that so(l) = l andsi(l) = 0. The event of the door remaining closed: the set of paths (SQ, . . . , s n ), for any η > 0 such that, for all i < n, S{(0) = O.2 The event of the room cooling. This is more difficult because of the inherent vagueness. But for a reasonably clear example of a path in this event, take any (s 0 ,... ,s n ) such that So(2) = 25 and s n (2) = 18 and furthermore, for all i < n, Sj(2) = s t+ i(2) and Sj(0) = Sj(l) = 0 and Si (3) = 15. It is worth emphasizing that, just as most events are instantiated by indefinitely many sequences, so most sequences instantiate indefinitely many events. Just being presented with a certain development of the system does not guarantee any understanding what is happening. In other words, a sequence does not provide its own interpretation; many interpretations are possible, and which to choose or choose between is a matter of perception, a creative task carried out by an observer - automatically in some cases, painstakingly in others. We define a state-of-affairs as a set of states-of-the-world. In general not every set of states-of-the-world will count as a possible state-of-affairs, so we 1 2

Here and below we shall assume that η > 0. Would η > 0 instead of η > 0 be an acceptable condition?

Results, Consequences, Intentions

149

assume that there is a set S consisting of all states-of-affairs. A state-ofaffairs S is true of or holds for or is realized by a state-of-the-world s if s € S. There are two trivial states-of-affairs, the set of all states-of-the-world and the empty set; the former is always realized, the latter never. Suppose, for some event e, that there is a state-of-affairs S that is true of σ(#), for every path σ e e. Then we might say that e results in S or that S is α result of e. If the set R(e) = n{T € S : V