New Essays on the Explanation of Action

New Essays on the Explanation of Action

Forthcoming By Constantine Sandis A Companion to the Philosophy of Action (Wiley-Blackwell) (co-edited with Timothy O’Connor)

New Essays on the Explanation of Action Edited by

Constantine Sandis Oxford Brookes University and NYU in London

Editorial selection and matter © Constantine Sandis 2009 Chapters © their individual authors 2009 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6-10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2009 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin's Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN-13: 978–0–230–52202–2 hardback ISBN-10: 0–230–52202–5 hardback This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. 10 9 8 7 6 5 4 3 2 1 18 17 16 15 14 13 12 11 10 09 Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne

This volume is dedicated to Elizabeth Alice Sandis With all my love, in the year of our wedding

This page intentionally left blank

Contents Notes on Contributors

ix

Introduction Constantine Sandis

1

Part I Reasons and Causes 1 What Must Actions Be for Reasons to Explain Them? Fred Dretske

13

2 What Are Reasons for Action? Stephen Everson

22

3 Was Sally’s Reason for Running from the Bear that She Thought it was Chasing Her? Rowland Stout

48

4 Con-reasons as Causes David-Hillel Ruben

62

5

75

Agential Reasons and the Explanation of Human Behaviour Peter Hacker

6 Reasons as Non-causal, Context-placing Explanations Julia Tanney 7

Interpretative Explanations G. F. Schueler

94 112

8 Anscombe on Expression of Intention Richard Moran and Martin J. Stone

132

9

Can One Act for a Reason without Acting Intentionally? Joshua Knobe and Sean D. Kelly

169

Reasons: Explanatory and Normative Joseph Raz

184

11 Reasons, Desires and Intentional Actions Maria Alvarez

203

10

12 A Niggle at Nagel: Causally Active Desires and the Explanation of Action Charles Pigden vii

220

viii

Contents

13

Acting in Character Annette Baier

241

14 Aquinas on Action and Action Explanation Stephen Boulter 15 Acting for Reasons – A Grass Root Approach Ralf Stoecker Part II 16

257

276

Agency and Moral Psychology

Sub-intentional Actions and the Over-mentalization of Agency Helen Steward

295

17

Determinism, Intentional Action, and Bodily Movements Frederick Stoutland

313

18

Free Agency, Causation and Action Explanation E. J. Lowe

338

19 Gods and Mental States: The Causation of Action in Ancient Tragedy and Modern Philosophy of Mind Constantine Sandis

356

20 Aristotle’s Conception of Practical Thinking A. W. Price

384

21

Action in Moral Metaphysics Jonathan Dancy

396

22

Non-cognitivism and Motivation Nick Zangwill

416

Index

425

Contributors Maria Alvarez is a philosophy lecturer at the University of Southampton, United Kingdom. She has published widely on actions, reasons, and their relation. Her most recent book is Acts and Facts: Reasons, Desires and the Explanation of Action (Oxford University Press, 2009). Annette Baier was educated at the Universities of Otago and Oxford. She taught at the universities of Aberdeen, Auckland, Sydney, Carnegie-Mellon, and Pittsburgh, and held visiting appointments at Florida Gainesville, City University of New York, and Michigan. Since retiring she lives in Queenstown, her birthplace, and Dunedin. She has published much about Hume, on trust, and about the philosophy of mind. Her books are Postures of the Mind (1985), A Progress of Sentiments (1991), Moral Prejudices (1994), and The Commons of the Mind (1997). (The last were her Carus Lectures. She also gave Tanner Lectures, on trust.) She has just published a new collection of essays entitled Death and Character: Future Reflection on Hume. Stephen Boulter is Senior Lecturer in Philosophy at Oxford Brookes University, United Kingdom. Prior to taking up his current post he was Gifford Research Fellow in the Department of Philosophy at the University of Glasgow in 1998–1999. He is the author of The Rediscovery of Common Sense Philosophy (Palgrave Macmillan, 2007) and is currently working on a book on Medieval Philosophy. Jonathan Dancy is Professor of Philosophy at the University of Reading, United Kingdom, and at the University of Texas at Austin. Previously he taught for 25 years at the University of Keele. He has written many articles on moral philosophy and on the philosophy of action, and his books include Moral Reasons (1993), Practical Reality (2000), and Ethics Without Principles (2004). He is currently working on a new account of the force of practical deliberation. Fred Dretske is Professor Emeritus at both Stanford University where he taught for 10 years and the University of Wisconsin where he taught for 25 years. He is currently a senior research professor (no teaching) at Duke University, United States. He has several books to his credit: He is the author of several books: Seeing and Knowing (1969), Knowledge and the Flow of Information (1981), Explaining Behavior (1988), and Naturalizing the Mind (1995) as well as numerous articles. Some of these articles have been collected in Perception, Knowledge, and Belief (2000). His current research interests centre on the nature of conscious experience and problems about self-knowledge.

ix

x

Contributors

Stephen Everson has taught at the Universities of Oxford, Cambridge and Michigan, Ann Arbor. He is currently a lecturer at the University of York, United Kingdom. He has published on various topics in ancient philosophy, ethics, and the philosophy of action. Peter Hacker is currently Emeritus Research Fellow of St John’s College, Oxford, United Kingdom, where he was a Tutorial Fellow from 1966–2006. He is author of numerous books on the philosophy of Wittgenstein, including the 4 volume Analytical Commentary on the Philosophical Investigations (1980–1996), the first two volumes of which were co-authored with Gordon Baker, Insight and Illusion (1972, 2nd edition 1986), Wittgenstein’s Place in Twentieth Century Analytic Philosophy (1996), and Wittgenstein: Connections and Controversies (2001). He has also written extensively, together with Max Bennett, on philosophy and cognitive neuroscience: Philosophical Foundations of Neuroscience (2003), and A History of Cognitive Neuroscience (forthcoming 2008). His most recent work is Human Nature – the Categorial Framework (2007), which is the first volume of a projected trilogy on human nature. Sean D. Kelly is Professor of Philosophy at Harvard University, United States. His work focuses on various aspects of the philosophical, phenomenological, and cognitive neuroscientific nature of experience. This gives him a broad forum: recent publications have addressed, for example, the experience of time, the possibility of demonstrating that monkeys have blindsighted experience, and the understanding of the sacred in Homer. He has taught courses on twentieth century French and German Philosophy, Philosophy of Mind, Philosophy of Cognitive Science, Philosophy of Perception, Imagination and Memory, Aesthetics, and Philosophy of Literature. He also runs the Philosophical Psychology Lab at Harvard, and has received fellowships from the Guggenheim Foundation, the NEH, the NSF and the James S. McDonnell Foundation, among others. Joshua Knobe is Assistant Professor, Department of Philosophy, at University of North Carolina-Chapel Hill, United States. He works primarily in experimental philosophy and is especially interested in the ways in which people’s moral judgments impact their use of various apparently non-moral concepts. His research thus far has investigated the concepts of intentional action, causation, and consciousness, as well as the distinction between doing and allowing and the practice of reason explanation. E. J. Lowe is Professor of Philosophy at Durham University, United Kingdom, specialising in metaphysics, the philosophy of mind and action, the philosophy of logic and language, and early modern philosophy. Author of, amongst other things, Kinds of Being (1989), Subjects of Experience (1996), The Possibility of Metaphysics (1998), An Introduction to the Philosophy of Mind (2000), A Survey of Metaphysics (2002), Locke (2005), The Four-Category Ontology (2006), and Personal Agency (2008).

Contributors

xi

Richard Moran is the Brian D. Young Professor of Philosophy at Harvard University, United States. He is the author of Authority and Estrangement: An Essay on Self-Knowledge (2001), and various articles the on philosophy of mind, action, aesthetics, and moral psychology. Charles Pigden was born in England and studied philosophy at Cambridge, before going on to do a PhD at La Trobe University, Melbourne, Australia. Since 1988 he has taught philosophy at Otago University, Dunedin, New Zealand. He has published on a wide range of topics from conspiracy theories to the reality of numbers, but, if pressed, will admit to being a meta-ethicist with special interests in Russell, Moore, and Hume. He is the author of the entry on Russell’s Moral Philosophy in the Stanford Online Encyclopedia of Philosophy and is one of the very few academics to have published a philosophical dialogue in blank verse. A. W. Price is Reader in Philosophy at Birkbeck College, University of London, United Kingdom. He works on ethics and moral psychology, ancient and contemporary. He is the author of Mental Conflict (1995), Love and Friendship in Plato and Aristotle (extended edition 1997), and Contextuality in Practical Reason (2008). Joseph Raz is professor at Columbia Law School, United States, and a Research Professor University of Oxford, United Kingdom. His books include The Authority of Law (1979), The Concept of a Legal System (2nd edn 1980), The Morality of Freedom (1986), Practical Reason and Norms ( 2nd edn 1990), Ethics in the Public Domain (1994), Engaging Reason (2000), Value Respect and Attachment (2001), and The Practice of Value (2003). He is also a Fellow of the British Academy and a Foreign Honorary Member of the American Academy of Arts and Sciences. David-Hillel Ruben has a BA in Philosophy from Dartmouth College, United States and a PhD from Harvard University. He is the author of Marxism and Materialism (1977 and 1979), The Metaphysics of the Social World (1985), Explaining Explanation (1990), Action and its Explanation (2003), numerous edited volumes, and many journal articles. Ruben was Professor of Philosophy for many years at the London School of Economics and is now Director of New York University in London and Professor of Philosophy at Birkbeck, University of London, United Kingdom. Constantine Sandis is a senior lecturer in Philosophy at Oxford Brookes University and New York University in London. He works on the philosophy of action and its explanation and is the author of numerous related articles and a forthcoming book. He is currently co-editing, with Timothy O’ Connor, A Companion to the Philosophy of Action. G. F. Schueler is Professor of Philosophy and Chair of the Philosophy Department at the University of Delaware, United States. He is the author of Reasons and Purposes (2003) and Desire (1995).

xii Contributors

Helen Steward did her graduate work at the University of Oxford before becoming a Fellow of Balliol College, Oxford in 1993. She moved to the University of Leeds in 2007. Her first book, The Ontology of Mind, was published in 1997. She is currently working on topics in the philosophy of action, specifically as they relate to the free will debate, and her next book, A Metaphysics for Freedom will be out in 2009. Ralf Stoecker is professor for philosophy, particularly applied ethics, at the University of Potsdam, Germany. Stoecker studied philosophy in Hamburg, Heidelberg and Bielefeld. He wrote a dissertation on the topic of events (Was sind Ereignisse?, Berlin, New York 1992) and a Habilitation on the brain death debate and its moral and metaphysical bearings (Der Hirntod, Freiburg 1999). His areas of specialisation are applied ethics, philosophy of personhood and action theory. Since the 1990s Stoecker published papers on various aspects of human action, in order to establish a radical non-standard account of agency that attempts to combine Davidson’s insights with those from, for example, Ryle and Wittgenstein. Martin J. Stone teaches Law at Cardozo Law School, Yeshiva University and Philosophy at the New School Graduate Faculty in New York, United States. He is the author of various articles on Wittgenstein, action, interpretation, and philosophy of law. Rowland Stout is Senior Lecturer in Philosophy at University College Dublin, Ireland. Prior to this he was based at Oxford then Manchester University. He works on problems in the philosophy of agency and the philosophy of mind, in particular on rationality, causation and emotion. He is the author of Things that Happen Because they Should (1996), Action (2005), and The Inner Life of a Rational Agent (2006). Frederick Stoutland is Professor of Philosophy Emeritus at St. Olaf College, Minnesota, United States and Permanent Visiting Professor at Uppsala University, Sweden. His main interests are in philosophy of mind and action. His book on Recent Theories of Meaning was published in German in 1997. He has published papers and reviews on various issues in philosophy of action, on aspects of von Wright’s life and philosophy, on Davidson’s philosophy, on Wittgestein, and on truth. Julia Tanney was educated at UCLA and the University of Michigan. Now living in Paris, she is a Senior Lecturer and presently Head of Philosophy at the University of Kent, United Kingdom and holds visiting and guest professorships at the Université de Picardie (Amiens) and the Université de Paris-IV (Sorbonne), and is on the Philosophy planning board at the Université de Paris-I (Sorbonne-Panthéon). She has written numerous articles in philosophy of mind, focussing especially on reason explanation, normativity, rule-following, and self-knowledge. While continuing to insist on the relevance of the later Wittgenstein’s philosophy to the contemporary debate,

Contributors

xiii

she has also contributed a number of pieces (in both French and English) – and is presently writing a book – on Gilbert Ryle’s philosophy of mind and language. Nick Zangwill is Professor of Philosophy at Durham University, United Kingdom. He is the author of The Metaphysics of Beauty (2001) and Aesthetic Creation (2007) as well as many papers on moral philosophy and philosophy of mind.

This page intentionally left blank

Introduction Constantine Sandis

The essays collected here are new not only in the sense that they are previously unpublished but, more importantly, in representing the latest thoughts of a growing stream of philosophers whose recent work has challenged some of the most popular ways of thinking about action and its explanation. While many of these dominant conceptions have roots in ancient, medieval, and early modern philosophy, their most direct inspiration is Donald Davidson’s groundbreaking 1963 article ‘Actions, Reasons, Causes’, itself strongly indebted to C.J. Ducasse’s ‘Explanation, Mechanism and Teleology’ (1925) and Carl G. Hempel’s ‘Rational Action’ (1961–62). The latter challenged the ‘strong neo-Wittgensteinian current of small red books’ and similar-minded works1 by arguing that (a) actions are events, (b) the reasons that ‘rationalize’ action are causes of the events in question and (c) the explanation of action makes reference to known, strict, psychophysical laws. In proposing an account of action individuation (and to some extent also of intention) that was heavily indebted to G.E.M. Anscombe,2 while also denying the truth of (c), Davidson may be seen as making a concession to the enemy. Be that as it may, his influential defense of (a) and (b) was so immense that it has come to represent a crucial turning point in philosophical history, as instructively captured by Julia Tanney in her contribution to this volume. Notwithstanding a continued flux of books that denied these claims,3 the Davidsonian view prevailed. In due course it came to be modified and developed in a variety of conflicting directions by Davidson himself as well as by naturalist philosophers of mind as diverse as David Armstrong, Michael Bratman, Myles Brand, Berent Enç, Jerry Fodor, Alvin Goldman, Ernest LePore, Kirk Ludwig, Al Mele, Hilary Putnam, R. Jay Wallace, and what at some point began to seem like half the philosophical world,4 infiltrating moral philosophy with the help of Michael Smith’s landmark 1987 paper ‘The Humean Theory of Motivation’ which would later play an important structural role in his The Moral Problem (Oxford: Blackwell, 1995).

1

2 Introduction

Both Davidson’s own thesis and the various views that it began rapidly evolved through rational selection to form what many (including Fred Dretske and Ralf Stoecker in their contributions to this volume) have come to call the ‘Standard View’. The view in question has two distinct parts: a theory of action and a thesis concerning the reasons for which actions are performed. The first maintains that actions are events that are identical to movements of the body caused, in a ‘non-deviant’ way, by a combination of beliefs and so-called ‘pro attitudes’; the second states that the primary reason for which an intentional action was performed is whichever combination caused the bodily movement in question. Each of the essays collected here focuses on one or more aspects of the Standard View. Many of them represent the culmination of an important fin de siècle movement of dissent. This included Rowland Stout’s Things that Happen Because they Should (1996), Bede Rundle’s Mind in Action (1997), Helen Steward’s The Ontology of Mind (1997), Jonathan Dancy’s Practical Reality (2000), Timothy O’ Connor’s Persons and Causes (2000), Paul Pietroski’s Causing Actions (2000), Rüdiger Bittner’s Doing Things for Reasons (2001), Jennifer Hornsby’s Simple Mindedness (2001), John Searle’s Rationality in Action (2001), David-Hillel Ruben’s Action and Its Explanation (2003), and G.F. Schueler’s Reasons and Purposes (2003), as well as additional papers by Arthur Collins, Giuseppina D’Oro, Dan Hutto, Bill Pollard, Josep Lluis Prades, Severin Schroeder, David Velleman and most of the contributors to this book (at least four of which are currently working on related monographs). Be that as it may, the essays are by no means the product of a single school of thought. Indeed, one cannot overemphasise the variety of backgrounds, stances, approaches, and methodologies that these contributors have brought with them. To give but a few examples, these include concerns relating to conceptual analysis and clarification (Alvarez, Everson, Hacker, Stoecker, Raz and Tanney), experimental philosophy (Knobe and Kelly, and Pigden), exegesis (Baier, Boulter, Moran and Stone, Price and Sandis), naturalism (Boulter, Dretske and Pigden), substance dualism (Lowe), and language use (Dancy, Dretske and Hacker). Nor are all of the contributors set on destroying the Standard View. Some merely wish to modify it, while others (such as Boulter and Pigden) defend aspects of it against criticism. Many bear testament to a recently revived interest in Anscombe’s work on the subject. The book divides into two named sections, each encompassing a further narrative of its own. The first deals primarily with the relation between actions and our reasons for performing them, the second with issues in moral psychology, focusing primarily on questions of control, agency and motivation. What follows is a brief overview demonstrating how these and other themes interrelate. Fred Dretske’s ‘What Must Actions be for Reasons to Explain Them?’ opens the volume with the important, but often neglected question, of what kinds of things actions must be if reasons are to explain them. He

Constantine Sandis

3

discuss various linguistic reasons for identifying actions with causal processes that (typically) have bodily movements as their result, contrasting it with the Davidsonian view (defended by Enç among others) that actions are these bodily movements, and concluding that this is the only way to give reasons explanatory purchase, thus also emphasising the dependence of a theory of action explanation on a theory of action: ‘if philosophers cannot agree about what action is, they cannot hope to understand the power of reasons to explain it’. Dretske’s analysis leads him to conclude that the reasons for which we act are to be identified with what we believe and desire, as opposed to our believing and desiring. In so doing he joins Raz, Dancy, Everson and others who form part of what the latter calls the ‘growing trend amongst philosophers of action to recognise that a reason for action is something that favours or makes valuable an action of the relevant kind, and that an action is not made valuable by the agent’s merely wanting to do it or believing that it would be good to do it’. This insight (whose flipside is Bernard Williams’ claim that ‘if there are reasons for action, it must be that people sometimes act for those reasons, and if they do their reasons must figure in some correct explanation of their action’5) invites questions regarding the ontological status of the so-called ‘contents’ of our beliefs and desires. It is with this in mind that in his essay ‘What Kind of Things are Reasons for Action?’ Everson revisits the debate between J.L. Austin and Peter Strawson on the differences and similarities between propositions, facts and states of affairs. He concludes that while ‘the concern to determine what reasons there are for some course of action is the concern to make sure that all relevant facts are taken into account’ which should not let this mislead us into thinking that the aim here is ‘to know how many reasons there are’. To think otherwise is to conflate two different notions of reason: a mass-notion and a count-notion. Stout’s ‘Was Sally’s Reason for Running from the Bear that She Thought it was Chasing Her?’ addresses a related worry concerning false belief that cuts across these ontological matters: if Sally believes falsely that the bear is chasing her, in what sense can the ‘content’ of her belief, however we wish to characterize it, be said to be her reason for running from it? Stout’s answer is that in such cases Sally is mistaken to think that she ran for the reason that she thought she had but that this does not imply that she did not run for a related explanatory reason, for example, because there was a noise in the undergrowth (which she rightly took as evidence for the belief that a bear was chasing her). This answers may be contrasted to Raz’s suggestion (in his contribution to this volume) that ‘when the belief is false ... as there is no reason making that belief true, no reason can be part of the explanation’. The issues that Everson and Stout’s essay devote themselves to are further discussed in the essays by Dancy, Hacker, Raz, and Stoecker. Another question one might ask about the reasons for which we act is whether or not these are best conceived of as causes of action. Peter Hacker, David-Hillel

4 Introduction

Ruben, G.F. Schueler and Julia Tanney all answer this question negatively. In ‘Con-reasons as Causes’ Ruben does so by introducing the concept of a ‘con-reason’ (a reason we have but which we do not act upon), persuasively arguing that if such reasons have no relevant causal effects then, given that they are ontologically identical to pro-reasons (reasons we act upon), the latter cannot have any either. The essays by Hacker, Tanney, and Schueler complement Ruben’s by offering various non-causal accounts of how actionexplanation works. These invariably revolve around the practice of placing action in intelligible contexts, but resist the notion that such contextplacement is in any way nomological. Thus, explanations of human behaviour in terms of agential reasons (Hacker outlines five other varieties) are what Tanney, in an essay which like Hacker’s also defends Wittgensteinian non-causalism, calls ‘non-causal context-placing explanations’. Likewise, Schueler maintains that such ‘interpretive explanations’ (to use his own term for them, which is also the title of his essay) ‘are both central to explanations of human action and irreducibly different in form from other commonsense explanations of events, as well as from explanations found in paradigm ‘hard’ sciences such as physics’, concluding that it is consequently ‘a mistake to think that interpretative explanations are somehow reducible to (or explicable in terms of) causal explanations’. One of the most common methods of rendering action intelligible is by redescribing it in a way that reveals the agent’s intention(s). This naturally leads to the task, famously associated with G.E.M. Anscombe, of providing an analysis of how to best capture the relation between action and intention. In ‘Anscombe on the Expression of Intention: An Exegesis’ Richard Moran and Martin J. Stone focus on some cryptic yet highly influential remarks by Anscombe. They work through various possible interpretations of what Anscombe might have had in mind, comparing her insights to those of Wittgenstein, Ryle, Davidson and Bratman, before settling for an account according to which Anscombe does not make the traditional behaviourist move of denying that intention is a mental state but rather denies only ‘that we understand how to apply the notion of a “state” here, on the basis of its application in other contexts like those of belief and desire’. In stressing, instead, the conceptual unity of ‘pure’ intending, intentional action, and intention-in-action, her achievement was to demonstrate that ‘no psychology will afford the right materials for explaining action which does not make use of a concept which applies throughout the spectrum of unfolding action, and which thus has same internal complexity as actions themselves’. Accordingly, Anscombe claimed that all action performed for a reason was intentional under some description (Intention, § 5) thus influencing Davidson6 who characteristically went further, adding that (conversely) all intentional action was done for reasons (a claim rejected by Anscombe and later also by Rosalind Hursthouse7). In ‘Can One Act for a Reason Without Acting Intentionally?’ Joshua Knobe and Sean D. Kelly apply the method of

Constantine Sandis

5

experimental philosophy with the aim of demonstrating that people’s intuitions do not conform to Anscombe’s thesis (that a behaviour cannot be performed for a reason unless that behaviour is performed intentionally). They next propose a system of principles that provides a better match for people’s intuitions. While they refrain from judging whether or not people’s intuitions in all these cases are actually correct they nonetheless take this to shift the burden of proof against Anscombe’s thesis. The somewhat weaker claim that ‘acting with an intention or a purpose is acting (as things appear to one) for a reason’ is defended by Joseph Raz in ‘Reasons: Explanatory and Normative’. As Raz notes, ‘while all actions with a purpose or intention are intentional actions, not all intentional actions can sensibly be said to be actions done with an intention’. In his essay he also argues that the normative reasons for which we act (which he identifies with facts) are also explanatory, their possession of the latter property being enabled by their possession of the former. In ‘Reasons, Desires and Intentional Actions’ Maria Alvarez distinguishes both normative and explanatory reasons from the reasons for which one acts (which, following Smith, Dancy, et al., she calls ‘motivating reasons’). Alvarez allows that a motivating reason may be both normative and explanatory, but argues that it need not be so. On her account, a reason for which I act need not explain why I act, nor need it be a reason for which I ought to act. She also denies, contra Raz, that a reason for which one acts that explains why the person acted need do so in virtue of (also) being a normative reason. Her main target, however, is the Humean view that ‘wanting something is part of the reason for which one acts’. While desire, on her view, is capable of explaining and/or motivating action, it cannot do so in virtue of being (even part of) a motivating reason for action since in all such cases the latter will be what she calls the desirability characterization of the action (viz. my reason for desiring to perform it), a point she traces back, through Anscombe, to the ancient and mediaeval view that what is wanted is always wanted under the aspect of the good. The weaker thesis that at least some desires are held for reasons forms part of Thomas Nagel’s famous argument for the view that beliefs can motivate alone (by producing ‘motivated desires’). In his ‘A Niggle at Nagel: Causally Active Desires and the Explanation of Action’ Charles Pigden defends a standard Humean position (not unlike the one attacked by Alvarez) from Nagel’s argument and the various uses it has been put to by Foot, McDowell, Cullity and Gaut, and Dancy. He does so by rejecting the suggestion that the alleged (conceptual) truth that ‘if I do X because I believe that it is likely to bring about Y, then I am acting out of a desire for Y’ is incompatible with the claim that the desire in question is causally efficacious. A separate question is whether Hume was himself a Humean about motivation (in any interesting sense). In her essay ‘Acting in Character’ Annette Baier describes a tension between Hume’s official account of

6 Introduction

action explanation and the numerous explanations of actual actions that he offers in his Histories. Its resolution, she maintains, lies in the realization that Hume’s conceptions of a ‘sentiments’ and ‘passions’ are far more inclusive than the modern notion of a pro-attitude ‘since they cannot all be characterized as “pro” or “con” something’. Baier also demonstrates that Hume interpreters have underestimated the explanatory role played by character in his overall system, arguing ‘that explanations that cite character to explain action do not necessarily reduce to belief-desire explanations’. So it is that she concludes that we would do best to abandon proto-Davidsonian readings of Hume, a far better modern counterpart being someone like (the non-reductive) David Velleman. Equally striking is Stephen Boulter’s account of ‘Aquinas on the Explanation of Action’. According to Boulter, Aquinas was a sophisticated causalist with a sufficiently rich account of the production of action to respond to the contemporary challenges to causalism raised by ‘deviant causal chains’ (discussed by Jonathan Lowe in his contribution to this volume) and the ‘normative constraint’ introduced above in relation to Dretske. Indeed, the overall account of action and its explanation presented here is not dissimilar to that of Dretske, although Aquinas’ hylomorphism also hints at the more ‘naïve’ naturalism of Jennifer Hornsby and John McDowell. Having explored a variety of issues and views relating to reasons and causes the first (and largest) half of the volume closes with a ‘back to basics’ contribution, Ralf Stoecker’s ‘Acting for Reasons – A Grass Root Approach’. Stoecker invites us to abandon our philosophical preconceptions and return to the questions we began with namely (1) what is it to act? and (2) what is it to act for reasons? Like Dretske, Stoecker rejects the Standard View of agency. He replaces it with the following, tentative, proposal: ‘To say that a person acts is to claim that a particular fact can be explained with recourse (i) to the person’s ability to behave as if she were constantly engaged in a public practical deliberation about what to do, and moreover (ii) by giving some of the reasons for which she acts, i.e. by mentioning arguments the person could have put forward in the virtual deliberation that she is acting on’. In so doing he paves the way for the discussions of agency, deliberation and responsibility that constitute the second half of the volume. The first of these, Helen Steward’s ‘Sub-intentional Actions and the Over-mentalization of Agency’ denies ‘that it is a necessary condition of something’s being an action that it be associated with a reason-giving explanation’ or indeed with any kind of psychological explanation. To think otherwise, Steward maintains, is to over-intellectualize agency, typically by overmentalizing it. Doing so runs the dual risk of underestimating the extent to which it should be conceived of as an animal power and overestimating our understanding of how it is that we have this basic ability to control certain movements of our bodies, a phenomenon with regard to which she wishes to remain agnostic.

Constantine Sandis

7

Rational explanation of those bodily movements involved in intentional action is, according to Frederick Stoutland, prior to neurophysiological explanation (of the same movements). In ‘Determinism, Intentional Action, and Bodily Movements’ he argues that we must consequently reject the thesis of the completeness of physics which states that all situations are governed by precise laws (be they deterministic or probabilistic). While this thesis is irrelevant to rational explanation per se, it is ruled out by Stoutland’s account of the relation between rational and neuroscientific explanation. In contrast to Stoutland, who sees intentional action as the fundamental issue here, in his essay ‘Free Agency, Causation, and Action Explanation’, Jonathan Lowe focuses on a threat which he takes the so-called ‘scientific world-view’ to pose to free agency. In response, he argues that intentional actions are ‘neither mere chance occurrences nor events that are wholly causally determined by prior events’, offering instead an account of free agency that relies on a notion of human intervention that is irreducible to event causation. Rejecting classical agent causalism as ‘an unstable half-way house’ Lowe favours the view that ‘all causation is fundamentally substance causation’. My own essay, ‘Gods and Mental States: The Causation of Action in Ancient Tragedy and Modern Philosophy of Mind’ also rejects the ‘scientific world-view’ typically found in naturalist philosophy. In it I argue that ancient tragedy presents us with an alternative understanding of human agency that, contrary to popular opinion, is preferable to both the Standard View and the various accounts of identification put forward by Frankfurt et al. Aristotle’s understanding of the relation of thought to action is the theme of Anthony Price’s essay ‘Aristotle’s Conception of Practical Thinking’. Price begins by outlining what Aristotle meant by an ‘intellect which reasons for the sake of something and is practical’ before asking whether we should conceive of deliberation as a kind of reasoning that is inherently practical and investigating how an action can stand as the conclusion of practical thinking. His interpretative aim is to demonstrate the plausibility of Aristotle’s work, thus motivating the conclusion that ‘there would be more reason to regret what he got wrong if philosophers had not often lost sight of what he got right’. An important subset of practical evaluation is moral evaluation. In ‘Action in Moral Metaphysics’ Jonathan Dancy brings action explanation to territory typically occupied by philosophers working in moral psychology. He asks what kinds of things actions must be if we are to at least be capable of relating to them morally, turning next to the question of what this tells us about explanation of both moral and non-moral action. Dancy defends a ‘deflationary’ account of actions, according to which actions, unlike events, ‘should not have identity criteria’. This is not to say, however, that there are no such things called ‘actions’ which people do for reasons (moral or otherwise), but only that these are not independent bearers of evaluative and explanatory properties.

8 Introduction

The volume closes with Nick Zangwill’s ‘Non-cognitivism and Motivation’, written in the Humean tradition of bridging the gap between moral psychology and meta-ethics. Zangwill, however, is no Humean. Indeed, he appeals to the phenomenon of ‘variable motivation’ – namely the fact that ‘the same moral judgements move different people differently, and they move the same person differently on different occasions’ – in support (through an inference to the best explanation) of cognitivist externalism with regard to moral motivation. In so doing, he also demonstrates just how far reaching the philosophical significance of work on action explanation can be. I would like to express the deepest of thanks to my co-contributors for their wonderful work, their enthusiasm and their helpful suggestions. In addition, I would like to thank Daniel Bunyard, Priyanka Pathak and Melanie Blair at Palgrave Macmillan: Dan for commissioning this volume and encouraging me to go ahead with it; Pri and Melanie for their subsequence patience and explanatory assistance. Thanks also to the three anonymous referees who read my original proposal and made numerous invaluable points and recommendations, and to Nafsika Athanassoulis, David Oderberg and John Shand for sound advice based on their own editing experiences. I’m also extremely grateful to Robert Vinten for preparing such a wonderful index and to Vidhya Jayaprakash and her team for all their hard work in production. Finally, I would like to thank The National Gallery of Ireland (and especially Camille Lynch) for the permission to reproduce Vermeer’s Woman Writing a Letter, with her Maid on the front cover. Never sold during Vermeer’s lifetime, it has apparently been stolen twice over the past thirty-five years, once by the IRA. Art critic Mark Harden writes that ‘the placid scene with its muted colours suggests no activity or hint of interruption’, but I prefer to think that the maid is observing someone through the window who has either just departed or is expected to arrive at any moment, as indicated by the empty seat. Perhaps this person is the cause and/or recipient of the letter that the woman is writing, no doubt with the intention of communicating reasons of some kind (an enigmatic still-life on the floor). CS Oxford, February 2008

Notes 1. ‘Hempel on Explaining Action’ (1976), reprinted in Davidson’s Essays on Actions and Events (Clarendon Press, 1980, p. 261). Most of the works in question are referred to in the first footnote of ‘Actions, Reasons, Causes’ (op. cit., p. 4). Some (most notably William Dray’s) were inspired by R.G. Collingwood as much as by Wittgenstein. 2. Cf. ‘Agency’ (1971) reprinted in Davidson 1980 (op. cit. p. 59, fn. 19) for the account of action individuation upon which the claim that whether or not an action is

Constantine Sandis

3.

4.

5. 6. 7.

9

intentional is a matter of description – as defended in ‘The logical Form of Action Sentences’ (1967, in op. cit., p. 121) – is based and ‘Actions, Reasons, Causes’ (op. cit., p. 5), for the related claim that all action done for reasons is intentional (under some description), discussed further below in relation to Knobe and Kelly’s essay. Davidson held on to these beliefs, all inherited from Anscombe, even after he came to reject her account of what it is to have an intention in 1978’s ‘Intending’ (1978, in op. cit., p. 83ff). Cf. Charles Taylor’s The Explanation of Behaviour (London: Routledge and Kegan Paul, 1964), A.R. Louch’s Explanation and Human Action (Oxford: Blackwell, 1966), Richard Taylor’s Action and Purpose (Prentice-Hall, 1966), R.S. Peters’ The Concept of Motivation (Routledge and Kegan Paul, 1967), Paul D.G. Brown’s Action (1968), G.H. von Wright’s Explanation and Understanding (Routledge and Kegan Paul, 1971), Rom Harré and Paul F. Secord’s The Explanation of Social Behaviour (Blackwell, 1972), as well as papers by G.E.M. Anscombe (whose complex work had also influenced Davidson), Roderick Chisholm, Anthony Kenny, Keith Lehrer, A.C. MacIntyre, Norman Malcolm, John McDowell, A.R. White, and many others. See, for example, Jerry Fodor’s Psychological Explanation: An Introduction to the Philosophy of Psychology (New York: Random House, 1968), Alvin Goldman’s A Theory of Human Action (Princeton, 1970), Myles Brand’s Intending and Acting: Toward a Naturalized Action Theory (MIT, 1984), Michael Bratman’s Intention, Plans, and Practical Reason (Harvard, 1988), Al Mele’s Springs of Action (OUP, 1992), and Berent Enç’s posthumous How We Act: Causes, Reasons, Intentions (OUP, 2003). Bernard Williams, ‘Internal and External Reasons’, reprinted in Moral Luck (Cambridge: CUP 1981) p. 102; quoted in Joseph Raz’s essay in this volume. See note 2 above (and all related passages). G.E.M. Ancombe, Intention (Oxford: Blackwell, 1957, § 17) and R. Hursthouse, ‘Arational Actions’, Journal of Philosophy, 1991, LXXXVIII, 2, pp. 57–68.

This page intentionally left blank

Part I Reasons and Causes

This page intentionally left blank

1 What Must Actions Be for Reasons to Explain Them? Fred Dretske

If philosophers cannot agree about what action is, they cannot hope to understand the power of reasons to explain it. Without such an understanding, the place of mind in a material world remains utterly mysterious.

1 The problem In a recent book, Berent Enç (2003) identifies actions with changes in the world that are caused by mental events. When reasons cause bodily movements or an external change of some sort in the right (nondeviant) way, we not only have behavior (snoring, blushing, hiccupping) but also action, and action is the movement or change that reasons cause.1 This Davidsonian view is widely accepted in philosophy of mind and action theory. I’ll call it “The Standard View.” The Standard View has much to recommend it. It fits nicely into a materialistic metaphysics. If beliefs, desires, and fears are physical events of some sort (presumably in the brain of the actor), as they certainly will be on a materialistic view of the world,2 then there is no problem in understanding how they produce the actions for which they provide (when the reasons are good) a rational justification. They cause them. Neurophysiologists provide, in increasingly rich detail, the electrical/chemical/mechanical details of just how this happens. So mental events are genuine players—via action—in the wider physical world in which they exist. However, this story doesn’t help one understand how reasons explain actions. As physical events in the brain, beliefs and desires can cause me to go to the fridge without my reasons for going to the fridge explaining my trip to the fridge. What I believe (that there is beer in the fridge) needn’t be true for this “fact” (that there is beer in the fridge) to be my reason, at least the reason I (wanting a beer) give, for going to the fridge. How does a false mental content—that there is beer in the fridge—explain a physical movement? If we do not understand how the content (whether true or false) 13

14 What Must Actions Be for Reasons to Explain Them?

of a mental state—the fact that I thought there was beer in the fridge— could explain a bodily movement, the fact that thoughts (being physical events) caused such movements provides absolutely no scientific or philosophical illumination. After all, we all know that spoken words—acoustic vibrations with meaning—cause things. They cause things in the same way as other sounds cause things. They exert pressure on them. That is not a problem. The problem, rather, is understanding whether—and if so, how— the meaning of these vibrations explains their effect. When I make some silly remark and my wife frowns, I understand reasonably well as to how my remark might have caused her to frown. My remark caused her eardrums to vibrate, a physical event that in turn caused electrical signals to be sent to her auditory cortex. I don’t know exactly how things proceed from this point forward, but it seems clear that eventually the motor cortex is activated and this, via electrical signals sent to facial muscles, produces a change of expression—a bodily change we describe as a frown. All this, in broad outline, seems reasonably clear and uncontroversial. What I don’t understand is how what I said, the meaning of the acoustic vibrations I produced, explains this effect. How does meaning explain her frown as it must if she is frowning not because I produced acoustic vibrations but because I said something silly? That, after all, is what she would give as her reason if asked to explain why she was frowning. Until we understand how meaning or content—what we say and believe—figures in the explanation of the physical events our utterances and beliefs produce (frowns and bodily movements), we will not understand the role of mind or meaning in the world. It is for this reason that I think it is a mistake to identify action with the external events (e.g., bodily movements) that are the effect of the reasons for which the action is performed. It makes it impossible for reasons to explain the action. Even if one assumed that bodily movements to the refrigerator are caused (in an appropriate way) by a belief and desire, these movements need not and, if we accept the scientific story, are not explained by what the person believed and desired. Not unless one assumes that the electrical/ chemical/mechanical explanations given by neurophysiologists are—and will always be—deficient or incomplete. If the neurophysiological story is correct, then actions must be something other than bodily movements if reasons are to explain them. Standard View theorists (like Enç) do, of course, carefully distinguish action from the result of action, the event or events which must occur for the action to occur but which can occur without any such action. The movement of my finger is the result of my moving my finger (the action), and there is a clear conceptual distinction between an action and its result. I can’t move my finger without my finger moving, but my finger can move without my moving it. You can move it. This distinction between an action and its result, although perfectly correct and useful, is of no help with the problem of understanding how reasons could explain actions. It is of no

Fred Dretske

15

help because, on The Standard View, the action (moving one’s finger) is still identified with the result (finger movements) when the result is brought about by reasons. So when you move your finger deliberately, for reasons, then if someone can explain why your finger moved, then, since the finger movement (result) is one and the same event as your moving it (the action), they have explained why you moved it. This, it seems to me, is still the wrong result. Neurophysiologists may be experts on why fingers and other bodily parts move as they do, but are they also experts on why people move them as they do? A husband is conceptually (type-type) distinct from a man, yes, but this doesn’t prevent (token) husbands from being identical with (token) men. Susan’s husband is a man. It follows that if that man is Susan’s husband, explanations of why that man got drunk are explanations of why Susan’s husband got drunk. Likewise, if actions are a special class of bodily movements, movements brought about by reasons, explanations of why a bodily movement occurred are, perforce, explanation of why an action occurred when that movement is caused by reasons (i.e., when it is an action). The Standard View, therefore, has this unfortunate result. When I go to the fridge to get a beer, an electrical/chemical/mechanical explanation of why the bodily movements occurred that got me to the fridge are, necessarily, explanations of why I went to the fridge. As far as I can see, this leaves the explanatory role of my reasons for going to the fridge epiphenomenal. Even if I have reasons, they don’t explain anything.

2

Temporal considerations: Basic vs. nonbasic actions

The Standard View has additional problematic consequences when applied to nonbasic actions. What do we identify as the action of, say, kicking a field goal in football? The result of this action is, presumably, the football flying between the goal posts. One hasn’t kicked a field goal unless this event occurs, but this event can occur without one kicking a field goal. One hasn’t kicked a field goal if one (or somebody else) throws the ball through the uprights. But, surely, the action of kicking a field goal cannot be identified with this result even when this result is caused in the right way (by, say, an intentional kick of the ball). If we identified the action with this result (when caused in the right way), then the action would occur after the agent did everything (kicked the ball in the right direction) he could do to bring about this result. In fact, the kicker could have dropped dead before he kicked the field goal if kicking a field goal is to be identified with the result of kicking a field goal. It would, I admit, have to be a long field goal and a quick death for this to happen, but it could happen. Can you do things? Intentionally do things, after you are dead? I know your actions can have effects (results) after you are dead, but can you actually do things after you are dead?

16 What Must Actions Be for Reasons to Explain Them?

Enç (2003) was fully aware of this anomaly and for that reason refused to identify nonbasic actions with their result. That identification only works for basic actions. For nonbasic actions like kicking a field goal the action is identified not with the result of the act (the ball flying through the uprights) but with the result of the basic action by means of which the nonbasic action is performed. Assuming, then, that the basic act by means of which one kicks a field goal is kicking the ball toward the goal post with the appropriate intention, the action—kicking a field goal—is identified with the result of the basic act—the movement of the leg by means of which one kicks the ball. Since basic acts are identified with their results when these results are caused in the right way (by an intention), kicking a field goal is identified with leg movements when those leg movements are caused by an intention to kick a field goal and these leg movements in turn cause the ball to fly between the goal posts. So on this account one avoids the paradoxical result that one can kick a field goal after one kicks the ball, but it has the odd result that kicking a field goal occurs before the ball goes through the uprights. Consider a different example. On this variation of The Standard View killing X (having the result that X dies) is not to be identified with the result of this act (X’s death), but with the result of the basic act by means of which one causes X’s death. So if one kills him by dropping poison in his morning coffee, then killing X is to be identified with the result of the basic action by means of which one kills him, the finger and arm movements executed in placing poison in the coffee. Once again, we get the result that the person is killed before the person dies—perhaps, depending on the effectiveness of the poison, long before he dies. This didn’t sound right to me. That is why I used such examples (in Dretske 1988) to argue that actions (behavior in general) should not be identified with bodily movements (or the more distant effects of bodily movement) but with a causal process, the process in which internal events (reasons in the case of actions) cause such events. Kicking a field goal, for instance, should not be identified with the ball flying through the uprights (the result of this action). Nor should it be identified with the bodily movements that cause the ball to fly through the uprights (the result of the basic action by means of which this action is executed). It should rather be identified with a temporally extended causal process in which one thing—leg and foot movements or, even earlier, the reasons those movements are executed— causes the ball to fly through the uprights. An action, being a causal process, takes time, and the time it takes is the interval from the occurrence of the internal cause (the reasons one has for performing the act) to the result of the action. One hasn’t kicked a field goal until the ball goes through the uprights. One has not killed X until X dies. One starts to do these things at the moment one’s reasons initiate the causal process that ends up with that result, but one hasn’t done it until the result occurs.

Fred Dretske

17

An interesting legal application of this distinction occurred in the state in which I reside. In the case of North Carolina State v. Detter (1979) the question arose as to whether a woman killed her husband at the time she administered the poison (February 1977) or at the time he died (June 1977). This turned out to be important since the law defining penalties for murder changed during this five-month interval. In February, the death sentence was not permitted. In June, it was permitted. So if the woman killed him in February, at the time she administered the poison, she could not be given the death penalty. If she killed him in June, when he died, she could. In an effort to comply with the prohibition against ex post facto application of the law, the court found that in this case the murder occurred at the time the woman administered the poison—in February—a time before the victim died. So the court held that she murdered him five months before he died. This, I admit, looks like the court is siding with The Standard View—at least Enc’s version of it. Nonetheless, the court also held that despite murder being a crime that requires both act (administration of poison in this case) and result (death), special considerations (of justice and fairness, presumably) may cause it to go one way rather than the other on this matter. In the case of North Carolina State v. Williams, for instance, when different legal considerations dominated (whether someone was an accessory after the fact to murder) the court identified the time of the murder with the time of death. So a person who rendered aid to the killer after the fatal blow was struck but before the resulting death could not be convicted of being an accessory after the fact to murder because the murder had not yet occurred (the victim had not yet died). The courts, we hope, are more interested in legal justice than terminological niceties, so I do not take this instance of judicial waffling as decisive. If it takes Sarah five months to kill Clyde, the courts are free to set the time of action wherever they please (in this five-month interval) to achieve fairness or justice in the application of the law. This is simply to say that courts aren’t in the business of making metaphysical judgments (although we hope that their judgments are in accord with commonsense metaphysics) anymore than the courts are in the business of making moral judgments (though, again, we hope their judgments are in accord with commonsense morality). The lesson of these examples, I’m afraid, is that the law isn’t of much help with this problem. It illustrates the problem; it doesn’t solve it. I don’t know how to adjudicate disagreements about ordinary language. It sounds wrong to me to say Sarah killed Clyde after Sarah was dead (which it makes sense to say if the time of Sarah’s action is the time of result—Clyde’s death). And it sounds wrong to say that she killed him before he died (which it makes sense to say if the time of action is the basic act Sarah performed that caused Clyde’s death). So I’m pushed toward the view that behavior is a causal process between basic act (or the mental antecedents of the basic act)

18 What Must Actions Be for Reasons to Explain Them?

and the result, a process that may take a long time. But others (like Enç), I know, see (and hear) things differently. So let me set aside this dispute about what it sounds odd to say and get to the real, the deeper, reason I think we have to identify behavior not (in the case of basic actions) with the results that reasons cause but with a causal process in which reasons cause these results. The real reason lies in the explanatory power of reasons. Without a process view of action, reasons are powerless to explain it.

3

Explaining action

I have been arguing that if basic action really is just bodily movement and change, then neuroscientists, not you and I, are (or will someday be) the authorities on why you and I do the things we do. Neuroscientists will tell us why our fingers and legs, our eyes and tongues, move the way they do because neuroscientists are, or will someday be, the experts on those electrical/chemical/mechanical processes that make these parts of our bodies move the way they do. If intentional psychology, explanations of behavior in terms of a person’s beliefs and desires, is competing with neuroscience for an explanation of bodily movement, then I, for one, don’t see much hope for intentional psychology. We should all pack up and move to San Diego. It will be neuroscientists (not the kicker) who will tell us why the kicker kicked the ball. It will be neuroscientists, not me, who will tell you why I went to the fridge. The only wayI know to avoid this result is to suppose that the sort of psychology that adverts to beliefs, desires, hopes, and fears in explaining (sometimes) why people behave the way they do is not playing the same explanatory game as neuroscience. Neuroscience may be able to tell us why our legs move the way they do, but we are better positioned, at least sometimes, to explain why we moved our legs that way. This strategy, though, requires distinguishing between the movement of a kicker’s legs and the kicker moving his legs. It requires distinguishing between the bodily movements that get me to the refrigerator and my going to the refrigerator. Once this distinction is in place, we can let neuroscientists explain the bodily movements. Reasons and the mental states that give us our reasons are playing a different explanatory game. Their business is the explanation of behavior. Nevertheless, how does this help? Whether we are talking about bodily movements or the causing of bodily movements, we are still talking about physical events (or processes) of some sort, and the question is: why are reasons any more relevant to the one than to the other. If neuroscientists are the experts on why our legs move, why aren’t they also the experts on why those events in the brain, whatever they are, cause the legs to move? I’ll give an analogy that I hope will be suggestive. It will be crude, but I hope the idea will come through.

Fred Dretske

19

Coins are deposited in vending machines and they cause the machine to behave in certain ways. You deposit US$1 worth of coins in the machine, push a button, and a coke comes tumbling down the chute. There is little doubt about what causes the coke to be ejected. It is the coins you put in it together with the button you pushed. If you had put only 85¢ worth of coins, nothing would have happened when you pushed the button. Cokes these days cost US$1. So we routinely explain the machine’s behavior—giving us a coke—by describing the monetary value of the objects, the coins, we put in it. You have to put in a dollar, we say, to get a coke. Just 85¢ isn’t enough. That is why you didn’t get a coke. Put in 15¢ more and the machine will behave the way you want it to when you push the button. Now we all know, or think we know, why the machine really behaves the way it does. Despite our ordinary way of talking it isn’t that the objects we deposit in them have a certain monetary value—US$1 or 85¢ as the case may be—that explains why a coke comes (or fails to come) out when we push the button. It is the fact that these coins have a certain size, shape, density, or whatever other intrinsic properties machines are built to respond to. Slugs of no (legal) monetary value whatsoever will have exactly the same effect on the machine if they have the same intrinsic properties—that is, if they are perfect counterfeits. Monetary value is causally irrelevant, epiphenomenal, as far as the machine-behavior is concerned, and it is epiphenomenal despite our ordinary practice of invoking monetary value in our explanations of machine-behavior. We say that the machine didn’t give us a coke because we didn’t deposit the required US$1 despite our knowledge that it isn’t having a monetary value of US$1 that explains the result. It is, and we all know it is, the intrinsic physical properties of these coins, not their monetary value, that explains the result. The coins (together with pushing the button) cause this result, yes, but what it is about the coins that explains this result is physics, not economics. It is facts about shape, size, and weight, not facts about monetary value. As far as vending machines are concerned, economics is epiphenomenal. We are merely pretending it isn’t when we explain machine-behavior by describing economic facts. So if machine-behavior (giving us a coke) is identified with the result of a machine’s basic actions, with the external events or changes (cokes coming down the chute) brought about (in the right way) by the appropriate internal events (the coins we deposit in them), we would have to conclude that physics, not economics, was the place to go for explanations of (say) why the machine didn’t give me a coke. Physics, not economics, would tell us why machines behave the way they do because physics, not economics, tells us what it is about the internal causes (the deposited coins) that explains their effect on machine movement. What happens, though, if we identify machine-behavior not with the result of its basic actions (the external events the deposited coins cause), but with the temporal process in which these internal objects (coins) bring

20 What Must Actions Be for Reasons to Explain Them?

about this result? What if we identify the machine’s action of giving us a coke not with a coke’s appearance, but with internal coins causing cokes to appear? Ask yourself, why do objects having the intrinsic properties of our nickels, dimes, and quarters have this effect on machines? Why, when deposited in coke machines, do they cause cokes to appear? Well, pretty clearly, vending machines are designed and manufactured to be sensitive to objects having the size and shape (intrinsic properties) of our nickels, dimes, and quarters because an important extrinsic property, monetary value, supervenes on these intrinsic properties. Business being what it is, machines that dispense commodities like cigarettes, food, and drink would not be designed to yield their contents to objects having these intrinsic properties unless objects having these intrinsic properties had monetary value, for the most part. Remove the fact of supervenience (as a result of widespread counterfeiting, say) and objects with these intrinsic properties, our nickels, dimes, and quarters, would soon lose their causal efficacy. They would no longer produce the effects they now produce. They would lose their causal efficacy because machines would no longer be built to respond to them. The causal efficacy of objects having those physical properties (on machines— not to mention people) depends on the supervenience of monetary value on those intrinsic properties. Let value supervene on a different set of intrinsic properties and these other objects will, quickly enough, assume the causal powers of our current nickels, dimes, and quarters. So if what we want to explain is not why a coke came sliding down the chute (the shape and size of the coins deposited will explain that), but why objects having the size and shape of nickels, dimes, and quarters cause cokes to appear, why objects of that sort have effects of this sort, the answer lies, in part at least, in economics, not physics. It lies in the fact that there is a reliable correlation—at least it is a reliable enough correlation—between intrinsic and extrinsic properties of the internal cause. When we turn to the mind–body case, this fact is suggestive. If we think of ourselves as “vending machines” whose internal causal structure is designed not, as with vending machines, by engineers, but by evolution and learning, we get the following result: although it is the “size” and “shape” (the syntax, as it were) of the internal causes that makes the body move the way it does it is, or may be, the fact that a certain extrinsic property supervenes on that neurological “size” and “shape” that explains why internal events having these intrinsic properties have the effects they have. What explains why certain neurological events in the visual cortex of a chicken— those caused by an overhead hawk—cause the bodily movements we think of as cowering and hiding is the fact that these neurological events have a significant (to chickens) extrinsic property—the property of normally being caused by predatory hawks. It is the possession of this extrinsic property (what the internal events indicate about external affairs) that explains why objects having those intrinsic properties cause what they do.

Fred Dretske

21

There is but a short step from here to the conclusion that it is the extrinsic, not the intrinsic, properties of internal events that causally explain behavior. All that is needed to execute this step is the premise that behavior is not the bodily movements that internal events cause, but the causing of these movements by internal events. All that is required is an appropriate distinction between the behavior that beliefs and desires explain and the bodily movements that are the result of that behavior. For, if moving your arms and legs (behavior) is not the same as the movements of the arms and legs, but rather an internal event causing the arms and legs to move, then although the intrinsic properties of our internal “coins” will explain (via activation of muscles) the movements of arms, legs, and tongue, the extrinsic properties, perhaps even the meaning or content, of these internal events will explain why we move them. Reasons get their hand on the explanatory wheel by distinguishing what reasons are supposed to explain—behavior—from the movements that are the result of that behavior. In the final analysis, then, it is not merely the temporal anomalies relating to when behavior occurs that is the reason we must insist on the distinction between behavior and the results of behavior, the distinction between moving your arm and the movement of your arm. Although these anomalies are, I think, important, they are not decisive. It is, at a deeper level, the need to understand the explanatory relevance of what we believe and desire, intend and fear that is the key.

Notes 1. I speak here of reasons causing actions. It would be better to speak of the mental states (belief, fear, desire, etc.) whose content constitutes our reasons as the cause (see Dancy 2000). When it is important to do so (it isn’t at the moment), I will be more careful in the way I describe these matters. 2. This is the unobjectionable (to most materialists) token-token identity, not the objectionable (to almost everyone) type-type identity where a belief in general (e.g., the belief that there is beer in the fridge) is identified with a certain type of brain state.

References Dancy, J. 2000. Practical Reality. Oxford: Oxford University Press. Dretske, F. 1988. Explaining Behavior. Cambridge, MA: MIT Press (A Bradford Book). Enç, B. 2003. How We Act. Oxford: Oxford University Press.

2 What Are Reasons for Action? Stephen Everson

1 When Donald Davidson published ‘Actions, Reasons, and Causes’ in 1963 what was most striking in his article was the claim that an agent’s beliefs and desires could be the causes of, and not merely the reasons for, his actions. In more recent years, however, what has become more controversial is whether one can properly take an agent’s propositional attitudes to be reasons for his actions. There has been a growing trend amongst philosophers of action to recognise that a reason for action is something that favours or makes valuable an action of the relevant kind, and that an action is not made valuable by the agent’s desire to do it or belief that it would be good to do it. One’s belief may be false and one’s desire unmotivated, and in such cases there may be nothing to be said for the action they lead to. For ease of labelling, let us say that an account of reasons such as Davidson’s is psychologistic and that the opposing account is rationalist. Now, an obvious way to contrast psychologism and rationalism is to say that whilst the former takes reasons to be psychological states of agents, the rationalist rather takes them to be facts. For the psychologist my reason for writing to the Vice Chancellor will be something like my belief that he needs telling, or my desire to tell him what he most needs to hear, or some suitable combination of beliefs and desires, whilst for the rationalist the reason will rather be the fact that he needs to be told. Although I think that rationalism is correct, I do not intend to provide support for it here. My interest is rather in taking up the question of what kind of thing we should take reasons to be if rationalism is true. For whilst commitment to psychologism does settle that question – reasons, on this view, are psychological states of agents – rationalists may yet disagree about what kind of things are reasons for action. That disagreement is not settled by accepting that reasons are facts, since there may be prior disagreement over the nature of facts. So, Joseph Raz in

22

Stephen Everson

23

defending the claim that reasons are facts rather than beliefs, proceeds to characterise facts in a rather Austinian manner: When saying that facts are reasons I am using the term ‘fact’ in an extended sense to designate that in virtue of which true or justified statements are true or justified. By ‘fact’ is meant simply that which can be designated by the use of the operator ‘the fact that ... .’ Similarly facts include the occurrence of events, processes, performances and activities. Given this wide use of ‘fact’, beliefs, though not their contents, are also facts.1 Others, however, will incline to a more Strawsonian understanding of facts, according to which ‘facts are what statements (when true) state; they are not what statements are about’2 – nor, we may add, are they what make a true statement true. On this view, facts will indeed be rather the contents of (true) beliefs than beliefs themselves, since they will be true propositions. There would, of course, be little point in pursuing a debate over which understanding of ‘fact’ gets closest to ordinary usage – it is unlikely that our ordinary talk of facts is, or needs to be, sufficiently coherent to allow any helpful theoretical morals to be drawn by reflecting on it. What is important is to note that one may talk about facts in these two different ways and to make sure that one does not slide between them. Having distinguished them, one may then consider the theoretical uses each is suited to – and, in this case, which of the two notions provides items suited to stand as reasons for action.3 Stipulatively, then, I shall follow Strawson and talk of facts as true propositions, using ‘state of affairs’ for the kind of thing taken by Raz to be reasons. The issue, then, is whether, having abandoned psychologism about reasons, the rationalist should think of reasons as facts or as states of affairs. Although Raz, as we have seen, effectively opts for the latter, this in fact constitutes little more than an aside in his argument, and he does not argue the point.4 Similarly, Jonathan Dancy in his recent Practical Reality, whilst declaring firmly that reasons for action, ‘are things like his self-satisfaction, her distress, yesterday’s bad weather, and the current state of the dollar. They cannot be abstract objects of the sort that propositions are generally supposed to be,5 does little to support this. He does say that propositions ‘are, as we might say, too thin or insubstantial to be able to make an action wrong’ – but does not then move to turn that metaphor into argument.6 This is disappointing, since the issue is not without interest. At the very least, working out what is at stake in the matter helps to clarify the different ways in which we talk of reasons. Although I favour a propositional account of reasons, I do not

24 What Are Reasons for Action?

pretend that I here provide anything like a knock-down argument against taking reasons to be states. My aim is to provide what might serve as a starting point for more serious discussion of the issue.

2 In renouncing the idea that one can identify agents’ reasons for action with their own propositional attitudes, the rationalist opens up the possibility that the reasons agents have to act are quite independent of their own motivational states. That is to say, it may be that the relation between a consideration, an agent (or agents) and an act type, such that the consideration is a reason for the agent to perform an action of that type, may now hold independently of the psychological condition of the agent.7 Certainly, on the rationalist’s account, one need make no mention of the agent’s propositional attitudes in specifying that this relation obtains.8 However, whether or not reasons are psychologically constrained – whether, to use the current terminology, we are to understand reasons internally or externally – they are nevertheless able to enter into our thought. To engage in practical deliberation one must reflect on the reasons for and against acting in whatever way; to criticise what someone does is to place his action in the context of the reasons he had for acting, and to explain someone’s action may be, in part, to specify the reasons for which they acted. All of these activities require that we should have at least limited cognitive access to the reasons that agents have to act, and such access has seemed to some to require that reasons must themselves be propositional in character. So, Stephen Darwall appeals to the deliberative role of reasons to deny the Davidsonian claim that we can take an agent’s propositional attitudes to be themselves reasons for him to act. Darwall’s objection is that beliefs and desires are not the right kind of thing to play the role in deliberation that reasons need to: since reasons ‘are what people are to take account of in evaluating choiceworthy alternatives, they must be the sort of thing that can be thought or said on behalf of an act’, and this requires that ‘they must be propositional in form and expressible with a “that” clause’. More generally considered, reasons for someone to do something are a subclass of the things that can be said asserted, considered, judged, thought, and so forth. To have a convenient tag to refer to such items, and to preserve the association of reasons with things that can be said on behalf of acting, I shall say that reasons for someone to do something are dicta.9 The implication of this for a thesis such as Davidson’s are obvious enough. If reasons must be dicta – or, as I shall rather say, propositions – then beliefs and desires are not of the ‘appropriate ontological category’ to be reasons,

Stephen Everson

25

and so the idea that ‘any reason for a person to do something must actually be a desire of his is therefore a nonstarter’.10 Although Darwall’s argument is directed against the psychologist about reasons, it functions as effectively against a rationalist who thinks that reasons are non-psychological states of affairs. Darwall’s argument is too hasty, however. Whilst it is certainly true that in practical deliberation one reflects on reasons for and against acting in whatever way, and so one must be able to think about reasons for acting, it is not a consequence of this that the reasons themselves must be able to stand as the contents of the thoughts one has about them, any more than it is a consequence of the need to think about the objects in one’s environment when deciding how to move around them that those objects should be propositional in character. An advocate of taking reasons to be states of affairs has no difficulty in describing what goes on when an agent deliberates. If it were Ian’s distress that favoured my telling him where I hid the wine, rather than the fact that he is distressed, then I can act on this reason by entertaining both the thought that Ian is distressed and that his distress favours my putting him out of his misery by revealing the wine. I need to think, that is, that the relevant state of affairs obtains – but I can do this consistently by thinking that the reason for revealing the wine is his distress itself. It is that state which I take to be the reason for acting – and since I think that it obtains and favours the action, I also think that I have a reason to reveal the wine. The point, then, is that whilst we of course need to be able to think about reasons, this does not mean that reasons must be what we think when we do so. We do not need reasons to be propositional in order to account for their role in deliberation or evaluation. They also, however, have a role in the explanation of action, and this would seem to provide more compelling grounds for taking them to be facts rather than states of affairs.

3 Consider what Bernard Williams gives as one of the two ‘fundamental motivations’ for his internalist account of reasons: It must be a mistake simply to separate explanatory and normative reasons. If it is true that A has a reason to w, then it must be possible that he should w for that reason; and if he does act for that reason, then that reason will be the explanation of his acting. So the claim that he has a reason to w – that is, the normative statement ‘He has a reason to w’ – introduces the possibility of that reason being an explanation; namely, if the agent accepts that claim (more precisely, if he accepts that he has more reason to w than to do anything else). This is a basic connection. When the reason is an explanation of his action, then of course it will be,

26 What Are Reasons for Action?

in some form, in his [actual motivations], because certainly – and nobody denies this – what he actually does has to be explained by his [actual motivations].11 If this argument were sound, and unless Williams is speaking loosely, it would settle the question of whether reasons are propositional. For explanations, or what are given in explanations, are facts, and so if reasons are to be able to be themselves explanations, they too would need to be facts – that is, true propositions.12 The idea that reasons are what one specifies in explaining actions is one that seems to have great intuitive appeal, and Williams’ argument here can be helpfully amplified by looking at John Searle’s more sustained treatment of these matters in his Nicod lectures. He states: The notion of a reason ... is embedded in at least three other notions, and the four can only be understood together as a family. The other notions are ‘why’, ‘because’, and ‘explanation’... . Stating a reason is typically giving an explanation or part of an explanation. Explanations are given in answer to the question ‘Why?’ and a form that is appropriate for the giving of a reason is ‘Because’.13 He then draws the conclusion from this that reasons must be propositional: The syntax of both ‘Why?’ questions and ‘Because’ answers, when fully spelled out, always requires an entire clause and not just a noun phrase. This syntactical observation suggests two semantic consequences. First the specification of both explanans and explanandum must have an entire propositional content, and second, there must be something outside the statement corresponding to that content. Reason-statements are statements, and hence linguistic entities, speech acts with certain sorts of propositional contents; but reasons themselves and the things they are reasons for are not typically linguistic entities. Reasons, then, are what reason-statements are true in virtue of – and there is ‘a general term to describe those features of the world that make statement or clauses true, or in virtue of which they are true, and that term is “fact” ’.14 However, action-explanations themselves show that one cannot maintain that all reasons are facts, since when the agent has false beliefs one cannot cite facts about the world to explain what he does. In those cases, one has to cite the belief itself as the reason. This, according to Searle, can still be accommodated within the general schema, since beliefs, like facts, have, he thinks, a propositional structure. ‘The formal constraint on being a reason is that an entity must have a propositional structure and must correspond to a reason statement’.15

Stephen Everson

27

That Searle is moved to treat both beliefs and their contents as reasons for action should alert us to the fact that something has gone wrong here. For in addition to any formal constraint on what kind of thing can a reason be there must also be a substantive constraint, and this is that anything that is a reason should be such as to favour the action for which it is a reason. The worry here is not that Searle is assuming a psychologistic understanding of reasons: in accepting facts as reasons at all, Searle distances himself from that position. The worry is rather the inappropriateness of thinking that, for instance, the fact that Ian is distressed and my believing that Ian is distressed should both be such as to stand in the same reason-giving relation to my revealing the wine. A clue to what is going wrong in Searle’s discussion can be found in how he supports his contention that the notion of a reason is tied to that of explanation: ‘To the question, “Why is it the case that p?” the answer, “Because it is the case that q” gives the reason why p, if q really explains, or partly explains p. That is the reason why all reasons are reasons why.’16 Now, it is certainly true that in giving any explanation one will cite a reason for what one explains, but it does not follow from this that all reasons are reason why. In effect, Searle is just equivocating on the notion of a reason. Care needs to be taken in this context to distinguish between different kinds of explanatory role that reasons for action might play. Williams and Searle commit themselves to the claim that reasons for action are themselves explanations, but this is clearly not the only way to allow such reasons to play a role in explanations. In an earlier article Williams placed a condition on something’s being a reason for action that it should be able to ‘figure’ in an explanation of action17 – and that condition is uncontroversial precisely because it is so vague. To see how it might be met without moving to accept the claim that reasons are explanations, it is helpful here to consider the parallel case of causal explanation. For one can certainly accept that it is a condition on taking one event to be a cause of another that the first should be able to figure in the explanation of the occurrence of its effect – one cannot have a causal explanation that does not make manifest to some degree the cause of what is explained – but clearly one should not be led from this to the thought that the cause will itself be the explanation of its effect: to use a slightly old-fashioned jargon, causation is a ‘natural’ relation that holds between events (or if one prefers between states or objects), whilst explanation is a ‘rational’ relation that holds between facts.18 In the case of causation and causal explanation, it is fortunately easy to distinguish these different relations, since we have the terminology to do so: the relevant relata of the two explanations are causes in the first relation and reasons in the second. Clear discussion of the explanatory role of reasons is made more difficult by the fact that ‘reason’, unlike ‘cause’, suffers from a straightforward ambiguity – and, moreover, an ambiguity that is, in this context, capable of misleading even the most alert. For there is a general

28

What Are Reasons for Action?

notion of a reason that permits us to say of an explanation of any type that it cites the reason for what it explains. The reason why my foot is swollen is that I dropped the dictionary on it; the reason why the train slid off the rails is that the points were loose. When the explanations are causal, we can readily distinguish between the reason which is explanans of the explanation and the cause of the effect whose occurrence we are explaining. The failure of the points was the cause of the derailment, whilst the reason the train was derailed was the fact that the points failed. When we come to rationalising explanations, in contrast, matters are terminologically more confusing, since one way such explanations work is by citing an agent’s reason for action. The notion of a reason here, however, is the notion of an item which stands in a justifying relation to an action, and this is at a level parallel to that of causes and not that of the ‘reasons’ of causal explanation. Let us say that a reason of this kind is a normative reason and that a reason of the other is an explanatory reason. Therefore, just as in the case of causal explanation, where we can say that one specifies the explanatory reason (some causally relevant fact) and in doing this cites the cause, in rational explanation one specifies the explanatory reason why someone did something, thereby citing their normative reason. It will now be seen just how treacherous it might be in this context to pose the question ‘What was the reason why a w-ed’, since this might properly be answered by citing either an explanatory or a normative reason, and it will be easy enough to be confused about which kind of reason has been offered. In this context also, then, it may seem less controversial than it should do to claim that someone’s (normative) reason for action will also be the explanation for his acting as he does. At least without further argument, there is no more reason to suppose that the normative reason for an action will be its explanation than there is to think that the cause of an event will be its. All explanatory reasons are reasons why, and to give the reason why someone did something may be to cite his reason for acting: but one can accept that all reasons why are facts whilst leaving it open whether reasons for are states of affairs or propositions.

4 Neither the role of reasons in deliberation nor in explanation, then, is such as to support taking them to be propositional in character. We certainly take reasons into account when deciding how to act, but this only requires that we are able to think about reasons and not that they should be themselves the contents of our thoughts when we do think about them. Similarly, although reasons for action do figure in our explanations of actions, they do not need to figure directly as one of the relata of an explanation, any more than do the actions whose occurrence is to be explained. Causal explanations similarly connect facts, but in doing so explain why some events come

Stephen Everson

29

about as the result of others. Indeed, the advocate of taking reasons to be states of affairs is likely to be encouraged by the comparison with causation and causal explanation, since to take normative reasons to be states of affairs will allow the two kinds of explanation to run on satisfyingly parallel lines. Each kind of explanation will connect facts, whilst its underlying relation will be between spatio-temporally located items – events in the case of causation and states of affairs (and events) in the case of rationalising explanation. As it stands, of course, this is no more than suggestive – and whilst it is clear why causes cannot be propositions (as they must be temporally related to their effects and so themselves be temporally located), it is at least not obvious that reasons too need to be temporally related to the actions they favour. I shall return to that issue in the section 7. First, however, I want to consider a more direct line of argument for taking reasons to be states of affairs. Consider here Joseph Raz’s characterisation of what he calls the ‘classical approach’ to agency: this ‘takes acting for a reason to be the distinctive and central case of human agency ... . Reason is then explained in part by invoking value: valuable aspects of the world constitute reasons’.19 Now, the idea that reasons are at least determined by valuable features of the world should be common ground to all except those who favour a psychologism about reasons. If our actions, or the states or events they give rise to, did not have good or bad features, then we should not have reasons to act at all. And whilst Raz is fairly relaxed about what can be reasons – these, he thinks, may include ‘events, acts, states and more’ – all the possibilities he gives are concrete rather than abstract items. Indeed, it may be the link between reasons and value that underlies Dancy’s rather pre-Socratic remark, cited earlier, that propositions cannot be reasons because they are ‘too thin or insubstantial to make an action wrong’. 20 Whatever this quite means, it is clear enough that Dancy thinks that something can only be a reason for action if it is the sort of thing that can make an action wrong (or right, presumably) – and that condition might well seem more compelling if one accepts the dependency of reasons on ‘valuable aspects of the world’. Certainly, if it is correct, it would seem to provide the basis for an argument that reasons are concrete rather than abstract items. For it is surely plausible that what have value or disvalue are the states of affairs that obtain and the events that occur in the physical world and not facts about those states and events. It would be at best odd to maintain that what made it wrong to tell Brad the truth was not the distress he suffered as a result or the pleasure I took in causing it but rather the fact that he was distressed or the fact that I took pleasure in causing his distress. If it makes sense at all to think of those facts as having disvalue – which is hardly evident – that could only be derived from the disvalue of the states they are facts about. Is this not sufficient, then, to secure the claim that reasons are indeed states? In fact, the argumentative gap between these considerations

30 What Are Reasons for Action?

and Dancy’s condition will be obvious enough.21 Reasons for action can be secured in the value and disvalue of events and states of affairs without their needing to be identified with the events and states that secure them. It is quite consistent to hold that whilst it is Brad’s distress which makes my action wrong and gives me a reason to comfort him, it is the fact that Brad is distressed that is the reason it gives. It is not as if there could be a relevant state or event without there being a relevant fact – and if it turns out that facts are better suited to stand as reasons for action than are events and states of affairs, this will in no way conflict with a classical approach to agency.

5 The discussion so far, then, has been both indirect and inconclusive. Various writers have looked to the relations between reasons and deliberation, reasons and explanation and reasons and value in the hope that these will show that reasons themselves must be either facts or states of affairs, but none of these has been sufficient to determine an answer. A different approach is needed – and to many the obvious strategy will be to investigate the semantic properties of the sentences we ordinarily use to ascribe reasons for action in the hope that these will favour setting one kind of item as reasons rather than the other. If one were to employ that method, this would probably favour facts over states of affairs. ‘One reason for you to give the money back is just that he saw you take it’ is, at least to my ear, much more natural than ‘One reason for you to give the money is his having seen you take it.’ Again, ‘You should take him out one night because he is miserable’ is more natural than ‘You should take him out one night because of his misery’. 22 Of course, two examples are certainly insufficient to make the general case, and even these may not sound so obvious to other ears. I shall not try to be persuasive on the point, however, since although I do think that if one were to treat ordinary usage as a metaphysical guide, this would support taking reasons to be facts, I also think that it would be madness to give it that role. The idea would have to be that at the time when usage was being fixed, speakers knew, or at least were somehow sensitive to, the nature of reasons for actions and adapted the way of ascribing reasons accordingly – and one only has to make that idea explicit to see just how implausible it is. The fact is that ordinarily people are pretty insensitive to the distinction between facts and states of affairs, as they are to that between facts and events, and there is no reason at all to think that, when those distinctions matter, the formal ontological commitments of everyday talk about reasons are more likely than not to be met. The method of truth is an important one in metaphysics, but unless one is careful to observe its limitations, its employment will result in a weirdly Jungian approach to determining the precise contours of

Stephen Everson

31

reality, secured in the expectation of finding metaphysical truths in the collective unconscious of English speakers. 23 The consequence of this, of course, is that a great many of the claims we make about reasons may well turn out to be false. If I say, for instance, that one reason you should give money to your old college is that it cannot afford to fund its scholarships, then (assuming that the function of that-clauses is indeed to specify propositions) what I say is true just if the proposition that the college cannot afford to fund its scholarships stands as a reason for you to give money to your old college. If the state-theorist is right, no proposition stands in that relation to you: being a reason for action is a relation between agents, act-types and states of affairs, not propositions. Since at least much of our everyday ascription of reasons is done by means of constructions that introduce sentences to specify propositions as reasons, it will indeed be a consequence of the state-theorist’s view that much of what we currently say about reasons is in fact false. The state-theorist need not be embarrassed by this but it does raise an obvious concern which he needs to satisfy: the starting point for any account of reasons must be our ordinary ascriptions of reasons to agents – and so a theory that is systematically or widely inconsistent with those judgements stands in danger of being insufficiently constrained to be credible. Now, what makes it inappropriate to look to ordinary usage to reveal the metaphysics of practical reason is that our everyday talk is careless of ontological niceties – and the state-theorist can quite properly try to appeal to this to explain why people say the particular false things that they do. What he needs to show is that whilst it may be false to say, for instance, that Erika should give money to her old college because it cannot afford to fund its scholarships, this is only because of a perfectly understandable confusion of states and facts. Here the example of causal relations is again helpful. If a doctor tells you that your illness was caused by the fact that you were exposed to radiation, it would be absurd to think that since he has said something false, he has provided no information about the aetiology of your illness. Knowing that doctors are no more careful about ontological categories than anyone else, one can readily construct a causal claim that the doctor’s report gives one good reason to believe is true: that your illness was caused by your exposure to radiation. Similarly, although (if reasons are states) it will be false to say that you should give money to your old college because it cannot fund its scholarships, one can also readily construct from it a reason-statement that is true: that you should give money to your old college because of its inability to fund its scholarships. So, assume for the moment that the state-theorist is right and that reasons are indeed states of affairs. Let us say that a reason-statement is acceptable just if either it is true or, if it is false, it specifies a proposition as a reason for someone to do something and were one to reformulate the statement by

32 What Are Reasons for Action?

replacing its specification of that proposition with a term that refers to a corresponding state of affairs, the revised statement would be true.24 Any false reason-statement that fails to meet this condition is unacceptable. Thus, when an acceptable reason-statement is false, this is only because it specifies a proposition as a reason when it should refer to a corresponding state of affairs. The statement ‘A reason for Erika to give money to your old college is that it cannot fund its scholarships’, though false, is acceptable. The challenge to the state-theorist is now to show that his account of reasons does not itself render unacceptable too much of what we think true about the reasons people have to do things.

6 What is at stake, then, is whether one could eschew all the propositionspecifying constructions that are used to ascribe reasons to people and still be able, in principle, to state what reasons people have to do things. Given a reason-statement of the offending kind that would ordinarily be accepted as true, the state-theorist should be able to show that it is acceptable because there is a suitable state or event to be the reason for the agent to do what, according to the statement, he has reason to do. Obviously, there are very many cases for which this presents no difficulty. If I should comfort Brad because he is distressed, my reason for action depends on the disvalue of his distress – had that state not been obtained, I would have no reason to comfort him. Again, if you should apologise because Brad was offended, your reason for action depends on the disvalue of Brad’s being offended. In such cases as these, the state-theorist will certainly have a candidate state of affairs which he can identify as the reason for action. It is easy enough, however, to find reason-statements that are much less tractable. Consider, for instance, ‘The reason for us to bury him in unconsecrated ground is that he died by his own hand’. Here it is an event, an action, that is the source of, and so the candidate to be, the reason for the action and the state-theorist can readily supply ‘The reason for us to bury him in unconsecrated ground is his dying by his own hand’ which, if true, will make the original statement acceptable. It will be true just if that event is a reason for those agents to bury him in unconsecrated ground – but in that case, the truth of the statement will not depend on how one refers to the agents or to the event. So, ‘The reason for us to bury him in unconsecrated ground is his dying’ will also be true, since his dying is the same event as his dying by his own hand. Similarly, ‘We should bury him in unconsecrated ground because he died by his own hand’ will be acceptable if ‘We should bury him in unconsecrated ground because of his dying by his own hand’ is true. But if that is true so is ‘We should bury him in unconsecrated ground because of his dying’. Now, the state-theorist might accept this – after all, the latter does indeed specify which event stands as the reason to act, even if it does

Stephen Everson

33

not do so in a way that reveals why it so stands. He could point out that in causal statements too, one may refer to the causally related events in ways that are unhelpful to conveying why the cause was such as to produce its effect without thereby affecting the truth of the statement itself. The difficulty here, however, is that it does not answer to the way in which we do single out reasons. For in the example it would be both intelligible and true to say that reason for us to bury him in unconsecrated ground is not that he died but that he died by his own hand, and it would seem that distinction between what is the reason and what is not requires that both be propositions. Now, whether the state-theorist can adequately respond to this depends on how relaxed is his notion of a state of affairs – for if states of affairs can include states of events, he could then distinguish the state of the man’s death’s being self-inflicted from the event of the man’s death. The man’s death is not the reason but its being self-inflicted is. Nor would the appeal to states of events be just an ad hoc response to this particular difficulty, since we often cite as reasons facts about events and states. To make acceptable such statements as ‘I should take a stronger pain-killer because the pain is intense’ or ‘You should take a book because his speech will be interminable’, one would need to be willing to allow as reasons the intensity of my pain and the interminability of his speech. Indeed, one will need to be ready to find states of affairs at higher levels than these. Maybe I should only take a mild pain-killer because the pain will be intense briefly: is the real reason to be the brevity of the intensity of my pain? What if I should tell the doctor because my pain was unexpectedly briefly intense – is the reason here to be the unexpectedness of the brevity of the intensity of my pain? No doubt with a certain ingenuity and patience one could come up with examples of even greater complexity and the state-theorist could come up with suitably complicated definite descriptions that purport to refer to states of affairs. At a certain point, however, the states that the theorist would claim to have discerned by means of the constructed definite descriptions look less like genuine individuals than gerrymandered items whose only role is to satisfy the demands of a particular theory of reasons. Indeed, if the state-theorist is to have any serious chance of discerning states of affairs to match the range of our judgements about reasons, he will need to count more dubious items than these as states. Consider here an example where the reason for action relates to a future state of affairs. So, when Jane is packing for a holiday in Brussels, one might say that she should pack an umbrella because it will certainly rain several times while she is there. Now, to many the need to find that future states are reasons for present actions will seem sufficient to reject the state-theory – requiring as it does that an event which has not yet occurred or a state that has not yet obtained already stands in the relevant relation to an agent. Even if one were to grant the coherence of this for the sake of argument, however, the state-theorist would have to explain how we are able to identify the reasons

34

What Are Reasons for Action?

that are, on his view, future states. Here the force of the move from a propositional specification of reasons to one that refers to corresponding states of affairs is important. If I say that the reason to lock the dogs inside is that it will thunder this afternoon, I make no reference to any particular event – for ‘it will thunder this afternoon’ to be true, there will need to be at least one clap or growl of thunder during the afternoon, but there may be several. Often, of course, what grounds the thought that an event of a certain kind has occurred, is occurring or even will occur is that we know that a particular event of that kind has occurred, is occurring or will occur. Certainly, we can identify particular future events and states: I may know that the forthcoming wedding of two friends will be expensive and their marriage short lived. More usually, however, our beliefs about what will happen in the future are not grounded in this way. So, at the time when the reason-statement is made, neither the speaker nor the state-theorist is in a position to single out any of the downpours that will occur during Jane’s trip and so although Jane then has a reason to pack her umbrella, the theorist cannot specify that reason. This will not matter if one relaxes the conditions for the acceptability of reason-statements so that the revised statement may quantify over states and events. So the original statement will be acceptable since it is true that there will be at least one downpour during Jane’s trip and any such downpour is a reason now for Jane to pack an umbrella. The difficulty, of course, is that this (potentially) multiplies the number of reasons for Jane to take her umbrella. Where originally there was one reason for Jane to do this – the fact that it would rain – the state-theorist’s version will produce as many reasons as there are downpours during her trip. Indeed, if the state-theorist is right, rather than knowing, as one thought one did, what reason there is to pack the umbrella, one knows only that there is at least one reason of a certain kind to do this, without knowing which reason any is nor how many of them there are. The state-theorist could avoid this if he were willing to count sets of events as events – for then the reason for Jane to pack the umbrella will be the event that is the set of downpours that occur in Brussels during her trip. This matches our initial intuition in finding a single reason and one that can be identified at the time of the original statement. What would be needed to secure this, however, is some independent motivation for taking sets of events to be themselves events (and presumably, to deal with parallel cases, sets of states of affairs to be themselves states of affairs), and it is at least dubious whether that could be found.25 Nor do the state-theorist’s problems stop there. For it would certainly be wrong to think that Jane only has reason to pack an umbrella if it will rain during her trip – perhaps she should pack it because it is likely to rain or even just because it might rain. If the reason to pack the umbrella is that it is likely to rain, then that reason holds independently of any actual future

Stephen Everson

35

downpours, of which there may be none. This, albeit by an all too circuitous route, brings one to the central difficulty for the state-theorist. Often, when one deliberates about how to act, one considers the reasons for and against acting in some way and considers both the value of what one will bring about by acting in that way and the value of how things will be if one does not act in that way: so, the reason to call my friend now is to find out which article he suggested I read and the reason not to call him is that it is the middle of the night and I would wake him up. In deliberating, I balance the value of my finding this out and the disvalue of his being woken up. I have a reason to call and a reason not to call, but neither depends on an actual future state of affairs. For Dancy, the main objection to the idea that reasons are facts is that ‘those who announce that all good reasons are propositions seem thereby to lose contact with the realities that call for action from us’.26 If to maintain the thesis that all reasons are not propositions but states, the state-theorist is forced not only to look to states of states of states and sets of states and events but also to states and events in close (and maybe distant) but not actual possible worlds in order to find the reasons we have to do things, that objection is unlikely to carry much weight.

7 One difficulty with taking states of affairs to be reasons, then, is that they would need to be reasons even before they obtain. Facts, in contrast, are not temporally located and this too might seem problematic, given that what is a reason for doing something at one time can cease to be a reason. If reasons are facts, this would require the possibility that abstract items should undergo change – that what is true of an abstract object at one time is not true of it at another – and that itself may be thought sufficiently unattractive a consequence still to favour states of affairs over facts, despite everything else. We may first question whether it is actually true that in taking reasons to be facts, one would thereby be committed to thinking that a fact may change in respect of its being a reason for action. Certainly, one needs to have some kind of temporal index in statements of reasons for action, but it is not obvious that this should be an index for when a fact is a reason. Consider the ambiguity of the sentence ‘There is a reason now to comfort Brad’. This might signify that some consideration is now, and may not have been before, a reason for comforting Brad, but it might also signify that that some consideration is a reason for comforting Brad now. So, one can distinguish between ‘There is reason now to comfort Brad tomorrow’ and ‘There is reason now to comfort Brad now’.27 Having made this distinction explicit, however, it is questionable whether one needs after all to have the ‘now’ in the first sentence or the first ‘now’ in the second. Let us say that it is the fact that Brad is distressed during the morning of the 23 July 2002 that is the

36 What Are Reasons for Action?

reason for comforting him: this fact will favour actions of comforting him during that time, but will not favour such actions done before or afterwards. We do not, then, need to say that this fact becomes a reason for comforting him on that morning, or that it ceases to be a reason after he has been cheered up by lunch. It will rather be timelessly true that the fact is a reason for comforting him during the morning of that day – although of course there is a limited period during which one can act on that reason. Similarly, the fact that Celia is upset on the morning of her fortieth birthday favours organising something before her birthday to cheer her up when she wakes, but does not favour organising anything once it is too late for any plan to come to fruition on her birthday. Again, it is not that the fact becomes a reason at some point before her birthday and ceases to be such just before she reaches forty: it is timelessly a reason for performing acts of organisation during some period before her birthday (and we can note that what period that is may vary depending upon which type of organisation is in question).28 If we assume a timeless pattern of facts and triples of act-types, agents and times then we do not need to allow that facts may become or cease to be reasons for action. If there is a difficulty to be found in this response, it will be that it requires what many will regard as an excessively robust view of future contingents. For if it is to work, one needs to find in this pattern not only facts about the present and the past, but contingent facts about the future as well. Say that on Monday morning I am invited to a party to be held many miles away on Tuesday evening. To get to the party, I shall have to drive for three hours and so shall need to set off on Tuesday afternoon. So far, this is readily accommodated within the present treatment: the fact that I am invited on Monday morning is a reason for me to go to the party on Tuesday evening, and so to set off for the party on Tuesday afternoon. Just as I am leaving, however, I am called and told that the house where the party was to be held has burnt down and the party is cancelled. It seems that the reason I had for setting out for the party has now lapsed, since there will be no party to go to. To admit this appearance, however, is to move back to thinking that facts may change in respect of their being reasons for action. It is still, of course, a fact that I was invited on Monday morning, but whilst that fact was a reason for my setting out on Tuesday afternoon, it is not so once the house has burned down. Now, it is indisputable, I think, that once the party has been cancelled, I do not have reason to set out, and so if the advocate of timeless reason-relations is to save his position, he needs to deny that the fact that I was invited to the party was ever a reason for going to it on Tuesday. This can be done readily enough if he can make use of the fact the party is cancelled on Tuesday afternoon: an invitation to a party that is not to happen provides no reason to set out for it. What would have been the reason-giving character of the fact that I am invited on Monday is cancelled by the fact that the party is called off on Tuesday. Since the party was never going to

Stephen Everson

37

happen, I had no reason to go to it, or to do anything to prepare for going to it. Until the phone call, of course, and not knowing the future, I believed that I had reason to go to the party, and this explains why I should have begun to get ready for it: my belief, however, was false, and made false at the time by the fact that the party would not take place. The vulnerability of this account lies in its reliance on facts about future contingents: for it to work, one has to assume that at the time of the invitation (and, indeed, before) it was a fact that the party would not take place.29 Whilst this is not obviously wrong, it is certainly controversial. Someone who believes that the truth-value of a proposition may vary over time, and who denies that bivalence holds for contingent propositions about the future will not accept it.30 To hold that view, however, is to undercut the motivation for moving to the tenseless understanding of reasons at all. For the anti-realist about future propositions precisely accepts the possibility of abstract objects’ changing over time: the whole point of the position is that a proposition can lack a truth-value at one time but not at another. Such a theorist cannot, then, find it worrying in principle that a proposition may at one time be a reason for someone to act in some way but not be such a reason at another. Indeed, the idea that a proposition may change in respect of being a reason is, if anything, less worrying than the idea that it may become true. For, depending on how one understands what it is for a proposition to be true, one will need to find that, in becoming true, a proposition changes in respect either of a non-relational property, or of a relation to another abstract object, that is, a truth-value.31 To be a reason, in contrast, a fact stands in relation to a concrete particular, that is, an agent, and to change in respect of this relation is correspondingly less problematic. Accepting this possibility is no more difficult than allowing that the same proposition may be believed by someone over a period of time, but rejected by him after that, or that whilst some positive number may be the number of books read by the president of the United States, it may yet cease to be such after a year or so. If we are going to allow the possibility of relations between abstract objects and concrete particulars at all, we have to allow for the possibility of changes in respect of those relations.

8 Given the variety of temporal relations between reasons and the actions they are reasons for, facts are thus better placed than concrete particulars to stand as reasons for action. If there are still concerns about the suitability of facts to stand as reasons for action, these will not be because facts are abstract items, but because of the kind of abstract item they are. If facts are true propositions, and have concepts as constituents, then in Fregean terms, they inhabit the realm of sense rather than that of reference – and it would not be mere prejudice to think that this is the wrong metaphysical kingdom to

38

What Are Reasons for Action?

find reasons for action. Holding to the view that reasons are states of affairs, one can take the relation between reasons and the actions they are reasons for to hold independently of how they are conceived, but this would seem to be compromised if one rather takes reasons to be facts. According to advocate of states of affairs, it is, say, the distress of that woman that is the reason for avoiding her. If that woman is Baroness Thatcher, then Baroness Thatcher’s distress is the reason for avoiding her, as is the distress of Dennis Thatcher’s wife. In discerning the reason to avoid her, it will not matter how we single out the relevant state of affairs: ‘C is a reason for a to w’ will, like ‘e is the cause of f’ to be extensional. In contrast, the fact that Dennis Thatcher’s wife is distressed and the fact that Baroness Thatcher is distressed are distinct – and this presents the apparently unwelcome possibility that one but not the other might be a reason for comforting the woman in question.32 This possibility would be unwelcome, for instance, if we thought that reasons should be what might be called ‘perspective-independent’.33 I take that term from Christopher Peacocke, and it is helpful here to compare an argument he gives for ‘the perspective-independence of truths about causation and explanation’: Suppose you ask me why there is a cloud of steam in the air, and I answer, pointing to a metal container, ‘Because that boiler is leaking’. We then walk around to the other side of the same boiler, and I add, ‘But the cloud isn’t there because that boiler is leaking’, while manifestly pointing to the same boiler. You would rightly be totally baffled. You would insist that what does the explaining is at the level of reference, not the level of sense, so that a true explanatory statement is insensitive to which modes of presentation are used to pick out the objects, events, properties, and relations being talked about.34 We can, it would seem, apply similar considerations to reason-statements. Let us take a reason-statement to be one that can figure in the context ‘A should w because ...’ to produce a true sentence.35 So, if you should see your mother because she is missing you, then ‘she is missing you’ states the reason for you to go to see her, and if you should save for a pension because state provision has been all but withdrawn, then ‘state provision has been all but withdrawn’ states the reason for you to save. 36 We could, then imagine an example parallel to that offered by Peacocke: walking over Hungerford Bridge, we find someone injured. I say ‘We should call an ambulance because that woman is hurt’. As we walk past her, I then say, manifestly indicating the same woman, ‘But it is not because that woman is hurt that we should call an ambulance’. You would undoubtedly be baffled, and if you were to respond in a similar spirit to Peacocke’s interlocutor, you would insist that reasons are at the level of reference, not the level of sense, so that a statement of a reason for action is insensitive to

Stephen Everson

39

which modes of presentation are used to pick out the objects, events, properties, and relations being talked about. At first sight, it might look as if this argument is guilty of a straightforward ignoratio elenchi: to claim that a context is intensional is to claim that substitution of a constituent term, preserving its reference but not its sense, can change the truth-value of the sentence of which it is a constituent. Clearly, one is not committed to claiming that such substitutions must change its truth-value (not least because there may be more than two terms which share reference but not sense). It is not, however, that we are supposed to take the first speaker – Peacocke, as it might be, in the example – as holding that in thinking of the boiler under a different demonstrative mode of presentation, he is not in a position to refer to the boiler in order to explain the steam. Rather, the idea must be that if we take the ‘because’ to create an intensional context, then we should always be able to make sense of someone’s taking opposed attitudes to sentences occurring within that context and which differ in that their constituent terms have the same referents but different senses. This is why it does not beg the question for Peacocke to specify that he points to what is manifestly the same boiler when he utters the two token sentences. So, I may know that Hesperus is Phosphorus and still find it intelligible that someone else believes that Hesperus is bright when he does not believe that Phosphorus is bright. In the case of the ascription of propositional attitudes, it does seem right to maintain that given any two terms T1 and T2, which differ in sense but not in reference, it will be intelligible to find that someone believes one proposition, expressible by a sentence S1 containing ‘T1’, but disbelieves another proposition, expressible by a sentence S2 that differs from S1 just in that it contains ‘T2’ where S1 contains ‘T1’. Care is needed in articulating this thesis, however. For there is obviously a crucial difference between thinking that, given any such pair of sentences, one can imagine circumstances in which a subject may hold one to be true but the other false, and thinking that this will make sense in any circumstance. So, our ability to find that someone believes that Hesperus is bright and that Phosphorus is not bright depends on our believing that he does not believe that Hesperus and Phosphorus are the same heavenly body (indeed, it will itself count as evidence for this). In Peacocke’s example, in contrast, both the speaker and his interlocutor know that the boiler demonstrated in the two utterances is the same. Now, this in itself is not problematic for Peacocke, as it would have been if the speaker had simply said ‘That boiler is leaking’ and then ‘That boiler is not leaking’. Someone who opposes Peacocke’s account of explanation and maintains that explanatory truths are perspective-dependent, must be prepared to accept that it can be true that p1 because S1 but false that p1 because S2, even though he knows that ‘T1’ and ‘T2’ refer to the same thing. What he does not need to accept, however, is that, for any such pair of

40 What Are Reasons for Action?

sentences, it will always be possible to find an explanandum such that if one explains it the other will not, nor that it will always be intelligible to ascribe the belief to someone that one sentence has an explanatory role which the other lacks. The point here is that distinctions in sense are secured by the need for cognitive differences between propositions, but there is no reason to think that explanatory relevance must cut as finely as that. One can deny that coreferring terms may always be substituted salva veritate in sentences occurring in explanatory contexts without thinking that any such substitution will be such as to produce a sentence that, in principle, has a different explanatory role, or could intelligibly be taken by someone to do so. So, in the example, a safer, because more cautious, response on the part of Peacocke’s interlocutor would be to express bafflement that Peacocke should think that the shift from one visually demonstrative mode of presentation of the boiler to another could affect the explanatory relevance of the resulting thought. If one may explain the fact that the room is steamy by singling out the responsible boiler by means of a visually demonstrative mode of presentation, then it should not matter which such mode of presentation is employed.37 For a theorist who maintains that the truth of explanatory statements is sensitive to the senses of the terms employed, a part of his theory will be to give some account of the kinds of change of modes of presentation that might be such as to change the truth-value of the explanation.38 Similarly in the case of reason-statements. That facts are individuated at the level of sense does not mean that if one takes reasons to be facts, one will thereby need to find intelligible someone’s holding that any two facts differ in respect of which actions they are reasons for. Again, the specification of one’s reason to call an ambulance will not depend on which perceptually demonstrative mode of presentation is used to pick out the injured woman, and one would be duly baffled if someone thought that it did. In other cases, however, it would be less baffling if someone were hesitant to allow substitution of co-referring terms when specifying reasons for action. Thus whilst Oedipus should go to the church because his fiancée is waiting there to marry him, it is dubious that he should go because his mother is waiting there to marry him.39 If there are such cases, then this will be sufficient to motivate the claim that reasons are to be found in the realm of sense – even if many reason statements will tolerate perfectly happily the substitution of co-referring terms.

9 These considerations are important because they bear on how we should think about the relation between accounts of reasons and theories of propositions. Consider a final argument against taking reasons to be propositions rather than states of affairs – that doing so will result in an absurd

Stephen Everson

41

inflation in the number of reasons one has to act. Allow that facts as propositions are to be individuated at the level of sense: then, if ‘S1’ and ‘S2’ differ in sense, the fact that S1 will differ from the fact that S2. Taking reasons to be facts that favour courses of actions, then if the fact that S1 favours my w-ing, that fact will be a reason for me to w. Similarly, if the fact that S2 favours my w-ing, this too will be a reason for me to w. It is, of course, unproblematic that someone may have more than one reason to do something, but we have here the materials for generating many more reasons for action than is tolerable. So, it may be that the fact that Jane is waiting at the church to marry Jack is a reason for him to go to the church. If Jane is Theresa’s daughter, then it will also be a fact that Theresa’s daughter is waiting at the church, and this too will be a reason to go to church. Do we want, however, to say that the fact that Theresa’s daughter is waiting at the church is a further reason for Jack to go there over and above the fact that Jane is waiting there for him? Should Jack reflect on the reasons he has not to pull out of the wedding, that doing so would distress his mother does seem to be an extra reason beyond the fact that it would distress Jane, whereas the fact that it would distress Theresa’s daughter does not. And, since no doubt Jane, like anyone else, will satisfy indefinitely many definite descriptions, this would mean that there are indefinitely many reasons for Jack to do the same thing. What this line of objection assumes is that, having decided that reasons are propositions, one will need to appeal to some prior account of propositions in order to individuate reasons. So, if the fact that p is a reason to w, and if the fact that q is a reason to w, in order to tell whether that p is the same reason to w as that of q, one will need to determine just whether the proposition that p is the same as the proposition that q. Identity and difference of reasons will simply fall out of an account of the identity and difference of propositions. A general theory of propositions, however, is answerable to the needs of a theory of thought: in so far as it is possible to believe that p but not to believe that q, ‘p’ and ‘q’ will express different propositions. We will, that is, also want to appeal to a theory of propositions to secure our judgements as to when people believe and assert the same things, as well as to when they have the same reasons for action. If it is possible to believe that Jane is waiting at the church but not to believe that Theresa’s daughter is waiting at the church, then the proposition that Jane is waiting at the church must be different from the proposition that Theresa’s daughter is waiting at the church. If, however, these are different propositions, and reasons are propositions, then they are different reasons and the theory runs into the problems just identified by the advocate of states of affairs. Even if we leave aside the need for propositions to serve as reasons for action, however, this model mis-represents the theoretical role of a theory of propositions. So, whether we take two people to assert the same proposition depends on whether we take them to say the same thing and this

42 What Are Reasons for Action?

depends on whether we take them to be, in Davidson’s terms, ‘same-sayers’. What the model just given rests on is the idea that there can be only one set of criteria relevant for judging whether people are same-sayers, or, if one likes, same-believers. That idea, however, is hardly self-evident. Sometimes our interests will be such that if people believe of the same thing that it has some property then this will be sufficient to find them same believers, even if they think of that thing in different ways. In other contexts, we should judge them to believe different things.40 It would be a mistake to think that the task of a theory of propositions must be to set out criteria of identity and difference for entities of a single kind: rather, it may be to bring some theoretical clarity to the different contexts where we will explain sameness of judgement or assertion by appealing to sameness of what is asserted or judged. Such clarity may be gained by finding that there are different kinds of proposition, because different criteria for identity and difference of propositions determined by different contexts in which we are interested in interpreting subjects and their utterances.41 It is one thing, however, to allow that there may be a kind of proposition – ‘robust propositions’, say – such that ‘the robust proposition that p’ allows for substitution of co-referring terms in the sentence which expresses it, and another to think that reasons for action will be propositions of that kind. At the end of the previous section, I suggested that whilst the fact that Oedipus’ fiancée is waiting at the church to marry him is a reason for him to go to the ceremony, the fact that Oedipus’ mother is waiting at the church to marry him is not. Since these two propositions are distinct only in virtue of differences in the senses of the terms in the sentences that express them, this would require that reasons are not robust. On reflection, however, it is less obvious that the two facts do differ in respect of what they are reasons for Oedipus to do: given that his fiancée is his mother, that she is waiting at the church to marry him is not a reason for him to go to the church – and certainly not to marry her. Since ‘Oedipus’ fiancée is waiting at the church to marry him’ and ‘Oedipus’ mother is waiting at the church to marry him’ express the same robust proposition, the example may after all be consistent with taking reasons to be robust propositions. There is clearly much more to be said about that issue, but in fact there is no need to settle it here. For even if it does turn out that reasons are not robust – and it is difficult to see how considerations that would support this could be of any comfort at all to a theorist who wanted to take reasons to be states of affairs – this is not in fact worrying in the way the objection claimed it to be. It would only be worrying if an aim of deliberation were to determine how many reasons there are for and against some possible action, but it is not. We operate with two different notions of reason: one is, as we might say, a count-notion whilst the other is a mass-notion. So, we can ask whether this reason for w-ing is the same as or different from that reason for w-ing, and we can also ask how much reason there is to w or whether there

Stephen Everson

43

is more reason to w than there is not to w. Certainly, when we decide how we should act, we will be concerned to determine what reasons there are for and against some course of action, but it would be a very crude mistake to think that if there are more reasons to w than there are not to, this would provide even prima facie support for inferring that there is more reason to w than not to w. The concern to determine what reasons there are for some course of action is the concern to make sure that all relevant facts are taken into account; it is not to know how many reasons there are. Rather, it is only when one has determined which facts are relevant that one can decide whether there is more or better reason to w than not to.42

Notes 1. J. Raz, Practical Reasons and Norms (Oxford: Oxford University Press, 1999), 15–16. This was originally published in 1975. For Austin’s views, see his ‘Unfair to Facts’, reprinted in S. Blackburn and K. Simmons (eds), Truth (Oxford: Oxford University Press, 1999), 183–99. See also G.H. von Wright, Norm and Action (London: Routledge and Kegan Paul, 1963), 25–7. 2. P.F. Strawson, ‘Truth’, reprinted in Blackburn and Simmons, Truth, 162–82, p. 167. 3. So Strawson’s complaint against Austin was not that he had misunderstood what a fact is, but rather that facts as Austin conceived them could not do the work he required of them in explicating truth. 4. Which is perhaps fair enough, since nothing in what he goes on to argue turns on this. Having formally identified ‘facts’ as what are designated by ‘The fact that ...’, he proceeds simply to cite facts using that-clauses so that neither his formulations nor his arguments are hostage to his understanding of what facts are. (Indeed, some of his formulations rather go against that understanding, as when he gives as a possible reason to drive slowly ‘the fact that unless one drives slowly an accident will occur’ (21–2) and it is not obvious how one will find a state of affairs to stand as the denotatum of that.) 5. Jonathan Dancy, Practical Reality (Oxford: Oxford University Press, 2000), 115. 6. In Dancy’s case, something does turn on his claim about the nature of reasons: he makes it as part of an attempt to rebut the idea that one might co-ordinate the psychological and rationalising explanation of actions by taking normative reasons for actions to be the contents of the beliefs that give rise to actions. In view of the importance of his claim it is unfortunate that he does so little to argue for it. 7. I use ‘consideration’ just to provide a term that is neutral between ‘fact’ and ‘state of affairs’. 8. Such independence is not, of course, an immediate consequence of rationalism: it will still be open to the theorist to argue that practical reasons are ‘internal’, in that whether something will count as a reason for someone to do something will depend on whether he believes that it is or could in some suitable way be brought to do so. For such an argument, see B. Williams, ‘Internal and External Reasons’, in his Moral Luck (Cambridge: Cambridge University Press, 1981), 101–13. An effective rebuttal of Williams’ argument is provided by Derek Parfit in his ‘Reasons and Motivation’, Proceedings of the Aristotelian Society, suppl. vol. 71 (1997), 99–130, 9. S.L. Darwall, Impartial Reason (Ithaca: Cornell University Press, 1983), 31. Similar claims are made by T. Scanlon, What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998), 56–7.

44 What Are Reasons for Action? 10. Impartial Reason, p. 33. 11. ‘Internal Reasons and the Obscurity of Blame’. In his Making Sense of Humanity (Cambridge: Cambridge University Press, 1995), 35–45, 39. Italics added. 12. At least that would seem to be the implication of this, given that the relata of an explanation are facts. There is a tension between this and his claim that reasons can ‘in some form’ themselves be in the agent’s motivational set – which might seem rather to suggest that he would agree with Davidson in taking them to be psychological states. Whether there is a conflict here will depend on how loose the conditions are for something’s being in the set: for instance, it might be that it can be there as the content of a belief rather than as a belief or desire itself. Williams himself is, for these purposes, unhelpfully vague about what kinds of things do make up a motivational set: ‘desires, evaluations, attitudes, projects, and so on’ (‘Internal Reasons and the Obscurity if Blame’, 35). It would be helpful here to be told whether something, perhaps a ‘project’, can be an element in someone’s motivational set without being an element in his psychology – and, if so, how. 13. J. Searle, Rationality in Action (Cambridge, MA: MIT Press, 2001), 100. It cannot be said, I think, that these claims support the idea that we can properly talk of the notions ‘Why” ’ and ‘Because’, but Searle’s argument can lose this idea without damage. 14. Rationality in Action, 101. 15. Rationality in Action, 103. 16. Rationality in Action, 100. Italics added. Taken strictly, this last claim seems inconsistent with Searle’s claim that one may have a good reason for doing something without knowing it: if one had a reason for doing something but did not know that one did, then one would not have acted on that reason, and so the reason would not have been a reason why one acted. Rather, then, than saying that all reasons are reasons why, it would be better to maintain with Williams that all reasons are such that they can be reasons why: this would allow Searle his formal constraint on being a reason, but would allow that not all reasons actually enter into explanations. 17. ‘Internal and External Reasons’, in his Moral Luck (Cambridge: Cambridge University Press, 1981), 101–13, 102. 18. See P.F. Strawson, ‘Causation and Explanation’, in B. Vermazen and M.B. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford: Oxford University Press, 1985), 115–36, 115. 19. ‘Agency, Reason, and the Good’, p. 22. 20. Dancy, Practical Reality, p. 115. Compare Raz: ‘reasons are facts in virtue of which those actions are good in some respect and to some degree’ (‘Agency, Reason, and the Good’, p. 23). Raz is here to be understood, I think, as again using ‘fact’ in an Austinian way: see above, p. 23. 21. This is not an argument that Dancy himself offers to support his view, however, and not all the examples he cites of states that can be reasons – ‘his self-satisfaction, her distress, yesterday’s bad weather, the current state of the dollar’ – are readily taken as having value or disvalue independently of particular goals. It would give him an argument for his position, though. 22. The first of these does not contain a that-clause, but it still uses a sentence to specify the reason. ‘That’ is not required here since ‘because’ itself connects to a sentence directly. Nor should one be misled by the ‘because’ into thinking that this is an explanation and so must relate facts: if it is an explanation it is along

Stephen Everson

23.

24.

25.

26. 27.

28.

29.

30.

45

the lines of an Aristotelian formal explanation so that the ‘because she is miserable’ has the force here of ‘in virtue of this reason: that she is miserable’. Dancy himself recognises that we use that-clauses in the relevant contexts: ‘Of course we do, or at least can, say such things as “That she is in distress is what made his action callous, and that she is in distress is a proposition.” ’ (pp. 115–6). He counters this by pointing out that ‘we also say things like “That the cliff was unstable was a consequence of the heavy rain, and that the cliff was unstable is a proposition” ’ and pointing out quite rightly that ‘we would be unwise to conclude from this that a proposition was a consequence of the heavy rain’. Disappointingly, rather than seeing that this shows ordinary usage to be, from a stern metaphysician’s point of view, worryingly lax about the categorical niceties, Dancy thinks that it shows that ‘though we use that-clauses to specify propositions, not all uses of such clauses are in the proposition-specifying business’. Clearly, if this notion of acceptability were made to bear any great theoretical weight, it would be necessary to define the relation of correspondence between facts and state of affairs that it relies on. I shall not attempt to do that here since if my argument is right, the prospects for a successful defence of a state-theory of reasons are so dim as to make it pointless. For a hostile discussion of that possibility in the case of events, see Jennifer Hornsby, ‘Physicalism, Events, and Part-Whole Relations’, in her Simple Mindedness (Cambridge, MA: Harvard University Press, 1997), 46–62. Dancy, Practical Reality, 114–5. Of course, it is unlikely that anyone would ever utter the second sentence, since there is a general presumption that when one says that someone has a reason to do something this will be reason for acting at the time of utterance or as soon as is possible or appropriate thereafter. One might wonder why the advocate of states-of-affairs should not help himself to the same kind of timeless relations in order to allow the possibility that future states might stand as reasons for present actions: but part of the point of taking reasons to be items that are spatio-temporally located is to disallow their standing in such timeless relations. Cp. E.J. Lowe, The Possibility of Metaphysics, (Cambridge: Cambridge University Press, 1998), 89. Note here Michael Dummett’s distinction between the use of the future tense to talk about present ‘tendencies’ and what he calls the ‘genuine future tense’. So, he finds the former in a standard form of newspaper announcement – ‘The marriage that was arranged between X and Y will not now take place’: ‘The presence of the “now” indicates a use of the future tense according to which, if anyone had said earlier, “They are going to get married”, he would have been right, even though their marriage never subsequently occurred.’ He contrasts this with the case where someone offers a bet by saying ‘ “I bet they will not be married on that date” ’, employing the genuine future tense. (‘Bringing about the Past’, in Truth and Other Enigmas (London: Duckworth, 1978), 333–50, 336; the same distinction is made by means of the same example in ‘Truth’, and in ‘Realism’, 21, 152). It must be the genuine future tense that would be in question for the example in the text. For a discussion of different possible semantic treatments of future-tense statements, see M. Dummett, Frege: Philosophy of Language, 2nd edn, (London: Duckworth, 1981), 390–400. See also ‘Realism’, 152–3.

46

What Are Reasons for Action?

31. Which may be why Frege was so opposed to allowing that a proposition might change its truth-value: see his Kleine Schriften, ed., I. Angelelli (Hildesheim: Georg Olms, 1967), 338. 32. Of course, seeing that some woman is distressed, I might think that the fact that that woman is distressed is a reason to comfort her, but, on discovering that that woman is Baroness Thatcher I might think that the fact that Baroness Thatcher is distressed is not a reason to comfort her (she might, for instance, think that such behaviour was feeble): this would, however, be to revise my original judgement. It is not that I am left thinking that the fact that that woman is distressed is a reason for helping her, but the fact that Mrs. Thatcher is distressed is not such a reason. Whether I have a reason here is dependent on who is distressed and not on how I identify or think about him. 33. This would not be to deny that reasons are relative to agents: it is quite consistent with this to maintain that some consideration can be a reason for one person to do something but not a reason for another to do it. The point will just be that if it is true that C is a reason for a to w, it will not matter how one thinks of the relata of that relation. 34. C. Peacocke, Being Known (Oxford: Clarendon Press. 1999), 73–4. 35. If the connective ‘because’ renders this an explanatory context, it will be one of formal or constitutive explanation: it is of the same type as, for instance, ‘This is a piece of furniture because it is a chair’ or ‘He is careless because he does not pay sufficient attention to semantic details’. 36. Note that it doesn’t follow from this that the reason itself is propositional, any more than it follows from the ability to say that the floor is wet because it’s raining that the cause of the wet floor is propositional. 37. Which is why one might think that Peacocke is right to say that explanatory truths are not perspective-dependent: the point would be that it is too hasty to move from this claim to the claim that they are at the level of reference rather than of sense. 38. Pursuing this matter further for the case of causal explanation will take me too far away from the topic of reason-statements. We should note, however, that an explanans of the kind that Peacocke gives in the example is indeed likely to be more robust than one that refers directly to a causally responsible event. Suspicion of a purely extensionalist account of explanation is likely to be motivated by the worry that a true explanation should be such that the knowledge of it will provide some level of causal understanding of what is explained, and that there are ways of thinking about the cause which will not be such as to provide this. So, to take a familiar example, one can reasonably doubt that knowledge that the event described on page four of the Times caused the event described on page nine of the Tribune is sufficient to provide the basis for knowledge of an explanation: to suppose that it is true that the event described on page four of the Times happened because the event described on page nine of the Tribune happened is to suppose that knowledge of this would constitute understanding of why the first event happened, and this is certainly deniable, as it would also be in the case of ‘The event described on page four of the Times occurred because its cause did’. The difficulty, of course, is that neither of these would give the least sense of what about the two events was relevant to their standing as cause to effect: in contrast, ‘the flooding occurred because the boiler exploded’ does help to make this clear, since the explanans here specifies that an event of the relevant kind occurred. Once this is

Stephen Everson

47

done, then there will be no difficulty in accepting that one might present that kind by means of a different mode of presentation, if one is available. 39. Compare here Anscombe on causal explanations: let us assume that ‘There is an international crisis because the President of France made a speech’ is true, and that de Gaulle, the President of France, is also the man with the biggest nose in France. Is ‘There is an international crisis because the man with the biggest nose in France made a speech’ also true. As Anscombe points out, our objection to this last would be ‘But surely not because he is the man with the biggest nose in France?’ As she also goes on to point out, a case of this kind ‘cries out for a Russellian kind of treatment’, so that one will distinguish between Concerning the man with the biggest nose: there was an international crisis because he ... and There was an international crisis because the man with the biggest nose ... ‘– the latter, since it is differentiated from the former, now being false – unless, as we say, “it was because his was the biggest nose” ’ (‘Causality and Extensionality’, in her Metaphysics and the Philosophy of Mind (Oxford: Blackwell, 1981), 173–9, 175–6. Similarly, if one says that Oedipus should go to the church because his mother is waiting there to marry him’, one’s doubt about the truth of this will be expressed by saying ‘But surely it is not because she is his mother that he should go to marry her?’ 40. In Joseph Moore’s interesting discussion of these matters (‘Propositions without Identity’, Nous 33 (1999), 1–29), he contrasts cases of reported speech where we are interested in the subject’s psychological states ‘only in so far that they carry useful information about the world outside’ and cases where one is rather interested in predicting the actions of the subject (pp. 7–9). Cases of the first kind will allow one to substitute co-referring terms in the sentences that report the belief, whilst cases of the second will not. 41. See Jane Heal, ‘On Speaking Thus: The Semantics of Indirect Discourse’, in her Mind, Reason and Imagination (Cambridge: Cambridge University Press, 2003), 174–95, pp. 189ff. 42. An earlier version of this paper was given to a conference organised at the University of York by Christian Piller. I am grateful to him and to others at the conference, especially David Velleman, for helpful discussion. I am also grateful to Joseph Raz, Tom Baldwin, Marie McGinn and Ian Rumfitt for commenting on a very early draft. I was led into thinking about these matters as part of working on a larger project about explanation and ontology in the Philosophy of Mind – a project whose progress has been greatly facilitated by a grant from the AHRC which allowed me a term’s relief from teaching.

3 Was Sally’s Reason for Running from the Bear that She Thought it was Chasing Her? Rowland Stout

Sally thinks she is being chased by a bear, and runs away. Let us suppose that running away makes sense in the circumstance.1 It seems clear that her reason for running is that a bear is chasing her. But it also seems that her reason for running is that she thinks a bear is chasing her.2 Indeed it is sometimes asserted that her real reason cannot be that a bear is chasing her, but must be merely that she thinks or believes that a bear is chasing her. For example, Michael Smith has argued as follows: Given that an agent who has a motivating reason to w is in a state that is in this way potentially explanatory of her w-ing, it is then natural to suppose that her motivating reason is itself psychologically real. ... By contrast with normative reasons, then, which seem to be truths ... motivating reasons would seem to be psychological states, states that play a certain explanatory role in providing action. (Smith 1994, p. 96) The question now for this psychological sort of approach to reasons for action is what sort of thing the reason is supposed to be. Is it a psychological state, as Smith has it here – for example, Sally’s state of believing that a bear is chasing her? Or is it the fact that she is in this psychological state – the fact that she believes a bear is chasing her? And is her belief that a bear is chasing her, the same as one of these or something else and perhaps a better candidate for being Sally’s reason for running. Smith (1994) does not distinguish between the belief and the psychological state. And John Searle makes nothing of the distinction between the belief and the fact that the subject has that belief. He claims (2001, p. 102) that ‘we can say either the fact that I believed it is the reason or my belief is the reason’. However Roger Crisp (2006, p. 38) is careful to say that my reason for acting is my belief but not my having that belief.

48

Rowland Stout 49

The advantage of thinking of Sally’s belief as her reason is that it seems to combine elements of both the fact that a bear is chasing her and the fact that she believes a bear is chasing her. Sally’s belief that a bear is chasing her has the same content as the fact that a bear is chasing her – the belief and the fact both are that a bear is chasing her. So one might think that they both justify her behaviour in the same way. At the same time it looks as though the belief is a better candidate for being her reason for running since it has that psychological element that Michael Smith is after. It seems that her belief that a bear is chasing her can both justify and explain her running. Nevertheless I think that the idea of Sally’s belief is metaphysically indeterminate in a way that gives the impression that it can have this dual role but really means that this impression is just a confusion. When we can say that Sally believes that a bear is chasing her we can then also say that she has the belief that a bear is chasing her. But saying that she has this belief is saying nothing more than that she believes it. There is no independently identifiable entity – her belief – that we can observe somehow in her possession and in virtue of which we say she has it Talking of Sally’s beliefs is completely derivative from talking of what Sally believes. Once we start talking about beliefs we can begin to think about identity conditions for those beliefs. But there are several choices here none of which is really forced on us. If Sally and I share a belief that a bear is chasing her does that mean that there is just one belief here that is literally being shared? If so, then, it is not psychologically real in the sense that Smith is after. If we say that there are two beliefs here – Sally’s and mine – what about their identity over time? Suppose Sally has a belief, stops having it, then believes exactly the same thing again. Does she have the same belief as she had before, or a different belief with the same content? Our talk of beliefs is not determinate with regard to these questions. (Nor do I think is our talk of states of believing.) Sally’s belief is a sort of hybrid between what she believes and the fact that she believes it. We might translate the phrase ‘Sally’s belief’ roughly as ‘what Sally believes being believed by Sally’. This can be true or false in line with whether what Sally believes is true or false but at the same time can explain things in line with whether the fact that Sally believes it explains these things. Given this derivativeness and indeterminacy in the idea of Sally’s belief I will not consider it as a candidate for what her reason for running might be but consider just two other candidates, the second of which represents the psychological approach: a. that she is being chased by a bear; b. that she thinks/believes that she is being chased by a bear.

50 Sally’s Reason for Running from the Bear

I will argue that there is a case against (b) being her reason for running away which is effective. The problem is that there also seems to be a pretty good case for thinking that (b) is her reason for running away. So we have an apparent contradiction. Sorting out this contradiction promises to provide some insight into the nature of reasons for action and the relationship between normative and motivating reasons.3 The first thing to consider is how the word ‘for’ is used in this context. As is often remarked, there are in fact at least two quite distinct uses corresponding respectively to the two phrases, ‘reason for an action’ and ‘acting for a reason’. We say that such-and-such is a reason for Sally to run. We also say that Sally runs for such-and-such a reason. And these mean quite different things. So consider the following: A. A reason for Sally to run is that such-and-such. B. Sally runs for the reason that such-and-such. In the first phrase, ‘reason for’ is similar in meaning to ‘reason in favour of’; ‘for’ is the opposite of ‘against’ here. There could be a reason for, or in favour of, Sally running even if she does not run. Such reasons do not need to be explanatory, since there may be such reasons even when there is nothing to explain – for example, when Sally does not run. A reason for Sally to run – one that favours her running – is often described as a normative reason. The existence of such reasons generates the possibility of Sally getting things right or wrong, succeeding or failing, acting correctly or making a mistake. We talk of reasons for action in order to set up norms for action. Rüdiger Bittner (2001) however denies that such reasons are really normative at all; they are merely states of affairs to which actions might be responses. This connection between reasons and responses is remarked independently by others too – in particular Stoutland (1998). But although the English language is not definitive about this the notion of a response itself appears to be normative; it is precisely responses rather than reactions which we can describe as appropriate or inappropriate. This does not mean that it follows from Sally having a reason to run that she should run. But it does mean at the very least that there is a systematic way of determining appropriate and inappropriate responses – a way of deriving recommendations for her actions – which given the reasons and in the absence of countervailing considerations yields the recommendation that she should run. This makes room for the possibility that the system of deriving recommendations is a bad one. For example, consider a West Ham football club hooligan’s system of justification, according to which someone being a Chelsea supporter is a reason to attack them. There are different sorts of claims that can be made here. We could make the existential claim that there is some system of justification relative to which it is recommended that he attack the Chelsea fan. This would be a very weak claim to make

Rowland Stout 51

however, since, if we allow bad systems of justification, then it would be true of just about every possible recommendation for action. Bittner takes reason-giving statements to be very weak existential claims of this sort. We could also say however that relative to a certain system of justification – perhaps the West Ham hooligan’s one – it is recommended that he attack the Chelsea fan. If we say of the hooligan that he has a reason to attack the Chelsea fan we may mean that according to his system of justification he has a reason to do so. Such a claim is relatively normative. I think that Bernard Williams (1981) is concerned with such claims in his discussion of internal and external reasons. If a reason is only a reason relative to the agent’s system of justification then in a sense it is internal to that agent’s motivational system. The third sort of claim that might be made here is an absolutely normative one. Instead of making a claim about a system of justification, we might be endorsing a particular system of justification (say the hooligan’s one) and making the absolute (and absolutely false) claim that attacking the fan is recommended. This sort of claim is straightforwardly normative. Saying that there is a reason to attack the Chelsea fan sounds like an example of this sort of claim. That the fan is a Chelsea supporter is in fact no reason to attack him in this sense. I think Parfit (1997) and Dancy (2000) are contemplating this sort of claim when they describe reasons as both external and normative. We can also use the phrase ‘reason for’ without any of these normative connotations. Consider the following: A*. A reason for the water running out of the tank is that there is a leak in the tap. There can be no such reason unless the water actually runs out of the tank, and there is no sense in which the water either succeeds or fails with respect to this reason. The fact that there is a leak in the tap favours the process of the water running out of the tank only in as much as we can infer that the water will run out of the tank from that fact. We can also say the following: A**. A reason for the water to run out of the tank is that there is a leak in the tap. To my ear at least there is an ambiguity in this sentence. On the one hand it might be used to state a possible explanation of something that may or may not happen. It is like A* but in a hypothetical mode. On the other hand the sentence might be used to express the absurd idea that the water might have a reason – a normative reason – for action. It is not difficult in any case to limit our use of the phrase ‘reason for’ to cases where what is described is a normative reason, and keep formulations such as A* and A** out of the picture. In B on the other hand the reason is simply explanatory. If Sally runs for the reason that a bear is chasing her or that she thinks a bear is chasing her

52

Sally’s Reason for Running from the Bear

then the fact that a bear is chasing her or that she thinks a bear is chasing her explains her running. In this phrase, ‘for such and such a reason’ means something like ‘on account of such and such a reason’. It is an awkward kind of expression, and we would usually employ one of the following expressions instead: ‘Sally runs because a bear is chasing her’; or ‘The reason why Sally run is that a bear is chasing her’. The target of my argument here is a different construction – a combination of A and B. C. Sally’s reason for running is that such-and-such. This means that a reason for Sally to run is that such-and-such and she runs for that reason. The reason simultaneously favours and accounts for her running. It is both normative and explanatory; in this respect it is often described as a motivating reason. And it is this joint role that gives rise to the philosophical tension that I described at the beginning. The consideration that favours her running away appears to be that a bear is chasing her, not that she believes that a bear is chasing her. But the consideration that appears to explain her running away – at least according to the line of thought I am rejecting in this essay – is that she believes a bear is chasing her rather than that a bear actually is chasing her. Let me begin with the question of whether the fact that Sally believes that a bear is chasing her is a (normative) reason for her to run away – whether the fact that she has that belief favours her running away. What can be accepted without much difficulty is that her having that belief makes her running away rationally intelligible. Learning that she thinks a bear is chasing her I can make sense of her running away; I can see that her behaviour is rational. But this does not in itself mean that the fact that she has that belief favours her running away. Learning that Sally has a pathological fear of furry animals makes her running away rationally intelligible; what she does makes sense in the light of her having this fear. But the fact that she has a fear of furry animals does not favour her running away; it is no reason for her to run away (assuming for the sake of argument that she is trying to fight this phobia rather than going with the flow). We could say that the fact that she has a fear of furry animals is a reason why she runs away rather than a reason for running away. This distinction is made by several people in the literature including Darwall (1983, p. 29) and Dancy (2000, pp. 5–6). Bernard Williams (1981) represents a classic source for this distinction. His example is someone putting tonic in a glass of petrol, thinking it is gin, and drinking it. Williams says that the man’s desire for gin is not a reason for him to drink that liquid, but if his belief that it was gin had been true, that desire would have been a reason.4 Yet for the man who inadvertently drinks petrol ‘we do not only have an explanation of his doing so (a reason why he did it), but we have such an explanation which is of the reason-for-action form’. (Williams 1981, p. 102). To summarise

Rowland Stout 53

Williams’s position here, there is an explanation of the man drinking that liquid, which is petrol. It cites a reason why the man drinks the liquid. And it is of a reason-for-action form. But it does not cite a reason for drinking the liquid. So it is of a reason-for-action form even though it does not appeal to anything that is actually a reason for action. It is not clear what Williams really means by the idea of an explanation of reason-for-action form that does not cite a reason for action. One possibility is that he is marking a distinction between reasons that are merely causal and reasons that put the action in a rational light, just as reasons for actions do. One reason why the man is drinking this liquid may be that he woke up late; this set off a causal chain that resulted in him being where he is with a glass of clear liquid in front of him. But this reason does not put his action in a rational light, whereas the facts that he believes the liquid is gin and that he wants to drink a gin and tonic form part of a causal explanation of his action that does put it in a rational light. Likewise, that Sally had a pathological fear of furry animals would put her running away in a rational light. What these reasons do not do is favour the action.5 What does favour the action is the fact that a bear is chasing her. There is a way of making recommendations for action – one that applies to Sally – that allows one to infer from the fact that a bear is chasing Sally that she should run away. In its simplest form we can think of that system just as the rule: if a bear is chasing you then run away. The rule can be justified by reference to a bit of bear psychology. It is a rule that Sally herself may acknowledge commitment to. Is Sally’s behaviour also responsive to a way of making recommendations that applies to her that allows one to infer from the fact that Sally believes that a bear is chasing her that she should run away – one that includes the rule: if you think a bear is chasing you then run away? If you were following such a rule you would have to establish first whether you thought a bear was chasing you and then act accordingly. But clearly Sally is not following nor should be following a rule which requires that of her. She should not attend to her psychological states in order to establish how to act; she should attend to the outside world – to whether a bear is actually chasing her. Of course it is possible to follow a rule which recommends action on the basis of whether you have a belief or not; but such a rule would have a very limited application. For example, there may be a rule: if you think that a bear is chasing you you should take your anti-psychotic medication. It does not matter that a bear is not chasing you; you should take your anti-psychotic medicine just in virtue of the fact that you think a bear is chasing you. But this is a peculiar sort of case.6 We can assume that in the case under consideration what does matter to Sally is whether a bear is actually chasing her. If a bear is not chasing her even though she believes one is she will make a mistake to run away. And if a bear is chasing her even though she does not believe it she will make a mistake not to run away. The rule that is operative

54

Sally’s Reason for Running from the Bear

for Sally in the case under consideration is one that makes reference to the fact that she is being chased by a bear, not to whether she believes it. Now one might respond that the fact that she believes a bear is chasing her itself makes it more likely that a bear is actually chasing her, assuming some degree of reliability in her belief-forming mechanisms, and so indirectly favours her running away. Suppose Sally is accompanied by Susan, who does not believe that a bear is chasing them but then sees that Sally does believe it. Susan may or may not go on to form the belief that a bear is chasing them. But in either case the fact that Sally believes a bear is chasing her favours running if the expected cost of running is outweighed by the expected cost of not running. Given that Sally believes a bear is chasing her, it may be a good bet to run away if Susan is not absolutely sure one way or the other. If Susan had such a reason for running then so presumably would Sally. But for Sally such a reason will always be redundant. If she has the belief for a reason then it is this reason for thinking that a bear is chasing her that is also her reason for running away; her having the belief does not add any reason to this. And if she does not have a reason for believing that a bear is chasing her then she has no reason to run away either. It might be possible to construct a bizarre scenario in which it is genuinely the fact that Sally believes that a bear is chasing her rather than the fact that a bear is chasing her that gives her a reason for running away. But even if such an example might be constructed it would not be possible to derive a general case for it always being the fact that she has the belief rather than the fact believed that is the reason for Sally to run. For, when Sally’s belief is not reliably formed the fact that Sally believes a bear is chasing her is not a reason for her to run. If Sally is walking through the English countryside, has just watched a film about someone being chased by a bear, and is a very suggestible person, then when she feels the hair rising on the back of her neck she may believe that a bear is chasing her. But this fact in these circumstances does not itself favour running away. The normative quality of reasons for action requires some separation of reasons from motivation. Kant expressed this by contrasting the way everything in nature works according to laws with the way that rational beings have the power to act according to the idea of laws – that is, according to principles (Groundwork, 36). Even if we accept the sort of internalism recommended by Williams (1981) and claim that reasons for Sally to act must be grounded in some way in her motivational state, we must still take them to exist independently of her motivational state. Her will must be able to respond to such reasons and be sensitive to them. And this means that she must be able to make mistakes about such reasons as well as being able to succeed in lining up her intentions according to them. If facts about what she believed constituted her normative reasons then they would not be properly independent of her motivational state.

Rowland Stout 55

Raz (1975) makes the same point when arguing that such facts cannot be used as guides for behaviour: It should be remembered that reasons are used to guide behaviour, and people are to be guided by what is the case, not be what they believe to be the case. To be sure, in order to be guided by what is the case a person must come to believe that it is the case. Nevertheless it is the fact and not his belief in it which should guide him and which is a reason. (1975, p. 17) Williams too is very clear about this. The fact that the man in Williams’s example believes that the stuff in his glass (petrol) is gin does not mean that he has a reason to add tonic to it and drink it, even if he wants a gin and tonic. If reasons for action – normative reasons – must be such that they may exist even when the agent is not aware of them, then the fact that Sally believes that a bear is chasing her is not such a reason. And most philosophers who make the distinction between normative and motivating reasons agree. Normative reasons are worldly considerations that favour a certain sort of action; they are not generally facts about an agent’s beliefs about such worldly considerations.7 However, many such philosophers go on to say that motivating reasons are different; they are taken to be facts about the agent’s beliefs and desires. If a motivating reason is just an explanatory reason, as for example Michael Smith takes it to be (Smith 1994), then this makes sense. But if a motivating reason is taken to be a reason for which one acts (as I have been taking it to be) then it is not clear how to make room for this possibility. For if the fact that Sally believes that a bear is chasing her is not a reason for her to run how can it still be her reason for running? The reasons given in formulations like C seem to be normative reasons that motivate. If the fact that Sally believes a bear is chasing her is not a normative reason for her to run given the reasoning system she is committed to in her motivational state, then it is not a motivating reason either. Derek Parfit considers the possibility that he falsely believes his hotel is on fire and so jumps into the canal. He baldly states that in this case ‘my motivating reason was provided by my belief; but I had no normative reason to jump. I merely thought I did’. (1997, p. 99) But he makes no attempt to explain why we should say that he is actually motivated by a reason in this case rather than just that he thinks he is. By allowing motivating reasons to be provided by beliefs, as he puts it, he is denying that motivating reasons are simply reasons for action for which one acts. He is extending the notion of motivating reasons, but it is not clear what guides this extension, assuming that the idea of a motivating reason is distinct from that of an explanatory reason. For R to be a motivating reason for S to w R must be S’s reason for w-ing. S must w in the light of R. R must be a reason or consideration that motivates

56 Sally’s Reason for Running from the Bear

S to w.8 For any of these conditions to be the case R must be a consideration that is taken by S to favour w-ing. For R to be S’s motivating reason for w-ing, R must be taken by S to be a normative reason for w-ing. It follows that motivating reasons must meet what Jonathan Dancy (2000, chapter 5) calls the normative constraint. This requires that a motivating reason, that in the light of which one acts, must be the sort of thing that is capable of being among the reasons in favour of so acting; it must in this sense be possible to act for a good reason. (2000, p. 103) If the fact that Sally believes a bear is chasing her were her reason for running then she would have to at least take it to be a reason for running, and this means that it would have to be the sort of thing that was capable of being among the reasons in favour of running. But, as we have seen, the fact that she believes a bear is chasing her is not, except in bizarre cases, capable of being among the reasons in favour of running. Sally could not take it to be such. So it cannot be her reason for running away. If motivating reasons must be potentially normative reasons, it cannot be the case that normative reasons are facts about the world and motivating reasons are only ever facts about the agent’s psychological state. So why is there such a powerful inclination to say that Sally’s reason for running is that she believes/thinks that a bear is chasing her? We have seen one source of this inclination right at the start in Michael Smith’s argument that motivating reasons must be psychologically real. What makes reasons psychologically real according to Smith is that they are beliefs. But in response to this argument one could propose instead that what makes a reason psychologically real is simply that it is believed rather than that it is itself a belief. The fact that a bear is chasing her can be psychologically real for Sally just by virtue of her believing it. Most people, presumably including Smith, accept that the fact that Sally is being chased by a bear can explain at the very least the fact that she believes a bear is chasing her. So why not also accept that it can explain her behaviour? There is another explanation for the strong inclination to think that it can only be the fact that Sally believes a bear is chasing her that is her reason for running. It is the plausibility of an argument we can call the Argument from False Belief. Jonathan Dancy articulates the argument clearly in the course of his rejection of it: The main reason for saying [that the things we believe cannot be the reasons that motivate us] is a worry about the case where things are not as the agent conceives them to be. Surely, in such a case, we cannot say that his reason for acting as he did was that p. We have to say that his reason for acting was that he believed that p. Accepting this for the case

Rowland Stout 57

where the relevant belief is false, then, we might still hope that ‘that p’ can indeed be the explanation where it is the case that p, but that where it is not the case that p the explanation can only be ‘that he believed that p’. But, as Bernard Williams puts it (1980: 102), the true–false distinction should not be allowed to affect the form of the relevant explanation. Supposing, therefore, that our explanation should take the same form whether it is or is not the case that p, and having already accepted that the correct explanation in cases where it is not the case that p, is ‘that he believed that p’, we are driven to say the same where the relevant belief is true rather than false. (2000, p. 121) This is my paraphrase applied to Sally’s case. 1. (Premise) When Sally’s belief that a bear is chasing her is false her reason for running is not that there is a bear chasing her. 2. (Interim conclusion) So, in the case where Sally’s belief that a bear is chasing her is false, Sally’s reason for running can only be that she believes a bear is chasing her. 3. (Premise – Williams’s Principle) The true–false distinction should not be allowed to affect the form of the relevant explanation. 4. (Conclusion) So, even when Sally’s belief that a bear is chasing her is true her reason for running away can only be that she believes a bear is chasing her and not that a bear is actually chasing her. Dancy rejects the conclusion and accepts everything else in the argument except the first premise. He makes the bold move that even when a bear is not chasing her Sally’s reason for running is that a bear is chasing her. This is bold because Dancy accepts a fairly standard way of linking reasons for action and explanations of action. This is that when someone acts for a reason then their reason for acting that way explains their acting that way. So Dancy’s denial of 1, commits him to accepting that even when no bear is chasing Sally it is still true that what explains her running away is that a bear is chasing her. So, as Dancy puts it, this sort of explanation is not factive. Dancy grants that this would not be acceptable if action explanation was causal. Causal explanations appeal to real things and genuine facts. But Dancy thinks action explanation is not causal, so it need not be factive. Still it must be considered a last resort to accept that that a bear is chasing her explains her running away even when no bear is chasing her. Can we find something else in the argument to attack? Let me start with the premise in line 3 which I call Williams’s principle. Williams produces this principle early on in his article ‘Internal and External Reasons’ in a section where he is making some fairly fine distinctions in order to clarify his terms before launching his main argument for the claim

58 Sally’s Reason for Running from the Bear

that reasons for acting must be internal in his sense. Dancy takes Williams’s principle to show that reasons for acting must be the same in the case where the belief is true and in the case where the belief is false. But this is certainly not what Williams takes his principle to show. Williams claims that an agent’s reasons for action are different depending on whether the belief is true or false, although the form of the explanation is the same. Certainly Williams’s principle should not be treated in the way Williams himself treats it if it is to serve Dancy’s purpose. Dancy is not concerned with the form of an explanation but with which reasons actually figure in an explanation. What he needs is the principle that the true–false distinction should not be allowed to affect what actually count as reasons for action. According to this principle, the reasons for running that Sally has when her belief that a bear is chasing her is false are still her reasons for running when the belief is true, and her reasons for running when a bear really is chasing her are still her reasons when the belief is false. So we should substitute 3* for 3, where 3* is as follows: 3*. (Premise) The truth or falsity of Sally’s beliefs does not affect her reasons for action. So Sally’s reasons for running when her belief that a bear is chasing her is false are also reasons for running when it is true. And her reasons for running when her belief is true are also reasons for running when it is false. Accepting this gives us just two options concerning Sally’s reasons for running. Either her reason for running whether the belief is true or false is that a bear is chasing her. Or her reason for running in neither case is that a bear is chasing her, which leaves the only plausible conclusion being that in both cases her reason for running is that she believes a bear is chasing her. So 3* gives Dancy his conclusion, although 3* should certainly not be attributed to Williams, since he clearly denies it.9 If, like Williams, we say that Sally has no reason for running if there is no bear then we must deny 3* (assuming that we ever have reasons for action). But note that 3* has two parts. It is the second part that would be rejected by Williams who thinks that the true believer may have a reason for acting that the false believer does not, but merely thinks they have. At the same time Williams would reject 2. Rejecting the first part of 3* too is a move that might be made by someone who was impressed with John McDowell’s (1982) disjunctive approach to appearances that is supposed to defeat the Argument from Illusion in epistemology and who thought it could be carried over to the philosophy of action to defeat the Argument from False Belief. The way the disjunctive approach might be applied here is as follows. When it is said that the reason why Sally is running is that she believes a bear is chasing her there are two distinct (but to Sally indistinguishable) things that might be going on and there is no highest common factor between them. Either Sally’s reason for

Rowland Stout 59

running actually is that a bear is chasing her or Sally’s reason is that she merely believes that a bear is chasing her. On this view the true believer is both better off and worse off than the false believer with respect to her reasons for action. She has the extra reason that a bear is chasing her, but she lacks the reason that she believes that a bear is chasing her. But this seems wrong too. It is easy to see how a true believer may be better off than a false believer; there is an extra reason in the world for her to be sensitive to. But how can the fact that the world turns out to be the way you think remove one of your reasons for action? A sort of inclusive disjunctive account might reject only the second part of 3* while holding on to 2. On this view the true believer has two reasons for running away – both that a bear is chasing her and that she believes a bear is chasing her, while the false believer’s reason for running is just that she believes that a bear is chasing her. This would not be vulnerable to the problem just raised since the true believer would not lack any reasons that the false believer had. But one might worry instead that there are too many reasons in this case. The true-believing Sally’s behaviour does not appear to be doubly justified. The fact that the bear actually is chasing her seems to be a quite redundant addition if the fact that she merely believes that a bear is chasing her is sufficient to count as a reason for her to run. In any case the arguments presented earlier against thinking of motivating reasons as facts about what agents believe apply whether the beliefs are true or false. This does not mean that Dancy’s rejection of 1 is after all the most reasonable way to respond to the Argument from False Belief given one has an argument against its conclusion. For we still have to look at the inference from 1 to 2. To invalidate this inference we need to show that when Sally’s belief is false it is possible for Sally’s reason for running to be neither that a bear is chasing her nor that she believes that a bear is chasing her. One way this might work is if, like Williams (1981) in the first two or three pages of ‘Internal and External Reasons’, we say that Sally has no reason for running when her belief that a bear is chasing her is false. The true-believing Sally’s reason for running is that a bear is chasing her. The false-believing Sally has no reason for running. This is akin to the disjunctive account in as much as it denies that there is a Highest Common Factor between the false-believing Sally and the true-believing Sally’s reasons for running. It does not have the problematic implication of the exclusive version of the disjunctive account that the true believer is worse off in some respect than the false believer with regard to their reasons for action. And it does not fall foul of the arguments that I have developed in this essay against thinking that facts about what Sally believes should count as reasons for her to act. Now if Sally has no reason for running it does not follow that she is irrational. There still may be reasons why she runs, reasons which put her behaviour in a rational light even if they do not justify or favour her behaving that way. She does not have a reason for running that she thought

60

Sally’s Reason for Running from the Bear

she had, but her behaviour can still be explained in a way which reveals her rationality. Sally, after she finds out that there was no bear, may feel foolish; she thought she had a reason for running away but it turns out she didn’t. It is also worth pointing out that even though the false-believing Sally’s reason for running was neither that a bear was chasing her nor that she believed a bear was chasing her, there may be other reasons we can find for her running. Her reason for running might be that there is evidence that there is a bear chasing her; for example, that there is a noise in the undergrowth might be her reason for running. This may be a slightly different explanation of her running than the one that cites the fact that there is a bear chasing her. Sally might say: I am not concerned with whether or not there really is a bear chasing me; the existence of this evidence is a good enough reason for me to run. Also, given that we are allowing bad systems of justification to provide reasons for action (reasons relative to those systems) we can say that there is a system of justification embedding the false belief that a bear is chasing Sally from which we can derive the recommendation that Sally should run given the fact that one should run when bears are chasing. So her reason for running might be that one should run away from chasing bears. This, though true (we are assuming), is a bad reason in the circumstances; but can still count as normative (in the relative sense described earlier) and motivating.

Notes 1. Had it been a grizzly bear running away would have been ill-advised, but with a black bear it might make sense. 2. For a small sample of moral philosophers who would take her reason for running to be that a bear is chasing her and not that she believes a bear is chasing her see Raz (1975, chapter 1), Williams (1981), Darwall (1983), Skorupski (1997). And although philosophers of action often take reasons for action to be facts about beliefs and desires, some – especially those represented in this volume – do not; apart from myself (Stout, 1996, 2006), see in particular Jonathan Dancy (2000), Rüdiger Bittner (2001) and Fred Stoutland (1998). 3. The real significance of this issue as far as I am concerned derives from the fact that agency essentially involves sensitivity to reasons for action. If we think of such reasons as psychological then we are led to the wrong conception (as I see it) of the relationship between action and the mind. See Stout (2006). 4. Like Dancy (2000, chapter 2) I do not think that the agent’s desires need be included as reasons for action even in the case where his beliefs are all true, but Williams’ claim is perfectly innocuous if all that he means is that the fact that that he felt like a glass of gin and tonic would have been a reason for him to drink that stuff if the glass had had gin in it. 5. It is precisely because it is not clear whether Donald Davidson (1980, essay 1) is talking about reasons for action or reasons why someone acts which are of a reasons-for-action form, when he claims in ‘Actions, Reasons and Causes’ that primary reasons are constituted from a belief and a pro-attitude, that I do not want to make his position the target of my attack in this essay although on the

Rowland Stout 61

6. 7. 8.

9.

face of it it does look like the sort of psychological approach to reasons that I am attacking. This example reflects Jonathan Dancy’s argument in (2000, p. 124). See Parfit (1997) for example. Notice there is a difference between some psychological state motivating you and the fact that you are in that state being a motivating reason. You may be motivated by greed or pathological fear; yet the fact that you have greed or pathological fear is not a reason that motivates you. It is wrong to say that you act in the light of having greed or fear or that your reason for what you do is that have that greed or fear. Likewise we could accept that Sally’s belief or state of believing motivates her while denying that the fact that she has the belief is a motivating reason. Bittner (2001) follows Dancy in misattributing this principle to Williams.

References Bittner, R. 2001, Doing Things for Reasons, Oxford: Oxford University Press. Crisp, R. 2006, Reason and the Good, Oxford: Oxford University Press. Dancy, J. 2000, Practical Reality, Oxford: Oxford University Press. Darwall, S. 1983, Impartial Reason, Ithaca: Cornell University Press. Davidson, D. 1980, Essays on Actions and Events, Oxford: Oxford University Press. Kant, I. 1948 edition, Groundwork of the Metaphysics of Morals, tr. Paton, H. as The Moral Law, London: Hutchinson. McDowell, J. 1982, ‘Criteria, Defeasibility and Knowledge’, Proceedings of the British Academy 68, 455–79. Parfit, 1997, ‘Reasons and Motivation’, Proceedings of the Aristotelian Society Supplementary Volume 71, 99–130. Raz, J. 1975, Practical Reason and Norms, London: Hutchinson. Searle, J. 2001, Rationality in Action, Cambridge MA: MIT Press. Skorupski, J. 1997, ‘Reasons and Reason’, in Gullity, G. and Gaut, B. (eds), Ethics and Practical Reasons, Oxford: Oxford University Press, 345–67. Smith, M. 1994, The Moral Problem, Oxford: Blackwell. Stout, R. 1996, Things That Happen Because They Should, Oxford: Oxford University Press. Stout, R. 2006, The Inner Life of a Rational Agent, Edinburgh: Edinburgh University Press. Stoutland, F. 1998, ‘The Real Reasons’, in Bransen, J. and Cuypers, S. (eds), Human Action, Deliberation and Causation, Dordrecht: Kluwer. Williams, B. 1981, Moral Luck, Cambridge: Cambridge University Press.

4 Con-reasons as Causes David-Hillel Ruben

An agent’s having of a reason for an action (hereafter, simply ‘a reason’) is often said to be amongst the causes or causal conditions of the action for which it is a reason (in this wide sense, ‘action’ includes many cases of inaction)?1 Hereafter, this view is referred to as (1). The reasons that (1) is about are sometimes called ‘explanatory’ reasons in contradistinction to ‘justifying’ reasons, or ‘motivating’ reasons in contradistinction to ‘normative’ reasons, or ‘internal’ reasons, in contradistinction to ‘external’ reasons.2 These are three different distinctions, related to one another in complicated ways. The reasons required by (1) are surely motivating: these are the reasons that have actual psychological ‘purchase’ on the agent, and not just merely in principle available to the agent in some wholly objective sense. A well-known view, which is also adopted by the position I am considering, asserts that the only reasons that could play the causal role required by the idea of motivation, and hence explain why the agent did what he did, are the agent’s internal psychological states. So the reasons required by (1), as I construe it, are internal, explanatory and motivating. According to (1), the idea of an agent’s being motivated by a reason, or a reason having ‘purchase’ on an agent, is a causal idea, and the explanation of action by citing the agent’s reason is causal explanation. But even if (1) divested itself of the idea that the explanatory causes of action were the agent’s internal psychological states, it is at least committed to the view that whatever it is that explains and motivates action must be some item which is capable of being caused and causing other things.3 (1) is consistent with almost any view about the mind–body relation. As far as (1) goes, that psychological event or state might (or might not) be identical to a physical one. An agent can, of course, have a reason for an action without its being a cause of that action, but in the case in which the agent performs the action because of that reason, the reason is said, on this view, to cause the action. Donald Davidson’s well-known argument attempts to demonstrate just this: ‘If ... causal explanations are ‘wholly irrelevant to the understanding we 62

David-Hillel Ruben

63

seek’ of human actions then we are without an analysis of the ‘because’ in ‘He did it because ...’, where we go on to name a reason’.4 The (motivating, explanatory, internal) reasons Davidson focuses on, in this argument and elsewhere, are reasons which function as ‘pro-reasons’, in two closely connected senses: (a) these are reasons relevant to the action the agent does in fact take; (b) they are reasons which favour that action. Davidson’s reasons consist of a belief and a desire, the desire being a pro-attitude towards the kind of action the agent does. Other views might identify intentions or some other mental items as these reasons,5 but precisely which causal items count as reasons will not concern me here. The literature in action theory has tended to overlook the fact that reasons function in another way too.6 (a*) One can have reasons for an action that one does not take; (b*) one can have reasons which disfavour an action taken (a ‘con attitude’, to parallel Davidson’s ‘pro attitude’). Sometimes a person has conflicting reasons for acting, one set of which is a set on which he does not act, and both sets of conflicting reasons can be rationally or deliberatively relevant to the same choice situation and rationally or deliberatively relevant to the same action finally chosen. Each set justifies or supports a different proposed action on the agent’s part, and the agent is not able to perform both actions, because it is impossible to act in both ways at the same time. So the agent chooses to act in one of those ways rather than in the other(s). The reasons might strongly conflict, in the way in which a reason to do some token act of type A (or, as I sometimes elliptically say, a reason to A) and a reason not to do any token act of type A conflict; or the reasons might weakly conflict, in the way in which a reason to do some token act of type A and a reason to do some act of type B (or, as I sometimes elliptically say, a reason to B) conflict on any occasion on which one cannot as a matter of fact do both. On an occasion on which one cannot do both, a reason to do an act of type B must also be a reason not to do an act of type A, but only modulo the additional information that one cannot do both acts in the circumstances. In the case of strong conflict, no additional information is similarly required. Suppose that X has a reason to do a token action of type A and a reason to do a token action of type B, where the two reasons weakly conflict. Suppose further that X chooses to perform a token action of type A. The first reason, which favoured doing A, was rationally or deliberatively weightier (in the circumstances, of course) than the second, which favoured doing B; the first counted for more, as far as X was concerned. As we say, all things considered, X chose to perform a token action of type A. Both the pro- and the con-reason are, as I shall say, ‘rationally or deliberatively relevant’ to, or bear on, the same eventual choice made or action taken. (I write as if choice precedes every action, but my argument would be unaltered if choice were not ubiquitous in this way.) The single final choice made or action taken is made

64

Con-reasons as Causes

or taken because of the pro-reason and in spite of the con-reason. Indeed, that is what ‘all things considered’ must mean: because of the one set of reasons and in spite of the other. As Dancy says about these con-reasons: ‘But still I was influenced by them [the con-reasons] and they do figure in my motivational economy’.7 A con-reason is also a pro-reason in its own right for the action not taken, and is a con-reason only in the sense that it counts against the action which was taken. Similarly, a pro-reason is only a pro-reason for the action taken and is itself also a con-reason for the action not taken. In what follows, to simplify terminology, I will only use the idea of a pro-reason to be the reason which counts for the action one takes, and the con-reason to be the reason which counts for the action one does not take, the reason which gets outweighed.8 In the light of this, it would be wrong to think of pro-reasons and con-reasons as two different sorts of reasons. I was careful above only to say that reasons can function in these two different ways, depending on context. We use the language of reasons, weights and strength, in describing our deliberations. Such language is metaphorical, but, metaphorical or not, it certainly seems irreplaceable. We can order reasons for action by their strengths: one reason can be stronger than another, weaker than a third. But the whole truth about reasons for action can’t be exhausted just by this type of relational, ordinal information. We can speak of a case in which we have only one reason to do something, and say of it that it is strong or weak. Even in a case of conflicting reasons with both a pro-reason and a con-reason, we can imagine another similar situation in which the pro-reason remains at the same strength but the con-reason gains in strength so that the latter will now outweigh the pro-reason. In such a case, the pro-reason will not have changed its strength, only the con-reason would have, and hence the ordinal facts will have changed. This suggests that the relative strength of reasons for action is grounded in some sort of intrinsic weight they have. However, for purposes of deliberation, it is the relational, ordinal information that is crucial, even though there seems to be an underlying reality about the strength of reasons on which that relational information is based. The notion that reasons are causes, as I have explained that view above, ties rationality and causality in an obvious way. On that view, if an agent has a pro-reason for doing A and a con-reason for doing B, and does an action token of type A because of that first reason, the strength of that rationally or deliberatively stronger reason on which he acts is reflected by the fact that it is that reason that is the reason that causes an action, his A-ing, and the relative rational weakness of the reason on which he does not act is reflected by that other reason’s failure to cause an action, in its failing to lead to his B-ing. For a proponent of (1), that one reason rationally outweighs another is reflected in the causal sufficiency of the first but not of the second for action. Rational or deliberative strength and causal strength are tied together.

David-Hillel Ruben

65

Weakness of the will constitutes an apparent exception to this claim, and indeed it is just this which makes weakness of the will an important topic for someone who upholds (1). The essay will not focus on questions about weakness of the will, so let us say that, with this admitted exception, (1) otherwise combines the idea of reasons as causes with the view, (1*) that the rational or deliberative strength or weakness of a reason is captured by the causal sufficiency of the reason for action, or the lack thereof. Rationally stronger reasons lead to the actions they support, and the reasons they are stronger than do not lead to the actions they support. Davidson himself says that ‘if reasons are causes, it is natural to suppose that the strongest reasons are the strongest causes’.9 I refer to the conjunction of (1) and (1*) as causalism (in the philosophy of action). I am not just asserting that (1*) makes the vacuous claim that stronger reasons, rather than the reasons they are stronger than, cause actions unless they don’t. I am saying that (1*) makes the substantive claim that stronger reasons cause actions and the ones weaker than them do not, except in cases of weakness of will. Weakness of the will is a specific and delineable phenomenon. I intend to show that (1) and (1*) are false, and my argument relies on certain general features of causation. It does not rely on, and I do not use any examples of, weakness of the will to show this. If the reader is happy to accept that weakness of the will would constitute a further argument against (1) and (1*), I am happy to accept that as well. My argument, if successful, shows something about (1) and (1*), in addition to whatever weakness of the will shows. What exactly is the connection between (1) and (1*)? Without claiming that (1) entails (1*) even for cases other than weakness of the will, I think that anyone who holds (1), when asked about the phenomenon of con-reasons, is bound to accept (1*) as well (in Davidson’s words, they ‘naturally’ go together). Surely, on this view, the pro-reason causes the action it supports, and the con-reason does not cause the action it supports, because the proreason is a rationally stronger or weightier reason than the con-reason (again, neglecting the case of weakness of the will, but I shall cease repeating this qualification hereafter). If degree of rational support and causation are not tied in the way in which (1*) proposes, then we are owed an alternative account by the proponent of (1) of why it is that a rationally weaker conreason does not typically or even always, and not just in limited cases of akrasia, cause the alternative action that it supports, when an agent has both a rationally stronger pro-reason and a weaker con-reason. The central question I wish to address in this essay is this: if reasons are causes of actions, and if the rational strength of a stronger reason is captured by its causal sufficiency for action, what can we make of the idea of the causality of rationally weaker con-reasons? The point that I shall want to make will be the reverse, in a sense, of Davidson’s famous argument, mentioned above, in which he claims that unless having reasons for acting is understood

66

Con-reasons as Causes

causally, there is no way to cash out the idea of the reasons because of which one acts. The argument advanced in this essay is twofold: (2) first, that there are at least some cases in which there is no way to understand as causes, modulo some other plausible assumptions to be adumbrated, the rationally weaker con-reasons that disfavour the actions one does take but that do favour the actions that one does not take; (3) second, if such con-reasons cannot be understood as causes, then neither can pro-reasons. So, to paraphrase Davidson and simplify somewhat, the argument here is that there is no way in which to cash out the idea of the causality of the rationally weaker reasons on or for which one does not act, and thus, no way to cash out causally the idea of acting for any reason at all. I start by arguing the first point, (2), which will occupy the bulk of the essay. (1) stated that an agent’s having of a pro-reason for an action is a cause or amongst the causal conditions of the action for which it is a reason. On (1), like a pro-reason, a con-reason (the agent’s having of that con-reason) is meant to be a state or event. This must be so, since pro- and con-reasons are not two distinct sorts of items. If a con-reason had been rationally weightier than it was, it would have caused the action for which it is a reason. But it can’t be that some item is not an event or state but would have been an event or state in other circumstances (e.g. if it had been weightier). So if a pro-reason is an event or state, so is a con-reason. Metaphysically, pro- and con-reasons must be cut from the same cloth. States and events have causes and cause things. Consider (4): if an agent’s having of a con-reason is an event or state, then it is a cause, either of that action for which it is a con-reason or a cause of something else. If a con-reason is part of the causal order of things at all, the con-reason must surely cause something (so I do not consider an acceptable response by the causalist to be that con-reasons have causes but in turn cause nothing). The con-reason will have some actual effects or other. It is true that the view encapsulated by the conjunction of (1) & (1*) by itself places no constraints on what it is that conreasons must cause. It is only committed to finding some effect(s) or other of the rationally weaker of the reasons that does not lead to action, effects which in some way hopefully mirror that relative rational weakness. Of course, nothing in what follows is meant to be inconsistent with the perfectly plausible view that if an agent has a con-reason, certain counterfactuals must be true in virtue of that fact. For instance, if the agent has a con-reason to B when he in fact A’s, then the having of that con-reason lowers the agent’s probability of doing A, even though he still does A, below what it would have been in the absence of the con-reason.10 But none of this need be given a causal reading and the argument that follows will try and convince the reader that it should not be given such a reading. On the causal story, there will certainly be a causal chain that leads from the pro-reason to the action the agent took. Clearly, the con-reason cannot cause the action that it favours, because that action never happened, and if

David-Hillel Ruben

67

it did not happen, nothing can cause it. The chain going from the con-reason to the action it favours is merely a counterfactual chain. However, there might be cases in which the con-reason also causally contributes to the action taken, the action which it disfavours (to repeat: NOT on the action it favours, which occurs only on the counterfactual chain but not on the actual one). Did his con-reason so modify his action that it was different from what it would have been had he only had a pro-reason to do it? Well, it might have. Perhaps if he had not had his con-reasons, he would have chosen or acted a bit differently – perhaps he would have chosen or acted more quickly, more assertively, more decisively, or less hesitantly. So the con-reason, we might suppose, affects the action the agent does take, by making it indecisive or hesitant or whatever. Buridan’s Ass is in fact a special case of this.11 On the Buridan supposition, had the ass only had reason to choose hay pile A (and hence had no reason to choose hay pile B), he would have chosen pile A. Had he only had reason to choose hay pile B (and hence had no reason to choose hay pile A), he would have chosen B. But when he has both reasons, the causal chains (if such there are) converge and cancel each other out, and he chooses neither, rather like the body which remains stationary under the effect of balancing forces. So in the special, ‘Buridan’ case, the reason to choose pile B would not merely modify his choice of pile A; it supposedly would eliminate it entirely, in favour of inaction. (The same can be said for his reason to choose hay pile A and his choice of pile B.) Hence, so the argument runs, there cannot be converging and mutually cancelling causal chains in such a case, since we know that even asses do not starve – they select one hay pile or another, long before starvation sets in. The problem with this as a general solution to our problem about the causal power of con-reasons is that although something like this might be true, and indeed no doubt is true in some cases, it need not be. It seems that there can certainly be intrinsically qualitatively identical choices or actions, differing only extrinsically in whether or not the chooser also had a reason to do something else. So, let it just be stipulated that we are considering a case in which the agent does something in the circumstances in which he does have conflicting reasons, but that he would have also done it in an intrinsically qualitatively identical manner had he only had the one set of reasons. His ‘opposing’ reason does not make him hesitate, or dither, in doing whatever it is that he does, in any way.12 There must, therefore, be cases in which the con-reason is not causally necessary for the actual action taken, if it is a cause at all; indeed, I speculate that this would be so in the vast majority of cases. If we assume that the con-reason does not also contribute to the causation of the action it disfavours, but rather would have to cause something else, there are any number of possible candidates as the effects of con-reasons available to the causalist. Perhaps a person’s con-reason directly causes

68

Con-reasons as Causes

regret,13 or causes some other change in his mental landscape (his dispositions to act, for example) or causes some psychological illness in him. He does the action favoured by the pro-reason but since he had reasons against it, his con-reason ends in him regretting what he did, or some such. Or perhaps the effect of the con-reason is not even at the personal level at all. Might its effect not be some physiological or brain event, one of which the actor is perhaps ignorant or unaware?14 (Or, ‘some further physiological or brain event’, if the having of a con-reason is such a physical event too.) The important feature of all these candidates is that they require a second causal chain, in addition to the one that goes from the stronger reason to the action taken. If so, there would be one causal chain leading from his having a pro-reason to his subsequent action. There would be another quite distinct causal chain leading from his con-reason to his subsequent regret, or illness, or to some (further) physiological or similar event. The causal chains would not converge causally on the final choice or action, as they would if both pro- and con-reasons causally contributed to the same action taken, as we sketched above. On this rather simple picture, the pro-reason initiates a causal chain leading to the action; the con-reason initiates a wholly independent, second causal chain, leading to the regret or brain state or whatever. One thing to note about this view is that it might not permit us to capture causally the idea that both pro- and con-reason are rationally or deliberatively relevant to the same token final choice or action. The con-reason might not be a reason against acting in a certain way in virtue of whatever causal role it plays. A con-reason would not be the con-reason it is (a reason not to do what was done) in virtue of its causing something else other than that action. At the level of reasons for choice and action, the two reasons bear differently on (one favours and the other disfavours) the same choice or action, but the causal story might not mirror this in any way. There are just two distinct causal chains, each of which leads to a different result; one leads to an action, the other to some psychological or neurophysiological or dispositional state. But perhaps a causal model of how pro- and con-reasons work in choice situations need not capture within the causal model this fact about the rational significance of both types of reasons to the same action or choice, so I don’t take this as an objection to the suggestion under discussion. The causal picture it suggests has so far been kept somewhat simplistic. This causal story seems to be exhausted by these two facts: the pro-reason causes the action via one causal chain, the con-reason causes something else via an independent causal chain leading to something else, whatever that might be. First, consider the actual situation, c. In c, the pro-reason to A is rationally weightier for the agent than the con-reason to B. Causally, therefore, if (1*) is true, it is the pro-reason that causes the agent to A, rather than the con-reason causing the agent to B (so the con-reason causes something else).

David-Hillel Ruben

69

But now consider a counterfactual situation, c*. c* is just like c, save in one feature, and whatever is a causal consequence of that one feature: in c*, although the pro-reason retains the same weight that it has in c, the con-reason becomes much weightier. I think that this sort of scenario is very common. At a later time, an agent can assess a reason as having more ‘gravitas’ than he earlier imagined it had. It might weigh more with him than it did before. So in c*, the con-reason counts more for the agent. The reason to B now rationally outweighs the reason to A in the agent’s deliberations, so the agent now B’s rather than A’s. At the level of decision, choice and reason, this is all straightforward. Notice how the deliberative story includes truths which are relational and comparative, as we described at the beginning of the essay: one reason is weightier than another, or is weightier than it was, counts more than another or more than it did. Even if underlying deliberation there are non-relational truths about the strength of reasons, it is relational, ordinal information about reasons that is crucial for understanding the deliberative story. If we have both proand con-reasons, we want to know which reason wins the deliberative contest. But causation is not relational; it does not come in degrees, nor is it comparative. A cause either causes an effect or it does not; a cause can’t cause something more or less, or more or less than something else. Something can’t be more of a cause than another thing. So how should we represent the subsequent causal facts of the matter in c* (in order to obtain a coherent causalist story)? In c*, since the reason to B has now become the strongest reason, according to (1*), the reason to B will cause the agent to B. In c*, there will be a causal chain leading from the reason to B all the way to the agent’s B-ing. But what does the reason to A now cause in c*? Remember that we are supposing that the only difference between c and c* is the increased intrinsic weight of the reason to B, and, as a consequence, the fact that the reason to B now outweighs the reason to A. Since the reason to A is just as strong in c* as it was in c (albeit now outweighed), and since the reason to A was sufficient to cause A in c, then the reason to A should cause the agent to A in c* as well (with one exception, described below). If it was strong enough in c to cause the agent to A and it has the same intrinsic strength in c* (albeit now outweighed), it should still cause in c* whatever it caused in c, given that there are no relevant differences between c and c* other than the increased weight of the reason to B and the relative weightings of the two reasons. In particular, the reason to A has the same intrinsic weight in c* as it had in c. If the reason to A has the same weight in c and c*, then its effects should be the same in both circumstances. So why doesn’t the agent do A in c*, just as he did in c? If the reason to A is strong enough in c to cause the agent to A, and if it has the same rational intrinsic strength in c* that it had in c, then it should still cause the agent to

70 Con-reasons as Causes

A in c*. True, the reason to B gains in intrinsic deliberative strength in c* (and so the ordinal facts about the relative strength of the reasons will change from c to c*) and hence what the reason to B causes will change from c to c*. So the reason to B should also cause the agent to B in c*. There should be, in c*, as far as we can tell, a stand-off: the agent should be caused both to do A and to do B. To be sure, the agent can’t do both A and B; by assumption, they are weakly incompatible. But in the counterfactual situation, causally speaking, there should be no grounds for thinking that the con-reason will now win out over the pro-reason. The con-reason is now strong enough to cause the agent to B, but the pro-reason remains at the same intrinsic strength and hence, if (1*) is true, is still strong enough to cause the agent to A. So why should we expect the agent to do one or the other? Why doesn’t the agent do A rather than B, even in the counterfactual situation, since his reason to do A remained in principle strong enough to cause him to do A, or why doesn’t he do nothing at all, as in a Buridan case, since the two causes might cancel themselves out? What we are finding is that there is a misfit between the non-relationality of causation and the relational, ordinal character of rational strength in deliberation. I mentioned one exception, above, to the claim that ‘since the reason to A was sufficient to cause A in c, then the reason to A should cause the agent to A in c* as well’. We need to take note of this qualification. Suppose that in c* the reason to B, in addition to causing the agent to B, is able to interrupt the causal chain that would otherwise lead from the reason to A to the agent’s A-ing, and that explains why the agent does not, after all, do A in c*. There would be some flexibility in deciding just where, in c*, the requisite inhibitor blocked or stopped the chain commencing with the reason to A from leading to its ‘natural’ conclusion, A, as long as the chain did not get all the way to that action. For the sake of argumentative simplicity, let us suppose that the reason to B inhibited the very next link on the chain. On such a chain, let m be the node that would have followed immediately after the reason to A. So let us say that, in the counterfactual situation, what happened is that the reason to B inhibited or prevented m from occurring, prevented or inhibited the reason to A from causing m, and hence prevented the action A. That is why the agent B’s instead of A’s in the counterfactual situation, and why his reason to A does not lead to his A-ing in c*, an explanation entirely consistent with (1) and (1*).15 The problem with this solution is simply that it is not true to the phenomenological facts of the case. Sometimes indeed an agent’s reason gets blocked from leading to the action to which it would otherwise have led. Even apart from cases of weakness of the will, there are an indefinitely large number of ways in which an agent’s deliberative decisions can be thwarted. Bad luck affects us all. A typical sign of this happening is agent frustration. If my reason to A does not lead to action only because the causal chain leading

David-Hillel Ruben

71

from it to action is blocked in some way, the agent will feel thwarted in doing what he really wants. Recall that on this view, the reason to A is meant to be as causally powerful in c* as it was in c, and so that reason would drive the agent to do A, in equal measure, both in c and c*. If the agent failed to A in c*, only because the causal efficacy of his reason to A had been blocked, even though by another reason, the agent would feel this as some sort of failure. But what we are trying to do is to give a causal model for the case in which the agent B’s, because his reason to B has become weightier even though his reason to A has retained its original weight. In this case, nothing needs to be thwarted and the agent need feel no frustration. The fact that his reason to B now outweighs in his own mind his reason to A is not a potential source of frustration to the agent. He is doing what he most wants, which is to B. He gladly surrenders his reason to A, at least in the circumstances, to his now-superior-because-weightier reason to B. It is not true that his reason to B prevents or blocks him from acting on his reason to A. In the case at hand, he chooses not to do A, because he takes his reason to do A as relatively of less importance or weight than his reason to do B, and in the case as we have constructed it, I do not see how this fact can be modelled causally. There is, I submit, no fully convincing way causally to model decisionmaking which includes con-reasons, for at least some cases. It is the element of weight, comparative strength, which cannot be captured causally, at least in those cases in which the con-reason does not contribute causally to the action taken. As long as one thinks only about pro-reasons for action causing the actions they favour, the point is not salient. But once con-reasons are introduced, it becomes clearer that there is no plausible causal modelling for all the ways in which con-reasons work in our deliberation scheme. Let me now move, briefly, to the second part of my overall argument. Suppose that I have showed some credibility in the view that there is no causal model for con-reasons, (2). (3) asserted that if con-reasons cannot be causes, then neither can pro-reasons be. Why not hold a divided view: Davidson’s argument only purported to show that the pro-reasons were causes; he did not address the issue of causality and con-reasons at all. Why can’t pro-reasons be causal items, but con-reasons play no part in the causal order at all? The answer to this is to remind ourselves that pro-reasons and con-reasons are not two different sorts of items. There is no such thing as a set of reasons which are always pro-reasons and a set of reasons which are always con-reasons. In different situations, in different contexts, a motivating reason will function now in one way, as a pro-reason, now in the other, as a con-reason. Thus, there are a series of true counterfactuals of the following type: Concerning some pro-reason r for an action A that an agent did, if r had been

72 Con-reasons as Causes

a less weighty reason, the agent might not have A-ed but have done something else, B-ed. In such a hypothetical situation, r might have been (only) a con-reason for his B-ing, that is, a con-reason for what the agent then would have done. Concerning some con-reason r* for an action A that an agent did, if r* would have been weightier, the agent might not have A-ed but done something else; he might have B-ed instead. In such a hypothetical situation, r* might have been a pro-reason for his B-ing, that is, a pro-reason for what the agent then would have done. It is this which makes, I think, divided views on the causality of reasons unacceptable. How could it be that a reason r plays a (or, no) causal role but had it been a weightier or less weighty reason, it would have lacked (or gained) a causal role, a place in the causal order? Had the con-reason had more or less motivational ‘oomph’, it would have gained or lost its status as a causal item. I know of no view about causation that would make that sort of divided story about reasons at all attractive. Changes in context can’t draw an item of some sort (think of numbers or propositions, for example) from being an a-causal item to being a causal one. I think this gives us reason for thinking that con-reasons, and hence reasons of both kinds, fail to be causes.16 If the having of reasons are not causes, what sort of item are they? I don’t intend to say much about this here. Some may want to persevere with the psychologistic view that reasons (or, the having of them) are psychological states of an agent, and the acquiring of the reasons or ceasing to have the reasons are psychological events which happen to the agent (even if they had no effects). Others will see the implication of my argument as demonstrating that the having of a reason cannot be any kind of event or state at all, and moving us to a non-psychologistic view of reasons and the having of them.17

Notes 1. (1) is understood here to speak of causation, not necessarily only of deterministic causation. The causation in question might be probabilistic or stochastic. What are causes? No paper can do everything and, with that, I intend to beg off any responsibility for explicating the idea of causation. I am presupposing a fairly standard account of causation, on which causes are token events or token states, and that a causal chain is a series of such. Finally, there is an important distinction between the full cause of something and merely a part of the full cause. The causal items I discuss here are surely causes only in the latter sense, causes in the presence of a large number of background conditions. Again, this distinction does not, I think, make any difference to what I say here. 2. On this distinction, see Williams, Bernard, (1981) ‘Internal and External Reasons’, reprinted in Moral Luck, Cambridge University Press, Cambridge. 3. Are states causes? I invoke Davidson’s rejoinder to this question in his ‘Actions, Reasons, and Causes’, reprinted widely and for example in White, Alan R., (1968) The Philosophy of Action, Oxford University Press, Oxford, pp. 87–8. States, dispositions, and conditions ‘are frequently named as the causes of events’. Still, there

David-Hillel Ruben

4. 5. 6.

7. 8.

9. 10. 11.

12.

13. 14.

15.

73

must be in addition to them a preceding event and ‘in many cases it is not difficult to find events very closely associated with’ the state or disposition. Onslaughts of states and dispositions, as well as items such as noticing, becoming aware, perceiving, learning, and remembering, provide obvious examples of associated events.’ Op.cit., pp. 86–7. See von Wright, G.H. ‘On So-called Practical Inference’, reprinted in Raz, Joseph, ed. (1978) Practical Reasoning, Oxford University Press, Oxford, pp. 46–62. Although Jonathan Dancy notes their existence. Dancy, Jonathan (2004) Practical Reality, Oxford University Press, Oxford, p. 4: ‘but still I will normally speak as if all the reasons that do motivate all pull in the same direction’. Ibid. In what follows, reasons have been so individuated that X has two sets, a pro-set and a con-set. It will make no difference to the argument if one speaks instead of one overall set of reasons with ‘pro’ and ‘con’ parts. The issues will be the same. It may appear that I assume that each set has only one reason as a member. More typically, this is not so and the set of pro- or con-reasons contains many members. Davidson, Donald (2001) ‘Introduction’, Essays on Actions and Events, Oxford University Press, Oxford, p. xvi. I am indebted to Richard Bradley, LSE, for helping me see this point. In fact, Buridan, it seems, was not responsible for the standard presentation of this case, wrongly named after him. For a discussion of Buridan-type cases, see Kane, Robert (1998) The Significance of Free Will, Oxford University Press, Oxford, pp. 198–9. Of course, the choice to A that he would have made or the A-ing he would have performed had he not had a reason to B must differ from the choice to A that he did actually make or the A-ing he actually did do in at least one way, simply in virtue of the fact that it would have been a choice made in the absence of having a conflicting reason to B. The qualification, ‘in some intrinsic way’, is meant to exclude such trivial differences. Bernard Williams, Moral Luck, p. 27 and ff. I do not think that one should underestimate the importance of the shift from the personal to the subpersonal level, in order to maintain (1) and (1*), broadened to include con-reasons. It is a major concession on the part of the causalist. I do not intend to develop the point here, but certainly the hope that lay behind the causalist programme for reasons for action was that reasons could be construed as causes, yet doing so was compatible with understanding reasons and actions in their own terms, sometimes called ‘the space of reasons’. This programme was not necessarily committed to construing reasons and actions as ‘really’ about brain states and gross behaviour (even if they turn out to be identical to brain states and gross behaviour). The language of psychology and action was meant to have an internal coherence and integrity all its own. To that extent, this option can easily take the causalist programme somewhere it had not intended to go. Note that this example is not one of pre-emption, as some have suggested to me. If it were a case of pre-emption, one would have two reasons both favouring the same line of action, the first of which causes the action and the other of which did not cause the action but would have caused the same action, had one not had the first reason. This is certainly not the case we are considering. But, arguably, all

74 Con-reasons as Causes cases of pre-emption involve some sort of causal inhibition or prevention, as does the case we are considering. In causal pre-emption, the inhibition or prevention is by the pre-empting cause of some node on the chain that would have led from the pre-empted cause to the effect. In the case under consideration, the inhibitor might belong to the first, ‘pro’ set of reasons but might also arise from some other quarter entirely. It will not matter which is true, for the argument. For the sake of the argument here, all that matters is that there be some inhibitor or other that interrupts the causal chain leading from the con-reason and which would have led to an alternative course of action had it not been blocked or inhibited. 16. In Goldman’s ‘A Causal Theory of Knowing’, Goldman expresses the hope that ‘inference is a causal process, that is, when someone bases his belief of one proposition on his belief of a set of other propositions, then his belief of the latter propositions can be considered a cause of his belief of the former proposition’ (p. 73). Goldman is speaking of reasons for belief, not for action, but I think many of the issues transpose to a Goldman-type programme for the causal modelling of reasons for belief from what I argue for causal modelling of reasons for action. In deciding what to believe, the epistemic agent is often confronted both by disconfirming as well as confirming evidential beliefs. Both the confirming and disconfirming evidence is deliberatively relevant to the same belief, and yet, if the confirming evidence causes my eventual belief, there is an analogous problem about what the disconfirming evidence is meant to cause. Goldman, Alvin (1967) ‘A Causal Theory of Knowledge’, Journal of Philosophy, 64, pp. 355–72. 17. Dancy, op.cit. chapters 4–6.

5 Agential Reasons and the Explanation of Human Behaviour Peter Hacker

1

Explaining human behaviour

Human behaviour is varied in type, reflecting both the manifold abilities of mankind and the complexity of social forms of life and associated institutions. The vocabulary of act-descriptions is correspondingly rich and variegated. We need, and often want, to understand our fellow human beings – for we are social creatures, living in close contact with, and in various forms of cooperative and competitive relationships with others. So, when their behaviour is not transparent, or when a tale of their behaviour is opaque, we ask for explanation. Correspondingly, we give others an account of our own behaviour. We strive, sometimes, to make ourselves intelligible to others. We tell others fragments of our autobiography – for amusement, to elicit sympathy or induce interest, to vindicate or exculpate, as example or counterexample. In the course of so doing, we explain our behaviour (as well as our thoughts and feelings) in a variety of ways, of which explanation in terms of agential reasons is but one. What we thus explain can be variously characterized in terms of control, intellect and will, on the one hand, and in terms of a variety of forms of context-sensitive descriptions on the other. These constrain in various ways the kinds of explanations that are possible and delimit the scope of explanation in terms of reasons. So, before we examine the categories of explanation of behaviour, we must first sketch the various ways of characterizing the explananda. (a) Behaviour: control, intellect and will – Human behaviour can be mere doing that falls short of action. We may slip, stumble, choke – these are not actions we perform voluntarily, not voluntarily or involuntarily. Some of the things we do are unchosen reactions, such as blushing, weeping, laughing, sneezing and trembling. Some of these, although not such as can be initiated at will, can often be stopped or partially suppressed at will. To that extent they verge upon the voluntary. The category of human action is the category of the kinds of things human beings can do or refrain from doing at will.1 75

76 Agential Reasons, Explanation of Human Behaviour

Accordingly, acts can be chosen, attempted and intended. One can deliberate whether to V, decide to V, be asked or ordered to V. Depending upon the disposition of one’s will and intellect, one can V impulsively, intentionally, deliberately, willingly or reluctantly, thoughtfully or thoughtlessly. Action can be voluntary without being intentional: (i) If it is an inhibitable reaction which one has not inhibited, such as a yawn or sneeze. (ii) If it is done knowingly but neither because one wanted to do it nor for a further reason, as when one whistles while one works or gestures while one speaks (and knows one is so doing). (iii) If it is knowingly done as an unwanted (or not wanted) consequence or by-product of intentional action, as when one crushes the grass beneath one’s feet while walking across the lawn or wakes one’s wife when putting the cat out at night. Intending belongs to a family of concepts that includes being about to, meaning to, having it in mind to, proposing to and planning to. The objects of intending are varied: one may intend to V; one may intend another to V if one has a form of authority over them; one may intend, or intend another, to be or become a so-and-so; and one may intend a certain state of affairs to obtain if one has appropriate control over whether it comes about or is maintained. Intentional actions, like voluntary actions, are an exercise of two-way powers. An intentional action need not be voluntary (action done under duress is not) but it cannot be involuntary. One may form an intention in advance of acting and then act on it (or fail to, because, e.g., one is prevented, one forgets or changes one’s mind). Such intention-formation may be the upshot of deliberation and decision. But one may act intentionally without any antecedent deliberation or decision. If an agent V-s intentionally, he must know that he is V-ing and must be doing so either because he simply wants to, feels inclined to or feels like it, or because he has a reason or further reason for V-ing. An agent may V with a further intention of X-ing. The further intention in the act may be merely concurrent (as when one goes to the cinema with the intention of having a drink afterwards), or V-ing may be a means to X-ing (as when one climbs up a ladder in order to change a light bulb). In the latter case, but not necessarily in the former, to X is A’s reason for V-ing, and ‘X-ing’ (or, in some cases, ‘trying to X’) is a further description of what A is doing. To say that A’s V-ing is intentional is not to give A’s reason for V-ing; but to say that A V-ed in order to X is. Here, to X is A’s purpose in V-ing and his reason for V-ing. One may also V with the intention of doing X, without V-ing intentionally, if one V-s accidentally or by mistake – as Tyrrel accidentally killed the king, intending to kill a stag. Clearly, one may V intentionally, on purpose, without V-ing for a purpose, as when one V-s with the intention of V-ing, as opposed to V-ing unintentionally

Peter Hacker

77

(unknowingly, by accident or by mistake) or not intentionally (knowing that one is V-ing but not wanting to V, as when one wakes up one’s wife while letting the cat out). That one V-s with the further intention of X-ing does not imply that one’s purpose in V-ing is to X. When one goes to the cinema with the intention of catching the 10:30 bus home, one’s purpose in going to the cinema is not to catch the bus home. So specifying a further intention with which one V-s need not give one’s reason for V-ing. Specifying one’s purpose in V-ing, however, does. For one cannot V with the purpose of V-ing, but only V with the purpose of doing something else. To specify the purpose for which one did something is to explain why one did it.2 Is every intentional act done for a reason? That depends on whether one takes V-ing ‘just because one feels like it’, or ‘because one feels inclined to’, as V-ing for a reason. If so, then all intentional action is done for reasons. If not, then one must distinguish between knowingly doing something just because one felt like it, ‘for no particular reason’ as we sometimes say, from acting intentionally for a reason. Certainly to feel a sudden urge or impulse to V is not necessarily to have a reason for V-ing, and to V because one feels an urge or impulse to is not necessarily to act for a reason. The urge, impulse or craving may be ‘quite mad’, that is, completely irrational. The consequences of our intentional actions may themselves be intentional, unintentional (if not foreseen) or not intentional (e.g. if foreseen but not part of the agent’s purpose) and the agent’s action may accordingly be redescribed in terms of its consequences. For example, if the agent opened the door, as a consequence of which the room cooled down, then he intentionally, unintentionally or not intentionally cooled the room down by opening the door. The act, thus described, obviously admits of various kinds of explanation, depending on whether it is intentional or not, and, if not intentional, whether done knowingly or unknowingly, voluntarily or not voluntarily, by accident or by mistake, and so forth. The cognitive and volitional character of the behaviour under consideration can often restrict the range or even determine the character of the explanation that might be offered for it. One cannot have a reason for doing something involuntarily or a motive for doing something inadvertently, and one cannot ask for the point of someone’s doing something unknowingly. One can ask for the purpose of an intentional action, but not of an unintentional one. (b) Behaviour: alternative descriptions – Human behaviour can be described in a large variety of ways. Often it can be described in terms of physical movements. Whether the movements are merely the movements of the agent’s limbs and body (‘his knee jerked’) or whether they are movements the agent makes (‘he moved his leg’) constrains the kinds of explanation appropriate. But there are many things human beings do, and for which we may want an explanation, which are not movement-involving, such as thinking of a solution or planning a move; looking at a painting, watching a scene, listening

78

Agential Reasons, Explanation of Human Behaviour

to or for a noise; waiting for a bus; keeping the door open, holding tight, gripping firmly; concentrating on a problem, a lecture or on what one is looking at; omitting, abstaining or refraining from doing something. Whether or not behaviour is movement-involving, it can be described or redescribed in many different ways, by reference to its antecedents (as is patent in such verbs as: to reply, repeat, re-enter, requite), its consequences (e.g. such causative verbs as to break, mend, dry, wet, open, close, pacify, irritate, console) or its circumstances. The varieties of ways in which acts can be described or redescribed in terms of their circumstances is large, and can be variously classified. Describing an act in terms of its circumstances specifies it, often as a species of a larger genus of a given kind of act. Commonly, actdescriptions can be analysed in terms of a simpler act and the specifying circumstance, as perjury, for example, is telling a lie in the circumstance that one is under oath, or bigamy getting married in the circumstance of already being married. The circumstances of an act can be construed narrowly, so as to exclude such subjective factors as voluntariness, knowledge and intention (and their negations), or it can be stretched to include these. Many moral and legal act-descriptions incorporate such factors: to steal is taking something, knowing one has no title to it; to lie is to tell an untruth with intent to deceive. Acts can also be described in terms of the manner of their performance (e.g. to mumble one’s thanks, to hurry to school), or in relation to the identity or some feature of the patient affected by their performance (suicide determines the identity of the patient with the agent, rape requires the unwillingness of the patient). A multitude of acts, as is evident from the foregoing examples, are described by reference to social, moral or legal norms that prohibit them (e.g. murder, trespass), require them (pay taxes, obtain a license), or make them possible (marry, make a will, score a goal, checkmate one’s opponent). The character of the description of the behaviour that requires explanation commonly excludes certain kinds of explanation. Acts that by their nature can be performed only once (such as suicide), cannot be explained in terms of tendency or frequency explanations, such as habit and custom (a suicidal tendency is not a tendency to commit suicide, but a standing temptation to do so). Acts that by definition or description are mistakes or accidents cannot be explained in terms of reasons, purposes or motives; nor can involuntary acts or doings that are not acts (such as slipping or stumbling). Acts that can be performed only intentionally, such as cheating, lying, forging exclude explanation in terms of inadvertence and inattention. Certain kinds of omissions can be explained by reference to lack of legal power or right. But possession of legal (or other normative) power does not explain why an agent exercised it on an occasion, it explains only that it was possible for him to do so. (c) The varieties of explanation – Citing an agent’s reasons for thinking, feeling or doing something explains why he thought, felt or did what he did. It

Peter Hacker

79

is one form, perhaps the most distinctive and important form, of explaining human thought, feeling and action. But it is not the only form, and before we examine its conceptual character in further detail, we should bring to mind the variety of explanations of human behaviour. ‘Why did A V?’ is the most general form of a request for an explanation of A’s doing what he did, and it can accordingly elicit any form of explanation of human behaviour, depending only upon the constraints imposed by the description of what the agent did. The kinds of answers that can be offered to explain behaviour are delimited not only by the description of the act, but also by the kind of question. For commonly, the speaker phrases his question in a manner that restricts the possible answers. ‘What did A V for?’ is a request for the point and purpose of an act and it presupposes its intentionality. It precludes various forms of efficient-causal explanation, as well as explanations in terms of ignorance and inadvertence. ‘What made A V?’ can be asked if V-ing was not an act of A, but something he did (such as stumble, slip or choke) or if it was an act, but one which was distinctly uninviting, unreasonable or irrational – something which A would not have done but for the explanatory factor. In the first kind of case, the question invites a causal explanation: A slipped because the pavement was icy; he fell because he was pushed; he choked because a piece of bread got stuck in his throat. In the second kind of case, an explanation in terms of an obliging factor is invited: he divulged the secret because he was threatened; he took the slow road because the motorway was closed. But a causal explanation is also possible in some cases, for example, he acted as a result of post-hypnotic suggestion. We can distinguish (non-exhaustively) the following forms of explanation of behaviour:3 i. Explanations in terms of constitutive redescription: We are sometimes puzzled at certain kinds of ritualized, semi-ritualized or otherwise rulegoverned activities. Why are these people gathered together and behaving thus? – They are engaged in a political demonstration, a trade union dispute, or celebrating a victory, commemorating an event, attending a wedding, a funeral or memorial service, playing a game and so forth. Here the explanation consists in redescribing the behaviour in its broader social setting and with reference to social norms. In this setting, behaving thus counts as doing so and so. Understanding what they are doing in these terms often resolves our puzzlement, for we may need no further explanation of why they are doing what they are doing thus described. ii. Explanation by polymorphous redescription: A different kind of explanation by redescription is explaining a specific activity in terms of its polymorph, as when we explain someone’s shutting the door as obeying an order, someone’s singing as practising for a recital, someone’s visiting as his keeping his promise.

80

Agential Reasons, Explanation of Human Behaviour

iii. Regularity explanations: We commonly explain behaviour as an instance of a regularity. Why is Emma going for a walk? – It is a habit of hers to go for a walk every afternoon. Why is Mr Darcy changing his clothes? – It is customary to dress for dinner at Pemberley. Why did Fanny Price hesitate to ask Sir Thomas? – Because she is timid (a disposition and character trait). It is noteworthy that these different kinds of regularity explanations are not causal and the explanation is not subsumption-theoretic despite the fact that the behaviour is explained as instantiating a regularity. They do not exclude explanations in terms of agential reasons, but they indicate that the behaviour is not exceptional. In many cases, no further explanation is required. iv. Inclination-explanations: We often explain behaviour as exemplifying a preference (that may be without reason) or liking either for the specific thing or activity in question, or for things or activities of the kind in question, or a dislike for some thing or activity. The inclination-explanation, in the cases of likings and dislikings, may also refer to a regularity. Sometimes we explain behaviour by reference to feelings that involve inclinations, as when we attribute A’s outburst to his feeling indignant (hence feeling inclined to protest) or greedy (hence feeling inclined to take a second helping of pudding). Occurrently felt emotions and attitudes such as anxiety, fear, tenderness, pity, amusement, curiosity are all associated with inclinations to behave, and conduct is commonly explained by reference to such a felt inclination. Hence, someone may be trembling out of fear, laughing with amusement, their tears may be tears of tenderness or pity. The behaviour that we explain thus is not caused by the emotion but is an expression of it. v. Causal explanations: Causal explanations obviously play a significant role in rendering human behaviour intelligible – but not the ubiquitous role that philosophers who conceive of reasons as causes suppose. We explain doings that are not actions, such as stumbling, slipping or blushing, in terms of causes. We explain many involuntary or only partially voluntary actions causally: he laughed because he was being tickled, cried out because he was hit, sneezed because he had taken snuff. Whether, in what sense and to what extent, intentional action can be explained causally will be examined in Section 3 below. vi. Explanations in terms of reasons: Intentional action is the primary object of explanation of behaviour in terms of agential reasons. Saying ‘I just intended to V’, unlike saying ‘I just wanted to’, does not explain one’s V-ing. But specifying one’s purpose (which is one kind of ‘intention with which’) does. However, there are many other kinds of factor that can be an agent’s reason for acting. We can classify them in numerous different ways: facts and values, norms and obligations, backward- and forward-looking reasons (past facts, prospective consequences, purposes and goals). It is a moot point whether desires, wants and preferences, on the one hand, and beliefs (the believing, not what is believed), on the other, should be regarded as reasons for doing anything. Some philosophers have argued that all reasons are

Peter Hacker

81

desires and beliefs, or combinations thereof. Others insist that desires are not reasons at all – it is the reason for wanting to V that is a reason for V-ing; and believing is rarely a reason for acting – it is what is believed that is normally a reason. We shall examine this below. vii. Motive explanations: Motive explanations are more specialised than explanations in terms of reasons.4 To give A’s motive for V-ing (e.g. revenge) need not be the same as giving his reason (e.g. that B killed his father). Acting out of a given motive is not the same as exemplifying a disposition – A may act out of revenge on a single occasion without being of a vengeful disposition. To act out of irritability, timidity or conceit is not to act from any particular motive. Motives are not feelings, even though a person’s behaviour can be explained by reference to the motive of pity or love. One may feel pity or love, overflow with compassion or affection – but one cannot feel one’s motives or overflow with them. To say that someone acted out of anger is not to explain his behaviour by reference to a motive, but by reference to an emotion. Motives, unlike feelings and emotions, cannot be pleasant or unpleasant, do not wax and wane, and do not occur at a given time or place. They are not kinds of mental event or state, but kinds of explanation. One salient type of motive-explanation specifies a pattern of backward- and forward-looking reasons. To say that A V-ed B out of gratitude is to explain A’s V-ing by reference to B’s having intentionally benefited A (a form of backward-looking reason), and A’s V-ing in order to requite B’s generosity. But A’s specific reason, for example, that A gave him a generous loan when he needed one, is not mentioned. Similarly, to explain A’s V-ing B as done out of revenge is to explain the action in terms of B’s having harmed A or an interest of his, and A’s acting in order to harm B in return. But A’s specific reason for V-ing B, for example, that B killed his father, is not specified. In many cases, the pattern is of a remedy-demanding antecedent circumstance that provides a backward-looking reason for acting, an action with a certain intention, and a prospective consequence that is thought to remedy the prior condition – thus providing a forward-looking reason. Differences between motives (e.g. between ambition, greed, patriotism, vindictiveness, gratitude) are then evident in the kinds of antecedent remedy-demanding condition and the kinds of prospective values or advantages aimed at. Some motives, such as revenge, friendship or gratitude, involve not merely acts done with an intention that exemplifies a pattern, but acts that can be done with the intention of exemplifying a pattern. Here one can speak of acting in order to wreak vengeance or to show friendship or gratitude. To these forms of explanation of action we may add further forms of explanation distinctive of inaction: i. inability, physical, technical (skill) or intellectual ii. prevention, natural or human

82 Agential Reasons, Explanation of Human Behaviour

iii. iv. v. vi. vii.

lack of equipment lack of opportunity ignorance of opportunity forgetting one’s intention lack of normative power

The question of why A did not V may be just as pressing as the question of why he did. It should be noted that some factors can be cited when offering one kind of explanation which may be logically excluded from another. So anger may be a factor in an explanation in terms of a passion, and timidity a factor in a regularity explanation, but neither can occur in a motive explanation. Values (fairness, filial piety, friendship) and norms (obligation, duty, rightness) may be cited as factors in explanations in terms of reasons, but not as factors in causal explanations.

2

Explanation in terms of agential reasons

A person’s reason for V-ing can be given by means of a nominal clause ‘that p’ or by means of infinitive clauses ‘in order to W’ or ‘to attain G’ (‘His reason for going to London was to see the Queen’, ‘To attain fame was Hume’s reason for writing the book’, ‘He went to Egypt for no other reason than to see the Great Pyramid’). We commonly use the infinitive clause when we explain why A V-ed by specifying his reason in the form of his goal or purpose (that for the sake of which he acted). But we also explain A’s behaviour in terms of his reasoning (actual or counterfactual). Here we give his reason in the form of a nominal clause that can appear, denominalized, as a premise in reasoning that supports the conclusion to V. So the infinitive clause that was used in explaining that he V-ed in order to attain G will find its correlate in a premise citing a goal or desire for an end. A’s reason for V-ing can be that p or in order to W or to attain G only if A knows or believes that p or that V-ing is a way of W-ing, or a means to W, or necessary to attain G.5 A reason, specified in sentential form, is a step in reasoning. If that R is A’s reason for V-ing, then R, in conjunction with other premises, must support, or be thought by A to support, the conclusion to V. But it would be mistaken to suppose that whenever A V-s for the reason that R, he goes through a process of reasoning ‘in his head’, let alone aloud. Often what is to be done is too obvious to require, or too pressing to allow, any deliberation, yet one can give reasons for one’s action. Often what is done is spontaneous – but not therefore without reason. To give one’s reasons for V-ing is not the same as reporting one’s reasoning. There are various criteria for something’s being A’s reason for V-ing. Typically, they consist in the antecedents and context of his action. Often a person’s reason for doing what he does is patent – seeing that it is raining,

Peter Hacker

83

he puts up his umbrella; having been told that a ticket costs a pound, he hands over a pound coin; wanting a drink, he says to the bartender ‘May I have a scotch’. There is nothing mysterious about our ability to discern immediately the reasons for much of our fellow human beings’ behaviour. We see the actions of other human beings, in the contexts in which they are embedded, as intentional, purposive and commonly conventional, that is, governed by constitutive rules that define the act and its meaning. We do not see them as mere ‘bodily movements’ (similarly, we see the world around us as consisting of relatively persistent coloured material substances in objective space and time – not as expanses of colour, here, now, in one’s ‘subjective’ visual field).6 But sometimes it is not obvious what another person is doing, or it is opaque what his reasons are for doing what he is doing. So we may ask him. If he sincerely says that his reason for V-ing is that R, then, normally, that R is his reason for V-ing. Looked at askew, this can seem mysterious. If the agent’s answer is not a report on the reasoning he has just gone through, on what he said aloud or to himself, what exactly is it? How does he know what his reason is or was? Is it a hypothesis of his that his reason was that R? Why should his word be accepted? What sort of authority (if any) does his word have, and whence does it derive? One way to approach these questions is ‘analytic-genetic’, that is, to reflect on how we might conceive of the roots of the language-game of giving reasons for our actions. A human child, like other young mammals, finds a multitude of things in its environment attractive or repulsive, sources of pleasure or of fear and dislike, and objects of curiosity with regard to his ‘interference’ in the course of things. He learns to act in pursuit of what he wants, and, as his rudimentary linguistic skills develop, he learns to supplement his frustrated strivings and screamings to get with a ‘Want!’, and subsequently to substitute for them an ‘I want’ and later still ‘I want the (or a) so-and-so’ (such utterances, it should be noted, are not reports on his mental state). As his motor skills develop he learns to manipulate objects in his environment, and learns what happens when he acts on (pokes, pushes, pulls, throws) them. Rapidly he, like any other young mammal, learns to bring about such-and-such by doing so-and-so. But unlike other young mammals, he is learning to speak. Once he has mastered the rudiments of speech, the child may be asked what he is doing or trying to do, why he wants to do what he is trying to do, what his purpose is in doing what he is doing. He learns to describe his actions. Later he learns a new language-game: the practice of announcing an action: ‘I’m going to V’ – for which he must learn that when one says ‘I’m going to V’ one must go on to V. Here lies the primitive roots of intention-formation in advance of action, of our non-inductive knowledge of our own future actions, and of the relative predictability of the action of others on the basis of announced intention. (If announced intentions for the immediate future were not typically followed by the intended action, the very concept of intention would lose its point.)

84 Agential Reasons, Explanation of Human Behaviour

Once this language-game is in place, there is room for further questions namely: ‘Why are you going to V?’, and for answers of the form ‘In order to W’, ‘To get X’ and ‘Because I want X’, that is, specification of one kind of reason, namely the purpose of V-ing, that for the sake of which the V-ing is to be done. The answer ‘Because I want X’ invites the question: ‘Why do you want X?’ and the child learns to specify what it is about X that is attractive or appealing – it learns to give a further kind of reason, namely to specify a desirability characteristic. And the child also learns to specify what it wants X for – for example, to eat it, or to play with it or as a means to a further end. Here we have the beginnings of reasoning and of giving reasons. It is noteworthy (and not contingent) that reasoning marches hand in hand with intention-formation. The transition from voluntary and purposive action that lies within the competence of mere animals to intended and deliberative action is an aspect of the emergence of rationality. The child’s emergent rationality is his growing ability to reason from given data to conclusions. A criterion for his reasoning thus is his answering such questions. To learn to answer such questions, to learn what counts as an adequate answer, is itself to learn to reason. In sincerely saying that his reason for V-ing is that R, the agent need not be recounting the reasoning he has just gone through – he may not have gone through any. So what makes R his reason for V-ing? He makes it his reason. For although one cannot choose to have a reason to V, one can, reflecting on the reasons one has, choose or decide to V, for such-and-such a reason (rather than from another reason (that one has)). It is this reason, not that one, that one takes to be decisive, or that one takes to justify, one’s V-ing. It is by reference to this factor, not that one, that one would justify one’s decision and that one would subsequently cite to explain why one did what one did. It is not that he decides to-V-for-the-reason-that-R, but rather for the reason that R (and not for the reason that S), he decides to V. If A is told to leave the room by B, he may say, ‘I am leaving the room because you told me’, or he may say ‘I am leaving the room, but not because you told me’.7 What his reason is for leaving the room is a matter of his reasoning – he decides to V for that reason, and his avowal of his reason characterizes his action. Each reply, in different scenarios, may indicate very different things. The first may indicate A’s compliance and willingness to obey; or it may signify resentment (since he would not have left the room otherwise); or it may be meant to shift responsibility for A’s absence onto B. The second reply may signify A’s indignation at something that preceded the order to leave; or outrage at B’s impertinence; or to notify B that A was going to leave anyway; or to remonstrate that despite appearances A would not dream of taking orders from B. In sincerely giving his reason, A provides others with an interpretation of his behaviour – a particular way of understanding it; but he does not interpret it. The way of understanding he thus provides them with is his way of

Peter Hacker

85

understanding his own behaviour – the way he conceives it. In saying that he is V-ing because it is the case that R, A is not merely citing R as a reason for him to V, he is declaring it to be his reason for V-ing, endorsing a particular teleological/hermeneutic explanation of his V-ing, and taking responsibility for it under the description ‘V-ing, for the reason that R’.8 Patently, his reason for V-ing is not something he discovers to be such, nor is it a hypothesis of his – although he may lie to others and may deceive himself. What is sometimes denominated ‘first-person authority’ in these contexts is mischaracterized thus. The agent is not ‘an authority’ on his own reasons for acting, as he might be an authority on Sanskrit literature. His sincere word is a criterion for his reason’s being that R, but not because he infers this from evidence he has which others lack. He has no epistemic authority, but rather verdictive power. This hermeneutic account illuminates cases of an agent’s giving his reasons for V-ing ex post actu where he went through no reasoning, and where his giving his reasons is neither recounting his reasoning nor reporting an explicit decision to V for the reason that R. Here, one might think, the agent’s ability to explain or justify his action by citing his reason is mysterious. For if there is nothing for him to remember by way of reasoning, how can he now say what his reason for V-ing was? One might suggest that in declaring R to have been his reason, he is averring that he would have cited R as his reason had he been asked. This can be misleading. Although it may be true that had he been asked, he would have given R as his reason, this counterfactual is not a hypothesis based on past experience. After all, the normal form of thus giving one’s reasons for a past action is not ‘I think my reason was that R’, or ‘Judging by my reason for V-ing last time, it is probable that my reason was that R’. Indeed, were someone to respond thus in standard cases we would doubt his sanity. Nor is saying that his reason was that R a matter of ‘inner observation’ – as if he were, so to speak, reporting a causal connection ‘seen from the inside’. For there is no such thing as introspectively observing his reason ‘at work’, so to speak, since introspection is not a form of observation, and reasons don’t ‘work’ – they are neither causal agents nor mental events. The ability to give one’s reasons ex post actu should be compared with the ability a speaker has to say what he meant by an utterance (e.g. ‘He is by the bank’) or whom or what he meant by a demonstrative or indexical in his utterance (e.g. ‘He’ll be there then’). The question of how the speaker knows what his reasons are or were, in the standard case, should be rejected as being as misconceived as the question of how a person knows what he means or meant. The agent’s avowal of a reason ex post actu must render intelligible the intention in the act. For it must present the reasoning that he holds to warrant the act, even though, ex hypothesi he went through no reasoning. It must provide his rationale for having performed it. So the reason he gives

86

Agential Reasons, Explanation of Human Behaviour

must cohere with the context of his action, the motivational background and his subsequent behaviour. (Failure so to cohere, and, in some cases, failure to cohere with his general behavioural dispositions, may cast doubt on his sincerity or on his self-understanding and self-knowledge.) The agent, we assume, knows now what he then knew or was aware of (which others, trying to understand why he acted as he did, may not know), and he knows now what he then believed (otherwise he may say that he cannot recollect why he V-ed). He also knew what he was doing (otherwise he was not acting intentionally) and still knows, and knows whether he succeeded or failed to do what he was aiming to do (so he can, in appropriate cases, say what he was trying to do). Thus citing his purpose as the reason why he V-ed is little more than being able to elaborate what his intention in the action was. Citing backward-looking reasons for his V-ing (e.g. that he was responding to B’s X-ing and not to C’s Y-ing), none of which he stated (either overtly or sotto voce) at the time since he acted purposefully without reflection or deliberation, is not to interpret his action. That is something others may do, or he may do under very special circumstances of self-doubt. Rather, it is to determine its character. In saying that his reason for V-ing was that R, he is making a connection between the facts as he knew them and his action (or his action and recollected purpose). In making that connection, he is endorsing a certain kind of description of his behaviour, and taking responsibility for it under that description. Where an agent has (or had) multiple reasons for V-ing and avows that he is V-ing (or V-ed) for one reason rather than another (because it would help Jack, not because Jill was looking; because it was right, not because it was also profitable), doubts may arise. He may be lying. He may be deceiving himself with regard to his own motivation. Or his motivation may be intrinsically, not epistemically, opaque. Here is an aspect of human behaviour and its understanding that can be constitutionally indeterminate. A may insist, in all sincerity, that he V-ed for the reason that p, but B may repudiate A’s explanation, insisting that he V-ed for the rather less reputable reason that q. A may persuade B of the correctness of his explanation (it never even occurred to him that q). Alternatively, B may convert A to seeing his own behaviour differently, perhaps in the light of facts about his character or facts about his past that he had not brought to mind. Or their disagreement may remain unresolved. Sometimes the failure of resolution may not be epistemic. The facts of the case may all be known, and yet lend themselves to more than one perfectly plausible story. In such cases, the disagreement may, sometimes tragically, be unresolvable.

3

Causal mythologies

We have displayed human action for reasons as a form of teleological behaviour (behaviour done for a purpose, for the sake of a goal), even

Peter Hacker

87

though it often incorporates backward-looking reasons. But even when an agent acts for backward-looking reasons, for example, thanks B because B did him a favour, or benefits B because B acted meritoriously, A does so in order to express gratitude or reward desert – that is his purpose. We have emphasized the role of reasons in reasoning, either in advance of acting or after the event. An agent’s citation of his reason for V-ing presents a warrant for his behaviour. But the most popular philosophical account of acting for reasons is causal (i.e. in terms of efficient causation). Reasons are held to be causes, and explanation in terms of reasons is argued to be a form of causal explanation.9 This conception needs to be confronted. It is bizarre to suppose that reasons are causes, since in reflecting on the reasons for and against doing something or other, one is surely not reflecting on what will cause one to do what reason dictates, but rather trying to decide what reason dictates. And in recounting, after the act, for what reason one V-ed, one is not describing or explaining what caused one to V, but what one’s rationale for V-ing was. Moreover, if one’s reason for acting were a cause of one’s behaviour, it would be puzzling how it is that one normally knows this cause (save in cases of ‘mixed motives’, self-deception or forms of pathology). For although one can know immediately what startled one (the sudden explosion made one jump) or what caused one to slip (the banana skin made one slip), it is far from obvious that knowing that one’s reason for going to London is to see the Turner exhibition (but nothing made one go) or knowing that one’s reason for staying home is that one promised to do so (but nothing made one stay at home), is comparable. The causal theorist cannot argue that when one cites one’s reason for V-ing, either in advance of, or after, acting, the reason is the cause of one’s acting. For it is clear that reasons are not causes. One’s reason for going to London may be in order to see the exhibition, but in order to see the exhibition cannot be a cause of one’s going; one’s reason for taking an umbrella may be that it is going to rain, but that it is going to rain cannot be the cause of one’s now taking an umbrella; one’s reason for writing ‘25’ in the questionnaire may be that √625 = 25, but that √625 = 25 cannot cause one to write ‘25’. We attribute reasons to an agent: a person can be said to have reasons for V-ing or for refraining from V-ing – but cannot be said to have a cause for V-ing. The reasons a person has may be good or poor, selfless or selfish, moral or immoral – but causes cannot be any such thing. Reasons justify or purport to justify the action for which they are a reason, causes do not even purport to do so. Accordingly, reasons may be convincing, defensible, weighty, compelling, persuasive, weak or barely acceptable. But causes can be none of these. The fact that p may not have been A’s reason for V-ing, yet it may be a reason for V-ing for all that; but one cannot say that if event e1 was not the cause of e2 it was a cause for e2 for all that. The causal theorist may shift ground and argue not that reasons are causes, but that explanation in terms of reasons is a form of causal explanation. His

88 Agential Reasons, Explanation of Human Behaviour

claim would then be that if an agent wanted goal G and believed that V-ing was a way to attain G, and V-ed for this reason, then his wants and beliefs conjunctively caused the agent to V. ‘If the connection between desire and action is not some kind of causal connection, what kind of connection can it be?’, the causal theorist may exclaim, ‘Perhaps the adoption of the causal framework is not a mistaken choice, not because it is the correct choice, but because it is inevitable’.10 It is inevitable, so it is claimed, because no noncausal analysis of acting for a reason can explain the difference between: i. A’s having a reason (construed as a combination of belief and desire) to V and Ving, but not V-ing for that reason. and ii. A’s having a reason to V and Ving for that very reason. Only a causal account of acting for a reason can make the relationship between an agent’s reason for V-ing and his V-ing for that reason intelligible.11 But, as should now be clear, that is wrong – the choice is not inevitable, only mistaken. The causal account, as its main defender admitted, cannot explain the relationship. But the hermeneutic account can. The causal theorist’s claim is that reasons are combinations of beliefs and desires. But that is mistaken. That I believe that p can be a reason for me to V, for example, if someone says ‘Would everyone here who believes that p please raise their hand’, then that I so believe may be a reason for me to raise my hand. But normally it is not my believing that p that is my reason for acting, it is what I believe, namely that p. Indeed, normally the kinds of factors I might cite as my reasons are not even what I believe but what I know (and it is debatable – though will not be debated here – whether, if I know that p, it follows that I believe that p). Nor is it true that my wanting to V, as such, is a reason for me to V. If I want to V, then my reason for V-ing is whatever reasons I may have for wanting to V. The causal theorist may claim that beliefs and desires are mental states, and it is they, or their ‘onset’ that cause an agent to act when he acts for a reason. But to believe something is not to be in a mental state. Felt desire (feeling thirsty, hungry, lustful) may qualify as a mental state, but the feeling is not itself a reason at all, although that one has it may be. And wanting to attain a certain goal is not a feeling. Mental states are paradigmatically moods, occurrent emotions, agitations, being in pain, concentrating on something. They obtain when one is awake and lapse with loss of consciousness. They have degrees of intensity, and may wax and wane. They can be interrupted by distraction of attention and later resumed. But if one believes that Hastings was fought in 1066, one does not cease to believe this when one falls asleep, there are no degrees of believing that Hastings was fought in 1066 (I can’t believe more than you, only be more or less convinced than you), and one’s belief cannot be interrupted by distraction of

Peter Hacker

89

attention and later resumed.12 If believing that p were a mental state, then it would make sense to say, ‘I believe that p [thus referring to my own mental state B], but it is not the case that p’. It would similarly make sense to say, ‘I believe that p [am in mental state B], but as to whether it is the case that p, that, as far as I am concerned, is an open question (or: that is a matter on which I take no stand)’. But these make no sense. For to assert that one believes that p is normally to take a stand on whether it is the case that p.13 So to believe something (a fortiori to know something) is not to be in any kind of mental state. The contemporary causal theorist may hold that the efficient causes of action are neurophysiological states and events that allegedly are (contingently token-identical with) mental states of believing and wanting. But believing that p could not be a neural state, since a neural state could not have the consequences of believing. Asserting that one is in a given neural state S is compatible with going on to deny that p, or withholding judgement on whether it is the case that p. But asserting that one believes that p is not. The only thing such neurophysiological states and events could explain are movements consequent upon muscular contractions, not actions. A neurophysiological explanation of why one’s fingers moved may be available, but not an explanation of why one was playing the piano or of why one was playing the Hammerklavier, let alone one that renders one’s playing intelligible in terms of polymorphous redescription, for example, that one is practising for tomorrow’s Beethoven concert or keeping a promise. A mere bodily movement is typically insufficient to determine what act was performed, and its neurophysiological explanation cannot explain whatever act was performed in or by moving voluntarily or intentionally. Reasons, by contrast, explain acts and omissions. It is true that the agent’s wanting to W or to attain G is commonly a factor in explaining his V-ing. But this explains the agent’s action not its constitutive movements. If we are puzzled by a chess-player’s gambit, we want an explanation of his move, not of his movement. Neural events in the brain may explain how it is that the chess-player is able to move his hand. If they did not occur as normal, he would not be able to move or control the movements of his hand. But neural events cannot explain his move – only those principles of chess strategy that he was aware of can do that. Nor can they warrant the redescription of his action in terms of its normative consequences, for example, that he checked his opponent. The notion of wanting is polysemic. ‘I want/He wants’ can be employed to signify the agent’s wish, inclination, failed attempt, felt desire (which may be an appetite, a drive, a craving, an urge), endeavour or purpose. The explanatory role of reference to what the agent wants or wanted can be correspondingly various. Sometimes ‘I did it because I wanted to’ serves merely to exclude other explanations, in particular causal ones. It excludes the possibility of my V-ing’s being an involuntary start, a nervous twitch or an

90 Agential Reasons, Explanation of Human Behaviour

electrical innervation. It identifies what I did as a voluntary action, therefore as something for which it makes sense to ask for my reason for doing it. It is no more a causal explanation of what I did than ‘I did it on purpose’. With such addenda as, ‘but not because you told me to’, it both identifies the action as voluntary and excludes one kind of reason. Often, ‘I V-ed because I wanted to V’ or ‘He is V-ing because he wants to’ intimates, in its context, that the agent finds V-ing in one way or another attractive or appealing, that he likes or enjoys V-ing. These desirability characteristics are reasons for wanting to V. So, if he V-ed because he wanted to, his V-ing was not instrumental – he was not V-ing with any further intent or purpose. His reason for V-ing is whatever reason he had for wanting to V. In such cases, normally, V-ing successfully yields satisfaction or gratification. In this use of ‘want’, a criterion for someone’s V-ing because he wants to V is that he is pleased to be V-ing or to have V-ed. (Sadly, it sometimes happens that we want to V and do so, only to find it disappointing. Here one’s want is satisfied, but that gives one no satisfaction at all.) ‘I V-ed because I wanted to attain G’ does explain the speaker’s V-ing by giving his reason. But the reason given is his purpose, that is, to attain G. ‘He V-ed because he wanted to attain G’ describes the aim of his act, not its psychological cause – one could just as well say ‘He V-ed in order to attain G’. And to say that he V-ed in order to attain G is not to specify a cause of anything. It characterizes his act of V-ing as a means (or supposed means) to a goal. It does not signify that the agent V-ed out of a desire to V. Here one can say ‘I didn’t want to V at all, I did it only in order to attain G’. Did he not ‘want the means’ (to V), given that he wanted the end (to attain G)? That would be misleading. Certainly the agent’s V-ing was an action of his, something he could have refrained from doing – but if it was done under duress, it was not voluntary at all (though not involuntary either). The victim did not want to hand over his money to the gunman – he chose to, in order to save his life. But there is no reason to assimilate choice to want. We often choose to do things we don’t want to do at all (and in the ‘thin’ sense in which a philosopher may insist that we want to do whatever we intentionally do, we may want not to do it much more than we ‘want’ to do it – judging by any normal criteria for strength of want.) So, when an agent V-es for a further reason, that does not mean that he V-es because he wants to V. As just noted, he may V under duress – his reason being to avoid the evil consequences of refusing to V. He may V because he has to (it is part of his job), because he promised to, because he was ordered to or because it is right. He may find V-ing odious, tiresome, may much prefer to do other things. But the reason for V-ing may outweigh all other considerations. He is not acting because he wants to V but because it is his duty. Does he not want to do his duty? No doubt; but that is not his reason for V-ing – which is rather that it is his duty to V.

Peter Hacker

91

Wanting can be a mental state, as when one is gripped by desire, overcome by appetite or beset by a craving. But then it is not a reason for acting, although to give quietus to one’s craving or appetite may be. More commonly, wanting is not a mental state at all – both when reference to it is merely indicative of purpose (in which case the verb can often be replaced by ‘in order to’), and when reference to it is indicative of liking or favouring. To explain an act by reference to wanting to do it, or wanting a further end, is neither to give a causal (nomological) explanation, nor is it to intimate that there is one. To explain that a person V-ed because he wanted to attain G, that is, that he V-ed in order to attain G, is not to explain his reason for aiming to attain G. That is explained by the desirability characteristics of G. The incoherence of the causal tale is made manifest by an argument of Wittgenstein’s.14 Suppose I form the decision to pull the bell rope at five o’clock (I want to call the butler and believe that by pulling the bell rope I shall do so). The clock strikes five. Should I now wait patiently for my arm to go up? If my wants and beliefs can be causes of my behaviour, then I should be able to sit back and let them bring about the movement of my arm. Should I, bearing in mind my antecedent decision, describe what happens with the words ‘and see! My arm goes up when the clock strikes five’? No; the upshot is not that my arm rises, but that I raise it – if I act in order to get what I want. Were the causal story true, I should, at least until the novelty has worn off, be surprised that when I have such-and-such wants and beliefs, such-and-such movements eventuate. Were the causal story true, how would an agent know whether he was acting voluntarily or moving involuntarily? For, as we have seen, one might have all the relevant beliefs and desires (the ‘primary reason’) and the appropriate movement might ensue, yet the question of whether I moved my limb would still be open. The causal story cannot distinguish between an agent’s limb’s moving because he has a certain reason and his moving his limb for a reason. But that is precisely what is explained by the altogether unmysterious rational-teleological account of human action in terms of reasons. Explanation of action in terms of agential reasons is a form of teleological explanation or an adjunct thereof. It enables us to understand the agent’s behaviour idiographically rather than nomothetically. Knowing his reasons for doing what he did, we may come to know what kinds of things weigh with him in his deliberations, and what kinds of considerations move him to act. We can see the extent of his rationality and the degree of his reasonableness, as well as the values for the sake of which he is prone to take action. Such explanations enable us not only to judge the agent and evaluate what he did, but also to judge his character. It enables us to understand our fellow human beings as persons.15

92

Agential Reasons, Explanation of Human Behaviour

Notes 1. So, perceiving in its various generic modes is not action, although watching, looking, gazing and peeking, and listening to, as well as listening for, are. Similarly, deciding, changing one’s mind, and in some cases, thinking and imagining are exercises of two-way powers and can be done at will (voluntarily, intentionally, on purpose). But we would not count them as forms of behaviour, and in some contexts not even as acts or action. It is noteworthy that certain kinds of success-descriptions, e.g. ‘win’, ‘solve’, ‘discover’, do not signify actions but only things done. Accordingly one can ask for the reason why A won, but one cannot ask for A’s reason for winning, solving or discovering (see A.R. White ed., The Philosophy of Action (Oxford University Press, Oxford, 1968),’Introduction’ p. 3). 2. See A.R. White, Grounds of Liability: An Introduction to the Philosophy of Law (Clarendon Press, Oxford, 1985), pp. 73–5. 3. For illuminating discussions of hermeneutics to which I am much indebted, see A.J.P. Kenny, Action, Emotion and the Will (Routledge and Kegan Paul, London, 1963) and Will, Freedom and Power (Blackwell, Oxford, 1975), Bede Rundle, Mind in Action (Clarendon Press, Oxford, 1997), G.H. von Wright, Explanation and Understanding (Routledge and Kegan Paul, London, 1971) and Practical Reason (Blackwell, Oxford, 1983), essays 1–5, and A.R. White, Philosophy of Mind (Prentice Hall, New York, 1967), chapter VI. 4. See Kenny, Action, Emotion and the Will, chapter 4 and White, The Philosophy of Mind, chapter 6. Note that motive explanations must not be confused with motive attributions which typically specify forward-looking reasons, as when we say that A had a motive for the crime, namely that he stood to inherit the victim’s estate. Having a motive, in this sense, amounts to there being a reason for the agent to V, and unlike acting for a reason, does not imply that the agent performed the deed at all. 5. In certain cases, that p may be A’s reason for V-ing even though it makes no sense for A to know (or not to know) or believe (or not to believe) that p, for example, if his reason is that he is in pain. For examination of these kinds of cases, see P.M.S. Hacker, ‘Of knowledge and of knowing that someone is pain’, in A. Pichler S. Säätelä eds, Wittgenstein: the Philosopher and his Works (WAB, Bergen, 2005), pp. 203–35. 6. In both cases, what is presented by philosophers as a foundation is actually an abstraction. 7. The example is Wittgenstein’s, see Investigations §487. 8. An account of this kind was intimated by Wittgenstein, Philosophical Investigations §§487–9 and developed by G.H. von Wright, ‘Of Human Freedom’, repr. in his In the Shadow of Descartes (Kluwer, Dordrecht, 1998), pp. 20–7. 9. The locus classicus for the causal conception is D. Davidson’s ‘Actions, Reasons and Causes’ of 1963, repr. in his Essays on Actions and Events (Clarendon Press, Oxford, 1980). He held that ‘failing a satisfactory alternative, the best argument for a [causal] scheme ... is that it alone promises to give an account of the “mysterious connection” between reasons and actions’ (ibid., p. 11). He did not consider the hermeneutical account we have proposed, but only the so-called ‘logical connection argument’.His paper re-instated the Vienna Circle’s methodological monism and the Cartesian/ empiricist account of action for reasons as movement caused by a mental act, event or state. The promise of an adequate causal account, however, was never fulfilled.

Peter Hacker

10. 11.

12.

13. 14. 15.

93

For Davidson ‘despaired’, as he put it, of distinguishing V-ing for the reason that R from V-ing because R (i.e. the right from the wrong causal chain), as when one’s desire to bring it about that p causes one to tremble, which brings it about that p. So ‘the best argument’ for a causal account of acting for a reason turned out to be no argument at all. D. Pears, ‘Sketch for a Causal Theory of Wanting and Doing’, repr. in his Questions in the Philosophy of Mind (Duckworth, London, 1975), p. 98. Davidson held that ‘cause and effect form the sort of pattern that explains the effect, in a sense of ‘explain’ that we understand as well as any. If reason and action illustrate a different pattern of explanation, that pattern must be identified’ (ibid., p. 10). But, as is evident from interminable controversies among philosophers about causation, the causal explanation is surely not generally understood ‘as well as any’, and the pattern of explanation in terms of reasons is no more problematic. It is no easier to gain an overview of the conceptual field of causes than to attain an overview of that of reasons. For detailed defence of the claim that believing that p is not a mental state, see P.M.S. Hacker, ‘Of the Ontology of Belief’, in Mark Siebel and Mark Textor, Semantik und Ontologie (Ontos Verlag, Frankfurt, 2004), pp. 194–202. See A.W. Collins, The Nature of Mental Things (University of Notre Dame Press, Notre Dame, Indiana, 1987), chapter 1–3. See Wittgenstein, Philosophical Investigations §627. This paper is a slightly modified fragment of chapter 7 of my book Human Nature: the Categorial Framework (Blackwell, Oxford, 2007).

6 Reasons as Non-causal, Context-placing Explanations Julia Tanney

I It is widely supposed that everyday explanations couched in terms of reasons, motives, intentions, and so on for an agent’s actions depend upon law-governed causal relations between states, events or properties which ordinary mental terms are alleged to pick out or in causal relations between to-be-discovered realisers of those supposed states.1 But this conception of the use of mental terms and of the kind of explanation they serve was disputed by philosophers influenced by Wittgenstein: those who conceived their task to be the untangling of philosophical perplexities thought to arise from inattention to logical or ‘grammatical’ detail. Such philosophers pointed to differences between the employment of mental concepts in everyday reason-explanations and of their (alleged) counterparts in psychological theories – those psychological theories, at least, which resonated to the ‘comfortingly causal talk characteristic of the hard sciences’.2 They argued that the everyday employment by teachers, lawyers, priests and doctors of mental concepts explain in a different sense of ‘explain’ from that favoured by the hard sciences. Wishing to emphasise the unlikeness between the two senses, these philosophers drew attention to the differences by contrasting reasonexplanation with causal-explanation and by insisting on important qualifications to the suggestion that reasons, motives and intentions are causes. It looked to some commentators that these philosophers were taking for granted a hopelessly simplistic and clearly mistaken view both about the concept of causation and about how causation works in non-mental domains. Thus, a number of philosophers today, though somewhat sympathetic with the writings of Wittgenstein and those he inspired, nonetheless refuse to fight the battle over causation.3 Although I shall not have much to say against these (in many ways kindred) positions, my hunch is that there may be good reasons for resisting the assimilation of importantly different senses of ‘explain’; such differences, I suspect, tend to be obscured by the appropriation by both camps of the concept of causation. 94

Julia Tanney 95

In this essay, I would like to trace the decline of the ill-understood Wittgensteinian perspective – paying close attention in particular to the writings of Melden, Ryle and Anscombe – in order to bring this particular orientation back into view. In Section II, I shall sketch my own understanding of the position, and, in so doing, answer some of the criticism of Davidson and Fodor. In Section III, I shall contrast the Wittgensteinian view with the instrumentalist and realist ones associated with Daniel Dennett as well as with the behaviourism (mistakenly) attributed to Ryle. I suggest that because on the Wittgensteinian/Rylean view mental concepts discharge their explanatory role other than by referring to a state, relation, event or property whose nature is in question, the position fails to find its place on the metaphysical map charted by realists and their irrealist opponents. In Section IV, I shall offer some hypotheses as to why reason-explanation as non-causal, contextplacing explanations have been resisted.

II Causation, Melden (1961) said, is one of the ‘snare’ words of philosophy. Looking carefully at how this word is used will not allow us to distinguish it from ‘reason’ or even from ‘explanation’: indeed, some use the words ‘causation’ and ‘explanation’ interchangeably.4 On this use, to deny that the mental is causally efficacious is to deny that the mental is explanatory and this (correctly) strikes most people as absurd. Nonetheless, there was (and, as we shall see, still is) a certain attraction to a broadly Humean view about causation; enough so that it will be worthwhile to bring out the differences between the explanations.5 On this broadly Humean understanding, causation is a relation between two logically and temporally distinguishable events. This is sometimes accompanied by the idea that the relation is explanatory insofar as it is subsumable under natural laws or law-like generalisations, which it is the business of the empirical sciences to discover. Let us agree to stipulate for the purposes here that we will understand a causal relation minimally as a relation between two logically and temporally distinguishable events. The position I wish to bring back into focus says that what it is for an action to be in execution of an intention or for it to be explicable by reasons is not a matter of there being a causal relation (in this sense) between intention or reasons and action. If causation is to be thus understood, the pattern in virtue of which a person’s intentions, motives or reasons explain her action is not eo ipso causal.6 For Melden, the motivation to construe motives, intentions and reasons as constituents in a causal-explanation of action is symptomatic of a misguided attempt to give an account of how an event construed as a mere bodily movement (an arm’s rising) can be construed as an action (someone’s raising his arm). No further description of the performance in respect of its

96

Reasons as Non-causal, Context-placing Explanations

properties as a bodily movement could possibly disclose that additional feature that makes it an action. This, Melden tells us, is for two reasons. First, the occurrence of a mere bodily happening (say, an event in the brain) does not have the logical force to turn a bodily movement into an action (and this would be so even if events of the one kind enter into lawful relations with events of the same type as the bodily movement to be explained). But, second, if we hypothesise the occurrence of something with the right sort of logical force (and call it an intention, a motive, a reason or a belief– desire pair) then we must, in doing so, presuppose the relation between it and the action that this mental occurrence was invoked to explain. Why? Because in order for motives, intentions and reasons to be explanatory, they must be motives, intentions or reasons for the action in question. But if motives, intentions and reasons are introduced in the first place to explain how a bodily movement (the arm’s rising) becomes an action (someone’s raising his arm) then the specification of the motive or intention cannot simply presuppose the action, on pain of circularity. A motive is a motive for some action either performed or considered; hence a motive, far from being a factor which when conjoined with any bodily movement thereby constitutes an action, actually presupposes the very concept of an action itself. (1961, 83) This is the kernel of Melden’s argument but it is not easy to understand and in any case is unlikely to move the contemporary philosopher of mind who learned in her first, introductory course on the subject that intentional states enter into both logical relations with other states (in virtue of their content) and causal relations with other states (in virtue of their form). To such a philosopher, Melden’s criticism amounts to the denial of a philosophical platitude. So more work is needed, it seems, in order to make the argument clearer. The gist of the argument I shall develop is this. Correctly to ascribe an intention, motive, or reason in such a way as to display its logico–grammatical relation to action is already to attempt an explanation of the action by putting it into a context that makes it understandable. To suppose that there are events that are designated by the reason- or motive-expression is not only unnecessary; it obscures the way reason-explanation functions. Davidson claims that it would be a mistake to conclude from the fact that placing the action in a larger pattern explains it, we now understand the sort of explanation involved, and that ‘cause and effect form the sort of pattern that explains the effect in the sense of “explain” that we understand as well as any’ (1980, 10). Davidson challenges the opponents of the causal view to identify what other pattern of explanation illustrates the relation between reason and action if they wish to sustain the claim that the pattern is not one of cause and effect. Let us try to meet this challenge. The

Julia Tanney 97

Wittgensteinian position starts out, I claim, by assuming that motives, intentions and reasons can be used successfully in explanations of actions and then it asks, when they are successful, how they explain. In many cases, attributions of motives, intentions and reasons explain a performance by characterising it as an action of a certain kind. This is already to distinguish an explanation in terms of motives and intentions from a causal-explanation, Melden tells us, since a causal-explanation suggested by the Humean picture usually takes it for granted that the event to be explained is already fully characterised as the kind of event it is; a causal-explanation offers us ‘an account of how it is that an event whose characteristics are already known is brought to pass’ (1961, 88). Of course it is consistent with this, as Melden himself immediately acknowledges, that the effect-event can be described in terms of its cause as an injury to the shoulder, say, might be described as a sunburn. But there must, on this view, be two logically independent (and therefore independently describable) events that enter into the causal relation. This is a condition that Davidson accepts;7 so, too, does Fodor: It is, of course, true that if X is the cause of Y, then there must be some description that is true of X and that is logically independent of the description ‘Y’s cause’, and there must be some description that is true of Y and that is logically independent of the description ‘X’s effect’. (1968, 35) Fodor adds, however, that this demand would be satisfied if the materially sufficient conditions for having a certain motive could be formulated in neurological terms; indeed, it would be satisfied by the existence of any state of affairs that is associated one to one with a psychological state by laws, empirical generalisations or even by accident. Thus, Fodor alleges, the appeal to Humean strictures is too weak for Melden’s purposes. But both Davidson and Fodor seem to interpret Melden’s claim that a cause must be ‘logically distinct from the alleged effect’ (1961, 52) as dictating the vocabulary that must be used to pick out the supposed mental event which – in order for it to count as a cause at all (it is agreed by everyone here concerned) – must have some logically independent description (whether we know what it is or not). But I read Melden, by contrast, as calling into question the idea that such a mental event or occurrence must exist. His argument, as I see it, is that the existence of such occurrences is not required for the concepts of intention, motive, reason and so on to discharge their explanatory role, thus throwing into question the whole idea that this explanatory role is causal. This, in any case, is the argument I shall develop. In order to bring to light some of the features of a contrasting, non-causal pattern of explanation, let us consider a simple case first – one removed from the context of reasons, intentions and motives. A chemistry student

98 Reasons as Non-causal, Context-placing Explanations

who had to leave the class early might find it puzzling why his teacher wrote CAT on the board. We can imagine his puzzlement relieved when his classmate explains, ‘Because she was writing “catalyst” – you left the room before she completed the word.’ This ‘because’ introduces an explanatory context, but it is not the sort of explanation in which one event (logically independent or not) follows another. Intuitively speaking, there is one event (the writing of ‘catalyst’) which has not been understood.8 The answer serves to re-characterise what happened so that it – as newly described – is no longer puzzling. The chemistry teacher’s writing ‘catalyst’ on the board is, I assume for the sake of the example, more understandable than her writing ‘c’, ‘a’ and ‘t’. This is not because we have now made out any mysterious connection between the occurrences of two contingently related events–the writing of ‘c’, ‘a’, and ‘t’, on the one hand and the writing of ‘catalyst’, on the other. For even if these were considered (implausibly) two distinct events, they would not be contingently related: writing the English word ‘catalyst’ entails writing the letters ‘c’, ‘a’ and ‘t’. Nor is there any reason to expect that we may find some other description of the performance of writing the letters that would qualify it as a logically independent event, in such a way that events of this newly described kind enter into a law-like connection with events typed as the writing of the word ‘catalyst’. The performance was puzzling only because it was conceived or described as the writing of the letters ‘c’, ‘a’ and ‘t’ or as the word ‘cat’ instead of as the writing of the word ‘catalyst’. The teacher’s writing ‘catalyst’ on the board is not puzzling, I assume, because it is part of a general pattern of behaviour that ‘belongs to’ or ‘is at home in’ a chemistry lesson. The ‘because’ in ‘She wrote the letters “c,” “a,” and “t” because she was writing “catalyst” ’, then, signals a different pattern of explanation from the causal pattern in which one event follows another. Here we have a clear-cut case of a non-causal, context-placing explanation.9 This case can be used as a model to develop an elucidation of the explanatory role of our concepts of intention, motive and reason. Melden’s famous example concerns a man who raises his arm. To the question, ‘Why did he raise his arm?’ the answer ‘In raising his arm, he intended to signal’ serves to re-characterise a performance first described as the driver’s raising his arm as an act of signalling. ‘Raising his arm’ and ‘signalling’ are different descriptions, each with different ‘implication threads’ (to borrow an expression from Ryle (1971)). Although there may be any number of (muscular, physiological, neuronal) events leading up to and forming part of a causal chain resulting in the arm’s rising there is no reason to characterise (or identify) any of these events as the motive, intention, will or reason to raise the arm. Such a characterisation in any case would not permit the redescription of the arm’s rising as either the driver’s raising his arm or as the act of signalling without adverting to the very background circumstances that I am here trying to show may be sufficient for such a redescription, and thus for a non-causal, context-placing explanation.

Julia Tanney 99

Wittgenstein thought that we tend to be misled into thinking that there must be such an event – even if hidden from view – because we are focussing on one way that language functions to the exclusion of others. The concepts in question do indeed allow us to speak, for instance, of a person who comes to a decision, forms an intention, or admits that such-and-such reasons for acting are overriding. These uses encourage the thought that having an intention or reasons results from having formed an intention or from having considered and accepted the reasons and these in turn are construed, reasonably enough, as mental events. Now it is true that a full elucidation of the concept of intention and of reason and its cognates would have to include these uses.10 The formation of an intention or the consideration and acceptance of reasons might also figure in a causal-explanation of an action.11 But it would be a mistake to form a general picture of the nature of intention or reason from this use alone, and require that every time we ascribe these concepts, there must have been a moment when the intention was formed or the reasons considered. Davidson argues for the former in his account of intention (1978);12 those working in the spirit of Davidson today have argued for the latter.13 We have seen that a rather different way of understanding the explanatory power of the response ‘He intended to signal’ is by the placement of the performance first described as the man’s raising his arm in the wider circumstances of his driving a car and being about to make a turn. The response will succeed in explaining the man’s raising his arm, however, only to the extent that a description that puts it into this context is more understandable than a description that leaves this context out. If the one who is puzzled does not understand our driving practices or why anyone who is driving and approaching a turn should signal, then this re-characterisation of the action will not satisfy her. Now Davidson acknowledges that a logico–grammatical relation is in place between the contents of the relevant attitudes (the belief and pro-attitude which, for him, constitute the agent’s reason) and the action-type that it recommends. He also holds – what I am here calling into question – that attitude or reason-ascriptions function by designating events (or standing states and triggering events). His (positive) argument for construing this relation as causal is that the logico–grammatical relation exhibited in the content-description is an ‘anaemic’ justificatory one: insufficient for accommodating the case in which the agent has a reason for acting in a certain way, acts in that way, but not because of that reason. For example, even though the driver had reason to signal – he was approaching his turn – he may have raised his arm for another reason – say, to wave to his friend. If so, then the re-characterisation of his action as a case of signalling will fail to explain the action. Davidson (1963, 11) presumably had this sort of case in mind as a counterexample to Melden.14 But the fact that a context may be imagined in which the redescription fails to explain the action presents no threat to the argument. For it is no

100 Reasons as Non-causal, Context-placing Explanations

part of Melden’s job to insist that every context-placing redescription will succeed. Melden need only argue that when such an explanation does succeed (because it enables the one who is puzzled to see the action in a new, sense-making light) it may be the kind of context-placing explanation just described; one that does not depend upon or cannot be understood as requiring the existence of mental events – let alone (in principle describable) logically independent ones – that are alleged to constitute the reason or intention. And, although it is true that the redescription would not succeed unless the requisite motives, intentions, beliefs and desires could also be ascribed, there is no obligation to construe the deployment of these related concepts as the identification of events or standing states; let alone (in principle describable) logically independent ones. When context-placing explanations such as these are unsuccessful we may need to probe further for a different or more far-reaching context-placing explanation that will succeed or possibly give up the initial expectation that the action can be explained by reasons – but not assume that having reasons, intentions and motives must be a matter of the instantiation of properties which may be in some sense hidden (e.g. tokenings of conscious or non-conscious mental events or of their alleged realisers).15 The problem in assuming that the motive, intention or reason is (in principle describable as) a logically independent, temporally antecedent, causally efficacious event (perhaps identified with its alleged ‘onset’) is that it mis-assigns the explanatory function of these concepts. The position commits us to postulating an event, unobservable to others and possibly even to the agent herself that would, if known, provide the sought-after reason-explanation for the agent’s action. In such cases, as Ryle insisted, an epistemological puzzle arises as to how anyone could ever know whether a person acts for reasons or what, if she does, her reasons are, since the hypothesis is not even in principle testable. Not only do we not, in everyday situations, have access to these hidden events, but even if we were, say, to monitor the neural activity of someone’s brain or access their stream of consciousness, we would never be able to set up the kinds of correlations that would establish a particular occurrence as an instance of a particular reason without already having a way of deciding whether someone acted for a particular reason in order to make the correlation. The foregoing considerations suggest that mental concepts such as intention, reason and motive operate very differently from causal concepts – say, that of a gene. We might say that Morgan’s concept of a gene was the concept of something whose nature was to be discovered, responsible for the transmission of heritable characteristics. The DNA molecule, it was later found, plays that role. But the argument of this essay is that the concepts of intention, reason and so on are not like this, for they discharge their explanatory role without designating anything; let alone causally efficacious states or events; let alone causally efficacious states or events whose nature awaits discovery.16

Julia Tanney 101

III This, I think, is the correct way to understand the arguments of Wittgenstein, Melden, Anscombe and Ryle. But it is difficult to know how to place this view within the contemporary debate in the philosophy of mind because it so rarely makes an appearance in today’s discussions. A student of philosophy of mind today might ask, for example, if this makes the view about intentions realist or irrealist, instrumentalist or behaviourist. In order to facilitate my aim of re-introducing Wittgensteinian territory into the contemporary landscape, it will be worthwhile taking a brief look at the temptation to plot this position with a particular metaphysical compass and suggest a reason why this temptation should be resisted. We were introduced to instrumentalism in the early work of Daniel Dennett. In ‘Intentional Systems’ (Dennett 1978) he describes the intentional stance by considering a chess-playing computer. In taking the intentional stance towards this machine, one is instrumentalist about propositional attitudes insofar as ‘we find it convenient, explanatory, [and] pragmatically necessary for prediction, to treat it as if it had beliefs and desires and was rational’.17 A machine for playing chess, however, is not like a man or animal: ‘its “rationality” is pinched and artificial’ (1978, 8). This was Dennett’s position in the 1970s. A decade or so later, his position seemed to change: [A]ll there is to being a true believer is being a system whose behaviour is reliably predictable via the intentional strategy, and hence all there is to really and truly believing that p (for any proposition p) is being an intentional system for which p occurs as a belief in the best (most predictive) interpretation. (1987, 29) These two characterisations are, on the face of it, inconsistent. On the first, there really is something to being a believer over and above being predictable by the intentional stance and on the second there is nothing to being a believer than being thus predictable. On the first characterisation there seems to be an implicit acceptance that mental terms like ‘believes’ pick out underlying, possibly causally efficacious states or events. According to this picture, there is a difference between low-grade computers and people: people really have the underlying states (etc.) to which mental terms purport to refer. Instrumentalism on this construal is like an ‘error theory’ in Mackie’s sense. Just as, for Mackie, moral terms purport to pick out moral facts, but do not really (and nonetheless serve their jobs) so do mental terms purport to pick out inner states, but sometimes do not really (and nonetheless serve their jobs). On the second characterisation, however, there may be no implied commitment to such underlying states in our use of mental expressions (on the contrary – this seems to be denied by the locution ‘all

102

Reasons as Non-causal, Context-placing Explanations

there is to being a believer...’). If there is no such commitment, this view would not be instrumentalist, since there would be nothing real to contrast with what is supposed to be instrumental. Is the second position a form of realism then? According to Devitt and Sterelny (1999, 293) Dennett’s later view is a form of ‘philosophical behaviourism’. Although usually construed as a type of fictionalism or instrumentalism about the mental, philosophical behaviourism is understood by them, in the context of discussing Dennett, as the realist doctrine that mental terms refer to (real) patterns of behaviour. But however appropriate or not this might be as a description of Dennett’s later position, it would not be a fair description of the view I have characterised as Wittgensteinian. On this latter view, to ascribe an intention, motive or reason for some particular action may not involve an attempt to refer to anything; the concepts may function rather to explain an action by placing it in a context that renders it less puzzling.18 Nor, for similar reasons, would it be fair to characterise Ryle’s view in The Concept of Mind as a form of philosophical behaviourism even though he is widely (and misleadingly) credited with introducing us to this doctrine.19 Ryle’s dispositions play the same role as the ‘sense-making pattern’ or the ‘wider circumstances’ play in the view I have just described. Ryle agrees with his interlocutor that when we use mental concepts to describe a performance, we are not merely taking into account ‘muscular behaviour’, because the same muscular behaviour in other circumstances could not be so described. A remark by a parrot, for example, could not be described as intelligent or witty. But it does not follow that in order to be credited with wit or intelligence the muscular behaviour must be accompanied by some mental act. In judging that a particular performance is intelligent, it is true that we look beyond the performance itself; but not into some ‘hidden counterpart’ performance, taking place behind the scenes. We are considering, according to Ryle, the abilities and propensities of which this particular performance was an actualisation. ‘Our inquiry is not into causes (and a fortiori not into occult causes), but into capacities, skills, habits, liabilities and bents’ (1949, 45). For Ryle, many of our ‘mental-conduct’ verbs are correctly applied to a performance or an action because it is the actualisation of a disposition. But talk of dispositions complicates the matter. According to Ryle, the particular mental conduct terms whose logical geography he was attempting to map discharge their explanatory role by helping to situate the agent and her actions within a pattern that can be articulated by an indefinitely long series of hypothetical (and mongrel-categorical) statements about what she could be expected to do, think, feel and so on, given her background (e.g. training) and the present circumstances. This is what he means in this context by noting that mental-conduct verbs are applied to performances that are actualisations of a disposition.

Julia Tanney 103

But the introduction of dispositions will, for others, take us right back to the realm of hidden, underlying causes.20 Quine (1977), for example, was dissatisfied with dispositional statements because they, like general causal statements, depend upon an intuitive or unanalysed notion of similarity or kind. Dispositional statements, best understood as subjunctive conditionals, are not amenable to paraphrase in the canonical (extensional) language in which Quine held that all serious scientific statements could be expressed.21 Quine suggested that when the disposition is of theoretical interest, then a mature science can dispose of this intuitive similarity notion by finding the underlying structure that will tell a more straightforward story in a way that renders the dispositional one obsolete. ‘Sometime, whether in terms of proteins or colloids or nerve nets or overt behaviour, the relevant branch of science may reach the stage where a similarity notion can be constructed capable of making even the notion of intelligence respectable. And superfluous’ (1977, 174). Attraction to some of the aspects of Quine’s programme may be among the reasons scientific realists look deeper than observable patterns of dispositions and search for a common underlying structure in the kinds of things that manifest those patterns. In certain cases, this may be essential: if the dispositional concept is a theoretical concept it arguably needs the discovery of a ‘realiser’ to vindicate it (as the discovery of the DNA molecule presumably was needed to vindicate the concept of gene). But it is precisely this way of construing mental concepts that is under dispute. When Ryle spoke of the ‘higher-grade’ dispositions of people as ‘multi-track dispositions the exercise of which are indefinitely heterogeneous’, and when he used the example of Jane Austen’s representation of pride ‘in the actions, words, thoughts and feelings of her heroine in a thousand different situations’ (1949, 44), he was reminding us of the patterns of conduct with which we are already familiar: indeed, he insisted that ‘the concepts of learning, practice, trying, heeding, pretending, wanting, pondering, arguing, shirking, watching, seeing and being perturbed are not technical concepts’ (1949, 319). It was not part of his project (any more than Jane Austen’s) to speculate about the underlying structure of the ‘systems’ or of people who exercised these dispositions. Nor did he think that such a scientific project could vindicate – let alone in principle replace – the everyday attributions effected by ordinary mental terms.

IV One used to hear the complaint made by philosophers impressed by Wittgenstein’s teaching that those who tried to treat mental terms as theoretical posits were guilty of changing the subject. This charge was rarely elaborated, and so it was dismissed on the grounds that it struck the opposition as sheer philistinism or brute prejudice against science. Now, this

104

Reasons as Non-causal, Context-placing Explanations

charge would sound philistine to those who have already accepted that psychology provides the science behind the ordinary use of mental concepts (together, perhaps, with an attraction to aspects of Quine’s programme). 22 But precisely these assumptions are denied by Wittgenstein and his followers by their insistence on distinguishing between different kinds of explanation. According to them, to ignore what is sense-making and observable in preference to what is underlying and whose nature awaits discovery is to misinterpret or ignore the role mental concepts play in our interpretive practices: to focus on underlying structures would force a change of subject by ignoring the way mental concepts normally discharge their role. The injunction, for example, to accept the observational-dispositional nature of mental terms for everyday use but to insist that their explanatory role depends on how well they interpret states within the system’s underlying structure does not make sense on the view I am recommending. This is because the explanatory function of reason concepts may be fully discharged by the placement of the action to be explained within the appropriate circumstances or wider context, or, as Anscombe suggests, so that a certain sense of the question ‘Why?’ is given application. For philosophers such as Melden, Anscombe and Ryle what it is to describe an action as one performed for such-and-such reasons or with such-and-such intention may simply involve an attempt to redescribe what in the context was puzzling with what in the new context is no longer puzzling. Insofar as the causal hypothesis forces us to construe the reason- or intention-ascription as functioning to designate an event, property, state, fact or condition of a person the mysterious nature of which is open to investigation, it mis-assigns the concepts’ explanatory role.23 In any case, the complaint of prejudice can be turned around against those who suppose that the role of truly explanatory concepts must be capturable in a language suitable for the aims of mathematics and logic. Quine, for his part, suggests that ‘amiable’ and ‘reprehensible’ are disposition terms that should draw on intuitive kinds. Why not suppose with the Wittgensteinians that mental conduct terms are like those?24 I have argued that it may be the familiarity, unsurprisingness or sensemaking aspect of the context, pattern or circumstances that perform the function of explaining the action and it is this pattern that is illuminated when the content of the reason, motive or intention is ascribed. This ‘sense-making’ criterion is closely related to, but sometimes conflicts with, another criterion we use for ascribing intentions, motives and reasons: namely, that which the subject herself says or would say about her reasons (etc.) for acting. 25 A fuller treatment of this will have to wait another time; a few brief (and no doubt provocative words) will have to do. Suppose we have recourse to ask Melden’s driver what he was doing: was he signalling or waving to a friend? Sometimes his answer will satisfy us; sometimes it

Julia Tanney 105

will not. But in answering us, he, too, is attempting to place his action within a sense-making context, for according to the view I am attempting to reintroduce, he will have mastered the relevant concepts – acquired the various skills – in the same way as everyone else. What if his answer does not make sense of his action? We have a choice. We may accept his answer because in asking for his reasons for acting, we may be seeking (on this occasion) his own conception of what he is or was doing, whether or not it satisfies us. When our way of making sense of him conflicts with his understanding of himself, or with his memory reports of what he was thinking at the time, we may reject his answer as inadequate. That is to say, this second (‘self-conception’) way of understanding the expression ‘his reason for acting’ may be set aside in favour of the other (‘sensemaking’) one. After all, what a person has to say is not always authoritative: as Anscombe reminds us, there are controls on someone’s proffered reasons, motives or intentions.26 Nonetheless, there may come a point when these controls have been exhausted: there is nothing occurring either before or after the action to check on an agent’s truthfulness, sincerity or lack of self-deception in declaring her reasons for acting. Where there is nothing in the circumstances either before or after the action to enable us to pinpoint her reason any better than she has been able to do, then perhaps we reach a point, says Anscombe (‘after much dispute and fine diagnosis of her genuineness’), when only the agent can say why she acts as she does. But this is not, as the traditional Cartesian would have it, because she has access to an intention, motive, or reason – now conceived as something interior – that is forever hidden from anyone else, but rather because no one else has any grounds for correcting her. As I shall put it on behalf of Wittgenstein, what the agent says is, in such (unusual) circumstances, the only criterion available – the only sense that can be given – to ‘her reasons for acting’. The arguments of this essay are intended to support the idea that the explanatory role of mental concepts is different from that supposed in contemporary philosophy of mind. The crux of the debate centres not only on whether mental concepts can be assimilated to theoretical terms: I have suggested that a full treatment of the Wittgensteinian position would involve denying that the predicative expressions in which mental concepts figure must involve the very notion of a reference or an extension that is at the root of the Carnap-Quine-Davidson programme. The arguments I put forward here call into question the idea that mental terms purport to function in the general case as referring expressions: that is, that their primary use or the way by which they discharge their explanatory role is to designate or name an event, state, object, property or relation.27 If I am right, then it would seem that a natural way of conceiving the dispute between mental realists and irrealists is based upon a category mistake from the outset.

106

Reasons as Non-causal, Context-placing Explanations

Notes An ancestor of this essay was presented in January 2004 to the Institut d’Histoire et de Philosophie des Sciences et des Techniques (IHPST) in Paris for a workshop on commonsense psychology. Thanks to Sandra Laugier, Daniel Andler, Pierre Henri Castel, Ruwen Ogien and Jean-Jacques Rosat for helpful discussion and to John Flower, Edward Harcourt and Richard Norman for their comments. A recent version was presented in 2007 as a Royal Institute of Philosophy lecture at the University of Keele. Thanks to Sorin Baiasu, Geraldine Coggins, Giuseppina d’Oro and James Tartaglia for their penetrating questions which have helped improve the text and to Constantine Sandis for suggestions that have helped me clarify the argument. 1. This is the root idea of functionalism. Originally, functionalism was proposed as a theory about the meaning of mental terms that are used everyday in non-theoretical discourse (Lewis 1972, 249–58). Our common sense or everyday concept of pain, on this account, is thought to pick out an (unspecified) inner state of an organism or system that occupies a certain causal role in mediating between other inner mental states, input and behavioural output. This causal role is to be specified by common sense platitudes: for example, that pain is likely to be caused by tissue damage and result in avoidance behaviour. Another version suggests that the role is rather to be found by the traditional methods of a priori philosophical analysis (Shoemaker 1984). This approach would, like the first, dovetail with the idea that mental concepts’ primary domain is common sense explanation but only if the methods of a priori philosophy are to make explicit what is already implicit in our (common) use of mental concepts. Psychofunctionalists, however, attempt to break the tie with ordinary concepts. They agree with other functionalists that ordinary mental concepts function to pick out a causal role realised by some or other inner (physical) state. They also agree that a complete, constitutive account of the second-order or functional states will be given by a story outlining the causal relations between the occupants of such states and their relation to input, output and other mental states. But they believe, not only that the occupants of this role, but the role itself are to be discovered by empirical psychology (Block 1978/1980, 268–305; Rey 1997, 187). 2. The expression is Fodor’s in (1968, xix). 3. Hornsby (2004), McDowell (1998) and Rudder-Baker (1995) are some examples. 4. Compare: ‘Words like “explanation,” “law,” “rule,” “principle,” “why,” “because,” “cause,” “reason,” “govern,” “necessitate,” etc. have a range of typically different senses. Mechanism seemed to be a menace [to the possibility of free will] because it was assumed that the use of these terms in mechanical theories is their sole use; that all “why” questions are answerable in terms of laws of motion’ (Ryle 1949, 76). Substitute ‘physics’ and ‘physical’ for ‘mechanism’ and ‘mechanical’ and ‘physical laws’ for ‘laws of motion’ and you have a nice statement of what we might call, with a nod to Ryle, the ‘bogey of physicalism’. 5. I think Anscombe (1983) is right in saying that the concept of causation is as general as the concept of a factor, so that it is misleading to talk about ‘the’ causal relation. Thus, like Melden, I am not particularly interested in defending Hume’s account, which has been disputed from every angle. It has, for example, been doubted whether causation is best seen as a relation between events (Hume said between two ‘objects’); it has been questioned whether necessity is involved, whether a singular causal statement implies a pattern or regularity of any kind, and so forth. But it is roughly this use of ‘cause’ which Davidson had in mind

Julia Tanney 107

6.

7. 8.

9.

10. 11.

when he argued that reasons are causes and is, for example, accepted by Fodor. (See the text below for more discussion.) And it is this use which tempts one to construe verbs such as ‘believes’, ‘thinks’, ‘wants’ as picking out mental events or occurrences conceived as hypothetical or theoretical entities. This latter picture is my quarry. This formulation allows us to concede that the concept of intention, for example, may be correctly applied, on some occasion, to designate an event that can legitimately be classed as ‘mental’: a state of a person or an event in her history that can form part of an intuitively plausible causal chain issuing in action (see text below and examples in note 11); it denies, pace Davidson, that the concept of intention’s explanatory value depends on the existence of any such causal relation. Note that these are reasons why the thesis I am defending here is not aptly described as denying either mental causation or the ‘reality’ of the mental: though of course it does deny aspects of mental realism as this is commonly understood. It should be noted, further, that on this view I am recommending the distinction between reasons and causes is not always a firm one. Following Anscombe we can agree that my hanging up my hat because my host said ‘Hang up your hat’ is one in which the intuitive distinction starts to vanish: it depends on the circumstances whether we would call this a cause or a reason (1957, §15). See (Davidson 1980, 12) where he suggests a number of candidates for such an event. The student may have construed it as the writing of the word ‘cat’ or as the writing of the sequence of letters ‘c’, ‘a’ and ‘t’. The performance may have been puzzling on either construal. If part of Davidson’s challenge here is to say how writing ‘catalyst’ on the board ‘belongs to’ or ‘is at home in’ a chemistry class, then it must be admitted that not much more can be said (except one that issues reminders about the kinds of things one studies in chemistry); at least there is no answer that can be given in more fundamental terms. One of the convictions of the position I am defending is that the ability to see actions as fitting into familiar patterns comes about through training and through a shared form of life and not in general through explicit instruction or through prior theoretical (rule-following) operations. Anscombe’s account of an intentional action as one for which a certain sense of the question ‘Why?’ is given application is a sophisticated attempt to describe in more detail what ‘belonging to’ or ‘being at home in’ involves in relating the action to the agent’s motives and reasons but it, too, appeals to our (considered) judgements about what makes sense without attempting to explain this. For a discussion of the process of acculturation that enables us to see actions in new ways, see the final chapters of Melden (1961). For an argument against explaining this ability in terms of prior theoretical operations, see Tanney (2000). See Anscombe (1957) for such an elucidation; see Tanney (2002) for my own attempt. Anscombe (1983) gives an example of such a story. She imagines a case in which she has a long-standing resolution never to grant interviews with members of the media. When someone asks why she refused to see the representative of Time magazine, he is told of this long-standing resolution, which ‘makes her reject such approaches without thinking about the particular case’. This explanation, involving the expression ‘makes her ...’ is causal, says Anscombe, in the sense

108 Reasons as Non-causal, Context-placing Explanations that it derives the action from a previous state. Or, to borrow an example of Rogers Albritton, my recognition of someone’s character, for example, might cause me to break off relations with him. But, warns Anscombe, [i]t is one thing to say that a distinct and identifiable state of a human being, namely his having a certain intention, may cause various things to happen, even including the doing of what the intention was an intention to do; and quite another to say that for an action to be done in fulfilment of a certain intention (which existed before the action) is eo ipso for it to be caused by that prior intention. In other words, an event (say) in an agent’s history that can legitimately be classed as mental (e.g. his having made the decision, in the light of various factors, that he must do such-and-such) may feed into a causal story of a subsequent action that is performed in execution of that intention. But it should not be inferred from this that what it is to act in execution of that intention is a matter of there being some causal relation between this event and the action. One unhappy consequence of making this inference would be the supposition that there must have been such an event – possibly hidden or non-conscious – even when there is no obvious, conscious candidate. The following closely related idea may be useful in helping to understand this point. It may be that in some particular performance that counts as following a rule, a person consults an expression of that rule and then acts as it mandates. But it should not be inferred from this that what it is to follow a rule is a matter of there being some (overt or hidden) consultation of a rule. I discuss this further in (2000) and (2008a). 12. In Davidson’s Actions, Reasons, and Causes, acting for a reason consists merely in having a belief and pro-attitude with the right sort of content, which cause the action. Davidson responds to Melden’s challenge to find the mental event that constitutes the reason (understood thus) by declaring: Of course there was a mental event; at some point the driver noticed (or thought he noticed) his turn coming up, and that is the moment he signalled. ... To dignify a driver’s awareness that his turn has come by calling it an experience, or even a feeling, is no doubt exaggerated, but whether it deserves a name or not, it had better be the reason why he raises his arm. (1963, 12–13) Thus the idea that the relevant concept’s function is to designate some kind of mental event is in place early in Davidson’s work. Incidentally, the driver’s noticing his turn coming up would (at best) be a reason why he chose that moment to signal. 13. Consider David Velleman: In order to have acted autonomously, the agent would need to have been actuated not only by the desire and belief mentioned in the story but also by the story itself, serving as his grasp of what he was doing – or, in other words, as his rationale. He would need, first, to have been inhibited from acting on his desire and belief until he knew what he was up to; and then guided to act on them once he had adopted this story. He would then have acted autonomously because he would have acted for a reason having been actuated in part by a rationale. (2000, 28) Michael Bratman’s (2001) claim that (full-blown) actions are caused by higher order reflexive policies is similar to Velleman’s. 14. I have suggested elsewhere (Tanney 1995) that what Davidson is really after is ‘causal cement’ for what he takes to be a logical gap between reasons and actions. (I have also suggested that his adherence to a Hempelian nomological-deductive

Julia Tanney 109

15. 16. 17.

18.

19. 20. 21.

22. 23.

24.

25.

model of explanations looms large in the background here.) On the view I am recommending this logical gap must not be closed. Reason does not determine, or provide a sufficient condition for, the action that it explains. Nor does a performance guarantee that a rationalising, reason-explanation is on offer. The relation between reason and action is more the relation between warrants and moves than the ‘determinate connection’ suggested in the causal, nomologicaldeductive account. See Tanney (1995, 2005a) for extended arguments for these claims. See Fodor (1968) and Putnam (1968) for early attempts to argue that mental concept should be construed by analogy with the concept of a gene. The discussion up until now has concerned intentions, reasons and motives whereas this paragraph introduces beliefs (and propositional attitudes in general). This is not the place to defend or elucidate the idea that the relation between these concepts and the concepts of reason, intention and action is a (logico-) grammatical one, but see Tanney (2005a) for one such discussion and the last chapters of Melden (1961) for a different discussion of how the concepts of agency, want and belief are thus connected. Note that to say that the concepts’ function is to explain in the way described above is very different from saying that the concepts discharge their role by designating a state, relation or process which is to be identified by its functional/ causal-explanatory role. I argue for this in more detail in Tanney (2005b, 2007). See Mumford (1998) who argues for a particular (functionalist) version of realism about dispositions. Quine’s claim that a serious scientific theory must be expressed in an extensional language bodes ill not only for unreconstructed dispositional statements but also, notoriously, for content and meaning in general. See, for example, Rey (1997). Compare: How does the philosophical problem about mental processes and states and about behaviourism arise? – The first step is the one that altogether escapes notice. We talk of processes and states and leave their nature undecided. Sometime perhaps we shall know more about them – we think. But that is just what commits us to a particular way of looking at the matter. For we have a definite concept of what it means to learn to know a process better. (The decisive movement in the conjuring trick has been made, and it was the very one that we thought quite innocent.) – And now the analogy which was to make us understand our thoughts falls to pieces. So we have to deny the yet uncomprehended process in the yet unexplored medium. And now it looks as if we had denied mental processes. And naturally we don’t want to deny them. (Wittgenstein 1953, §308) The Quinean motivation for reducing dispositions to underlying structures is dubious in any case. Even if they could be accommodated by logical and mathematical methods, these methods themselves would have to rely on intuitive notions of similarity: this is, after all, the lesson of Wittgenstein’s discussion of rules. See Wittgenstein (1953, §§143–55, §185ff). For an admirable discussion of this point, see Dilman (1978/1979, 35–58). This may involve either a pronouncement on her motives, intentions or reasons construed as her own way of making sense of her action, or as a (memory) report of events that may be accompanying or have preceded the action (e.g. a sudden decision or realization), or both.

110

Reasons as Non-causal, Context-placing Explanations

26. Anscombe (1957, §25). 27. See Ryle (1971) for his explorations on this theme. A number of moves made in metaphysics, epistemology and in philosophy of language and mind in the 1960s and 1970s – ones that are presupposed in most of the work in these fields today – would be thrown into question as well if this idea is correct. If they are not referring expressions, then the construal of them by scientific realists on the analogy with natural kind of terms such as ‘gold’ or ‘tiger’ – whose essence is a matter for science to discover – is a non-starter. This is, for example, the treatment that (early) Putnam (1968, 1–19) suggests for the concept of pain. The diagnosis would also cast doubt on Armstrong’s (1980, 16–17) characterisation of conceptual elucidation and ontology as the investigation of second- and firstorder questions, respectively. So too, of course, would it help to define the real trouble with functionalism. The functionalists (and conceptual-role theorists) were right to focus on the importance of function, role or use. But the explanatory role is played, I have argued, by the way the concept is wielded in redescribing the behaviour of a system. It is not played by referring (however obliquely) to the system’s internal (first- or second-order) states. And finally the view that ‘criteriological’ investigations modelled after Wittgenstein’s are about justification and therefore about epistemology and not metaphysics would founder as well. See the Preface to Shoemaker (1984) for a biographical account of how his acceptance of this distinction led him to abandon criteriological investigations for causal accounts.

References Anscombe, G.E.M. 1957. Intention, Basil Blackwell, Oxford. Anscombe, G.E.M. –1983. ‘The Causation of Action’, in Knowledge and Mind: Philosophical Essays, C. Ginet and S. Shoemaker (eds), Oxford University Press, New York, 174–90. Armstrong, D. 1980. The Nature of Mind, University of Queensland Press, Queensland. Block, N. 1978. ‘Troubles with Functionalism’, reprinted in Block, 1980. Readings in Philosophy of Psychology, vol. 1, Harvard University Press, Cambridge, MA. Block, N. –1980. Readings in Philosophy of Psychology, vol. 1, Harvard University Press, Cambridge, MA. Bratman, M. 2001. ‘Two Problems about Human Agency’, in Proceedings of the Aristotelian Society, vol. 101, issue 3, 309–26. Davidson, D. 1963. ‘Actions, Reasons, and Causes’, originally published in Journal of Philosophy 60 and republished as essay 1 in Essays on Actions and Events, Clarendon Press, Oxford, 3–19. Davidson, D. –1976. ‘Hempel on Explaining Action’, originally published in Erkenntnis 10, 239–53 and republished as essay 14, in Essays on Actions and Events, Clarendon Press, Oxford, 261–75. Davidson, D. 1978. ‘Intending’, originally published in Philosophy of History and Action, Yirmiaku Yovel, ed. (D. Reidel and The Magnes Press, The Hebrew University). Reprinted as essay 5, in Essays on Actions and Events, Clarendon Press, Oxford, 83–102. Davidson, D. 1980. Essays on Actions and Events, Clarendon Press, Oxford. Dennett, D. 1978. Brainstorms – Philosophical Essays on Mind and Psychology, Bradford Books, Montgomery VT. Dennett, D. 1987. The Intentional Stance, A Bradford Book, MIT Press, Cambridge, MA.

Julia Tanney 111 Devitt, M. and Sterelny, K. 1999. Language and Reality – An Introduction to Philosophy of Language (2nd edn), Blackwells, Oxford. Dilman, I. 1978/79. ‘Universals: Bambrough on Wittgenstein’, Proceedings of the Aristotelian Society, vol. LXXIX. Fodor, J.A. 1968. Psychological Explanation: An Introduction to the Philosophy of Psychology, Random House, New York. Hornsby, J. 2004. ‘Agency and Actions’, in Agency and Action, John Hyman and Helen Steward, eds, Royal Institute of Philosophy Supplement, 55, Cambridge University Press. Lewis, D. 1972. ‘Psychophysical and Theoretical Identifications’, Australasian Journal of Philosophy, vol. 50, no.3. McDowell, J. 1998. ‘Might there be external reasons?’, in Mind, Value, and Reality, Harvard University Press. First published in J. E. J. Altham and Ross Harrison, ed., 1995. World, Mind, and Ethics: Essays on the Ethical Philosophy of Bernard Williams (Cambridge: Cambridge University Press), 387–98. Melden, A. I. 1961. Free Action, Routledge & Kegan Paul, London. Mumford, S. 1998. Dispositions, Oxford, Oxford University Press. Putnam, H. 1968. ‘Brains and Behaviour’, in R.J. Butler, ed. Analytical Philosophy, vol. 11, Basil Blackwell, Oxford. Quine, W.V.O. 1977. Originally published as ‘Natural Kinds’, in Essays in Honour of Carl G. Hempel, ed. N. Rescher, et al. The page reference is to the republication in Naming, Necessity, and Natural Kinds, ed. by Stephen P. Schwartz, Cornell University Press, Ithaca. Rey, G. 1997. Contemporary Philosophy of Mind – A Contentiously Classical Approach, Blackwell, Cambridge, MA and Oxford, UK. Rudder-Baker, L. 1995. Explaining Attitudes – A Practical Approach to the Mind, Cambridge University Press, Cambridge. Ryle, G. 1949. The Concept of Mind, Hutchinson, London. The references are to the republication by Penguin, London, 2000. Ryle, G. 1971. Collected Papers, vol. 1, Hutchinson, London. Shoemaker, S. 1984. Identity, Cause, and Mind, Cambridge University Press, Cambridge. Tanney, J. 1995. ‘Why Reasons May Not be Causes’, Mind & Language, vol. 10, nos. 1,2, 103–26. Tanney, J. 2000. ‘Playing the Rule-Following Game’, Philosophy vol. 75, no. 292, 203–24. Tanney, J. 2002. ‘Self-Knowledge, Normativity, and Construction’, Logic, Thought and Language, (Royal Institute of Philosophy Supplement: 51), Cambridge University Press, 37–55. Tanney, J. 2005a. ‘Reason-explanation and the Contents of the Mind’, Ratio, vol. XVII, no. 3, 338–51. Tanney, J. 2005b. “Une Cartographie des Concepts Mentaux”, Critical Introduction to Gilbert Ryle’s La Notion d’Esprit (The Concept of Mind), Payot, Paris, 7–70. Tanney, J. 2007. ‘Gilbert Ryle’, Stanford Encyclopedia of Philosophy, http://plato.stanford.edu/entries/ryle/ Tanney, J. 2008a. ‘Real Rules’, Synthese, xxxx. Velleman, D. 2000. The Possibility of Practical Reason, Oxford University Press, Oxford. Wittgenstein, L. 1953. Philosophical Investigations, translated by G.E.M. Anscombe, Blackwells, Oxford.

7 Interpretative Explanations G. F. Schueler

Commonsense explanations of actions, in terms of the agent’s reasons, hopes, desires and the like, are on their face frequently teleological in form. They specify the goals, purposes or points of the things we do. In this they seem sharply different from other sorts of commonsense explanations of events, as well as from the sorts of explanations found in sciences such as physics and chemistry, all of which are causal, and of course not teleological. But actions are often simply constituted by events involving the agent of the action. And these events are obviously open to causal explanation as long as we describe them in terms of their physical or chemical makeup. So there is a puzzle here. How can commonsense explanations of actions, which are apparently teleological and hence not causal in form, actually explain these actions? In this essay I will argue that what I will call ‘interpretative explanations’ are both central to explanations of human action and irreducibly different in form from other commonsense explanations of events, as well as from explanations found in paradigm ‘hard’ sciences such as physics. If this is right it turns out that, as a consequence of this different form, it is a mistake to think that interpretative explanations are somehow reducible to (or explicable in terms of) causal explanations. What I mean by an ‘interpretative explanation’ will be brought out in the course of the discussion. But we can start with an example.

1 We sometimes misinterpret what others are doing. Many years ago Andy Griffith did a comic routine where he described something he had witnessed on a college campus. Two groups of students, each dressed in colourful costumes, were performing some sort of ritual in a cow pasture. Each group would have a short meeting to discuss and vote on some topic, and then the ones selected to present the conclusions of the group would line up facing the other group. After a brief moment of silence, one person on each side 112

G. F. Schueler

113

would yell out its opinion and then a fight would break out which had to be broken up by people in striped shirts. Then the whole thing was repeated. The title of Griffith’s piece was ‘What It Was, Was Football’.1 Griffith was just being funny, of course, but the possibility of misunderstanding in this way is a real one. Finding out that ‘what it was, was football’ would explain the events on the field to a foreigner who really was unaware that this was what was being witnessed, in a perfectly ordinary sense of ‘explain’. It is that sense of this term that I will say involves giving an interpretation and that this essay will explore. In his routine, Griffith describes the actions of the players as if he does not know that they are playing football, but he knows they are doing something. That allows him to pretend to understand the players as performing intentional actions, just not the ones characteristic of football. He pretends to misinterpret what they are doing. But we can imagine an observer who does even worse than that. Suppose that the observer is unaware not just that it is a game that is being played but even that the organisms she is watching are performing any intentional actions at all. She is, let’s suppose, an alien from outer space (flown in especially to work in philosophical examples) who sees the events on the field simply as very complex interactions of some of the local fauna.2 Of course these events really are complex interactions of some of the local fauna. So this will not prevent her from describing with complete accuracy, and to any level of detail her observational powers allow, everything that happens on the field. It is just that she won’t describe them as intentional actions. She won’t interpret what she sees in this way. This suggests that there are at least two rather different kinds of mistakes one could make here. In the case satirized by Griffith the observer sees that he is observing people who are performing intentional actions. He simply fails to realize what actions they are performing. But one might also make the more serious mistake of not realizing that intentional actions were being performed at all. This would be to understand the behaviour being observed in the way we often look on the behaviour of lower animals, insects for instance: complicated behaviour produced by complex brain responses to the environment but not intentional actions. If that were the only correct way to look at behaviour, as some philosophers have held, it would follow that the mistake satirized by Griffith would not be any more mistaken than any other interpretation. If absolutely no intentional characterizations correctly apply to anything, then those students on the field are no more playing football than they are having brief discussions and then fighting with each other. On such a view both characterizations of what is going on are equally mistaken. Rather than pursuing this issue now,3 however, I will start by assuming the reality of the mistake satirized by Griffith, where the form of the mistake seems to be that the observer misinterprets the actions she is observing while realizing that they are indeed intentional actions.

114

Interpretative Explanations

So what would have gone wrong if an observer, seeing what is in fact a football game, takes it as some sort of ritualized debate followed by fisticuffs, in the way Griffith pretended to? Some of the errors Griffith pretended to make can just be set aside. We need to distinguish errors of interpretation from those based on mistakes about the underlying facts being interpreted. Here is an example. Suppose I am at what seems to me a very boring party. I manage to catch the eye of my wife, who is across the room, and she gives me the sort of ‘rolling back of the eyes’ look that I take to mean that she can hardly wait to leave. So I invent an excuse to give the hosts and drag her away. Once we are out the door though she is incensed; she was having a great time. I was mistaken in thinking she wanted to leave. One of two things might have happened. It could be that she rolled her eyes all right but she wasn’t thereby signalling that she wanted to leave. (Maybe she just at that moment noticed the chandelier above her head.) The other thing that could have happened is that she didn’t roll her eyes at all. A trick of the light only made me think she had. It was not that I misinterpreted what I saw. Rather I did not see what I thought I did. This second sort of error, where she did not in fact roll her eyes, is not an error of interpretation on my part but a factual error about what I saw. The first sort of error though was an error of interpretation. The most straightforward way to draw this distinction is by saying that the first sort of error involves misattributing at least one intentional state, such as my wife meaning something by rolling her eyes, while the second sort need not. The second sort might involve only misattributions of nonintentional states, such as whether her eyes moved in a certain way. Our outer space visitor, who never attributes any intentional states to the objects she observes on this backward planet, might still make no mistakes of the second sort. Depending on her observational powers, she might be completely accurate in her description of non-intentional states, properties and the like. As I described Griffith’s story, it involves lots of errors of the second, noninterpretative sort. The football itself for example, does not even get mentioned.4 So to have an example of a purely interpretative mistake of the sort I want to discuss we will either need to do some re-working of Griffith’s story or just use another example, such as my misinterpreting my wife’s rolling of her eyes, or perhaps Wittgenstein’s example of a set of yells and foot stampings, performed by members of some foreign culture, which can be interpreted as moves in a chess game.5 I am just going to assume here that at least sometimes all the non-interpretative mistakes can be eliminated by adjusting the alternative story. That is I am going to assume for now that there can be purely interpretative mistakes.6 The question I want to ask is what has gone wrong when the observer makes such a completely interpretative mistake, that is, where she gets none of the underlying facts wrong but still misinterprets what is going on.

G. F. Schueler

115

An interpretive mistake of this sort will at least involve misattributions of some intentional states to the people on the field. For two teams to be playing a game of football, the players must have many of a very large but indefinite set of intentional states. Similarly, for two groups to be engaging in a ritualized form of debate which involves short statements of position followed by fights, a very different set of intentional states is required. In specifying that only interpretative errors are involved I am supposing that none of the ‘underlying’ physical states, movements and the like have been mistaken by the observer. So none of the things our space alien observes, such as the movements the players make, or the sounds that come from their mouths, are in dispute between the correct interpretation and the mistaken one. Though what actions these movements constitute and what these sounds mean will be of course different in the two interpretations. What the football interpretation holds is the quarterback calling signals, for instance, the ritualized-debate interpretation presumably will have to say is some sort of reference to a text or debate position. So I am assuming that the two competing interpretations are consistent with, and intended to be based on, exactly the same set of ‘underlying’ facts, events, states of the players and so on Of course while much of each interpretation will involve assigning different intentional states to the people involved, there will also be other intentional states of the various agents that are the same in each of the two interpretations, such as beliefs about the colour of the grass. That is, both interpretations will assign them (though of course not the space alien, who doesn’t assign any intentional states to the objects she sees). But the point is that the various beliefs, actions and thoughts ascribed to the players, coaches and officials by each of these two interpretations will be claimed to supervene on the same set of underlying facts, which will include only movements, sounds and the like.7

2 I will explain below why I think this is not problematic, indeed not even uncommon, that is, why it is not always the case that a mistake about one or another underlying fact will serve to distinguish the correct from the mistaken interpretation. But first it might be worth examining whether what I am assuming violates the principle of supervenience as philosophers have used it. Even if that were true I cannot see that it affects the argument I want to make, but in any case it is not true. To say that one state supervenes on some other states, as when a mental state is claimed to supervene on some physical states of the brain, is to say that there can be no difference in the supervening state without some difference in the underlying states on which it supervenes. ‘A set of properties A supervenes upon another set B just in case no two things can differ with respect to A-properties without differing with respect to their B-properties.’8

116

Interpretative Explanations

Someone might think that this means that the elements of the football interpretation and the elements of the ritualized debate interpretation could not possibly be held to supervene on the same underlying facts. Since the first interpretation is correct and the second incorrect, the thought would be, there must be some difference in the underlying facts that distinguishes the two. But that would be a mistake. It is true that there can’t be two sets of events which are exactly the same in all relevant respects but one of which is correctly described as a football game and the other of which is not. But we are not dealing with two (correctly interpreted) sets of events here, only one set, interpreted in two different ways, one of which is mistaken. To see that this difference is important it might help to recall that a claim of supervenience is not the same as a claim of entailment, or indeed of any other regular connection such as would hold if the underlying properties were connected to the supervening property by a scientific law. A claim that one set of properties supervenes on another set is merely a claim about a certain relation between those sets of properties. It says nothing about why this relation holds. As Kim says at one point, the mere fact of ‘supervenience leaves open the question of what grounds or accounts for it ... Supervenience is not a metaphysically deep, explanatory relation; it is merely a phenomenological relation about patterns of property covariation.’9 If there is a nomological connection, or even a logical entailment, between the underlying and supervening properties, then of course that would be explanatory as well, but such connections go beyond mere supervenience. And if there is no such definitional or nomological connection between underlying and supervening facts, the mere claim that the one supervenes on the other carries with it no requirement that denying a supervening fact one must deny one of the underlying facts. The requirement is there, when it is, only because of the connection that explains the supervenience, not because of the supervenience relation itself. On exactly these grounds, I want to claim that so far as the relation of supervenience goes someone can without logical or nomological error deny, for instance, that what she is observing is a football game and yet accept all the underlying facts on which its being a football game supervene. If she is making an error, which in this case she is, it need not be that error. It might help to take a different sort of case, one where it seems clearer that there really is no logical or nomological connection between the underlying facts and the supervening one. So suppose that you and I both find ourselves in court, facing the same judge, charged with the same crime. Discussing our cases, we discover that the various circumstances of our crimes are exactly the same in all relevant respects. Each of us is charged with doing something unfair to a student, lets suppose, and it turns out to be exactly the same sort of thing in exactly the same sort of class to exactly the same sort of student (etc.). Your case is called first, all the relevant facts

G. F. Schueler

117

come out, and you are found not guilty. My case is next, all the relevant facts are brought out again, but I am found guilty. Considerations of ‘cosmic’ or ‘poetic’ justice aside, something must have gone wrong. The judge has been inconsistent. If all the relevant facts are the same in both cases then either both of us have been unfair to our student or neither has been. Fairness and unfairness supervene on the facts. There cannot be a difference as to the fairness of how we treated our students without some difference in the relevant facts of our two cases. Notice however that this tells us nothing about whether what you and I have done is actually unfair. The fact that fairness and unfairness supervene on the facts, and that the facts are the same in each case, entails that either we both treated our student unfairly or that neither of us did. But nothing in this says which it is. The judge would have been consistent, and not violated any consideration of supervenience, whichever decision she had made, as long as she decided both cases the same way. By the same token, two judges, both looking at exactly the same set of underlying facts, and in complete agreement as to what those facts are, can still disagree as to whether the correct interpretation of the law and of the applicable principles of fairness require a verdict of ‘fair’ or ‘unfair’ in our two cases. The judge hearing our cases is mistaken about one case and since the facts of our two cases are the same what makes the one ruling mistaken and the other one correct cannot be a mistake about any of those facts. That is what I am assuming to happen between the football and ritualized debate interpretations of the events Griffith witnessed. Some have held that the supervenience of the moral on the non-moral, and the resulting possibility of such disagreements in moral evaluations, argues for non-cognitivism about the ascriptions of moral concepts.10 But that doesn’t seem at all plausible if, as I am claiming, exactly the same thing applies to football games.11 This assumption about the two interpretations by itself yields an interesting conclusion, which is part of the reason it will be worth looking at it more carefully below. Since the underlying states and events will of course be held to interact causally in exactly the same way under both these interpretations, the difference between the two interpretations – what makes one true and the other false – cannot be any causal factor, any more than it can be a physical or chemical one. Just as both interpretations will be consistent with exactly the same number of people on the field, the same colours of clothing, and the like, so both will be consistent with, because they will be claimed to supervene on, exactly the same set of underlying causal relations among the various events that take place. What the ritualized debate interpretation sees as part of a fight, the football interpretation will see as tackling the tailback for a three-yard gain. But the causal interactions between the events involving the participants will be the same under each interpretation. So the picture is this. We have two completely different interpretations of exactly the same set of underlying facts. One, the correct one, says that a

118

Interpretative Explanations

football game is in progress. The other one says that it is a ritualized series of debates, each of which is followed by fighting. We are supposing that there is no disagreement at all about the underlying facts. And my claim is that, given all this, since the elements of both interpretations will be claimed to supervene on exactly the same set of underlying facts, the difference between the correct interpretation and the incorrect one cannot be found in those underlying facts. In particular it cannot be found in any of the causal relations included in those underlying facts. The elements of both interpretations are claimed to supervene on exactly the same set of facts, which include exactly the same causal interactions between the same events and so on. So whatever it is about these two interpretations that makes the one correct and the other incorrect, that is not where we will find it.

3 Of course, at this point I am really only assuming this is possible in this case, even though it seems a plausible assumption. Still, how could it be so? It will help to contrast the interpretative explanations we are discussing with a situation where it is not so, that is, where a different explanatory theory requires some difference in the facts on which it is based. So consider the difference between two theories supposed to describe the motion of some object through space. Suppose we are technicians looking through the records of radar scans taken on some remote island, covering some part of the sky for the last few minutes. The radar is part of an environmental monitoring programme and our job is to check these records. We notice markings indicating that the radar has detected something, but we don’t know what. Maybe it is a weather balloon, maybe a rocket, maybe only a bird flying in front of the radar. That is what we need to figure out. At first we have recorded only a relatively small number of observations of whatever this object is, four or five. So all we really know is the object’s position at those times. On the basis of these observations we formulate two theories of this object’s motion based on what it might be, theory R, that the object is a rocket, with a smooth path (which, unknown to us, is correct) and theory B, that it is a bird, with a much more erratic path. Since both these theories are consistent with all the observations of this object that we have when we start trying to figure out what it is, there is no evidence from these observations that supports theory R over theory B, or vice versa. So in that respect these two theories are analogous to our football versus ritualized debate interpretations of what is happening on that football field. At the same time, both theories R and B will of course be ‘underdetermined by the data’ which supports them. Both theories make far more predictions about the position of the object in question than anyone has yet actually checked or, really, could ever check, since there will never be more than a finite number of observations and each theory makes predictions for

G. F. Schueler

119

the positions of the object at every point in time, not just the times when actual observations are made. To see which theory is correct we have to look at the predictions each theory makes about the as yet unobserved positions of each object. Both theories are consistent with all observations so far. But the predictions the two theories make about where the object will be observed apply to all possible observations, not just the ones already made. When a new observation is made, say by making another pass with our radar, the position of the object is consistent with the predictions of theory R but not those of theory B. Of course it will always be possible to add a new feature to theory B, an ‘epicycle’ for instance in which the alleged bird flutters into just that position, which adjusts it to the new observation. The resulting, adjusted theory (B2) will once again be consistent with all the observations yet made but it will have the same fate as B when yet another observation is made. And then it too will have been refuted, or at least it is no longer consistent with the observations. This process can continue, of course, but if we stick with any one of these theories, we keep finding we have to abandon or revise it as soon as new observations are made. Contrast this with our two interpretations of what is happening on the football field. Here too both interpretations are underdetermined by the data, that is, there are lots of facts that are ‘brute’ relative to these interpretations but that have not yet been observed. Each interpretation makes predictions about what will happen, say, in the next minute. But, if I am right, in contrast to the two theories of the object detected by our radar, none of these facts, either the ones already checked or the ones ‘predicted’ by the two interpretations, need be inconsistent with either interpretation. How can that be? I suggest that it is because, unlike our two theories of the moving object, neither of the two interpretations of the events on the football field is completely determinate with respect to the underlying facts on which it is based. Both the theories of the object moving in front of the radar and the interpretations of the events on the football field are underdetermined by the data that supports them. They make predictions about much more than has yet been observed. But, unlike the theories of the moving object, the interpretations of the events on the football field are indeterminate with respect to some of the underlying facts on which they supervene. Each leaves lots of possibilities open, even for the underlying facts relevant to the interpretation. That is quite different from the two theories of the moving object. Both theories R and B make predictions about the exact positions of the moving object for every point in time. That is because, whether that object is a rocket or a bird, there will not be any ‘gaps’ in its movement. Because of that, when conjoined to the initial observations about the positions of the object, each theory entails that the object will occupy a specific portion of space. And so each will entail specific, though of course different, claims about what observations will be made. In short, within the parameters

120 Interpretative Explanations

of the theory, each of these theories is determinate in its predictions. For each area of space, each point in time and so on each theory either predicts the object occupies that part of space or that it does not. Nothing like this is true for the two contrasting interpretations of the events on the football field. Even though each will claim to supervene on the same set of actual underlying events, each interpretation is indeterminate within a range of possible events. Each leaves plenty of things ‘open’. For each interpretation there are lots of underlying facts, relevant to the interpretation, which can either obtain or not without affecting the truth of the interpretation. If that sounds mysterious, think of how many open choices there are for those participating in either a football game or a debate tournament. Whether the team on offence calls a running play or a pass play, whether the player with the ball cuts to the left or the right, whether the defence rushes all its linebackers or drops them back in pass coverage, it is still a football game between two specific teams and so on. Similarly for the sort of debate tournament we are supposing for the alternative interpretation. Which specific debate position gets supported by the vote of the team members, for instance, would be a matter of how the members choose to vote. So for each of these two opposing interpretations there will be plenty of underlying facts, facts on which the interpretation in question supervenes, with respect to which the interpretation is indeterminate. Whether the quarterback decides to run or throw the ball, whether the receiver gets tackled or manages to score, it is still a football game. So there is a difference between saying that some theory is underdetermined by the data and saying that an interpretation is indeterminate. Being underdetermined by the data just means that the implications of the theory go beyond the evidence for it. The theory entails claims about the world for which as yet there is no evidence one way or the other. This is as true of both the two sets of cases we have looked at, the theories about the moving object and the interpretations of what is happing on that field. In saying that the two interpretations of what is happening on that field are indeterminate, however, I am saying something different. An interpretation is indeterminate if so far as the interpretation is concerned the underlying facts being interpreted can be of various different sorts without being evidence against the interpretation. Suppose you have a complete physiological theory of how human bodies work. In order to be complete your theory will have as a consequence the proposition that under some circumstances the muscles in the running back’s legs will cause him to move to the right rather than the left. If in the course of a football game these exact circumstances obtain for some running back and yet he moves to the left, your theory will be refuted, or at least have significant evidence against it. Like theory B in the radar example, it will need some revision. But the interpretation of these same events that says that this is a running back carrying the ball in a

G. F. Schueler

121

football game has no such problem. It is indeterminate as to which way the running back moves. Of course to say that an interpretation is indeterminate is not to say that anything goes. It would be better to say that an interpretation specifies a range of possible facts, with things inside that range consistent with the interpretation, things outside inconsistent with it. If it is a football game, ball carriers can run to the left or to the right but they can’t sit down and start working crossword puzzles. But it can still turn out that each of the two interpretations of some specific set of events leave open underlying facts within some range and that the actual events at issue fall into that range for each interpretation. If that happens, then whatever exactly these facts turn out to be, they are consistent with both interpretations. That is what we are assuming for the two interpretations of what happens on that football field. That is why it is unobjectionable to assume that both the correct, football, interpretation and the incorrect, ritualized debate, interpretation can agree completely about the underlying facts on which each supervenes. Both interpretations can supervene on the same set of underlying facts because, as we can put it, their ranges of indeterminacy happen to overlap in such a way that the actual sequence of events on the field falls within both. That won’t always be the case with any two interpretations. Saying there is indeterminacy in interpretations does not mean that nothing falls outside the range of indeterminacy. If that were true then every interpretation would be consistent with every possible set of underlying facts. Suppose that Griffith, instead of interpreting what he saw on the field to be a ritualized debate tournament, had thought he was witnessing a horse race. Horse races are indeterminate in the same sense football games or debate tournaments are since jockeys can manoeuvre their horses in different ways, for instance. But for these two interpretations it is hard to see how the ranges of indeterminacy could overlap. There will be some underlying facts that are allowed by one interpretation and not by the other. Interpreting some set of events as a horse race, for instance, is not consistent with a complete lack of horses, though that is allowed by a football game interpretation. So if Griffith had thought he was witnessing a horse race rather than a football game, his mistake could have been traced to a mistake about this underlying fact.

4 In the football example the indeterminacy arises from the fact that the events being interpreted involve groups of people and numerous choices on the part of those involved. At those places where a choice is possible for someone, each interpretation allows alternatives, each of which is consistent with the interpretation. That is why the underlying facts are neither nomologically, nor definitionally, connected to either interpretation. But it would be a mistake to think that indeterminacy only arises where the events

122 Interpretative Explanations

being interpreted involve groups of people. This indeterminacy is equally characteristic of any explanation that appeals to an agent’s reasons for doing whatever she did. Explanations of actions in terms of agents’ reasons are also interpretative explanations in my sense. The football example is only a special case that happens to involve more than one person. To see this, consider cases where the agent needs to make a choice but can see no reason for choosing one way rather than another (so-called Buridan cases). Suppose I am running some evening, being chased by some bad guys, and I come to a fork in the road. I can see no reason for going left rather than right or vice versa. Still, I need to keep running. I don’t want to get caught. So I just make a choice and go, lets say, left. Clearly in this situation turning left is something I do intentionally but it seems false to say that I have a reason for doing it. I might neither have, nor think I have, reason to go left rather than right, though I have reason for continuing to run. And to do that I must go one way or the other. So I have a reason for choosing one or the other direction. But though I intentionally turn left, it is not true that I have a reason for turning left rather than right. So not all intentional actions are done for reasons. Explanations of actions in terms of the agent’s reasons do not cover everything agents do intentionally. There are some choices one makes, and sometimes in fact must make, where one does not oneself think one has a reason to choose one way rather than another. And Buridan cases of this sort are common. Most cases of doing things for reasons ‘contain’ intentional actions of this sort. When I turn down the road to the left I am of course doing something for a reason. I am running away from those bad guys. But that would also have been my reason had I turned down the road to the right. My action of ‘running away from the bad guys’ itself involves other intentional actions some of which, like turning down the road on the left rather than the one on the right, involve choices between different things which are, relative to my goal, equally ‘reasonable’. And in all such situations the choices of each of those things are typically not done for reasons. Many of the so-called ‘basic actions’ by which one performs the (less basic) actions which one performs for reasons are still intentional actions.12 But there are often numerous possibilities and for the most part the choice of one of these rather than another is not something one does for a reason, like the choice to turn left rather than right. So explanations of actions in terms of the agent’s reasons are frequently indeterminate in the same way an interpretation of those events on that field as a football game is indeterminate, and for the same reason. In each case the explanation (or interpretation) is consistent with various choices on the part of the agents involved. Within some range these choices can go in quite different ways and still be consistent with the interpretation in question. All this argues that the same conclusions about the possibility of an alternative interpretation supervening on the same set of underlying facts can be

G. F. Schueler

123

drawn for explanations of actions generally that we saw followed for the interpretation of the football game. Suppose Andy Griffith sees me running, turning left at the fork in the road, and interprets what he sees as just another jogger, out to get some exercise. He would have misinterpreted what he saw. What I am really doing is running away from those bad guys. So here again we would have two interpretations, each of which is (or at least could be) consistent with the underlying facts on which its elements supervene, because each is indeterminate with respect to numerous open choices the agent in question can make, including the choice of whether to turn left or right at that fork in the road. This means that, as before, the difference between the correct and the incorrect interpretation may not be found in any of the underlying facts on which the elements of these two interpretations supervene. In particular it may not be found in some causal connection which one interpretation uses or presupposes and the other does not. The difference between the two interpretations is not ‘causal’ in this way since each interpretation might supervene on exactly the same set of underlying facts, including facts about the causal connections among the various events involved, such as the muscle contractions in my legs that propel me to the left rather than the right when I arrive at that fork in the road.

5 An obvious question remains. The correct and incorrect interpretations are both consistent with the same set of underlying facts and yet one is correct and the other not. How can that be? The answer, I suggest, is that what the correct interpretation includes, and the mistaken one misses, is the actual point or purpose of what the agent or agents are doing. Consider the running example again. Even if all my movements, even all my thoughts, would be the same whether I were merely out jogging or trying to escape some bad guys, the point of what I am doing would be quite different in the two cases. I am not merely trying to get some exercise; I am trying to save my skin. Perhaps when I first encounter those bad guys I reason that prudence is the better part of valour and decide to run away, heading with my usual opening sprint down the road I ordinarily take, in exactly the way I have begun my evening run every day for months. A few blocks along, just as I come to a fork in the road, I pass Andy Griffith, who thinks I am out for my usual evening run. But he is mistaken, even though he is correct about all the facts about my leg movements, speed, direction and so on. That is, the elements of his incorrect interpretation of what I am doing supervene on exactly the same set of underlying facts as do the elements of the correct interpretation, which is that I am running away from those bad guys. (They are exactly the elements that our space alien, had we enlisted her at this point, would have observed.)

124 Interpretative Explanations

Nor can we say that because my conscious decision would be different in each case there must be a difference in the underlying facts for the two interpretations. Even leaving aside the fact that I need not have made any conscious decision, the purpose of what I am doing is not always the same as the explicit decision I come to, or even my belief I have about what I am doing. Akrasia and self-deception are always possible. Not only all my physical movements but even my conscious reasoning and resulting decision, in fact even my own belief as to what I was doing, might be the same whether I was just jogging or was actually running away.13 The difference would be that my real purpose was to get away, no matter what I or anyone else thought I was doing. The possibility of weakness and self-deception shows that both the correct and incorrect interpretation of what the agent is really doing are consistent with any explicit reasoning or choice the agent makes. If the purpose of the action is what determines whether an interpretation is correct, we can see at least one reason why interpretations are indeterminate. I have been arguing that interpretations are indeterminate, whether they are about the actions of individual agents or about events that involve cooperation among several agents. They supervene on the underlying facts but they allow ranges of facts, rather than specifying specific underlying facts at every point, as determinate theories do. That is why there is no nomological connection between the underlying facts and the supervening, interpretive claim. Different interpretations can be perfectly consistent with the same set of underlying facts. And in the cases we have considered so far this is apparently because both interpretations allow open choices for the agent or agents in question. If this is right it tells us how the sort of indeterminacy I am claiming for these interpretations is possible, what it consists in so to speak, at least in these cases. But it does not explain why these, or any, interpretations have this feature. The answer to that question, I think, is to be found in the same thing that makes one interpretation correct and another one mistaken. The difference is that the correct interpretation includes, and incorrect ones miss, the actual purpose (or purposes) of the action or actions. Purposes necessarily involve the sort of indeterminacy we have seen in the interpretations we have looked at. It may be easiest to see this, and to see that having further open choices is not essential, if we shift for a moment from actions that have purposes to objects that do. So consider the large rock that rests at the corner of my friend Steve’s driveway, just where it meets the road. This rock has a purpose. Steve’s house is on a hillside and his driveway is rather steep. It crosses a ditch (via a culvert) as soon as it leaves the road and then immediately makes a left turn downhill. The driveway is also narrow enough that if you drive in then, when you want to leave, you have to back out, since it is very difficult to turn around. The purpose of the rock is to keep people who are backing out of the driveway from accidentally going off into the ditch.

G. F. Schueler

125

There are a few things to notice here before explaining how indeterminacy enters into this story. First, obviously this rock has this purpose only because Steve has a purpose for it. In this it differs from what biologists sometimes call ‘functions’, which are a result (roughly) of the evolutionary history of the ancestors of the thing that has the function.14 Such functions can be discovered but they are not assigned. Even non-sentient things such as flower petals can have them. Rocks can’t have functions of that sort, not being organisms or parts of organisms.15 But they can have purposes, because people can have purposes for them. Purposes involve indeterminacy in at least three ways. First, there are plenty of features of that rock that have nothing directly to do with its purpose of keeping people from driving into that ditch when backing out of Steve’s driveway. To serve its purpose of course it must have some colour, for instance. But within limits, it probably does not matter what colour the rock is. Similarly for size and shape. In general when objects have purposes those purposes are served by specific features of the objects, such as, in the case of this rock, the fact that it might do some damage to one’s car to hit it. But objects always have plenty of other features than the ones which serve the purpose in question, and (perhaps within some limits) those can be anything at all and the purpose will still be served. Second, nothing about a thing’s purpose by itself specifies how it is going to be achieved. Even for that rock, its purpose might be achieved in more than one way. People might see it in their mirrors and turn slightly to miss it. Or they might hit it with a tire and change course slightly. If either happens the rock will have served its purpose. But, so far as this purpose is concerned, it does not matter which happens. Nothing about a thing’s purpose requires that it be achieved in a specific way.16 But, third, the fact that something has a purpose in no way insures that this purpose will actually be achieved, or even that it can be. The fact that the purpose of that rock is to keep people from driving into the ditch is perfectly consistent with its having no effect whatsoever on the people backing out of Steve’s driveway. Imagine that it is in fact a very small rock that no one even notices. Its purpose could still be to keep people from driving into the ditch. All these sources of indeterminacy are, I think, consequences of the fact that purposes are ‘intentional states’ in something at least very like the way beliefs and desires are, which is why inanimate things such as rocks can have purposes only if someone has a purpose for them. But, given that, the rock still has a purpose and that purpose creates an ‘intensional’ context in the sentences in which it is referred to, just as any other intentional state does. For example, the purpose of that rock is to keep people from going into the ditch when backing out of Steve’s driveway. It is also true as a matter of fact that keeping people from going into the ditch when backing out of Steve’s driveway is a saving of the amount of gasoline needed to hitch them up to

126 Interpretative Explanations

Steve’s pickup to pull them out. But it doesn’t follow that the purpose of that rock is to save this gasoline, though of course it might have been. If we return now to actions and the events that constitute actions, we shouldn’t be surprised that we find these same sorts of indeterminacy. Human actions are events that have purposes supplied by the agents of the actions.17 So for actions, or at least for most of them, figuring out what the purpose of the action is or was is essential to figuring out what the action is or was. To do that is to give what I am calling an interpretative explanation.

6 Purposes, I am claiming, explain actions but always involve indeterminacy. More than one purpose or set of purposes is always consistent with the actual underlying facts about the objects or events on which the purpose supervenes. That is the fundamental reason why the underlying facts for any interpretation need not entail or even be nomologically connected to just one interpretation. But the question remains, what makes one assignment of purposes correct and another not if it is not the underlying facts being interpreted? I think the straightforward answer is simply that the action or actions being interpreted really do have the purposes assigned by the correct interpretation. To see what this comes to it will help to distinguish two different sorts of questions. There is a difference between asking how we know what purposes are (at all, so to speak) and asking how we know when people really have certain purposes. These can get conflated if we think that figuring out when others really have a certain purpose must be in the end a matter of reducing purposes to their constitutive elements and then doing an investigation of when the actions of others possess those elements. But this is not how it is. We should distinguish between saying that some concept ‘applies to’ some data and saying that a concept is ‘based on’ certain data. Consider some theoretical entity, such as an electron. How do we know that there is any such thing as an electron? The answer is that electrons are hypothesized by empirically very well-established physical theory. And the evidence for the theory is also evidence for electrons, in fact this is all the evidence for electrons that there is. Electrons have precise, detailed roles in explanations of lots of physical phenomena, including electricity, chemical bonds and many others. They contribute to these explanations, that is, the idea that there are electrons is empirically applicable. But at the same time electrons are only known to exist because of their place in these explanations. So the idea of an electron, besides being applicable to phenomena, is also based on exactly the same phenomena in the sense that electrons are essentially theoretical entities. If different theories which do not use the concept of an electron are found to do a better job explaining the same phenomena, that will be taken to

G. F. Schueler

127

show that we were wrong to think that there ever were any such things as electrons. Or we might in that case end up saying that there turned out to be several different sorts of electrons, perhaps, or that, besides electrons, there were other particles that were previously thought to be electrons but were not. The point is that there is no other reason to think that there are electrons at all beside their usefulness in the theories in which they appear. If that usefulness turns out to have been illusory, so will electrons. The fate imagined here for the notion of an electron is in essence the fate predicted by eliminative materialism for all intentionalistic and purposive concepts. The idea is that once the underlying neurophysiological mechanisms are understood, the much cruder ‘folk’ theoretical concepts such as purpose and intention used in commonsense explanations of action will be seen to have been illusory. This claim presupposes that these concepts, like that of an electron, are not only applicable to the phenomena with which they deal but also that they are based on these phenomena in the sense that we have no further reason to think the things they refer to actually exist beyond that provided by whatever evidence supports the theories in which they appear. But not all concepts that are applicable to the phenomena they explain are like this. In particular the idea of the purpose of an action is not like this. Our grounds for having these concepts, that is, for thinking that they apply at all, are not based on the phenomena to which they apply but arise independently of the explanations in which they are used. Think about our space alien again. She suspends judgment as to whether the complex organisms she encounters on this planet have any intentional states, including any aims or purposes. But that does not mean she needs to suspend judgment about whether she herself has goals or intentional states, and she need not suspend judgment about this even if she goes on to accept the version of solipsism that actually denies there are any other minds than her own. Such a solipsist would simply deny the applicability of intentional or purposive predicates to others than herself. This is, or at least seems to be, a coherent position. (It is for instance the position dualists seem forced into by ‘other minds skepticism’). It is not incoherent even if it seems very implausible. The thought would seem to have to be that one knows directly from one’s own case what, for example, purposes and intentions are, while for others one only knows for sure about the various movements and sounds their bodies produce. There are two things to notice about this position. First, this position is very implausible. Once one agrees that there are such things as intentional states, purposes and the like, the evidence that others have these states is overwhelming. Someone who refused to make use of purposive explanations of others, including, importantly, purposive explanations of what they are doing in making sounds come from their mouths, would find it virtually impossible to make any sense at all of human activities. She would thus find it impossible to engage in any sort of distinctively human interactions. At

128 Interpretative Explanations

the same time, regarding these activities as purposive, intentionally contentful and the like would completely solve this problem.18 So, while coherent, this form of solipsism would seem to be profoundly unempirical. Once one sees that it is possible for something to have purposes and other intentional states, the evidence is overwhelmingly in favour of the claim that others do indeed have them. The second thing to notice here is that the coherence of this form of solipsism presupposes that concepts such as ‘purpose’ can be known from, and applied to, oneself independently of their applicability to others.19 If it is coherent to suppose that one might oneself be the only purposive agent in existence, then the concept of a purposive agent is not dependent on having a place in theories explaining the behaviour of other people. So if this is right the knowledge of what purposive agency is, and the knowledge that there are purposive agents at all, could survive the discovery that a complete explanation of the behaviour of others, by neurophysiology for instance, had no place for the concept of purposive agency. But of course neurophysiology (eventually, when complete, etc.) is supposed to explain the behaviour of everyone, oneself included, not just of ‘others’. So the form of solipsism that holds that I know from my own case that I myself have purposes, intentional states and the like, but I don’t know whether anyone else has such states, cannot be allowed by eliminativists of the sort mentioned at the beginning of this essay. An eliminativist will have to hold that the form of solipsism we are considering here is in fact not coherent since it does not apply neurophysiological explanations to everyone it should apply to, oneself as well as others. It tries to make an exception for the first person case. An eliminativist will have to hold that observing oneself no more reveals, or provides grounds for hypothesizing, purposes or intentions than observing others does. This is the heart of the issue. When I ‘learn from my own case’ that, say, my purpose in running down the street is to get away from those bad guys, does that mean that I somehow ‘observe’ myself internally and then on some grounds or other attribute such a purpose to myself, more or less in the same way someone else who is observing or thinking about me might? An eliminativist, who holds that purposive concepts are (supposed to be) based on the evidence they (try to) explain, will have to answer yes to this question. For an eliminativist, purposes are simply crude or defective empirically based concepts. But the answer has to be no. Having a purpose, acting with some purpose, is itself a ‘state’ of an agent, frequently perfectly conscious, different from merely attributing such a state to someone, even oneself. And given the sort of state it is, there has to be an element of selfawareness involved. In the normal situation at least, someone who is acting with some purpose must by the very fact of acting on it, realize what this purpose is. Being an agent, acting with some purpose, is itself a certain sort of ‘mental’ state. So ‘having a purpose’ is not a theoretical concept, like

G. F. Schueler

129

‘electron’, that depends for its use (and one might say, existence) completely on some explanatory theory. Thinking that notions such as purpose, intention and the like are theoretical concepts analogous to ‘electron’ is similar to the mistake pointed out long ago by J. L. Austin of thinking that all language use is descriptive. Not all mental states have their content exhausted in being ‘about’ something else in the way beliefs and desires are about something else. In particular doing something intentionally is not like this. The paradigm first person example of purposes is surely not just describing one’s own purposes to oneself, that is, thinking that one has a purpose of some sort. It is actually having that purpose. Otherwise, if ‘purpose’ were a purely third person, explanatory concept the content of which was unavailable to whomever actually had the purpose just in virtue of having it, while we could interpret or think about our actions, it is hard to see how we could actually perform actions. It would be as if we were condemned to being merely internal observers of the motions and sounds we were making, forever trying to figure out what we ourselves were doing. This is not just comical; it is incoherent. If it were correct then it is hard to see how anyone could ever actually do anything at all, since just having a purpose, without also reflecting on it, wouldn’t by itself be enough to let the agent know what her purpose actually was. So it is hard to see how she could actually be pursuing it. At best she would, like everyone else, have to try to figure out what she was doing by thinking about the movements and sounds she was making. But of course ‘trying to figure out what I am doing’ is itself a purposive activity. That is why this picture is incoherent. It has to presuppose that the internal observer is herself acting with some purpose of which she is aware, that is, in trying to figure out the purposes of the motions and sounds she observes herself making. If this purposive activity, trying to figure out the purposes of the motions and sounds she is making, isn’t something the content of which is available to her just by having it then presumably yet another observational level will be needed to try to figure out the purpose of this one. But obviously this does no good since exactly the same issue would arise in exactly the same way all over again. It follows that just having a purpose includes, typically at least, awareness of that purpose, automatically so to speak, without the need for further interpretation and without anything analogous to ‘observation’. The idea that I have to interpret my own actions (or movements) in the way I interpret those of others is not coherent. It leads to a regress, since interpretation is itself a purposive activity. This is a way of saying that ‘purposiveness’ is not a theoretical concept, used in and dependant on explanations of behaviour. It is indeed applicable to behaviour but it is not based on behaviour. It is a concept we bring to explanations of behaviour from the fact of our own agency. These last two points together give us at least the basic elements of an answer to the question of what makes an interpretation, an assignment of

130 Interpretative Explanations

purposes, correct. The answer is that this is an empirical issue that turns, like any such issue, on the explanatory power of the interpretation proposed, but the essential concept involved, that of purposiveness, is not itself based on evidence in the way other theoretical concepts are. We understand purposiveness, acquire the very idea so to speak, from the fact that we are ourselves purposive agents. Once we have that idea, however, we can apply it like any theoretical concept to the events and behaviour that confront us.

Notes 1. The original version included lots more detail and was of course much funnier. It is available at http://www.carolinafan.com/ar/02/020926_what_it_was.html 2. She is probably a relative of the Martian described by Daniel Dennett in ‘True Believers’, in The Intentional Stance (Cambridge, MIT Press, 1987), pp. 13–42. 3. We will return to this issue below. 4. Though this is true of my version of the story, all these elements are in fact included in Griffith’s actual routine. 5. Wittgenstein, Philosophical Investigations, paragraph 200. 6. This assumption will be defended below. 7. This is the actual claim, but I will sometimes abbreviate this by saying that the two interpretations supervene on the same underlying facts. 8. McLaughlin, Brian and Karen Bennett, “Supervenience,” The Stanford Encyclopedia of Philosophy (Fall 2006 Edition), Edward N. Zalta (ed.), URL = . 9. Quoted in Scott Sehon, Teleological Realism (Cambridge, MIT, 2005), p. 117. 10. See for instance, Simon Blackburn, ‘Moral Realism’ in Morality and Moral Reasoning ed. J. Casey (Methuen, London, 1973) reprinted in his Essays In Quasi-Realism (New York, Oxford University Press, 1993). 11. On the general question of whether there can be supervenience without either logical or nomological ‘reduction’ see Sehon (2005), Chap. 8. Sehon gives several examples, based on making the supervening property a noncomputable number, where there is supervenience but no possibility of entailment of the supervening facts by the underlying facts no matter what extra scientific law is supposed to connect the two. 12. A ‘basic action’ is something one does, such as raising ones arm, but not ‘by means of’ doing something else. To take the earlier example, one might signal one’s boredom by means of rolling one’s eyes back. So signalling boredom would not be a basic action but rolling ones eyes back presumably is. 13. According to Wittgenstein, ‘It is, of course, imaginable that two people belonging to a tribe unacquainted with games should sit at a chess-board and go through the moves of a game of chess; and even with all the appropriate mental accompaniments’ (Philosophical Investigations, § 200). 14. I am simplifying things here since there is another account of function which doesn’t depend on evolution, roughly the ‘causal role’ account. Though using that account would complicate the argument here, so far as I can tell it makes no essential difference. See my Reasons and Purposes (Oxford, Oxford University Press, 2003), Chap. 1 for a fuller discussion of these two accounts. 15. Even a rock could have a function in the ‘causal role’ sense though. A rock might function to keep moisture in the soil under it from evaporating for instance.

G. F. Schueler

131

16. Of course whoever gave the object its purpose might have believed or even intended that the purpose would be achieved in a specific way. But that is not strictly required for a thing’s having a purpose. 17. I don’t intend this to be a definition. It seems to me to be true, but if there are actions of which it is not true, then obviously what I say here won’t apply to them. Of course ‘supplying’ a purpose need not be, in fact cannot be, an intentional action, since that would just lead to a regress. 18. Dennett imagines a ‘predicting contest’ between a human using the ordinary, purposive interpretations and an outer space alien such as the one imagined above who knows all the underlying physically described facts but uses no purposive or intentional concepts. Without the intentional and purposive concepts, the outer space visitor of course losses badly. See Dennett (1987). 19. This of course was famously questioned by Wittgenstein and his followers.

8 Anscombe on Expression of Intention Richard Moran and Martin J. Stone

Of course in every act of this kind, there remains the possibility of putting this act into question – insofar as it refers to more distant, more essential ends ... . For example the sentence which I write is the meaning of the letters I trace, but the whole work I wish to produce is the meaning of the sentence. And this work is a possibility in connection with which I can feel anguish; it is truly my possibility ... tomorrow in relation to it my freedom can exercise its nihilating power. Sartre, Being and Nothingness1 There might be a verb with the meaning: to formulate one’s intention in words or other signs, out loud or in one’s thoughts. This verb would not be equivalent in meaning to our ‘intend’. There might be a verb with the meaning: to act according to intention; and this would also not mean the same as ‘to intend’. Yet another might mean: to brood over an intention; or to turn it over and over in one’s head. Wittgenstein, Remarks on Philosophical Psychology I2

I The Problem Anscombe begins her monograph Intention by recalling three familiar contexts in which, as she says, we ‘employ a concept of intention’ (§1): (Case 1) Someone says ‘I’m going to walk to the store’: An expression of intention, she says. (Case 2) Someone is walking (or has walked) to the store: An intentional action.

132

Richard Moran and Martin J. Stone

133

(Case 3) ‘Why are you walking to the store?’ – ‘To get some milk’: The question seeks – and the answer provides – the intention with which something is done. 3 This isn’t philosophy yet, only its raw material. Anscombe will shortly suggest the need for a philosophical investigation by suggesting that we have trouble seeing how one concept of intention finds application in these cases.4 This set-up seems straightforward, yet on closer inspection someone might understandably object: ‘What an odd mix of cases!’ Case 1 is a type of act (viz. one of speaking), and, as such, an instance of the second. Doubtless, when performed, it is also done with some further intention (Case 3). So there is overlap here. But apart from this, it might be wondered: Why should expression of intention appear in an initial division of the subject at all? Sure, we put ‘a concept of intention’ to work in this context. But we put ‘a concept of emotion’ to work in speaking of someone’s expressing emotion; and that seems unlikely – save in a world described by Borges – to head the opening divisions in a book called ‘Emotion’. In general, that ‘expressions of Ø’ will be pertinent in studying a psychological concept Ø isn’t simply to be taken for granted. Exceptions would include those concepts taken up in performative verbs, where (e.g.) to command or promise is to formulate something in words, to give it expression. But that to intend isn’t a performative is seen among other ways in this, that we can speak of expressions of it, though not of a command;5 and from this point of view, it seems doubtful that ‘expressing an intention’ has any greater claim to pertinence than, say, ‘brooding over an intention’ or ‘concealing an intention’, and so on. Yet this is Anscombe’s first sentence: ‘Very often, when a man says “I am going to do such-and-such,” we should say that this was an expression of intention’ (p. 1). Anscombe does not pause here to explain why she draws attention to ‘expression’, but instead turns to distinguishing two different uses to which (e.g.) ‘I am going to fail in this exam’ might be put – either (1) an expression of intention, or (2) the speaker’s estimate of her chances, a prediction (p. 2). Distinguishing expressions of intention from predictions suggests that the two were liable to be confused. So it is in fact a particular sort of ‘expression’ which comes up here, a verbal statement of fact. To explain, we express ourselves, and our states and attitudes express themselves, in a variety of ways – through what we do and say, and through how we do these things. Since we are thus bound to be expressing things continuously, a central distinction in this area will be between expression in the impersonal sense (the manifestation of some state or condition) and expression in the personal sense (the intentional act of one person directed to another).6 Evidently, Anscombe is thinking of the personal sense, for her examples are all imperfect statements of fact (‘I’m going to ...’)7 – items which can do double duty for predictions – notwithstanding that intentions are, it would seem, expressed in other ways

134

Anscombe on Expression of Intention

as well. ‘It would seem’: In fact, in a passage occurring just after the opening, Anscombe will be found denying that intentions, in contrast to other states of the person, are ever impersonally expressed (p. 5). (The meaning of this strange and unintuitive doctrine will occupy us in much of this essay.) To say that Anscombe has inherited this focus on ‘expression’ from Wittgenstein, though correct, is obviously not the sort of explanation needed here.8 And in truth, if the idea were merely to locate the topic of intention by remembering some main uses of ‘intention’, something seems forced here. For we don’t really ‘very often’ speak of ‘expression of intention’ in characterizing what people say. Philosophical purposes aside, we don’t generally speak this way unless the context gives special consequence to the distinction between the two kinds of expression – for example, in the law, which sometimes asks whether an intention (e.g. to take possession) has been expressed (the personal sense), and not merely whether it was evident under the circumstances to others. What, then, is the meaning of Anscombe’s initial emphasis on ‘expression’? We aim to show the work this notion is performing throughout Intention. But the background to this task is that Anscombe’s first division – expression of intention – seems incongruous. It seems to belong only to a very different catalogue of divisions, one featuring such items as symptoms or indications of intention, or intentions which are disguised or merely passively revealed; or perhaps in a catalogue of speech-acts – expressions of belief, predictions, commands, promises and so on. Consider now that rather strange and unintuitive passage. ‘Expression’ recurs emphatically here, in the form of a claim made by Anscombe, that while non-human animals (brutes) have intentions, they do not express them: Intention appears to be something that we can express, but which brutes (which, e.g. do not give orders) can have, though lacking any distinct expression of intention. For a cat’s movements in stalking a bird are hardly to be called an expression of intention. One might as well call a car’s stalling the expression of its being about to stop. Intention is unlike emotion in this respect, that the expression of it is purely conventional; we might say linguistic if we will allow certain bodily movements with a conventional meaning to be included in language. Wittgenstein seems to me to have gone wrong in speaking of the ‘natural expression of an intention.’ (Philosophical Investigations (§ 647) (p. 5, original emphasis) Someone might read past this unhindered, because they might understandably take Anscombe’s point to be merely the anodyne one that brutes do not tell us of their intentions (i.e. express them in the personal sense). But that can’t be her point. For clearly what this passage is after is a contrast between intentions and states of emotion with respect to their possibilities of expression

Richard Moran and Martin J. Stone

135

(‘Intention is unlike emotion in this respect ...’). That anodyne rendering would undo the contrast. For the natural behaviour of brutes does express emotion (or so Anscombe allows), yet brutes do not tell us of their emotions either. So Anscombe’s point is best represented – as she herself represents it – like this: There is no such thing as the natural expression of intention, as there is of emotion; the expression of intention is always conventional or linguistic. First appearances notwithstanding, this is not essentially – but only by application – a point about non-human animals. Anscombe might also have said (she should be committed to saying), ‘The natural behaviour of human beings is no ‘expression’ of their intention either.’ But the question is, why not? Why say this? The problem which emerges here, beyond Anscombe’s making ‘expressions’ one of the topic’s divisions, is what the relevant notion of ‘expression’ might be, such that it has no application to the intentions manifested by an agent’s stalking movements, while still finding purchase on the non-verbal manifestations of other states like emotion.9 The solution to both problems involves seeing Anscombe’s emphasis on ‘expression’ – and more generally her explanation of intention ‘in terms of language’ (p. 86) – as part of a distinctive strategy for elucidating the unity of the uses of intention. We will call this strategy one of immediate elucidation, and will contrast it with ‘connective strategies’. An immediate elucidation exhibits the divisions of intention as inflections of a single form. It thereby also helps reveal how the unity of ‘intention’ has become linguistically submerged, hence lost to a philosophically unassisted view. To explain this, however, it seems best to begin by recalling the shape of a prominent strand in contemporary, post-Anscombian philosophy of action, since connective strategies appear to be the main ones, or the only ones, imagined there.

II Background: the transformation of Intention A familiar homage says that many action theorists follow Anscombe at least in (1) associating acting intentionally with acting for reasons, (2) treating the topic of intention as comprising this and at least two other divisions (‘intending to act’ and ‘intention with which’) and (3) requiring philosophy to explain how these notions are connected. This account of the influence of Anscombe’s work is only roughly correct, however. A sign of inexactness is just the way her first division (expression of intention) is apt to be remembered – for example here, by Davidson, in recounting his own theory: [Earlier] I believed that of the three main uses of the concept of intention distinguished by Anscombe (acting with an intention, acting intentionally, and intending to act), the first was the most basic. Acting intentionally, I argued ... was just acting with some intention. That left intending, which I somehow thought would be simple to understand

136

Anscombe on Expression of Intention

in terms of the others. I was wrong. When I finally came to work on it, I found it the hardest of the three; contrary to my original view, it came to seem the basic notion on which the others depend; and what progress I made with it partially undermined an important theme... – that ‘the intention with which the action was done’ does not refer to an entity or state of any kind.10 Losing reference to ‘expression’, Anscombe’s first division has become ‘intending to act’ – a hard notion for action theory, as Davidson avers, partly because, being potentially free of any contamination by action, it seems to refer to an as yet unanalysed state of mind. Moreover, given the general drift here – an organization of ‘intention’ around a distinction between worldly events and autonomous mental states – one might understandably speak of just two main Anscombian divisions. Thus, Michael Bratman: [W]e use the notion of intention to characterize both people’s actions and their minds. Thus, I might intentionally pump the water into the house, and pump it with the intention of poisoning the inhabitants. Here intention characterizes my action. But I might also intend this morning to pump the water (and poison the inhabitants) when I get to the pump this afternoon. And here intention characterizes my mind. (p. 1) ... [O]ur common sense psychological scheme admits of intentions as states of mind; and it also allows us to characterize actions as done intentionally, or with a certain intention. A theory of intention must address both kinds of phenomena and explain how they are related. A natural approach, the one I will be taking here, is to begin with the state of intending to act. (p. 3)11 And Bratman distinguishes this approach from that of other theorists as follows: Instead of beginning with the state of intending to act [some theorists] turn immediately to intention as it appears in action: [they] turn directly to acting intentionally and acting with a certain intention .... This is, for example, the strategy followed by Elizabeth Anscombe in her ground-breaking monograph, Intention.12 On this reception of Anscombe, the question naturally arises: Which is the more ‘basic’ notion of intention? For Bratman, as for Davidson, it is ‘intending to act’, though in analysing this state Bratman goes boldly (where Davidson had gone only reluctantly) beyond mere ‘desire’ and ‘belief’ to a much richer psychology of states, one more adequate to the complexities of action.13 The significance of this development will come in for interpretation later on.14 What matters for the moment, however, are only two apparent

Richard Moran and Martin J. Stone

137

commonplaces heard in these passages: first, that ‘intending to act’ is some state of mind; and second, that action theory should explain how this state is related to ‘intentional’ as it characterizes things in the ‘world’ – what people do or cause to happen. We call the framework comprised by these points Transformed Anscombe (TA). The steps leading from Anscombe to TA might be thought of as follows: 1. Noting that Anscombe sometimes calls her first division ‘expression of intention for the future’, and sensing perhaps that the invocation of ‘expression’ is not essential here, the later term is (understandably) factored-out, leaving ‘intention for the future’ (‘prior-intention’ or ‘intending to act’) as the item Anscombe meant to distinguish.15 2. This effect of this – and perhaps one of its motivations too – is to give the list a new sharpness. For occupants of the new category – for example, someone’s intention to fly to Boston next week – seem, as such, to be neither intentional acts nor (supposing the agent has not yet done anything to realize her intention) any intention in acting. Naturally, such an intention may be taken up – or become present in some way – in her flying to Boston, and in the intention with which she does other things, like packing her suitcase.16 But it need not be. For she may change her mind, or something may interrupt her plans, and then her intention to fly to Boston will remain ‘pure’. 3. Given the possibility of ‘pure’ intending, it becomes hard to see how this category could fail to designate a mental state, attitude or disposition of some kind.17 So the divisions of ‘intention’ now take shape around the philosophical polestar of the division between mind and world: two notions of intention find purchase only where there is behaviour causing things to happen; a third refers to a mental state, attitude or disposition which, though in some way present in such behaviour, is also abstractable from it and capable of existing on its own. 4. The theoretical elucidation of ‘intention’ is now apt to be organized around two tasks: (1) an analysis of ‘intention’, conceived as a (potentially ‘pure’) mental state or attitude, and (2) an explanation of the other behaviour-dependant applications of ‘intention’ in terms of (1). Theorists are of course apt to disagree about how to carry out these tasks: What other mental states or attitudes are entailed by intention, whether some reduction – ontological or otherwise – is possible, and the relation between the relevant state of intending and intentional action, are familiar points of controversy. 5. Generally, in TA: (A) An event is an action when it is intentional under some description; (B) an action meets this requirement when it is done for reasons; (C) this means it is susceptible to a form of rationalization (a special sense of ‘He did it because ...’); and (D) such rationalizations elucidate action through a movement from inner to outer: from the

138

Anscombe on Expression of Intention

agent’s beliefs, desires or other states to something happening outside him. (Whether either such a movement or its explanation is – in some previously recognized sense – a causal one, is another hub of controversy.) From within this framework, what differentiates Anscombe is commonly thought to be the ‘methodological priority’ she gives to intention as it appears in action and – curiously, as she does not discuss this – her denying that the relation between the mind and action sides of the rationalizing ‘because’ is a causal one.18 Bratman in fact portrays Anscombe as accepting everything in (5), with one qualification: Her view of whether ‘intending’ is an independent psychological state (not reducible to appropriate desires and beliefs) cannot be made out because she says too little about this state.19 This story of progress within a single framework of action theory has its obvious satisfactions. Nonetheless, a distortion is present when the contemporary theorist credits Anscombe with having discerned the starting points of TA, as described here. This comes out in the following puzzle: While Anscombe does contemplate intending in its putatively ‘pure’ or unworldly form – ‘[A] man can form an intention which he then does nothing to carry out, either because he is prevented or because he changes his mind: but the intention itself can be complete, although it remains a purely interior thing (p. 9)’ – she offers, as Bratman observes, almost no account of it. In fact, returning late in the book to her first division, Anscombe merely remarks briefly that what has been ‘said about intention in acting applies also to intention in a proposed action’ (p. 90). Here a large gap in her account must appear from the contemporary point of view: How can her previous teachings about intentional actions, things done, simply now ‘apply’ to an intention in a proposed action, conceived as something which may remain ‘a purely interior thing’? By TA’s lights, Anscombe’s previous remarks cover 5 A–D. Beyond identifying the central case of intentional action with action for a reason, they explain how when an agent so acts he has a further intention, which often furnishes a wider description of what he is doing – and related matters. But surely some further explanation of ‘intention in a proposed action’ – and not merely an application of these doctrines – was needed, even by Anscombe’s standard of compression. For we cannot understand, say, ‘an agents intention to fly to Boston next week’ merely on the anaemic basis of an agent’s having a reason to do as much, however this notion is analysed. 20 Among the problems, there is the common fact that the agent may simultaneously have a reason not to fly to Boston without it being the case that she both intends to fly to Boston and intends not to. Intentions apparently stand open to contradiction in a way which mere reasons do not. Since what has been said about intention ‘in action’ thus seems

Richard Moran and Martin J. Stone

139

insufficient to cope with intention ‘in a proposed action’, the question arises: Having distinguished these notions of intention, why would Anscombe omit any significant analysis of the latter? Two answers appear in the literature. First, no reason: Anscombe simply omits to discuss intending, leaving it to us. This seems implausible: One should really choose between making Anscombe the founder of the Three Divisions and having her omit to discuss the first – the combination amounts to philosophical malpractice. Second, her reason is behaviourist: She thinks we will grasp how to explain intending as some function or complication of intentional action, which she takes to be more basic. This too is implausible: and are we really to suppose that Anscombe (a student of St. Thomas, after all) seeks, with the behaviourist, to solve intellectual problems by collapsing spirit into nature? Significant internal difficulties block the attribution of behaviourism, in any case. Anscombe’s forthright talk of intention as ‘a purely interior thing’ is one difficulty. Her pivotal remark that what was said about intention in acting ‘applies also to intention in a proposed action’ is an even greater one. For a genuine behaviourist does not talk that way. He does not say that his account of (e.g.) pain-behaviour applies also to pain! The words ‘applies also’ evidently say that a distinctive kind of unity is available here: that which consists in seeing a group of items as falling under – or engaging application of – a single idea, form or pattern. In contrast, what the behaviourist needs to say at this point is that the relevant mind-characterizing notions can be analysed or explained in terms of other items. Otherwise put, the behaviourist is someone trafficking in the divisions emerging from TA. His problem is therefore to connect the different notions of ‘intention’ through a strategy of explanatory extension; and, in this at least, he will differ from other contemporary theorists only in taking for secondary those notions which they take for basic. But in whichever direction it runs, such a connective explanation will be something different from what Anscombe appears to contemplate: an apprehension of the divisions of intention as instances of a single form.21 This explains why the best evidence of behaviourist sympathies in Intention is bound to be inconclusive. For the best (of the thin) evidence must be what Anscombe says just after mentioning the possibility of pure intention: [T]his conspires to make us think that if we want to know a man’s intentions it is into the contents of his mind, and only into these, that we must enquire; and hence, that if we wish to understand what intention is, we must be investigating something whose existence is purely in the sphere of the mind; and that although intention issues in actions, and the way this happens also presents interesting questions, still what physically takes place, that is, what a man actually does, is the very last

140

Anscombe on Expression of Intention

thing we need consider in our enquiry. Whereas I wish to say that it is the first. (p. 9) ‘Some say A then B, but I say first B then A.’ No doubt, there is a genuine issue of starting points here, of what is to be modelled on what. Anscombe’s remark will suggest behaviourism, however, only to someone already sure of the theoretical options: Either the application of ‘intention’ to what people do will have to be explained in terms of its use to characterize a state of mind, or vice versa. There is another possibility, however. Say that Anscombe’s aim is to exhibit the unity of intention directly, by subsuming the three divisions under a single form. Once properly in view, it should be possible to see what this form applies to, without further complication. On this reading, ‘First B...’ says only that one does not come to see what formally unites the ‘interior thing’ with ‘intention in acting’ by asking first what properties characterize someone’s psychological state as one of intending; it is rather in action that the genus comes most perspicuously into view. While this might prove an unsatisfying thesis about intention, it isn’t exactly novel to suggest that members of a kind have asymmetrical powers of exhibiting it. If you want to see why bad arguments are arguments, for example, it is best to start by examining the valid kind, this being more basic. Or, again, successfully murdering is more basic than attempting the same. To see what kind of wrongdoing unites these, it seems best to start with the infringement of someone’s right in the completed act; and from here, one can discern the wrong in the more attenuated case, without the aid of the sort of theories which would have to say (starting at the other end) that, for example, the wrong in every act of murder is ‘really’ only that of someone’s intention to murder, plus some causal assistance from the world. Similarly, on the present thesis, what kind of entity or state this is – intending to write the word ‘action’ – will best be exhibited in the performance of writing the word ‘action’.22 This needn’t incur any commitment to the behaviourist’s denials. In support of this: Anscombe speaks of ‘intention in a proposed action’ (where TA is apt to write ‘intending’ or ‘future-directed intention’). Her point is that the distinction in question here is one between two intentions in action – items sharing an underlying structure or form, but differing in their positions along a spectrum of presence. On this view, a future intention to write the word action is structurally a variant of writing the word the action; it is distinguished only by its remoteness, by the intended intervention being, as it were, not-yet present. This will need further development: Just what Anscombe takes this common structure or form to be – the one best exhibited by intention in present action – remains to be articulated. But it shouldn’t be difficult to guess that the answer will have something to do with the attention she gives to her special question ‘Why?’ This is evidently the central ‘device’ (p. 80) in her ‘non-connective’ explanation of the unity

Richard Moran and Martin J. Stone

141

of intention. It also goes hand-in-hand, as will become plain, with her emphasis on ‘expression of intention’. Here, however, we reach ahead. To summarize, the divisions of ‘intention’ being connected in TA aren’t, on their face, Anscombe’s divisions. One can of course ask why Anscombe omits to discuss intending, and offers little analysis of this putative state at just the point – intention in a proposed action – where analysis seems needed. But this problem might also be turned around. Despite the standard homage, few theorists actually find much use for Anscombe’s first heading, at least as she inscribes it. Reading them, one would never suspect that ‘expression of intention’ figured centrally in her discussion.

III Why ‘Expression’: three clues Turning back to the problem of the prominence of ‘expression’ in Anscombe’s initial set-up, we find two explicit clues in Intention: (1) Ignoring the expression of intention in favour of what it is an expression of is apt to lead to wrong notions: ‘e.g., psychological jargon about “drives” and “sets”; reduction of intention to a species of desire, i.e, a kind of emotion; or irreducible intuition of the meaning of “I intend” (pp. 5–6)’. And (2), it is also apt to obscure just how different expressions of intention are from paradigm cases of expressions of states of mind. (1) Anscombe’s reference to Wittgenstein’s discussion of ‘I was going to....’ – a ‘pure’ intention again, only the time for action is past – pinpoints at least one ‘wrong notion’: namely that the truth or assertability of “I’m going [was going] to w” consists in some occurrent words, images, sensations, or feelings.23 Against this: A person’s emotions, desires or drives may lead in contrary directions without hint of irrationality; not so their intentions. In addition, one can have an intention over a period of time (e.g. to see a friend) while seldom thinking of it; and, in recalling an intention, whatever ‘scanty’ mental items memory presents as having occurred ‘do not add up to’ – they aren’t necessary or sufficient for – having that intention. (Nonetheless one can easily enough recall what one was going to do – one simply gives the words which express it.)24 All this suggests that (unlike, e.g. ‘I have an itch’ or ‘It feels like going down in a lift’), ‘I intend to ...’ neither reports an experience-content nor requires an experiential vehicle. To adapt a remark of Anscombe’s adapting a remark of Wittgenstein’s: No experience could be an intention, because no experience could have the consequences of intending.25 (2) Considering such paradigm cases of the ‘expression’ of an experience or state as ‘Ouch!’ ‘I’m in pain’, or ‘Foiled again!’, it might be said: Someone who tells us what they are going to do isn’t in that way ‘expressing’ an intention; they aren’t simply venting. Indeed, ‘I’m going for a walk’ no more looks like a communication about the speaker’s inner state than does,

142

Anscombe on Expression of Intention

say, her belief that the store closes at 8:00, as expressed by means of an appropriate declarative sentence. Only in special cases does an expression of belief aim to inform us about the speaker – that is, when it is about her. So too, when someone (actively) expresses an intention, Anscombe says, they give us information – right or wrong – about what is going to happen. (Of course, biographical facts will get passively expressed or manifested as well.) Herein lies one point of likeness with predictions: An expression of intention such as ‘I’m going for a walk’ is true or false according to whether the speaker goes for a walk.26 Hence, like a prediction, it can serve to give someone direct knowledge of what is going to happen.27 It is not a mere indication of the speaker’s psychic condition. These considerations – in brief, the dissimilarity of intention to a psychic state on the model of emotion – do suggest some motivation for Anscombe’s emphasis on ‘expression’. For a personal ‘expression of intention’ clearly makes a kind of claim, and has consequences, recognition of which will serve to place intention outside this model. But without more, these points suggest that Anscombe’s peculiar emphasis might be only a dispensable device for avoiding certain philosophical errors – ‘particular dead-ends’, as she calls them (p. 6). To see why these points do not go far enough, remember that the personal expression of belief is also to be understood as presenting a proposition of fact (an answer to the question ‘What did he/you say?’); it too is no mere indication of mental goings-ons. Yet, presumably, this isn’t likely to suggest that ‘expressions’ will be a main division in studying belief; that wrong notion, should it arise, could simply be dealt with directly. Why then – apart from the pro-active avoidance of errors – should ‘expressions’ figure as one of the main divisions of the topic of intention? That was the problem (§1), and the problem remains. The main clue needed here must evidently concern features of intention which are special to it, and which therefore generate a special problem. And there is such a clue, in the low-level linguistic facts which suggest divisions of the topic in the first place. What matters most is just that applications of ‘intention’ are spread along a spectrum extending from what is, naively speaking, ‘in the mind’ to what is ‘in the world’. Picturing a line, on the far left will be found pure intentions, defined as cases in which the agent intends to do something but hasn’t yet done anything else in order to do that. Moving rightward, the agent has more worldly deeds to show for his intentions: If he is described as Ø-ing or as intending to Ø, then, at this point, it will be correct to say that he is also doing various things in order to Ø, or because he intends (wants, aims) to Ø. At the far right, his performance is fully unfolded and finds description in the past tense: ‘He Ø-ed’ (or ‘He has Ø-ed’) will now be true, and not merely – what holds anywhere between these end points – ‘He was Ø-ing.’ The special problem, as all agree, is that of exhibiting the unity of the notions of intention which appear here.

Richard Moran and Martin J. Stone

143

This progressive structure has no parallel in the case of emotions, desires or beliefs. Generalizing this contrast, we may say: It begins to specify what is meant by a ‘state’ to note that the progressive form of the relevant verb – for example, ‘to believe’, ‘to be hungry’, ‘to be taller than’ – isn’t used. To join a subject to such a verb, no grammatical discriminations of progressive versus perfected aspect (‘was Ø-ing’ vs. ‘Ø-ed’) are needed, only those of tense; and there is always a good inference from the present (‘X is hungry...’) to the perfect (‘X was hungry...’ said at some future time). States, in short, are static; and this exhibits, by way of contrast, two related features of the spectrum of uses of intention. First, with anything an agent does, there are, in principle, any number of purposive sub-parts: things he must do in order to do that. This is because his performance takes time:28 It may be done quickly (taking little time) or slowly, and we can mark its progress – and thereby distinguish further points in our spectrum – by speaking of someone’s ‘just starting to Ø’, ‘being nearly halfway done’, ‘just finishing up’ and so on. Performances unfold: They involve a diminishing future and a swelling past, of what the agent needs to do in order to do, or have done, what he is doing. States, in contrast, merely go on for a time, without unfolding: For example, no one will be found slowly desiring a drink, or almost finished being the tallest boy scout even if his desire is soon to be quenched, or others are about to grow.29 Second, the intentional object of ‘He intends...’ is given by a performance verb (‘to Ø’), so that if someone intends to Ø, then what fits this notional ‘state’ of his is another performance to which the spectrum of the concept of intention in principle applies (viz., his Ø-ing, or his Z-ing in order to Ø), and not, apart from this, any extensionally equivalent state of the world. Hence the object of the attitude of intention is another intention or a performance, whereas what makes a belief true (or satisfies a desire) can be propositionally rendered, and is only in special cases characterizable through another application of these same concepts: Not all beliefs are about other beliefs, as not all desires could be merely for other desires. And whatever their causal contribution to a particular action, beliefs and desires do not come to serve as qualifications of the action itself (e.g. as ‘beliefish’ or ‘desirous’). Here, then, it looks as if ‘intention’ earns its literal, archaic sense – a ‘stretching forward’ – for it is structured as action itself is. Not that this is surprising: It is location on a spectrum of unfolding action which fits an event to be described using concepts of intention in the first place;30 the kind of thing an intention is to be explained in terms of a concept which applies throughout the spectrum. Anything this wasn’t true of – which didn’t characterize both a person’s attitude as well as the object of that attitude – wouldn’t be our (i.e. a) concept of intention. Anscombe asks: Would intentional actions still have the characteristic ‘intentional’, if there were no such thing as expression of intention for the future, or as

144

Anscombe on Expression of Intention

further intention in acting? I.e. is ‘intentional’ a characteristic of the actions that have it, which is formally independent of those other occurrences of the concept of intention? (p. 30)31 And she answers: ‘This supposition [that intention only occurred as it occurs in “intentional-action”] ... carries a suggestion that “intentional action” means as it were “intentious action”... that an action’s being intentional is rather like a facial expression’s being sad.’ This remark traces the consequences of losing the unity of the concept of intention across its different contexts: Cut ‘intention’ loose from its unfolding on a spectrum, consider it only as a qualification of action, and ‘intentional’ becomes ‘intentious’, essentially the name of a state, like sad or angry. By the same token, if TA has recently tended to discover that an adequate account of action must go beyond states of ‘having reasons’ conceived as propositional belief – desire pairs, what it has caught sight of may be structurally expressed thus: No psychology will afford the right materials for explaining action which does not make use of a concept which applies throughout the spectrum of unfolding action, and which thus has same internal complexity as actions themselves. This said, the image of a performance as a line touching points of ‘pure’ and ‘perfected’ intention is to be taken with a grain a salt. What is represented here is not a performance, but only the kind of thing a performance is, its structural possibilities. A few disclaimers will help to bring the spectrum into clearer focus: First, no implication arises that every action touches the right-most point. To the contrary, as J. L. Austin has observed, a mark of any action is its exposure to the risk of failure or incompleteness.32 No inference is available from the progressive (‘He is Ø-ing’) to the perfect (‘He Ø-ed’). Likewise, no implication arises that every action begins at the left-most point. Not all intentional action is the execution of a prior intention. Many intentions – for example, to role out of bed in the morning, to change speed according to traffic – never exist apart from the things one does. Third, where there are pure intentions, no implication arises that they precede acting, except relative to some descriptions of the action in question. To illustrate, someone may have now a pure intention to build a tree house, but only – as is likely – because he is already underway with something else of which building a tree house is a part or phase: for example, he is raising his kids to enjoy the outdoors, or improving the property before he sells it so he can retire and finish writing his novel. Likewise, no implication arises that in achieving the right-most point, there remains nothing virtual or pure about what the agent is doing. To illustrate, even after our agent’s intention to build a tree house has become impure, the point arrives where he forms (what is now) a pure intention to

Richard Moran and Martin J. Stone

145

buy lumber; failure on this score will mean he never builds a tree house, and perhaps never does the things of which this was a part. From the above, it appears that wherever ‘purity’, as defined here, is used to characterize intentions, it must be possible to apply the term in the same sense to actions themselves. Consider how it is with our agent who, far along in building his tree house, forms an intention to drive in a nail to fasten two boards together. At the moment, his intention is a ‘pure’ one, for he has taken a break and lies on the grass. What is this agent, with his nailand plank-regarding intentions, now doing? Well, among other things, it would be correct to say he is building a tree house (he has been at it since winter), only, at the moment, this action of his remains pure, for he is not now doing anything in order to do that. Any action of significant duration is apt to have moments of pure intending and pure acting among its innumerable parts. Notice that our agent’s nail-regarding intention may at present be pure (just like his action of building a tree house), though his intention to fix together two boards is impure (he has placed them next to each other) – all this, notwithstanding that hammering the nail, fixing the boards, and building a tree house are, as Anscombe teaches, the same action under different descriptions. So although pure intending comprises the left-limit case, and perfected action the right-limit one, this does not entail that what the person purely intends to do isn’t, under a wider description, an action which is already underway and uncertain of completion. Davidson’s own example of a pure intention – writing the word ‘action’ – illustrates the point: I have formed this intention because I’m already engaged in writing a sentence, and this with a view to writing a book, the second in a series, and so on. Further, once we are able to locate impure intentions in this way – by enlarging the frame of pure ones – it should naturally be possible also to find pure ones within impure ones, for the principle is the same. 33 If there is a tendency to think of action in terms of a one-way sequence, beginning with intending and moving through acting to having acted – that is, as something beginning in the mind and ending in a state of the world – an illusion is present which perhaps arises from the philosopher’s focus on unmotivated or point-like actions (the lifting of a finger) strangely abstracted from the vita activa. Now, returning to the question of Anscombe’s divisions, we venture the following thesis. What makes the phrase ‘expression of intention’ look like an outlier in her list – its reference to speech – is also what allows it to stand as a natural example of each division, and thus, in a sense, to subsume the entire list: (Case 1) Someone says ‘I’m buying some shoes’: An expression of intention for the future (perhaps even a pure one).

146

Anscombe on Expression of Intention

(Case 2) Someone says ‘I’m buying some shoes’: An intentional act (e.g. when the speaker is asked what he is doing). (Case 3) Someone says ‘I’m buying some shoes’: A further intention in acting (e.g. when this is the answer to the question ‘Why are you walking out the door?’) Not that this is unexpected, as if ‘expression of intention’ should somehow be proprietary to the special case of pure intending. The spectrum of intention – which exhibits that as a limit case – already suggests this wouldn’t be so. And consider the linguistic facts: One who offers, for example, ‘I intend to take the train to Boston tomorrow’ (a pure intention, let us suppose) might also express themselves thus: ‘I am taking the train to Boston – tomorrow.’ And having begun to pack, they might explain what they are doing by reference to the partial presence of that larger performance – ‘I’m taking the train to Boston’ – but they might also revert to the notionally ‘psychic’ explanation: ‘I intend to take the train to Boston.’ If any of these are ‘expressions of intention’, surely all must be. The principle of this linguistic sharing of labour is seen, according to Anscombe, when the use of the progressive is denied: [T]he less normal it would be to take the achievement of the objective as a matter of course, the more the objective gets expressed only by ‘in order to’ E.g., ‘I am going to London in order to make my uncle change his will’; not ‘I am making my uncle change his will.’ (p. 40) Likewise, the less normal it would be to take the achievement as a matter of course, the more the objective (‘in order to Ø’) gets expressed with ‘because’ followed by a proposition which couples the performance verb (to Ø) with some overtly psychic form (‘I intend ...’, ‘I want ...’, ‘I plan ...’, etc.): ‘I am going to London because I intend [want, will try, etc.] to make my uncle change his will’. Where ordinary language demands them, such ‘psychic’ expressions are markers of absence or remoteness in an unfolding performance. But in this, they are exactly like the idiomatic use of the simple progressive (‘I am Ø-ing’), which itself always conveys imperfect aspect, incompleteness and the risk of failure (only to a different degree). All expressions of intention, then, are such as to explain action by locating it within a larger action in progress. Where ‘psychic’ markers of remoteness are available, however, we can expect the use of the simple progressive to be correspondingly confined to a narrower range of cases exhibiting relative presence or proximity.34 This explains a critical remark of Anscombe’s. After (1) connecting the notion of intentional action to the applicability of a special question ‘Why?’, and (2) showing how an answer can furnish terms for a wider description of what an agent is doing (i.e. just when the question has the right sense), and (3) noting that this leads, chain-wise and eventually, to

Richard Moran and Martin J. Stone

147

‘a break’, to an act-description characterizable, on account of remoteness, only as something the agent is now going (or – we may add – intending, wanting, planning, etc.) to do, she writes: I do not think it is a quite sharp break. E.g., is there much to choose [i.e., in answer to why someone is putting on the kettle] between ‘She is making tea’ and ‘She is putting on the kettle in order to make tea’ [and, we may add, ‘She intends to make tea’? – our note] – i.e., ‘She is going to make tea’? Obviously not. (p. 40) The implication of this is that the notionally mind-characterizing uses of intention differ from the others only in articulating a greater degree of remoteness or uncertainty. That is, they continue – at a different point in the spectrum – the same form of explanation which was most perspicuously seen in the earlier links of the chain: essentially, the fitting of an action into a larger, presently incomplete whole. Call this the teleological structure of action: Action is the kind of thing which rationalizes its sub-parts (those actions done ‘in order to’ do it). As Michael Thompson has argued, this suggests that it is fundamentally intentional actions in progress which explain actions, and that it is only on the basis of this primitive structure – an action as a space of reasons – that a more sophisticated development becomes possible: namely, the joining of a psychic expression with a performance verb to create an etiolated form of the same structure in the interest of articulating relative non-presence, remoteness or uncertainty.35 In support of this, it bears remembering that the psychological items under consideration are someone’s intending, planning or wanting to do something (to Ø) and not, say, someone’s wanting or desiring that something or other happen or be the case. The progress noted earlier in TA towards more committed or articulated psychological states (beyond belief and desire) was just the recognition that only states of mind having a role in the unfolding of action itself – which means only ‘states’ subject to qualification (i.e. to placement on the spectrum) as ‘pure/impure’ – can effectively explain action by reference to the agent’s attitudes.36 But this just means that such states – intending to, planning to and so on – inherit through their objects (a performance-form: to Ø) the distinctive teleological structure characteristic of intentional action. And this suggests that no matter how rich a psychology we employ, we do not attain to an understanding of the behaviour-characterizing uses of ‘intention’ on the basis of relation to a psychological state, except by helping ourselves to a notion of a ‘state’ informed by a prior understanding of the concept of intentional action (as in intending or planning to do something). To summarize: ‘I intend (plan, want, etc.) to make tea’ expresses the same incompleteness of action seen – at earlier points in the chain of answers, where achievement is more ‘a matter of course’ – in the use of a performance

148

Anscombe on Expression of Intention

verb by itself: e.g., ‘Why are you boiling water?’ –‘I’m making tea’, ‘And why are you making tea?’ –‘I intend to serve the guests.’ Hence all such psychic forms are performance modifiers: Insofar as they are employable in actionexplaining answers to the question ‘Why?’, they express forms of being onthe-way-to-but not-yet having Ø-ed, of already stretching oneself towards this end. So expressions of intention ‘for the future’ are variations on a common theme: They are structurally of a piece with the simple progressive, only further to the left on the spectrum, where they permit a more refined articulation of the imperfection – an agent’s being committedly underway – informing all action explanation. In expressing imperfection, they are just like any other suitable answer to the question ‘Why?’, whether this be a declaration of what one will do, a description of what one is currently doing, an account of the intention with which one is doing something, an explanation of what one desires or is trying or endeavouring to do, or – think of Anscombe’s shopping list (p. 56) – a specification of what one is to do. The logical relations between these expressions of intention aren’t themselves very important. As a matter of ordinary language, they admit of no exact relative placement on the spectrum, though they are of course open to the stipulations of the philosopher whose purposes require more precision.

IV Expressions of intention: the general and the special use Conceived as an answer to the question ‘Why?’, ‘expression of intention’ is clearly a capacious notion (any item on the spectrum can be represented as such an answer),37 and it is this capaciousness (the power to exhibit what is structurally common to any use of intention) which recommends ‘expression of intention’ for Anscombe’s purposes. On this conception, there could be no question of finding out what an expression of intention is by first investigating the properties which characterize a psychological state as one of intention, as if the powers this state might have to rationalize what someone does might then remain open for investigation. Anscombe’s focus makes us take things the other way around: What can and does count as an agent’s ‘expression of intention’ is determined by its availability to enter into an elucidative account of action (fitting it into a whole in progress), its suitability, in other words, to be given to another who asks what is going on. An intention is whatever can be given to another in an expression suited to play this role. The special problem addressed by this account arises from the tendency for the continuity of uses of ‘intention’ to get disguised: ‘Intend’ looks like a state (like ‘believe’) and this sets the problem of its connection to ‘intentional’ as applied to things getting done. Indeed, by at least one criterion, ‘to intend to Ø’ should designate a state, since an inference from the present (‘He intends to Ø’) to the perfect (‘He intended to Ø’) does hold, and since, unlike a performance, intending to Ø does not – notwithstanding the colloquial use

Richard Moran and Martin J. Stone

149

of the progressive – take time.38 On the other hand, the fact that ‘intending’ (like any psychic verb put to employment in explaining action) takes a performance-form rather than a proposition as its object seems to mark a decisive difference. This is related to two features of ‘intending’ which would make it quite special among states, but which are commonplaces of performances. First, we can ask ‘why’ (for what purpose, with what intention) someone intends to Ø, just as we can ask this about action itself. (All intentions are like actions in this way: they are explained by – and they explain – other actions and other intentions.) Second, intending to Ø, being thus something voluntary, is subject to being commanded, like any action.39 Taking it that such differences are what matter (and not what label one applies), we propose the following account of the significance of ‘expression of intention’ for Anscombe. First, although Anscombe quickly leaves her first division (‘expression of intention for the future’) aside till later, a distinctive form of verbal exchange remains a pervasive feature of her exposition: ‘Why are you lying there?’ – ‘I’m doing Yoga’; ‘Why did you pump water this morning?’ – ‘To poison that lot, don’t you know?’; ‘Why worry about them?’ – ‘Those people have strangled the country long enough, I intend to get the good people in.’ All of the positive answers in such interrogations are ‘expressions of intention’, not just the last answer (an intention for the future): [I]f a description of some future state of affairs makes sense just by itself as an answer to the [‘Why’] question, then it is an expression of intention. But there are other expressions of the intention with which a man is doing something: for example, a wider description of what he is doing. For example, someone comes into a room, sees me lying on a bed and asks ‘What are you doing?’ The answer ‘lying on a bed’ would be received with just irritation; an answer like ‘Resting’ or ‘Doing Yoga’, which would be a description of what I am doing in lying on my bed, would be an expression of intention. (pp. 34–35) So ‘expression of intention for the future’ is one species of a common genus, that of positive or action-elucidating answers to the special question ‘Why?’ Recall the role of that question. It provides a definition of an ‘intentional act’ in terms of the applicability of a question to which the agent can always give some knowledgeable response. Given this, it should be possible to see that ‘expression of intention’ (in the wide, generic sense) works to exhibit the unity of the three, initially disparate-seeming divisions. This is just what Anscombe says: To a certain extent the three divisions of the subject made in § 1, are simply equivalent. That is to say, where the answers ‘I’m going to fetch my camera’ [expression of intention for the future], ‘I am fetching my

150

Anscombe on Expression of Intention

camera’ [intentional act] and ‘in order to fetch the camera’ [further intention in acting] are all interchangeable as answers to the question ‘Why’. (p. 40)40 On this account, the unity of the divisions lies both in the applicability of the question ‘why’ to the material of each, and in the suitability of each to itself rationalize action – that is, to figure in an agent’s answer to the question ‘why’, asked about something else he is doing (in the present example, his going upstairs). That is, the unity of the divisions is seen in the fitness of their corresponding linguistic expressions – ‘I’m going to Ø’, ‘I’m Ø-ing’, ‘ ...in order to Ø’ – to provide elucidatory responses in a special interrogation of the agent. ‘Expression of intention’ thus finds a narrower and a broader use in Intention: narrower, as one of the headings – ‘expressions of intention for the future’ – which brings the topic provisionally into view; more broadly, as the genus – comprising answers to the question ‘Why?’ – which formally unites the three divisions. This way of putting things seems to reverse Anscombe’s better-known formula, according to which an intentional action is one subject to the question ‘Why?’. But the possibility of such reversal is implicit in the chainlike structure of what Anscombe calls ‘the ABCD form’ (p. 45), whereby a positive answer to the question ‘Why?’ is itself the description of an intentional action, and, as such, subject to that question. Starting, then, with an intentional action, we can move forward along the chain by interrogating the action – ‘Why?’ For just when this question has the relevant sense, positive answers to it are themselves expressions of the agent’s intention.41 But, starting from the same point, we can also move backwards along the chain, by remembering that an action, suitably described, is something which can be an agent’s answer to the question ‘Why?’ (i.e. asked about something else he is doing). Anscombe’s ABCD form thus pictures intentional action as both ground and grounded, explanans and explanandum, something both subject to, and responsive to, a distinctive interrogation. And this is represented by saying that the unity of the three divisions lies in their fitness to appear in ‘expressions of intention’ in the generic sense: answers to the question ‘Why?’. Summarizing, Anscombe’s basic idea – her general strategy for discharging the explanatory task set in §1 of Intention – is this: (1) The applicability of the relevant question ‘Why’ is what marks anything out as an expression of intention (p. 90); (2) Any expression of intention thus subject to the question ‘Why’ is fit also to answer a serially related question ‘Why’; (3) The unity of the trinity – intention in a proposed action, intention in a present action and intentional action – is seen in this, that each is capable of being represented in an ‘expression of intention’, furnishing an answer to a question ‘Why’. Thus represented (as answers in the interrogation of action),

Richard Moran and Martin J. Stone

151

the three divisions are generally interchangeable: Usually, ‘there isn’t much to choose between them’ (p. 40).

V

‘In terms of language’

At this point, a contemporary action theorist might still ask: Why define intentional action ‘in terms of language’ (p. 86) – a chain of questions ‘Why?’ Obviously, the content of such exchanges – for example, ‘Why are you lying on the bed?’ – ‘I’m doing Yoga’ – might always be exhibited, apart from the interrogative context, in a variety of ways. For example, ‘A is lying on the bed because she is doing (she intends/wants to do) Yoga.’ Such actions can and do take place without anyone making any speeches!42 Wouldn’t it be less obscure to factor out reference to questions and answers, and speak instead of intentional action as (e.g.) ‘action for a reason’? ‘Expression of intention’, in that case, need not get any special emphasis. One answer to this challenge lies simply in the fruitfulness of Anscombe’s expository procedure in revealing unity in the otherwise diverse materials of action (cf. p. 80). Still, confidence in this procedure would improve if it became clearer how the specifically linguistic representation of intention really is an aspect of the thing represented. A further answer might run: ‘Intentions, like other conditions of the person, may intelligibly be attributed only insofar as they are publicly expressed.’ This is true enough. But Anscombe’s expository procedure involves ‘expressions’ of a quite special sort, those suited, we may now say, to be a person’s answers to a question addressed to him. No one will be tempted to think of ‘expressions’ in this sense as the only or canonical way in which various other conditions of the person get revealed. Is there a reason to think that intention is different? Return to that strange passage concerning animals (§ I). Here Anscombe does appear to say that intention is different in this respect, and she even pauses to correct Wittgenstein on this point. Wittgenstein had written: ‘What is the natural expression [Ausdruck] of an intention? – Look at a cat when it stalks a bird; or a beast when it wants to escape. (Connection with propositions about sensations.)43 But this ‘goes wrong,’ according to Anscombe: The expression of intention ‘is purely conventional’ or ‘linguistic’. Animals thus ‘have intentions’ but, lacking language, don’t express them.44 What makes this proposition strange is that while Anscombe (1) recognizes that intentions are manifest in non-verbal behaviour without further gloss;45 and (2) allows the term ‘expression’ application to the natural manifestation of other states (like emotion); she nonetheless (3) challenges Wittgenstein on this point: An animal’s behaviour isn’t any expression of its intention. Taking it that ‘expression of intention’ is a device for exhibiting the unity of ‘intention’,

152

Anscombe on Expression of Intention

this further problem is raised here: Is there a way of seeing Anscombe’s Proposition? AP: There is no such thing as the natural (only the linguistic) expression of intention as informative rather than merely stipulative? A positive answer would make clearer how the representation of intention ‘in linguistic terms’ comes, so to speak, with the matter itself. To begin with, however, a few words about what this question is not. As was said, questions concerning an animal’s capacities are not essential here. We are familiar in the human case with the distinctions – expression versus other indications, natural versus conventional expressions – invoked by Anscombe. Among us, there are both verbal and non-verbal expressions of fear, for instance. So if the expression of intention is purely linguistic, then it must be true of the man, and not just of the beast, that the movements of stalking a prey (or catching it and hauling it back to camp, etc.) are not to be considered ‘expressions’ of intention, notwithstanding that these movements make intentions manifest. Anscombe’s point applies to all animals if it applies to any. Similarly, questions concerning whether non-linguistic creatures can be credited with having intentions at all remain bracketed here. Negative judgments might seem to support Anscombe’s linguistic procedure, since they often derive from a focus on the formation of plans, or standing intentions – cases in which the basis for attributing intentions must go beyond natural behaviour.46 But not all intentions are standing ones or the result of making plans. And Anscombe’s challenge to Wittgenstein, in any case, does not concern some purer type of standing intention, but just the immediate intentional action of the cat stalking the bird. To speak of ‘expression’ in this case would be as inapt – she says – as calling ‘a car’s stalling the expression of its being about to stop’. Unobvious as this thesis is, it expresses no skepticism about the car’s being about to stop, or about the engine evincing this fact. Likewise, Anscombe nowhere denies (rather, she consistently asserts) that the structures of action uncovered in her investigation apply beyond the human world (pp. 86–87). Finally, it is worth noting that Anscombe’s remarks on animals run contrary to recent arguments which enlist one of her ideas – namely, that actions are intentional ‘under a description’ – in recommending the conclusion that neither intentions nor propositional attitudes are possible for creatures without a language. Davidson writes: One can intend to bite into the apple in the hand without intending to bite into the only apple with a worm in it [...] The intensionality we make so much of in the attribution of thoughts is very hard to make much of when speech is not present. The dog, we say, knows that its master is

Richard Moran and Martin J. Stone

153

home. But does it know that Mr. Smith (who is his master) or that the president of the bank (who is that same master), is home? We have no real idea how to settle, or make sense of, these questions.47 The intensionality of intentions does suggest that the capacity to recognize intentions in others is exclusive to the talking creatures, or those who interpret the talk of others. But if a creature does not recognize intentions, must it therefore have none of its own?48 Notably, Anscombe insists that whatever may be ‘language-centered’ in her account, and whatever the role of particular descriptions in delineating the intentional aspects of action, we do discern intentions in non-talking creatures: It sounds as if the agent had a thought about a description. But now let’s suppose that a bird is landing on a twig so as to peck at bird-seed, but also that the twig is smeared with bird-lime. The bird wanted to land on the twig all right, but it did not want to land on a twig smeared with birdlime. [...] Landing on the twig was landing on bird-lime – we aren’t considering two different landings. So, if we form definite descriptions, ‘the action (then) of landing on the twig’, ‘the action (then) of landing on a twig with bird-lime on it’, we must say they are definite descriptions satisfied by the same occurrence, which was something that the bird did, but under the one description it was intentional, under the other unintentional. That the bird is not a language-user has no bearing on this.49 We say the bird is intentionally landing on the twig (but not on the limetwig) because we can see that ‘landing on a twig to peck at bird-seed’ answers to purposes the bird is assumed to have, whereas ‘...a twig smeared with lime’ does not, even though we are talking about one and the same twig. We seek a description which makes what the bird did comprehensible within what we know the bird is seeking. A reference to the agent and its good does operate as a constraint here, but this need not be a reference to the agent’s psychology narrowly conceived, that is, to what is thought or said. Rather, we apply the descriptions under which the creature’s action is intentional, without supposing that the creature has such descriptions ‘in mind’. In sum, various questions concerning the attribution of intention to animals don’t shed light on Anscombe’s strictures on what an expression of intention can be. What could be at stake for her in refusing to apply the innocuous-sounding phrase ‘the natural expression of intention’ both to humans and other animals? That is the nub of the problem.

VI What natural indications can’t do: contradiction, commitment, impugning the facts One possibility should be ruled out. Anscombe had better not be found saying that by ‘expression’ she simply means expression in the personal

154

Anscombe on Expression of Intention

sense. From this it would of course follow that animals do not express intention, for to express intention (or anything else) you’ve got to use language. But this would be mere stipulation.50 Against this, let it be clear that anyone can call non-verbal manifestations of intention ‘expressions’ if they like – what to call things isn’t the question. The question is what distinction Anscombe seeks to mark by refusing application of ‘expression’ to natural manifestations of intention.51 It looks like the answer must be: some distinction between the verbal and non-verbal manifestation of (1) intention on the one hand, and of (2) emotion (and other states) on the other. But what distinction? If a creature’s stalking its pray is not to be considered an expression of intention, the interest of this thesis – what raises it above empty stipulation – must derive from a distinction we recognize here. But there is a distinction to recognize here: Between natural behaviour which manifests intention and the overt verbal expression of intention, there exists a gap in logical powers, without parallel in the case of emotion or other states. To explain, it seems natural to speak contrastingly of natural and conventional ‘expressions’ of, say, fear, precisely because, in this case, conventional expression can take up or perform the same work as natural expression – there is no gap. Otherwise put, the contrast (natural/conventional) is at home within a space of common functions, where verbal expression continues (while also enhancing and rendering more precise) the same functions more primitively available by natural means.52 The other’s frightened look and his ‘I’m afraid’ can convey the same thing – his fear. This isn’t to say that natural expressions of fear are given to another or meant to inform him. But conventional expressions of fear are not always given either.53 When they are, however – and this is what matters here – what they overtly convey is what might also be read in the speaker’s non-linguistic behaviour. Verbal expressions of intention, in contrast, do not ever stand in for intentional behaviour in this way. Certainly it would be wild to suggest that they are ‘learned as a substitute’ for intentional behaviour – one point subserved by the phrase ‘natural expression’ as applied by Wittgenstein to emotions and sensations.54 And as Wittgenstein himself notes, there isn’t any distinctive behavioural repertoire of intentional action, as there is of emotional states and feelings.55 No, the relation between verbal expressions of intention and intentional behaviour is different from the relation between (e.g.) verbal expressions of fear and natural fear-behaviour. It is this: When an intention is verbalized, it specifies the performance to which the agent is committed in the future, or in which he is already engaged, and that performance may then be judged correct or mistaken in light of what is expressed. So Anscombe is right to mark a difference here. Rather than standing in for performances in either a logical or developmental sense, expressions of intention have a force which no bit of natural behaviour could have.

Richard Moran and Martin J. Stone

155

Specifically, they make contradictable claims, and they require that something else one does then be regarded as correct or mistaken. Of course, we can sometimes see directly that someone is naturally barking up the wrong tree – say, merely running the brush along the wall when they mean to be painting the room yellow. But here it is important that we grasp a particular description of what they are doing or intending to do, one which a human being could give in answer to the question ‘Why?’56 To develop the implications of this, a number of contrasts, so far mainly implicit in our discussion, may now be made more explicit. The upshot is that a number of related features – contradictability, claiming, commitment and the impunging of performances – can be seen to make intention fit for a notion of ‘expression’ which lacks a non–linguistic counterpart. 1. Intentions are contradicted by other intentions. Midway through the book, Anscombe asks, ‘What is the contradictory of a description of one’s own intentional action?’ and she answers: The contradiction of ‘I’m replenishing the house water supply’ is not ‘You aren’t, since there is a hole in the pipe’, but ‘Oh, no, you aren’t’ said by someone who thereupon sets out e.g., to make a hole in the pipe with a pick-axe. And similarly, if a person says ‘I am going to bed at midnight’, the contradiction of this is not: ‘You won’t, for you never keep such resolutions’, but ‘You won’t, for I am going to stop you’. (p. 55) To contradict an expression of intention is intentionally to oppose the act which the agent declares herself to be engaged in – that is, what the agent is intent upon – rather than to assess the agent’s states, dispositions, or other conditions affecting the likeliness of performance. Verbal expressions of emotions and desires can also be contradicted of course, but what is denied, in that case, is the claim that the speaker has the item in question. Contradicting an agent’s expression of intention to ␸, in contrast, leaves no doubt – indeed, it presupposes – that she has the intention. 2. Like a belief, the expression of intention makes a claim upon the world. From (1), it follows that when Anscombe speaks of ‘expression of intention’ as something that can be true or false, this means a true or false claim about what one is doing or will do, not a good or a bad indication of one’s state of mind. Herein lies the aspect of intention which leads some theorists to see it as a kind of belief. If an ordinary assertion (‘The train just left’) is one canonical form of the expression of belief, its contradiction is the denial of what is claimed (‘No, it hasn’t left yet’), not the denial that this expression manifests the speaker’s state of mind. Expressions of both belief and intention make claims which can be countered by the denial that things are or will be as they are declared to be.

156

Anscombe on Expression of Intention

3. Yet, unlike a belief, the expression of intention can impugn what one does. Against this background, Anscombe distinguishes an expression of intention, not only in terms of the applicability of the question ‘Why?’ (p. 90), but also, initially, in terms of the impugning of performances: If I don’t do what I said, what I said was not true. [ ... ] But ... this falsehood does not necessarily impugn what I said. In some cases the facts are, so to speak, impugned for not being in accordance with the words, rather than vice versa. This is sometimes so when I change my mind; but another case of it occurs when, e.g., I write something other than I think I am writing: as Theophrastus says, the mistake here is one of performance, not of judgment.’ (pp. 4–5). Putting this in terms of the previous point concerning contradiction, we may say: To contradict a prediction (or other expression of belief) is merely to deny a proposition of fact; to contradict an expression of intention is also to oppose what someone is doing or will do. This specifies a discontinuity between expressions and mere behavioural ‘indications’ of intention. Anything which serves to indicate something (like the animal’s intention to escape) will itself be the thing faulted or impugned when it fails to conform to what was putatively indicated. Where there is a clash between an indicator and what it purports to indicate, it is the indicator which stands to be corrected. 4. Practical knowledge: expressions of intention are distinguished by the possibility of ‘mistakes in performance’ and not otherwise by their ‘direction of fit’. The last point touches on a larger theme of Anscombe’s: No statement will count as an expression of intention unless it expresses ‘practical knowledge’. If, for example, the speaker says ‘I’m going to crush the snail’ (or ‘I am crushing the snail’) on the basis of his observation of forces impelling him to move, then this is no expression of his intention. Now in speaking of expressions of intention as expressing practical knowledge, Anscombe might be taken to mean, in part, that the fit between such knowledge and what is known runs in the opposite direction than it does in cases of belief. This comes out when the speaker is not in fact doing what he takes himself to be doing. Here, if the speakers words are an expression of intention (rather than, e.g. a prediction), then the mistake will lie not in what the speaker thinks or says, but in what he is doing (‘the mistake here is one of performance not of judgment’).57 Understandably, then, an enduring legacy of Anscombe’s monograph is the idea of ‘direction of fit’. The only disappointing aspect of this legacy is that the phrase ‘direction of fit’ doesn’t actually occur at its source – and for good reason: While any talk of ‘knowledge’ must find room for application of an idea of fit or accord, the fit present in cases of practical knowledge (or

Richard Moran and Martin J. Stone

157

absent in cases of its failure) isn’t simply a matter of reversing the priority between the same two items which figure in cases of speculative knowledge; practical knowledge involves a distinctive class of items known. Unsurprisingly, then, Anscombe’s distinction bears little resemblance to the contemporary one between the functional roles of belief and desire, defined in terms of different directions of fit between such states and ‘the world’. Instead of ‘two directions’, Anscombe speaks of a difference between a mistake in what was said and a mistake in performance. What turns on this? Three things: (A) From expression of intention to desire. The ‘fit’ which concerns Anscombe involves conformity between an expression of intention – something said – and what the person actually does. In the contemporary functionalist account, by contrast, the focus is no longer on verbal expressions, but on states of belief and desire, the later being defined by the requirement of the world (or ‘the facts’) conforming to fit it rather than vice versa. (B) From action to ‘the world’. With this shift in focus from intentions to desires goes naturally a transformation in how the other side of the relation – the world – is understood as well. When Anscombe speaks of ‘the facts [being] ... impugned for not being in accordance with the words’, the facts in question are someone’s performance. But in the contemporary account, instead of a ‘mistake in performance’ – an action characterized as needing correction – we are referred to some state of the world itself (e.g. the absence of the cool drink I am longing for) that is to be altered, made to conform to the state of desire. (C) Disappearance of the notion of ‘mistake’. With the foregoing shifts in the nature of the relata (a person’s state rather than his statements, this state’s relation to some state of the world rather than to the person’s actions), the idea of ‘mistake’ or ‘correction’ disappears from the analysis. For there is no mistake on anyone’s part, mine or God’s, when something I desire is out of reach; and the facts comprising this aspect of the world are in no way impugned by my desire or its expression. Rather, my having this desire simply means, other things being equal, that I will strive to alter these facts in order to satisfy my desire. Given this, it is perhaps easy to see why the contemporary functional account of desire in terms of ‘direction of fit’ between a person’s state and the world – does not appear in Anscombe, though it is widely attributed to her. Very simply, it is difficult to see how we are to apply normative notions to either item. In contrast, the application of ‘mistake’ to actions and statements – things done and said – remains straightforward: These are the very things to which notions of mistake, correction, rightness and so on primarily apply.58

158

Anscombe on Expression of Intention

5. Commitment. Various theorists have seen in intentions a form of ‘commitment’, even though ‘I’m going to...’ is not a performative utterance, and the relevant notion of commitment is different from promissory commitment. (Anscombe’s denial that there are natural expressions of intention would be straightforward if expressions of intention were performatives, speech-acts dependent on conventional means). But although expressions of intention are not performatives, there is nonetheless a sense in which all intentions involve performances. For from (1) above – to contradict an intention is to oppose an agent’s action – it follows that an agent’s expression of intention is itself a kind of engagement to act, not a mere revelation of his state of mind. (Contradiction is symmetrical.) Self-contradiction in intentions exhibits the same pattern. If my intention to fly to Boston conflicts with my intention to stay home this weekend, this is not because these two intentions characterize my state of mind as one which dooms me to some frustration whatever happens (as happens with conflicting desires). It is rather that I engage myself in both doing and not doing the same thing. The way even pure intentions involve commitment is thus perhaps best understood in terms of the commisive aspect of action itself. Consider someone now engaged in writing the word ‘action’. That is how someone who merely intends to write the word ‘action’ is also engaged: He stretches himself toward the act, awaits himself in its successful completion.59 The structure of intending to act, this is to say, is that of a performance, and, as such, something continuous with intentional action itself. Of course, an intention for the future may be cancelled, blocked or otherwise never realized. Hence an objection arises here. What sort of ‘commitment’ is it which can be unilaterally rescinded by the agent, without penalty, simply by a change of mind?60 This argument proves too much, however. If it implies there is no sense of ‘commitment’ independent of obligation, it will follow that even someone now doing X – say, conducting a war in a foreign country – hasn’t committed themselves to anything. For this action may also be cancelled or blocked at any stage before its perfection. The completion of any extended action requires the agent’s continuing assent; as Sartre puts it, there always remains the possibility of ‘putting [the] act into question’.61 So the fact that an agent may change his mind does not distinguish the commisive quality of pure intending from those paradigmatic commitments which are his temporally extended actions. Remembering the idea of spectrum at work here, we therefore affirm that a good analogue of ‘intending’ can be found in someone’s doing something.62 Less metaphorically, the point is just that intention in a future action does not differ fundamentally from intention in (a present) action, or from intentional action. All alike, as Anscombe taught, are fit to be taken up in ‘expressions of intention’, conceived as answers to the question ‘Why?’. All alike are

Richard Moran and Martin J. Stone

159

engagements of agency, and enter into the structure of commitment, contradiction and impugning which characterizes performances as opposed to paradigmatic ‘states’. All involve an agent stretching towards a describable future which is not-yet.63 To summarize: With the verbal expression of intention a discontinuity with natural manifestation arises, and it is this which Anscombe seeks to mark. Such expressions introduce something new: a characterization of what one is doing – what larger action one’s actions are part of or toward which they are aimed. Expressions of intention are thus ‘world-directed’, but not just in the way that expressions of states like belief or hope are: They make possible the application of the notion of ‘mistake’ to performances, and they express practical commitments.

VII

Homage revisited

In conclusion, (1) three overlapping ways of making more precise the necessity of representing intention in terms of its verbal expression, and (2) a comment on one prominent theme. (1) First, it might be said that only verbal expressions of intention exhibit the responsiveness of the person to the special question ‘Why’. This question may apply to the bird pecking at the lime twig as much as to the man putting on his coat. The man, however, is in a position not only to answer the question, but also to refuse its application in a given formulation, as when he says that, as described in those terms, he was not aware he was doing that. (The bird is not expected to play any role in refusing application to the question ‘Why?’) Both acknowledging and refusing application of a certain question ‘Why?’ are part of the agent’s understanding of his action as goal directed. Refusing one application of the question makes the agent subject to some other application of it. (‘If you weren’t intentionally sawing the last plank of oak, then what were you doing? What did you take yourself to be doing?’) The answer to such a question gives the terms in which the action is to be seen as intentional (the point of the activity, the good of it, what is being pursued). The role of the refusal of application to a particular ‘Why?’ question marks the fact that an action will have conditions of success or failure only as described in certain ways and not others. In answering this question, an expression of intention spells out the aim that some piece of behaviour is guided by and to which the person is seeking to conform the rest of his action (i.e. the larger actions in Anscombe’s ABCD structure). Aims have a linguistic structure in this sense, that the objects we handle in the world (and the movements we make in handling them) are multiply describable, and only a fraction of these descriptions will be relevant to what makes the results something aimed for.

160

Anscombe on Expression of Intention

Second, only when intention is not merely indicated, but verbalized as a statement of what one is doing, do we see how its expression could be something in light of which a performance might be mistaken or corrected. For only a statement – a claim upon the world established independently of the regularities it seeks to track – could be in a position to impugn the facts. The content of an expression of intention will have to be conventionally or linguistically determined if it is to serve as an independent standard of this sort. Finally, only a verbal expression of intention can directly display the unity of ‘intention’ as it occurs in ‘intending to X’ and ‘an intentional action’. For what is needed here is essentially the notion of expression we see at work in the description of what one is doing as, for example, replenishing the water supply. When it comes to human actions, ‘the description of what we are interested in is a type of description that would not exist if our question ‘Why?’ did not’ (p. 83). Developing this formula, we might say: For every answer to the special ‘Why?’ question there is a complementary answer to a special question ‘What?’, applied to someone’s performance. That is, because ‘intentional’ applies to action itself (and not just to something in the mental history of the agent causing action), a true and positive answer to the question ‘Why?’ tells us not merely why some event is taking place, but also what is happening, in terms of an action being performed. Thus, building on Anscombe’s remark that there isn’t ‘much to choose between [the answers] “She is making tea” and “She is putting on the kettle in order to make tea’ – i.e., ‘She is going to make tea’?,’ we might add: There often isn’t much to choose between the questions ‘Why are you messing about with the kettle’ and ‘What are you doing messing about with the kettle?’. The answer to either question will be an expression of intention in the sense canvassed here in terms of the predicates of ‘contradictability’, ‘commitment’, ‘world-directedness’ and ‘mistake in performance’. That the concept of intention applies across the performative spectrum, on the one hand, and that the answer to the special question ‘Why?’ tells us also what is happening, amount here to the same thing. In action, the What appears together with – is already made for – an answer to the question ‘Why?’. This serves to distinguish action from other phenomena in nature, where the identification of <What is to be explained> is independent of how its interrogation in terms of ‘Why?’ may turn out.64 (2) Anachronistically put, Anscombe shows what is at best optional in the contemporary view that we understand what intention is only by asking what the mental state of intending is and how it could causally contribute to the production of intentional action. In place of this, she stresses the conceptual unity of a certain trinity: a single concept at work along a spectrum of cases, including ‘pure’ intending, intentions in action and intentional action. There is irony in the fact that one of her aims was to break up the sense of sharp distinctions among the divisions she is credited

Richard Moran and Martin J. Stone

161

with having discerned. There is also a danger that recognition of this point will involve attributing to Anscombe the thesis that ‘intention’ isn’t a mental state at all, as happens in the behaviourist reading. But what Anscombe denies is only that we understand how to apply the notion of a ‘state’ here, on the basis of its application in other contexts like those of belief and desire. In fact, TA inches towards the same conclusion, for its more recent discoveries might be represented like this: No psychology will afford the right materials for explaining action which does not make use of a concept which applies throughout the spectrum of unfolding action, that is, which has the same internal complexity as action itself. Intuiting this, it becomes natural, within the framework of TA, to seek an enhanced psychology of states, one which might include such items as intending, planning or even believing or wanting in some special sense. No doubt, there are such states of mind, in the anodyne sense that human agents do intend, plan, want and so on. But the point to grasp is that leading one’s sense of the psychological materials needed (for connecting the various uses of intention) is a prior understanding of what intentional action is. To make that understanding explicit was Anscombe’s problem. To elaborate, suppose the contemporary action theorist inspired, by the felt necessities of his material, to introduce a special psychological state called ‘intending’ or ‘planning’ (call it X-ing: the name does not matter). His avowed task is to analyse it and to explain the behaviour-involving uses of intention on the basis of their relation to it. But if X-ing is even to seem to be fit for this employment, it had better admit of internally nested relations (parts and wholes) of the sort exhibited in explanations like, ‘A is X-ing to Ø because he is ␸-ing’ (A is planning to buy lumber because he is building a tree house: the ‘state’ of X-ing explained in terms of a larger whole of action.) ‘A is X-ing to Ø because he is X-ing to C’ (A is planning to buy lumber because he is planning to build a tree house: the ‘state’ of X-ing explained by reference to a different occurrence of X-ing, one directed upon a wider description of the action.) ‘A is V-ing because he is X-ing to Ø’ (A is getting his car keys because he is planning to buy lumber: an action explained by reference to the ‘state’ of X-ing.) The nature of the pressure to endow action theory with an enhanced psychology of states can be represented concisely, in Anscombe’s terms like this: However we wish to understand the relevant action-theoretic state of X-ing, it ought to be the sort of thing about which it makes sense to ask ‘why’ (i.e. for what purpose) one is X-ing-to-do-something, and to answer

162

Anscombe on Expression of Intention

this by reference to other things one is X-ing-to-do or other things one is doing. And of course the expression of this ‘state’ will have to partake of the structure of contradiction, action-characterization and correction of performances (as identified through X-characterizing descriptions). ‘Intending’ and ‘planning,’ in idiomatic employments, do fit this special bill of requirements. But, as our discussion should make clear that is because the basic psychological item needed must be: X-ing to do something. That is, X-ing, whatever it is, must inherit through its object (a performance), just the distinctive structure characteristic of intentional action. And as Anscombe’s problem was just to make that basic structure explicit, this suggests a route by which TA – after discoveries about the enhanced psychology of states it requires – might at length pay a more unqualified and accurate homage to Anscombe.

Notes A special thanks to Mathew Boyle and Doug Lavin for a series of conversations about action at the Garden of Eden in Boston, and for their comments on an earlier draft; to Jonathan Lear, Candace Vogler and Constantine Sandis for their comments on an earlier draft; to David Velleman, for his comments at the 2006 Central Division APA, and to members of workshops at University of Chicago, University of Pittsburgh and University of Southern California, where this essay was discussed. 1. Sartre, Jean-Paul (1956), Being and Nothingness, translated by Hazel Barnes (1993) (Washington: Washington Square Press) p. 74. 2. Wittgenstein, Ludwig (1980), Remarks on the Philosophy of Psychology, Vol. I, translated by G.E.M. Anscombe (Oxford: Blackwell) §830. 3. Or a ‘further intention in acting’. These divisions play a structuring role in Anscombe’s discussion. They explicitly appear at p. 1 (statement of the headings), p. 9 (the transition from ‘expression of intention’ to ‘intentional action’), p. 40 (the unity of the three divisions) and p. 90 (return to ‘expression of intention for the future’). And they are drawn on elsewhere in Anscombe’s discussion. 4. ‘Where we are tempted to speak of “different senses” of a word which is clearly not equivocal, we may infer that we are in fact pretty much in the dark about the character of the concept which it represents.’ Intention, p. 1. In this essay, we approach Anscombe’s ambitions in Intention in light of her sense of why the concept of ‘intention’ calls for philosophy at all. The answer evidently refers to a submerged unity in our otherwise familiar employments of ‘intention’. What makes these all cases of ‘intention’ does not immediately appear. 5. Cf. Intention, p. 5. 6. On this distinction, see Richard Moran, ‘Problems of Sincerity’, Proceedings of the Aristotelian Society, Vol. 105, pp. 341–361 (December 2005). 7. Or future statements of fact – for example, ‘Nurse will take you to the operating theatre’: p. 3. 8. With Anscombe’s discussion of expression of intention versus prediction, compare, for example, Wittgenstein, Philosophical Investigations, p. 224. 9. A similar claim – though perhaps on different grounds – is made by Keith Donnellan, ‘Knowing What I am Doing’, The Journal of Philosophy, Vol. 60, No. 14 (1963), pp. 401–409 at p. 409. But Anscombe’s passage is obscure, even by

Richard Moran and Martin J. Stone

10. 11.

12.

13.

14. 15. 16.

17. 18.

19. 20.

163

Anscombian standards. The term ‘expression’ is idiomatically multivalent enough to embrace Wittgenstein’s talk of natural behaviour as ‘expressions’ of intention. So it is hardly obvious that what Anscombe wishes to deny must be the same thing Wittgenstein is asserting. Further, Anscombe herself, turning to the question of ‘how ... we tell someone’s intentions’, will point out that intentions are often legible in someone’s behaviour: ‘You will have a strong chance of success [at this] if you mention what he actually did or is doing’ (pp. 7–8). This only makes more pressing the question of what could be stake in her denial that natural behavioural manifestations of intentions are proper ‘expressions’ of it. See §§ V–VI below. Donald Davidson, ‘Introduction’, Essays on Actions and Events (Oxford: Oxford University Press, 1985), p. xiii. Michael Bratman, Intention, Plans and Practical Reason (Cambridge: Harvard University Press, 1987), pp. 1, 3. Bratman’s footnote attributes this distinction between ‘mind-’ and ‘world-’ characterizing uses of intention to Anscombe. Ibid., p. 5. Cf. H.L.A. Hart, ‘Intention and Punishment’, in Punishment and Responsibility (New York: Oxford University Press, 1968), p. 117: ‘Intention is to be divided into three related parts ... The first I shall call “intentionally doing something”; the second “doing something with a further intention,” and the third “bare intention” because it is the case of intending to do something in the future without doing anything to execute this intention now’. ‘Reluctantly’: Cf. Davidson, ‘Intending’ in Essays on Actions and Events, p. 88. If reasons, conceived as belief/desire pairs, seem adequate to explaining what it is to act intentionally, they appear immediately hopeless when it comes to ‘intending to act’. One main problem is that the familiar conflicts which are present among an agent’s desires seem intolerable when it comes to her intentions; intentions seem to ‘commit’ the agent in a way which mere desires do not. This and other difficulties with belief/desire psychology in the theory of intention are discussed by Bratman in Intention, Plans and Practical Reason (Ibid.). On the apparently commissive aspect of ‘intending’, see Bratman’s discussion on pp. 4–5, and our discussion infra, §VI. Davidson’s reluctance, as opposed to Bratman’s boldness, will comprise only a superficial difference here. For the reluctance is only about recognizing ‘intention as a sui generis state of mind (one not ‘ontologically reducible’ to beliefs and desires: see Davidson ‘Intending’, pp. 88, 83). Both take for granted, however, the explanatory framework described in this section. See infra, pp. 144, 147, 161–162. We shall generally use ‘intending’ or ‘intention to act’ for this category. ‘It would be astonishing if that extra element were foreign to our understanding of intentional action’. Davidson, Intending, p. 88; and see p. 89: ‘[T]here is no reason not to allow that intention of exactly the same kind is also present when the intended action eventuates’. See Davidson, Essays on Actions and Events p. xiii, and ‘Intending’, pp. 87–88. See Bratman, op. cit., pp. 5–6. Anscombe says almost nothing in Intention about whether action explanations mention causes, save for an occasional suggestion that the relevant notions of ‘a cause’, and ‘causal’ would have to be made clearer for us to understand what this question is about. See e.g. §§ 5, 9–11. Bratman, Intention, Plans and Practical Reason, p. 7. Nor can we understand this on the basis of the agent’s having done something for such a reason (say bought a ticket); for, by hypothesis, the intention in question may remain pure, the action merely proposed.

164

Anscombe on Expression of Intention

21. Cf. Intention, p. 84: ‘[T]he term “intentional” has reference to a form of description of events’ (Anscombe’s emphasis). 22. One feature of this analogy seems worth highlighting. Someone who, endeavouring to understand ‘wrongful’, started with the case in which it characterizes an agent’s plan would find themselves having to connect this to other cases through a story about such a plan bringing about –’in the right way’ of course – a situation in which a different, world-involving notion of wrongful finds application (e.g. to another’s loss or injury). (And they might feel puzzled over how applications of the world-involving notion of ‘wrongful’ – for example, wrongful injury – could involve a greater degree of culpability, given the fortuitous role which factors beyond the agents control are bound to play in this extended story; but this is a distinguishable problem.) Starting at the other end, one sees what makes a plan wrongful by seeing what happens when it succeeds. This does not occlude the possibility of causal explanations, but it suggests that the unity of the divisions can and perhaps must be grasped before they get under way. 23. See Ludwig Wittgenstein, Philosophical Investigations §§629–660 (esp. 635, 645–646), pp. 216–217; Zettel 44. Anscombe alludes to this discussion on p. 6 of Intention. 24. For a related discussion, see Anscombe, ‘Events in the Mind’ in Metaphysics and the Philosophy of Mind: Collected Papers Volume II, esp. pp. 57–61. 25. Cf. Intention, p. 77 on the absurdity of accepting both the empiricist idea of pleasure as an impression of some kind and seeing it as ‘quite generally the point of doing anything’. Anscombe is adapting Investigations §218: ‘Meaning is not a process which accompanies a word. For no process could have the consequences of meaning’. 26. Cf. Intention, p. 92: ‘Nor can we say: But in the expression of intention one isn’t saying anything is going to happen! Otherwise, when I had said “I’m just going to get up,” it would be unreasonable later to ask “Why didn’t you get up?.” I could reply: “I wasn’t talking about a future happening, so why do you mention such irrelevancies?” ’ On Anscombe’s account, an expression of intention differs from a prediction in not being founded for the speaker on evidence or observation, as well as in the particular notion of ‘mistakenness’ we apply in connection with it. 27. ‘Direct’: that is, knowledge not merely on the basis of an inference from how it is with the speaker. Cf. Intention, p. 3: ‘[N]or does the patient normally infer the information from the fact that the doctor said that; he would say that the doctor told him’. 28. See Michael Thompson, ‘Naive Action Theory’, in Life and Action (Cambridge, Mass: Harvard University Press, (forthcoming). Our presentation of the spectrum of intention owes a general debt to Thompson’s paper. 29. On the contrast between states and performances (and taking time vs. going on for a time), see A. Kenny, Action, Emotion and Will (London: Routledge and Kegan Paul, 1963), Chapter 8, to which we are indebted here. 30. The use of intention which arises in considering how to understand a speaker (especially when the speaker is absent, for example, – ‘the author’s intention’) would appear to call for separate treatment, as it does not involve a notion of unfolding action. In light of this case, it appears that Anscombe’s divisions are neither exclusive nor exhaustive. 31. Anscombe is exploring here the disabling consequences of isolating one of her divisions – intentional action – from the others. However, her overall point is

Richard Moran and Martin J. Stone

32. 33.

34.

35. 36.

37.

38. 39.

40. 41.

42.

43. 44.

165

more general: Other distortions would arise from the isolation of any of the divisions. See J.L. Austin, ‘A Plea For Excuses’, in Philosophical Papers, ed. J.O. Urmson and G.J. Warnock (Oxford: Clarendon, 1961). Thus, someone writing the word ‘action’ may have a pure intention, at this moment, to write the letter ‘c’. But asking why they intend to write ‘c’ will disclose this intention to be grounded in an action already underway: the pure intention to write ‘c’ stands to writing the word ‘action’ (once they have begun on the ‘a’) as writing the word ‘action’ does to writing the sentence ‘An intention can exist ...’ and so on. So pure intentions are everywhere. But that is because actions, underway but not yet completed, are everywhere; such actions are their grounds. See Michael Thompson, ‘Naive Action Theory, op. cit.; on the ‘openness’ of the progressive, cf. Kevin Falvey, ‘Knowledge in Intention’, 99 Philosophical Studies, pp. 21–44, esp. p. 26 (2000). Thompson, ‘Naive Action Theory’, op. cit. Michael Bratman’s master term – planning – fits the bill of particulars here. Unlike beliefs and (appetitive) desires, plans are wholes which rationalize their subparts, can be pure or impure, take a performance from as their object (‘I plan to Ø’), are subject to the question ‘why’ in the relevant sense (‘for what purpose?’), can be commanded (‘Plan to be there at 4:00!’) and so on. Planning, in short, shares in the structure of action. Nothing this wasn’t true of would even seem to be a good candidate for an intention-explaining psychological state. The instructive exception to this is the right-limit case, which, by definition, does not express imperfection and hence does not enter into action-explanation: For example, in answer to ‘Why?’ never the past-perfect (I Ø-ed), but only the past progressive (I was Ø-ing). See the previous discussion above, p. 143. This is sometimes missed because there is no imperative-form (‘Intend to Ø!’). However, nothing is easier than making someone’s intention – for example, to return the book – the object of a command: One simply orders them to return the book. It might be objected that, were intending really an etiolated form of performance, it ought to be possible to command someone to ‘intend to return the book’ in perfect purity, that is, ‘Plan to return the book, but don’t actually return it!’ But the answer to this is that it is also impossible to command someone ‘to go ahead with returning the book without actually returning it’. Intending to Ø stands to the progressive Ø-ing just as Ø-ing (or doing things in order to Ø) stands to successfully Ø-ing or having Ø-ed: None can be commanded apart from the others. The italics are our notes. The relevant sense of the question ‘Why’, and the fact that positive answers to it are themselves expressions of intention, are mutually defining notions for Anscombe. Cf. Intention, p. 80: Like Aristotle’s ‘practical reasoning’, the ‘order of questions “Why”? can be looked at as a device which reveals the order’ in the diverse materials of action. But Anscombe also points out: It is ‘as artificial as Aristotle’s [construction]; for a series of questions “Why?”... with the appropriate answers, cannot occur very often.’ Wittgenstein, Philosophical Investigations, §647. This point appears twice in Intention: pp. 5, 86–87.

166

Anscombe on Expression of Intention

45. Cf. p. 8: ‘[I]f you want to say at least some true things about a man’s intentions, you will have a strong chance of success if you mention what he actually did or is doing.’ If intentions were not so legible, their personal expression would lack much of its point. Such expression gives another warrant to expect behaviour of a certain describable shape, and this implies a general capacity to recognize another’s behaviour, when the time comes, as satisfying (or frustrating) those expectations. 46. See, for example, Stuart Hampshire, Thought and Action, pp. 97–98. Relatedly, it is sometimes said that only linguistic creatures initiate action by deciding on it from among a range of alternatives. And, no doubt, we speak of ‘decisions’, just where someone has, in effect, answered a question, or resolved their intentions against the background of other prospects. But these points are moot here, for the reasons explained in the text. 47. Davidson, ‘Thought and Talk’, in Truth and Interpretation, p. 63. 48. Indeed, Michael Tomasello’s reading of the empirical research on non-human primates defends just this position: While brutes have intentions, they fail to recognize intentions – and hence to distinguish ends from means, and both from the upshots of what is done – in others. See The Cultural Origins of Human Cognition (Harvard: Harvard University Press, 2001). 49. Anscombe, ‘Under a Description’, in Metaphysics and the Philosophy of Mind, pp. 208–219. This is in further explanation of her remark in Intention, pp. 86–87: ‘[W]e certainly ascribe intention to animals. The reason is that we describe what they do in a manner perfectly characteristic of the use of intention concepts: we describe what further they are doing in doing something ... . We do this although the cat can utter no thoughts, and cannot give expression to any knowledge of its own action, or to any intentions either.’ 50. This is just what is happening, according to Donald Gustafson, one of the few commentators to notice the problem. Concerned mainly to defend Wittgenstein, Gustafson criticizes Anscombe for overlooking this: ‘[T]hat a person’s face has a determined look, while it does not imply that he expressed determination in the sense that his saying he is determined does so, is [nonetheless] an expression of determination or a determined expression’ The implication must be that AP joins no issue, since the impersonal sense in which Wittgenstein is asserting that the stalking cat expresses an intention isn’t the personal sense in which Anscombe would be denying it. This resolution comes at the cost of making Anscombe incoherent, however. For if her point were merely that animals do not express their intentions in the personal sense – that is, don’t tell us of them – then she ought to have said the reverse of what she does: Intention is just like emotion in this respect, for animals don’t tell us of their emotions, or of their hopes and fears, either. 51. In point of fact, ordinary usage gives AP some support. One can express one’s intention to turn right by making a hand gesture (since this is a ‘bodily movement with a conventional meaning’: p. 5); but someone who begins to turn right is not ‘expressing an intention’ to do so, even though that same movement might express his determination or his fear, and even though it may make his intentions (e.g. to turn right, to confuse the enemy, etc.) apparent to an observer. We need to see more clearly, however, what distinction language is harbouring here. 52. Of course, numerous states of the person – for example, a dull throbbing in one’s right knee, a slow, spreading fear in the pit of one’s stomach – are manifested

Richard Moran and Martin J. Stone

53. 54. 55. 56.

57.

58.

167

only in creatures who have the linguistic means to differentiate and report them. So conventional expression goes beyond natural expression. But the present point isn’t that conventional expression only deputizes for what is already expressible by non-verbal behaviour, only that it sometimes does. Cf. Wittgenstein, Philosophical Investigations, p. 189. Cf. Ludwig Wittgenstein, Philosophical Investigations § 242, Zettel §545. See Wittgenstein, Remarks on the Philosophy of Psychology, Vol. II, § 179: ‘There is no cry of intention’ – that is, as there is a cry of pain, joy, grief and so on. Even God’s intentions, in so far as they are not expressed but merely revealed by history, do not impugn anything. For example, the regularity of the seasons does not afford a basis – as a verbal expression would – for regarding unseasonable rain as a divine mistake. See Anscombe, ‘Rule, Rights and Promises’, in Ethics, Religion and Politics: Collected Philosophical Papers Vol. III (Oxford: Blackwell, 1981), p. 99: ‘God himself can make no promises to man except in a human language.’ The point of Anscombe’s denying that an incipient course of action expresses intention (as a face or tone of voice expresses anger) might be approached another way, by considering what might serve to distinguish impersonal ‘expressions’ from other signs or indications of a person’s state. As Anscombe elsewhere remarks, ‘A man could be said not to have given expression to his anger at all – he merely brought it about that the man who had offended him was ruined or hanged.’ ‘Pretending’, in Metaphysics and the Philosophy of Mind: Collected Philosophical Papers, Vol. II (Minnesota: University of Minnesota Press, 1981), p. 88. According to this, an expression is not a mere sign or indication, but what exactly is the difference? It seems correct to say: An expression allows for the possibility of retaining the concept in cases where what is expressed is absent; a mere sign or indication does not. Thus, we may say that the relation ‘X expresses Y’ excludes cases in which the truth of the statement depends on the truth or actuality of Y. To illustrate: Ruining him is not an expression of anger, for it is not anger which is expressed at all unless the person really is angry. In contrast, if X expresses Y, then Y is itself present in the expression X, as for example, anger is present in the angry furrows in the face, whatever the person’s state of mind. Hence it is possible to be surprised that (e.g.) a face expresses anger (e.g. since this person has no cause for complaint), whereas – since a person’s action only manifests an intention to do what they are really doing or intending to do – there can’t be any surprise that a bit of behaviour manifests an intention to do such and such. Considered apart from their verbal expressability, intentions are sunk in facticity. This is why Anscombe suggests an analogy between the movements of the cat (as a basis for attributing intentions) and the car’s stalling engine. Just as the engine’s behaviour indicates that the car is going to stop only if the car is going to stop, so the movements of the cat indicate only what it actually goes on to do. For more on the notion of practical knowledge, see Richard Moran, “Anscombe on ‘Practical Knowledge’,” in Agency and Action, eds. J. Hyman & H. Steward (Cambridge: Cambridge University Press, 2004), pp. 43–68 (Royal Institute of Philosophy Supplement: 55). The source of the phrase ‘direction of fit’ as attributed to Anscombe seems to be Mark Platts, Ways of Meaning (London: Routledge and Kegan Paul, 1979), pp. 256–257. The transformation of her idea is already complete in this short passage, which re-writes her original thought about the relation between an expression of intention and what the person does as a claim about the relation between a state (desire) and the world: ‘Miss Anscombe, in her work on inten-

168

59.

60. 61. 62.

63.

64.

Anscombe on Expression of Intention tion, has drawn a broad distinction between two kinds of mental state, factual belief being the prime exemplar of one kind and desire a prime exemplar of the other (Anscombe, Intention, § 2). The distinction is in terms of the direction of fit of mental states with the world.’ See Luca Ferrero, ‘Intending and Doing’ (manuscript), which stresses the active quality of such ‘waiting’ (monitoring for interferences, commitment of resources, etc.) in the usual case. See Donald Davidson, ‘Intending’, in Essays on Actions and Events, p. 90. See the opening quotation from Sartre, Being and Nothingness, p. 74. Or better, their ‘being engaged in doing something’: the word ‘engaged’ straddles the sense of (1) doing something and (2) being committed – in a non-promissory sense – to doing something. We agree with Ferrero’s conclusion (note 51 above) that intending is a kind of performance which is continuous in structure with intentional action, so that to intend to F to is be (already) engaged in F-ing. As Ferrero puts it, ‘[F]uture directed intending is not a truly separate phenomenon from either the intending in action or the acting itself. Ultimately, all intentions are in action, or better still, in extended courses of action.’ In this connection, see recent work of Pamela Hieronymi, where actions as well as attitudes are not only understood as embodying reasons, but more specifically where the relevant notion of ‘reason’ is the more articulated one of ‘a consideration that bears on a question’, as distinct from the more primitive one of ‘a consideration in favor of’ (which might apply equally to considerations in favor of having some belief as well as to considerations in favor of its truth). In ‘The Wrong Kind of Reason’, The Journal of Philosophy, Vol. 102, No. 9 (September 2005), pp. 437–457; and ‘Controlling Attitudes’, Pacific Philosophical Quarterly, Vol. 87, No. 1 (March 2006), pp. 45–74, the appeal to the applicability of a certain range of questions as constituting the action or the attitude is motivated in part by reference to Anscombe’s ‘Why?’ question. The systematic unity of action’s ‘what’ and ‘why’ might be thought of as Anscombe’s central theme. She shows that: (1) An expression of what an agent is doing characterizes an intentional action only insofar as there is an answer to the question of why she is doing it (in the relevant, reason-involving sense); (2) positive answers to the interrogation why themselves characterize what the agent is intentionally doing (at least up to the point where use of the progressive gets linguistically denied); and (3) any expression of what an agent is doing is itself the answer to a question why, directed at something else she is doing.

9 Can One Act for a Reason without Acting Intentionally? Joshua Knobe and Sean D. Kelly

Since the important work of Elizabeth Anscombe (1957), philosophers have been almost unanimous in accepting the claim that every behavior performed for a reason is performed intentionally. We develop and discuss here an apparent counterexample to this claim—a case in which people are inclined to say that an agent has performed a behavior for a reason, but are not inclined to say that he has performed that behavior intentionally.

I It is widely agreed that the concept of acting intentionally is closely linked in some way to the concept of acting for a reason, but there has been a certain amount of disagreement about the precise nature of this link. We therefore begin by distinguishing two related theses: (Davidson’s Thesis) Every behavior that is performed intentionally is performed for a reason. (Anscombe’s Thesis) Every behavior that is performed for a reason is performed intentionally.1 Taken together, Davidson’s Thesis and Anscombe’s Thesis entail the biconditional conclusion that a behavior is performed intentionally if and only if it is performed for a reason. This is a conclusion of great theoretical power and simplicity, and it has probably been discussed more widely than either of the two individual theses taken separately. Still, it is important to remember that the two theses are logically independent. Arguments against one of the two theses will not usually count as arguments against the other. Although Davidson’s Thesis has received vigorous support from a number of philosophers of action (e.g. Davidson 1963; Goldman 1970; Mele 1992), it has also occasionally been called into question. Hursthouse (1991), for example, claims that spontaneous expressions of emotion are not performed 169

170 Act for a Reason without Acting Intentionally?

for any reason but still count as intentional actions. Consider the agent who gets very angry and ends up punching the wall. It seems clear that such an agent has punched the wall intentionally, but on Hursthouse’s view at least, he or she has not punched the wall for any reason. Here, however, our primary concern is not with Davidson’s Thesis but with Anscombe’s. This latter thesis has not been particularly controversial. It is upheld by Goldman (1970), Malle, Knobe, O’Laughlin, Pearce and Nelson (2000), Mele (1992) and, of course, Anscombe (1957). We know of no arguments against it.2 A key source of evidence for Anscombe’s Thesis is our ordinary practice of giving reason explanations. Specifically, it is thought that people do not ordinarily accept reason explanations for behaviors that they do not regard as intentional. If a speaker thinks that Jane bumped into Bob unintentionally, that speaker will not normally find it acceptable to use a reason explanation like “Jane bumped into Bob in order to get his attention.” As long as Jane’s bumping into Bob is regarded as an unintentional behavior, it will be felt that Jane did not bump into Bob for any reason at all and hence that it would be wrong to say that Jane bumped into Bob “in order to” fulfill any kind of purpose. Similarly, with many other types of unintentional behavior. If Jane unintentionally shivers or trips or starts crying, it will be felt that she is not performing these behaviors for any reason and that it would therefore be wrong to explain her behavior using phrases like “in order to” that are ordinarily reserved for reason explanations. But, of course, accidental behaviors and uncontrollable reflexes are only the simplest forms of unintentional behavior. The true test of Anscombe’s Thesis is its ability to handle more complex cases—cases in which a behavior would normally be considered unintentional even though it was foreseen and produced by a deliberate choice.

II We turn, therefore, to the difficult problem of side effects. An outcome can be considered a side effect if and only if (1) an agent chooses to perform a behavior, foreseeing that she will thereby (probably) bring about the outcome, but (2) the agent does not actually perform the behavior for the purpose of bringing about the outcome. Thus, suppose that the chairman of the board of a company decides to implement a particular program. He knows that the program will lead to some outcome x. But he does not care at all about outcome x; he has only chosen to implement the program because he thinks that he can thereby

Joshua Knobe and Sean Kelly

171

increase profits. Here we can say that outcome x is a “side effect” of the chairman’s behavior. A question now arises as to whether side effects can ever be brought about intentionally. Here, opinions differ. Adams (1986) says that side effects are never brought about intentionally; Ginet (1990) says that side effects are always brought about intentionally; Harman (1976) and Bratman (1984; 1987) say that some side effects are brought about intentionally and others are brought about unintentionally. As we shall see, people’s ordinary use of the word “intentionally” follows this third view. People are willing to apply the word to some side effects but not to others. Thus, there can be no single general answer as to whether or not people think that the chairman intentionally brought about “some outcome x.” People’s answers will depend in a crucial way on what outcome x happens to be. This fact about people’s use of “intentionally” gives us a valuable opportunity to put Anscombe’s Thesis to the test. We can check to see whether people’s use of reason explanations fits with their use of “intentionally” in the way Anscombe’s Thesis seems to suggest. Of course, Anscombe’s Thesis could potentially be true even if it does not comport well with ordinary language, but it seems clear that ordinary language provides at least prima facie evidence regarding difficult questions like this one. Suppose we found that people were sometimes willing to accept reason explanations when they felt that a side effect had been brought about intentionally but that they were never willing to accept reason explanations when they felt that a side effect had been brought about unintentionally. Surely, this finding would provide a kind of evidence for Anscombe’s Thesis—not conclusive evidence, of course, but evidence nonetheless. And now consider another possible result. Suppose we found that people sometimes accepted reason explanations for side effects that they regarded as unintentional. We would then have a very strong argument against Anscombe’s Thesis. Indeed, the only way to defend Anscombe’s Thesis against the argument would be to show either that ordinary English speakers were mistaken in their use of the word “intentionally” or that they were mistaken in their use of reason explanations. Although it might be possible to show that ordinary speakers are mistaken in one of these ways, we take it that there is a very strong presumption in favor of the view that ordinary speakers are speaking correctly. With these issues in mind, let us consider two specific cases in which an agent brings about a side effect. First, a case in which the side effect is environmental harm: The vice-president of a company went to the chairman of the board and said, “We are thinking of starting a new program. It will help us increase profits, but it will also harm the environment.”

172

Act for a Reason without Acting Intentionally?

The chairman of the board answered, “I don’t care at all about harming the environment. I just want to make as much profit as I can. Let’s start the new program.” They started the new program. Sure enough, the environment was harmed. And now consider what happens when we modify this case by replacing “harm” with “help”: The vice-president of a company went to the chairman of the board and said, “We are thinking of starting a new program. It will help us increase profits, and it will also help the environment.” The chairman of the board answered, “I don’t care at all about helping the environment. I just want to make as much profit as I can. Let’s start the new program.” They started the new program. Sure enough, the environment was helped. It has been shown experimentally that these two cases yield two radically different patterns of intuition about whether or not the agent acted intentionally (Knobe 2003). Most people who receive the first case say that the chairman intentionally harmed the environment, but most people who receive the second case say that the chairman unintentionally helped the environment. Presumably, the key difference between these two cases has something to do with the normative status of the two outcomes. Because people see environmental harm as a bad thing and environmental help as a good thing, they end up concluding that the former was brought about intentionally, the latter unintentionally. As of yet, there is no consensus on the question as to why normative considerations affect people’s responses to this pair of cases. Some researchers say that normative considerations actually play a role in people’s concept of intentional action (Knobe forthcoming; Mele 2003); others say that people are being unduly influenced by feelings of blame (Alicke forthcoming; Malle & Nelson 2003; Nadelhoffer forthcoming); and still others say that the effect can be explained in terms of conversational pragmatics (Adams & Steadman 2004a; 2004b).3 We will make no attempt to resolve this controversy here. Instead, we will focus on the question as to whether people’s intuitions about intentional action—however these intuitions might have arisen—end up fitting with their intuitions about reason explanations in the way that Anscombe’s Thesis seems to suggest. Assuming that people’s intuitions accord with Anscombe’s Thesis, they shouldn’t feel that the chairman helped the environment for a reason.

Joshua Knobe and Sean Kelly

173

Recall that Anscombe’s Thesis states that every behavior performed for a reason is performed intentionally. If the chairman helped the environment unintentionally, as subjects’ judge he did, then according to Anscombe’s Thesis he did not help the environment for a reason. In the case of harming the environment, Anscombe’s thesis makes no prediction. It does allow, however, that the chairman may have harmed the environment for a reason. In fact, people’s intuitions fit this pattern perfectly. Experimental results show that people think it sounds right to say “The chairman harmed the environment in order to increase profits” but that people do not think it sounds right to say “The chairman helped the environment in order to increase profits” (Knobe 2004). This is truly an impressive victory for Anscombe’s Thesis. Indeed, it might be thought that Anscombe’s Thesis provides the only natural way of explaining the asymmetry in people’s use of “in order to” in these two cases.

III The careful reader may by now have surmised that our argument takes the form of a trap. We began by describing a phenomenon that seems to provide evidence in favor of Anscombe’s Thesis, but it will later become clear that this very phenomenon, when properly understood, actually provides evidence for the view that a behavior can be performed for a reason even if it is not performed intentionally. If one accepts what we have said thus far and takes the data presented above to be legitimate evidence in favor of Anscombe’s Thesis, it will be extremely difficult to resist the conclusions we draw below. Anticipating this turn of events, a defender of Anscombe’s Thesis might offer certain objections to the argument presented in the previous section. We should emphasize, however, that no such objections have actually been put forward in the existing philosophical literature. Philosophers have offered competing explanations of the findings described above, but all of the philosophers who have discussed these findings have concluded that people’s use of reason explanations is sensitive in some way to moral considerations and that this sensitivity is to be understood in terms of people’s tacit grasp of the relationship between reason explanations and intentional action (Adams & Steadman 2004b; Knobe forthcoming; Nadelhoffer 2006; Turner forthcoming). In other words, nothing in what we have said so far is actually controversial among philosophers working on these issues. Still, it may be helpful to pause briefly here to allay some lingering doubts.

174

Act for a Reason without Acting Intentionally?

We turn now to the first potential objection: In the cases under discussion, one finds three basic elements—a behavior, an aim and a side effect. These three elements can be depicted as follows: Aim

Side effect

Behavior

Now, the phrase “in order to” is used to indicate a means-end relationship. In other words, it is used to indicate the kind of relationship that can only obtain between a behavior and an aim. Strictly speaking, then, the phrase cannot be applied to the bringing about of a side effect. So what people really mean to say is not that “the agent brought about the side effect in order to bring about the aim” but rather that “the agent performed a behavior (namely, the behavior that brought about the side effect) in order to bring about the aim.” To the extent that people sometimes fail to express this point with sufficient precision, it is only because they are being sloppy. Nothing of importance is thereby revealed about the nature of reason explanation itself. The key claim behind this objection is that the phrase “in order to” can only be correctly applied in cases that meet a specific criterion. But what the data show is that people sometimes use that phrase in cases that do not meet this criterion. So the objection posits a standard of “correctness” that departs in certain respects from ordinary usage. One way to make this objection compelling would be to show that people find it difficult to keep track of all the complex conceptual distinctions that seem to be in play here. For example, it could be argued that people have a hard time getting a handle on the distinction between saying “The agent brought about the side-effect in order to ...” and saying “The agent performed a behavior (namely, the behavior that brought about the sideeffect) in order to...” But the data seem to suggest that people actually are capable of using this distinction. In particular, people seem not to be willing to say that the agent brought about a side effect for a reason when that side effect is a good one. The objection therefore faces a major hurdle. It seems unlikely that people are able to make all of the necessary conceptual distinctions as long as the side effect is a good one but that they somehow find these distinctions too difficult when the side effect is a bad one. A more plausible explanation would be that they make all the necessary

Joshua Knobe and Sean Kelly

175

conceptual distinctions in both cases but that, when the side effect is a bad one, they regard reason explanations as acceptable. Of course, there are various strategies one might use to argue for a standard of “correct” use that diverges in certain respects from the use one finds among ordinary speakers. But we take it that there is at least a strong presumption in favor of the view that native speakers of English are speaking English correctly. The burden of proof thus seems to be on those who would deny that it is correct to use sentences similar to the ones people deemed acceptable in the studies described above. We turn therefore to a second objection: The sentence “He harmed the environment in order to increase profits” involves an important ambiguity. On one interpretation, the sentence means something like “There exists a behavior that is a harming of the environment, and he performed this behavior in order to increase profits.” On this interpretation, the sentence is true. After all, the behavior of harming the environment simply is the behavior of implementing the policy. (What we have here are two descriptions of the very same behavior.) Moreover, it is clear that he implemented the policy in order to increase profits. So there is a sense in which the behavior of harming the environment was a behavior performed in order to increase profits. But there is also an interpretation on which the sentence comes out false. On this latter interpretation, the sentence says that the behavior was specifically performed for a reason under the description “harming the environment.” But this is clearly not the case. The behavior was not performed for a reason under that description but rather under the description “implementing the policy.” In short, it can be correct to regard a sentence like this one either as true or as false, depending on which interpretation one happens to adopt. This objection takes us into difficult territory in the theory of action. Some theories of action do indeed imply that the harming of the environment and the implementing of the policy are the same behavior under two different descriptions (Anscombe 1957; Davidson 1967), but others imply that what we have here are two distinct behaviors that just happen to involve the same bodily motions (Goldman 1970). We cannot hope to resolve this issue here. Fortunately, however, it is possible to address the central questions of the present essay without taking any stand on the difficult issue of action individuation. In fact, it seems to us that that whole issue is a red herring in the present context. What needs to be explained is the asymmetry between people’s judgments regarding good side effects and their judg-

176

Act for a Reason without Acting Intentionally?

ments regarding bad side effects. The alleged ambiguity here seems not to provide an explanation of that phenomenon unless we assume that there is a systematic pattern whereby people resolve the ambiguity differently depending on the goodness or badness of the side effect. We see no grounds for positing such a systematic pattern. In the remainder of the essay, we therefore ignore these objections and proceed on the view that has become the consensus among researchers working on this issue. That is to say, we will proceed on the view that a certain kind of normative judgment—a judgment about the goodness or badness of a side effect—really does play a role in people’s use of reason explanations.

IV Although the cases we have just discussed seem to be evidence for Anscombe’s Thesis, there is even stronger evidence that the thesis is false. One can construct cases in which subjects judge that an agent has performed an action for a reason even though they also judge that he did not perform it intentionally. In the cases of the chairman and the environment, as we have seen, judgments about acting for a reason and acting intentionally line up nicely. But this is only because in these cases there is no radical divergence between two different kinds of normative evaluation that the subject can make. When we devise examples that pull these normative evaluations apart, we get robust evidence against Anscombe’s Thesis. The two normative evaluations in question are the evaluations of the moral goodness of the side effect as understood by the subject, on the one hand, and as understood by the agent, on the other. In the cases discussed in Section II, these evaluations line up fairly well; or at least they do not diverge radically. To see this, consider the actions of helping and harming the environment from the point of view of normal subjects and of the chairman in the two scenarios. We may assume that in general subjects think that helping the environment is better than harming it. This is a fairly prevalent view in our society, and it seems reasonable to think that most subjects have something like this as a background assumption. (In future experiments, of course, this is the kind of thing that might be tested for explicitly.) Likewise, in the two chairman scenarios there is no strong reason to think that the chairman believes that harming the environment is actually better than helping it. He is, of course, at the very least callous in the case in which he goes ahead and harms the environment for profit. But he is not malicious; he is not represented as taking any joy in harming the environment, and indeed he says that he doesn’t care one way or the other about it. But it is possible to devise scenarios in which the agent’s evaluation of the moral worth of a side effect is represented as explicitly contravening the subject’s (assumed) evaluation of the very same act. This will work best if

Joshua Knobe and Sean Kelly

177

the agent is represented as someone with moral values that are diametrically opposed to our own. Consider, for instance, the following scenario. A terrorist discovers that someone has planted a bomb in a nightclub. There are lots of Americans in the nightclub who will be injured or killed if the bomb goes off. The terrorist says to himself, “Whoever planted that bomb in the nightclub did a good thing. Americans are evil! The world will be a better place when more of them are injured or dead.” Later, the terrorist discovers that his only son, whom he loves dearly, is in the nightclub as well. If the bomb goes off, his son will certainly be injured or killed. The terrorist then says to himself, “The only way I can save my son is to defuse the bomb. But if I defuse the bomb, I’ll be saving those evil Americans as well... What should I do?” After carefully considering the matter, he thinks to himself, “I know it is wrong to save Americans, but I can’t rescue my son without saving those Americans as well. I guess I’ll just have to defuse the bomb.” He defuses the bomb, and all of the Americans are saved. In this scenario, the side effect is that the Americans are saved. Unlike in the chairman cases, however, it seems here that the agent’s evaluation of the moral worth of this side effect explicitly contravenes the evaluation of normal American subjects. For although we can assume that a typical sample of American subjects will judge that it is a good thing to save the lives of innocent Americans, the terrorist is represented as believing that this side effect is morally bad. Now we can ask two distinct questions about this side effect. First, does the terrorist intentionally save the Americans? And second, does the terrorist save the Americans for a reason (namely in order to rescue his son)? Among ordinary speakers of the English language the answers to these questions diverge dramatically. Most people judge that the terrorist saved the Americans in order to save his son, but that he did not save them intentionally. (This claim about people’s responses strikes us as correct on an immediate and intuitive level, but we ran a quick experiment to verify it.4) The divergence of responses in the two conditions is striking. They seem to provide an example in which it is considered acceptable to say that an agent has acted for a reason but in which it is also considered acceptable to say that he has not acted intentionally. In other words, they provide an apparent counterexample to Anscombe’s Thesis. In the following section we will analyze these results.

V We have seen that normative judgments influence people’s intuitions about the application of two distinct action-theoretic concepts: the concept of

178 Act for a Reason without Acting Intentionally?

acting intentionally and the concept of acting for a reason. But we have also seen that a subject’s intuition about how to apply each of these concepts is influenced by a different normative judgment: (1) Intuitions about whether a side effect was brought about intentionally are influenced by the subject’s judgment about its normative status. (2) Intuitions about whether a side effect was brought about for a reason are influenced by what the subject believes to be the agent’s judgment about its normative status. Although different normative judgments are relevant to subjects’ intuitions about acting intentionally and acting for a reason, there is a general principle about how the normative status of a side effect influences the application of action-theoretic concepts. This principle is general in the sense that it applies to both of our action-theoretic concepts—acting intentionally and acting for a reason. Once we have determined which normative judgment is relevant, the principle predicts whether or not a subject will attribute the corresponding action-theoretic concept to a given side effect. The general principle in question is: Principle of Un-charity: A subject is more willing to attribute the action-theoretic concept to a side effect if the side effect violates the relevant norm. When combined with (1) and (2) above, the Principle of Un-charity yields definite predictions about whether or not a subject will attribute each of the action-theoretic concepts to a given side effect: if the side effect violates the subject’s norms for behavior, he will judge it to have been brought about intentionally; if it violates the agent’s norms for behavior, he will judge it to have been brought about for a reason. In determining whether a side effect was brought about intentionally or for a reason, the subject seems to go through a two-step process. First he chooses the normative evaluation of the side effect that is most relevant: his own or the agent’s. Next, he uses that normative evaluation in accordance with the Principle of Un-charity: if the side effect violates the relevant norm, he is more willing to attribute the action-theoretic concept in question. Let us consider each of these steps in turn. First, the choice of the relevant norm. Suppose the subject is considering whether a side effect was brought about intentionally. Why should the subject’s normative evaluation be relevant in this case? There are many possible explanations here, but one plausible account goes like this. When a subject is trying to determine whether a behavior was performed intentionally, it is often because he is trying to decide whether the agent deserves praise or blame for that behavior. In determining whether a behavior is praiseworthy

Joshua Knobe and Sean Kelly

179

or blameworthy, moreover, the subject’s normative evaluation of the behavior is the most salient one. (Would you really praise an agent for performing an action that you yourself didn’t believe was good?) For this reason, the subject’s normative evaluation of the behavior is the most salient one in determining whether the behavior was performed intentionally. Suppose, now, that the subject is considering whether a side effect was brought about for a reason. We have seen in this case that the subject considers the agent’s norm to be the relevant one. Why should this be so? Again, there are many possible explanations. Consider the following plausible account. When a subject is trying to determine whether a behavior was performed for a reason, it is natural for her to look at the behavior from the agent’s point of view. The question whether the behavior was done for a reason is understood as a question about the agent’s reasons for performing it. Therefore, the agent’s normative evaluation of the behavior is the most salient one in determining whether the behavior was performed for a reason. Let us turn now to the Principle of Un-charity. As we have seen, this principle states that a subject is more willing to attribute certain actiontheoretic concepts to a side effect if that side effect violates a relevant norm. The principle seems to apply both to the concept of intentional action and to the concept of acting for a reason. Why should it apply to both of these concepts? First, the case of intentional action. The Principle of Un-charity implies an asymmetry here: side effects that the subject regards as bad are judged to have been brought about intentionally while side effects that the subject regards as good are judged to have been brought about unintentionally. One plausible explanation for this asymmetry in people’s judgments about intentional action ties it to a prior asymmetry in judgments of praiseworthiness and blameworthiness. Subjects are generally more willing to assign blame for side effects that they regard as bad than they are to assign praise for side effects that they regard as good. This may be a purely contingent fact about our assignments of praise and blame; but it is a conspicuous and notable fact nonetheless. Recall the scenarios from Section II. Subjects are much more willing to blame the chairman when he harms the environment (a bad side effect) than they are to praise him when he manages to help it (a good side effect). Similarly, subjects do not regard the terrorist as worthy of praise, even though his behavior brings about the good side effect of saving the Americans.5 As we have seen already, there are a variety of different theories that attempt to explain the interaction between judgments of praise and blame on the one hand, and judgments of intentional action on the other. But all of these theories admit that there is an interaction between these judgments, and on all of these theories the asymmetry that we have noted between judgments of praise and blame can be used to explain the asymmetry in judgments about intentional action.

180

Act for a Reason without Acting Intentionally?

Finally, the case of acting for a reason. We need to explain why subjects are more willing to say that a side effect was brought about for a reason when the side effect violates the agent’s norms. We have seen that the question whether the side effect was brought about for a reason amounts, in this case, to a question about the agent’s own reasons for bringing it about. The agent will feel especially compelled to provide such reasons when his behavior brings about a side effect that he considers bad. Reason explanations, in cases like this, can serve to justify the side effect that the agent’s behavior brings about. A subject who is trying to determine whether a behavior was performed for a reason will naturally think of the situation from the agent’s point of view. She will therefore be more willing to provide reason explanations in cases in which the agent needs to justify his action.

VI The experiments reported here suggest that subjects are sometimes willing to say that an agent has performed a behavior for a reason even though they are not also willing to say that he performed the behavior intentionally. We believe that these results count as evidence against Anscombe’s Thesis—the thesis that every behavior performed for a reason is performed intentionally. If subjects are systematically willing to use the expression “in order to” in cases in which they are not also willing to use the expression “intentionally,” then subjects’ use of these terms seems to be evidence against Anscombe’s Thesis. How might an objector reply to our claims? If our experiments have been conducted properly, they do seem to establish an interesting fact about normal linguistic use. It remains open to an objector, however, to insist that facts about normal linguistic use are no sure guide to the truth of Anscombe’s Thesis. The thesis that every behavior performed for a reason is performed intentionally, the objector might say, is not a thesis about how people use the relevant expressions but a thesis about what is true of our behaviors. Take, by analogy, the case of arthritis. If we want to understand the true nature of this disease, it would be folly to proceed by investigating the ordinary use of the term “arthritis.” So too, the objector will say, facts about the ordinary use of expressions like “in order to” are simply irrelevant to questions about the true nature of human action.6 An objection like this does present a legitimate challenge to our view. In particular, it shows that our argument against Anscombe’s Thesis is not conclusive. Still, we believe that the evidence we have presented here serves at the very least to shift the burden of proof. For now a person who wants to hold on to Anscombe’s Thesis needs to tell a convincing story about why normal linguistic use diverges from it so systematically. Such a story might be forthcoming, but in its absence, the evidence from linguistic use seems hard to ignore.

Joshua Knobe and Sean Kelly

181

Notes 1. See Anscombe (1957, §5), Davidson (1963, p. 3). The key difference between the views of these two authors arises in cases where the agent might respond to a request for reasons by saying, “No particular reason. I guess I just felt like it.” In cases of this type, Anscombe (1957, §17) says that the agent does not act for a reason even though the behavior is fully intentional, whereas Davidson (1963, p. 3) says that the agent is simply indicating that she has no further reason but that she nonetheless has a certain kind of reason, namely, a desire to perform the behavior. 2. However, for an argument in the general area see Mele (1992). Mele claims that in some lucky achievement cases—cases of simple luck and cases of causal deviance— people may have middling opinions on average about whether the action was done intentionally but be strongly inclined to see the action as done for a reason. For example, in a deviance case subjects may have middling opinions about whether the shooter hit the bull’s-eye intentionally but be strongly inclined to say that he hit it in order to win the prize. It would be interesting to test such cases. 3. Note that although some of these theories assert that people incorrectly classify immoral behaviors as intentional, none of them asserts that people incorrectly classify morally good behaviors as unintentional. The counterexample we propose in Section III involves the classification of a certain behavior as unintentional. Thus there is nothing in any of these theories to suggest that that classification is incorrect. 4. To demonstrate this we conducted a simple experiment. Subjects were 62 people spending time in New York City parks. All subjects were asked to read the above scenario and fill out a brief questionnaire. We randomly assigned subjects either to the “reason” condition or to the “intentionally” condition. Subjects in each condition were asked to evaluate a certain sentence. In the “reason” condition, subjects evaluated the sentence “The terrorist saved the Americans in order to rescue his son.” In the “intentionally” condition, subjects evaluated the sentence “The terrorist saved the Americans intentionally.” Evaluations consisted in rating the sentence on a scale from –3 (“sounds wrong”) to +3 (“sounds right”), with the 0 point marked “in between.” Subjects in both conditions were also asked to evaluate two sentences concerning the normative status of the terrorist’s actions. These sentences were (1) “The terrorist deserves praise for saving the Americans” and (2) “The terrorist deserves blame for violating his own moral code.” The average rating for the sentence “The terrorist saved the Americans in order to rescue his son” was 1.8. The average rating for the sentence “The terrorist saved the Americans intentionally” was –0.5. This difference was highly statistically significant, t (61) = 4.13, p<.001. 5. Indeed, this tendency was apparent within the experiment itself. Subjects were asked to evaluate the sentence “The terrorist deserves praise for saving the Americans.” On a scale from –3 (“sounds wrong”) to +3 (“sounds right”), they gave an average rating of –1.6. One additional fact is worth reporting. Subjects tended not to agree with the statement “The terrorist deserves blame for violating his own moral code.” On a scale of –3 to +3 they gave this statement an average rating of –1.7. As will become clear, we propose that the subject thinks of the terrorist as having had an internal reason for performing what he (the terrorist) understands to be a morally bad act. In other words, he is understood as having an excuse. Whether in addition he

182

Act for a Reason without Acting Intentionally?

deserves blame for having performed (what he understands to be) a morally bad act depends upon whether that excuse was a good one. Our view takes no stand on this question. 6. See Weatherson (2003), who argues that people’s intuitions about the application of certain predicates can systematically diverge from the conditions under which those predicates are properly applied.

Bibliography Adams, Fred (1986). Intention and Intentional Action: The Simple View. Mind and Language, 1, 281–301. Adams, Fred and Annie Steadman (2004a). Intentional Action in Ordinary Language: Core Concept or Pragmatic Understanding? Analysis, 64, 173–181. Adams, Fred and Annie Steadman (2004b). Intentional Actions and Moral Considerations: Still Pragmatic. Analysis, 64, 268–276. Alicke, M. (forthcoming). Blaming Badly. Journal of Cognition and Culture. Anscombe, G. E. M. (1957). Intention. Ithaca: Cornell University Press. Bratman, Michael (1984). Two Faces of Intention. Philosophical Review, 93, 375–405. Bratman, Michael (1987). Intention, Plans, and Practical Reason. Cambridge, MA: Harvard University Press. Davidson, Donald (1963). Actions, Reasons and Causes. Journal of Philosophy, 60, 685–700. Reprinted in Davidson, D. 1980. Essays on Actions and Events, Oxford: Clarendon Press, 3–19. Davidson, Donald (1967). The Logical Form of Action Sentences. In N. Rescher (ed.), The Logic of Decision and Action. Pittsburgh: University of Pittsburgh Press. Ginet, Carl (1990). On Action. Cambridge: Cambridge University Press. Goldman, Alvin (1970). A Theory of Human Action. Englewood Cliffs: Prentice Hall. Harman, Gilbert (1976). Practical Reasoning. Review of Metaphysics, 29, 431–463. Hursthouse, Rosalind (1991). Arational Actions. Journal of Philosophy, LXXXVIII, 2, 57–68. Knobe, Joshua (2006). Intentional Action: A Case Study in the Uses of Folk Psychology. Philosophical Studies, 130, 203–231. Knobe, Joshua (2004). Intention, Intentional Action and Moral Considerations. Analysis, 64, 181–187. Knobe, Joshua (2003). Intentional Action and Side Effects in Ordinary Language. Analysis, 63, 190–193. Malle, Bertram F. and Sarah E. Nelson (2003). Judging Mens Rea: The Tension Between Folk Concepts and Legal Concepts of Intentionality. Behavioral Sciences and the Law, 21, 563–580. Malle, Bertram F., Joshua Knobe, Matthew O’Laughlin, Gale Pearce and Sarah E. Nelson (2000). Conceptual Structure and Social Functions of Behavior Explanations. Journal of Personality and Social Psychology, 79, 309–326. Mele, Alfred (1992). Acting for Reasons and Acting Intentionally. Pacific Philosophical Quarterly, 73, 355–374. Mele, Alfred (2003). Intentional Action: Controversies, Data, and Core Hypotheses. Philosophical Psychology, 16, 325–340. Nadelhoffer, Thomas (2004). Blame, Badness, and Intentional Action: A Reply to Knobe and Mendlow. Journal of Theoretical and Philosophical Psychology, 24, 257–269.

Joshua Knobe and Sean Kelly

183

Nadelhoffer, Thomas (2006). On Trying to Save the Simple View. Mind & Language, 21, 565–585. Turner, Jason (forthcoming). Folk Intuitions, Asymmetry, and Intentional Side Effects. Journal of Theoretical and Philosophical Psychology. Weatherson, Brian (2003). What Good are Counterexamples? Philosophical Studies, 115, 1–31.

10 Reasons: Explanatory and Normative Joseph Raz

A thesis familiar through being as often disputed as defended has it that intentional action is action for a reason. This essay contributes, through a partial elucidation, though not by arguing for it, to the defence of a weaker version of the thesis, namely: Acting with an intention or a purpose is acting (as things appear to one) for a reason. This thesis is weaker in two respects: (a) One would be Fing intentionally if one Fs for a reason. And one Fs for a reason if one Fs because, as one believes, there is a reason to F (even if the belief is false); and (b) While all actions with a purpose or intention are intentional actions, not all intentional actions can sensibly be said to be actions done with an intention.1 The reasons referred to above are normative reasons. A normative reason is a fact which, when one acts for it, gives a point or a purpose to one’s action, and the action is undertaken for the sake of or in pursuit of that point or purpose. Reasons, and this is the common view among writers on the subject, have a dual role here. They are both normative and explanatory. They are normative in as much as they should guide decision and action, and form a basis for their evaluation. They are explanatory in that when there is a normative reason for which an agent acted then that reason explains (i.e. it features in an explanation of) that action. ‘Reasons’ has two meanings. When the term is used in one meaning it refers to a normative consideration, when used in the other it refers to an explanatory factor. I will start by trying to make good this claim (Sections 1 and 2), and (in Section 3) contrast it with Broome’s account of reasons, a discussion which focuses on the relations between what one has reason to do and what one ought to do. Finally, I will consider the way normative reasons figure in the explanations of action (Section 4).

1

On explanatory reasons

Whatever provides a (correct) answer to questions about the reason why things are as they are, become what they become, or to any other reason-why 184

Joseph Raz

185

question is a Reason. Reason-why questions seek explanations and whatever provides or constitutes an explanation is the reason why whatever it explains is as it is. Needless to say I am not proposing a grammatical test. Reason-why questions can be asked without using those words. We can ask ‘what is the reason for the deformation?’, or ‘what explains the deformation?’, or use other words. What is important is the distinction between providing (or purporting to provide) information (‘It is four p.m.’, ‘She is in Sydney’) and providing (or purporting to provide) explanations. Reasons provide explanations. Some writers take propositions to provide explanations, and therefore to be reasons. As false propositions explain nothing I will join those who take facts to be explanatory reasons. One reason to take propositions (rather than facts) to be explanatory reasons is that logical and conceptual relations hold among propositions regardless of their truth. But, as such relations also hold among facts, it is well worth preserving the core idea (that reasons explain) even at the cost of occasional complexity (due to the inability to rely on unrestricted inferential relations) or awkwardness of expression. Facts are reasons why; that is, they are not reasons in themselves, but reasons why something is thus and so. They are reasons in as much as they provide an explanation. Possibly, any fact is a reason for something or other. For every fact there may be a reason-why question, in a correct reply to which it figures non-redundantly. To refer to a fact as an explanatory reason is to refer, at least implicitly, to a relation it has to something else: it is a reason why this or that happened, and so on. Arguably, explanations are also relative to the person(s) for whom they are intended. An explanation is a good one if it explains what it sets out to explain in a way which is accessible to its addressees, that is in a way that the addressees could understand were they minded to do so, given who they are and what they could reasonably be expected to do in order to understand it. However, there is a clear distinction between the two relativities. No useful information is conveyed by a proposition of the form: this fact is a reason. One needs to specify something about what it explains to convey any useful information (e.g. ‘this fact explains something about the origin of life’). On the other hand, while the criteria for an explanation being a good one are relative to its addressees, its character as an explanation is not. An explanation of the nature of laser radiation suitable for university students is an explanation of laser radiation, even when addressed to primary school children, for whom it is not a good explanation. Explanatory reasons are so in virtue of their relations to what they explain, and stating that a fact is a reason is stating that it stands in the explanatory relation to what it is a reason for. As can be seen I am using ‘reason’ to refer to any fact which figures (nonredundantly) in an explanation and not merely to the totality of facts all of which figure (non-redundantly) in an explanation. It is tempting to call the

186

Reasons: Explanatory and Normative

totality of all the facts which figure non-redundantly in an explanation a complete reason. I may occasionally use the term in order to avoid complex formulations. But if taken literally it implies more than is established and possibly more than is true: It implies that there is at least one comprehensive way of individuating facts, such that relative to any such scheme of individuation and object of explanation, it is either true or not, regarding each fact, that it belongs to the explanation of that object. There is reason to doubt that the explanation relation is such that it is ever true that regarding any object of explanation there is a set of explanatory facts such that it explains that object, and that adding any other fact to it is redundant so far as that explanation goes. It seems that our ways of individuating facts and the notion of explanation are such that any explanation can always be nonredundantly amplified, clarified and expanded.2 We should therefore take talk of complete explanation with a pinch of salt. There are, however, two important points. First, that even if there are no ‘complete reasons’ there may be alternative reasons, that is alternative successful explanations of the same object, which are not in competition with each other, so that each is an independent explanation of the object. Second, that normally in advancing or citing reasons, non-trivial parts of ‘complete’ reasons are cited as (asserted to be) reasons, and by so citing them the speaker implicitly refers to ‘a complete reason’ (or to a disjunction of complete reasons), of which they are a constituent part, as the reason (for whatever they are meant to explain). We can state this point while avoiding reference to the completeness of any reason: In stating that R is the reason for P we refer to a (possibly complex) fact {R} which includes R, and state that it explains P. Suppose I say, ‘the heat wave was the reason for his collapsing’, and you reply, ‘that is not so’. He would not have collapsed had it been less humid. What sort of disagreement is this? You are probably pointing out that the heat does not explain his collapse by itself. It explains it only in the context of certain other facts, and it may be useful to mention some other of them (or not, as the case may be). So we do not disagree about the explanation, merely about which features of it are worth mentioning. It would have been otherwise had you said: ‘No. He collapsed because he was struck by a bullet.’ In that case we would have been advancing rival explanations. As it is we both referred to the same explanation by citing different parts of it.3

2 Normative reasons The preceding observations explain why explanatory reasons are not much discussed by philosophers. Whatever one can say about them is better explored when studying explanations, a voluminous philosophical subject. Explanatory reasons are mostly discussed, or at least mentioned,

Joseph Raz

187

by philosophers interested in normativity, who consider whether there is a second sense to ‘reasons’, such that in that sense ‘reasons’ refer or purport to refer to what I will call normative reasons. Is there a second sense to ‘reasons’, and if so are there such reasons? Put in different terms: are there normative reasons, and are normative reasons, if there are such, reasons independently of being explanatory reasons? Are they reasons of a different kind? This is not the same as to ask whether all reasons are explanatory reasons. I have already acknowledged that they are: it is likely that all facts, I said, can figure in some explanation or another. I will continue to assume that all reasons are facts, and when we refer to other things as reasons, the references can be recast as references to facts,4 hence all reasons are explanatory reasons. That does not, however, establish the univocality of ‘reasons’. It is possible that there are facts which are reasons in a different sense while being also explanatory reasons. That they are reasons in a different sense can perhaps be established by the fact that they can explain (at least some of) what they can explain because they are reasons in a different sense of the word. I will argue that there is a second sense to ‘reasons’. When the context requires disambiguating my meaning, I will refer to reasons in this second sense as normative reasons. I will suggest that their character as normative reasons enables them to play a certain explanatory role, and thus that the way they function as explanatory reasons presupposes that they are also reasons in a different sense. It is generally agreed that the notion of a normative reason cannot be explained through an eliminative definition.5 That is, any explanation of it in which the word ‘reason’ does not occur will include another term or phrase whose meaning is close to that of ‘a reason’ so that those who puzzle over the nature of reasons will not be helped by the definition. It will raise similar puzzles in their minds. We explain the notion of a normative reason by setting out its complex inter-relations to other concepts. Not to explain, but to minimally locate what we are talking about, we can say that normative reasons, if there are such, count in favour of that for which they are reasons. They have the potential to (i.e. they may) justify and require that which they favour. Those who wish to deny that normative reasons are a distinct kind of reasons may claim that normative reasons are simply explanatory reasons that differ from others in providing explanations of a special kind of facts. After all explanatory reasons are often classified by what they explain: individual events, or laws of nature; motivations or pain an so on. The distinctness of the object of explanation does not require different senses of ‘reasons’. Reasons are just the facts which explain. Different kinds of facts may explain different kinds of phenomena but they explain them, and are reasons, in the same sense of ‘explanation’ and of ‘reasons’.

188

Reasons: Explanatory and Normative

If so-called normative reasons can favour, justify or require, is it not simply a way of saying that they are the facts which explain why it is a fact that something is favoured, justified or required? That is, they are facts explaining why the proposition that this or that is favoured, justified or required is true. If so does it not follow that there is no different sense of reason here, only a different object of explanation? But normative reasons do not always justify or require what they favour. Nor is it always the case that what they favour is (non-relationally) favoured. That depends on what else is true of it. When they neither justify nor require it, they cannot explain why it is justified or required. And when what they favour is not (non-relationally) favoured they cannot explain why it is favoured. Can we do better by modifying the suggestion: would not those reasons explain why what they favour is favoured by them? Only if one thinks that a fact can explain itself; or if you like, only in the pickwickian sense in which that A favours B explains why B is favoured by A. Clearly here talking of explanation is otiose. In any case it does not dispose of the claim that normative reasons do something other than explain. Their explanatory use is secondary, and depends on the fact that they favour what they favour, a fact which sets them apart from other explanatory reasons. The existence of a normative relation – that one thing is a reason for another – is, on this suggestion, the object of the explanation. But for there to be something to explain there must be normative reasons, that is reasons in a sense which is independent of the explanatory sense of ‘reasons’. Does one fare any better by suggesting that ‘P is a normative reason for A’ just means that P explains why A is favoured (to a degree) by something? And does it explain this by displaying what it is that favours A, namely P? Again, that style of explanation presupposes that there is a normative relation of being a reason for something (in this case for A) and states what that reason is in the instant case (namely that P).6 Or consider another tack: if reasons for belief are reasons only in being explanatory reasons, what do they explain? The answer may appear obvious: they explain why the person who has the belief, has it. They explain his believing. If that is meant as a universal truth about explanations of believing, that is of why people who have a belief believe as they do, then it is false. For example, the reason Jamie believes a certain proposition may be neither the reasons there are for his belief, nor the reasons for it of whose existence he is or can become aware. The reason for his believing may be that he was hypnotised to believe it. But hypnotic suggestions are not reasons for belief. So while the reason for his believing may be that he was hypnotised that is not a reason for the belief, it is not a normative reason for believing that proposition. Nor is it necessarily the case that the reasons to believe a proposition are the facts which explain that belief (that believing) if the belief is rational or rationally held. The reason which explains the believing looks back to

Joseph Raz 189

its causes (the causes of having it or of still having it). The rationality of believing depends on one’s openness to critical evaluation of the belief, one’s ability and willingness to revise or reject it were the evidence to point that way. One may say that reasons for a belief are those facts which explain the believing, meaning the acquisition of the belief when it was rationally induced. But this view allows that ‘reasons’ is ambiguous between explanatory reasons, which, presumably, can explain all beliefs, and normative reasons for belief, which also explain those beliefs which were rationally arrived at, that is beliefs arrived at because of reasons for the beliefs (and the same can be said of the explanation of why one rationally sustains certain beliefs when the explanation invokes reasons for those beliefs).7 Regarding the latter kind of reasons their ability to explain the believing depends on the fact that they are normative reasons, reasons which can justify a belief, whether or not they also explain it, and which explain beliefs as rational or justified because they are normative reasons. It is relevant here that we regularly refer to reasons for belief independently of any explanatory context, that is, when reasoning about what to believe, which is not the same as reasoning about what would explain the belief once we have it (and remember that – for reasons given – reasoning what to believe is not to be confused with reasoning about what one ought to believe).

3 Normative reasons and ought-propositions – Broome’s reasons A concept which is beyond doubt a normative concept is that expressed by ‘ought’ (when used in its primary meaning). It is therefore not surprising that if one thinks that reasons are explanations, and nothing but, then one would be tempted to claim that they are nothing but facts which can explain what one ought to do, to believe and so on. That is the line of argument taken by John Broome, and I will consider that hypothesis by considering his view. My discussion of Broome’s views has two objects: first, to understand why ought is not the basic concept; second, to explain why Broome’s understanding of reasons is partial and misleading.8 Broome’s view revolves around two theses:9 (A) Some reasons are perfect reasons: ‘A perfect reason for you to F is ... a fact that explains why you ought to F’. (B) Other reasons are pro tanto reasons: ‘A pro tanto reason for you to F is a fact that plays the for-F role in a potential or actual weighing explanation of why you ought to F, or in a potential or actual weighing explanation of why it is not the case that you ought to F and not the case that you ought not to F’. (p. 41)

190 Reasons: Explanatory and Normative

The first thing to note about the theses is that they are consistent with the view that ‘reason’ has two meanings: an explanatory and a normative one. Arguably only facts which constitute normative reasons can explain true ought-propositions.10 Furthermore, my view of normative reasons is more than merely consistent with the claim that normative reasons explain (that they are parts of an explanation of) the truth of true ought-propositions; it entails it. Broome seems to think that the reasons normally thought of as normative reasons are facts which explain ought propositions in a special way. He calls them ‘pro tanto’ reasons and claims that they feature in weighting explanations: [T]here are reasons for you to F and reasons for you not to F. Each reason is associated with a number that represents its weight. The numbers associated with the reasons to F add up to more than the numbers associated with the reasons not to F. That is why you ought to F. (pp. 36–37) He relaxes these conditions: weight need not be exact and the function of total weight from component weights need not be additive. Even so, Broome’s explanation of pro tanto reasons combines two elements. First, these reasons come with weights, and what one ought to do is some function of those weights. Second, there being reasons for something entails the possibility of the existence of a reason against that very thing. I am not aware of any ought-proposition which can be explained by reference to weights in the way suggested by Broome. The very notion of an associated weight is hard to make sense of. Be that as it may, it seems plain that weights play no role in our understanding of many reasons for action which are not conclusive ones, but which determine what to do. Broome does not offer any justification for the weight-related view of pro tanto reasons, unless one takes his reference to the use of weighing and balancing metaphors by various writers as a justification. But metaphors are exactly that. I regularly write about one reason defeating another. But that is hardly a reason to attribute to me the view that conflicting reasons are fighters, who engage in some form of combat. I have suggested that the nonmetaphorical point is that in deliberating about what one ought to do propositions about various reasons are relevant, and it is by reasoning from them, and about their interrelations and so on, that the conclusion is drawn.11 Given that the claim about weights is not essential to the thesis that reasons are just explanatory facts, I will ignore it. The second element in the definition of pro tanto reasons is that they are normative reasons which can be defeated. Broome rightly points to a difference in the logical standing of (what I regard as) different normative reasons. Some are such that there cannot be any reasons which conflict with them. Others, his ‘pro tanto reasons’, do not have that feature, that is there being a reason for

Joseph Raz

191

something is consistent with there being a reason against that very thing. The contrast is real and significant. But so are many other distinctions between types of reasons. Arguably, the difference between epistemic and practical reasons is more fundamental, but both are normative reasons.12 I conclude that Broome’s distinction between pro tanto and perfect reasons does not contribute to our understanding of normative reasons, and need not be further considered here. I shall argue that Broome’s account misconceives the relations of ought and reasons. First, according to Broome normative reasons are mere explanations of why ought-propositions are or are not true. It is worth remembering that, as Broome rightly points out, the explanation relation between explaining and explained facts is not itself normative. But according to him there is nothing normative in the explaining fact either. Reasons are correctly labelled ‘normative’ when and because they feature in the explanation of ought-facts, and nothing more. Their normativity, as it were, exhausts itself in being such explanations. One way of describing the point is to say that according to Broome there are no reasons which are in themselves normative. Second, and perhaps as a result, Broome misunderstands the nature of ought-propositions. Instead of a direct explanation of the ways normative reasons are normative, too large a task for the present essay, I will illustrate Broome’s mistake by pointing to two contexts in which the point of referring to reasons does not depend on their explanatory role. First, in cases of akrasia people act for what they believe to be the lesser reason. For akrasia to be possible it must be possible that they are right in that belief.13 So imagine cases in which they are right. I knowingly act for a lesser reason. I am not acting as I ought to act, and I know it. But I am acting for a (genuine) reason. It is merely one which, as I am aware, is defeated in the circumstances. The reason for which I act is a normative reason, and it explains my action. Its being a normative reason and its success in explaining my action, which is due to its ability to motivate me, as normative reasons can, do not depend on its contribution to the explanation of any ought-proposition, not even to the explanation of the falsity of any such proposition. This shows that, even though normative reasons may contribute to the explanation of some true ought-propositions, their relevance in cases of akrasia goes beyond any such contribution. The second example is drawn from a case recognised by Broome. He mentions that sometimes it is not the case that one ought to F, nor is it the case that one ought not to F. As he says such cases may belong to different subcategories (pp. 38–9). In some there are no normative reasons either for or against F-ing. In others there are reasons pro and con F-ing which do not defeat one another. Here the existence of normative reasons is essential to elucidate the difference between these two types of case, though there is no difference between them regarding which ought-propositions apply to

192 Reasons: Explanatory and Normative

them. Again, we see that normative reasons do more than explain oughtpropositions. These reflections point to the nature of the relations between reasons and ought-propositions, which I will sketch in rough metaphorical terms. Normative reasons are facts which have normative bearing or force. They are called ‘reasons’ because they can serve as stepping stones in reasoning about what to believe or what to do. Deliberating from the reasons which apply to us we become aware of the attractions and drawbacks of options. We may reach a variety of conclusions: that we have a duty to do something (and we may have duties which we ought not to fulfil); that we have a right to do something; that certain options are acceptable (we may have a right to do unacceptable things); that it would be prudent to take some actions; that it would be irrational not to take them; and others. One such conclusion, different from any of the ones I mentioned, is that we ought to take some action. Ought-propositions are not the centre of practical thought. Nor are they the foundations on the basis of which we can understand reasons. Rather, they are one of a variety of propositions whose truth conditions are the existence or absence of some normative reasons or others. There is one particular factor which complicates the explanation of oughtpropositions. Let us focus on propositions of the form: ‘When C, P ought to F’ (where ‘C’ stands for circumstances, ‘P’ for a person, or a set of people, and ‘F’ for an action, a doing or an omission). When such propositions are very specific they state what one has conclusive reason to do. For example: that John Doe ought to give Jane Roe £5 by midnight is true only if there is a conclusive reason for John Doe to give Jane Roe £5 by midnight.14 But this is not true of general ought-propositions. This is particularly clear regarding universal ought-propositions. Here are a few examples: ‘People ought to pay their debts punctually’; ‘People ought to be kind to their grandparents’; ‘One ought to vote in parliamentary elections’. Those who believe these and similar propositions do not necessarily believe that there are always conclusive reasons to be kind to one’s grandparents, to repay debts punctually or to vote in parliamentary elections. At the very least it should be clear that compliance with these three ought-propositions may on occasion conflict: On a particular day it may be the case that one would not be able to vote in the election if one were to pay one’s debt punctually and vice versa. There are two kinds of universal ought-propositions which are true only if there are conclusive reasons to behave as they indicate: First, there are conceptually true propositions of that kind. If murder is unjustified intentional homicide then ‘One ought not to murder’ is true because there is always a conclusive reason not to murder, and, on our assumption, that is a conceptually necessary truth. Second, there may be some so-called moral absolutes (or absolute reasons of other kinds) namely reasons (or combinations of reasons) which defeat all possible (combinations of) conflicting

Joseph Raz 193

reasons. For example, possibly it is true that ‘In all states the law ought to prohibit torture, without any qualification’, meaning among other things that all governments ought to see to it that their law prohibits torture. But most universal ought-propositions do not belong to either kind. Are they all false? One cannot respond that they are implicitly relative. ‘One ought to vote in general elections’ and ‘so far as one’s civic duties are concerned one ought to vote in general elections’ are distinct propositions, and in any case even the relativised proposition could be true even though it is not the case that in so far as one’s civic duties only are concerned one always has a conclusive reason to vote in general elections. Sometimes one’s civic duties require abstaining. The same will be true of any non-conceptually necessary relativisation. Nor can one maintain that universal ought-propositions are abbreviations of very detailed propositions which include a complete list of exceptions, all the circumstances in which one does not have a conclusive reason to act as the proposition indicates. It is plausible to think that while one can have complete knowledge and understanding of (the content of) ought-propositions, no one can know all the exceptions (i.e. that no one can know either a list of them or a generalisation stating them, which is not true merely for logical or conceptual grounds). Hence, the propositions normally expressed by universal ought-propositions are what they appear to be. They are not (identical with) detailed propositions listing all the cases, or circumstances, in which one need not do what the proposition states that one ought to do. It is one of the virtues of the concept of normative reasons that it enables us to think about normatively complex and indefinitely changeable situations, helping us to marshal their normatively significant features into forms which facilitate coherent deliberation. So how are universal ought-propositions related to specific ones and to reasons? I think that propositions of the form ‘When C, P ought to F’ are true only in the case that, and because, there is a reason (or a number of reasons) which applies whenever C is the case, and which in at least some instances of C is a conclusive reason for P to F.15 On this view it is a conceptual truth that there are normative reasons which explain why one ought to F, when one ought to. This account of the truth conditions of practical ought-propositions16 specifies the same truth conditions for specific and for universal ought-propositions. It is merely that as specific propositions apply only to one occasion they are true only if on that occasion there is a conclusive reason to do as they indicate. It is far from clear whether this account provides an adequate explanation of ought-propositions. For one thing it is not clear whether it applies to epistemic and other ought-propositions, which are not based on practical reasons. It may have other limitations as well. It is too simple to capture the nuanced ways in which ‘ought’ is standardly used, and therefore to account for the nuanced differences among propositions in whose expression it

194 Reasons: Explanatory and Normative

features, though many of them may be due to conversational implicatures. It does explain, however, why practical ought-propositions cannot play any foundational role in understanding practical thought.

4

The normative/explanatory nexus

It seems plausible to assume that reasons in both senses are called ‘reasons’ because of their connection to Reason. But there is a closer connection between them which explains the common name. Briefly, it is that normative reasons provide the standard explanations of beliefs and of actions done with an intention or a purpose. Moreover, it is a necessary condition of any fact being a reason that, when conditions are appropriate, it provides such an explanation. Put another way, epistemic reasons can explain (or figure in an explanation of) beliefs, and practical reasons can explain (or figure in an explanation of) actions performed with an intention or purpose. This point is generally recognised, though sometimes neglected. It expresses the thought that normative reasons can guide agents, that is that they can move agents, who are aware of them, to action, belief and the like. Hence, they can feature in explanations of such actions, beliefs and the like.17 In further exploring that idea I will not be looking for a characterisation of the causal or other mechanisms on the existence of which these explanations depend. I will merely try to characterise the kind of explanation involved. We can begin the exploration in the company of Bernard Williams, since the point was crucial to his argument for reason internalism.18 How do normative reasons explain and what do they explain? Following Williams I will explore this regarding practical reasons only. Similar considerations apply to epistemic reasons. Obviously, reasons for an action do not always explain the action, even when it was performed. It may have happened accidentally, and even when intentional and performed for a reason, the intention may have been motivated by something else, either by some other reason for that action or by a mistaken belief that there is some other reason. So the point is not that whenever one does what there is a reason to do one acts for that reason. Nor is it that there are no other, non-reason-related, explanations for an action (hypnosis, statistical explanations and others). Rather the point is that normative reasons must be capable of providing an explanation of an action: If that R is a reason to F then it must be possible that people F for the reason that R and when they do, that explains (is part of an explanation of) their action. Or, as Williams puts the point: ‘If there are reasons for action, it must be that people sometimes act for those reasons, and if they do their reasons must figure in some correct explanation of their action.’19 Furthermore, the role reasons play in the explanation must be of a certain form. If that R is a reason to F, then it must be possible that awareness that

Joseph Raz 195

R motivated the agent to F. Sometimes the phrase ‘motivating reasons’ is invoked in such contexts. I will not use it myself, for it is liable to confuse. Sometimes the phrase is used to refer to a kind of explanatory reasons for actions, those which explain them by pointing out that they were motivated by belief in the existence of a (normative) reason. That sense is much narrower than the natural understanding of the phrase (motivating reasons being reasons explaining actions by their motivations).20 Back to business: what is important for our purpose is not that facts that are normative reasons can explain (that they can figure in the explanation of) actions. Just about any fact can (given appropriate circumstances) figure in some explanation of some actions. The normative/explanatory nexus requires that the potential explanatory role of facts which are normative reasons depends on and presupposes their normative force: it has to be that they can explain because they are normative reasons. That I promised can explain my promise-keeping action in a way in which the impact of low atmospheric pressure cannot. No doubt being in the mountains where atmospheric pressure is low can explain some aspects of my conduct, but the explanation (assuming the normal scenario) is of a completely different kind. What is this special way in which normative facts figure in explanations? I take my promise as a reason for the action, recognise it as such even before the action, and take the action to be sensible because it is supported, perhaps even required, by my promise. The low pressure affects me independently of whether I am aware of it, and regardless of what I think of it. We can baptise this thought ‘normative/explanatory nexus’, namely that regarding every normative reason, it is possible for it to feature in an explanation of the action for which it is a reason as a fact whose recognition motivated the agent to perform it, and guided him in its performance. This last condition, that awareness of the reason guides the performance, and does not merely initiate it, is taken for granted in common reasonexplanations of actions, but is sometimes ignored to their detriment by more formal philosophical accounts. Harry Frankfurt pointed out its importance, and Kieran Setiya21 suggested that it solves the problem of deviant causation (so far as basic actions are concerned) which afflicts accounts of intentional action such as Davidson’s. Williams explains an important aspect of this nexus. To see it we need to remind ourselves how Williams’s view differs from that of some other Humeans. First, Williams finds no problem in the fact that reasons may be facts about how things are in the world and not only about the agents’ beliefs and desires. Of course, in that case they could figure in explanations of the agents’ conduct only if those agents are aware of them. But that is consistent with the normative/explanatory nexus as he (and I) understand it. Furthermore, Williams does not object to the thought that beliefs can motivate. He does not endorse the view that only desires can be reasons, since only desires motivate, and therefore only they can explain actions in the

196 Reasons: Explanatory and Normative

right way. This is neither true, nor is it true to Williams’s account. He allows, for example, that ‘there are some cases of an agent’s F-ing because he believes that there is a reason for him to F, while he does not have any belief about what that reason is’ (p. 107). More generally he asks: ‘Does believing that a particular consideration is a reason to act in a particular way provide, or indeed constitute, a motivation to act? ... Let us grant that it does – this claim indeed seems plausible ...’ (ibid.). The crux for Williams is not directly in the possibility of being motivated by one’s beliefs but in the way one could acquire such motivating beliefs: the basic point lies in recognising that the external reasons theorist must conceive in a special way the connexion between acquiring a motivation and coming to believe the reason statement. For of course there are various means by which the agent could come to have the motivation and also to believe the reason statement, but which are the wrong kind of means to interest the external reasons theorist. Owen [Wingrave – in James’s story] might be so persuaded by his father’s moving rhetoric that he acquired both the motivation and the belief. But this excludes an element which the external reasons theorist essentially wants, that the agent should acquire the motivation because he comes to believe the reason statement, and that he should do the latter, moreover, because in some way he is considering the matter aright. (pp. 108–9) This is indeed a required element. The initial thought that normative reasons must be capable of explaining is not that (taking the matter beyond Williams’ example, but in the spirit of his remarks) one could accidentally come to be motivated by awareness of the reason. Awareness of the reason must non-accidentally motivate. It must motivate, as Williams puts it, because the agent ‘is considering the matter aright’. But what is that way? Here we have to go beyond Williams, though without conflicting with what he says in the quoted passages. Williams’s phrase ‘considering ... aright’ suggests, first, that the explanation relates to rational agents, and depends on their exercise of their rational powers, and therefore (given the implausibility that the motivation can be generated in the right way by some external circumstances surrounding the believing) second, that the way that the belief has to explain the motivation is by having the content it has, by what it is a belief about. This leads in several steps to the requirement that the reason itself figure in the explanation. First, the belief must be belief in the fact which is a reason, and include belief in its character as a reason. Motivations to perform a particular act would not be reliably and rationally brought about (or constituted) by a belief unless the belief was a belief about a reason to perform that act. Second, the belief itself must be explained by the existence of the

Joseph Raz 197

reason, and it must be acquired or maintained in a rational way. Therefore, third, the reason itself figures in the explanation of the action, in being part of the explanation of the belief which motivated the action. And finally, fourth, the belief, the awareness of the reason (where it is rational and true), must not only prompt, but guide the action. A typical objection relies on the fact that, as was pointed out at the beginning of the essay, an act is intentional and done with a purpose even if the belief which motivated it is false. It follows, goes the objection, that the fact which renders the belief true (when it is true) cannot be part of the explanation of the action. It has, of course, to be admitted that when the belief is false (a) the action can be explained, and (b) its explanation as intentional must include reference to the belief that there was a reason for it (as was argued in the first point above), and (c) as there is no reason making that belief true, no reason can be part of the explanation. But that applies to the explanation of intentional actions based on false beliefs. In such cases the explanation does not refer to a normative reason for the action. Even if there were such reasons they do not explain why the action was performed. However, the objection continues, if the belief alone is sufficient to explain intentional actions when it is false it must also be sufficient to explain actions when it is true. The further factor, the existence of the reason, is not necessary to the explanation which is, as is shown by the case of false beliefs, adequate without it. It has to be admitted that citing the belief, without adding that it is rational and true, does explain the action, and shows it to be intentional. Furthermore, it is plausible to think that ‘being an explanation of’’ is not a transitive relation. Sometimes even if C explains B and B explains A, C does not explain A. So even when the existence of the reason explains awareness of it as a reason, we need something additional to show that it can also explain action for that reason. But that is consistent with the possibility that an explanation which includes the reason among the explanatory factors is a better explanation of the intentional actions to which it applies. To be sure the reason is not part of the explanation of the action just by being an element in the explanation of the belief which prompts the action. It has to play a role in the explanation of the action itself, especially in its explanation as intentional. The practice of explanation shows that in fact the reason does figure in explanations of actions: Why did I go to Chamonix for my holidays? Because it is so beautiful there. Why am I rushing to my office? Because I promised to meet a student there in ten minutes’ time and so on. It is, however, one thing to know that reasons can figure in explanations of action, it is another to understand why this is so. Analogous considerations show that epistemic reasons can be part of the explanation of the belief (the believing) that they are reasons for, and that practical reasons can be part of the explanation of the action for which they are reasons.

198 Reasons: Explanatory and Normative

Where a belief is held for reasons these reasons are relevant to its evaluation as warranted or otherwise, and therefore properly belong with the explanation of the believing. Analogously, when an action is taken for a reason that reason is relevant to the evaluation of the action, as an acceptable action, a rational or irrational one, a sensible or a silly one, and to the evaluation of the action in light of its consequences: is it a step in a direction which should be followed, or discontinued, should one apologise or compensate for it?, and so on. Hence, where an action is taken for reasons the reasons contribute to its explanation, deepen our understanding of the action in ways which are relevant for its evaluation. It cannot be objected that explanations by reference to reasons are not part of ordinary explanations of believing or acting, that they are special explanations relevant only to those interested in the rationality or justification or some other evaluative aspect of the belief or action. On the contrary, reasonexplanations explain action and belief by reference to their inherent features. After all, it is inherent to beliefs that those having them take them to be warranted, and would abandon them had they thought that they were unwarranted. Similarly, by their nature intentions to act involve belief in reasons for the intended action. Hence, reason-explanations deepen our understanding of intentions, actions and beliefs, by contributing to an understanding of whether they have the features which they purport to have. It may clarify things if we return to explanations of actions done in the belief that there is reason (or that there is undefeated reason) for them when that belief is false. Observing common explanations offered in such cases we distinguish several types: (a) The agent did it because he believed that R (b) The agent did it because he mistakenly believed that R (where ‘R’ has the required content for a reason explanation). It would be a mistake to think that (b) is the same explanation as (a) as the reference to the fact that there was no R is explanatorily idle. Rather (b) is a more comprehensive explanation than (a). (a) is adequate to a certain range of interests in why the agent so acted. (b) is adequate for a wider, perhaps one may say here, deeper range of interests. Similarly, when the belief which led to the action was true we still have two possible types of explanation: (a) The agent did it because he believed that R (c) The agent did it for R ((c) entails that he believed that R). (a) is still a good explanation, but for the reasons explained (c) is a better, more comprehensive explanation, one which answers a wider range of interests in the action.

Joseph Raz 199

Finally, to the fourth claim, that awareness of the reason must guide the action. I do not mean that otherwise the motivation would be irrational. There are many possible causal routes from a belief in a reason to motivation, which while not irrational, are adventitious. If normative reasons are to meet a meaningful explanatory potential requirement they must be capable of explaining through belief in their existence qua reasons. The element of guidance can be understood by analogy to a negative feedback mechanism: we, automatically and normally without being conscious of the fact, monitor the performance of the intentional action such that if it deviates from the course we implicitly take to lead to its successful completion we correct the performance, bringing it back to the correct path, or interrupt it, when we fail to correct it. So, the claim is that one’s action is guided by a reason just in the case that one is motivated by the reason, through awareness of it, in a way which is manifested by the (normally unconscious) self-correcting process of tracking the success of the process of performing the action. Another possible worry may be generated by my claim that reasons explain actions through the mediation of belief in the reasons as reasons. It may be thought that this implies possessing concepts and having beliefs which many people who act for reasons do not have. This worry, if justified, may not disprove the letter of the normative/explanatory nexus. After all it requires nothing more than the possibility of certain explanations. But it would go against its spirit. For surely, the nexus is meant to relate to a standard form of explanation of actions undertaken for reasons. The worry is, however, unjustified. Saying of people that they have certain concepts may mean either of two things. Perhaps it is better to say that there are two standards by which such claims are judged true. The lower standard requires having thoughts whose description involves the concepts, recognising the implications of the thoughts, and being guided by them, reacting to them appropriately and trying to adjust our responses when becoming aware that they are inappropriate. The inferential connections between concepts and between thoughts can be spelt out in general terms. We may say that there are principles which spell out these connections. Knowledge of such principles, however, is not required to meet this basic standard for concept possession. The higher standard of concept possession is one which requires ability to explain the concept, to articulate its implications and inferential connections. That is, it requires knowledge of the principles constitutive of the concepts. These are matters of degree. However, the basic standard for concept possession does not require such knowledge. This means that the gap between having concepts and knowing how to employ them, on the one hand, and being able to think about the principles which govern one’s understanding of those concepts, on the other, is greater than may appear. I can treat a promise as a reason, and as we may say, implicitly know that it is a reason, without being able to understand any statement of the principle which sets out what

200

Reasons: Explanatory and Normative

I understand when I understand the notion of a promise. We are here in territory which was explored by Brandom, and using his terminology we may say that making things explicit is more difficult than is sometimes thought, for it may require additional concepts, concepts which one need not have to have the implicit knowledge. Applying the distinction to normative reasons we could say that action for a reason requires ability to take certain facts to be reasons (e.g. it may require taking promises to be reasons, at least for those having the concept of a promise). It does not require being able to think about reasons in the abstract. I therefore conclude that it is justified to say that the normative/ explanatory nexus does not require excessive conceptual mastery, nor excessive conceptual knowledge. As was already explained, the requirement that normative reasons explain through agents being aware of reasons as reasons is necessary to ensure that we refer to the right kind of explanation, and the right kind of explanation is explanation mediated by our rational faculties. A final possible dissatisfaction I will mention here is that none of the above constitutes an explanation of the productive process which leads to awareness of the reason and from there to the motivation and the action. That is, of course, true, but I doubt that it is a drawback. Welcome as such explanations are, they are not needed for an understanding of the normative/ explanatory nexus. Indeed so far as that goes they may do too much. As stated the nexus is sufficient, for those who understand it, to distinguish cases in which an action is done for a normative reason (and can be explained relying on the nexus) from other cases. It can be rightly pointed out that the ability to distinguish those cases made possible by the statement of the nexus is not sharp, that it leaves us undecided in many cases in ways which no further explanation can resolve. But that is just how things should be. The phenomena are not sharp because they are defined by our concepts, and our concepts are vague, leaving the phenomena they apply to vague. So, the normative/explanatory nexus states that necessarily normative reasons can explain the actions, beliefs and the like of rational agents. The relevant explanations explain agents’ beliefs or actions and so on as a result of their exercise of their rational powers, leading to awareness of the facts which are reasons qua reasons, and to rational reaction to this awareness.

Notes I am grateful to David Enoch for helpful comments on an earlier draft, and to John Broome for helping to clarify the points on which we disagree. 1. The point is familiar. I discuss it further in ‘The Guise of the Good’ in S. Tenenbaum (ed.) Desire and the Good (Oxford: OUP forthcoming). 2. I am assuming further, unstated, restrictions on what may count as a complete explanation. They will be needed to avoid having trivial complete explanations, as for example, that for any fact all the other facts constitute its complete explanation.

Joseph Raz

201

3 All these considerations apply, mutatis mutandis, to normative reasons. It is possible to try to identify different types of explanatory roles for different elements in an explanation. I will not attempt such classifications. 4. I do not mean to suggest that we do or should refer only to facts as reasons. I follow this usage for convenience’ sake only, but by the same token not only facts can be taken to be explanatory reasons. For example, I refer to hypnosis as a reason in the text below. 5. Broome (see below) being one notable exception. 6. The same point applies to the suggestion that reasons are facts which would explain why justified or required actions are justified or required. Here too the explanation would presuppose that the reasons which feature in the explanations, or some of them, are themselves normative reasons. 7. This remark assumes that rationality has to do with the ability to discern and respond to reasons, a view which cannot be argued for here. 8. Broome formulates his theses with reference to ought-facts. I will use the more common terminology of true ought-propositions, meaning true propositions which can be expressed in sentences containing an ought operator (used in its primary meaning). 9. Broome, J. ‘Reasons’, in Wallace, R.J., Pettit, P., Scheffler, S. & Smith, M., Reason and Value (Oxford: OUP 2004). 10. It is true, though, that all perfect reasons, that is, all explanations of why one ought to do this or that or to believe this or that, will include elements which are not normative reasons. As I will explain below, they all require closure propositions, that is, propositions to the effect that nothing defeats the conclusion, and closure propositions are not themselves statements of reasons. So, if, as Broome stipulates, ‘perfect reasons’ explain true ought-propositions then – and this is my view – they include some normative reasons, and some other elements. 11. Raz, J. Practical Reasons and Norms (Oxford: OUP 1999, first published 1975). 12. See my ‘Reasons: Practical and Adaptive’, Sobel, D. & Wall, S. (eds) Practical Reasons (forthcoming). 13. While a person who acts on what he thinks is the lesser reason but is mistaken about this, and the reason he follows is the better reason, is akratic, it seems pointless if not meaningless to attribute akrasia to creatures who are unable to tell which reasons are better, and are inevitably mistaken in all their beliefs about the relative strength of reasons. 14. On many occasions there is no conclusive reason to do any of the actions open to us (as naturally presented when debating what to do). We may have undefeated reasons to stay at home and carry on with the novel we started last week, but also undefeated reasons to visit our cousin, or to perform any of a number of other feasible actions. In such cases there is nothing we ought to do. 15. Note that I am referring to simple unqualified ought-propositions (displaying the general form ‘X ought to F’ or ‘when C, X ought to F). Their meaning varies when qualified: one always ought to F may mean that one’s reason for F-ing is always conclusive and so on. 16. First suggested by me in the Introduction to Raz (ed.) Practical Reasoning (Oxford: OUP 1978). 17. The characterisation given is imprecise. To give but one example: suppose that a fact cannot guide because it is impossible for people to believe in it, but would be capable of guiding, and therefore of figuring in an explanation of actions, were it possible to believe in it. Is it a reason (at least so far as the stated condition is

202

18.

19. 20.

21.

Reasons: Explanatory and Normative concerned)? Whichever conclusion one comes to on this and similar issues is likely to be stipulative, or if you like, explicative, that is going beyond explaining existing concepts, and involving their modification for theoretical purposes. It need be none the worse for that. The important point is that to say of a fact that it is a reason for an action is not merely to say that it shows the action to have some good, some point to it. It is to say something like that it can rationally guide an agent towards that action. I will not consider the merits of any form of internalism or externalism about reasons, nor Williams’s own argument for internalism. Like some other writers I think that the contrast is more confusing than helpful. My own view will be clear enough. Its classification as a form of internalism or externalism is immaterial. Williams, B. ‘Internal and External Reasons’, Moral Luck (Cambridge: CUP 1981) p. 102. See Dancy, J. Practical Reality (Oxford: OUP 2000) p. 6: ‘I have characterised the distinction between the reasons why we do things and the reasons in favour of doing them in terms of the motivating and normative. In doing so I have tried to avoid any suggestion that we are dealing here with two sorts of reasons. ... the same reason can be both motivating and normative. A reason for acting can be the reason why one acted’. See Frankfurt, H. ‘The Problem of Action’ in The Importance of What We Care About (Cambridge: CUP 1988) pp. 69, 72ff, and Setiya, K. Reasons Without Rationalism (Princeton: Princeton UP 2007) pp. 31–32.

11 Reasons, Desires and Intentional Actions Maria Alvarez

Introduction There are two fundamentally different ways of understanding the relation between desires and intentional actions. According to one, whenever someone does something for a reason, some desire of his (or ‘want’ – I use these terms interchangeably here) is always part of the reason for which he does it. Since this view is associated with Hume, I shall call this the ‘Humean’ view of reasons. Humeans argue that without desires there is no motivation in action and that, therefore, desires are an essential part of the reasons for which we act and of any explanation of an intentional action. There is an opposing view that says that desires are not, or not normally, part of someone’s reason for acting – call this the non-Humean view. Defenders of this alternative view maintain that the Humean position cannot be right, not least because beliefs can also motivate someone to act and, so, these philosophers play down, or even eliminate, the role of desires in motivation and in the explanation of action. In the first part of this essay (Sections I to IV) I shall defend the view that desires are not normally part of the reasons for which we act. Nonetheless, I believe that the Humean insistence on the importance of desires contains important truths – truths that both Humeans and their opponents have failed properly to understand. The failure is due, I shall argue, to a lack of conceptual clarity in these discussions. In the second part of the essay (Sections V to VI) I shall outline some of the necessary conceptual distinctions and then provide a picture of the relation between desires and intentional actions that seeks to bring out the truths in the Humean position. None of these tasks, however, can be embarked upon without first making a few preliminary points. First, in the introductory paragraph I characterized the Humean view as one which states that desires are always part of the reason for which someone acts because. I said ‘part of the reason’ because, although Humeans place great emphasis on the role of desires, they tend to hold that the reason 203

204

Reasons, Desires and Intentional Actions

for which someone acts is normally a combination of a desire and a belief. The focus of this essay is on desires, in particular, on the question whether desires are part of the reasons for which we act. However, for the sake of simplicity, I shall often drop the qualification ‘part of’ and talk of whether desires are reasons for which we act. Second, some recent discussions about desires, reasons and actions have centred on whether desires are or provide reasons for acting – in the current terminology in the literature, whether desires are ‘normative reasons’: the reasons for which we ought to act. However, my concern in this essay is not with normative but rather with what are called ‘motivating reasons’: the reasons for which we actually act. So my question is whether desires are motivating reasons, that is whether desires are reasons for which we act – even if they are not reasons for which we ought to act. Third, the term ‘motivating reason’ is a term of art normally used to mean a reason that motivates someone to act, although what motivating amounts to in this context tends to be left unspecified. By ‘motivating reason’ I shall mean a reason that, in the agent’s eyes, makes the action right or appealing for him.1 What makes the action seem right or appealing to an agent can relate to different kinds of value: moral, hedonic, prudential, aesthetic and so on. For instance, what makes buying a particular new car seem right to me might be that it’ll be great fun to drive, or that it’ll be more fuel-efficient than my current car, or that it has a beautiful design. So any of those things might be a reason to motivate me to buy a car. Finally, a comment on the concept of wanting or desiring. It is possible to distinguish two senses of wanting or desiring – as we might put it, a ‘weak’ and a ‘strong’ sense. In the weak sense, wanting something is compatible with finding the thing wanted unappealing or even repugnant. This is the sense in which most of us might sometimes be said to ‘want’ to go to the dentist, to have an injection and so on. The stronger, perhaps more common, sense of ‘desire’ is associated with preference and pleasure, and contrasted with, for example, duty or necessity. 2 The claim that wants or desires are reasons for which we act is normally meant by those who advance it to involve the first, weak, sense of want or desire; and that is how I shall use it here.3

I

Intentional actions and acting for a reason

Like the term ‘motivating reason’, ‘intentional action’ is a term of art. It may be found outside of philosophy, although it is much more likely that the relevant ideas be articulated in terms of whether someone did something intentionally, or whether it was intentional of him or her to have done suchand-such a thing. In philosophy, however, the term (and hence the concept) ‘intentional action’ have occupied centre stage for over fifty years now, particularly since Anscombe discussed the concept in some detail in

Maria Alvarez 205

Intention, and many of the questions that have preoccupied philosophers of action during that period have focused on exploring intentional actions. An intentional action is often characterized in contemporary philosophy as ‘something done for a reason’.4 The characterization is not uncontroversial as it is plausible to argue that there are some things one does intentionally but not for a reason, such as doodling while talking on the phone, crossing and uncrossing one’s legs and so on (and perhaps actions which are expressive of emotions and passions: e.g. smashing a plate ‘in rage’.)5 On the other hand, if ‘intentional’ is taken to mean ‘done with an intention’ or ‘something the agent intended’, it may seem more accurate to say that these actions are not intentional since, typically, they are not done with any particular intention nor are they things that are clearly ‘intended’ by the agent.6 The issue seems largely a matter of definition. If ‘done intentionally’ is defined as merely done knowingly and voluntarily, then the actions mentioned above are typically intentional – or at least on the periphery of the intentional because sometimes one hardly realizes one is doing them. If, on the other hand, ‘intentional’ is defined as ‘done for a reason’, then these actions would not seem to be intentional because one does not normally do such things, for example, doodle or smash a plate in rage, for a reason, even though there is a reason why one does them.7 This would not imply that actions like doodling or crossing one’s legs are normally unintentional or accidental or coerced; we might say that such actions are voluntary but not intentional. For the purposes of this discussion, I shall use the second definition, and regard something someone does as an intentional action of his if and only if it was something he did for a reason. And the reason for which someone does something I shall refer to as the ‘motivating reason’ for that action. So the question I want to explore in this essay is whether, whenever someone does something for a reason, his desires are motivating reasons, that is, (part of) the reason for which he does that thing.

II

Wanting something and acting for a reason

As I said above, the majority of contemporary philosophers endorse the Humean view that my wanting something, together with my believing something, constitutes the reason for which I act.8 So, Humeans hold that if, for instance, I revise hard in order to pass my exam, my reason for revising is my wanting to pass my exam and my believing that my revising will result in my doing so.9 But although this is the dominant view, some have rejected it and argued that one’s wanting something is very rarely the reason for which one does it. The latter is a view that Anscombe expresses, though not explicitly, in Intention, so I shall start with an argument against the Humean view that can be found in that book. The argument I have in mind relates to Anscombe’s discussion of wanting. According to Anscombe, everything wanted has a ‘desirability characteristic’

206

Reasons, Desires and Intentional Actions

in the eyes of the person who wants it – if a person wants something, then there will be something the person sees in what he wants, some feature that the person takes the thing to have, which makes it desirable to him. For Anscombe this is a conceptual truth, so that if a person claims to want something but denies that there is any ‘desirability characterisation’ to be given of that wanted thing, for example, that it is beautiful, reassuring, rare, useful, completes their collection and so on, their claim to want it becomes unintelligible.10 The desirability characterization that a thing has for someone would constitute the agent’s answer to the question: ‘What do you want it for?’ and the answer will always, she argues, relate either directly or indirectly, to some form of the good (of human beings). Another way to put the point is that the agent’s desirability characterization of the thing wanted shows what, in the agent’s eyes, the good of the thing is for him (see Intention, pp. 77–8). In saying this, Anscombe is endorsing an ancient and mediaeval view that what is wanted is always wanted sub ratione boni (‘under the aspect of the good’).11 As she puts it, ‘good is the object of wanting’ as ‘truth is the object of judgment’ (p. 77); and she adds: it does not follow from this that everything judged must be true, or that everything wanted must be good [ ... ] the notion of ‘good’ that has to be introduced in an account of wanting is not that of what is really good but of what the agent conceives to be good (Ibid.).12 Now, the desirability characteristic that the thing has for an agent is the reason (real or apparent) for which that agent wants that thing: for example, that it is profitable, or beautiful, fun, restful, it gives one a sense of peace and so on. Since the things we want may be wanted instrumentally (i.e. for the sake of something else), or intrinsically (i.e. for their own sake), the reasons for which we want something may also be instrumental or intrinsic. When something is wanted instrumentally, the reason for wanting it is, precisely, that it is a means of achieving some good (or apparent good) one wants. When something is wanted for its own sake it is still something one wants for a reason, namely that it is good in some respect. For something I want for its own sake is something that I regard as good in itself – where, as I said above, its goodness may relate to a variety of criteria: moral, prudential, aesthetic, hedonic, legal and so on. And the good that I see in what I desire (that it is pleasant, my duty, elegant, an act of friendship, etc.) is my reason for wanting it: it provides the desirability characterization of the thing wanted. And, in as much as I want for its own sake, it is something I see as an instance of some form of the good. Because of this, as Anscombe says, when something is wanted for its own sake, the agent’s answer to what he wants it for will be ‘a desirability characterisation which makes an end of the questions “What for?” ’ (Intention, p. 74).

Maria Alvarez 207

The traditional view that good is the object of wanting endorsed by Anscombe has been disputed, for instance with the objection that one can want what is bad, and want it precisely because it is bad – to adapt the medieval terminology, to want it ‘under the aspect of the bad’.13 I find this objection unpersuasive since it is plausible to claim that if one wants something because it is bad, this is because of some feature that its being bad has that makes it seem good to the agent – though this claim is only plausible, admittedly, so long as ‘good’ is not restricted here to what is morally or ethically good but includes what is good from a variety of perspectives:14 hedonic, aesthetic, instrumental, moral, prudential and so on, that relate to a variety of ends that a person can have.15 For example, one may want to do what is morally bad for the thrill of it, or as a way of defying authority, and thus showing that one is free, or powerful an so on. As Anscombe puts it, Satan’s exhortation in Paradise Lost, ‘Evil be thou my good’ is open, in order for it to be intelligible as a statement of what is wanted, to the question: ‘What is the good of evil?’ And she suggests that a plausible answer might be ‘the condemnation of good as impotent, slavish and inglorious’ (Intention, p. 75). The view that there is a conceptual connection between wanting and the good, namely that, as Anscombe puts it, good is the object of wanting as truth is the object of judgment, seems to me right. For whenever someone’s desirability characterization of what he wants involves something bad for that person (say something that he sees as painful, humiliating, hurtful, etc. to himself) it is possible to make sense of that person’s claim to want it only on the assumption that its being bad in those ways serves, or appears to him to serve, some other end that is itself a form of the good (e.g. health, power, friendship, pleasure, etc.). In Anscombe’s words, ‘the good (perhaps falsely) conceived by the agent to characterize the thing must really be one of the many forms of good’ (Intention, p. 76). And those who question the connection between wanting and the good would have to explain why it is that, if someone claims to want something because it’ll be, say, boring, or painful, or humiliating, we must either imagine that he thinks some good might come out of this bad thing (perhaps he sees it as penance, or as ‘good for the soul’, or takes, as we say, perverse pleasure in such things), or must remained baffled by their avowal. Since I am persuaded by the traditional view, I shall argue on the basis that it is right and that the desirability characterization that an agent would give of what he wants relates, as Anscombe claims it does, to the good broadly conceived. But even if the view that there is a conceptual connection between wanting and the good were wrong or an exaggeration, and the connection turned out to be simply contingent, Anscombe’s point that wanting requires a desirability characterization on the agent’s part would still stand – even if what makes it desirable might occasionally be, as her opponents claim, that what is wanted is ‘bad’.

208 Reasons, Desires and Intentional Actions

The relevance of this claim about wanting, the good, and the need for a desirability characterization to the question of motivating reasons is as follows. I said above that a motivating reason is the reason that motivates me to act because it is something that, in my eyes, makes the prospective action appealing or right for me, when I consider it from some perspective: prudential, hedonic and so on. And we have just seen that what makes an action appealing in my eyes is, precisely, the desirability characterization that I’d give of the action: what good I see in the action, whether in itself or as a means to a further end I desire in itself, which makes me want to perform the action. And this desirability characterization I’d give of the action is both, my reason for wanting to do that thing, and the reason for which I do it – if I do it for a reason. In other words, my reason for doing it is the same as my reason for wanting to do it: whatever good I see in my doing that thing. If this is right, then it seems that my reason for doing something is not that I want to do it but rather whatever reason I have for wanting to do it. So, for instance, suppose I sell my car for a reason. Since I sell it for a reason, then we might say that I wanted to sell my car. But it would be a mistake to conclude that the reason for which I sell my car is that I wanted to sell it. Rather, my reason for selling it is my reason for wanting to sell it, for my reason for selling it will be something that makes selling the car seem appealing to me – the desirability characterization that I’d give of the action of selling my car. Let’s imagine that the desirability characterization I’d give is that it’ll save me quite a lot of money (we might think of this as a prudential instrumental reason). The fact that selling my car will save me money is, to use Anscombe’s phrase, the desirability characteristic that selling my car has for me: what speaks to me in favour of selling it, and that is the reason that motivates me to sell it. My reason for selling it is not that I want to sell it, for my wanting to sell it says nothing about what, in my eyes, is appealing about selling the car. Of course, in these circumstances, I do want to sell the car, but I want to sell it precisely because I shall thereby save money. So, that it’ll save me money is both my reason for wanting to sell it and my reason for selling it. So it seems that my reason for doing A is not (normally) that I want to do A but something about doing A that seems to me good, worthwhile, or otherwise appealing.16

III Doing A because one wants to One may be prepared to accept that when I have a reason to want to do something and do that thing, then, typically, my reason for doing it will be my reason for wanting to do it. Nonetheless, one may argue, there are times when I do something and the only answer to the question why I did it that I could sincerely give is simply that I wanted to do it, perhaps that I felt like doing it. Maybe if others insisted I could come up with some kind of

Maria Alvarez 209

‘desirability characterization’ that doing that thing seemed to have (it was quirky, or fun, etc.) but it seems mere stipulation to say that I did it for that reason rather than, as I would naturally say, that I simply wanted to do it. If so, then this would seem a case where my reason for doing A was that I wanted to do A.17 But, although it is undeniable that we often claim to do things simply ‘because we want’, there is no reason to assume that in saying this, we are giving our reason for doing that thing: the reason that motivated us to do it. For there are other ways of construing cases where we do A simply ‘because we want to’. For example, if when asked why I sold my car I answered that I did it ‘Because I wanted to’, my interlocutor is likely to hear this response as an injunction to mind his own business, not as a statement of my reason for selling the car. So sometimes expressions such as ‘Because I want to’ are used to convey the fact that there is a reason that one does not care to reveal. And there are other possibilities also. For example, suppose that Andrew says that he returned some stolen property ‘because he wanted to’. One thing he might mean by this statement is that nobody forced him to do it. To say this is not to give his reason for returning the money, it is rather to exclude a range of reasons for which he acted – a range of reasons connected to external coercion, commands, duty and son on: it tells us that Andrew did not act because he had been commanded to do so, or threatened, or because he felt obliged or compelled to do it. Another thing he might mean, which is compatible with the above, is that he did not do it for some ulterior motive, for example, to avoid jail or to appear to be good, but rather he did it ‘for its own sake’ – he saw a point in returning the stolen stuff which is not instrumental to some other end. Here, the reason for which Andrew acted was some good that he saw in his acting – for example, that it was the decent thing to do, or whatever, which appealed to him. Finally, in saying that he returned some stolen property because he wanted to, Andrew might mean that he returned the money simply because he felt like it, that is, because he felt an impulse to do so. In saying this, he would be explaining why he returned the money and also probably implying that he did not do it for a reason. For doing something simply because you feel like doing it is not doing it for a reason: your feeling like doing it is not your reason for doing it, although it is the reason why you do it. In denying that wanting to do A is the reason for which one does A, I am not claiming that one never does things just because one wants to, or because one feels like doing them. My claim is that, when one does do something just because one wants to, and there is no more to say about the issue, that action will not be an action done for a reason, even though it will be voluntary, and even though there will be a reason why one did it – namely that one wanted to do it:18 one’s wanting to do it makes it prima facie intelligible that one should do that thing, that is, it explains one’s doing it, but

210

Reasons, Desires and Intentional Actions

that does not mean that it is the reason that motivated one to do it. (I return to this below, in Section VI). It should be noted that doing something because you want to, or because you feel like doing it, is different from doing something because you like doing that thing. If you give as your reason for doing something that you like doing that thing, then your reason for doing that thing is that you will enjoy it, which is perfectly intelligible as your reason for acting – though, as Anscombe says (Intention, p. 75), there are limits here. The limits are not those of morality or propriety but of intelligibility, for sometimes it is difficult to see what the pleasure of doing a particular thing might be – for instance, the pleasure of sitting through an eight-hour-long speech by Fidel Castro. But, though this is a case when you act for a reason, it is quite different from doing something simply because you want to or feel like doing it, which is not. Let me mention a kind of example that I think is the closest to a case where wanting to do something might seem to be your reason for doing it. Suppose that you have a persistent want to do something – to remove a thread from the jacket of the person sitting in front of you at a lecture. Now, you may or may not be able to give a desirability characterization for it (it could be an aesthetic desirability characterization, or the thought that the world will be a tidier place if you remove the thread). But suppose that though initially you decide against doing it (perhaps you don’t know the person at all), the desire to do it is so persistent and is so distracting that eventually you conclude that unless you do it you won’t be able to concentrate on the lecture, and you end up removing the irritating thread. Here it seems plausible to say that your only reason for removing the thread was that you wanted to do so, and indeed that you wanted to do it very much. But in fact I think that is not quite right. Your reason for removing the thread, if we agree that you did it for a reason, is rather captured by the following. You judge that you have good reason not to remove the thread – it may disturb the person in front of you – you may decide that you have better reason to do it because unless you do it, you won’t be able to concentrate on the lecture. That, namely that it’ll enable you to concentrate of the lecture, is the desirability characterization for you that removing the thread has (given that we agree that you did it for a reason), and not the mere fact that you wanted to do it; the fact that you very much want to do something does not make doing that thing desirable, it makes it desired – which is quite a different thing.19 To sum up, none of the interpretations examined of what doing something ‘because one wants to’ are cases where that one’s reason for acting was that one wanted to do it: things done ‘because one wants to’ are either things done for some reason that has not been stated, and is perhaps obscure and subconscious – and hence, a fortiori, not for the reason that one wanted to do them, or things not done for a reason – and hence, and

Maria Alvarez 211

also a fortiori, the fact that one wanted to do it was not one’s reason for doing it.

IV Doing A because one wants to do B Defenders of the view that wanting is a motivating reason might think that my arguments have been aimed at the wrong target. For they do not claim that, typically, my reason for doing A is that I want to do A. Rather, their claim is that my reason for doing A is, typically, that I want something else (perhaps x, or to do B), and believe that I shall get that thing I want (shall get x, or to do B) by doing A.20 So, according to them, in the example above my reason for selling my car would be that I want to save money and believe that selling my car is a way of saving money. This suggestion seems plausible but on examination it is also unconvincing. My reason for selling the car, I suggest, is not that I want to save money but rather that selling the car will save me money. For that is the aspect of my action of selling the car that speaks to me in favour of doing so – it is the desirability characteristic it has for me. It is true that this is a reason that motivates me only because I want to save money – that is, because saving money has some desirability characteristic for me. But this does not suggest that my wanting to save money is the reason that motivates me to sell my car. What it suggests is rather that my wanting to save money is a condition for those other things to be reasons that motivate me. But we must not conflate a condition with what it is a condition for. That it will save me money was a reason that motivated me to sell the car only because I wanted to save money but it does not follow that my wanting to save money is also my reason for selling the car. This may be difficult to see because we often say ‘my reason for doing A was that I wanted to do B’. For instance, my reason for running was that I wanted to catch the train; my reason for studying is that I want to become a barrister, and so on. And this certainly appears to imply that my wanting the end was my reason for taking the means. However, I want to suggest that those expressions do not give the reason that motivates the agent to act; rather they provide the agent’s goal: for example, to catch the train, to become a barrister and so on. Outside of philosophy, goals, as well as intentions and motives are called ‘reasons’. However, when trying to map out the conceptual territory of reasons, desires and so on we must see what distinguishes these concepts. And these apparent expressions of reason are really statements of purpose or goals. Consider a peculiarity that these expressions have. If I say that my reason for running was that I wanted to catch the train, then it follows that I ran in order to catch the train. However, if I say that my reason for running was that I was late (and running was my only chance of catching the train), then it does not make sense to say that I ran in order that I was late. That is

212

Reasons, Desires and Intentional Actions

because the first kind of expression, unlike the second, does not really state a reason but a goal, namely to catch the train. And my reasons for running relate to the means to achieve that goal and to the value that goal has for me. (I return to the notion of a goal shortly.) So some things are motivating reasons only for those agents who want certain things, that is, for those agents who have certain goals. Nonetheless, this does not make the goals those agents have the reasons that motivate them to act. So far I have argued against the Humean view that desires are reasons for which we act. But, as I said in the Introduction, I think that Humeans are right to insist on the importance of desires in motivation and in the explanation of action. In the remainder of the essay I shall explain what role desires play in motivating action, and in the explanation of action. In order to do so, I need to make an important conceptual distinction concerning the term ‘desire’.

V Desires and motivation In the Introduction I articulated the Humean position as the view that desires are reasons for which we act. But so stated, there is an ambiguity in the claim, corresponding to an ambiguity concerning the term ‘desire’, which arises from the fact that the term ‘desire’ has two possible uses: one to refer to someone’s desiring something, and the other to refer to what is desired. 21 And although the question I intend to address now is not whether desires are motivating reasons (since I concluded above that they are not) but what role, if any, they play in motivation and in the explanation of action, the ambiguity inherent in the term ‘desire’ infects those issues too. These two uses of the term ‘desire’ just mentioned correspond to the act/ object distinction; a distinction that can be brought out by considering that there are things which are true of the one (i.e. the object) that need not be true of the other (the act). For instance, what A desires, let’s say, to own a limitededition Ferrari, may be expensive, beautiful and perhaps very desirable, while A’s desiring it need not be any of those things. Consider the difference between an intense desire and an unattainable desire. This difference is reflected in the fact that a desire of each kind is characteristically qualified by means of an adverb and an adjective, respectively. So, typically, we use an adverb if we are qualifying the act – that is, the desiring; and tend to use an adjective when qualifying the object – that is, what is desired. For instance, if I have an intense desire to see Mt Everest, it is my desiring, not what I desire, that is intense. And this is expressed by means of an adverb that qualifies the act: ‘my intensely desiring to see Mt Everest’ (and not my desiring to intensely see Mt Everest). However, if I have an unattainable desire, for example, to win an Olympic

Maria Alvarez 213

medal, then it is what I desire, not my desiring it, that is unattainable, and this is expressed by means of an adjective that qualifies what is desired: ‘Her desire, namely to win an Olympic medal, is unattainable’. This difference between desiring something and what is desired is a familiar distinction which nonetheless tends to be neglected in discussions of desires and reasons. And yet I think the distinction is crucial in order to understand the role of desires in motivation and in the explanation of action. Before we do this, however, we need to examine what these things desired are. The things we desire, though seemingly very varied, can be thought of under two main headings: (i) for oneself, or for someone or something else, to F (where F-ing need not be an action); or (ii) for oneself, or for someone or something else to be a certain way (or the negation of these: not to F, not to be a certain way). So they can all be expressed by means of an infinitival clause: I desire x to F, or I desire x to be F. Examples of these are: I can desire or want to be more patient; to drink water; to own a horse; to run a mile in four minutes; my daughter to come home; cold-callers to stop calling; tomorrow to be a sunny day, poverty to become history and so on. (An apparent exception is that we may desire a thing or a person, but it is only an apparent exception because to desire a thing or a person is to desire to own, or consume, or somehow use or interact with that thing or person – i.e. to desire to F it.)22 Now, what is desired can play the role of a purpose or goal for the sake of which we act: that is, given a goal and certain conditions, one may act in the way that one believes will bring it about that one’s goal is realized, whether that goal is for one to do something, or for others, or for things to be a certain way and so on. Suppose that I want my neighbours to turn down their music. Then, my goal in acting will be that my neighbours turn down their music, and I shall act so as to bring it about that they do, for example, by knocking on the wall, or going round and asking them to do so. So what is wanted, though not a reason, does motivate – it is that for the sake of which we act. And as Aristotle put it, ‘t `o o’ r«kto´n’, what is wanted, is the starting point of motivation and of practical reasoning. Thus, Aristotle asks What is the source of (local) movement for us; and his answer is that Both of these then are capable of originating local movement, thought and appetite; thought, that is, which calculates means to an end, i.e. practical thought (it differs from speculative thought in the character of its end); while appetite is in every form of it relative to an end; for that which is the object of appetite is the stimulant of practical thought (De Anima, Book I, section 10; 433a, 10–18. My italics). So, according to Aristotle, on the one hand there are goals, which are ‘the objects of appetite’ and ‘the stimulant of practical thought’. On the other

214

Reasons, Desires and Intentional Actions

hand, there is practical thought, which involves means-ends reasoning about how to achieve those goals, as well as about how the satisfaction of one goal may affect other goals one also has.23 In the case of humans, not every thing wanted is made into a goal of action. I may feel a desire to eat but reason that I ought not to, because I need to fast for an operation. Or I may have a desire to visit Rome but reason that it will be too hot and decide against visiting it. Here these things are not adopted as a goal (at least pro tem). However, when something wanted is adopted as a goal, then it can motivate, and be the beginning of practical reasoning about how to achieve that goal; as Anscombe, para’ phrasing Aristotle, puts it: ‘the arxh ´ (starting point) [or practical reasoning] is t`o o’ r«kto´n (the things wanted) (Intention, p. 63). And she adds: The rôle of ‘wanting’ in the practical syllogism is quite different from that of a premise. It is that whatever is described in the proposition that is the starting-point of the argument must be wanted in order for the reasoning to lead to any action (Intention, p. 66). Thus, what is desired is clearly crucial in motivation. It is important to note here that Humeans might agree with this, for this claim is one of their main contentions. For example, in his defence of the Humean thesis about motivation, David Lewis says: A Humean thesis about motivation says that we are moved entirely by desire: we are disposed to do what will serve our desires according to our beliefs. If there were no desires to serve, we would never be moved more to do one thing rather than another. (‘Desire as Belief’, Mind, 97, 1988, pp. 323–32, p. 323). The Humean mistake is to think that from this we can conclude the following two theses. One, that it is states of desire (rather than the ‘objects’ of desire) that motivate us to act. Second, that those states of desire are motivating reasons. It is mistaken because the Humean thesis that ‘we are moved entirely by desire’ seems compelling only if we interpret ‘desire’ to mean ‘what we desire’ (as indeed Lewis’s passage suggests). And we have seen that the role of what is desired in motivation is that of a goal and not that of a reason. So if one has a desire which one has made into a goal, then one is motivated to act but what motivates one to act is not having the desire, but the desire one has, together with whatever reason one might have for having that desire, and for how to act in pursuit of that desire. So we might say that desires, in the sense of what is desired, motivate actions but not because they are reasons to act but because they are the goals we pursue in acting. On the other hand, desiring something does not motivate either as a reason, or as a goal. The state of motivation that we are

Maria Alvarez 215

in when we desire something can be said to be a state of being inclined to act in order to obtain what we desire. If I desire to eat, or to become Prime Minister, then I am inclined to act so as to satisfy those desires. What is desired (to eat, to be Prime Minister) is motivating: it is, or can be unless one decides against pursuing the desire, the goal towards which our actions aim. But the state of desiring is not; though it is a state of being motivated. Sometimes, wanting something is characterized as a motivating reason because, it is argued, wanting something is a ‘motivational state’.24 But the claim that wanting something is a motivational state is somewhat ambiguous, as it can be read as a claim that wanting something is a motivating state, or as a claim that wanting something is a state of being motivated. Consider an analogy. As Aristotle points out, the term ‘healthy’ can be applied to something in two different though related senses: a thing can be said to be healthy on account of its promoting health, or on account of its having health. Thus exercise is said to be healthy in the first sense, while people are (generally) said to be healthy in the second. For my wanting something to be a motivating reason, my wanting something must be motivating in the first sense: it must be a motivating state, a state that motivates me to act. However, it seems that a state of wanting something is motivational in the second sense: it is a state of being motivated to act in the way, if any, that one believes will bring it about that one gets what one wants. Thus, a state of being motivated is not what motivates and a fortiori it is not a motivating reason.

VI

Desires and the explanation of action

I have already noted that Humeans are right to emphasize that desires are important in motivating action but I have also explained that there is an ambiguity in the term ‘desire’ and that the Humean thesis is plausible only if we construe it in terms of what is desired. So, contrary to what Humeans argue, this prominent role of desires in motivation has no tendency to show that our desires, conceived as states of wanting, are reasons for which we act. Nonetheless, Humeans are right in saying that desires play an important role in motivation, and this is one of the truths contained in the Humean position. The Humean emphasis on desires contains another truth. But just as the previous one could not be brought out without disambiguating the term ‘desire’, this one cannot be made clear without making another conceptual distinction – though I should note that this distinction is not one I have ever seen made and may prove more controversial. In the literature, motivating reasons are often identified with explanatory reasons – the reasons that explain why we act. In my view, that identification is most unhelpful because a reason is called ‘motivating’ or ‘explanatory’ respectively on different grounds: namely, on the grounds that it motivates,

216

Reasons, Desires and Intentional Actions

or on the grounds that it explains an action, respectively. It is true that the same reason can do both – that is, the same reason can motivate someone to act and explain his action. For example, that I enjoy concerts could be a reason that motivates me to go to a concert and it could also be the reason that explains why I go to a concert. In that case the reason, that I enjoy concerts, will be both a motivating reason (a reason for which I act) and an explanatory reason (a reason that explains my action). But the fact that the same reason can play both roles does not obliterate the difference between the two roles, and hence between the two kinds of reason. Because of this I think that we ought to distinguish between three kinds of reason: normative reasons, which are the reasons for which one ought to act; motivating reasons, which are the reasons for which one actually acts; and explanatory reasons, which are the reasons cited in explaining one’s action. The same reason may play all three roles for a particular action but it need not. To see the difference between the motivating and the explanatory roles that reasons play more clearly consider that, although a reason that motivates an action can always explain it, the converse is not true: the reason that explains an action need not be the reason that motivated the agent to act. For example, the reason that explains why Fred gives much of his money to charities may be that he’s a generous man; but that he’s a generous man is not the reason that motivates Fred to give money to charities (the fact that he is generous is not what, in his eyes, makes his giving money right or appealing). And, a different kind of example, the reason that explains why Sarah bought a new mobile phone is that she thought hers had been stolen; but the reason that motivated her was not her thinking that her phone had been stolen. And yet a different kind of case, the reason why Angie missed the party may be that she forgot all about it, but that she forgot is not a reason that motivated her to miss the party. So in these cases, the reasons that explain why Fred, Sarah and Angie f (gives money to charity; bought a new mobile phone; missed the party) are not the reasons that motivated them, and therefore they are not motivating reasons.25 They are only explanatory reasons: they explain why these agents did what they did (or omitted to do what they omitted to do). Therefore, I think we ought to use each of those labels depending on whether we are referring to a reason in so far as it plays a motivating or an explanatory role. I, at any rate, shall do that here, and so I shall call a reason for which someone acts a ‘motivating reason’; and a reason why someone acts an ‘explanatory reason’, without assuming that these will always be the same for a particular action – but allowing that they may be. Humeans are impressed by the fact that when we explain intentional actions, we often, or perhaps always, implicitly or explicitly make reference to the fact that the agent wanted certain things. From this they tend to conclude that one’s wanting something is (part of) the reason for which

Maria Alvarez 217

someone acts – that is, is part of a motivating reason. However, it ought to be clear from the discussion so far that what the point just mentioned suggests is that an agent’s wanting something is (part of) the explanatory reason – the reason that explains why that agent acted.26 So neither the Humean thesis that without desires there is no motivation, nor the view that actions are explained by reference to the fact that the agent desired something, is grounds for concluding that wanting something is part of the reason for which one acts, that is, that it is part of a motivating reason. As we have seen, although desires play an important role in motivating and in explaining actions, they are not motivating reasons.

Notes I am grateful to the Arts and Humanities Research Council who funded the period of research leave during which I completed the final draft of this essay. I am also grateful to my colleagues at Southampton and to the editor of this collection for their very helpful comments on earlier drafts. 1. Motivating reasons are premises in the agent’s implicit or explicit practical reasoning. I shall return to this below. 2. This strong sense of desire corresponds roughly to Aristotle’s ‘epithymia’ and to the Medieval concept of concupiscentia. G. F. Schueler, in his Reasons and Purposes: Human Rationality and the Teleological Explanation of Action, Oxford University Press, 2003, ch. 2, calls what is desired in the stronger sense ‘proper desires’. Perhaps this is justified because in ordinary use we tend to restrict the term ‘desire’ for the strong sense, and use ‘want’ when the weaker one is at issue, although this usage is not consistent. 3. So I shall not rely on an argument against the Humean view defended by, for example, Schueler (Schueler, 2003, ch. 2), that seems to depend on construing the Humean claim to involve a stronger sense of ‘want’. 4. Most philosophers accept a point made explicit by Anscombe in Intention, Blackwell, 1957, that actions are intentional only ‘under a description’. Although this way of talking is widespread, I believe Anscombe’s point is better put using the adverbial phrase ‘A f-ed intentionally’ since, as Anscombe herself complains, the phrase ‘under a description’ encourages confusion, for example, it encourages talk of things happening, or being done, or being performed ‘under descriptions’ but, as she says: no natural sense suggests itself for ‘happening’ or ‘being done’ or ‘being performed’ together with the phrase ‘under the description d’. At best, the phrase seems redundant – one might say: What happens happens under every description that is true of it! (Anscombe, ‘Under a Description’, in her Metaphysics and the Philosophy of Mind: Collected Papers Volume II, Blackwell, 1981, pp. 208–19, p. 209). Nonetheless, I shall occasionally use Anscombe’s phrase while making sure that such confusion is avoided. 5. For discussions of this issue see R. Hurtshouse, ‘Arational Action’, Journal of Philosophy, 83, 1991, pp. 291–5, and J. Raz, ‘Agency, Reason and the Good’, in his Engaging Reason: On The Theory of Value and Action, Oxford University Press, 1999, pp. 22–45.

218 Reasons, Desires and Intentional Actions 6. This would involve a rejection of the view that there is an action only when there is some description of what the person did ‘under which’ it was intentional, for on this view a person could act, that is perform an action, even though there is no description that applies to her action such that, so described, it would be right to say that she acted intentionally. 7. On this criterion, there may be actions of a different kind that are not intentional, for example things one does because one feels like doing them, such as whistling, or giving a little jump. Many claim that such things are done for the reason that one felt like them. But, I shall argue, it is plausible to say that such actions are not done for a reason at all. If that is right, then these would be things done voluntarily but not intentionally. 8. Notable among them, because of his influence in establishing this view, is Donald Davidson. He encapsulated the Humean view in his characterization of a ‘primary reason’ for an action in his paper ‘Actions, Reasons and Causes’. According to Davidson, C1. R is a primary reason why an agent performed the action A under the description d only if R consists of a pro attitude of the agent towards actions with a certain property, and a belief of the agent that A, under the description d, has that property (Davidson, ‘Actions, Reasons and Causes’, 1963, reprinted in his Essays on Actions and Events, Oxford University Press, 1980, pp. 1–19, p. 4). Davidson holds that a primary reason, that is, the reason for which someone does something, is also the cause of that action. 9. Most of those who endorse the Humean position, think of these reasons as mental states of the agent’s but I shall not assess that claim here. 10. For an excellent defence of this claim, see W. Quinn, ‘Putting Rationality in its Place’ in his Morality and Action, Cambridge University Press, 1993, pp. 228–55. See also Raz, 1999; and T. Scanlon, What we Owe to Each Other, Harvard University Press, 1998. 11. Thomas Aquinas, Summa Theologiae, ed. Thomas Gilby, Blackfriars, 1960–73, 1a2ae, q.8, a.1. See also Aristotle’s Physics, II, 3, 195a26; and Nichomachean Ethics, I, 1, 1094a3. The Complete Works of Aristotle. The Revised Oxford Translation, ed. Jonathan Barnes, Princeton University Press, 1984. 12. Aristotle says that the object of desire ‘may be either the real or the apparent good’ (De Anima III, 433a28). And Francis Bacon says that ‘There is no man doth a wrong for the wrong’s sake, but thereby to purchase himself profit, or pleasure, or honours, or the like’ (Book IV, Of Revenge Francis Bacon, Essays, Civil and Moral, The Harvard Classics, 1909–14). 13. On this see M. Stocker, ‘Desiring the Bad: An Essay on Moral Psychology’, Journal of Philosophy, 76, 1979, pp. 738–53, and ‘Raz on the Intelligibility of Bad Acts’ in Wallace et al., 2004, pp. 303–32; and Raz’s ‘Agency, Reason and the Good’ in Raz, 1999. See also D. Velleman, ‘The Guise of the Good’, 1992, reprinted in Velleman, 2000, pp. 99–122; and U. Heuer, ‘Reasons for Actions and Desires’, Philosophical Studies, 121, 2004, pp. 43–63. 14. So, for instance, Velleman’s objection to Anscombe’s view seems to me to depend on implicitly thinking that the good must be some kind of ‘ethical good’, for he says: ‘Anscombe’s Satan can desire evil only by judging it to be good, and so remains at heart, a lover of the good and the desirable – a rather sappy Satan’ (Velleman, 2000, p. 119). But this objection only has bite, I think, if Satan must remain a lover of the ethically good as only that would make him a ‘sappy Satan’ but not if he is a lover (and desirer) of evil, that is, if he has, as

Maria Alvarez 219

15. 16.

17. 18.

19. 20. 21.

22. 23. 24.

25.

26.

the Satan of Paradise Lost (John Milton, Paradise Lost, Book IV, Penguin, 2003, 110) seems to have done, made evil his good. This relates to another medieval notion that Anscombe endorses, namely that bonum est multiplex. If this is right, how can we respond to the problem raised by akrasia, where one acts for a reason that is not the reason that makes the thing done appear right or worthwhile: I see two possible responses here? either (i) we act for a reason – that the ‘irrational’ act is pleasant – even though it is a reason we think is less good than the reason not to do it; or (ii) we do not act for no reason, but there is a reason why we act: that doing the thing one fails to do requires effort while that the thing one ends up doing doesn’t. A.R. Mele, Motivation and Agency, Oxford University Press, 2003, pp. 82ff has an argument to this effect. Perhaps some people, persuaded by Freud, might claim that things we do ostensibly for no reason, or claim to do just because we feel like doing them, are in fact always done for some reason – but they are reasons that the agent is unable (psychologically, not epistemically, speaking) or unwilling to identify; reasons that may require a great deal of soul-searching or professional help to be unearthed (a project which in most cases may be felt to be neither necessary nor profitable). If this is right, then there may be reasons for which we do those things but, if so, our reasons for doing those things are not that we feel like doing them or that we want to do them but rather those other facts that psychoanalysis will unearth. A variant of this case is examples where the fact that one wants to, for example, marry one’s mother is one’s reason for visiting a psychoanalyst. Which roughly conforms with Davidson’s characterization of a ‘primary reason’. The same is true of the term ‘belief’: it can be used to refer to my believing something, or to what I believe. I discuss the relation between beliefs and reasons in ‘Reasons and the Ambiguity of “Belief” ’, Philosophical Explorations, 11, 2008, pp. 1–13 On this issue see A.J.P. Kenny, Action, Emotion and Will, Routledge, 1963, 112ff. On this see also Aquinas Summa Theologiae, 1a2ae, q.1 For instance, Audi says that wanting may be considered as ‘the most representative motivational element’ (R. Audi, The Architecture of Reason, Oxford University Press, 2001, p. 66); and Mele says that states of desiring are states ‘that encompass motivation’ (Mele, 2004, p. 16). In Angie’s case there was no motivating reason and her missing the party was not something she did intentionally. In Sarah’s case, she thought she acted for a reason but she didn’t. In Fred’s case, we can assume that he did act for a reason but we have not been told it explicitly. Indeed, it is precisely because being in a state of desiring something is normally being in a state of being inclined to act that it is possible to explain why someone acted by citing the fact that they desired something, that is, that they were in that state of motivation.

12 A Niggle at Nagel: Causally Active Desires and the Explanation of Action Charles Pigden

1

Introduction

My aims in this essay are less modest than the title suggests. True, the essay is a niggle at Nagel, that is, a nit-picking criticism of one very brief passage in Possibility of Altruism (Nagel 1970, pp. 29–30). But, brief as it is, that passage has had an enormous and, in my view, an undeserved influence. Philosophy, said Wittgenstein, is the struggle of our intellect against the bewitchments of language. More often, so it seems to me, it is the struggle of our intellects against the bewitchments of bad arguments. And a great many eminent philosophers have allowed their intellects to be bewitched by the bad argument in this passage. It is a ring of power whose magic has been used to erect mighty Barad-dûrs of intellectual error. They are very edifying errors on the whole – the Barad-dûrs appear to their proprietors to be Minas-Tiriths of reason and virtue – but the errors are errors nonetheless. My aim in the essay is to destroy the ring and perhaps to bring those Baraddûrs of error crashing down. Whether I succeed in the second aim depends upon the degree to which those Barad-dûrs are themselves dependent on the ring. But of my success in the first aim I have no doubt. The ring of Nagel’s argument can definitely be destroyed. For despite its mesmeric power, it is definitely a dud. The most it can prove is that a certain argument for Humeanism – and a rather silly one at that – is a failure. And it can only prove that on condition that our ordinary notions of desire and want are not causal concepts – a condition which in my opinion is not met. Nagel’s argument cannot prove the substantive conclusion that Humeanism is false or that a causally active desire for the end Y is not required if the belief that X is the means to Y is to motivate the performance of X. And if it cannot establish this substantive conclusion, it cannot provide a foundation for the Barad-dûrs of error.

220

Charles Pigden

221

First the ring, then the Barad-dûrs, then finally the Cracks of Doom. It is my mission to dissolve the ring of Nagel’s argument in the lavas of logical analysis.

2

Nagel’s argument

After making his famous distinction between motivated and unmotivated desires Nagel goes on to state: The claim that a desire underlies every act is true only if desires are taken to include motivated as well as unmotivated desires, and it is true only in the sense that whatever may be the motivation for someone’s intentional pursuit of a goal, it becomes in virtue of his pursuit ipso facto appropriate to ascribe to him a desire for that goal. But if the desire is a motivated one, the explanation of it will be the same as the explanation of his pursuit, and it is by no means obvious that a desire must enter into this further explanation. Although it will no doubt be generally admitted that some desires are motivated, the issue is whether another desire always lies behind the motivated one, or whether sometimes the motivation of the initial desire involves no reference to another, unmotivated desire. Therefore, it may be admitted as trivial that, for example, considerations about my future welfare or about the interests of others cannot motivate me to act without a desire begin present at the moment of action. That I have the appropriate desire simply follows from the fact that these considerations motivate me; if the likelihood that an act will promote my future happiness motivates me to perform it now, then it is appropriate to ascribe to me a desire for my own future happiness. But nothing follows about the role of the desire as a condition contributing to the motivational efficacy of those considerations. It is a necessary condition of their efficacy to be sure, but only a logically necessary condition. It is not necessary either as a contributing influence, or as a causal condition. (Nagel 1970, pp. 29–30) For Nagel, motivated desires are, very roughly, the desires that are generated as the result of means/end calculations. This is rough because an unmotivated desire to have a good time tonight can generate a motivated desire to go to the movies, not because going to the movies is a means to having a good time, but because going to the movies is constitutive of having a good time. So motivated desires are derived desires that we arrive at by some process of deliberation, whereas unmotivated desires are the starting points in the deliberative process.1 These are the ‘passions’ to which reason is supposed to be the slave, since, according to the Humeans, the role of reason (our beliefforming faculty) is to show us the way to realize our unmotivated desires.

222

A Niggle at Nagel

But Nagel wants to argue that although when I do X intentionally, I want (or have the desire) to do X, the motivated desire to do X need not be the product of an unmotivated desire in any substantial sense. Nagel seems to be conceding two things: (a) that intentional action is driven by desire (so that whatever I do intentionally, I want to do) and (b) that it is some kind of conceptual truth – even a necessary one – that if I do X because of a belief that doing X is likely to promote Y, then I desire Y. But, having made these concessions, he then pulls the rug from under his Humean opponent. Because thesis 2) I have a desire to bring about Y, follows from thesis 1) I do X because I believe that doing X is likely to bring about Y, thesis 2) means no more than thesis 1). And if thesis 2) means no more than thesis 1) – if, that is, it is conceptually contained in thesis 1) – then the unmotivated desire to bring about Y that it ascribes to me is, as it were, a logical or conceptual shadow, not a substantial ‘biffy’ something or an independent causal factor in the situation. Of course, I may have a genuine or causally active desire to bring about Y – Nagel’s argument does not exclude unmotivated desires with genuine biff – but there is no reason to think that such an unmotivated desire underlies every action. To be more precise, what Nagel’s argument purports to prove is that although it is conceptually necessary that if anyone does X in the belief that doing X will bring about Y, then they have a desire to bring about Y, this conceptual necessity is quite compatible with the non-existence of a biffy or causally active desire to bring about Y. Thus Nagel is arguing for the following thesis: N*) A person can be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y. So although people often act on the belief that their actions will bring about some end, we only have reason to believe in desires for those ends if those desires are construed as ‘consequential’; that is, logical or conceptual shadows cast by the belief and the fact that the agent has acted on the belief. Indeed, it is very clear that Nagel wants to go further. He seems to think that, in many cases, when someone is motivated to do X by the belief that X leads to Y, there is, in fact, no causally active desire for Y, and that it is not necessary in any sense that there should be such a causally active desire. Thus he would deny that it is naturally necessary, in virtue of the human constitution, that if you are motivated to do X by the belief that X leads to Y, then you have a causally active desire for Y. What is a little unclear – it

Charles Pigden

223

seems to puzzle Dancy for instance – is whether Nagel wants to give motivated desires the same treatment that he gives to unmotivated desires. For an analogous argument is clearly in the offing. Since thesis 4) I want or desire to do X, follows from thesis 3) I am doing X intentionally, Thesis 4) means no more than thesis 3). And if thesis 4) means no more than 3) – if, that is, it is conceptually contained in thesis 3) – then the motivated desire to do X that it ascribes to me is, as it were, a logical or conceptual shadow, not a substantial something or an independent causal factor in the situation. However that may be, it is clearly Nagel’s claim that beliefs alone can motivate by producing motivated desires, and therefore, ultimately, action. (It is right to say that the agent has – indeed must have – a [relatively] unmotivated desire for the end he is trying to bring about, but this is a concession that concedes nothing of substance, since his argument shows that this desire need be nothing more than a conceptual shadow.) Why does Nagel think this a happy conclusion? Well, if beliefs could only motivate with the aid of substantial or biffy desires, then a belief by itself would not constitute a motivating reason for anyone to do anything (or perhaps we should say that it would constitute a reason for some people but not for others). The belief that there is ice-cream in the fridge would constitute a reason for action for me, but not for you, because I want ice-cream and you don’t. And by extension the fact believed (or the fact that in some sense one ought to believe) would not constitute a reason either, unless it connected with the agent’s wants. So the fact that you are in pain (say) would not constitute a reason for me to alleviate your suffering unless I happened to like you or to care about you, which I might very well not do without being irrational. For Nagel, the idea that I could be callous without making some kind of mistake is intolerable. So in order to make room for the possibility of altruism – or rather to prove that altruism is rationally required – he has to suppose that biffy desires are not required to motivate, which is exactly the conclusion of his argument. There is another, Kantian, reason why one might object to the Humean thesis that beliefs cannot motivate without the aid of biffy desires. Kant seems to think that it is a criterion of a genuine moral fact that is it necessarily motivating to any rational being that becomes aware of it, whatever that being’s inclinations or desires. But if beliefs cannot motivate without the aid of biffy desires (and if there are no desires that are constitutive of rationality) then there may be no facts that are necessarily motivating to any rational being. For given any moral ‘fact’, we always run the risk of

224 A Niggle at Nagel

meeting a rational being who regards it with indifference even though he or she is well aware of it, because he, she or it lacks the relevant desires. And if there are no facts that are necessarily motivating to any rational being then, by the Kantian criterion, there are no moral facts. (Icy shivers down the spine!) But if beliefs can motivate without the aid of biffy desires then it seems we can forestall this appalling possibility. Our deviant rational being – I tend to think of a rational mantis from Mars with a taste for human brains – would not require biffy desires to be motivated by the moral facts. Belief or awareness would suffice. Thus the mantis’s awareness that is wrong for her to eat my brains – however, exactly that belief is to be cashed out – could motivate her to refrain, despite the fact that she considers my brains very tasty, and despite the fact that she has no biffy desire to consider my interests or even to do the right thing. And if, despite all that, she remains unmotivated, then she would not be really rational, since a propensity to be motivated in right way is constitutive of [practical] rationality. Whatever the fate of my brains, the moral facts would be safe – safe, that is, from the threat of non-existence. Thus there seems to be a lot riding on Nagel’s argument. The possibility of a rationally mandatory altruism – and maybe the very existence of moral facts – would appear to be at stake.

3

Barad-dûrs of error

Nagel’s argument was endorsed by Philippa Foot in her article, which, ‘Reasons for Actions and Desires?’ (reprinted in Foot 1978, pp. 148–156): Yet surely we cannot deny that when a man goes shopping today because otherwise he will be hungry tomorrow he wants, or has a desire to, avoid being hungry? This is true, but an analysis of the use of the expressions such as ‘wants’ and ‘has a desire to’ in such contexts shows that these ‘desires’ cannot be the basis of the reason for acting. Thomas Nagel in an excellent discussion of prudence has explained the matter in the following way: [there follows the crucial passage from Nagel]. What we have here is a use of ‘desire’ which indicates a motivational direction and nothing more. One may compare it with the use of ‘want’ in ‘I want to f’ where only intentionality is implied. Can wanting in this sense create a reason for acting? It seems that it cannot (Foot 1978, p. 149). Foot’s claim that Nagel’s argument is based on ‘an analysis of the use of the expressions such as ‘wants’ and ‘has a desire to’ in such contexts’ is surely somewhat exaggerated. Whatever his faults, Nagel is not really an analyst of the use of expressions – he has much bigger, metaphysical fish to fry. But Foot, I think, is trying to pay him a rather old-fashioned compliment. Good philosophy consists in analysing the use of expressions (which is what she

Charles Pigden

225

was taught when she was young); Nagel has produced some good philosophy; therefore he must have been analysing the use of certain expressions. The upshot, however, is plain. Since desires are merely ‘consequential’ and consequently mere shadows, they cannot constitute reasons for action. Foot continues to think highly of Nagel, down to the present day. In her ‘Locke, Hume and Modern Moral Theory’ (Foot 2002, pp. 117–145), in which she explicitly attacks the thesis that biffy desires are required to explain action, she praises ‘the pioneering work on the subject of action and desire” in Thomas Nagel’s (1970) The Possibility of Altruism, citing in particular chapter V, which is where the crucial paragraph occurs. In Natural Goodness she proclaims her admiration once again: That prudence on its own can motivate seems to me to have been demonstrated by Thomas Nagel many years ago in The Possibility of Altruism, chapters V and VI. If philosophers still insist that only the presence of what they call a ‘conative state’ can explain an action, they are, to my mind, ignoring this lesson (Foot 2001, p. 61n). This footnote occurs in a passage where she is attacking the thesis that if someone does what they think is right because they think it right, then we have to posit a ‘conative state’ of wanting to do the right thing in order to adequately explain the action. No doubt they did want to do the right thing, but this is just another way of saying that they did it because they thought it was right. More generally, she is arguing that you can have a practical reason to do something even if you don’t have a biffy desire to do it. Thus is it no objection to her conception of natural goodness that considerations about what is naturally good may have no bite for the rational gangster. If the gangster is unmoved by considerations about natural goodness, then what this shows is that he isn’t really rational after all, since being rational entails being moved (or being moveable) by such considerations. Foot has a lot of praise for Quinn (also an admirer of Nagel) who opened her eyes to the possibility of defining her way to victory in this cheap and easy manner. Quinn’s method, to be sure, has many advantages for the moral realist – they are analogous to the advantages of theft over honest toil. How far is the theory of Natural Goodness reliant on Nagel’s argument? I am not at all sure. You can certainly be a virtue-theorist and a Humean about motivation, since Hume himself was both a Human about motivation and a virtue-theorist – and that without any obvious inconsistency. However, if we don’t cheat by defining [practical] rationality as a propensity to be moved by considerations about morality or natural goodness, if we insist that a belief or a fact does not constitute a reason for someone to act unless it marries up with causally active desires, and if we insist that an analysis of the right and the good is not adequate unless it provides reasons for action to all or most human beings, then Foot’s theory is inadequate. For she fails

226

A Niggle at Nagel

to supply a motivating reason to be ‘naturally good’.2 And it is not just gangsters she has to worry about either. I have no desire to be ‘naturally’ good in her sense (though I do want to do some of the things she regards as good or right). Foot retracted her opinion that moral requirements are hypothetical imperatives in part because of McDowell’s criticisms in ‘Are Moral Requirements Hypothetical Imperatives?’. McDowell notes her allegiance to Nagel, endorses Nagel’s argument, and then goes on to turn it against her (McDowell 1978, p. 15): Suppose for instance, that we explain a person’s performance of a certain action by crediting him with awareness of some fact that makes it likely (in his view) that acting in that way will be conducive to his interest. Adverting to his view of the facts, may suffice, of its own, to show us the favourable light to which his action appeared to him. No doubt we credit him with an appropriate desire, perhaps for his own future happiness. But the commitment to ascribe such a desire is simply consequential on our taking him to act as he does for the reason we cite; the desire does not function as an independent extra component in a full specification of his reason, hither to omitted by an understandable ellipse of the obvious, but strictly necessary in order to show how it is that the reason can motivate him. Properly understood, his belief does that on its own. (See, Nagel 1970, pp. 29–30) Though couched as an ad hominem attack on Foot, this paper is in fact the first step in the construction of McDowell’s own moral philosophy3 according to which ‘moral requirements are not conditional at all: neither upon desires nor upon the absence of other reasons’ (McDowell 1978, p. 29). How far is his position dependent on Nagel’s argument? So far as I can see he does not have any other argument for the thesis that beliefs can motivate without the aid of (non-consequential) desires. (Though, of course this deficiency may have been remedied in subsequent writings.). He illustrates the thesis and answers some objections but gives no other reason for supposing it to be true. Does he need this thesis to support his particular brand of virtue ethics according to which moral requirements are not hypothetical imperatives? I think so, yes. The reason that moral requirements are not hypothetical imperatives – that they do not depend for their reason-giving force on the desires or propensities of the agent – is that one ‘one cannot share a virtuous person’s view of a situation in which it seems to him that virtue requires some action, but see no reason to act in that way’ (McDowell 1978, p. 26). But if beliefs cannot motivate without the aid of pre-existing biffy desires, then it is indeed possible to share the virtuous person’s view of a situation but see no reason to act as virtue requires, since you may not share the virtuous person’s desires. If this is right, then McDowell’s Barad-dûr of virtue is

Charles Pigden

227

heavily dependent on the power of Nagel’s argument and would begin to totter if deprived of its support. The case is clearer with respect to Jonathan Dancy. In Moral Reasons (1993) Dancy develops an extreme form of the Nagel–McDowell thesis according to which all desire ascriptions are ‘consequential’ in McDowell’s sense, and causally active desires play no part whatsoever in the generation of action (Dancy 1993, p. 9).4 So far as I can see, Dancy accepts both of Nagel’s arguments, the one he explicitly advances and the one that appears to be in the offing. That is, he thinks that because thesis 2) I have a desire to bring about Y, follows from thesis 1) I do X because I believe that doing X is likely to bring about Y, thesis 2) means no more than thesis 1). In which case there is no reason to posit an unmotivated desire for Y as an independent causal factor in the situation. But Dancy also thinks that because thesis 4) I want or desire to do X, follows from thesis 3) I am doing X intentionally, thesis 4) means no more than thesis 3). In which case there is no need to posit a motivated desire to do X as an independent causal factor in the situation either. Fortified by this argument, Dancy dismisses all desires en bloc contending that they are merely consequential. Thus Dancy’s theory is much more extreme (or, as he puts it, ‘pure’) than the theories of Nagel, Foot and McDowell who are all prepared to admit that at least sometimes genuinely biffy desires do a certain amount of work. Is Dancy dependent on Nagel? Very much so. He says himself that he has ‘said little in favour of the pure theory’ and that he will ‘in fact never offer an explicit argument in favour of that theory nor argue directly against its Humean rival’ (Dancy 1993, p. 20). Instead, he quotes and endorses Nagel’s paragraph (Dancy 1993, p. 8) and then goes on to contend that, given Nagel’s argument, his pure theory is better than Nagel’s hybrid theory. Thus without Nagel’s argument, he has got nothing except the intrinsic plausibility of the pure theory. Which is to say that he has got nothing. Finally, I turn to Cullity and Gaut. In the introduction to their anthology Ethics and Practical Reason, they endeavour to depict the state of play with respect to these topics as of 1997. They paraphrase and endorse Nagel’s argument, which they clearly regard as a major contribution to the debate. What is the upshot? That ‘Nagel shows that the neo-Humean argument fails to establish the conditionality of normative reasons upon the agent’s desires’ (Cullity and Gaut 1997, pp. 8–9). Read one way this is half-right, read another it is wholly wrong. For there is, I think, an ‘ordinary language’ argument for neo-Humeanism that Nagel’s argument does discredit though only on the

228 A Niggle at Nagel

(false) condition that desire is not a causal concept. But if they mean to suggest that Nagel shows that normative reasons are not conditional on the agents’ desires (since beliefs by themselves, or maybe even the facts believed can motivate agents in the absence causally active desires) then they are just mistaken. If this is the consensus among the practical reason crowd, then it is a consensus of error.

4

The Cracks of Doom I: the substantive thesis

It is now time to justify my big talk. Is Nagel’s argument really as bad as I have suggested? As we have seen, the key premise of Nagel’s argument is the following claim: I) Thesis 2) I have a desire to bring about Y, follows from thesis 1) I do X because I believe that doing X is likely to bring about Y. But in one sense I) is simply false. Thesis 2) does not follow logically from thesis 1): the one is not a logical consequence of the other. In logic (very roughly) you don’t get out what you haven’t put in. Yet 2) contains new matter – the concept of desire – and it seems inconceivable that this new relation or affirmation could be a logical deduction from another which is entirely different from it. (There is no mention of desire in the premise of this supposed inference.) Well, of course thesis 2) is not a logical consequence of 1)! What Nagel and his cohorts meant to say is that 2) is an analytic consequence of 1): that is that 2) can be logically derived from 1) with the aid of some uncontroversial analytic truth or truths. Before exploring this option I want to discuss another reason for thinking that 2) means no more than 1) which lurks at the back of people’s minds greasing the wheels of fallacy. Perhaps 2) means no more than 1) because they both have the same verification condition. What verifies 2), that I have a desire to bring about Y is precisely what verifies 1) that I do X because I believe that doing X is likely to bring about Y. Since they both have the same verification conditions the two claims are equivalent. And if 2) is equivalent to 1), then 2) means no more than 1) in which case the desire for Y is not a real thing but a conceptual chimera. To say that I desired Y is simply to say that I did X in the belief that it was likely to bring about Y. Let me stress that nobody states this argument explicitly, certainly not McDowell, who won his spurs refuting the verificationism of Dummett. Yet some expositions of Nagel’s argument have a gamey whiff of verificationism about them, and I can’t help thinking that residual verificationist intuitions give the argument more plausibility than it deserves. So perhaps it is worth explaining just why this verificationist version of Nagel’s argument is worthless.

Charles Pigden

229

To begin with verificationism is false. The meaning of a proposition is not its method of verification. Hence two propositions with the same verification conditions need not be equivalent. There are many reasons for thinking that verificationism is false, not least the fact that it is impossible to isolate little nuggets of potential experience to constitute the verification conditions of individual sentences. What counts as confirming an individual sentence depends upon what else is believed. As the Man said, ‘our statements about the external world [or, as he might have added, the internal world] face the tribunal of experience not individually, but only as a corporate body’ (Quine 2004, p. 49). Second, even if verificationism were true, theses 1) and 2) would not have the same verification conditions. This is partly because 2) is embedded in a big complex theory, belief/ desire psychology, which has to be verified or falsified in terms of its overall success in explaining human conduct. But even if we waive that point, it is clear that the verification conditions of 1) and 2) are not the same. Even if we admit that 2), that I have a desire to bring about Y, would be verified by whatever verifies 1) that I do X because I believe that doing X is likely to bring about Y, 2) could also be verified by many other phenomena, for instance the phenomena that verify 2’) that I do Z because I believe that doing Z is likely to bring about Y or 2’’) that I do W because I believe that doing W is likely to bring about Y and so on. Thus even if verificationism were true, 2) would not be equivalent to 1) which means that talk of unmotivated desires could not be regarded as a terminological variant of talk of intentional action. The verificationist version of Nagel’s argument, in so far as it exists, is utterly hopeless. So let us get back to the real argument. As I understand it, it goes something like this: Thesis 1) I do X because I believe that doing X is likely to bring about Y, plus thesis A) It is conceptually necessary that if I do X because I believe that it is likely to bring about Y, then I have a desire to bring about Y, entail thesis 2) I have a desire to bring about Y. THEREFORE N) Thesis 2) means no more than thesis 1) and there is no need to posit a desire for Y as a causally active ingredient in the situation. I do not accuse Nagel of reasoning like this: Thesis 1) I do X because I believe that doing X is likely to bring about Y, plus thesis

230 A Niggle at Nagel

A) It is conceptually necessary that if I do X because I believe that it is likely to bring about Y, then I have a desire to bring about Y, entail thesis 2*) It is conceptually necessary that I have a desire to bring about Y, THEREFORE N^) Because 2*) is conceptually necessary, that is analytic, it says nothing substantial about my mind, and there is no need to posit a desire for Y as a causally active ingredient in the situation. This argument derives the necessity of the consequent from the necessity of the consequence: that is from premises of the form P and Necessarily (If P then Q) it derives the conclusion Necessarily Q – a well-known and fallacious form of inference. This error is compounded by making the assumption that, if Q is conceptually necessary, it says nothing substantial about the world, an assumption that seems to me distinctly dubious. But though this argument too may be at the back of people’s minds facilitating the fallacy, I do not think it is Nagel’s official position. He is not claiming that thesis 2) says nothing about the world because it is conceptually necessary or analytic: he is claiming that it says no more about the world than thesis 1). It is for that reason that we don’t have to posit a desire as a causally active ingredient in the situation. Thus it is the first version of the argument that corresponds to his intentions. Nagel thinks he can concede that it is a conceptual truth that if I do X because I believe that it is likely to bring about Y, then I am acting out of a desire for Y. But (so he argues) it is precisely because it follows that I am acting out of a desire for Y, that the desire for Y cannot be an independent factor in the situation. This shows that he does not really understand the nature of conceptual truths. Consider, A) It is conceptually necessary that if I do X because I believe that X is likely to bring about Y, then I have a desire to bring about Y. This amounts to the following: A’) Our concepts of belief and desire are such that if I do X out of a belief that X is likely to bring about Y, it does not really count as a belief unless I do X because of a desire for Y. Now suppose that I do X. And suppose too that I do X in the belief that doing X is likely to bring about Y. In what sense does it follow that I desire Y? It follows in this sense. On the assumption that I am doing X out of a genuine or desire-entailing belief that doing X will bring about Y, that is a belief such that if I act on it then I will have the corresponding desire, then I will have the corresponding desire. But this certainly does not mean that the desire is not, or need not be, an independent causal factor in the situation.

Charles Pigden

231

Consider the parallel case. The following is a conceptual claim: B) It is conceptually necessary that if John is a son he has (or has had) parents. This amounts to the following: B’) Our concepts of ‘son’ and ‘parents’ are such that John does not really count as a son unless he has (or has had) parents. In what sense does it follow from the fact that John is a son that he has parents? It follows in this sense. On the assumption that John is a son, that is the kind of being such that, if he exists, then he has or has had parents, then he has or has had parents. But is certainly does not follow from this that John’s parents were causally inactive in the production of John. On the contrary, their causal intervention was essential. If an item a is of a certain kind W, and if it is a conceptual truth that things of kind W have characteristics K, then a will have characteristics K. But it does not follow from this that the characteristics K are unreal, formal or causally inactive. In truly calling a a W, we have already and ex hypothesi TRULY called it a thing with characteristics K. If we have truly called John a son we have already and ex hypothesi TRULY called him a person who has (or has had) parents. And if the conceptual truth that Nagel concedes is indeed a conceptual truth, then in claiming truly that I did X out of the belief that X would lead to Y in this sense, we have truly claimed that I did it out of a desire for Y. But neither in this case nor in the others does this mean that the desire was causally inactive. On the contrary, if our concept of a desire is a causal concept – and nothing so far has suggested that it is not – then we have claimed, and claimed truly, that a desire for Y was causally responsible for my doing X. Indeed, given the conceptual truth the Nagel concedes, if there is not a genuine desire for Y, then I did not really act out of the belief that X is likely to bring about Y! It is true in a sense that if 1) I do X because I believe that doing X is likely to bring about Y, analytically entails 2) I have a desire to bring about Y , then 2) means no more than 1). But this is not because there is less to desiring than meets the eye – which is what Nagel thinks – but because there is more to believing (or rather to acting on a belief). If A) is really true – and it must be true for 2) to follow analytically from 1) – then I cannot really act on the belief that doing X will bring about Y, unless I desire Y. And there is nothing in the argument so far to show that desiring Y is not a very biffy thing indeed. Is there anything we can do to restore the argument to validity? Yes. We can alter A to

232 A Niggle at Nagel

A’’) It is conceptually necessary that if I do X because I believe that X is likely to bring about Y, then I have a merely consequential and noncausal desire to bring about Y. This gives us the following argument: Thesis 1) I do X because I believe that doing X is likely to bring about Y, plus thesis A’’) It is conceptually necessary that if I do X because I believe that it is likely to bring about Y, then I have a merely consequential and noncausal desire to bring about Y, entail thesis 2’’) I have a merely consequential and non-causal desire to bring about Y. THEREFORE N’’) thesis 2) means no more than thesis 1) and there is no need to posit a desire for Y as a causally active ingredient in the situation. But this argument suffers from two defects. A’’), so it seems to me, is pretty clearly false, and even it were true, the argument would be redundant. If it really is a conceptual truth that unmotivated desires are consequential and non-causal – which is what, in effect, A’’) claims – then we don’t need the elaborate rigmarole Nagel’s argument to prove the point. Nagel’s argument then fails to prove the substantive conclusion: N*) A person can be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y. Thus in so far as the Barad-dûrs of error depend upon the substantive thesis N*) they are in big trouble. For Nagel has given us no good reason to believe it.

5 Objection5 But maybe I have misrepresented the argument. The point was not to prove that the substantive thesis N*). The point was rather to refute an argument which purports to disprove N*), on the basis of supposed analytic truths. Thus what Nagel was really arguing for was this: N#) It is not possible to disprove N*) by appealing to analytic truths. Or perhaps this: N##) It is not an analytically or conceptually impossible for a person to be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y.

Charles Pigden

233

But N##) entails N###) It is conceptually possible (not excluded by our common concepts of belief and desire) for a person to be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y. Now N###) is much weaker thesis than N*). For there are many things that are conceptually possible that are otherwise impossible (physically, psychologically or historically impossible, for instance). There is nothing about the concept of a pig or the concept of flight that excludes the possibility of flying pigs. Nevertheless, pigs can’t fly. ‘If I could turn back time’, sings Cher, ‘I’d take back those words that hurt you, and you’d stay’, implying thereby that there is nothing incoherent, nothing conceptually impossible, about turning back time. Nevertheless, as the lyric makes abundantly clear she just can’t ‘find a way’ to do it, presumably because it is physically, if not metaphysically, impossible. But though N###) is much weaker thesis than N*), it may be that the two have been confounded in the minds of Nagel’s admirers. And it may be that N#) was what Nagel was arguing for all along, in which case N##) and N###) follow as corollaries. After all, The Possibility of Altruism is based on Nagel’s B.Phil thesis, which was drafted at Oxford in the heyday of ordinary language philosophy, when conceptual analysis was all the rage. (My late supervisor, H.J. McCloskey, was able to publish his paper ‘The Nature of Moral Obligation’ by changing its title to ‘The Concept of Moral Obligation’, a change he rather objected to since in his view moral obligation had a nature independent of our conceptions.) Now in the Oxford of the 1950s a Humean philosopher might be inclined to dispute N*) on the basis of conceptual considerations. Hence the first order of business for a defender of N*) would be to defuse the conceptual arguments against it. And that is all that Nagel is trying to do in this famous passage.

6

Reply

Maybe so, but that still leaves N*) unproven. And it seems to me that N*) is what Nagel and his followers really need. As we have seen, Nagel’s title is a bit misleading. He does not merely want to prove that altruism is possible – obviously it is, since some of us have altruistic desires. What he wants to prove is that altruism is rationally required, and that selfish people are making some sort of a mistake. To do this he must prove that other people’s suffering constitutes a reason for action for all rational human beings whatever their psychological quirks, a reason that they cannot neglect without intellectual error of some kind. Now suppose that N*) is false. Suppose, that is, that it is psychologically impossible for someone to be motivated by the belief that X leads to Y unless there is a pre-existing desire for Y, where a desire is construed as a causally active ingredient in the situation. Suppose

234 A Niggle at Nagel

too that a fact only constitutes a reason for action for a person P if it is irrational for that person to believe the fact whilst remaining unmoved (a thesis to which Nagel would presumably agree). And suppose it is a fact – and a fact that I believe – that I can alleviate your suffering by lending you a helping hand. Unfortunately I have no desire whatsoever to alleviate your suffering, perhaps because I don’t like you, or perhaps because I am simply callous. Let us add that, deplorable at it may be, there is nothing particularly irrational about having a callous disposition. Whatever the cause in nature that makes these hard hearts it does not seem to be a defect in our rational capacities. Then (absent a change of heart) it will be psychologically impossible for me to be moved by the belief that I can alleviate your suffering by lending you a helping hand since I lack the desire to alleviate your suffering. Thus, (for the moment, at any rate) the fact that lending you a helping hand would alleviate your suffering does not constitute a reason for action for me, since in remaining unmoved I am not being irrational but merely hardhearted. (How can I be rationally required to do what it is psychologically impossible for me to do, given that my psychology is not irrational?) Thus if N*) is false, Nagel cannot show that altruism is a rational requirement binding on all human beings whatever their dispositions and desires. He cannot show that selfish people are making some kind of mistake. For if N*) is false, the facts which constitute reasons for action for the altruistic, do not qualify as reasons for action for the selfish. But N*) can be false and N###) true. Hence establishing N###) is not enough. However, I am inclined to think that Nagel’s style of argument cannot establish theses N#) – N###) without the aid of a question-begging assumption which is probably false, even though theses N#) – N###) are probably true. And an argument that cannot prove what is probably true is obviously a little lacking. Why do I think that theses N#) – N###) are probably true? Because I reject the principle A), that Nagel implicitly endorses, that it is conceptually necessary that if I do X because I believe that it is likely to bring about Y, then I have a desire to bring about Y. Desires are complicated things, but I am inclined to analyse them along the following lines. A desire for the state of affairs Y (and all desires are, in my view, really desires for states of affairs), is the ground of a complex disposition, the disposition (ceteris paribus) to perform the action X or desire the state of affairs X given the belief that X leads to Y, to perform the action Z or desire the state of affairs Z given the belief that Z leads to Y and so on and to desire that not-R given the belief that the state of affairs R would prevent or inhibit Y, to desire that not-Q given the belief that the state of affairs Q would prevent or inhibit Y and so on. Even in the degenerate case where I desire to do something (i.e. the state of my doing something) that is well within my powers, such as whistling while I work, the desire isn’t simply the disposition to whistle while I work. For if I really want to whistle, then I am disposed (ceteris paribus) to remove any

Charles Pigden

235

obstacles that would prevent me whistling, such as chewing gum or a gag. But if this is roughly right, then it is conceptually possible for me to do X in consequence of the belief that X leads to Y without the aid of such a complex dispositional state. I could do X in the belief that X leads to Y without being disposed (ceteris paribus) to do W if I came to believe that W led to Y. This would be psychologically weird, and perhaps deeply irrational, but conceptually, at least, it seems perfectly feasible. But if A) is false, then a conceptual argument to the negation of N*) is not going to get off the ground. And if there is no conceptual argument to the effect that N*) is false, then our concepts do not exclude the possibility that a person might be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y. In other words N##) and N###) are both true.

7 The Cracks of Doom II: the conceptual/logical thesis But let us suspend our disbelief in A) and go back in time to an era when giant herds of conceptual analysts roamed the earth (or at any rate Oxford). How would a Humean conceptual analyst argue against N*)? Perhaps like this: A’’’) It is conceptually necessary that if a person does X because they believe that X is likely to bring about Y, then they have a desire to bring about Y. THEREFORE ~N*) It is not conceptually possible for a person be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y. (Of course in the 1950s the argument would not have been formulated in such a carefully non-sexist way.) ‘No, no!’ says Nagel. ‘I grant the truth of A). If someone does X because of a belief that X leads to Y then it follows analytically that the person has a desire for Y. But it does not follow that that desire is biffy. It does not follow that the desire is a condition contributing to the motivational efficacy of the consideration that X leads to Y. It is a necessary condition of its efficacy to be sure, but only a logically necessary condition. It is not necessary either as a contributing influence or as a causal condition.’ Is this a good argument? Consider the following parallel. Suppose that we had a fifties philosopher who was arguing, rather redundantly, that there cannot be sons without the causal intervention of parents. But like a true fifties philosopher, he stakes his case on a conceptual claim. B’’’) It is conceptually necessary that if a person X is a son, then he has or as had parents.

236

A Niggle at Nagel

THEREFORE ~S*) It is not conceptually possible for a person to be a son without two parents helping to cause his existence via sex (though we don’t exclude such offbeat variations of sex as in vitro fertilization). Next, imagine a parent-Nagel, someone who takes the same line with respect to parents as Nagel does towards desire. ‘No, no!’ says parent-Nagel. ‘I grant the truth of B’’’). If someone is a son then it follows analytically that the person has, or has had, parents. But it does not follow that the parents were causally biffy. It does not follow that the parents were a causally efficacious condition of the person’s being a son. To be sure the prior existence of parents is a necessary condition of the person’s being a son, but only a logically necessary condition. Their existence or activity is not necessary either as a contributing influence or as a causal condition.’ Why is parent-Nagel’s response ridiculous? Because parenthood is a causal concept. To be a parent is to play a certain kind of causal role with respect to ones sons and daughters. A parent is a kind of cause just as a son is a thing caused in a certain kind of way. Thus if it is conceptually impossible for someone to be a son without having had parents, it is conceptually impossible to be son without the parents having played the relevant causal role. Now Nagel’s response to the Humean conceptual analyst is not as ridiculous as that. For it is not so obvious that ‘desire’ is a causal concept. But if ‘desire’ is a causal concept then Nagel’s response is equally unsuccessful. Remember that we are suspending disbelief in premise A’’’). We are granting (for the moment) that it is conceptually necessary that if a person does X because they believe that X is likely to bring about Y then they have a desire to bring about Y. But if this is the case, and if having a desire to bring about Y consists, in part, in having a causal disposition to do X if you believe that doing X leads to Y, then, in endorsing A’’’), we have endorsed the following thesis: ~N**) It is conceptually necessary that if a person does X because they believe that X is likely to bring about Y then they have a causal disposition (inter alia) to do X if they believe that doing X leads to Y, that is, a causally active desire for Y. What is the upshot? That Nagel’s style of argument (which crucially grants the various variants of principle A)) can only fault the conceptual argument that a person cannot be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y, on condition that ‘want’ and desire’ are not causal concepts. But this is a large part of what he wanted to prove. Thus Nagel cannot even prove N#) – that it is not possible to disprove N*) by appealing to analytic truths – without circularity. And

Charles Pigden

237

the failure here is all the more abject because the thesis that he cannot prove without begging the question, is, in fact, trivially true. Things are even worse if ‘desire’ is in fact a causal concept. For in that case Nagel cannot even establish the truth that N*) cannot be disproved on conceptual grounds without the aid of a question-begging assumption that is in fact false.

8

Conclusion

What Nagel’s argument requires and what he seems to want to prove is the following thesis: N*) A person can be motivated to perform X by the belief that X is likely to bring about Y, without a causally active or biffy desire for Y. But despite its enormous influence, his argument for this claim is a complete failure. A fallback position is that Nagel might have been arguing for N#) It is not possible to disprove N*) by appealing to analytic truths. But although this claim is true, Nagel’s argument only succeeds given the question-begging assumption that ‘desire’ is not a causal concept, an assumption that is probably false.

9

Coda: a spot of experimental philosophy

But is it false? When I claimed rather dogmatically that ‘desire’ is a causal concept Severin Schroeder demanded, with a touch of asperity, whether I had done any surveys of common usage. Well, I had not then, but I have now. Here is a questionnaire administered to innocent first-year students to test their conceptual intuitions about the nature of desire. Remember the point is not to determine what wants or desires are but to determine what they are commonly conceived to be, and specifically whether they are commonly conceived as causes.

Conceptual intuitions survey This questionnaire is designed to find out about your conceptual intuitions, that is, what you are naturally inclined to say about certain imaginary situations, because of your understanding of certain key concepts that we all employ in everyday life. (We are not going to say which concepts since that would contaminate your response!) There is no right answer here – we want to know what you think. Furthermore, we want to know what you think without too much thinking. You may feel that what you are naturally inclined to say is blindingly obvious and that we cannot be wanting you to say such a blindingly obvious thing, thus leading you to think up a more

238 A Niggle at Nagel sophisticated answer. Not so! If you think that what you are at first inclined to say is blindingly obvious, then please go ahead and say the blindingly obvious thing. Don’t over-think it! It’s the ‘common-sense’ obvious answer (or what you think is the ‘common sense’ obvious answer) that we want. Example 1 Alice flips the switch, believing that by flipping the switch she will turn on the light. Question: Did a desire or want for turned-on light cause Alice to flip the switch? Please circle ‘Yes’, ‘No’ or ‘Not sure’. Yes No Not sure ---------------------------------------------------------------------------------------------------------Example 2 Stalin tells his henchman Khruschev to dance the gopak, a very vigorous peasant dance which the rather stout Khruschev would otherwise much prefer not to perform. Khruschev dances the gopak, believing that if he does not, Stalin will have him arrested and perhaps shot. Question: Did a desire or want not to be arrested and shot cause Khruschev to dance the gopak? Please circle ‘Yes’, ‘No’ or ‘Not sure’. Yes No Not sure ---------------------------------------------------------------------------------------------------------Example 3 Simon and Soraya are both traffic cops. Each of them separately stops a speeding car which is driven by an unscrupulous billionaire. The billionaires each offer Simon and Soraya a large bribe to let them off without a ticket. Both Simon and Soraya think it is wrong for police officers to take bribes. But each of them is absolutely sure that he or she could get away with it if they took the bribe. And neither of them believes in an all-seeing God. Simon takes the bribe but Soraya does not. Question: Is this correct – Simon’s act of taking the bribe was caused by the fact that his desire for the money was stronger than his desire to do his duty whilst Soraya’s refusing to take the bribe was caused by the fact that her desire to do her duty was stronger than her desire to for the money? Please circle ‘Yes’, ‘No’ or ‘Not sure’. Yes No Not sure

-----------------------------------------------------------------------------------------------

Charles Pigden

239

The survey was administered to 65 students at Oxford Brookes University in England (not Oxford, where the conceptual intuitions of incoming undergraduates would be more likely to be contaminated by contentious theory) and 74 students at Otago University in New Zealand. None of the students were mine nor had I spoken to any of them. In both cases the majority (though not all) would have been native English speakers. At Oxford Brookes 46 per cent answered ‘Yes’ to all three questions, 38.5 per cent answered ‘Yes’ to two questions and ‘No’ or ‘Not Sure’ to one, 6.7 per cent answered ‘Yes’ to one question and ‘No’ or ‘Not Sure’ to two and 6.7 per cent answered ‘No’ or ‘Not Sure’ to them all. At Otago 56.7 per cent answered ‘Yes’ to all three questions, 35.1 per cent answered ‘Yes’ to two questions and ‘No’ or ‘Not Sure’ to one, 6.7 per cent answered ‘Yes’ to one question and ‘No’ or ‘Not Sure’ to two and 1.3 per cent (one student) answered ‘No’ or ‘Not Sure’ to them all. Thus in both cases the majority of students are inclined to think of desires as causes at least most of the time, though the New Zealanders are rather more causally inclined than the English. For each question the percentages were as follows: Example

Yes

No

Not sure

1

89.47%

6.57%

3.94%

2

85.52%

11.84%

2.63%

3

71.05%

22.36%

6.52%

1

71.64%

16.41%

11.94%

2

85.07%

8.95%

5.97%

3

65.67%

17.91%

13.43%

Otago

Oxford Brookes

Notes 1. This is still a bit problematic as the starting points of one deliberative process may be the end points of another and vice versa. I may want a good time tonight because I think it the best way to escape the feelings of futility and depression that have been dogging me for the last few days as a consequence of reading Nagelinspired ethical theories. And once I have decided to go to the movies I need to deliberate about which movie to see. 2. There is also the problem that, despite her big talk about biology, there is nothing particularly natural about Foot’s conception of goodness – but let’s not go there. 3. I was present as an undergraduate in 1978 when McDowell read the paper at a meeting at the Moral Sciences Club in Cambridge. Though he kept insisting that his argument was ad hominem, the aged R.B Braithwaite, who was also present,

240

A Niggle at Nagel

suggested that this ploy was a little evasive, and that since McDowell obviously had something new to say, he should come right out and say it. 4. Dancy’s attitude in his subsequent book, Practical Reality is rather more equivocal, and I can’t work out whether he wants to endorse Nagel’s argument or not. He seems to be saying that if I do X intentionally then I do have a desire to do X, but that this desire simply consists in my being motivated to do X. But that desire (or that state) can be wholly explained by my beliefs. If I form the desire to do X on the basis of my belief that doing X will bring about Y, no substantial desire to bring about Y is required. Why not? Presumably because such a desire is merely ‘consequential’ on my being motivated to do X by my belief that it is likely to bring about Y. Which suggests that he endorses Nagel’s argument after all. See Dancy, 2000, pp. 79–84. 5. This objection is due to the highly critical philosophers of Reading where I gave an earlier version of this essay in 2006. Papers are unlikely to improve if they are only read to people who can be expected to agree with them and thanks are due to the philosophers of Reading, who fully lived up to my expectations of disagreement. They have done this essay a power of good. I would like particularly to thank Phillip Stratton-Lake and Severin Schroeder.

References G. Cullity and B. Gaut eds (2000) Ethics and Practical Reason (Oxford: Oxford University Press).

J. Dancy (1993) Moral Reasons (Oxford: Blackwell). J. Dancy (2000) Practical Reality (Oxford: Oxford University Press). P. Foot (1972) ‘Reasons for Action and Desires’, Aristotelian Society, Supplementary Volume, 46, 203–210. Reprinted in Foot (1978), 148–156, all references to this reprint. P. Foot (2001) Natural Goodness (Oxford: Oxford University Press). P. Foot (1978) Virtues and Vices and Other Essays in Moral Philosophy (Oxford: Blackwell). P. Foot (2002) Moral Dilemmas (Oxford: Oxford University Press). J. McDowell (1978) ‘Are Moral Requirements Hypothetical Imperatives?’ Proceedings of the Aristotelian Society, Supplementary Volume, 52, 13–29. T. Nagel (1970) The Possibility of Altruism (Princeton: Princeton University Press). W.V.O. Quine (2004) Quintessence, ed. R.Gibson (Cambridge, MA: Harvard University Press).

13 Acting in Character Annette Baier

No union can be more constant and certain, than that of some actions with some motives and characters. (David Hume, T 2, 3, 1, 12)1 Hume is often taken to have claimed, in his Treatise of Human Nature, that human action, like all animal action, is caused by the agent’s strongest desire, directed by her beliefs about how best to satisfy it. “Reason is, and ought only to be, the slave of the passions, and has no other office than to serve and obey them” (T 2, 3, 3, 4). Of course our passions include the love of truth, concern for good reputation, and all the passions we come to have by sympathy with others, and include also our moral sentiments, so any picture of the Humean agent as a ruthless desire-satisfier without any intellectual or humanitarian concerns would be a crude caricature. Hume’s moral psychology makes room for a great variety of kinds of human agent, with different ruling passions, different habits of belief acquisition and different degrees of that “strength of mind” which enables a person to gain control over what in the Treatise he termed her “violent” and unruly passions, in favor of her calmer less disruptive ones. Violent passions, for Hume, are the ones unfit for even temporary rule, despite the fact that the office of passions is to rule. They are the dysfunctional ones in a person’s repertoire of passions, the ones least likely to make good use of their servant, that information-gathering and long-sighted reason which would, if listened to, point out the consequences of the rash satisfaction of violent passions, such as anger at insult, or a craving for appetizing-looking, forbidden and indeed poisoned fruit, offered by sinuous tempters. How a person manages her different and often conflicting passions, all keen to rule, and how she uses her intelligence to serve her prevailing passions, as well as which passions those are, will be a matter of character.2 Hume regards character as “of vast weight and importance.” He uses these words for “our reputation, our character, our name” (T 2, 1, 11, 1) and the three, for him, are closely linked. Character is not some hidden inner constitution of a person, it is the outward expressive 241

242 Acting in Character

face of that inner nature, which helps determine the role a person plays among other persons, and the reputation thereby acquired. The original sense of the word referred to a stamp externally applied. Samuel Johnston’s, 1755 Dictionary of the English Language, gives that as his first sense, moves on to “an account of anything, as good or bad,” and then to “a person, with his assemblage of qualities.” Hume speaks of a person having “a character for honesty” or whatever, where this is very close to reputation. He wrote an essay, “Of National Character” which is as much about reputation as it is about true nature, and another, entitled, “A Character of Sir Robert Walpole,” which was what we would call a character-sketch of the prime minister, an account of him, as good and bad. It was published in 1741, shortly before Walpole was forced from office, and Hume reduced it to a footnote to the essay, “That politics may be reduced to a science” in 1748, after Walpole’s death, and withdrew it altogether in 1770, after expressing uncertainty about its fairness. (It had by no means been all bad, but had criticized Walpole for increasing the national debt.) “A character” may be unfair, as reputation may be ill deserved, and a bad name unfairly or mistakenly given, though Hume never retracted his judgment that Walpole as a minister had lacked “frugality.” What we today call “character” will be what it is, regardless of how it may be represented or misrepresented. For Hume, character is a person’s nature, as in principle perceivable by fellow persons. A prime minister’s economic policy is not inscrutable, and Hume regards our character as, in most respects, public and private, quite scrutable to those who know us. Normally, thanks partly to our mirror neurons,3 we perceive each other’s emotions and immediate intentions fairly correctly, and fairly immediately, and it is these emotions and intentions, as much as our actual actions, which exhibit our private characters. Character is what we are like, as shown in how we reveal ourselves to others. Character is largely a matter of what our passions and intentions are, and what habits of belief-acquisition and belief-correction we have, but whereas beliefs may change fairly rapidly, passions are “slow and restive” and character is taken by Hume to be even slower to change. He wrote, arguably mistakenly,4 that we hold a person responsible for what she did only if the action flowed from something “durable and constant” in the agent, some “cause in the characters and disposition of the person,” (T 2, 3, 2, 6. Note the plural “characters” equivalent to our “character traits.”) Here he is looking mainly at criminal responsibility, not at honors lists or medals for bravery, and thinks magistrates are less likely to punish a person for an uncharacteristic act than for one that is in character. It is interesting that this durable and constant character of a person was quite blocked from Hume’s view when he searches, in the Treatise, in “Of Personal Identity” and again in the Appendix,5 for some reason to believe that we remain the same persons throughout life. But once he is considering us in our social interactions, in Books 2 and 3, character comes fully

Annette Baier

243

into his view. Hume ends almost every paper of his History of England with a paragraph labeled, “death and character,” in which he sums up his judgment of how the deceased monarch is to be viewed. Sometimes he had cited this relatively long-lasting aspect of a person, her character, to explain what she did, as well as to justify holding her responsible for that, and to serve as the proper object of moral evaluation. Here his use of “character” seems almost the same as ours. In this essay I shall be largely concerned with character’s explanatory role, for Hume and for us, so looking at “character” in our sense. When it comes to persons, coming to understand sometimes cannot be sharply separated from coming to judge. Certainly Hume tends to elide the two operations, in his History. He wrote, in the early essay, “Of the Study of History” (later withdrawn, because of its condescending tone to women, whom he had urged to read history rather than novels and romances), of the appeal of history as lying not just in “the rise, progress, declension, and finally extinction of the most flourishing empires,” but also “to see the whole human race, from the beginning of time, pass, as it were, in review before us; appearing in their true colors, without any of those disguises, which during their lifetimes, so perplexed the judgment of beholders.”6 The historian is at a temporal distance, Hume says, so can in principle be disinterested in his judgment of those about whom he writes. His own History of England ended in 1688, and he refused suggestions that he continue it up to his own time. His typical criticism of other historians, such as Rapin and Clarendon, was that they lacked this impartiality which the historian should have, that they were Whigs, or Jacobites, seeing the past as leading up to present political struggles. Someone like Clarendon who had been closely involved in the events whose history he wrote about (“The Rebellion”) can provide important material for the historian, but can scarcely let the people he wrote about, such as Oliver Cromwell, pass in review before an unprejudiced eye. This early essay makes it very clear that Hume sees the historian’s job to include a discovery, in his sense of an uncovering, of true character. David Velleman says that we explain human actions sometimes by pointing out the motivation for the action, sometimes by showing it as done from habit, sometimes by reference to the agent’s character, sometimes by citing the emotion expressed by what the person was doing. He sees no reason to reduce the latter three sorts of explanation to the first, so rejects the Davidsonian theory that a desire (or “pro-attitude”) plus a belief about how to satisfy it are always the main constituents of any satisfactory explanation of what is intentionally done.7 Davidson is often taken to be agreeing with Hume in this account of what causes action, so such desire plus belief explanations have been called “Humean.” I shall be looking at explanations that cite character trait, and looking especially at some of Hume’s explanations of why the people whose actions he relates in his History of England did what they did.

244 Acting in Character

When he is relating the various mis-moves made by Charles 1, leading up to the civil war, Hume seems to appeal to weaknesses in the king’s character to explain what he did. In the character sketch he gives when telling of his early reliance on Buckingham, he diagnoses one such weakness: Charles’ bad judgment in choice of advisors, leading to his adopting dangerous and inconsistent means to his ends. He was very steady, and even obstinate in his purpose; and he was easily governed, by means of his facility, and of his deference to men inferior to himself both in morals and in understanding. His great ends he inflexibly maintained; But the means of attaining them, he readily received from his ministers and favorites, though not always fortunate in his choice. (H 5, LVI, p. 175)8 His temporizing pacification of the Scots in 1639 was due to his not being sufficiently “vigorous or decisive,” to his being apt, from facility, “to embrace hasty counsels.”(H 5, LIII p. 268) (Here we see how entangled explanation and evaluation are, for Hume.) Charles had gone to war against the Scots to force them to accept Laud’s liturgy, and episcopal authority, but made peace without securing those goals. “To advance so far, without bringing the rebels to a total submission, or at least reasonable concessions, was to promise them, in all future time, an impunity for rebellion.”(H 5, LIII p. 266) Charles did this because he feared the support that the Scots might get from English sympathizers, if he continued to fight them. The rebellion might then spread to England. This was, as future events proved, a not-unreasonable worry. And he had “a natural propension towards the misguided subjects of his native kingdom,” (H 5, LIII, p. 268) so disliked being at war with them. A more sagacious leader would not have temporized as Charles did on this occasion. But Charles was not sagacious. He acted stupidly, acceded in measures, such as that abolishing episcopacy in Scotland, that he soon found he could not really accept, so had to resume fighting, then call a parliament in England to try to get supplies to pay his armies. They refused, partly for the reasons Charles had foreseen, sympathy with the anti-Laud rebels. He had vacillated in Scotland, with fatal effects. Then in 1642, faced with another obstreperous Puritan-led English parliament who passed the “Remonstrance,” then impeached the protesting bishops for treason, while crowds in London shouted against “bishops and rotten-hearted lords,” he retaliated by impeaching Lord Kimbolton, and five members of the House of Commons, leaders of the remonstrating and anti-episcopal faction. This, Hume writes, was, on Charles’ part, “an indiscretion, to which all the ensuing disorders and civil wars ought immediately and directly be ascribed” (p. 364). He had been urged to take a firm stand against the “insolence” of the commons by his queen, and several

Annette Baier

245

courtiers. The articles of impeachment accused Kimbolton and the five members of traitorously endeavoring to subvert the fundamental laws and government of the kingdom, to deprive the king of his regal power, and to impose on his subjects an arbitrary and tyrannical authority, even to raise tumults against the king, and levy war against him. (H 5, LV, p. 365) This charge, itself echoing the one made by parliament against Strafford, was, after nine years real warfare, to rebound on Charles, providing almost the model, with significant omissions, for the charges brought against him in 1649.9 Hume says, of Charles’ move to impeach the parliamentary leaders, “the whole world stood amazed at this important accusation, so suddenly entered upon, without concert, deliberation, or reflection” (H 5, LV, p. 365). The king came with 400 armed courtiers to parliament to arrest the accused, but they were not there, and soon it was the king who had to flee. The royal impeachment lapsed, civil war broke out, and the next impeachment was of the king himself. It was the strange case of Rex v Rex, of the office of king-in-parliament, against Charles Stuart, who had filled that office, but taken up arms against his parliament. That Hume can distinguish between a good reason and a poor but operative one is evident from this account of Charles 1’s temporizing pacification of the Scots in 1639. “What were the reasons, which engaged the king to admit such strange articles of peace, it is in vain to enquire; For there scarcely could be any. The causes of that event may admit of a more easy explication” (H 5, LIII p. 267). Charles’ reasons, awareness of the sympathy felt by some of the English for the “Scottish malcontents,” were so poor, Hume judges, as not to deserve the label “reasons.” Obviously he has adopted the “vulgar” honorific use of “reason” as something that should be listened to, and something that can oppose mere passion-causes. “Reason” here seems to be the calm desire for the greatest good, “reasons” to be reliable indicators of where it lies.10 Oliver Cromwell, who unlike Charles was “sagacious,” is allowed by Hume to have had reasons, not just psychological causes, of his decisions, even when some critics may query his judgment. Hume certainly queried his judgment that Charles, seen as a tyrant, must be removed without possibility of return. And Cromwell’s reasons for refusing to take Charles’ place on the throne might also be queried, but he did have reasons, not just causes, for his decision. That, as king, he would lack the support he needed, and that he would make himself and his old army followers look like hypocrites, were good reasons to refuse the crown. That the succession would be peacefully fixed if he accepted it was one reason for acceptance, and so one clause in the new version of his rights as protector gave him the right to nominate his successor. Cromwell’s six-week-long agony and perplexity, before refusing the crown, was not just a psychological conflict of causal forces in his soul, it was a difficult judgment about which reasons were the best reasons. His ambitious character explains much that he did, and his ruthless character

246

Acting in Character

certainly has to be invoked, in addition, to explain his determination to have Charles tried and executed. Hume invokes Cromwell’s ruthlessness to explain his regicidal actions, and his sagacity to explain his renouncing the chance to be king, and he invokes the weaknesses of Charles’ character to explain his fatally imprudent actions. Charles’ aims, of maintaining his own regal power and privilege, and having uniform Anglican worship in England and Scotland, were not enough to explain the measures he took. They could have led him to keep fighting the Scots in 1638, and be less belligerent with the English Remonstrants a few years later. Nor will it do just to add the beliefs accompanying his aims, in these two years – the belief that, had he not made peace in 1639, the English might have come to the aid of the rebel Scots; the belief that he could bully these same English, in 1642. The beliefs themselves need explanation, and that takes us back to Charles’ reliance on bad advisors. I here assume that a satisfactory explanation will relate the relatively puzzling thing to be explained to something(s) less puzzling. To explain a puzzling action by an even more puzzling belief will be to fail as an explainer, or at least to leave oneself with outstanding debts.11 So although Charles’ aims and beliefs of course figure in the full explanation (by which I mean an explanation which works for the most ignorant) of his pacification of the Scots, and of his impeachment and attempted arrest of the English parliamentary leaders, the key factor, Hume thinks, was Charles’ weak character, his bad judgment, his reliance on bad advisors, his vacillation between “facility” and arrogance. Does this mean that every explanation in terms of beliefs and desires is incomplete, until the source of those beliefs and desires has been given? No. Some beliefs are self-explanatory (their explanation can be taken for granted), some desires natural and normal. It is when either the motivation or the beliefs of the agent themselves cry out for explanation that citing them does not satisfy our wish to understand, nor end our puzzlement. Then a description of the agent can help. Just as with animal behavior we sometimes explain just by saying, “he’s a cat, so of course he pounces on a mouse, even if he has no need for it as food,” so we can say of Charles, “he’s a king whose power is threatened, so of course he attacks those who threaten it.” And even with cats, character can come in. “He’s a fearless old cat, who forgets he is toothless, so still attacks dogs.” And so with kings, when they break off the fighting too soon, or start a fight they cannot expect to win, we try to explain what they did to advance their not-at-all puzzling ends by looking to see who advised them, and why they heeded that advice. The kind of person one is helps determine both the sort of desires one has and acts on, and the way one selects advisors and forms one’s beliefs.12 As Velleman points out, it is not just character but sometimes it is self-perceived character that explains one’s actions. Cromwell saw himself as called to be chief sheriff of his country, and so declined to

Annette Baier

247

wear the crown. Even if one misperceives one’s own character, the misperception can explain what one does. Cromwell might have made a good king, and felt fulfilled in that role. He certainly would have been less likely to posthumously lose his head to Stuart avengers. (The English respect the corpses of their monarchs, even their beheaded monarchs. Charles I was buried in the royal chapel at Windsor, on some stories with his head reattached, so that his family could view his remains.) Of course, just as desires and beliefs can be puzzling so can character and self-image. There likely was in Charles 1’s past an explanation for both his firm belief in his divine right to rule England, and for his reliance on unwise advisers on how best to protect that right against challengers. From age 12 his father had instilled in him his divine right to rule, once it became clear that it was not God’s will that his older brother live to take the throne. The very belatedness of his regal calling doubtless made this lesson especially firmly taught. And he inherited his father’s favorite, Buckingham, as his first bad advisor. So character traits can get as well as give explanations. The character of an agent, and his age, make a difference to what he considers doing, and to what he decides to do. Cromwell was not too old to consider taking the crown he had longed for, but Hume says he was by then unwell and tired. Henry II’s imprisonment of his queen, Eleanor, for the last 16 years of his reign was due to her support of her son’s rebellion against her husband. He had expected her to forgive his many infidelities, but could not forgive what he saw as her treachery. When younger he had chosen her as his bride despite her record of sleeping with her first husband’s enemy, a handsome Saracen, met when accompanying her husband, Louis VII of France, on a crusade. But when he was older, Henry could not forgive her for supporting his enemies, even when they were their own sons. Hume had noted in his first Enquiry that tiredness and illness affect what we feel like doing. What he calls “self-command” over violent passions such as anger is lessened by exhaustion, and never expected of the very young. So the full explanation of what persons do may bring in their age, health, and character, to explain the puzzling desires or beliefs behind their action. Eating when hungry and drinking when thirsty do not call for such extended explanations, unless they are done on unsuitable occasions. If someone starts eating a chocolate bar during a wedding ceremony, the explanation may be that he is younger than his size would lead one to believe, so prone to such immediate gratification of desires. There are a whole range of ordinary actions, like getting up, breakfasting, taking out the garbage, going to work as usual, which are self-explanatory. Should anyone be puzzled at them, their puzzlement is what would call for explanation—are they from Mars? Or do they know something we do not, say that the one leaving for work was sacked yesterday, so has no work to go to? What calls for explanation does so against the background of normal activity that does not. Scientists may formulate the laws governing such normal

248

Acting in Character

activity, since they cultivate puzzlement at the familiar, so explain what ordinary people do not find puzzling. I am speaking of ordinary explanations, not of scientific ones, and such explanations need not explain by subsuming the event to be explained under some law. For actions, tracing causes may assume causal laws, but they rarely do the main explanatory job—that is done by supplying the facts, facts about pre-existent desires, about the agent’s beliefs, about her fears, about her character. When the explanation takes the form of citing the agent’s character, what it points to will not be something universally shared by all human agents, such as periodic hunger, thirst, sleepiness and so on. Wanting regular food, liquid, and sleep is part of human nature, not a character trait. Of course any character trait for which we have a name, such as greed, laziness, self-indulgence, arrogance, will be shared with many others, but not with all. So citing the agent’s character to explain her action is classifying her as a person of a particular sort, not just emphasizing that she is only human. Are desires and beliefs privileged, among what can explain an action, in that they will always be present, whereas other factors are only sometimes relevant? Emotion and habit seem only occasional explainers, since some actions are quite unemotional and many are not cases of habit. Actions from habit or routine need involve no occurrent desires or beliefs, but may be done semi-automatically. But character, like human nature, seems ever present, and always to some degree relevant. How could what the agent is like fail to be a factor in what she does? As Hume says, in his essay “The Sceptic,” we are “almost entirely guided by constitution and temper,” and “the fabric or constitution of our mind no more depends on our choice than that of the body.” Some character traits attribute particular strength to some desire, so that invoking that trait, say ambition, to explain action is no different from citing the desire. Others, like irascibility, point to the proneness of the agent to some passion, in this case anger, itself shown in typical action, and giving rise to typical desires. But other traits, like laziness, or pride, can explain without attributing any desire. If my pride makes me refuse an invitation to join some select society I had hoped to be asked to join 10 years earlier, the pride itself explains that. It is not that I want to offend them, although I might want that, nor is it that I want to be a nonmember—my earlier wish to become a member would make that claim hypocritical. The proud person expects the invitation at the right time, not belatedly. The lazy person does not want to bestir herself, which is not the same as wanting to do nothing. When urged to get up and get on with some task, she may believe that would indeed be a good idea, but still not muster the will to make the effort. She sinks back into inactivity, not because that is the state she enjoys most—she may be a guilt-ridden lazy person, who hates herself for her inactive life. The listless or depressed person is inactive because of the lack of normal motivation. Apathy is not a peculiar passion, it is the absence of normal passion.

Annette Baier

249

For Hume both our intelligence and our passions are a special case of animal nature, and throughout his Treatise of Human Nature he regularly pauses to glance at “Reason of animals,” “Pride of animals, “Love of animals,” and so on. (He notes that the automatic sympathy, contagion of mood, which he believes we display, is strikingly evident in dog packs.) There is sympathy among animals, but no “morals of animals,” let alone justice of animals, for that requires a reflection of passion back on passion, something which they do not seem to go in for. Explaining human actions is a special case of explaining animal behavior, itself a special case of explaining the movement and non-movement of the things around us. The branches grow that way to get the light. The mice play because the cat is away. The nature of the thing for whose behavior we seek an explanation is always a relevant explanatory factor. With animal behavior, we look to the instincts, typical wants and dispositions of that kind of animal. Only if it is doing something atypical will we need to look not just at its species but also at its individual character—“he’s a fearless old cat, especially intolerant of dogs, and does not realize they have dangerous teeth, while he has no teeth.” With human beings, we expect some standing wants and dislikes, as part of human nature. Treatises of human nature, which spell out the wants and beliefmaking habits of our species, contain few surprises. John Passmore went as far as to say that Hume, as psychologist, came out only with commonplaces.13 What may surprise us are the theoretical trappings in which Hume in the Treatise garbed his commonplace truths, which can be as strange14 as some of the variations that anthropologists find. Hume hopes his account of human habits and passions will cover not only the variations but the variations of the variations. There are, he believes, some constants, such as believing our eyes, feeling periodic hunger, thirst, and lust. “Is it more certain that two flat pieces of marble will unite together than that two young savages of different sexes will copulate?” (T 2, 3, 1, 8). There are variable factors, like causal inference based on experienced conjunctions, which some do more exclusively and carefully than others, believing what others tell us, and love of the marvelous, which inclines some to believe in tales of miracles. There are some, like sympathy, and kindness to children, which operate unless some opposing factor comes in. Belief in miracles, distrust of testimony, malice, cruelty to children, do happen, and, when they do, need more explanation than their contraries. Most human action is boringly normal, and does not call for any explanation, given accepted views on human nature. When the motivating influences on the will are standard desires in standard circumstances, and ordinary not at all puzzling beliefs about how to satisfy them, there will be no need to bring in reference to the agent’s character, should some perverse philosopher pretending to be a Martian (or a scientist) want the full explanation spelled out. It is when there are unusual desires or strange beliefs behind the action, that reference, say to the agent’s superstition, may be needed.

250 Acting in Character

Of course a good story teller can make quite ordinary desires look interesting in sequence. Life stories are unique, however common the actions and motives going into them. For the outcomes of actions are often unexpected, and it is the situations agents find themselves in, as much as their individual characters, which provide the interesting variety we find in life histories. As Hume says, it is “temper and situation” that regularly determine human action (T 2, 3, 1, 10). When he relates the life histories of the monarchs of England, and other leaders, it is situation that introduces most of the novelty. As far as temper goes, he finds few “extraordinary personages.” (Becket and Oliver Cromwell were “extraordinary,” partly in rising so high above their lowly origins.) Most monarchs are found to have been fairly ordinary persons, with the usual repertoire of passions, falling within the usual range of temper, showing the usual mix of virtues and vices. But they were often cast in especially demanding roles, so their lives may be more interesting than yours or mine, as well as affecting more people. Nor is the demanding role of monarch fixed. It developed and changed, so that the imperiousness that Elizabeth got away with was found intolerable in Charles I. Situation change, partly due to role change, is one main factor diversifying human lives, though as Hume noted, variations in “temper,” and the normal interplay of calm and violent passions, also contribute to the diversity (T 2, 3, 8, 12). Character too may change, and some like Becket make themselves into wholly new personages, when they take on new roles. Such abrupt changes usually arouse suspicions of pretence and hypocrisy, and Becket’s character gets from Hume a special examination, to see if the charge of hypocrisy is warranted.15 He finds in Becket’s favor, on this count, although plenty other grave charges remain. Abrupt changes can occur, repentance does happen, but usually something in the lasting character will explain the change. Hume himself appeared in different guises.16 His diplomatic personage in Paris was different from that of historian-librarian, which was itself a bit different from that of enquirer into human understanding, natural historian, and essayist. As one ages and has new experiences, old passions sometimes get tempered, old frenzies calmed. Hume the Edinburgh resident of St Andrews Square was not as hostile to religion as he had been when he lived in Jack’s Land. (When an evangelical woman, a candle-makers wife, came to try to get him to repent and die a Christian, he received her kindly, and ordered a large supply of candles from her.) Character traits such as kindliness may persist, while ambitions fade, or shift their focus. I have so far argued that the sort of agent whose actions we want to understand is always a relevant explanatory factor for understanding human actions, so individual character always counts. Character may develop and improve, as Hume saw Cromwell’s to have done, and choices may play some role in character change. The young Hume who wrote the Treatise was pessimistic on that score, seeing it as more likely that any changes for the better will be due to improved “situation.” When character explains action,

Annette Baier

251

the character trait cited, such as Cromwell’s ambition, may specify which desires are likely to be strong. Sometimes the character trait cited will say something about the agent’s beliefs (superstitious, far-sighted, bone-headed). But some character traits, such as slyness, laziness, hyperactivity, arrogance, diffidence, impatience, vacillation, will explain without attributing any particular desires or beliefs to the agent. So Velleman is right that explanations that cite character to explain action do not necessarily reduce to beliefdesire explanations. And often they explain puzzling operative desires and beliefs, so go further than explanations in terms of them can go. We all most of the time necessarily act in character, so character is usually relevant to explain what we do. The necessarily unusual cases where someone acts out of character are the cases where the need for explanation is most strongly felt, at least by those who know the agent. They may be just the cases where belief and desire will do the explaining. Why did Henry II make Becket, who was proving an excellent chancellor, archbishop of Canterbury? He was usually a good judge of character, and knew his old friend Becket very well, yet apparently did not foresee that Becket, once head of the church in England, would seize the opportunity to make it, and himself at its head, a rival power to the throne. It was unlike Henry to show such bad judgment. What explains his out-of-character action? He believed that Becket agreed with him about the proper role of the church in England, and had good enough reasons to think that, since Becket, before donning sacerdotal robes, had helped him oppose the church’s pretensions. Becket was his old friend, and he wanted to load him with dignities. So he made him primate, and lived to rue the day. In a way it was in character for Henry to expect the best of his old friend, to be initially unbelieving of his perceived treachery. He was similarly outraged at his loved sons’ rebellion against him, and his queen’s siding with them. Trusting those he was fond of was in character for him. With Becket, whom he had “honoured with his friendship and intimacy,” his trust was misplaced. Hume writes: “never prince of so great penetration, appeared, in the issue, to have so little understood the genius and character of his minister.” (H 1, VIII, p. 309) The out-of-character misjudgment of how Becket would act as primate is explained by his affection for Becket, his wish to heap honors on him, and his belief that Becket agreed with him about the proper subordination of princes of the church to the civil magistrate. These desires and beliefs were not surprising, given the past friendship between the two, and Becket’s seeming agreement with Henry, while he was chancellor, on the proper role of the church. He did not foresee Becket’s total alteration in manner of life, and in ambitions for the clergy, once he headed the church. He found himself “grievously mistaken.” Another out-of-character action of Henry’s was his hypocritical show of penitence at Becket’s tomb. And this is easily explained by his desire, after Becket’s murder, to avert the “thunders of the Vatican,” and get on better

252

Acting in Character

terms with the English bishops. His usual good judgment on matters of policy was at work here, and he was willing to sacrifice his usual openness and lack of guile to secure better relations with the church. Both these cases of out-of-character action were also in character, in another respect. In the first case, reliance on friendship was in character, misjudgment of Becket’s ambition out of character. In the second case, Henry’s usual good judgment was in play, along with an unusual willingness to pretend to feelings he lacked, in a spectacular public show—having his bare back whipped by monks, after fasting and praying all night at Becket’s tomb. And this four years after the murder. That, when he first heard of the murder, he should feel some remorse for his expression of anger at Becket, which led his knights to set off to murder the prelate, was understandable. He did fast and pray in solitude then for three days. But then he went about his and his nation’s business for four years, quelling riots in Ireland, Scotland, and Normandy, before feeling the need to put on his penitential act at the new saint’s tomb. Its timing was due to unrest in England, led by bishops, as well as in Scotland, and dangers in Normandy. His army’s victory over the Scots on the day after his penance “was regarded as an earnest of his final reconciliation with heaven and with Thomas a Becket” (H 1, IX, p. 355). The ambitious Oliver Cromwell’s refusal of the crown of England might also be seen as out of character.17 Had he taken it, Hume says, he would have been seen as a “perfidious traitor” to the anti-monarchical cause. Velleman postulates, as distinctive of and intrinsic to human agency, a need to make sense of what we are doing, where making sense is fitting into some intelligible ongoing narrative, consistent with a certain self-conception. Reasons for action provide this sort of coherence. “We make sense of what we do by doing what makes sense.”18 If Cromwell’s self-conception was as constable, keeping the peace of the parish, at times by the use of force, then his refusal of the crown fits into this narrative about one called to be a chief-constable. But of course there can be more than one narrative option. Cromwell’s life would have made sense, but a different sense, had he become king. He could have seen it as another crowning mercy, this time literally crowning. He would not necessarily have had an “identity crisis,” of the sort that according to Velleman threatens those who act against their own self-image.19 A little “cognitive dissonance,”20 may have been felt, if the one who had had a Stuart king deposed and executed, the one who, with the help of John de Witt, had tricked the Dutch into agreeing, in the treaty of Westminster, 1654, to exclude the house of Orange from ruling again, were now, in 1657, to inaugurate in Britain the house of Cromwell. But dissimulation, on Hume’s story, had been Cromwell’s trademark all along, so this could be have been the great dissimulator’s finest hour. (“Had you fooled, you gullible anti-monarchists, didn’t I?”) The alternative story would, however, need a bit of self-interpretive work, and Cromwell may not have been quite up to it. He liked to echo the Bible, especially the Old Testament, so preferred to

Annette Baier

253

have some readymade script to follow, rather than write his own. Hume says that his chaplain, Hugh Peters, often took as his text, Psalm 149, referring to a twofold sword in the hands of the saints, to execute vengeance on the heathen, to bind their king with chains and their nobles with fetters of iron (H 5, LIX, p. 515, note). When he refused the crown, he likened the defeat of Charles I to the falling of the walls of Jericho, and said he would not try to rebuild them. And he did have, to guide him, the prophecy that Clarendon refers to, that he would become the greatest man in England and would be near to becoming king. To refuse the crown, so that the prophecy might be fulfilled, would be one way to give one’s life what Velleman terms “sense,” to see it as following a script, a book of life, already written. Cromwell’s ambition had shown no signs of flagging before his remarkable refusal of the crown. But it had already been well satisfied. Is satiety a sufficient explanation of why one refuses a second helping? “He’d had enough” is an under-recognized explanation of action. Then there is the force of inertia to be considered, the disinclination to make a change in the ways things are, especially if things are going fairly well. Hume of all people knows the force of habit in a person’s life, and the older we grow the more disinclined we often become to make new departures, as long as we are fairly content with how things are. To insist on formulating this sort of reason for refusing some new opportunity as a “desire” to keep things as they are seems to debase the meaning of “desire.”21 Hume uses the term for the direct passion that has some not-yet-possessed pleasuregiving thing, such as an attractive fruit, or a jeweled crown, as its object. Cromwell’s longing to be king was a desire. But if it “languished,” unseconded, or seconded by the wrong people, what would take its place would be habit, and pleasant facility in keeping doing what he knew how to do. Normal human activity, keeping doing one’s job, needs no explanation. We do not “desire” to get up as usual, eat breakfast, and go to work. We may occasionally desire not to, and that is when explanation by citing a desire has its place. It is self-proclaimed “Humeans,” not Hume himself, who have no room for any explanation of what we are doing except desire satisfaction. Hume in an Enquiry Concerning the Principles of Morals, Appendix 1 says that sentiments and affections set all the ends of human action, and even in the Treatise it was “the passions,” not just the direct passion of desire, that reason was to be slave to, and that were said to be needed to motivate action. Passions are more inclusive than Davidson’s “pro-attitudes” since they cannot all be characterized as “pro” or “con” something, so modern notions of “directions of fit” do not apply to some of them (pride, for example). Some, like love, often give rise to desires, but not always to the same desire. Hume sees no contradiction in supposing that love might lead us to torment the beloved. Hume says love can show itself in a hundred ways in addition to wanting to make the loved person happy (T 2, 2, 6, 5–6). Esteem for the rich and powerful need not motivate

254 Acting in Character

any action toward them, except behaving in a deferential way in their presence. Passions, for Hume, include such things as esteem, respect, contempt, admiration, envy, resentment, despair, world-weariness, antipathy to treachery, love of family and friends, love of truth, every one of which can show itself in a hundred different ways. Then there are all the passions we come to share by sympathy. Did Cromwell, late in life, come to share by sympathy the distress of his youngest daughter, Elizabeth Claypoole, at the amount of bloodshed her father was responsible for? Did he remember his own threat that he would have Charles beheaded with the crown on his head, and dread the same fate? Many passions may have joined forces to make Cromwell refuse the crown. Cromwell’s refusal of the crown may have shown more “sagacity” and less ruthless ambition than some of his earlier actions, but was not wholly out of character. It is most unlikely that any of us ever act totally out of character. Even Becket, when he donned the character of sanctity and started mortifying his hitherto well-indulged flesh, still showed his old concern for how he looked—sackcloth peeping out from his archbishop’s robes. When we act out of character in some respect, it will be still in character in another respect. Still, if the usual explanation of out of character action is to cite some belief and some desire, themselves not out of character, then that might explain why beliefs and desires have been the preferred philosophical explainers of action.22 For we are more likely to look for explanations of what seems out of character than for what is in character. But the explainers of the exceptional need not be the best explainers of the less exceptional. Explaining by citing a character trait of the agent is appropriate where the action explained is not so ordinary and everyday as to be self-explanatory, nor so unusual, for this particular agent, as to seem out of character. My unexciting conclusion is that character can explain action, as long as that is not out of character.23 And even then, the beliefs and desires that do explain it may be quite in character. Beliefs and desires can explain out-of-character actions, but in other cases often themselves call for explanation, and can get it by citing character. The chain of explanations ends in different places in different cases. What counts as a good explanation of an action, in ordinary life, is relative to the prior knowledge and present puzzlement of those to whom the explanation is offered.24

Notes 1. References to Hume’s A Treatise of Human Nature abbreviate the title to T, followed by book number, part number, section number, paragraph number, as numbered in the Norton and Norton Clarendon edition. 2. I briefly discuss the views of those who think there is a fundamental error in attributing lasting character traits to people in “Demoralization, Trust, and the Virtues,” in Setting the Moral Compass edited by Cheshire Calhoun, (Oxford, 2004) pp. 176–188.

Annette Baier

255

3. For mirror neurons, see Giacomo Rizzolati, Leonardo Fogassi, and Virttorio Gallese, “Mirrors in the Mind,” Scientific American, November 2006. 4. Philippa Foot, in “Freewill as Involving Determinism,” challenged it, and noted how odd a view it was for Hume to hold, if he believed a person is no more than the series of her momentary perceptions. See Virtues and Vices (Oxford, Blackwell, 1978), pp. 70–71. 5. I discuss this tunnel vision of Hume’s in “The Life and Mortality of the Mind,” in Anette C. Baier, Death and Character: Further Reflections on Hume (Cambridge, MA, Harvard University Press, 2008). 6. David Hume, Essays, Moral, Political and Literary, edited by Eugene Miller, (Liberty Classics, Indianapolis, 1985) 566. 7. See J. David Velleman, Practical Reflection (Princeton, Princeton University Press, 1989), 249. 8. Reference is to Volume 5, Chapter LVI, p. 175 of the Liberty Press edition of Hume’s History of England. Subsequent references will be given in the same way. 9. Geoffrey Robertson, in The Tyrannicide Brief: The Story of the Man who sent Charles 1 to the Scaffold, (Vintage Books, London, 2006) p. 62. 10. Was that most famous passage of the Treatise, that it is not contrary to reason to prefer one’s lesser good to one’s greater (T 2, 3, 3, 6), one reason why Hume asked us not to read his first work? He certainly has not had his request granted, as far as that passage goes! The claim is not repeated in his dissertation, “Of the Passions,” nor in his second Enquiry. 11. Alan Musgrave, in “Explanation, Description, and Scientific Realism,” Scientia 727–740, objects to the assumption that a good explanation has to be an “ultimate” one. I am not assuming that, merely assuming that it not leave us more puzzled than we were before the explanation was offered. Musgrave also rejects the “subjective” view I am adopting, that explanations succeed when puzzlement is ended. For some Yorkshire people, he tells me, any human action can be “explained” in my weak sense by “there’s nowt so queer as folks.” But I am requiring an explanation of action to cite well-based beliefs about what the agent is like, or well-based beliefs about what beliefs, desires, and emotions the agent had. We may indeed be all “queer,” but character explanations specify the particular brand of queerness that was operative. Of course if we are wrong about the agent’s character, the explanation will fail, just as any explanation fails if it invokes a pseudo-law or pseudo-fact to do the explaining. 12. Once when a friend tried to persuade me to join a health club to which both she and my husband belonged, pointing out how much good it might do me, and how easily I could accompany them there, my friend remembers that I replied, “No thanks. I am not the kind of person who joins health clubs.” (Nor apparently one who takes much advice, even from good friends.) The kind of person I was and saw myself as settled the matter. 13. John Passmore, Hume’s Intentions (Cambridge, Cambridge University Press, 1952) p. 157. 14. I explore his metaphors and theories in “Hume’s Impressions and his other Metaphors,” and “The Life and Mortality of the Mind.” 15. I discuss this case, and that of Henry II, referred to later, in “Hume’s Excellent Hypocrites,” New Essays on David Hume, edited by Emilio Mazza and Emanuele Ronchetti, (Filiosfia e scienza nell’eta moderne, Francoangeli, Milan) pp. 267–286. 16. M.A. Box, in the Suasive Art of David Hume, writing about Hume’s early essays, says “those who had worked their way through the profundities of the Treatise

256 Acting in Character

17.

18. 19. 20. 21. 22.

23.

24.

might find it difficult to accept that the same man had written both books.” (p. 118) I look at some of these changes, and the unity within the life they occurred in, in “Hume’s curriculum vitae, his own life, written by himself.” I discuss this in “Hume’s Treatment of Oliver Cromwell,” in Anette C. Baier, Death and Character: Further Reflections on Hume (Cambridge, MA, Harvard University Press, 2008). J. David Velleman, Self to Self: Selected Essays (Cambridge, Cambridge University Press, 2006) p. 283. Ibid., p. 4. Ibid., p. 227 ff. See my “The Ambiguous Limits of Desire,” The Ways of Desire, edited by Joel Marks (Precedent Publishers, 1986). I wrote a paper, for a festschrift for Carl Hempel, called “Explaining the Actions of the Explainers,” Erkenntnis 22 (Reidel Publishing Co., 1985) pp. 155–173, so it is quite in character for me to turn explanation on explanation. Like Hume, I enjoy making a reflexive turn. What are we to make of the older Hume’s action in disowning his Treatise? Was that in character or out of it? Since he owned it again, in My Own Life, I regard it as out of character, explained by fairly lasting spleen over critics like Beattie who ignored his request that the Enquiries be taken to express his considered views. The spleen lasted quite a while, for a man famous for his good nature, and for his difficulty in cultivating the hypocrisy that he saw to be needed to pass through the world without bringing oneself and one’s companions unnecessary distress. For it was a touch hypocritical to hide his responsibility for the Treatise behind the fact that his name had not appeared on it in its first edition, and false to say he had never acknowledged it, since he had eagerly acknowledged it to such people as Pierre Desmaizeau, Henry Home, Frances Hutcheson, and others whose opinion he respected. His strong desire to divert attention from it to his underdiscussed Enquiries, and his belief that vociferous critics of the Treatise such as Beattie were “bigotted” and “silly,” explain why he did an uncharacteristically splenetic and even false thing. I thank Alan Musgrave for his skeptical reactions to an earlier version, and Constantine Sandis for constructive suggestions.

14 Aquinas on Action and Action Explanation1 Stephen Boulter

Introduction Perhaps no other major figure in the West has devoted as much time and space to the development of a theory of action and action explanation as Thomas Aquinas. At a minimum a complete study of his views would require close attention to several magisterial ‘mini’ treatises, including the Treatise on Man (Summa Theol. I-I, q. 77–89), the Treatise on Human Acts (Summa Theol. I-II, q. 1–21), the Treatise on Habits (Summa Theol. I-II, q. 49–89) and the Treatise on Law (Summa Theol. I-II, q. 90–108). What is more, no such study would be complete without consideration of Aquinas’ lengthy and detailed commentaries on Aristotle’s Nicomachean Ethics and De Anima. Of course there is no question of doing justice to all of this material here. And since it would be unreasonable to assume a familiarity on the part of most readers of this volume with the general shape of Aquinas’ thought, let alone the details of any position he might have entertained, it would be inappropriate to enter here upon a sustained discussion of any particular aspects of Aquinas’ views on action, however tempting that might be. Consequently, the purely exegetical aspects of this essay will be confined to a presentation of the general plan or schema of Aquinas’ account of action, the many details of which we must let pass sous silence. However, my main concern in this essay is not primarily exegetical but rather to bring the outlines of Aquinas’ thought into contact with contemporary work within what is roughly called the analytical tradition. Aquinas develops a sophisticated version of causalism which is sufficiently distant to us in historical time and intellectual context to cast a distinctive light on otherwise very familiar territory. Seeing an old subject with new eyes is always rewarding and often highly suggestive. To this end, and in keeping with the spirit of Analytical Thomism, I want to consider how Aquinas might cope with several puzzles that currently confound contemporary causalists. It is here that Aquinas is perhaps at his most intriguing. The ensuing centuries have seen many of Aquinas’ views on important matters dropping 257

258

Aquinas on Action and Action Explanation

out of favour with the philosophical community. These views – on topics ranging from metaphysics and methodology, the philosophy of mind, and the nature of explanations – have been replaced by a mind-set entirely foreign to his way of thinking. Moreover, it is precisely these developments which have resulted in the emergence of certain puzzles relating to action explanation, puzzles which simply do not arise for the Dominican. Whether we should revert to the views of Aquinas is of course another matter; but the fact that his account of action and action explanation steers clear of not insignificant difficulties should be enough to win him serious attention.

On the centrality of action to Aquinas’ thought A concern with action and action explanation is far from being an incidental ingredient in Aquinas’s thought. In fact the various notions of action to be found in Aquinas are central to some of his most important projects, both philosophical and theological. In order to properly understand these concepts, however, it is important to understand why action and action explanation was such a central concern for Aquinas. In order to avoid anachronistic readings of Aquinas, we need to be clear on the precise nature of the problems Aquinas hoped to address with his theory of action. That action is no incidental side issue for Aquinas is made clear from the opening remarks of his mini treatise on human action: Since happiness is to be gained by means of certain acts, we must as a consequence consider human acts in order to know by what acts we may obtain happiness, and by what acts we are prevented from obtaining it ... . In treating of what is universal in human acts, the points that offer themselves for our consideration are (1) human acts themselves; (2) their principles. (Summa Theol. I-II, preamble to q. 6) This concern for human happiness, no small matter for a theologian or philosopher, is one of the key motivations behind the development of an elaborate theory of action and action explanation. But as we shall see, the concept of action is also central to his metaphysics, methodology and philosophy of mind. In the first and widest sense of action, anything which has a causal influence on things outside itself can be said to act. It is in this widest sense of action – roughly equivalent to an unqualified ‘doing something’ – that inanimate as well as animate objects can be said to act. It is this sense of action that permits Aquinas to speak of ‘natural agents’, and to say, for instance, that physical bodies act on one another inasmuch as they attract each other through the force of gravity. It is in this sense that Aquinas says that particular kinds of inanimate bodies have their own particular actions. Examples of this kind of action are frequently taken from chemistry, where

Stephen Boulter

259

particular substances are said to have their own distinctive causal role to play in chemical processes. In this sense one can say that rust is the result of the action of the oxygen on iron, or that the action of Aqua Regis on gold is to dissolve it. One of Aquinas’ standard examples of the actions of ‘natural agents’ of this sort is that of fire heating something.2 Aquinas also uses a notion of action with a narrower sense which is restricted in application to animate or minded objects. These actions are characterised by the fact that their explanation requires the attribution of mental states to the agent. In this sense of action, anything which moves itself in response to internal and external stimulus is said to act. It is this sense of action that applies to both human and non-human agents to the exclusion of inanimate objects. Finally, Aquinas has a concept of action which is entirely restricted to human agents. Aquinas distinguishes between mere bodily movements and what he considers to be properly human actions (Commentary on the Nicomachean Ethics, Book I. Lect. 1). Movements such as absent mindedly scratching one’s beard or automatically brushing one’s teeth are mere doings to be distinguished from properly human actions, the latter being movements or behaviours lying under the control of the agent and proceeding from a deliberate will. In this most restricted sense actions require deliberate attentional resources, and usually involve (a) planning and decision-making, (b) problem solving and (c) a departure from habitual and automatic behaviours. Actions in this sense are not found among the other animals and are the characteristic preserve of human beings. For ease of exposition in what follows I will call such ‘properly human actions’ ‘human actions’ for short, but it should be kept in mind that I am using this term to refer only to an important sub-set of human actions. Each of these concepts of action has an important role to play in Aquinas’ philosophical framework. The first sense of action, which applies to animate and inanimate objects alike, lies at the very heart of his method in epistemology. Because this is likely to strike many readers as novel it is worth pausing for a moment to consider this aspect of Aquinas’ methodology. Following Aristotle (Post An. II, chs 1–2), Aquinas thinks that a full understanding of any particular object of scientific knowledge consists in our being able to provide answers to four distinct but related questions. Of any given X we may ask (1) Does X exist?, (2) Does X have property Y?, (3) What is X? and finally (4) Why does X have property Y? The important distinction here for present purposes is that between knowledge of particular facts known via sense perception and knowledge of ‘reasoned facts’, that is, the distinction between being able to state that such and such is the case and being able to explain why such and such is the case. It is not enough, for example, to know that rust is the result of the action of oxygen on iron; for Aristotle and Aquinas’ full scientific understanding is achieved only when we have arrived at explanations, that is, at accounts which make clear why things have to be the way they are.3

260 Aquinas on Action and Action Explanation

In addition to this emphasis on the importance of explanations, both Aristotle and Aquinas share the belief that answers to questions of type (4), questions of the form ‘Why does X have property Y?’, can be answered (if they can be answered) only after one has answered questions of type (2), that is, the ‘What is X?’ question. The leading idea is that one will understand why X is Y when one can see that X’s being Y follows necessarily from the essential nature of X. Thus answers to questions of type (4) are to be given the following form: X is Y because it is of the nature of X to be Y, or: X is Y under these specific conditions because of the nature of X. The key idea here is that each and every determinate object has an essence (a core set of properties which together make X the sort of thing X is and allow one to track X through its various changes) and this essence has an explanatory role to play in our efforts to understand the object. Given this framework identifying an object’s essence is the sine qua non of understanding that object. We are then left with an obvious (and infamous) question: How does one go about identifying an object’s essence when essences are not directly observable but are thought to lie ‘behind’ or ‘underneath’ an object’s phenomenal properties? Many have thought the task impossible. Essences are ‘that I know not what’, some mysterious and unknowable entity postulated to unify the disparate phenomenal properties of things. There is no doubt that Aquinas would have sympathised with some of the concerns familiar in the early modern period, and it is clear that he did not think we could ever come to know completely the essence of anything.4 But Aquinas did not consider the task hopeless, and it is here that his concern with action in the widest sense takes centre stage. A fundamental methodological rule which Aquinas applies in all contexts where the study of essences is concerned is the following: The nature or essence of any X is its set of capacities or capabilities. That is, what something is essentially is determined by what it can do, and so X is defined by what actions it is able to perform (where action’ is taken in the widest sense). Moreover, X’s capacities or capabilities are identified in the first instance by looking at its actual acts, the activities which it is actually engaged in.5 Now if one takes this functionalist approach to essences then the essence of X does not appear to be epistemically unapproachable despite the fact that it is not directly observable. But this epistemological gain is bought at a price – one must be able to identify and understand an object’s acts and be able to distinguish these from events in which the object just happens to be involved. Thus a theory of action is required if Aquinas’ hopes in metaphysics and epistemology are to be realised. A theory of action is also an essential ingredient of Aquinas’ philosophy of mind, both as a topic of discussion in its own right, but also as the basis of a distinct methodological approach to the study of the mind. On the first point, it was commonplace among the ancients that a concept of mind is required in order to account for two striking features of living organisms,

Stephen Boulter

261

namely, their ability to discern elements of their environment, and to move themselves about in this environment. The primary goal of a philosophy of mind is thus to provide complete accounts of cognition and action (where action is taken is the second sense outlined above), for these constitute mindedness itself (Commentary on the De Anima, Book III, Lect. XIV). Thus action is a central topic of discussion in Aquinas’ philosophy of mind. Action is important to Aquinas’ philosophy of mind for methodological reasons as well. There is a difficulty in studying the mind, for minds, like essences, are not directly observable (Summa Theol. I, q. 87, a.1). But as expected, Aquinas employs the methodological rule noted above to provide the initial leverage on the topic as follows: [T]he process of self-knowledge has to start from the exterior things whence the mind draws the intelligible concepts in which it perceives itself; so we proceed from objects to acts, from acts to faculties, and from faculties to essence. But if the soul could know its essence directly it would be better to follow the reverse procedure; for in that case the closer anything was to the soul’s essence, the more directly could it be known by the soul. (Commentary on the De Anima, Book II, Lect. VI. 308, 97–98) According to Aquinas we are able to come to understand something of the nature of our minds only indirectly via a combination of introspective and behaviourist approaches by reflecting on the mind’s own activities. (Here action is taken all three senses simultaneously.) To sum up: Aquinas works with a least three distinct notions of action. First, there is the widest sense of ‘action’ which is equivalent to ‘causally affecting something external to oneself’, a sense of ‘action’ applicable to inanimate as well as animate objects; second, there is a narrower sense of ‘action’ in which it is equivalent to ‘the agent’s moving of her body’, a sense applicable only to animate or minded objects; third, there is the most restricted sense of action in which it is roughly equivalent to a ‘bodily movement under the deliberate control of the agent’. The first sense is central to Aquinas’ methodology in metaphysics and epistemology; the second and the third are central to Aquinas’ philosophical psychology; and the third is central to Aquinas’ ethics, politics and his concern for eudaimonia in general.

Aquinas’ analysis of properly human action For the rest of this essay I will confine myself to Aquinas’ account of action in the third, most restricted sense of action. It is best to begin an account of Aquinas’ theory of properly human action with that element of his thought which is most familiar to those raised on the Humean belief/desire theory of motivation. Following Aristotle, Aquinas claims that the soul has two

262

Aquinas on Action and Action Explanation

motive forces by which it moves itself, the intellect (ratio) and appetency (voluntas) respectively ((Commentary on De Anima, Book III, Lect. XV, Commentary on the Nicomachean Ethics, Book I, Lect. 1)). On this view the intellect and appetency together are both causes of, and reasons for, particular actions. Much of what Aquinas has to say about beliefs and desires is familiar enough from Hume’s much discussed belief/desire theory of motivation. For instance Aquinas holds that the intellect is powerless to produce bodily movement on its own, while some forms of appetition can move independently of reason.6 However, we shall see shortly that Aquinas’ understanding of the nature of intellect and appetency as motive forces differs in important respects from Hume’s. In keeping with this view of the motive powers of the soul is Aquinas’ basic intuition regarding human action and action explanation, namely, that all properly human action is action for some end. Aquinas writes: ‘... every agent intends an end while acting, which end is sometimes the action itself, sometimes a thing made by the action’ (Summa Contra Gentiles, Book III, ch. 2). In short, agents act as they do because their beliefs and desires lead them to adopt certain goals, and these goals constitute their reasons for actions. The importance of the agent’s intentions in determining which behaviours count as human actions is made perfectly clear from the following: Now there can be no doubt that those which act by intellect act for an end, since they act with an intellectual preconception of what they attain by their action, and they act through such a preconception; for this is to act by intellect. (Summa Contra Gentiles, Book III, ch. 2) It is important to emphasise even at this early stage that the manner in which an agent’s intentions cause the agent to move her body is just as important in determining whether a behaviour counts as an action as the fact that these intentions do cause the agent so to move her body. With these points in mind it seems reasonable to assume that Aquinas would be able to accept, at least as a provisional definition of action, the modern causalist definition which maintains that human actions are bodily movements intentional under some description that are caused in the right way by certain pro-attitudes (Summa Theol. I-II, q. 1, a. 1, 3). But this familiar causalist claim, central though it is to Aquinas’ account of action, is subject to considerable elaboration. First, Aquinas goes well beyond either Aristotle or Hume in spelling out in no small detail precisely how pro-attitudes bring about the agent’s moving of her body in such a way as to allow these movements to be classed as actions. Second, Aquinas embeds the causalist claim within a wider theoretical framework which embraces (a) ‘external’ as well ‘internal’ principles of action, and (b) an elaborate theory of human nature. According to Aquinas, any complete account

Stephen Boulter

263

of human action will have to appeal to external as well as internal principles of action, as well as a theory of human nature. Nonetheless, this central causalist definition of action is an important starting point in our account of Aquinas’ theory because it specifies the kind of theory Aquinas is offering. What is crucial to this definition of action is the claim that it is the causal history of an event which distinguishes properly human actions from events and mere doings, and it is this claim which marks Aquinas as a causalist. I want to dwell on this point for a moment in order to draw out two important points. If one assumes that a human action is an event with this sort of causal history, then there are important implications for the theory of action explanation. First, since Aquinas assumes that human actions have an internal principle, he assumes that it is within the control of the agent to perform or to refrain from performing any such action.7 If this is so then any complete explanation of a human action must provide answers to two distinct questions. We will want to know in the first instance why the agent exercised her capacities for action when she did rather than remain at rest; and then we will want to know why, having decided to exercise her capacities for action, she performed this particular action rather than some alternative. Providing answers to these two questions is the task of a theory of action explanation (Summa Theol. I-II, q. 9, a.1). A second consequence of the causalist definition of action is the particular slant it suggests on action explanation. If human actions are events caused by an agent’s pro-attitudes, and if one explains something by identifying its causes (rather than, say, subsuming it under a law of some kind) then properly human actions are to be explained by reference to an agent’s beliefs, desires, decisions and choices because these are (among) the causes, ‘motive-forces’ or ‘principles’ of the behaviour in question. (Summa Theol. I-II, q. 1, a. 1. See also Aristotle’s De motu animalium, ch. 6). Of course this is to say that, according to Aquinas, an agent’s reasons do not simply render an act intelligible in the sense that they allow one to understand why the agent might have acted as she did; since the reasons number among the causes of her behaviour they will figure in the explanation of her action as well. What is more controversial is that, according to Aquinas, if the reasons are of the right sort, they can also serve to justify the action. That is to say, in Aquinas, psychological and normative reasons can be, and often are, identical. This is a point to which I shall return below. Apart from the identification of psychological and normative reasons Aquinas’s theory so far differs little from most versions of causalism. But as mentioned above, this central claim is only one element of Aquinas’ account of action. For in addition to these familiar internal principles Aquinas posits what he calls external as well as internal principles of action. It is not just beliefs and desires of the agent that figure in an account of

264 Aquinas on Action and Action Explanation

action. Objects and states of affairs in the world also have a central role to play. In fact it would not be unreasonable to assert that for Aquinas these external principles are more fundamental to action explanation than the internal principles favoured by Hume, a point to which we shall return in due course.8 The combination of internal and external principles, the interdependency of beliefs and desires, and an appeal to a theory of human nature are important features of Aquinas’ theory of action explanation which render his version of causalism distinct from those more familiar versions developed in modern times. These extra ingredients can be introduced in a schematic presentation of Aquinas’ general theory as applied to any particular action. If we want to know why an agent performed a given properly human action on a particular occasion then the complete answer we are looking for will rely implicitly or explicitly on the following points: 1. There is an external world, populated by objects with their particular properties and layout, which constitutes the theatre of action in which human agents find themselves. 2. The human agent has an ordered set of standing natural inclinations characteristic of the species, the satisfaction of which constitutes the achievement of the agent’s general good qua human being. These inclinations are to (i) basic self-preservation, and to the means necessary for the preservation of one’s being (an inclination shared with all substances); (ii) to the union of male and female, and to the rearing and education of children (an inclination shared with the animals); and finally we have inclinations based on our specifically rational capacities, in particular: (iiia) to seek knowledge and avoid ignorance, and (iiib) to live peaceably in society. These inclinations point to or indicate those ends towards which our properly human actions are ordered. 3. The human agent, unlike other animals, has a standing natural (although often incomplete) understanding or apprehension of what her general welfare consists in, and of what is required for the satisfaction of her basic needs. This understanding is founded on the following basic principle: ‘... all those things to which man has a natural inclination are naturally apprehended by reason as being good, and consequently as objects of pursuit, and their contraries as evil, and objects of avoidance’. Points (2) and (3) are the basic elements of Aquinas’ theory of natural law (Summa Theol. I-II, q. 94, a. 2). A law, according to Aquinas, is ‘a rule and measure of acts, whereby man is induced to act or is restrained from acting; for lex [law] is derived from ligare [to bind], because it binds one to act’ (Summa Theol. I-II, q. 90, a. 1). The point about natural law, however, is that there are principles by which human beings are induced to act or are

Stephen Boulter

265

restrained from acting which, while discoverable by reason and ultimately in accordance with reason, are ‘consequent upon’ on human nature. By calling these inclinations ‘natural’ Aquinas is insisting that they are not optional. All creatures, including human agents, cannot but desire their general welfare and the means to its attainment. Indeed we are determined by our constitution to desire (1–3b), and to find the satisfaction of these inclinations to be reasonable and good. A result of this is the following: If an object or end were to appear good to an agent from all perspectives and in all respects, then we could not but desire that object or end and see the satisfaction of that desire as reasonable. As it happens, no particular object or end (apart from God) ever appears to be good in all respects and from all perspectives. So while we are determined by our nature to seek our general welfare as far as it is known to us, we are not determined to any particular object or act. Our freedom thus lies not in the choice of our ultimate ends, but in our choice of the means by which we pursue these ends. We cannot pursue the details of Aquinas’ theory of natural law here, but what is crucial for present purposes is the appreciation of Aquinas’ basic claim, namely, that human beings have a set of natural inclinations geared to what is objectively good for us given the sort of animal we are, and that everything towards which we have a natural inclination is naturally perceived by reason to be good and something to be pursued, while those objects from which we naturally recoil are perceived by reason as objects to be avoided. Points (1)–(3) constitute the general background against which particular actions on particular occasions are to be considered. These are perfectly general points which Aquinas takes to apply to all psychologically healthy human agents at all times. Moving now from this general background to actions performed on particular occasions, Aquinas maintains that a particular action begins with, 4. The agent’s apprehension of some particular object or end which is understood to be a good for that agent. Aquinas calls this object or end the ‘desirable object’. The desirable object is the ultimate external principle of particular actions on particular occasions since it is an ‘unmoved mover’ inasmuch as it brings about a change in the psychological state of the agent without itself being changed. It is thus the first link in the causal chain leading to a particular action. (Summa Theol. I-II, q. 6, a. 2). Note that the causal chain leading to bodily movement begins with an internal, non-transitive act of the intellect which Aquinas calls apprehensio finis. 5. The agent’s apprehension of the desirable object brings about a particular desire for that object in the agent and a (revocable) intention to pursue it (intentio). Aquinas’ point here is that it is only when a particular external

266

6.

7.

8.

9.

10.

11.

12.

13. 14.

15.

Aquinas on Action and Action Explanation

object or end is understood to be a good that the agent comes to desire it (Summa Theol. I-II, q. 12, a. 1). So although there is a standing general desire for the satisfaction of one’s needs (voluntas simplex), one is not able to desire anything in particular without first having appreciated its status as a good to be pursued. The intention to pursue an end is followed by what Aquinas calls ‘counsel’ (consilium). At this stage the agent begins to deliberate about the available means that can be employed in the pursuit of the intended object or end (Summa Theol. I-II, q. 14). Council is followed by the judgment that actions x, y and z are acceptable means which can be employed in pursuit of the intended end (iudicia) (Summa Theol. I-II, q. 13). If there is only one such action the agent proceeds to Step (9), otherwise her judgment is followed by. The agent’s giving consent to the set of means (consensus). At this stage the agent accepts that actions x, y and z are all feasible and acceptable ways of achieving her end, and so each is considered to be a ‘live option’ and the agent consents to take each option seriously (Summa Theol. I-II, q. 15). The agent’s consent is followed by choice (iudicum electionis). At this stage the agent plumps for one option in preference to the others. (If the agent judges that there is only one acceptable course of action open to her then consent and choice are identical.) (Summa Theol. I-II, q. 13). The agent’s choosing a course of action is followed by the formation of a specific intention to adopt that course of action (electio) as a means to the desired end (Summa Theol. I-II, q. 13). The formation of the intention to pursue a particular course of action is then followed by self-direction, that is, a command given to the self to perform that action (imperium) (Summa Theol. I-II, q. 17). A command to the self is then followed by an exercise of one’s physical capacities to carry out the command (usus) (Summa Theol. I-II, q. 16). Technically usus is encountered at earlier stages in the causal chain as the agent employs her capacities throughout the course of the causal chain leading to action. What is different here is that we move from the exercise of strictly internal or psychological capacities to the application of one’s bodily members to the execution of the action. Usus is then followed by. The external action itself (a moving of the agent’s body). After which, if all goes well, the agent apprehends that the action has been accomplished (cognitio finis in actu), (Summa Theol. I-II, q. 11, a. 1 ad 3). Which in turn is followed by. Satisfaction at the achievement of the desired end (fruitio). (Summa Theol. I-II, q. 11).

Stephen Boulter

267

This is Aquinas’ account (in schematic form) of the paradigm case of properly human action. Of course many actions will fail to conform in all details to this reconstruction of the psychological antecedents of action, for being a properly human action is a matter of degree. But what is clear is that a bodily movement will not count as a human action simply because it is caused by the pro-attitudes of the agent. These pro-attitudes must cause the bodily movement in something approaching the manner outlined above if one’s moving of one’s body is to count as a properly human action, a point to which we shall return below. Aquinas’s extended discussion of these various stages in an action is very often taken up with the issue as to whether the stage in question belongs properly to the intellective or to the appetitive power. The details of these discussions will have to be set aside, but two important points need to be emphasised for present purposes: First, the appetitive power, often referred to as the will, is not an irrational or arational motive force as it often appears to be in the work of Hume. The basic instinctual drives – hunger, thirst, sex, self-preservation and so on – are indeed arational, but they are not specifically characteristic of human beings. A bodily movement driven entirely by instinct is not a properly human action. The crucial point for both Aristotle and Aquinas is that the appetitive power or will peculiar to our species is not blind instinct but rather our capacity to be responsive to reasons. Arational animals do not act for reasons, they simply follow their natural instincts. Of course human beings often behave instinctively; but we have the additional distinctly human cognitive capacity of reason which allows us to act not merely from inclination but for reasons the force of which we are able to appreciate. It is precisely this ability to perform an action because we understand that it will further our interests – even if it first appears to be counter to our initial, and perhaps habitual inclinations – that distinguishes us from the other animals. The second point follows closely on the first, namely, that the relationship between the products of the intellect and will is one of close and inextricable interdependency in the psychologically healthy human being inasmuch as what one can intelligibly desire very much depends on what one believes, a point that Hume is often taken to deny.9 As one proceeds through Aquinas’ account of the psychological antecedent leading to action one is struck by the fact that an act of the intellect gives rise to a response on the part of the appetitive power, which in turn leads to a further act of the intellect, which in turn leads to a further desire, and so on and so forth until the process results in an intentional bodily movement under the deliberate control of the agent. Remove any of the acts of the intellect in this causal chain and the subsequent acts of the will will fail to materialise. What is more, only certain responses on the part of the will will be intelligible given what the agent has already

268

Aquinas on Action and Action Explanation

accepted intellectually. [Finnis presents this point graphically with the follow chart (taken from his 1998): REASON (ratio) One’s understanding and reasoning

WILL (voluntas) One’s responsiveness to reasons

(1) Intellectus finium Understanding basic ends/goods (3) Apprehensio finis Envisaging a possible purpose for action (5) Consilium Deliberation (when needed) about means – devising of possible options

(2) Voluntas simplex Basic openness to these goods (4) Intentio Interest in pursuing that purpose (6) Consensus (initial) Assent to the inherent interest or ‘liveliness’ of all the options available as means (8) Consensus (firm or final) Assent to options x, y and z as sufficiently interesting to be considered live options

(7) Iudicia A judgment is made that options x, y and z are each practical/suitable/ appropriate – with none being ‘dominant’ (9) Iudicium electionis A judgment of preference made in choosing (this, x, is what I ought to do) (11) Imperium Self-direction by using the chosen proposal as a ‘rule of action’ and directive to act (13) Cognitio finis in actu Knowledge of action’s success in achieving its end

(10) Electio Formation of a definitive intention and adoption of one proposal by choice (12) Usus Exercise of one’s physical capacities and skills in carrying out the external act of choice when the occasion arises or presents itself (14) Fruitio Taking satisfaction in action’s achieving its end

Aquinas and the contemporary challenges to causalism Let this suffice for the moment as a schematic account of Aquinas’s theory of action and action explanation. It can be elaborated upon further to good effect if we now consider how Aquinas might respond to three of the most pressing challenges facing contemporary causalists. I will be most concerned to augment the above account precisely when and where Aquinas has the resources that appear to contain at least the germ of a response to these challenges. One of the most familiar challenges causalists face is the problem of ‘deviant’ causal chains. Consider Davidson’s example of a climber in trouble on the slopes: Imagine that a climber’s desire to let go of a rope combined with her belief that this is the right course of action together produce the physiological result that her palms becoming sweaty. The resulting slickness

Stephen Boulter

269

of the skin leads to the rope slipping out of her hands. It seems natural to say in this case that the belief/desire pair caused the slipping of the rope from her hand; but we would not want to say that this was an action despite the fact that it was caused by the relevant belief/desire pair since the slipping of the rope from her hands did not proceed from her pro-attitudes in the appropriate manner. Two problems emerge from cases of deviancy like this. First, the causalist’s definition of action thus appears to be too broad, and some means of narrowing the definition while respecting its core intuition is needed. A further complication arising from cases of causal waywardness, at least for Davidson, is to explain how pro-attitudes which render a behaviour intelligible and causally bring that behaviour about can still fail to explain the behaviour ‘in the right way’. The solution to both problems would appear to lie in spelling out precisely the way in which pro-attitudes must cause bodily movements so as to render them intelligible; but stating the challenge has been far easier than meeting it. It is here that Aquinas appears to offer some help. His elaborate account of the psychological antecedents of action allows one to spell out precisely what the ‘right way’ is, at least in the paradigm case of properly human actions. The problem with the case of the climber is that it is very difficult to present it in such a way as to make it fit naturally into the stages of a paradigmatic human action. At the very least crucially important steps between the formation of the relevant pro-attitudes and the ensuing behaviour are entirely absent. And this, I suggest, is enough to rule it out as an action despite the fact that the agent’s beliefs and desires did cause the rope to slip from the climber’s hand. The account of the psychological antecedents also spells out precisely how pro-attitudes must cause behaviour in order to explain the behaviour in the appropriate manner. This failure of fit is painfully obvious if one tries to reconstruct the climber example using Aquinas’ model as a guide. Following Aquinas one begins by assuming that the climber is psychologically normal and healthy10 and so has a standing desire for self-preservation, and an ability to appreciate self-preservation as a good. Now, having appreciated that she is in a perilous situation which is exacerbated by the fact that she is holding a rope at the other end of which is another climber she quickly forms the intention seriously to consider letting go of the rope along with any other possible course of action which might ameliorate her situation.11 Assuming there are other options before her, she then deliberates (perhaps very quickly) about the feasibility of the various courses of action, and then reaches a judgment that indeed there are various options open to her. She then consents to the appropriateness of the various possibilities. Let us then say that of the various options she elects to let go of the rope. Assuming that this reconstruction is not already wildly implausible, at this point we could say, in the cruder terminology of the belief/desire theory of motivation, that the climber has the belief/desire pair that, if followed by the appropriate steps, could lead to her genuinely performing the

270

Aquinas on Action and Action Explanation

action of letting go of the rope. But notice that these steps are not satisfied in our example. The climber would have to form the specific intention to let go of the rope; issue a command to herself to let go of the rope; and then exercise her physical capacities in letting go of the rope, none of which occur in our hypothetical case. The rope slips from her hands on account of the slickness of her palms. The slipping of the rope from her hands was indeed caused by her pro-attitudes, but not in such a way as to constitute a properly human action. It might be thought that all of these extra steps that Aquinas spells out could themselves lead to the climber getting nervous, thereby causing the rope to slip out of her hands in a casually deviant manner. Is there anything to suggest that Aquinas’ psychological antecedents rule out the possibility of causal waywardness creeping back in? I think the crucial step in Aquinas’ account of action which does rule out the possibility of deviancy is step (12) where the command issued to the self is then followed by an exercise of one’s physical capacities to carry out the command (usus). If a bodily movement does not arise as a result of the agent exercising her physical capacities in the carrying out of a command to the self, then (a) the bodily movement is not an action, and (b) the agent’s pro-attitudes do not explain the bodily movement despite the fact that they cause that movement and render it intelligible. One immediate response to this line of thought might be that it entirely misses a deeper point lurking behind the initial objection. It might be thought that Aquinas’ account of the psychological antecedents of properly human actions is perfectly in order, but an objection arises at a different level because the account relies on the content of different psychological states of the climber to play a causal role in the production of bodily movement. Here the issue is how the content of psychological states can be causally related to bodily movements. The problem arises because no one doubts that neurological events must figure in the explanation of an agent’s moving of her body. This raises a question regarding the relation between the content of mental states and physiological states of brains. If they are not identical one is faced with the problem of overdetermination if one continues to think that psychological states have a genuine causal role. But if they are identical, one is then inclined to ask whether the mind/brain states cause behaviour qua brain state or qua mental state, with the threat of epiphenomenalism looming in the background. The point for present purposes is that if one is inclined to believe that brain states qua brain states are the true cause of bodily movements then Aquinas’ elaborate account of the psychological antecedents of action is entirely beside the point. This problem faces anybody who wishes to include intentional states of agents within their accounts of action, not just Aquinas. But Aquinas would have thought that he had at least the form the solution to this problem has to take. The answer lies in his hylomorphism, in particular, the view that

Stephen Boulter

271

the mind, soul or psyche is the form of the body (Summa Theol. I, q. 76, a. 1, Commentary on the De Anima, Book II, Lect. 2). On this view the physiology of a brain state and its intentional content are two aspects of one and the same thing. The central idea is that content of a thought is not some extra epiphenomenal ingredient that accompanies a brain state but is encoded in either a sentence-like structure or a map-like structure (or both) in the very arrangement of the material components that make up the brain state.12 This is important because the brain state has the causal powers it does (and is the brain state that it is) in virtue of the arrangement of its physical components, which is to say, in virtue of its intentional content. Thus on this view the question as to what wears the causal trousers, the brain state or the intentional content, makes as much sense as asking what it is about a knife that allows it to cut, the fact that it is made of steal or the fact that it is sharp. Clearly it is both; but the steal and the sharpness are not two distinct things (which would lead to the problem of overdetermination) but rather two aspects of one item, the knife, aspects which are separable only in thought.13 There is much to be said both for and against Aquinas’ theory that the mind, soul or psyche is the form of the body, but these are issues into which we cannot enter here.14 It is enough for present purposes simply to point out how Aquinas might handle puzzles facing current forms of causalism. I want to finish then with another sketch of a solution to a further problem facing causalists, this time a problem emanating from metaethics. Jonathan Dancy has argued that the standard Humean belief/desire theory of motivation is incompatible with moral realism. Since Aquinas is a moral realist it is a challenge he would want to meet. The challenge can be stated as follows: It is assumed by moral realists that agents are able, at least on occasion, to perform morally acceptable actions for the right reasons, where ‘right’ is taken in a normative sense. These normative reasons, according to the realist, are to be found in states of affairs of the world. But most contemporary versions of causalism maintain that agents act on account of their beliefs and desires. But psychological states are not normative reasons, for what makes an action right in the normative sense is not the agent’s psychological state (cognitive or otherwise) but some salient state of affairs in the world. So if causalism is right it would appear that we cannot act for a normatively good reason, a result Dancy quite rightly finds intolerable. Aquinas would be in general sympathy with Dancy, but his version of causalism, combined with his theory of cognition, provides a number of lines of thought which can be brought to bear on this particular issue. In line with contemporary versions of causalism, Aquinas would insist that an agent’s action cannot be adequately explained without reference to the agent’s psychological states. But as we have seen, he departs from these versions by insisting that actions cannot be explained solely by reference

272 Aquinas on Action and Action Explanation

to internal principles. There can be no complete explanation of an agent’s action without some prior reference to an external ‘desired object’, that is, that object or state of affairs which the agent first apprehends as a good and subsequently desires. So Aquinas would quite happily maintain that any version of causalism which fails to include reference to external principles will be inadequate; but it is only these incomplete versions of causalism which lead to the difficulty Dancy has identified. For on Aquinas’ view the causes of an agent’s actions will include both internal ‘motivating’ reasons and normative reasons if the act’s external principle, its desired object, not only plays a role in bringing the action about (indirectly, by producing the internal motivating’ reasons) but also justifies it. And there is no reason to think that this happy coincidence of justifying and causing does not often arise. Thus on this line of thought there is no need for a causalist to fear that we cannot, at least on occasion, act for a normatively good reason. This is the first (simple) solution Aquinas might be tempted to offer in the face of Dancy’s challenge. But there is another line of thought which is also worth considering. The solution considered above assumes that both internal and external principles have a distinct and independent causal role to play in the production of action. But there is reason to believe that this distinction is not as hard and fast as one might think. This possibility is suggested in the first instance by Aquinas’ claim that the psychological states which follow upon the initial apprehension of the desired object have the causal power they do only in virtue of the causal efficacy of external desired object. Aquinas writes: [If] the moving principles are considered formally and specifically they are reducible to one, to the object of desire or appetite; for this is the absolute starting point of movement, inasmuch as, being itself unmoved, it initiates movement through the mind or the imagination. And because the secondary motive-principles only move in virtue of their share in the primary one, therefore they all as such partake of the nature of this primary one. And yet, though specifically one, they are numerically many’ (Commentary on the De Anima, Book III, Lect. XV) The point for present purposes is that psychological reasons have the causal powers they do only because they ‘partake of the nature’ of the ‘primary motive principle’, the external desired object. This implies that internal principles of action do not have an independent causal role to play in the production of action. In the context of Dancy’s challenge, Aquinas can thus be read as saying that, when push comes to shove, we really act as we do because of the external desired object. If Aquinas is read in this fashion then Dancy’s challenge is met by denying that the real causes of our actions are ever internal psychological states of agents; and since the real causes of our

Stephen Boulter

273

actions are external to us there is no reason in principle why motivating reasons should never also be normative reasons. The problem with this second line of thought is that is denies any independent causal role to internal psychological states of agents, a claim which does not sit well with Aquinas’ own account of the psychological antecedents of action. A final consideration is worth mentioning at this point. One might square this circle, and solve Dancy’s problem, by identifying external and internal principles in some fashion. And there is good reason to believe that Aquinas would be willing to consider this possibility. Aquinas states quite specifically that, while numerically many, internal and external principles are ‘specifically one’. And this line of thought fits well with Aquinas’ general theory of cognition, sometimes referred to as the ‘mind/world’ identity theory.15 According to this theory, again developed on Aristotelian lines, ‘something is cognized by someone in virtue of its somehow being within the one cognising’ (Disputed Questions on the Nature of Truth, 2.2c). More precisely, the intellect is able to be directed towards and understand particular things in the external world in virtue of its ability to take on the specific form of the object cognised. If the intellect understands a stone, say, it is because the form of the stone comes to reside in the intellect, resulting in the formal identity of cognising intellect and object cognised. To underline the seriousness with which this identity is taken it is worth noting that Aquinas writes: ‘[T]he intellect in act is said to be the thing understood in act’ (Summa Theol. I, q. 55, a. 1 ad. 2) and “[T]he actually understood object and the actually understanding subject are one being ... . So the understanding and the understood are one being’ (Commentary on the De Anima, Book III, Lect. IX). The interest of this theory of cognition for present purposes is that the distinction between normative and psychological reasons breaks down in the case of moral actions. If the salient features of a moral situation are such as to demand a particular course of action (thereby justifying it) and an agent apprehends those features and comes to form the desire to perform the action, the salient features of the moral situation and the agent’s psychological states are formally identical although numerically many. On this line Aquinas would be quite prepared to say that an agent can perform an action for the right, that is, normative, reasons despite the fact that the immediate cause of the action is the set of psychological states of the agent. The normative and psychological reasons are, in this instance, formally identical. Now while this means that one can act for a

274 Aquinas on Action and Action Explanation

normatively good reason, some might wonder whether this solution comes at the cost of sacrificing moral realism since normative reasons are now formally identical to psychological states. But I think this worry makes sense only when operating with a different mind-set, one in which psychological states are radically distinct from states of affairs in the world. In Aquinas’ scenario, this radical distinctness is precisely what is denied, so there is little to fear from identifying normative reasons with psychological states – the objectivity of the normative reasons is not compromised by the fact that those reasons can in some sense take up residence within the mind of an agent.

Notes 1. Many thanks to Constantine Sandis for very helpful comments on an earlier draft. 2. Aquinas writes: ‘Now the movement of every agent tends to something determinate, since it is not from any force that any action proceeds, but heating proceeds from heat, and cooling from cold’ (Summa Contra Gentiles, Book III, ch. 2) The Summa Theol. and Summa Contra Gentiles passages are found in The Basic Writings of Saint Thomas Aquinas. Volumes 1 and 2. Pegis (ed.), Indianapolis, Hackett Publishing, 1997. 3. This marks a significant difference between the ancient and medieval view of explanations from that of later periods. Scientific knowledge is strictly speaking only possible of what is necessary. Contingent facts are not amenable to scientific understanding in the strict sense. 4. Aquinas writes concerning the limitations of our knowledge that ‘The essential principles of things are hidden from us’ (Commentary on Aristotle’s De Anima, Translated by Foster and Humphries. Notre Dame, Indiana, Dumb Ox Books, 1994. Book I, Lect. 1); ‘We do not know even the essence of a fly’ (In Symbolo apostolorum I). 5. As Aquinas says, ‘[T]he nature of each thing is shown by its operation’ (Summa Theol. I, q. 76, a. 1). 6. Aquinas writes: ‘Now it cannot be said that appetite is a moving principle through sharing the specific nature of intellect, but rather e converse; for intellect only moves anything in virtue of appetition. It moves by means of the will, which is a sort of appetition ... . Appetition, on the other hand, can move to action independently of reason, as we see in the case of the concupiscible desire which is a sort of appetite’ (Commentary on the De Anima, Book III, Lect. XV). However, appetency is not always sufficient on its own to produce action. As Aristotle says, ‘not even appetition imperates such movement as this. For the self-restrained do not act according to their desires even while they are actually wanting and desiring; instead they obey reason’ (De Anima, 433a1–6). 7. Not all human actions are voluntary actions, but all human action are uncoersed. An action counts as voluntary for Aquinas if (a) the action stems from a principle within the agent, and (b) the agent is aware of all the relevant circumstances pertaining to the action, and (possibly) (c) the agent is not suffering from any mental disorder. See chapter 8 of my The Rediscovery of Common Sense Philosophy (Houndmills, Palgrave Macmillan) for further discussion.

Stephen Boulter

275

8. Aquinas writes: ‘it is reasonable to assert that both appetition and the practical reason are motive-principles; for the object desired certainly incites to action, and it is also what the practical reason first considers; so that the latter is said to impel to action because the starting point of its deliberations, the object of desire, does so ... . So it is clear that there is ultimately one mover, the object desired’. (Commentary on the De Anima, Book III, Lect. XV. Emphasis added). 9. Hume is often taken to suggest that there can be no rational critique of desires because these are logically independent of beliefs. Whether this is an accurate reading of Hume or not, it is not a position Aquinas would endorse. 10. Admittedly a rather dubious assumption in this case. 11. Here we encounter the first problem. If there are no other feasible options by means of which she could save her skin, she is left with either letting go of the rope or risking death by continuing to hold onto it. But it is far from clear that this is any option at all, particularly given the fact that humans are determined to desire and see as reasonable whatever is required for self-preservation. While she did have the option of holding on and then dieing, this is not an option Aristotle would consider to be a live one, since choosing death in such circumstances is beyond human endurance. 12. Aquinas is not entirely clear as to the nature of the structure of representational thought. At times he suggests that thought cannot take place without pictures, but he also suggests that the mind must be able to form mental propositions. See Robert Pasnau’s (1998) ‘Aquinas and the Content Fallacy’, Modern Schoolman 75: 293–314, for a discussion of this point. 13. At this point one is likely to be tempted to counter this line of thought with the following standard objection: Causal explanations are law-like. Reasonexplanations are not. So reason-explanations cannot be causal explanations, and so beliefs and desires and the like cannot be causes. I cannot enter here into the details of a Thomist response to this objection, but it would certainly be based on the view that nothing is explained in virtue of laws, so if a causal explanation is successful it is not in virtue of its being law-like. In short, Aquinas would question whether the alleged disparity between causal and reason explanations has been established. 14. See Kathleen Wilkes (‘Psuchi versus the Mind?’, in Essays on Aristotle’s De anima, Nussbaum and Rorty (eds), New York, Cambridge University Press, 1992) for an enthusiastic contemporary appraisal of this theory. Jonathan Barnes is less impressed (‘Metaphysics’, in The Cambridge Companion to Aristotle, Barnes (ed.), New York, Cambridge University Press, 1995). 15. See Haldane’s ‘Aquinas’ Mind/World Identity Theory and the Anti-Realist Challenge’ for an extended statement of this theory. It is interesting to note that contemporary thinkers have been tempted by similar ideas. See John McDowell’s Mind and World, Lect. 2, section 3 for an example.

Bibliography Commentary on Aristotle’s Nicomachian Ethics. Translated by Litzinger. Notre Dame, Indiana, Dumb Ox Books, 1993. Finnis, John Aquinas: Moral, Political and Legal Philosophy. Oxford, Oxford University Press, 1998.

15 Acting for Reasons – A Grass Root Approach Ralf Stoecker

There are many accounts of what it is to act for a reason. Yet, most of these accounts are committed to what might be called the standard theory of human agency. According to the standard theory actions are events that result from the agent’s having mental attitudes of a specific kind (e.g. a pair of beliefs and desires or a particular intention), which on the one hand cause the event and on the other hand show it to be reasonable from the perspective of the agent.1 To my mind this standard picture of agency is fundamentally misguided, and consequently the accounts of acting for a reason that build on it cannot be correct either. Therefore, although in this essay I shall not venture to show the misgivings of the standard approach, I feel justified to take up a strategy that may be worth pursuing in philosophy in any case, the strategy to go down to the roots and to figure out how far the advanced debates have a firm base. Hence, I shall present some exercises in what might be called grass root action theory, which will finally answer the question of what it is to act for reasons. The essay will be structured as follows: First I shall ask what it is to act at all (1). Then, since the answer will immediately connect agency with reasons, I shall question what reasons are (2). The resulting account of reasons will in turn lead me to ask what acting for reasons amounts to (3), and what kind of explanation reasons might provide for actions (4). Reason-explanations, it will turn out, work very differently from what the adherents of the standard approach think.

1 In order to understand what it is to act for reasons it is advisable to start with the question of what it is to act at all. The first step in answering this question is to emphasize the relational character of agency: (1) To say that a person acts is to claim that the person stands in a certain relation to something. 276

Ralf Stoecker 277

This necessary condition of agency is hardly momentous, but neither is it trivial. First of all, not every sentence expresses a relation. If one says, for example: (1)

Hanna is sick,

one does not relate Hanna to something at all. More importantly, even some typical action sentences don’t look as if they were ascribing a relation to a person, either. Take as an example: (2)

Dora is singing,

which is at least superficially similar to (1). Hence, if one wants to stick to claim (I), it will in the end be necessary to show that such supposed counterexamples are relational, too. For the moment, though, I shall follow most action theorists in taking (I) for granted. Sentence (I) is highly plausible, because we usually work on the assumption that if someone acts then his or her acting will make a difference to the world, which certainly entails a sort of relation (to ‘the world’). But what kind of relation is it that is entailed by agency? The claim that acting makes a difference to the world is in a sense metaphorical. What we mean is that some truths, some facts, are due to agents. Hence we get facts as the corresponding relata of the relation. (II) To say that a person acts is to claim that the person stands in a certain relation to a particular fact. Imagine a cook who fires the oven in order to bake a cake. To say that she acts by firing the oven is to relate her to the fact that the oven is getting warm. This claim is not uncontroversial any more. Usually agency is construed as a relation not between a person and a fact but between a person and an event, which is then taken to be the agent’s action. And although one might identify the fact that the oven is getting warm with the event of its getting warm, this event certainly isn’t the cook’s action. According to the standard theory, it is only a consequence, an effect, of the action. Yet, the standard view that actions are events is problematic in several respects. First, although it may be granted that usually when we act something happens, occasionally nothing happens at all, particularly when we omit to do something.2 Omissions form the first problem for the received view. Second, the received view is confronted with frequently discussed difficulties concerning the spatio-temporal location of actions; these, in turn, quickly lead into the dilemma between fine-grained and coursegrained methods of action individuation (neither of which may be happily

278

Acting for Reasons

embraced). Third, it does not even sound right to say that actions are events, since actions are things that we perform, and nobody performs events. Although there are numerous responses to these objections which attempt to defend the received view, I take the difficulties as a sufficient justification for pursuing my grass root attempt of finding a satisfactory account of acting for a reason without presuming that it is crucial for agency that the agent is related to an action which is an event. In the end it will turn out that we can do without either part of the received view: actions are not events, and play no important role for understanding agency at all. Given (II), according to which saying that a person acts is to claim that the person stands in a certain relation to a particular fact, the next thing to figure out is what kind of relation it is which holds between persons and facts that grounds agency. Here we are back on conventional ground. The common assumption that agency is supposed to make a difference to the world already gives us a clue about the special character of this relationship: it has to be in some sense explanatory. Since a fact (a ‘difference’) is said to be due to the agent, we should expect to learn something about why it obtains from looking at the agent. The agent bears an explanation for the fact. For example, in a case where a cook is firing an oven, it is worthwhile to turn to the cook if one wants to know why the oven is getting warm. (III) To say that a person acts is to claim that a particular fact can be explained with recourse to the person. It should be noted in passing that the idea that an explanatory relationship is essential for the attribution of agency fits very well with the claim that the second relatum of an agency ascription is a fact, since facts arguably are what is stated in true assertive sentences and therefore are adequate explananda for an explanation. Yet, although sentence (III) describes a necessary condition of agency, it is obviously not a sufficient one. A lot of facts are explainable with recourse to us, without constituting cases of agency. Take for example, (3)

In the course of her birth Dina caused her mother a hard time.

Although the pain of Dina’s mother is explainable with recourse to Dina, Dina did not act when she caused her mother pain. Sentences like (3) show that (III) is not sufficient to account for agency. It is interesting, though, that it is far from easy to find examples like (3), that is, sentences that fulfil the condition mentioned in (III) but obviously do not refer to an instance of agency. Take for example, Yesterday evening, Rita caused her mother a hard time.

Ralf Stoecker 279

Perhaps Rita fell down the stairs and broke her leg, causing her mother a lot of worries. Alternatively, Rita may have been fooling around all night, thereby making her mother angry. Only in the latter kind of case is agency involved. The example shows not just that not every instance of something’s being explainable with recourse to a person constitutes a case of agency, but also that many sentences which describe a fact (e.g. the fact that a certain mother had a hard time) as being explainable with recourse to a person, leave it open whether they refer to instances of agency. Sentences we use to ascribe agency to persons only rarely imply that there was an action. Talk of ‘action sentences’ is, in this respect, ambiguous. This label may be applied either to sentences which entail that what they refer to is an instance of agency, or to sentences which only in certain contexts refer to instances of agency de facto (and in other contexts do not).3 In any case, what all these sentences have in common, even if they do not necessarily refer to instances of agency, is that they describe a person as doing something. To say of a person, or of just about anything at all, that he, she or it does something is to state that a fact is explainable with recourse to him, to her or to it. When we say, for example, that hurricane Katrina destroyed New Orleans, we say that there was a fact (namely that New Orleans was destroyed) that could be accounted for by looking at Katrina. A common way of ascribing agency to a person then is to describe the person as doing something, leaving it to the context or leaving it open at all, whether what the person does is an instance of agency or not. But although we do not always make it explicit whether we are talking about a person as acting or as merely doing something, there is certainly a difference between merely doing something and acting, and it is almost unanimously agreed that the difference is to be found in the special way in which facts can be explained with recourse to the agent. Agency is a distinctive type of making a difference to the world. Hence, (IV) To say that a person acts is to claim that a particular fact can be explained with recourse to the person in a way specific only for agency. The crucial question is how to specify the right kind of explanation. A first answer might be as follows: (V) To say that a person acts is to claim that a particular fact can be explained with recourse to the person as being reasonable from the perspective of the person. This answer cannot be quite right, though, since there are cases where, from the perspective of the agent, it is not the fact itself which has its attraction but the agent’s producing that fact. For example, an athlete will be less

280 Acting for Reasons

interested in a javelin flying for more than 100 meters than in his being the one who throws it that far. Still, since (V) is prima facie plausible and widely accepted, I shall postpone this difficulty until Section 4. Given the widespread assumption that the explanatory backbone of agency is tied to the feature of being reasonable, one has to ask what may be meant by ‘being reasonable’. Evidently, being reasonable will have to be somehow connected with having reasons, and the straightforward account is as follows: (VI) To say that a person acts is to claim that a particular fact can be explained with recourse to the person by giving reasons the person has for it. This is a good first answer although there are obvious potential counterexamples. On the one hand we sometimes say that a person acted on no reason at all (e.g. if she just sings and dances in the rain), and on the other hand we talk about the reasons of mere doings (e.g. the reason why a hurricane suddenly changed its direction or why my brand new bicycle produces such a squealing noise). Despite these difficulties I suppose that (VI) is an extremely plausible starting point. The question is only what is meant when we say that a fact may be explained by giving the reasons the agent has. In order to answer this question we have to consider what reasons are in the first place. (Not forgetting that I aim to move along at grass root level.)

2 Reasons are basically something put forward or given. Someone gives a reason for or against doing something. The cook in the example may give several different reasons for firing the oven, for example: 1. 2. 3. 4. 5. 6.

It saves energy to preheat the oven. The cake will rise if the oven is already warmed up. It is my duty to make sure that the guests get cake at tea time. I like working in the kitchen. I don’t want to be told off by the boss again. I intend to bake a cake.

These are reasons for firing the oven because they speak in favour of doing so, in the sense they provide the cook with arguments for her acting. Hence we may conclude that reasons are basically arguments for doing something. As one can see from the example, reasons as arguments can take very different forms: • Sometimes reasons state (presumed) facts (e.g. that preheating the oven saves energy).

Ralf Stoecker 281

• Sometimes reasons allude to future states (e.g. that the cake will rise). • Sometimes reasons refer to values or norms (e.g. that it is the cook’s duty to make sure that the guests get cake at tea time). • Sometimes reasons express emotions, feelings and so on (e.g. her liking to work in the kitchen) • Finally (and philosophically most interestingly) reasons can also express intentional attitudes, like beliefs, desires, hopes, expectations and so on (e.g. the cook’s want not to be criticized by the boss and her intention to bake a cake). An immediate advantage of taking reasons as arguments is that it fits well with the idea that agency consist in an agent’s being explanatorily related to a fact. Arguments are of a linguistic form and therefore could easily be explanantia for the obtaining of facts. Still, if reasons basically are arguments, we would like to know how a person is to be related to an argument if we are to explain something by giving the reason the person has. The idea that reasons are arguments was based on the claim that reasons are something we put forward. Hence, an obvious answer to the question how an agent should be related to an argument in order to facilitate an explanation of what she does is that the agent must have uttered the argument (or expressed it otherwise):4 (VII) To say that a person acts is to claim that a particular fact can be explained with recourse to the person by giving reasons the person has for it, that is, by relating her to an argument uttered by her. This first proposal for amending (VI) has the advantage that it provides us with a straightforward idea of how stating the reasons someone has could provide an explanation: utterances are events and therefore can have a causal impact. Moreover we know instances of such a causal influence from our everyday experience of people who talk with themselves, who for example, obey their own explicit orders, or answer to questions they have put to themselves. So, in the example the cook might have given herself the heartening self advice: ‘The guests need a cake for teatime’ which then caused her to switch on the oven. Mentioning what the agent said to herself in such cases is a way of explaining something, which is due to her. Nonetheless, it is evident that (VII) cannot be correct as it stands, since agents only very rarely utter the arguments that speak in favour of what they do before they act (whether aloud or even silently in their heads). Cooks normally do not and need not order themselves to switch on the oven. So we must look for a better account of how the arguments should be related to the agent in order to account for the explanatory power of giving the reasons an agent has.

282

Acting for Reasons

Once again, it is quite obvious how we should proceed. It is true, one might say, that the cook need not have uttered ‘It is my duty to prepare a cake for tea time’ when she was firing the oven, but she obviously has to believe it. It must have been in her mind. So, according to this suggestion, an agent has a reason by having an argument in her mind. I take it that this is what we would usually say and, moreover, that it is true. It is true only as far as it goes, though, since strictly speaking it just gives a metaphorical reformulation of our question about the relationship between an agent and her reasons: Yes, she has to have the arguments in her mind, but ‘in’ is not a spatial relation and ‘mind’ is not a location. So, we are still left with the question of how to cash out these metaphors in a way which explains how being related to an argument could account for the obtaining of facts in the world. At this point, there are (at least) two strategies available. One is the conventional way of the standard theory the other one follows the grass root strategy. According to the standard theory the metaphorical character of the spatial characterizations should not be exaggerated. In order to account for what an agent does the arguments have to be at least somehow in the agent, they have to be represented in her, which in the end must mean: in her brain. This is the second proposal for amending (VI): (VIII) To say that a person acts is to claim that a particular fact can be explained with recourse to the person by giving reasons the person has for it, that is, by relating her to an argument represented in her brain. According to this view, agents have arguments in their minds in the sense of having them represented in their heads or brains. These ‘mental representations’ can cause the agent to behave in certain ways. Hence, we can explain their behaviour by referring to the essential feature of the mental representation, that is, the argument represented, which is also called its ‘content’ or ‘proposition’. And the way we refer to its content is by stating a sentence which expresses the same proposition, as the one represented in the agent.5 The cook believes that it is her duty to bake a cake, hence she has a mental representation in her head the content of which may be expressed by the English sentence: ‘It is my duty to bake a cake’. Referring to this belief provides us with an explanation of her firing the oven, because the respective mental representation causes the event of her doing so. All this is a very familiar part not only of today’s action theory but also of today’s philosophy of mind. It should be noted, though, that it is also at the bottom of some of the most vexing problems in these areas, for example, the problem of how, to put it in Daniel Dennett’s terms,6 a syntactical engine could ever be a semantical engine, or, more specifically, how mental structures, hidden in the brain, could have ever gained a semantics, although they do not have any pragmatics. Moreover, reducing the explanatory role

Ralf Stoecker 283

of reasons to special features of neuronal structures leads into the debate on the problems of causal exclusion and the prospects of non-reductive physicalism.7 Obviously, there is insufficient space to discuss all of these problems here. All I want to claim is that these difficulties threaten every attempt of reducing the explanatory value of reasons to the causal influence of mental representations and hence spoil the prima facie plausibility of (VIII). It is therefore worthwhile to look at a possible alternative to the mental representation approach to reasons for actions. As I have already noticed, the alternative third proposal follows the grass root strategy. It tries to keep close to the way we actually talk about these matters. Given the idea that reasons are arguments, and given that we may explain something by giving the reasons an agent has, the easiest way to localize these explanatory valuable arguments is in the explanation itself, or rather in what we utter when we give the explanation. If, for example, I say that the cook fired the oven because it was her duty to provide the guests with cake, it is me who states the argument, while the cook probably did not say a word before switching on the oven. According to this proposal reasonexplanations are based on the agent being related to an argument which is stated by the person who gives the explanation. (IX) To say that a person acts is to claim that a particular fact can be explained with recourse to the person by giving reasons the person has for it, that is, by relating her to an argument expressed in the explanation. This proposal has a couple of advantages. First, since reason-explanations usually state the arguments that speak in favour of what is done it has the advantage compared to (VII) that for every instance of reason explanation there is an argument to which one can relate the agent. Second, it can, unlike (VIII), accept the metaphorical character of the spatio-temporal localization of mental attitudes instead of assuming hidden entities or phenomena in the agent’s brain. Yet there seems to be a crucial shortcoming in so far as it may appear to be totally mysterious how reason-explanations could be explanatory at all. Let us assume that the cook we talk about was a monolingual citizen of Munich who died in the 1950s some time before I was born. Would it be not a sheer mystery to assume that a relation of this Bavarian woman to my utterance more than 50 years later, in quite another part of Germany, and in a language she did not understand, could explain that the oven in her kitchen was getting warm?! I take it that something like this idea is behind most attempts to find the reasons on which someone acts somewhere at, in or near to the agent. But, in fact it is not mysterious at all to ground an explanation on a wayward relation like the one between a monolingual Bavarian cook from the early 1950s to an utterance of an English sentence by a Prussian philosopher at

284

Acting for Reasons

the outset of the twenty-first century. Consider the following analogous example from a newspaper report. In 2005, English scientists discovered that, contrary to a received hypothesis, the tragic failure of a very early Mount Everest expedition which took place in 1924 could not have been due to insufficient clothes, since (as the scientists found out in a simulation of the expedition) the clothes which the mountaineers had worn in the 1920s were even warmer and more convenient than the modern synthetic outfits worn by today’s mountaineers.8 What is interesting in this story is not so much the refutation but the refuted hypothesis. According to this received hypothesis, the expedition failed and the participants died because their clothes were not as warm and as convenient as modern outfits. Although the hypothesis finally turned out to be wrong, it was a plausible story for a long time. What this story did, though, was to explain a historical fact (namely, the failure of an expedition) with recourse to a relation between something in history (the old clothes) and something of today (the modern suits), namely that the new ones were allegedly much better. And it is also easy to see how this explanation worked: the modern items were employed as a standard of comparison. The explanation was based on two assumptions, first that clothes which are less convenient than modern outfits would make it impossible for mountaineers to endure the strains of climbing Mount Everest, and second that the clothes in the 1920s were less convenient than the modern ones. From these premises it followed that the expedition had to founder. Two things can be learned from the analogy. First, we need not be worried by the assumption that reason-explanations may be based on a remote relation. Hence it is not absurd to claim with proposal (IX) that reasons are arguments we ascribe to an agent. The argument that it was the cook’s duty to fire the oven could very well be something which I put forward (and the cook doesn’t). The second lesson to learn from the Mount Everest analogy is that these remote explanations also work as standards of comparison as well. In order to validate these suggestions we have to inquire more closely into the question how those agent-specific explanations work, in which we relate people (who may no longer even exist) to arguments we put forward in order to explain facts as it were located in their vicinity.

3 The key to understanding how reason-explanations work is to be found in the first and obviously insufficient suggestion (VII) that the agent’s arguments gain their explanatory force from being uttered. The proposal was unsatisfactory because most instances of agency are not preceded by something the agent says. Yet, as I want to show in what follows, the proposal has an important kernel of truth: in order for the arguments to play an explanatory role in agency, at all, they must be somehow dependent not on actually being put

Ralf Stoecker 285

forward, but on the characteristic feature that they can be put forward by the agent. In other words, it is a good guess that reason-explanations work because agents can give reasons, because they can argue for what they do. Reasonexplanations work because reasons are tied to reasoning, and because reasoning in turn is basically conversational. Human beings can act (as opposed to simply doing something) because they have learned to talk about what to do and what to achieve and because they have learned to behave accordingly. They have acquired the skill to participate in what I want to call the practice of public practical deliberation.9 Public practical deliberation is the way we talk about what to do. This practice has a number of crucial features. First and foremost it is bound to have practical consequences, that is, usually it should result in someone doing something. A conversation about either going to the theatre or going to a pub is part of a practical deliberation only if it is to be expected that it will probably end up in someone’s going to the theatre or to the pub. (Though, needless to say, our deliberations do not always have results.) The second feature of public practical deliberation is that it does not merely tend to result in people doing something. When we talk about what to do we usually also talk about who has to do it. We are not just arguing about how nice it were if this or that would be the case but sooner or later we have to decide whose job it is to ensure that it will be the case. But how can people be ensured that something will be the case? Part of the answer is to be found in a third feature of public practical deliberation that is of utmost importance for our active life. Practical deliberation grows action trees. Following Alvin Goldman, I shall use the expression ‘action tree’ to describe the entangled pattern of action characterizations that may be combined with ‘by’.10 The cook satisfied her boss by baking the cake, she baked the cake by first firing the oven (and then doing other things), and she fired the oven by turning the knob at the oven’s panel. These different action characterizations form an action tree, provoking the question of how to describe the relation which holds between its branches. At this point the debate on action individuation mentioned above usually sets in: Do the descriptions located on the different branches of the action tree refer to different actions or are they merely descriptions of the same action (or is there a middle ground and part of the branches describe the same action, while others don’t)? But now we are in the position to understand the real importance of these action trees. They stand for our ability to control things that are beyond our proximate vicinity, beyond the reach of our bodies and limbs. While we are far from being Jedi Knights (having no telekinetic powers and being only able to learn to move our bodies in certain limited ways) the practice of public deliberation allows us to connect our bodily movements with distant effects. And since we know so much about the ways the world goes, we can direct our bodily movements very reliably to the goals we regard as worthwhile. This is the reason why we can

286

Acting for Reasons

base our explanations of remote facts on the explanation of proximate facts about the agent’s bodily behaviour, and this is what we allude to, when we say that someone has done something by doing something else, for example, that a cook baked a cake by (among other things) firing the oven which she in turn did by turning a knob at the oven’s panel.11 Taken so far, the emerging picture still seems to be the insufficient one of proposal (VII), though: Reasons are arguments put forward in a public practical deliberation, and they can explain those facts that result from a decision that a participant of the conversation should modify her bodily behaviour in a way to let this fact obtain. Children learn to act by learning to engage in such a deliberation and to adapt their bodily behaviour accordingly. Unfortunately though, we rarely talk before we act. What is missing is a further, fourth feature of public practical deliberation. Practical deliberation is not just a handy means in order to trigger convenient behaviour occasionally; it bears a strong normative constraint on us. Within certain boundaries we are supposed to take care of what is the case in a way that could stand the trial of practical deliberation. In short, we are responsible for what we do. Responsibility is not only etymologically connected with the duty to provide an answer. At the very basis of agency there is a demand to accord to a potential conversation about what to do, in the sense that we are expected to do what one should do according to such a conversation. Hence learning to act on reasons is more than learning to respond to a piece of practical talk, it is learning to behave as if one were continuously responding to such a public practical deliberation. This is an ambitious task, since there is obviously no way of accompanying all of our behaviour with extensive practical conversations. Hence as children we had to learn to behave as if such a deliberation had taken place without actually performing the deliberation. I shall not venture to engage in psychological speculations about how people master this task, but what seems clear to me is that the way we have learned it is first by learning to lead such a conversation with ourselves in soliloqui, second by establishing very complex routines that spare us the trouble of deliberating about what to do in standard cases (e.g. when as drivers we are confronted with a red traffic light or as cooks with a cold oven) and third by establishing a fine sense at which point we should better change from the second to the first skill, that is, when we had better start actually thinking again. Learning to participate in the practice of public practical deliberation means taking responsibility, but of course only for facts within a certain domain. We are not responsible for everything that happens in the world. We are not even responsible for everything we can influence. What we are responsible for is what we are expected to bring about. The boundaries of what we are supposed to give answers about are part of our practice of public practical deliberation that grounds agency. This is most obvious in cases of so-called negative actions where someone omits doing something or lets

Ralf Stoecker 287

something happen. In these cases we expect that the agent could give us arguments that support the decision not to intervene into the course of events. He is the agent of these acts, although he did not bend a finger, because it is his responsibility to take care that what happens in the respective domain is justifiable by a practical deliberation. From the point of view of the agent the practice of public practical deliberation bears normative constraints. From the point of view of other people it is of great prognostic value. It gives us a means of estimating how people in our surrounding will behave. Since we can work on the assumption that within their domain of responsibility they probably behave as if they had thought it over in a public practical deliberation we can easily carry out such a deliberation ourselves and see what it recommends that the agent should do. This is the basic idea for an account of the explanatory power of reasons. These reasons are arguments in a practical deliberation that we expect an agent to accord to. Although the agent need not perform the deliberation, as long as he acts at all, we expect him to behave as if he had performed it. Reason-explanations are based on the assumption that the agent fulfils this expectation. (X) To say that a person acts is to claim that a particular fact can be explained with recourse to the person’s ability to behave as if she were constantly engaged in a public practical deliberation about what to do. This is the basic idea, but it has to be modified, since it may seem that we won’t need reasons at all for our explanatory purposes. This will be the topic of the following section. I end this section by taking up two worries that were left dangling for some time. First there was the worry conjoined with proposal (V) that what may be reasonable from the perspective of the agent need not necessarily be some fact but may very well be the agent’s bringing about of the fact instead, for example, it need not be the fact that the javelin flies for 100 meters but it may be the athlete’s throwing of the javelin that far. If reason-explanations are based on the practice of public practical deliberation in the way described, these two different directions of being reasonable are easy to understand. Usually, when we talk about what to do, the arguments concern the state of affairs that we aim at, for example, whether it should be an evening in the theatre or in the pub, hence this is what has to be reasonable. Sometimes though what we primarily argue for is that it should be oneself who takes care of something. In the latter case what the reason-explanation is based on is not that a fact itself is reasonable from the agent’s perspective but that it is this agent’s being responsible for this fact which is reasonable from his or her perspective. The second worry came up with proposition (IX). Recall that the cook was a monolingual Bavarian who would never have dreamt of deliberating in

288

Acting for Reasons

English. So how can we still explain the fact that the oven was getting warm with recourse to the argument that it was her duty to prepare a cake for the guests? In order to meet this objection it is necessary to emphasize that according to the proposal reason-explanations are basically a social art. They rely on the practice of public deliberation and its normative character. We can explain what a person does because we can go on the assumption that, in principle, we could always call her to account for what she does. The virtual public practical deliberation according to which the agent behaves is a virtual conversation with us. Now, evidently we are not in the position actually to demand every agent to justify himself or herself, but as long as these agents grow up and live in a social setting similar to ours, we can rely on the assumption that their social mates would ensure their reasonable behaviour, that is, that they would guarantee that these agents behave as if they were constantly deliberating with us about what to do. The same goes for people from other linguistic, cultural or historical realms. We can still operate on the assumption that the very fact of a public practice of practical deliberation in their social life allows us to explain their behaviour as an instance of agency according to a virtual conversation with us. (Evidently, this assumption weakens and the explanations get worse in proportion to the cultural distance between us and them.) Hence, even if we know very well that the cook only spoke German we can assume that she behaved as if she had put forward the sentence, ‘It is my duty to bake a cake’ in an argument with us, since she behaved as if she had actually entertained the German sentence: ‘Es ist meine Pflicht, einen Kuchen zu backen’, and this sentence just means: it is my duty to bake a cake. Therefore, we may well relate the agent to an argument in English, without diminishing the explanatory value of the reason explanation. Relating an agent to an utterance of one’s own is explanatorily powerful because one’s own utterance serves as a standard of comparison (like the modern clothes in the Himalaya explanation) or more precisely as a sample of how the agent could have argued in her virtual deliberation on what to do in conversation with the explaining person.12

4 In the previous section I defended the claim that reason-explanations are based on the assumption that the agent fulfils the expectation to behave as if she has performed a practical deliberation in advance of her behaviour. But now one might worry whether given this assumption reason-explanations could still have a point. After all, it may now seem that all the explanatory power in attributing agency is confined to the information that the fact to be explained (e.g. that the cook’s oven is getting warm) is due to the agent’s disposition to act as if she would deliberate all the time. So, what is gained by actually mentioning some of the arguments the agent accords to? Why

Ralf Stoecker 289

should we ever mention parts of this virtual deliberation? What more can we learn from a reason-explanation that is not already implicit in the claim that the cook’s firing the oven was an act of her? Interestingly enough, there is not just one answer but there are a number of complementary answers to this question, which may be illustrated by the different reason-explanations in the cook example. First of all, the audience of the explanation might not dispose all of the relevant information to simulate the deliberation of the agent. The audience might, for example, not know that the woman in question was a cook and therefore had the duty to take care for the cakes of the guests. To such an ignorant audience we might explain the fact that the oven was getting warm by suggesting, on the one hand, that the oven was getting warm because the woman was firing the oven, that is, because it was part of an instance of agency (the cook behaved as if she had performed a practical deliberation), and moreover, on the other hand, by adding that one of the factors to be taken into account was that she was the cook and hence had a duty to the guests. But not only might the receiver of the reason-explanation be insufficiently informed, the agent herself might be so too. Although in instances of agency the agent is expected to act according to a practical deliberation, we cannot always assume that she could base this on perfect knowledge about the relevant circumstances. The cook was certainly wrong that preheating the oven would save energy (quite the opposite is true). Still, we can explain the fact that the oven was getting warm by presupposing that the cook acted as if she had performed a practical deliberation, yet we must now put in an idiosyncratic strain into the virtual deliberation she exemplified: she behaved as if she had deliberated extensively about what to do – but: instead of considering the argument that firing the oven would be a waste of energy (which is true), we take it that she would have argued that it saves energy. And we even have a special linguistic device for indicating that our explanation refers to such an idiosyncratic twist in her virtual deliberation. In order to mark that we either know that the agent is wrong in this respect or are at least unsure, we usually include a term of reservation, we say ‘She believes’ or ‘she thinks’, for example, ‘she believed that preheating the oven would save energy’. But an idiosyncratic strain in a deliberation need not be a mistake. We can even learn more about the behaviour of an omniscient by a reason-explanation than simply by being told that it acts – at least if it isn’t a pure intellect in Kant’s sense. So long as it is not such a ghostly existence it will have wants, preferences, needs, lusts, emotional bindings, moods and so on. At this point we encounter a fifth feature of public practical deliberation, the agent relativity of some of the arguments. With respect to some of the most influential arguments in practical conversations it is not even supposed that they should be brought to conformity. The cook might enter

290 Acting for Reasons

discussions over whether she will save energy or not, and if in the end we do not agree with her, we take this to be an unwelcome result of either her or our epistemic deficits. But we do not have such an expectation with regard to her liking of kitchen work or even her dislike for being criticized by her boss. Our practice of public practical deliberation has different levels of tolerance, so to speak. And since it is possible to start a perfectly good practical deliberation with the argument that it would be nice to do some work in the kitchen, as well as to start one with the argument that there isn’t anything worse than kitchen work, simply being told that a particular fact was due to an instance of agency is explanatorily weak. We understand much better, why this fact obtains, when we learn more about the different idiosyncratic features of a certain agent’s ability to act as if she had deliberated about what to do. And this, I take it, is the rational of the abundance of mental attitude ascriptions we are familiar with. We use them to express such idiosyncratic features. By ascribing them to a person we express characteristic tunes or twists in the virtual deliberations they accord to in their behaviour. Finally one must take into account that practical deliberations are not just flat logical operations on sentences. When we talk about what to do, we encounter that some arguments are more or less convincing. Logic can tell us that some of these relations between arguments are of necessity tight, but only a small and uninteresting part of practical deliberation moves on secure logical trails. With respect to other steps in our reasoning there is a much broader scope for idiosyncracy and hence again more need for explanation. Even if the cook had not been fond of working in the kitchen she might have fired the oven anyhow, and we can explain this, by saying, for example, that her desire not to be criticized by the boss was stronger than her dislike of kitchen work. The qualitative modifications of reasons, their respective strength or weakness, also mirror features of the public practice.

5 Let me sum up this speculative race through what I take to be the most basic elements of what it is to act for reasons. Agency is based on our public communicative practice of deliberating about what to do. Although we are only sporadically involved in such deliberative discourses we live under the normative constraint to ensure that within our domain of responsibility things are as one would expect them to be if the agent had deliberated about them. Luckily enough we have learned to meet this demand very reliably. This leads me to my final proposal: (XI) To say that a person acts is to claim that a particular fact can be explained with recourse (i) to the person’s ability to behave as if she were constantly engaged in a public practical deliberation about what to do, and moreover (ii) by giving some of the reasons for which she acts, that

Ralf Stoecker 291

is, by mentioning arguments the person could have put forward in the virtual deliberation that she is acting on. These reasons minimally make sure that what the agent does is an instance of agency, though usually the reasons we mention have an additional explanatory weight because they describe the idiosyncratic character of the agent’s ability, which sets it apart from the corresponding ability of other agents. To take up Kant’s term again: agency is based on treating ourselves as intellects; reason-explanations reveal where the spots of impurity are.

Notes I would like to thank Jens Kulenkampff, Constantine Sandis and Thomas Spitzley for their helpful remarks on earlier versions of the text. 1. What is usually considered to be the classical expression of the standard theory is Davidson 1963. For some reservations regarding the common reading of Davidson see Stoecker 1993. 2. Other cases where we arguably sometimes act without anything happening are: (unsuccessful) attempts, bodily acts like staying put, and mental acts like watching something. Of course, a possible (though in my eyes Procrustean) way out of these difficulties would be to doubt that these are instances of agency at all. 3. This is a lesson that can already be found in Davidson 1971, p. 44 ff. 4. The argument may, for example, have been written down on a sheet of paper, typed as SMS etc. Whatever I shall say about utterances will be true mutatis mutandis for these other kinds of expressions, too. 5. I here skip various differences and niceties of the debate (e.g. concerning indexical contents). 6. Cf. Dennett, 1981, p. 61. 7. Cf. the contributions in Heil and Mele 1993. 8. Cf. BBC News 13 June 2006, http://news.bbc.co.uk/2/hi/science/nature/ 5076634.stm. 9. I have dealt more extensively with the matters of this section in Stoecker 2001. 10. Cf. Goldman, 1970, Ch. 2. 11. For a more comprehensive account of this suggestion cf. Stoecker 2005. 12. The claims that reason explanations relate an agent to an utterance of the explainer and that these utterances work as a standard of comparison draw heavily on Davidson’s paratactic account of ‘that’-sentences and also on his thesis that we employ the content of propositional attitudes almost as measurements (cf. Davidson 1968, pp. 104 ff., and Davidson 1989, pp. 59 ff.). I have elaborated my approach in much more detail in Stoecker 2003a and 2003b.

References Davidson, Donald (1963), ‘Actions, Reasons, and Causes’, in: Davidson, Essays on Actions and Events, 2nd edn, Oxford: Clarendon, 2001. Davidson, Donald (1968), ‘On Saying That’, in: Davidson, Inquiries into Truth and Interpretation, 2nd edn, Oxford: Clarendon, 2001.

292

Acting for Reasons

Davidson, Donald (1971), ‘Agency’, in: Davidson, Essays on Actions and Events, 2nd edn, Oxford: Clarendon 2001. Davidson, Donald (1989), ‘What Is Present to the Mind?’, in: Davidson, Subjective, Intersubjective, Objective, Oxford: Clarendon, 2001. Dennett, Daniel (1987), ‘Three Kinds of Intentional Psychology’, in: Dennett, The Intentional Stance, Cambridge (Mass.), London: MIT Press. Goldman, Alvin (1970), A Theory of Human Action, Princeton (NJ): Princeton UP. J. Heil, A. Mele (eds) (1993), Mental Causation, Oxford: Clarendon. Stoecker, Ralf (1993), ‘Reasons, Actions, and Their Relationship’, in: R. Stoecker (ed.), Reflecting Davidson, Berlin, New York: De Gruyter. Stoecker, Ralf (2001), ‘Agents in Action’, Grazer Philosophische Studien, 61, 21–42. Stoecker, Ralf (2003a), ‘First Person Authority and the Merits of Minimal Monism’, in: A. Baechli, K. Petrus (eds), Monism, Frankfurt am Main, New York: Ontos. Stoecker, Ralf (2003b), ‘Climbers, Pigs and Wiggled Ears – The Problem of Waywardness in Action Theory’, in: S. Walter, D. Heckmann (eds), Physicalism and Mental Causation, Exeter, Charlottesville: Imprint Academic. Stoecker, Ralf (2005), ‘By “by” ’, in: João Sàágua (Hg.), “A Explicação da Interpretação Humana”, Lisbon.

Part II Agency and Moral Psychology

This page intentionally left blank

16 Sub-intentional Actions and the Over-mentalization of Agency Helen Steward

It is remarkably difficult to say anything that is both informative and accurate about what actions are. Though it is generally conceded on all hands that actions have something important to do with agents, it is very hard to say anything that is both correct and illuminating about what that important something is. The natural ideas that it might be said that they are doings of things by agents or that perhaps they are causings of events by agents, turn out to be insufficiently specific – for there are many doings and causings by agents which are not actions.1 I may be said to do something when I trip over, for example (I do something silly or careless, perhaps), but tripping over is not an action. And if I break a vase as I trip, I cause an event – the breaking of the vase – but my causing of this event is not an action. Some have also argued, more controversially perhaps, that there are actions which are not causings – Carl Ginet (1990), for instance, has suggested that saying the French word peu mentally to oneself is an action – but that there is nothing here which could count as the event which is effected by that causing.2 And if that were so,3 it would scupper entirely what one might have thought initially were good prospects for attempting to build on the causal approach by saying something more specific about the distinctive nature of the causation which is involved when agents act – such as, for instance, that an action of S’s is a causing of something by S such that that causing may be described as S’s w-ing, where S w-ed intentionally. For if Ginet is right, there are actions which simply do not fall into the category of causings at all. One approach to the characterization of actions which might seem to enable one to avoid any possible pitfalls that might be involved in the assumption that to act is always to cause something, focuses instead on the idea that what is crucial to an action is that it is a distinctive kind of effortful undertaking (which may, but need not necessarily involve the bringing about of an effect such as a bodily movement). Actions are characterized, on views of this type, as endeavours in which, when successful, one gets a body (or perhaps even a mind) initially at rest into motion (or its mental analogue), or perhaps keeps it in motion, or alters the nature of that motion, 295

296 Sub-intentional Actions and the Over-mentalization of Agency

by exerting oneself in a distinctive way – actions have been said, for example, to be strivings, or tryings4 – and it might be thought that mental actions of the kind Ginet mentions might involve an effort of the requisite sort, even if they are not causings. Perhaps if I am to succeed in mentally saying the word peu I must at least try mentally to say it, and perhaps it is this effortful character which means that my mentally saying it constitutes an action. The word might on other occasions pop up in my head unbidden, as it were – but in that case, its occurring there would not constitute action on my part because I had not tried mentally to say it. It might be argued, indeed, that Ginet’s own account of the essential features of actions rests, at bottom, on this idea about the effort which characterizes active processes. Ginet suggests that actions are either simple mental occurrences having what he calls ‘the actish phenomenal quality’ or else events which consist in the causing of further events by such simple mental occurrences.5 And although this account makes no explicit use of the concept of trying, and highlights the phenomenology of action rather than its striving nature, it might nevertheless be argued to belong in some ways with views of the striving/trying sort, because it is hard to see what the ‘actish phenomenal quality’ might be, if it is not something which has to do with the awareness of an effort. But approaches of this type have met with various objections, too. Some opponents have complained that to characterize actions as strivings or tryings in this way is to mistake a particular feature of a number of special sorts of action for the distinguishing feature of the whole class – it is only occasionally, they say, that we have to try or strive to do the things we do. Others feel worried that to think of actions as tryings must be to think of them as ‘inner’ – although, as I have argued elsewhere, it is not obvious, actually, why that should be thought to follow.6 And with respect to Ginet’s account, it has been suggested that it must be unsatisfactory to try to locate the distinctive feature of actions in their phenomenology. For one thing, it is not clear to everyone (it is certainly not clear to me) that actions have a distinctive phenomenology. Actions form a very broad and heterogeneous class – running is very different from talking, for example – and it is not easy to become convinced that there is anything phenomenological in common between them which might explain why we think of them both as active. Perhaps it is not surprising, therefore, in view of the difficulties that have been met with by accounts such as these, which seek to characterize actions intrinsically, that it is a more extrinsic approach that now finds itself at the heart of much contemporary philosophy of action. The distinctive character of actions is to be sought, according to this approach, in the sort of explanations which (it is said) they invite – explanations in which the agents’ reasons are cited. Where there is an action, there is something which is to be explained in terms of an agent’s purposes or goals or desires, together

Helen Steward

297

with his beliefs about how these purposes or goals are to be achieved, or the desires satisfied. Some philosophers assume in addition that these explanations are causal explanations – but I shall not be concerned with this debate here. What I want to focus on in this essay is the very idea that we can say something about the essential character of actions by saying something about the sort of explanations with which they are associated.7 I want to argue that it is false that it is a necessary condition of something’s being an action that it be associated with a reason-giving explanation. There are, of course, differing accounts of reasons in the philosophical literature – but I want to deny that the availability of a reason-giving explanation is essential to an action on any of the main accounts of what a reason is. Moreover, I shall deny also that it is even a necessary condition of something’s being an action that it have a psychological explanation, as that concept is traditionally understood. The argument I shall offer is based simply on the existence of counterexamples: there are many actions, I shall insist, that are not the doings of things for reasons, and moreover which have no roots at all in any of the related mental phenomena by means of which it might be hoped to bring them somehow within the fold of reasons perhaps more loosely construed – such as motives, desires, purposes or intentions.8 The counterexamples are not new – the existence of the phenomena I shall remark upon has been acknowledged and discussed many times before – but it seems to me that their importance has been enormously underestimated. This is mainly, I think, because it is not sufficiently widely accepted that the counterexamples are counterexamples. Some simply deny that the sorts of phenomena I shall describe are actions in the first place, preferring to reject the examples, rather than the theory with which they seem to conflict; whereas others insist, after engaging in various procrustean manoeuvrings, that these actions are after all associated with the sorts of explanations demanded by the standard model. But I shall try to argue that neither expedient is satisfactory. The counterexamples therefore stand; and we must look elsewhere for our account of the essential hallmarks of agency. In the second and third parts of the essay, I shall try to say something about the significance of the counterexamples for the philosophy of action. It is easier, I think, to see what sorts of accounts are falsified by their existence, than it is to supply an illuminating account which will serve properly to include them, but I shall try to say something to address both questions. On the negative side, I shall try to suggest that the existence of the counterexamples should alert us to a certain over-intellectualization – or perhaps it would be better to say overmentalization – which is involved in many dominant theories according to which actions are linked essentially to paradigmatically mental states like intentions and desires – and to a parallel over-mentalization of the concept of an agent. Encouraged by the correct thought that where actions occur, we agents play a very special role in the unfolding of reality, we have supposed that we must always play that role in the way we sometimes play it – by attempting to

298

Sub-intentional Actions and the Over-mentalization of Agency

bring it about that our purposes are served. But agency, I shall suggest, is a more basic phenomenon than teleological endeavour – it is an animal power that may be harnessed to serve an animal’s ends, and no doubt exists in some sense only because it may be so harnessed – but it is not the same power. And on the positive side, I want to try to suggest that there was, after all, an important insight in the idea that an action is an effortful undertaking. The notion of effort is not quite the right notion to use in order to express the insight, however, for it is itself associated with a certain mistake of over-mentalization. I shall therefore also try to say something about the notion with which I think it should be replaced.

1 Sub-intentional actions The counterexamples I want to consider in this section are usually termed ‘sub-intentional actions’. The term ‘sub-intentional action’ was coined originally, I believe, by Brian O’Shaughnessy (1980) in his two volume book The Will9 as a way of referring to a sub-class of our doings which seem not to be redescribable as w-ings such that any sentence of the form ‘S w-ed intentionally’ is true.10 During any reasonably extended period, for example, one will be moving around in various ways – slightly rearranging oneself, absent-mindedly scratching one’s head, fiddling with one’s jewellery, leaning a bit more this way or that, turning one’s head slightly, jiggling one’s foot. The resultant movements often occur below the level of our conscious notice, and when they do, it seems to be impossible to characterize the events by means of which they are produced as w-ings of any kind such that ‘S w-ed intentionally’ is true – not, at any rate, without doing quite considerable violence to the concept of intention. And certainly it does not seem as though we normally have reasons for doing the things we do sub-intentionally – at any rate, the special sense of the question ‘Why?’ which Anscombe noted surely has no application to our subintentional actions. Anscombe (1957) explicitly notes that the special sense of the question in which she is interested ‘is refused application by the answer “I was not aware I was doing that” ’, and if it is right to suppose that we are often not aware that we are engaged in the jigglings and twiddlings which constitute our sub-intentional actions, then, self-evidently, they are not associated with reason-giving explanations in Anscombe’s special sense.11 On most dominant conceptions of what an action is, therefore – someone’s doing something intentionally, for example, or someone’s doing something for a reason – it would seem that subintentional actions are not really actions at all. It might perhaps be retorted that there may be reasons for sub-intentional actions which are of a somewhat different type from those highlighted by Anscombe’s special ‘why?’ question – for instance, it might be said that my shifting around in my seat serves to relieve subliminal discomfort – and that

Helen Steward

299

this is the reason for it. But first, this is not at all the type of reason-giving explanation which most philosophers have had in mind when they associate actions with reasons, and so it does not seem as though this stratagem would in any sense save the standard view. Second, there would seem to be reasons of the same type (that is to say purposes of which I may or may not be aware) for all sorts of bodily processes that are not actions, for example blinking, so that this conception of there being a reason for something is not a promising means of distinguishing actions from other, non-actional processes. And third, it seems highly doubtful in any case that purposes of this type really are served by all sub-intentional actions – there seem to exist many subintentional actions which have no such underlying point. I conclude, therefore, that there is little hope of salvaging any useful version of the thought that sub-intentional actions are associated with things done for a reason by means of this expedient. I should like to agree with O’Shaughnessy, though, that despite the lack of any associated reason-giving explanations, at least some kinds of sub-intentional actions are actions. But what reasons might there be for insisting here, that the examples should be permitted to trump a rather wellestablished and popular theory about what actions essentially are? – rather than rejecting them on the basis of the well-established and popular theory? – or at any rate on the basis of the motivations for the well-established and popular theory? Hornsby, for example, considers very briefly in a footnote the suggestion that perhaps sub-intentional actions ought to be admitted to the class of actions proper on the grounds that they can be revelatory of a person’s psychological states and responds as follows: To this, the reply can only be that it is doubtful whether we can straightforwardly effect any cut-off between movements which do and movements which do not have a psychological history ... and that we have to draw boundaries where we can find significant boundaries to draw. If we take an interest in regarding ourselves as rational creatures much of whose behaviour receives a satisfying explanation when it is seen to issue from our beliefs and desires, interests and concerns, then the restriction of actions to events that can be characterized in such a way that we can see that something was intentionally done when they occurred will enable us to deal with a vast area that we take to matter.12 In other words, according to Hornsby, we have to decide where to draw the boundaries between movements whose production amounts to an instance of the phenomenon we call action and movements whose production does not – and the dividing line between movements whose production issues from ‘our beliefs and desires, interests and concerns’ and those whose production does not issue from such things is as good a dividing line as any on which to rely. But that dividing line excludes sub-intentional actions from the class.

300 Sub-intentional Actions and the Over-mentalization of Agency

Some of what Hornsby says here is undoubtedly true. I am sure that it is true, for instance, that there is no clear dividing line between movements which do and movements which do not have a psychological history; and it is also true that it is up to us to delineate concepts which draw boundaries in interesting and significant places. But the case for the inclusion of subintentional actions within the category of actions need not rest on the idea that they are revelatory of an agent’s psychology.13 Indeed, it seems to me highly doubtful that more than a small fraction of sub-intentional actions would meet this criterion in any case – there may be a few jitterings and jigglings which can be ascribed to nervousness, for instance – but it seems implausible that the vast generality of the class of sub-intentional actions have any convincing psychological explanation. Rather, the best case for including sub-intentional actions within the class of actions proper, it seems to me, rests on the naturalness with which we ascribe the production of the resulting movements to ourselves. When I fiddle with my jewellery, it seems to me, it is me who is fiddling with it, even if I am not aware that I am doing so. And though it must be admitted that no great weight can be placed merely on the fact that this seems to be a natural thing to say, since, as I have already noted, I can sometimes be said to have done things even in cases where my doings of those things certainly did not count as actions of mine (as, e.g. when I trip over – it was me that did the tripping!). I seem to be involved in the production of my fiddling in a much more intimate way than in the tripping – indeed, I do not seem really to be involved in the production of the tripping at all. I am active in the fiddling, though not in the tripping.14 The fiddling seems to be something which is under my control, and I seem to control it in very much the same way that I control many of the processes which constitute my intentional actions (although in the sub-intentional case, the control is not exercised in the service of an end). Bodily actions, one might say, occur when I make my body move in a certain distinctive direct kind of way – a way in which I am normally able to make it move in virtue of my possession of the capacity for bodily control. Often, when I make my body move in this distinctive direct kind of way, I do so knowingly and intentionally in order to carry out some plan or purpose that I have. But the knowingness and the intentionality, I should like to say, are not themselves the things which make it correct to say that it is me who is moving my body in these cases – for it seems to me correct to say that I am moving my body in the distinctive way, even when this knowingness and intentionality are absent. When I fiddle with my jewellery or jiggle my foot absent-mindedly from side to side as I type, or rub my chin while in thought, it seems to me right to say that I am moving my body. I am involved and in a sense clearly present in these processes in a way in which I am clearly not involved in my tripping, or in my digestive processes, or in the reflex that occurs when my knee is hit with a hammer, say. It seems to me that I am producing these movements in the distinctive direct

Helen Steward

301

kind of way characteristic of the phenomenon of bodily action, even when I fail to notice that I am doing so and even when I am not producing them intentionally, when I am surprised, for instance, to have it pointed out that I am producing them. That is why they must count as my actions. Perhaps it will be said that it is not me, but rather some sub-personal process going on inside me that must be producing and controlling the fiddling – since there is no connection between the fiddling and any of my conscious purposes and plans, and need not even be any awareness, on my part, of the fiddling. But this seems to me to be quite wrong. Suppose, for instance, that I suddenly become aware of one of these sub-intentional actions – perhaps because I am thinking about the phenomenology of such actions, so that I can say something about it in the essay I am writing, and so decide to see whether there are any going on right now. I might realize, for instance, that I have been twiddling with a piece of my hair for some time. And I might now continue to twiddle with it for a bit, this time attending closely to what I am doing and continuing the process intentionally because I am suddenly interested in it. Something has changed – I am now attending to what I am doing, and I am now acting intentionally. But it seems to me most unnatural to say that the process which was going on while I was twiddling my hair before I focused attention on it was not a process which I was orchestrating and only became one which I was orchestrating once I had focused attention upon it and became possessed of a reason to continue with it – that I was not acting, initially, but then began to act once I began to intend that the process should continue. On the contrary, the natural thing seems to me to be to say that I was orchestrating the process all along, that I was acting from the start – and am now simply newly focused on an activity in which I have been engaged for some time, and have a new reason, which I lacked before, for continuing with it. Reflection on the phenomenology of intentional action may help to make it seem less peculiar than it may seem at first to suggest that sub-intentional actions are manifestations of the very same active powers which come to be utilized for our own ends in situations in which we act intentionally. For all intentional bodily agency involves the interweaving of conscious systems of bodily control with more basic, effectively automated or non-intentional systems. When I type, for instance, although I may decide consciously which words I shall use, I do not need to engage in any conscious supervision of my fingers – they just get on with the job by themselves, as it were, now that I have learned to type. An enormous amount of the control of movement in which we are engaged as agents is delegated, inevitably, to processes which are very ill-described as the causing of motions by mental states or events such as desires, choosings, intendings and the like. But this does not imply that what is controlled by those processes to which the agent delegates is not at the same time controlled by the agent. What the agent normally does in the case of intentional action, it would seem, is to organize

302

Sub-intentional Actions and the Over-mentalization of Agency

and harness a variety of lower-level bodily powers, skills and so on, ensuring that they are smoothly exercised co-operatively in order to achieve certain ends. What I am suggesting is merely that it may be the exercise of those lower-level bodily powers rather than the organizing and harnessing for ends that is really the mark of agency. What is truly essential to the power of action is something which, as it were, lies beneath the capacity of an animal intentionally to bring about certain movements of its body, rather than being simply constituted by that capacity. What can it be, though, it might be asked, which makes it seem so right to suppose that a sub-intentional action bears the intimate relation to the self, the agent, which is the mark of action generally? One might think initially that the key to explaining why it seems so plausible to regard such sub-intentional actions as actions, even though they are not done intentionally, or even knowingly, might lie in the fact that in these sorts of cases I at least have the capacity to prevent altogether, stop in its tracks, reverse, alter, change the direction and speed of, or otherwise affect the motion in question, whether or not I actually choose to exercise it in a given case. This, one might think, will certainly serve to distinguish my fiddlings and jigglings from such things as digestive processes and reflexes, over which I have no direct control at all. But in fact, I do not think this initially attractive suggestion can, in the end, be made to work, just as it stands. For there seem to be processes which I have the capacity to affect, but which nevertheless do not, under normal circumstances, count as exercises of bodily control on my part. Consider processes such as breathing or blinking, for example. I can certainly deliberately affect the processes of breathing or blinking if I choose to do so – deciding, for example, to take particularly deep breaths, or to blink every five seconds. But it does not seem to me to be true that breathing and blinking when they are carried out in the normal, automatic way are actional processes at all. Breathing is not normally something I make happen or control, even though I can control it for a while, if I choose to do so. But if I am not active when I breathe, it cannot be the capacity to affect and to some extent control a process which accounts for the agential character of sub-intentional actions. It may be a necessary, but it cannot be a sufficient condition for an actional process. What is the difference, then, between fiddling and breathing, in virtue of which I have suggested fiddling should normally counts as acting and breathing normally should not? O’Shaughnessy (1980) at one point makes the suggestion, which it might be thought we could utilize here, that we should look to the concept of desire to explain why sub-intentional actions ought to be accounted actions; and certainly, if it were true that desire played a role in the causation or explanation of all actions, including the sub-intentional, it would enable one to make the wanted separation here between sub-intentional actions and automated yet controllable processes

Helen Steward

303

like breathing. This stratagem would also offer the prospect of retaining a weak version of the idea that wherever there is an action, there is something that was done for a reason – for it might be argued that desires just are (a kind of) reason.15 However, it seems to me unappealing to insist that all sub-intentional actions are the products of desires, or desire-like phenomena, such as urges. It just does not seem to me very plausible that all the small unintentional movements of which I am the agent are things which in any sense are the products of desires, urges or like phenomena – though of course, some may be. I do not want to be fiddling with my jewellery; indeed, I may positively not want to be doing so – I may be aware that it is a rather bad habit of mine, which others find irritating. We might try saying, I suppose, that the desire in question is unconscious, or speak of subliminal urges – but there would seem to be no real justification for this stipulation, other than the hope of preserving the thought that it is distinctive of actions that they have psychological causes of a desire-like sort. And the question is, why must we suppose this? Why must we be able to trace the production of a movement to a mental or psychic event or property if it is to be an action? There is an answer to this question which, I take it, goes something like this: unless there is some reason to suppose that a movement is in some sense the product of something mental, there can be no reason to think it should be associated in any special way with the self, with the agent. In the absence of any mental causation, its connection with the agent could only be the boring one that it is produced by processes which occur in that agent’s body. Unless my mind is somehow involved, the thought goes, I could not be involved either. In the following section of the essay, I want to examine and challenge this line of thinking, which I shall argue is rooted in two deep prejudices of which I think we will have to try to rid ourselves if we are ever going to understand actions properly – one, a Cartesian prejudice about the nature of the self, the other a prejudice about how to conceptualize causation.

2

Mentalistic conceptions of action and their source

The concept of the self is apt sometimes to make philosophers feel uncomfortable, particularly in the context of philosophy of action, where the idea that the self might be able to produce bodily movements is apt to prompt ghost-in-the-machine type imagery and related unease. But let me be clear that in speaking of the self or the agent, I do not mean to speak of a ghost in a machine. I mean merely to speak of an entity such as myself – and that is to say, a human being, an animal of a certain kind, an essentially embodied entity. And I should like to insist that the idea that an animal might be able to produce a bodily movement, so far from being a strange piece of metaphysical lunacy seems to be part and parcel of an everyday

304 Sub-intentional Actions and the Over-mentalization of Agency

picture of the world with which we are very comfortable. It is not at all obvious that there must be something deeply wrong with it. Animals have many powers – what is so strange about the idea that one of the types of powers of which they are possessed is the power to control in certain respects movements (and other changes) in their own bodies? It is true that a certain form of dualism pervades this commonsensical picture. But it is not dualism of the Cartesian kind. We could call it animalbody dualism; the two distinct things which it recognizes are not the animal’s mind and the animal’s body, but rather the animal and its body. We think and speak of animals – especially human ones – as possessed of their bodies, and to a certain extent, as controllers of them. No doubt such animal-body dualism, were it to be transformed from a folk psychological structure into a metaphysical position would be a controversial view, and it would need defending against reductionist onslaughts of various kinds.16 But my present point is merely that animal-body dualism and mind-body dualism are different kinds of dualism – and so that in speaking of an animal’s production and control of its own movements, we need not be speaking of an animal mind’s production and control of those same movements (whatever sense we think we can give to that idea of an animal mind). For an animal is not identical with its mind – how could it be, indeed, when an animal is a substance and a mind is not? It is no doubt true that animal-body dualism only appeals in the first place as a way of thinking about certain biological systems because it appears necessary to suppose that those biological systems have mental lives. We do not think that tables or chairs have bodies, nor even that trees do – and if we ask why that is so, the answer must presumably be that there is not that complexity in the behaviour of a table, chair or tree that demands we take what Dennett has called the intentional stance towards it. As we ascend the scale of biological complexity, however, a whole host of related ideas begin to press themselves upon us as requisite for the proper understanding of animal activity. We become increasingly inclined, for example, to imagine the creatures we encounter as we ascend this hierarchy as possessed of mental lives, of rudimentary forms of consciousness, of points of view from which the world can be assessed; we start to wonder whether there is something it is like to be them. We attribute to them such things as simple plans and designs. And as it becomes more plausible to attribute such things as consciousness, such things as plans and intentions to a creature, so it becomes more plausible to think of that creature not merely as identical with, but rather as possessed of its body. For a creature which it makes sense to regard as having some kind of mind, it starts to make sense, too, to regard as something whose body is at its disposal – a tool for the execution of its plans, a bit of the world which can be got to do its bidding.

Helen Steward

305

The idea of body possession and the idea of mindedness, then, are certainly closely connected with one another; it is only of systems that we think of as possessing a capacity to represent the world and to respond by acting to those representations, that we accord the special privilege of body possession, and therefore conceptualize in accordance with the dualistic agent/body scheme. But it does not follow from this that each individual exercise of power by an animal must be an exercise of the power of control over its body by something we are inclined to think of as its mind. That there is a general connection between the attribution of mindedness and the attribution of the power of agency does not imply that each manifestation of that power must imply that causation which is in some sense mental must be always involved whenever the agent acts. No doubt it is normally true that exercises of agency involve such things as intentions, motives, purposes and desires; there is not much point, after all, in having the power to move and control a body unless one is able to do so in the service of some of one’s ends. But must it always be true? Why might it not be that sometimes this power, a power possessed by all animals which have attained a certain degree of complexity, is exercised for no particular end? Why might it not just be exercised, on occasion, for no reason and out of no desire? Surely we are not creatures so totally governed by rationality and teleology that it is straightforwardly inconceivable that we might exercise a power we have without any purposive intent at all? What stops us from being easily able to accept this thought, I think, is that the philosophy of causation which modern philosophy has mostly come to embrace actually has no proper place for the idea, of which I have just made use, of an exercise of a power by a substance.17 An exercise of power has to be conceptualized, on the dominant Humean view of causation, according to which the only true relata of the causal relation are such things as events and perhaps also states, as itself an event, and therefore as something which must itself be caused by prior events and states. The exercise cannot just be thought of, unreduced, as just that, an exercise of a power by a substance, for the notion of an exercise of power by a substance is deemed unintelligible, unless is can be shown to be equivalent to the notion of an event occurring within that substance producing an effect. Our inclination to suppose that we sometimes make things happen, therefore, gets reinterpreted in the terms of this conception of causation – we may say that we sometimes cause movements of our own bodies, if we like to say so, but we must recognize that in saying this, we are really talking about certain movements in our bodies being made to happen by such events as our choices and decisions, or perhaps ultimately by such states as our beliefs and desires and intentions. That is what the causal role of the agent has to consist in because ultimately, to talk of a substance exercising a power to produce an effect is unintelligible. Truly irreducible substance causation is not permitted in the ontological scheme of most modern metaphysics.

306

Sub-intentional Actions and the Over-mentalization of Agency

It will be allowed, of course, that we humans (and perhaps other animals) are different in many respects from other substances which have powers, and that these differences may even legitimate talk of the exercise of powers, to some extent, in our case. Paper, for example, has the power to burn – but paper cannot choose whether or not to burn; it cannot decide to burn; it cannot burn if and only if it wants to burn more than it wants to do anything else. But where we have the power to bring about events of certain kinds, whether or not they will occur depends on such things as our choices, our decisions, our wishes. We can exercise our powers, then, it might be said, because we can choose whether or not the events which we have the power to produce will occur or not. I can choose, for example, whether to raise my arm or not – and because I have that power of choice, I am not like paper. This account is likely initially to raise worries about what the causes of the choice itself might be – but those worries are supposed to be allayed when it is said that the choices are caused, when things are as we want them to be, by our reasons (which are in turn to be thought of as such things as beliefs and desires). But this is of course a way of dealing with the distinction between our actions and the actualizations of the powers of paper mentalistically. It is to reduce the concept of an exercise of power by S to the concept of a causal relation between causes which are thought of as events and states – such as beliefs, desires, intentions, choices and so on and an event-like effect, a movement of some sort. And it is because of the mentalistic nature of these causes, it will be said, that we are inclined to suppose that we are involved in what goes on when an action occurs. What I think is so valuable, though, about reflection on the existence of sub-intentional actions is that it reveals that this answer must be wrong. We do not choose to act when we act sub-intentionally; we do not decide to do so; we do not intend or want or wish to do so. We do not act intentionally, and there is no reason-giving explanation for what we have done. We just act. There need be no mental cause of the resultant movement at all – unless of course one wishes to say, as perhaps it would be perfectly proper to say, that the mental cause of that movement is myself – an animal with mental properties. The power of action might indeed still be accounted a mental power – for we might, if we wished, choose to associate the concept of mentality with the concept of the self, rather than with the standard range of states and events by reference to which what is mental is typically defined which one finds strewn across the literature in philosophy of mind. But the point is that we seem to need in order to account for these actions, the concept of an exercise of power by an animal which does not reduce down to the concept of causation of a movement or change in that animal by one or more of that animal’s mental states or events. And that, I suggest, is a powerful and radical realization to which the philosophy of action ought to respond.

Helen Steward

307

3 Actions as exercises of the two-way power of bodily control But how might it respond? If we cannot say anything about what actions essentially are by reference to the sorts of explanations they introduce, or by appealing to the idea that when an action occurs, a bodily movement is made to happen by some kind of mental cause, what on earth is there left to say? What are actions? It is at this point that I think we need to reach for the insight that is present in views which take note of the striving character of action. Let us return to the phenomenon of breathing. Breathing is not normally a form of action. Why not? Not because I am not breathing for a reason; not because I am not breathing intentionally; not because my breathing has no mental cause – for all these things may be said also of many actions – namely, the sub-intentional ones. But what about the idea that there is no trying, no striving in normal breathing? There is something about this idea that seems quite plausible. Breathing, one might say, is something I do, but it is not something I (normally) have to do.18 I do not (in normal circumstances) have to try or strive to do it. Could that be the reason it does not count as a form of agency? As it stands, this cannot quite be the right answer. I do not try or strive to fiddle with my jewellery – indeed, I may be trying quite hard not to do so – and yet when I do so sub-intentionally, I act. So trying or striving as such cannot be the mark of agency. Nevertheless, I think there is a notion in the vicinity of trying which may serve our purposes. The trouble with trying is that it would seem that one cannot really try to do something without knowing one is trying to do something. One can, I suppose, try to w without knowing one is trying to w – for example, I may be trying to break up my former boyfriend’s relationship, though I am unaware that that is what I am trying to do when I keep attempting to persuade his girlfriend to go back to live in the US where I know he will not want to follow her. But I can hardly be trying to do this if I am unaware even that I am attempting to communicate with his girlfriend. There surely has to be something that I know (or at least believe) I am doing, or trying to do, if I really am trying. And so the notion of trying seems unsuitable for capturing what it is that seems to make our sub-intentional actions actions. For I may be engaged in sub-intentional activity without being aware at the time that any relevant activity is going on at all. Like other notions that we have discussed, the notion of trying is too mentalistic to constitute the basic characteristic which all actions share. But there is something present, I think, which at any rate resembles trying, or striving, when we act sub-intentionally. What is present is an exercise of the power we have to move or change our own bodies, and an exercise of a power is a bit like an attempt, in that it is up to the agent (though not necessarily, of course, up to the mentalistically construed agent)

308 Sub-intentional Actions and the Over-mentalization of Agency

whether or not it occurs. That power is exercised and not merely actualized for it does not have to be exercised at the time at which it is in fact exercised, in the way that the power of paper to burn has to be actualized when a flame is present under the right conditions. The power in question is what Reid called an active or two-way power, the power to do something or not, a type of power which is never possessed by any non-minded agent. And it is because of the presence of this factor, I believe, that we are inclined to ascribe these actions to ourselves, which makes us feel that they are, in the requisite way, our doings, even though they are not the products of our intentions. What an action is, then, it might be said, according to this view, is an exercise of bodily control on the part of its agent – the bringing about of a bodily movement or change by that agent by means of the exercise of its two-way power to do so. I see no reason to suppose that this account will not serve also to encompass mental actions such as mentally saying the word ‘peu’, for it would seem that when I do so, I exercise a certain power over the part of my body which is my brain, though I have less insight, in this case, into the physical nature of the change that I have effected. And the point on which I should like to insist is that the exercise of such bodily control by the agent need not involve specifically mental causation at all. The picture I would like to endorse attempts to take more seriously the thought that the agent who is doing the controlling is a bodily, as well, of course, as a mental being, equipped in fact with a multitude of means by which to effect its own movements, as well as certain sorts of changes within itself, not all of which are equally deeply connected to that agent’s conscious life. For an animal to have control over its body is merely for it to be able, in the actual context in which it finds itself, both to bring about some particular movement of its body, and to be able not to bring it about. But nothing is said or implied by this conception of bodily control specifically about any antecedent thinkings, wishings, plannings or the like. There is nothing as yet in this picture of what control involves to imply that something which genuinely counts as the animal’s controlling things cannot take place by means of bodily systems which do not involve any antecedent role in the voluntary causation of bodily movement for mental states, as those things are usually conceived. The animal body, on this conception, is then not merely the instructed instrument of that animal’s will. On the contrary, an important sub-set of the complex set of embodied systems which enliven it become constitutive themselves of the phenomenon of willing. The agent that controls the body is not to be conceived of as a pure will – not even a physical version of one which can be located, roughly, in the physical brain. It is to be conceived of as a whole, functioning animal whose systems of agent control are various, and only some of which involve the paradigmatically mental phenomena often said to be essential to the causation of action.

Helen Steward

309

It might, perhaps with some justice be said that it is not very illuminating to say that an action is an exercise of a two-way power of bodily control by an animal since even if it is right to say so, the idea of an exercise of power is connected via a very small circle to the concept of action, and so that we have not really got anywhere. For what, it might be asked, is an exercise of the power of bodily control by an animal? Does not that word ‘exercise’ just contain all the mystery that was originally contained in the concept of action? In some ways, it seems to me that this accusation is perfectly fair – I think it is true that the concept of an exercise of a two-way power is not much of an analytical advance on the concept of an action. But I want to end by defending myself against the allegation that to say that an action is an exercise of a two-way power to move or not to move, or change or not to change, some part or aspect of one’s body is really to say nothing that is of any use to the philosophy of action at all. The first thing to say, I think, is that is it easy to be beguiled into supposing that theories of action which forsake the concept of power for the concept of cause are not subject to a similar difficulty. We forget that if the question were posed: But what is it for one event to cause another? we would likely have nothing to say. We have become happy to accept the thought that the concept of cause is a basic one about which we can say very little, and which will not submit to any illuminating analysis. But then why might we not say the same about the idea of agency, about the idea of an exercise of power by an animal? I am inclined to think that it, too, is a basic notion, one in terms of which we cannot help but think about the production of events in the animal – and especially the human – world. And if it were thus basic, it would be wrong to expect that there would be very much in the way of analytical decomposition to be done, and so the criticism that not much had been done would be moot. Still, it might be said, we do not want two basic notions floating around in an area where we might have one. And surely, we animals are just mechanisms, at the end of the day – vastly complex machines, in which causal processes ultimately explain everything there is to be explained. An exercise of causal power on the part of the animal just has to be identical with a neural firing, or something similar – it has to be a mere event, like any other, something with its own causes and effects. To suppose that we need an unreduced notion of agency alongside the notion of cause is, then, surely a ludicrously anti-naturalistic suggestion, with about as much to recommend it as a proposal to restore dormitive virtues to their proper place in the explanatory resources of biochemistry. It seems to me, though, that this argument prejudges the question how biology will ultimately explain the remarkable phenomenon which is constituted by animal activity. That a whole system may affect its own parts is an idea that is coming slowly to gain acceptance in the biological sciences – and so the idea that an exercise of power on the part of an agent

310

Sub-intentional Actions and the Over-mentalization of Agency

must be identical with some small neural occurrence is by no means mandatory. Biology may find that it, too, requires, in the end, a notion of agency, which it will understand as the highest level manifestation of a general biological phenomenon of downward causation which it is already beginning to think it needs.19 I do not doubt that scientific light will come to be shed on the question how it is possible for an agent to exercise a power to affect its own body by the developing biological sciences. What I object to, though, is the thought that we philosophers should tidy up the area beforehand, by translating the notion of agency into what is alleged to be the scientifically preferable notion of event causation. We do not know how it is that we have the power to control movement and change of certain sorts in our own bodies – so let us not second-guess the question what resources will turn out to be necessary to explain it by insisting on forcing the phenomenon of action into an event-causal mould which simply does not fit it. I have tried to suggest, then, that we should take a good deal more heed than we have so far done of the phenomenon of sub-intentional actions. We should not do with them what Hume famously did with his ‘missing shade of blue’ example, and be content to regard them as marginal phenomena which ought not to be allowed to overturn a theory for which there is so much, otherwise, to be said.20 For they show us something important about actions; they show us, indeed, that most of what we have said about what is truly essential to them is incorrect, in that it imports into their characterization what I have called an over-mentalistic conception of the agent. And though I concede that I have defended a conclusion which is in some ways disappointing – that is, that it is not possible for philosophers to say a very great deal about what is essential to actions without invoking concepts to which the concept of action bears in any case a very tight relation – it is better to say a small amount that is true, than a very great deal that is false.

Notes 1. See e.g. Donald Davidson (1980). 2. Ginet (1990), pp. 11–12. It might perhaps be doubted whether mentally saying the word peu always constitutes an action – but it seems to me incontrovertible that it sometimes may. I might, for instance, resolve that I will mentally say the word peu 10 seconds from now – then, if I do so in response to that prior resolution, I shall surely have acted. 3. As will become clear later, I do not agree with Ginet that there is nothing which could count as the event which is caused by the agent in a case like this. 4. The term ‘striving’ is associated, in particular, with O’Shaughnessy (1980); the view that actions are tryings is defended by Hornsby (1980) and Pietroski (2000). McCann (1974) also defends what may be regarded as a variant of the view. 5. Ginet (1990), pp. 11–22. 6. For an argument that it does not follow, see my (2000), pp. 122–4.

Helen Steward

311

7. I speak of actions being ‘associated’ with reason-giving explanations, rather than being explained by such explanations, because I do not accept that typical reason-giving explanations are explanations of actions, considered as particular events. A reason-giving explanation is typically an explanation of why some agent S ␸-ed. But this is an explanation of a fact – that S ␸-ed – not an explanation of a particular. The connection between reasons and actions, then, if there is one, must be this: that where there is an action, there is something describable as S’s ␸-ing such that a true, reason-giving explanation of why S ␸-ed is available. This is the relationship I mean to encode by speaking of actions being ‘associated with’ reason-giving explanations. 8. It is therefore a different sort of objection from that which is made to the standard view by Rosalind Hursthouse (1991), for Hursthouse’s arational actions remain intentional. 9. O’Shaughnessy (1980), Vol. II, ch. 10. 10. Or, as the (inaccurate) jargon has it, which are not intentional ‘under any description’. The jargon is inaccurate because it does not make any sense to characterize an action itself as intentional or unintentional, so I have employed a more laborious formulation. 11. See Anscombe (1957), pp. 9–15. 12. Hornsby (1980), p. 37. 13. Indeed, there is a sense in which the idea that this would have to be the basis on which they are to be included is guilty of the same mistake of over-mentalization which I will shortly discuss. 14. Though I may quickly become active after the initial stumble, as I try to save myself, recover my balance and so on. 15. It would only be a weak version because, as I remarked earlier, the special sense of the question ‘why’ which Anscombe helpfully elucidates has no application here. We are looking precisely for a reason which is not a reason in Anscombe’s sense – since we may not be aware of our sub-intentional actions. 16. One might hope, in doing so, to take a leaf or two from the work of those who have defended what has come to be known as animalism in the field of personal identity, for example, Wiggins (1980), Snowdon (1995, 1996), for in that area too, a distinction is often made between an animal and its body. 17. I mean to use the word ‘substance’ here in its Aristotelian sense – that is, to encompass such things as horses, and oak trees, not such things as milk and castor oil. 18. Of course, the word ‘do’ is only of limited help here – but emphasized in the way I have emphasized it it can perhaps serve to elicit some of the wanted intuitions. 19. As recognized by the newly burgeoning field known as ‘systems biology’. 20. See Hume (1975), p.21:’This instance is so singular, that it is scarce worth our observing, and does not merit that for it alone we should alter our general maxim’.

References Anscombe, G.E.M. (1957): Intention (Oxford: Basil Blackwell). Davidson, D. (1980): ‘Agency’, in Essays on Actions and Events (Oxford: OUP), pp. 43–61. Ginet, C. (1990): On Action (Cambridge: CUP). Hornsby, J. (1980): Actions (London: Routledge & Kegan Paul).

312

Sub-intentional Actions and the Over-mentalization of Agency

Hume, D. (1975): Enquiries Concerning Human Understanding and Concerning the Principles of Morals ed. L.A. Selby-Bigge (Oxford: OUP). Hursthouse, R. (1991): ‘Arational Actions’, in Journal of Philosophy 88, pp. 57–68. McCann, H. (1974): ‘Volition and Basic Action’, in Philosophical Review 83, pp. 451–473. O’Shaughnessy, B. (1980): The Will, 2 vols. (Cambridge: CUP). Pietroski, P. (2000): Causing Actions (Oxford: OUP). Reid, T. (1969): Essays on the Active Powers of the Human Mind. ed. Baruch A. Brody (Cambridge, MA: MIT Press). Snowdon, P. (1995): ‘Persons, Animals and Bodies’, in J. Bermudez, A. Marcel and N. Eilan (eds), The Body and the Self (Cambridge, MA: MIT Press). Snowdon, P. (1996): ‘Persons and Personal Identity’, in S. Lovibond and S. Williams (eds) Essays for David Wiggins: Identity, Truth and Value (Oxford, Blackwell), pp. 33–48. Steward, H. (2000): ‘Do Actions Occur Inside the Body?’, in Mind and Society 1, pp. 107–125.

17 Determinism, Intentional Action, and Bodily Movements Frederick Stoutland

Here is one formulation of determinism: “The world is governed by (or is under the sway of) determinism if and only if, given a specified way things are at a time t, the way things go thereafter is fixed as a matter of natural law.”1 It is a strong thesis and there are many reasons for rejecting it, including, in my view, its ruling out an acceptable account of human action. But rejecting it is one thing and establishing how that bears on an account of action is another. I want to consider the role such rejection might have in a plausible account of how an agent’s reasons explain her intentional action. Such explanation—“rational explanation”—is constitutive of human agency in that to be capable of acting intentionally as a human being is to have the capacity to give and receive reasons for the actions of oneself and others. This is not a specialized capacity for experts but the ground of a human life where one takes responsibility for what one does intentionally but not for (most) other things one does, where one can distinguish between things that just happen and mistakes one makes, where one knows the difference between remorse and regret. The fundamental issue is not free but intentional action. Human action is intentional in a way the behavior of non-human animals is not. Both act intentionally when they respond to reasons, but only human beings respond to reasons as such. A lion perceives an antelope in the vicinity, and that is the reason it stalks the animal, but it does not perceive the presence of the antelope as a reason for its stalking. Human beings, however, can apprehend a situation as a reason for them to act in a certain way and may act because of that reason, in which case what they took to be a reason for them to act explained why they acted. Such action may be free but it may be unfree because coerced by threat, required by one’s position, restricted to only one alternative, and so on. It cannot, however, be either free or unfree unless it manifests a responsiveness to reasons as such, and it is, consequently, intentional in a sense the actions of brute animals are not. I shall first distinguish between nomic and rational explanation, arguing that the thesis of determinism is irrelevant to rational explanation per se. 313

314

Determinism, Intentional Action, Bodily Movements

Then I will argue that rational explanation involves non-necessitating causes and that the indeterminist-libertarian account of such causes should be rejected. I next consider the relation between rational and nomic (here neurophysiological) explanations where they intersect, namely in explaining the bodily movements involved in intentional action, and I argue that rational explanation is there prior to neurophysiological. I argue, finally, that making sense of this requires that we reject not only determinism but a broader thesis, the “completeness of physics.”

1 Wilfred Sellars distinguished between the logical space of causes and the logical space of reasons, but in order not to beg the question about the nature and status of causation, it is better expressed as one between the logical space of laws and the logical space of reasons, that is to say between nomic explanations and rational explanations.2 Nomic explanations that appeal to natural laws are at the heart of the physical sciences even though not all their explanations are nomic. The biological sciences in particular offer numerous teleological explanations that are not nomic (nor rational as I use the term). But the formulation and application of natural law explanations are central to the physical sciences and to any inquiry that seeks to utilize their methods. Two features of such explanations are essential. The first is that they exclude normative concepts. Giving a scientific explanation is, of course, a normative activity, but the explanations given are not. We may explain an eclipse of the sun by saying that the reason for its occurrence was the natural laws governing the solar system, but this use of “reason” has no normative dimension. We use the term because reference to the laws of the solar system makes it intelligible why the eclipse occurred, but this is an intelligibility of laws, which would not claim the solar system favored such an eclipse—found it appropriate, correct or justified—a type of explanation that has been excluded from the physical sciences since the demise of Aristotelian physics. The other is that nomic explanations are not first person, by which I mean that the concepts used are not implicit in the behavior, attitudes, or mental states being explained but are created by scientific investigators and belong to their domain.3 This is obvious when the objects of explanation have no mastery of concepts, but it applies also to the behavior of human beings, who do. An explanation of human behavior in the logical space of laws uses concepts only experts need know anything about, which are not implicit in the constitutive, commonsense conception of ourselves and the world in which we live. This second feature means that they aim at something like a view from nowhere—at an understanding and explanation of the world that is

Frederick Stoutland

315

accessible to anyone regardless of her particular point of view. Although it is dubious that there is a view literally from nowhere, the point reminds us that the normative significance of the world—its status as favoring some actions (and disfavoring others), as rendering some claims correct or justified (and others not)—is accessible only to persons who occupy a particular point of view in virtue of their culture, education, perceptual capacities, practical competencies, needs, aims, sensibilities, and so on. To discover the normative significance of a situation for an agent requires a grasp of how the world presents itself to her; that is possible without sharing her point of view, but only if one can grasp what it is. These two features are, therefore, interrelated: nomic explanations are not normative since they do not involve first-person concepts; and because the latter are normative, they do not figure in the precise generalizations of natural laws.4 Formulating those generalizations is very difficult because it requires concepts quite unlike those involved in the everyday understanding of our thought and action, because we must describe human behavior in terms of technical concepts that belong to a theory that enables mathematically formulated generalizations and inferences. It is against this background that determinism is a live option. Given a range of phenomena describable in terms that permit precise generalizations, it is possible to formulate natural laws that govern that range of phenomena, so that “given a specified way things are at a time t, the way things go thereafter is fixed as a matter of natural law.” Since the sixteenth century, the motions of bodies in the solar system have been taken to constitute such a range of phenomena, which means the solar system is a deterministic system.5 Determinism more generally would be the claim that all phenomena can be described in terms of concepts that are instances of natural laws.

2 Because rational explanations exclude the precise generalizations presumed by natural laws, the question whether all or only some of them are deterministic is moot since they all require the normative and first-person concepts that explanations in the logical space of laws exclude. They appeal to considerations that favor one’s acting in a certain way,6 sometimes in the strong sense of requiring it (on moral grounds, because of one’s role or profession, and so on), but more often in the weaker sense of rendering it appropriate, beneficial, justified, and so on, the latter being normative in the broad sense that includes the evaluative. Rational explanations can, of course, appeal to considerations that do not actually favor the action. Jacob walked across the street because someone appeared lost on the busy street corner and needed help; but she may in fact have only been observing the passing scene. His reason for walking across the street did not, therefore, favor his doing so, but it nevertheless explained it. Ascribing the reason to

316

Determinism, Intentional Action, Bodily Movements

him, however, required grasping his point of view, that is, understanding how the situation presented itself to him, namely, as someone who was lost and needed help. Although rational explanations can explain a variety of matters—why an agent took a state of affairs to favor his acting, why he regarded one such reason as stronger than another, why he concluded that, all things considered, this is what he should do—my concern is explanation of why an agent acted as he did (the reason for him to act having become the reason because of which he acted). Knowing that Jacob walked across the street because someone appeared in need of help renders it intelligible why that action—that event—occurred and if, as many philosophers maintain, the explanation of the occurrence of an event is causal, then this one is causal. But this is an intelligibility of reasons, and hence the causality is also a matter of the logical space of reasons not of laws. Since the explanation does not render the action intelligible by showing that it was an instance of a natural law, the reason for which he acted may be said to have causally explained his action but not to necessitate it.7 In this essay I shall always use “cause” to denote causal explanations and not causal relations. Causal explanations, like explanations generally, are intensional: explanations are always of phenomena as such and such—under such and such a description—and hence “cause” denotes a relation between sentences (propositions, facts).8 Causal relations, by contrast, hold between particular events no matter how described, so that their ascriptions are extensional. My claim is that rational explanations are causal explanations in the logical space of reasons, which cannot be a claim about causal relations, for they hold no matter how described and hence belong to no particular logical space. Nor do they necessarily yield an explanation of phenomena since they may be true even if the cause and effect are described in terms with no explanatory import and, consequently, do not render the occurrence of an event intelligible. Davidson also held that explanations of intentional action are not nomic, but he thought further that where there is causal explanation, there must be a causal relation. He maintained, for instance, that beliefs and desires are not events and hence not causally related to actions, but that in order to explain actions, they must be associated with events to which the actions are causally related. He also held that wherever there is a causal relation, there are descriptions of the related events that are instances of a law of physics.9 I think Davidson was mistaken in thinking a causal explanation requires an associated causal relation, but I also believe that what is acceptable and important in his view does not require that claim—certainly not that it entails a law of physics. If (but only if) Davidson is read in that way, his view of rational explanation as causal is similar to mine.10 It was reading Anscombe that convinced me that a reason may causally explain an action but not necessitate it. She argued that a cause can bring

Frederick Stoutland

317

about an effect even if there are no natural laws at any level that connect the cause with the effect. The fact that C explains why E occurred on this occasion (given the circumstances) by being its sufficient cause does not entail that when C occurs again, E must, as a matter of natural law, also occur—even if the situations are the same. The claim that C was sufficient for E means that C was enough to explain why E occurred—that it was a sufficient condition in that sense—but it does not follow that it was a sufficient condition in the logico-philosophical sense of whenever C occurs, then E occurs or (better) if C were to occur, then E would (have to) occur.11 Rational explanations are like that: someone who appeared lost was sufficient reason for Jacob to walk across the street to help her—it explained why he did so—but it doesn’t follow that if he were to perceive such a person again, he would come to her aid, even if the circumstances were the same. Let me say more in defense of this notion. We can characterize a reason as a cause if it explains why an agent acted intentionally as he did, hence why the event that was his acting occurred, but it is non-necessitating in that it does not show that his acting was an instance of a natural law. A rational explanation does not exhibit such necessity because it requires that the action be characterized in the logical space of reasons, hence in terms of normative and first-person concepts, which rule out the precise generalizations required for natural laws. A non-necessitating cause does not violate the same cause-same effect principle that if reason R explains why an agent acted in a certain way on one occasion, then R explains why she acted in that way on other occasions similar to it. While the principle is true, it does not entail that if R is a sufficient reason for an agent to A, she must A intentionally. Assume that R explains why she A-ed in these circumstances; if, under the same circumstances, she did not A, R may still have been a sufficient reason for her to have A-ed even though it did not explain why she A-ed (since she did not A). Being able to infer from what an agent did intentionally to the reason for which she did it, does not entail that we can infer from what was a sufficient reason for her to act in a certain way to her acting in that way. To put the point in another way: R can explain why an agent A-ed on an occasion without a second-order explanation of why R explained why she A-ed. There need, therefore, be no explanation of why R, although a sufficient reason for her to have A-ed, did not, on an occasion, causally explain her having A-ed (if she did not A). This means that establishing which reason explained an agent’s having acted is not a matter of verifying the generalization that whenever R is a sufficient reason for an agent to A, given R she will A. Verifying such a generalization is a common way of establishing causal explanation in the logical space of laws where causal explanations are typically taken to express causes that necessitate.12 One example is the use of Mill’s methods of sameness and difference; another is the appeal to manipulation or intervention: if we could by manipulating C bring about a change in E, then C is the (necessitating)

318 Determinism, Intentional Action, Bodily Movements

cause of E. These methods are, however, irrelevant to establishing causal explanation in the logical space of reasons since causal explanation there neither necessitates nor presumes generalizations. We have to use an interpretive-holistic procedure that considers not only what occurred before the agent acted but what she did afterwards, how she acted on other occasions, what the course of her life had been in various respects, what it became later, and so on. Such procedures may not establish with certainty the reason for which she acted—there might not be a fact of the matter—but they are, in general, sufficient to distinguish the reason which explained her action from reasons that merely appeared to explain it.13

3 The way many indeterminist libertarians deal with the notion of a non-necessitating cause seems to me mistaken. These indeterminists (as I shall call them) are concerned with free action, which they think can be explained by an agent’s reasons only if they causally explain the action but do not necessitate it. I object, not to the latter, but to the way indeterminists construe it in thinking the crucial issue is what they call the agent’s “active control” over her acting. “An agents exercising active control,” Randolph Clarke suggests, “consists in her action’s being caused by her, or in her actions being caused, in an appropriate way, by certain events involving her, such as her having certain reasons, and a certain intention.”14 The causation here is, of course, non-necessitating—“non-deterministic” as Clarke calls it—but it is construed not simply as denying the relevance of determinism to rational explanation but as affirming indeterminism in the quantum physics sense. A non-deterministic cause is one that fixes the effect necessarily, but probabilistically rather than universally. This worries indeterminists, for they think it amounts to an agent’s having “diminished control” over her action. If causal explanation of her action were deterministic so that the effect always followed the cause, then the agent would have complete control over her action. But if causal explanation is non-deterministic, then the agent has only diminished control since the reason that purportedly explains her action may not always produce it, and the less often it produces, the less control she has over it. This creates a dilemma: deterministic causation allows for full control but not for free action, whereas indeterministic causation introduces an element of luck or chance that may mean insufficient control for an action to be intentional. The view assumes that if a cause explains an effect, the effect must follow on the cause universally because that accounts for the explanatory power of the cause. Since free action involves a reason that is its causal explanation but is not universally followed by the action, the reason does not fully explain why the action was performed. The indeterminist view is that to the

Frederick Stoutland

319

extent to which we do not have a deterministic causal explanation, to that extent the effect occurs as a matter of chance or luck. I would reject the dichotomy, either universality or chance, on the ground that a reason explains an action without any universal connection between the sufficient reason for the action and the action. That someone was, in Jacob’s view, lost and in need of directions is sufficient reason for him to cross the street; the reason explains his acting but presumes no universal generalization. That does not mean it is mere luck or chance that he crossed the street—that his doing so lacked sufficient explanation—but that the reason that explained why he acted on this occasion need not result in (and hence will not explain) his acting in that way on other like occasions. To accept the dichotomy is to put rational explanation in the logical space, not of reasons, but of laws, for the latter is the home of that dichotomy. However, an account of how actions are explained by reasons without being necessitated requires recognition that rational explanations are a distinct type to which neither determinism nor indeterminism is relevant. Such explanations are not nomic, but that does not mean they are indeterministic in the sense of quantum physics. This latter point is captured by what philosophers of science call the “completeness of physics,” a thesis that claims that all phenomena can be accounted for by nomic generalizations, which may be either universal (hence deterministic) or probabilistic (hence indeterministic). The latter is also a form of necessity, for in physics probability is precisely fixed by law. The thesis is that all events can be described so that their occurrence, under that description, is explained as the result of a (perhaps very complex) network of laws whose vectorial combination explains and fixes (perhaps probabilistically) those events. The claim is that whether or not we know how those laws interact to fix events, they nevertheless operate underneath.15 Libertarian indeterminists accept something like the completeness of physics, although without affirming that there is a precise mathematical measure for all indeterministic causes. Reasons that do not causally explain actions deterministically do so indeterministically, which entails that rational explanations of actions are more or less incomplete depending on the degree of probability of the relevant generalizations. But if the completeness of physics is irrelevant to rational explanations per se, then the latter, although non-necessitating, are not for that reason incomplete, and the intentional actions they explain are not a matter of luck or chance, even though they presume no precise generalizations of any kind.

4 I have claimed that nomic explanations can render phenomena intelligible by showing them to be instances of natural law, whereas rational explanations can render phenomena intelligible by showing they are done for

320 Determinism, Intentional Action, Bodily Movements

reasons. The question remains whether the same phenomena can be rendered intelligible in both ways. Can an agent’s behavior be explained not only as his acting intentionally under a description, but also as non-intentional movements of his body and limbs that result from muscle contractions and neural processes. It is a conceptual truth that acting intentionally can (in general) be given a rational explanation since it is essentially acting for a reason,16 and while it is contingent that an agent’s behavior can be given a neuroscientific explanation, nothing rules it out. The question is whether and in what sense the same behavior can be explained in both ways. Jaegwon Kim argues that it cannot: “A ‘purposive’ explanation of human action in terms of the agent’s ‘reasons’ and a ‘mechanistic’ (e.g. neurobiological) explanation of it in terms of physiological mechanism must be regarded as incompatible and mutually exclusionary—unless we accept an appropriate reductive relationship between intentional states and underlying biological processes.” His objection appeals to what the “principle of explanatory exclusion”: “there can be no more than one ‘complete’ and ‘independent’ explanation for any single explanandum.”17 Let us assume that the two types of explanation are independent because the states to which rational explanations appeal cannot be reduced to “underlying biological processes.” Let us also assume that each is complete in that one makes it intelligible why the agent acted intentionally as he did, the other why his bodily movements described non-intentionally occurred as they did.18 This does not violate Kim’s principle since two such independent and complete explanations of the same action are different types of explanation. We do not have a single explanandum because explaining action requires that we explain it as described—in this case, either as in rational explanation or as in neurophysiology. Since these different types of description belong to different logical spaces, they are compatible, and so are the explanations. This raises the problem of why there should be what I have called congruence between the two types of explanation. The rational explanation of Jacob’s walking across the street is his taking someone to be lost and in need of help. Let us assume we can also explain his behavior, described as non-intentional bodily movements, in terms of muscle contractions and neural processes. Why is what he did for a reason congruent with what occurred because of muscle contractions and neural processes? Why would his reasons explain his intentionally walking across the street, while neurophysiology would explain the non-intentional movements that occur in his acting? There are two approaches to this question I take to be mistaken. The first maintains that once it is recognized how different the two kinds of explanations are, the very notion of congruence dissolves. Neurophysiology specifies the causes of bodily movements, whereas to give an agent’s reasons for acting is to explain what his acting means. On this view, neurophysiology sets

Frederick Stoutland

321

action in the context of causal laws that show how it came about, whereas a rational explanation re-describes the action that came about so as to reveal its meaning. This was the view of anti-causalist philosophers of action in the 1960s,19 who held that an agent’s action is independent of his bodily movements in that, on the one hand, different bodily movements will have (in different contexts) the same meaning, and, on the other hand, the same bodily movements will mean different things depending on the context in which the agent acts. As a result, the very notion of congruence dissolves if we think through what it is supposed to mean. The objection to this is that it fails to note that “bodily movements” is used in two ways. On the one hand, an agent’s bodily movements may be described as the intentional bodily movements entailed by the primitive description of his acting. If an agent intentionally moves his arm, then his arm intentionally moves; if he intentionally walks, then his legs (and other parts of his body) intentionally move so that he walks (always at least one foot on the ground, for instance), and so on. On the other hand, his bodily movements may be described non-intentionally and given quite specific descriptions that vary with each instance of his walking. Described in that way, they can be explained by neurophysiology but not by a rational explanation. It is true that an agent’s intentional action is independent of his nonintentional bodily movements in that what explains his acting intentionally does not explain his bodily movements described as in neurophysiology. But it is false that bodily movements described in that way have the same meaning in different contexts or different meanings in the same context: meaning is irrelevant to bodily movements described nonintentionally. While it is relevant to them described as intentional, so described they are explained by the reasons that explain intentional actions, and such explanations, I contend, are causal in making it intelligible not only what an agent did but why he acted as he did, and hence they explain why he intentionally moved his body and why his body intentionally moved. That means both kinds of explanations explain bodily movements, though under different descriptions, and hence the question remains of why there is congruence between the two kinds of movements. The other approach is just the opposite. Rather than trying to dissolve the problem by separating the types of explanation, it attempts to unify then in one explanatory theory. This is the approach of Hartrey Field’ (his term is “mesh”), who requires that we “explain why our neurophysiological laws and our psychological laws never come into conflict—that we show that our neurophysiology and our psychology ‘mesh’.” He goes on: “Whenever we employ laws at different levels, there is a prima facie possibility of their coming into conflict, and it is eminently reasonable to

322 Determinism, Intentional Action, Bodily Movements

want an explanation of why such conflict does not arise ... . [Hence] a main advantage of reducing psychology to lower-level science ... is that doing so one would be able to explain the mesh between psychology and the lower-level sciences.”20 This reductivist account of congruence assumes that the reasons explaining an agent’s intentional action are her beliefs and desires, which are conceived as inner states that are both reasons for her action and causes of the non-intentional bodily movements involved. Given the conviction that causation must be physicalistically based, it is then argued that beliefs and desires cause these bodily movements because of their place in the logical space of laws, which they have because they are either identical with or superveniently determined by neural states. I cannot give here the detailed critique these physicalistic theories merit so I shall restrict myself to two general objections. First, to claim that what are in the logical space of reasons are either identical with or determined by what are in the logical space of laws assimilates the two types of explanation much too closely. Rational explanations are normative and first person and hence explain actions and bodily movements as described in those ways, whereas neurophysiology explains them as non-intentional movements described in technical terms devoid of normative significance. Combining two such different sets of descriptive terms and explanatory schemes in one unified explanatory theory is a not coherent project. Supervenient determination may seem more plausible than identity, but only if a psychological state were taken to supervene, not on a specific neural state, but on a spatiotemporally discontinuous collection of neural states. That, however, would have no explanatory value, certainly none for explaining why congruence obtains.21 The other objection to this approach is that it takes for granted that the only alternative to the first is to unify the two types of explanation. There is a third alternative, however, which maintains that not only formulating an explanation is a mode of intentional action, so is explaining. Explanation does not exist in the nature of things to be discovered, with formulation being our only contribution, but is something we do. We explain by making things intelligible—by fitting various factors together in a pattern that shows how or why certain things came to be, ceased to be, changed, and so on. In many cases the relevant factors are already known, the explanatory task being to fit them together to render them intelligible. In other cases, we have to discover such factors, but they have explanatory value only against the background of theories, generalizations, interpretations, narratives, and so on, all of which are human creations. They are not always created intentionally, of course—those in the logical space of reasons typically are not—but those in the logical space of laws usually are. To understand why congruence obtains, we should reflect on how the two ways of explaining human behavior are interrelated.

Frederick Stoutland

323

5 The problem of congruence arises at the level of bodily movements because that is where rational and neuroscientific explanations converge. The notion of bodily movements is very complex, however, and in need of more discussion. I begin with the notion of a basic (or primitive) act, which Arthur Danto introduced as an act not done by doing some other act. He recognized that whenever we act intentionally, we perform a basic act because otherwise there would be an infinite regress of our not acting until we had already performed another act. He also assumed that basic acts are numerically distinct from acts done by doing the basic act, an assumption Davidson rejected in taking over Anscombe’s view (which I take to be correct) that an agent whose act has many results acts only once although her act has as many description as it has results. A basic act is, therefore, not numerically distinct from the acts done by performing it: whether an act is basic (“primitive” as Davidson puts it) depends on how it is described, which means the notion is intensional. Jacob walked across the street, thereby helping someone who appeared lost, and also frightened the pigeons. Those are three descriptions of the same acting: the second and third describe what he did by walking across the street, but the first describes his acting neither as a result of anything he did nor in terms of anything it resulted in,22 and hence describes it as a primitive act. Described as primitive, an act can have a rational explanation (Jacob walked across the street in order to help someone) but it cannot be described in terms of what it resulted in. Davidson wrote in a well-known passage that “our primitive actions, the ones we do not by doing something else, mere movements of the body— these are all the actions there are. We never do more than move our bodies; the rest is up to nature.”23 My concern here is not with what Davidson meant by this passage but with the way the “standard story” of action understands it, namely, that an action consists of non-intentional bodily movements, the kind neurophysiology describes and explains, which are what a human action is.24 Although rational explanations do not typically describe an agent’s actions in neurophysiological terms, what are described are her nonintentional bodily movements; they are, in Quine’s terms, the ontology of action, while everything else is ideology. What on this view is special about primitive action descriptions is that they explicitly describe actions in terms of non-intentional bodily movements and thereby directly describe, indeed, characterize, the ontology of action, which is essentially the ontology of neurophysiology. In order to constitute an action, non-intentional bodily movements must have a special kind of cause—the standard view is that they must be caused (in the right way) by an agent’s beliefs and desires—but being so caused does not change their nature as non-intentional bodily movements. Whether they are so caused and hence

324

Determinism, Intentional Action, Bodily Movements

constitute an action, or are not so caused and have only a neurophysiological explanation, they are essentially mere bodily movements. This mistaken approach appears plausible in part because it focuses on atypical examples of primitive action descriptions such as “raising an arm” or “moving a leg.” These are primitive action descriptions, but they are misleading examples. It may seem plausible to say that raising an arm consists of the mere bodily movements of one’s arm rising (if caused in the right way) because “S raised his arm” entails that his arm rose, and the latter sounds like a mere bodily movement as described in neurophysiology. Or “S moved his leg,” which entails that S’s leg moved, may be taken to mean that the action consists of a mere bodily movement that essentially differs from a knee jerk or a cramp only in its cause. But many other examples of primitive action descriptions are not like these: “S walked, jumped, pushed, skipped, scampered, chewed, climbed, spoke, typed, sat down, wrestled, wrote, smiled, laughed, wept, lifted, threw, pulled, kicked, strolled, scratched, bent, hit.” These describe an agent intentionally moving her body and hence entail descriptions of her body’s intentionally moving: “S walked” entails that her legs intentionally moved so that she walked (she qualified for a walking race), “S intentionally chewed” entails that her jaw intentionally moved up and down, “S typed” entails that her fingers intentionally moved as typists’ do. These are not, however, non-intentional bodily movement descriptions suitable for neurophysiology, but intentional descriptions conceptually dependent on—unintelligible apart from—primitive action descriptions.25 This is also true of simple primitive action descriptions like “raising an arm” or “moving a leg,” although not as obvious. An agent can intentionally raise her arm only if her arm rises, which means that “S is raising her arm” entails “S’s arm is rising.” The latter, however, describes her bodily movements as intentional, as her arm intentionally rising in the way it does when an agent intentionally raises it; it does not describe those bodily movements as non-intentional. One could in principle describe each instance of her arm intentionally rising as a collection of non-intentional bodily movements but those descriptions (although not entailed by the agent’s raising her arm) would presuppose that the intentional bodily movement had been identified, since it is the movement described as intentional that is now to be (re) described as non-intentional bodily movements. Arms also move non-intentionally when they move as a result of a neurological disease or electric shock, and the like, and not because agents intentionally move them. Such bodily movements are intentional under no description, and I shall, therefore, call them intentionless bodily movements. They contrast with non-intentional bodily movements, which are bodily movements described in non-intentional ways but which are intentional under another description, for instance, the one entailed by the primitive action description of the action. A primitive action description is

Frederick Stoutland

325

transitive because it describes the action as an agent moving his body, whereas an intentional bodily movement description is intransitive (although it entails that some agent moved her body in that way). A nonintentional bodily movement description is always intransitive, although those movements also have a description under which they are intentional. An intentionless description of an agent’s bodily movements is always intransitive, since such bodily movements are not involved in any intentional action. Does the fact that S intentionally raised her arm entails that her arm intentionally rose, mean that raising her arm explains her arm’s intentionally rising? While that is often said, the relation between them is too close for an explanatory relation. One might ask in a meeting why Smith’s arm rose and get the answer “because she raised it,” but that has a point only if it is general knowledge that her arm is defective: it goes up by itself randomly or is so weak that others usually push it up. But if the movement is intentional, then to ask why Smith’s arm rose is just to ask why she raised it. The rational explanation of bodily movements described as intentional and of the action that entails them as so described are the same: the reason Smith’s arm intentionally rose was that she wanted to ask a question, which is just the reason she raised it.

6 It might be responded as follows. Even if primitive action descriptions are not in terms of non-intentional bodily movements suitable for neurophysiology, those movements do occur when we act and are what our acting consists of. That is enough to save the claim that the ontology of action is exhausted by non-intentional bodily movements, while the rest is ideology—how things are described. On this view, ontology is a matter of what is described and ideology a matter of descriptions that denote the features of what is described. Action consists of non-intentional bodily movements, for those are what action descriptions are true of. This, however, simply denies that, although what happens, happens under any description, whether it is intentional depends on how it is described. Bodily movements that are not intentional under neurophysiological descriptions may be intentional under other descriptions. The claim that non-intentional movements are what actions consist of can only mean that descriptions of them as non-intentional are more basic than descriptions of them as intentional—that they yield the essential nature of bodily movements. I think that is false but before giving my reasons, let me make two points. One is to note Anscombe’s way of saying that actions do not consist of non-intentional bodily movements because they are not what we describe in describing action. “The proper answer to ‘What is the action, which has all

326 Determinism, Intentional Action, Bodily Movements

these descriptions?’ is to give one of the descriptions. Any one, it does not matter which; or perhaps it would be best to offer a choice, saying ‘Take whichever you prefer’.”26 We may describe an action in terms of the nonintentional bodily movements involved, but there is no reason to give such description priority; moreover, to describe it that way is not to describe is as action. The other point concerns the kind of ontology we would have if we grant that the issue is not what actions consists of beneath any descriptions but how best to describe them. That ontology is, in my view, like Davidson’s (which he took to be Aristotle’s and Spinoza’s): “Ontological monism accompanied by an uneliminable dualism of conceptual apparatus ... . There is only one [kind of] substance [but] the mental and the physical are irreducibly different modes of apprehending, describing, and explaining what happens in nature.”27 This is the view that there are no non-physical entities, that is, none that does not have a physical description, because to have a physical description is to be physical. That is not physicalism because many entities also have mental descriptions, and to have a mental description is to be mental. Davidson had an expansive view of a mental description: “To say of an event, for example, an intentional action, that it is mental is simply to say that we can describe it in a certain vocabulary—and the mark of that vocabulary is semantic intentionality.”28 On this view, intentional actions are mental events, although many are, as we say, physical rather than mental acts. To articulate this view, we should distinguish different senses of a physical description. Some are physicalistic, that is, involve only basic concepts of physics, the aim of neuroscience being to construct explanations that use only such descriptions. There are also descriptions that are chemical and others that are biological, neither of which is reducible to the physicalistic. There are, in addition, descriptions of actions and movements as intentional, which Davidson called “mental.” Anscombe called primitive action descriptions “vital”: “These descriptions are all basically at least animal. The “characteristically animal movements” are movements with a normal role in the sensitive, and therefore, appetive, life of animals.”29 She also wrote that “a great many of our descriptions of events effected by human beings are formally descriptions of executed intentions,” and these might be thought of as another category of physical descriptions. Ontological monism in this sense presumes “irreducibly different modes of apprehending, describing, and explaining what happens in nature,” and although what happens in nature happens no matter how described, it is apprehended, explained, or intentional (or non-intentional) only as described in one or another of these different modes. Understanding action may involve several of them, but since the physicalistic mode is by no means basic to that understanding, the ontology of action is not the physicalistic ontology of mere bodily movements.

Frederick Stoutland

327

7 Let us return to congruence. Consider a relatively simple example of intentional action—Paul’s pulling on a bell rope in order to alert the gate-keeper to open the gate. His acting is intentional under the primitive description “moving his arm down,” and it has a rational explanation, namely to alert the gate-keeper. Let us assume we also have a neurophysiological explanation of the non-intentional bodily movements involved in pulling on the rope. What the two kinds explain are, therefore, congruent, because the intentional action explained by the first, and the non-intentional bodily movements explained by the second, not only were not in conflict but occurred harmoniously: Paul pulled on the rope smoothly without jerking, his body working as it should. What kind of account of this is required? I contend that it does not require a substantive explanation. The description of Paul’s behavior as intentionally pulling on the rope entails the description of his arm as intentionally moving downwards. A neurophysiological explanation, by contrast, describes it as non-intentional bodily movements (“colorless movements”). It is evident that a substantive explanation of why congruence obtained would require a theory about how two different kinds of explanation can result in movement that mesh rather than conflict. That is what Field projects—a theory that reduces rational explanations to neurophysiology, or at least shows how intentional movements are determined by non-intentional ones. But surely there is neither need nor place for such explanation of how the intentional movements of Paul’s arm are congruent with the nonintentional movements of his arm, for those are the same bodily movements differently described. What are described as intentional movements in the logical space of reasons are what are described as non-intentional movements in the logical space of laws. The descriptions are quite different, but as long as they are true of the same bodily movements, they cannot be in conflict, and hence their congruence neither has nor requires a substantive explanation.30 That does not mean that it must be possible to explain an agent’s bodily movements under descriptions of both kinds. If his bodily movements are the result of a spasm or electric shock, they are not intentional under any description and cannot be rationally explained; and if we are unable to describe them in non-intentional terms, we cannot explain them in neurophysiology. But if we do succeed in describing in non-intentional terms the bodily movements involved in an agent’s intentional action, the descriptions must describe the same movements that are intentional under a different description. If Paul intentionally moved his arm to pull on the bell rope, then any adequate neurophysiological explanation of the nonintentional movements of his arm must yield movements congruent with

328

Determinism, Intentional Action, Bodily Movements

the intentional movements of his arm, for those are the same movements, though explained in different logical spaces. Nor does it mean that an agent’s intentional and non-intentional bodily movements cannot fall apart. If Paul set out to ring the bell but his upper body (unknowingly) became paralyzed, the bodily movements required for his being able to pull on the bell rope would not have occurred, and he could not have pulled the rope. The question of congruence would be moot, however, because his behavior could not be described as intentionally pulling on the rope. There would be no congruence between bodily movements intentional under a description and those not intentional because there would be no movements intentional under a description: his behavior could be explained only as intentionless (rather than non-intentional) bodily movements in the logical space of laws. Consider Paul as not wholly paralyzed but unable, because of some neural defect, to move his arms and hence to pull on the bell with either arm. He might use his teeth, however, and then a rational explanation of his acting would cite his intentionally clenching his teeth on the bell rope in order to pull it by moving his head. That would entail the intentional movement of his teeth and head, movements that might also be described in non-intentional terms and given a neurophysiological explanation. But the latter would be acceptable only if the bodily movements it explained under non-intentional descriptions were the same bodily movements that were intentional under a description entailed by his intentional action. Paul managed to pull on the bell rope even though he could not move his arms, which means that the intentional movements entailed by his pulling on the rope were congruent with the non-intentional movements of his limbs necessary for him to be able to pull on the rope. What accounts for the congruence is that both kinds of explanation explain the same movements differently described, not now of Lars’ arms, but of his teeth and head. That is surely not a substantive explanation of congruence. Congruence, therefore, obtains necessarily if the agent’s behavior can be given a rational explanation. If she failed to act intentionally as expected because she could not move her body and limbs in the expected manner but was able to act in a different way, then the description of her bodily movements as intentional (entailed by the description of her action) would also be different. But so would the description of her bodily movements as non-intentional, and a neurophysiological explanation of them would be adequate only if it yielded the same movements described as intentional. The agent’s movements as explained by neurophysiology will necessarily be congruent with the rationally explained movements since those are the same movements. The notion of the same bodily movements being intentional under some descriptions and not under others is like the notion of the same action having different descriptions. It does not imply that there is a unique description of

Frederick Stoutland

329

what bodily movements descriptions describe; to specify what is described, cite any (true) description. In the overall sense any description will do, though there are contexts in which one is better than others. It is true that a description under which bodily movements are intentional individuates differently from a description under which they are not intentional. Describing intentional bodily movements as “walking” gives them a unity that is lacking in describing them as non-intentional movements that enable me to walk. But if bodily movements are thought of as collections of movements, then different descriptions describe the same collection even if they individuate its members differently. There is no such thing as the correct or the basic way to individuate them; individuation is not context-free.

8 This account of congruence takes rational explanations as basic and requires that neurophysiological explanations conform to them. It will be objected that this is unacceptable to anyone sympathetic to the achievement and status of the natural sciences. My response is to reiterate that both kinds of explanation render the occurrence of phenomena intelligible. A rational explanation renders human behavior intelligible in the logical space of reasons, while a neurophysiological explanation renders it intelligible in the logical space of laws. Both explain why behavior, thus described, occurred but neither is basic in an absolute sense because priority depends on the context. Rational explanations have priority when our concern is with the relation of the two ways of explaining action. We have an action only if we have an event that is intentional under some description and hence was done for a reason. To consider the role of neurophysiological processes in explaining action, therefore, is to consider their role in explaining something that has a rational explanation. What neurophysiology seeks to explain in this context are bodily movements described non-intentionally that are also intentional under (another) description. It may be objected that we should begin the inquiry with non-intentional bodily movements explicable in neurophysiology, and then add on the features necessary for them to be actions. The “standard story” of action proceeds in this way, taking the add-on feature to be beliefs and desires that, in causing the bodily movements, constitute them as action. This gives neurophysiological explanation priority over rational explanation, and hence an account of why congruence obtains requires a theory about the nature of reasons for action that construes them as identical with, or determined by, neurophysiological causes.31 This assumes that the relevant bodily movements can be identified independently of whatever action they are involved in. This is possible for bodily movements that result from neurological defects, electric shock,

330

Determinism, Intentional Action, Bodily Movements

artificial stimulations of the brain, and so on and hence have only neurophysiological explanations. But intentionless bodily movements are unlike those involved in intentional action, for they are not the movements of healthy bodies and limbs but of diseased, degenerate, or disabled ones. Such movements are not essential to the human body or its limbs, for even if the latter never moved in intentionless ways, they could still move in the intentional ways essential to their nature. Explanation of intentionless movements is relevant to understanding the nature of action and intentional movements, but only in the indirect way in which explanation of an unhealthy (or a dead) body is relevant to understanding a healthy (or a living) body. An unhealthy body is a degenerate form of a healthy body; a healthy body is not an unhealthy body plus something else (nor is a living body a dead body plus something else). So intentionless bodily movements are a degenerate form of bodily movements and unlike non-intentional bodily movements, which are intentional under some description. The latter are the movements of a body and limbs whose nature it is to move intentionally: they are not intentionless movements plus something else. Nonintentional bodily movements involved in intentional action must be precisely those that are so involved, hence must be intentional under some description, and hence can be identified only by identifying intentional action. Anscombe argued (now 50 years ago) that it is not a proper aim of philosophy of action to specify the additional features that distinguish action from mere bodily movements: “We do not add anything attaching to the action at the time it is done by describing it as intentional. To call it intentional is to assign it to the class of intentional actions, and so to indicate that we should consider the question “Why?” relevant to it in the sense I have described.”32 The argument Anscombe attached to that claim is obscure, but the claim itself is clear enough to be defensible, and it is important. She meant by “we should consider the question ‘Why?’ relevant” that an action is intentional if done for a reason.33 Her overall claim is twofold. First, philosophy of action begins with neither mere bodily movements nor action in general but with intentional action, because action is essentially intentional: lesser forms of action are to be understood in terms of intentional action, which is not a species of action in general. Insofar as action is inadvertent, accidental, or by mistake, it is not intentional, but it is action only if it is intentional under some description: intentionality is essential to its being action. Even bodily movements not intentional under any description are conceptually dependent on intentional action and movement if they are the movements of fingers, hands, arms, legs, or other limbs: limbs are the parts of our body that we can move intentionally—whose nature is to be intentionally moved. We cannot intentionally move our fingernails, hearts, kidneys, or brains, which is why they are not limbs nor are their movements bodily movements in the usual sense.34 Movements that result

Frederick Stoutland

331

from neurological damage or artificial stimulation are bodily movements only if they are the movements of limbs whose essential nature is to move intentionally but whose movements are now, because of special circumstances, intentionless. Second, asserting an action is intentional if done for a reason is not to characterize it in terms of an additional feature (or property), because neither being intentional nor being done for a reason are features. An action is intentional under some descriptions and not under others, and done for a reason under some but not other descriptions; hence, if being intentional or being done for a reason were features, actions would both have and not have those features, which is a reductio of the claim. The crucial distinction for our purpose is not between intentional and intentionless bodily movements, for the latter are intentional under no description and hence are not movements that enable agents to act intentionally, but between the intentional and non-intentional movements that occur in acting intentionally, which is a distinction between bodily movements described in intentional and in non-intentional ways. But it is evident that to consider the relation between movements thus distinguished, we must look to where they intersect, and we cannot focus on that unless we begin with, and give priority to, rational explanation. There are, of course, contexts in which neurophysiological explanations are prior, for example, explaining why agents are incapable of certain actions. Rational explanation cannot explain such incapacities because explaining an agent’s action as intentional presupposes that the agent is capable of acting and not subject to various kinds of abnormalities. The incongruence that consists in one’s body and limbs not moving in the way required for action requires explanation from the logical space of laws. Neurophysiological explanations are also relevant to explaining how agents develop the abilities to act in various ways—to walk, climb, throw, type, play the piano, and so on—some of which develop naturally while others require careful training. Moreover, there are scientific explanations of why there are normal agents who have acquired the capacity for intentional action and whose bodies and limbs are by their nature capable of intentional movements. Given our post-Darwinian science, explaining why there are such agents is a task of evolutionary considerations that are in the logical space of laws: there are clear survival benefits to beings capable of action that can (often) be given a rational explanation. Of course, such beings might never have come to exist, in which case there would have been no intentional actions or bodily movements that could be explained in the logical space of reasons. Such an impoverished world illustrates another context in which neurophysiological explanations are basic, namely, what would have been explained had there been explainers can exist without normal agents. That is one way of expressing global supervenience of the intentional on the physical: destroy the

332

Determinism, Intentional Action, Bodily Movements

physicalistic and you destroy everything else, but not the reverse. Accepting that has no consequences for the nature and function of rational explanation.

9 Only now are we in a position to consider the question of the role of the rejection of determinism in an account of rational explanation of human action. I have argued that the question be altered to ask, not about the rejection of determinism, but about the rejection of the broader thesis of the completeness of physics. The claim of determinists that all things are fixed as a matter of natural laws is not rejected by indeterminists but modified to claim that all things are fixed as a matter of natural laws, either universal or probabilistic. As John Dupré puts it, the completeness of physics thesis asserts that “there is some quantitatively precise law governing the development of every situation,” where the law can be either universal (deterministic) or probabilistic (indeterministic).35 Including indeterministic laws rules out predicting the development of every situation, but leaves open the completeness of nomic explanation. I have argued that it is irrelevant to rational explanations per se whether one accepts the completeness of physics since they are not a matter of natural laws of either kind. If we deny that rational explanations are reducible to or determined by explanations in the logical space of laws, the completeness of physics thesis is directly relevant only to explanations in the latter, its claim being that all phenomena so describable can be explained as instances of mathematically precise laws of physics. The claim is not that all phenomena can be explained as instances of one or another of such laws, but rather that they can be explained as instances of vectorial combinations of them. While there is much that we do not know how to explain in that way, the thesis is that all phenomena are nevertheless governed by natural laws interacting beneath the surface. The completeness of physics thesis has been under sustained attack by philosophers of science in recent years for presenting an idealized picture of the physical sciences that confuses models of the phenomena with the actual workings of nature. Scientific models are abstractions, either theoretical structures or experimental set-ups constructed by investigators, to which natural laws may apply with great exactness. But these natural laws do not apply in any straightforward way to the workings of nature in real life, even if we add numerous ceteris paribus conditions. We must not assume, writes Nancy Cartwright, that “facts about behavior in highly structured, manufactured environments like a spark chamber ... have a special and privileged status [and] are exemplary of the rest, and the others be made to conform to them.”36 While the laws of nature are indispensable to explaining phenomena in the logical space of laws, opponents of the completeness of physics contend that

Frederick Stoutland

333

it is empirically unjustified—mere dogma—to assert that all such phenomena can be explained by them. John Dupré puts the point forcefully: Far from knowing that these laws [of Newtonian mechanics] are universally true, we know that they are generally false. The assumption that the laws of Newtonian mechanics are, in some sense, carrying on regardless under the overlay of increasingly many interfering and counteracting forces is not merely sheer speculation, but actually of dubious intelligibility. What are these laws supposed to be doing, given that the objects, subject to such diverse other influences, are not behaving in any sense in accord with them? Certainly this can hardly be a good empirical ground for the alleged universality of microphysical laws.37 This critique of the completeness of physics, although not relevant to an account of rational explanation per se, is relevant to the account of the relation between rational and neuroscientific explanation that emerged in my discussion of congruence. A neuroscientific explanation of the bodily movements involved in an agent’s acting must begin with the agent’s having intentionally moved his body, then describe in neuroscientific terms the resulting movements, and then proceed backwards to the muscle contractions and neural processes that explain the movements thus described. It is a criterion for which muscle contractions and neural processes explain the agent’s behavior (in the logical space of laws) that they yield behavior that is also explained (in the logical space of reasons) as action done for a reason. In this context neuroscientific explanations necessarily conform to rational explanations. The completeness of physics thesis is relevant at this point. If all things are fixed by natural law, then neurophysiological explanations of an agent’s behavior will not necessarily conform to rational explanations since the former will appeal to factors that have been fixed from an indefinitely distant past. It may be logically possible that the factors so fixed will yield behavior that is also (describable as) the agent’s intentional bodily movements, but it cannot be a criterion for which neural processes explain an agent’s behavior that they yield behavior that is also intelligible as his intentional action. Given the completeness of physics, rational explanations will not be more basic than neuroscientific ones in the context of explaining intentional action. If all things are fixed by natural law, then they have always been fixed, not in a wholly deterministic sense, but nevertheless so as to give explanation in terms of natural laws absolute priority over explanations in terms of reasons. The problem is not the fixing per se since neuroscientific explanations of bodily movements will appeal to neural processes that fix them as described neuroscientifically. The problem is the regress into the indefinite past, which rules out what the fact of congruence clearly shows, namely, that the criterion for an explanation that renders non-intentional bodily movement intelligible must begin with one that renders intentional

334

Determinism, Intentional Action, Bodily Movements

action and movements intelligible. This gives us reason to reject the completeness of physics thesis, one which fits nicely with the criticism it has received from philosophers of science who appeal to the procedures and results of physics itself.38 This rejection is not grounded in reflection on our capacity to act freely. Acting freely requires acting intentionally but acting intentionally need not be acting freely. Whether a particular act is free is a question in the logical space of reasons not in the logical space of laws. An adequate account of intentional action, however, requires considering the relation between explanations of human action in both types of logical space, and it is reflection on that relation that supports the rejection of both determinism and the completeness of physics.39

Notes 1. Carl Hoefer, “Causal Determinism,” in Stanford Online Encyclopedia of Philosophy. 2. Sellar’s distinction is in his “Empiricism and the Philosophy of Mind,” first published in Herbert Feigl and Michael Scriven, eds, Minnesota Studies in the Philosophy of Science (Minneapolis: University of Minnesota Press, 1956). The suggestion that the distinction be altered is John McDowell’s in Mind and World (Harvard: Harvard University Press, 1994, p. 70f.). It changes Sellers’ intent in an important way because he thought of the logical space of reasons as having normative import as opposed to having causal efficacy. 3. I do not mean by a first-person explanation one that appeals to an agent’s introspective knowledge of herself. I mean one that uses concepts of how the world presents itself to her, which is not something she knows through introspection and which others may, through interpretation, know as well as she does. 4. The argument I would give that they do not so figure is that ascribing psychological states to an agent requires that the agent be at least minimally rational, which entails giving up generalizations relating her actions, reasons, and psychological states to each other if that is required to preserve her normative status as a rational agent, and that is something we could not do if those generalizations were natural laws. Cf. Donald Davidson, Essays on Actions and Events (Oxford: Clarendon Press, 1980), p. 222. 5. This oversimplifies the situation: corrections have to be made periodically in calculating the orbits of the planets because we are unable to make the calculations required by the interaction of three (or more) bodies. 6. Or not acting in a certain way; I will ignore negative reasons. 7. There is much more to be said about the nature of reasons and their relation to psychological states, for instance, about the difference between acting because of a reason and acting because of what one takes to be a reason. But my views on that issue are largely irrelevant to this essay, and I will not discuss them here. Persons who are interested might see my “Responsive Action and the Belief-Desire Model,” Graz Philosophischen Studien, 61 (2001), pp. 83–106; or “Reasons for Action and Psychological States” (in Anton Leist, ed., Action in Context (de Gruyter/Mouton, forthcoming). 8. Cf. Davidson, Essays on Actions and Events, p. 242: “Laws (and nomological explanations) do not deal directly with events, but rather with events as described in one way or another.”

Frederick Stoutland

335

9. This may be why he generally used “cause” to denote both causal explanations and relations and why it is not always evident which he meant. 10. This also means that Davidson’s view is not very different from von Wright’s. Von Wright always used the term “causal” in a nomological sense and argued that actions explanations were not causal in that sense. Davidson agreed with him on that negative point; the differences between their positive views of action explanation are primarily in language, not in substance. I have developed these claims in “Intentionalists and Davidson on Rational Explanations,” in G. Meggle, ed., Actions, Norms and Values: Discussions with Georg Henrik von Wright (Walter de Gruyter, 1999), pp. 191–208 and in “Interpreting Davidson on Intentional Action” forthcoming in a volume on Davidson edited by Jeff Malpas. 11. Anscombe wrote: ‘ “Sufficient condition” sounds like “enough,” and one certainly can ask: “May there not be enough to have made something happen—and yet it not have happened?” ’ [‘Under a Description’ in Metaphysics and the Philosophy of Mind, (Minneapolis: University of Minnesota Press, 1981), p. 135.] 12. Such explanations have traditionally been called ‘causal’ though Davidson, while speaking of causal laws—he means what I have called ‘natural laws’—as central to natural science, contends that causal explanations drop out of a discipline as it becomes more of an advanced science. He says that precisely because he thinks that rational explanations must do without precise generalizations, and hence without necessitating explanations. I will not conform to Davidson’s terminology but will instead speak of different kinds of causal explanation, depending on whether they belong to the space of laws or of reasons. 13. My topic here is establishing the reason for which an agent acted, which is not the same as knowing the reason: the latter is possible simply by observing, without interpreting, her action. 14. Randolph Clarke, “Incompatibilist (Non-deterministic) Theories of Free Will,” Stanford Online Encyclopedia of Philosophy. This useful article is a main source for the view I discuss in this section. Because philosophers who hold the view differ among themselves, my criticism may be too narrowly focused, but it nevertheless applies in some form to all versions of the theory. 15. For criticism of the thesis see Nancy Cartwright, The Dappled World (Cambridge: Cambridge University Press, 1999) or John Dupré, Human Nature and the Limits of Science (Oxford: Oxford University Press, 2001). For illuminating discussion of this criticism see Paul Needham, Law and Order (Stockholm: Stockholm University Department of Philosophy, 2003). 16. There are intentional actions that are done for no reason, for example, idly whistling or humming, or simply strolling along the beach. But these are borderline intentional, and they are intelligible because they could have been done for a reason and because most intentional actions are done for a reason. Anscombe says that if it is asked why someone is acting, where “Why?” means “For what reason?,” the reply, “For no reason,” is not a rejection of the question. “The question is not refused application because the answer to it says that there is no reason, any more than the question how much money I have in my pocket is refused application by the answer ‘none’.” Intention (Oxford: Basil Blackwell, 1957), p. 25. 17. Jaegwon Kim, Supervenience and Mind (Cambridge: Cambridge University Press, 1993), p. xiii. Norman Malcolm defended a similar claim in “The Conceivability of Mechanism,” in Philosophical Review, 77 (1968), 45–72. 18. Neither is complete, of course, in giving an explanation of the explanation and so on, but that is not relevant to the argument.

336 Determinism, Intentional Action, Bodily Movements 19. It has been defended more recently by, for example, Robert Audi in Action, Intention and Reason (Ithaca: Cornell University Press, 1993, p. 13f. A variant is developed by Scott Sehon in Teleological Realism: Mind, Agency and Explanation (Massachusetts: MIT Press, 2005). 20. “Physicalism” in Inference, Explanation, and Other Frustrations, ed. by John Earman (California: University of California Press, 1992), 285. 21. I have discussed this issue in my “Review of Wm. Child, Causality, Interpretation, and the Mind,” Philosophy and Phenomenological Research (Sept., 1998), 711–715. 22. These are equivalent minimal descriptions of an action. The description “Jacob walked” does not describe his action either in terms of what resulted in it or what it resulted in. 23. Donald Davidson, “Agency” in Essays on Actions and Events, p. 59. 24. Davidson later admitted that what he wrote here was misleading. My view is that the standard way of reading this passage badly distorts Davidson’s account of action, a view I defend in “Interpreting Davidson’s Philosophy of Action.” 25. Moreover, many of them are more than bodily movement descriptions in that they entail features of the environment beyond the agent’s skin. Walking or running entails a surface on which to walk or run (‘space-walking’ is metaphorical; it’s just floating), jumping requires a base from which to jump, climbing requires something to climb, swimming entails a liquid in which to swim, and so on. 26. “Metaphysics and the Philosophy of Mind” p. 209. 27. Donald Davidson, “Aristotle’s Action” in Truth, Language and History (Oxford: Clarendon Press, 2005), p. 290. 28. “Problems in the Explanation of Action” in Davidson’s Problems of Rationality (Oxford: Clarendon Press, 2004), p. 114. 29. Intention, p. 86. 30. Consider an analogy. An explanation of why an iron sphere of a certain mass dropped from a certain distance hits the ground with a certain force is not an explanation of why a cannon ball from the Civil War that belongs to the Smithsonian dropped from a certain distance hits the ground with a certain force. But if those are two descriptions of the same event, then what is explained by the first—the iron sphere hit the ground with a certain force—must be congruent with what is explained by the second—the Civil War cannon ball hit the ground with a certain force. 31. Fred Dretske’s admirable Explaining Behavior (Massachusetts: MIT Press, 1988) is a nice example. He writes (p. ix), “If, then my body, and I are not to march off in different directions, we must suppose that my reason for going into the kitchen—to get a drink—is, or is intimately related to, those events in my central nervous system that cause my limbs to move so as to bring me into the kitchen. My reasons, my beliefs, desires, purposes, and intentions are, indeed they must be—the cause of my body’s movements.” 32. Intention, p. 28. 33. See note 16 for a qualification. 34. St. Augustine held that before the fall of man, the penis was a limb whose nature was to move intentionally. Cf. The City of God, Book XIV, Chapter 17. 35. Human Nature and the Limits of Science, p. 157. 36. ‘Fundamentalism vs. the Patchwork of Laws’ in Proceedings of the Aristotelian Society, New Series, xcix, p. 281. 37. Human Nature and the Limits of Science, p. 166.

Frederick Stoutland

337

38. Cf Anscombe, “Under a description,” p. 208: “Since most human action consists in bodily movements it couldn’t be that the actions were free (not necessitated) though the movements were determined (necessitated).’ 39. This essay benefited greatly from Anton Ford’s “Action Unqualified” (not yet published) and from correspondence with him, which gave me the notion of bodily movements as intentional and essential to the human body and limbs, and clarified, among other things, my understanding of the centrality of intentional action. It was improved by conversations with Peter Myrdal and Paul Needham.

18 Free Agency, Causation and Action Explanation E. J. Lowe

Part of our intuitive sense of freedom is our commonsense conviction that we, as free agents, can intervene in the natural world and make a difference to the historical course of events. The supposedly ‘scientific’ picture of the natural world painted by some physicalist philosophers appears to threaten this conviction in various ways. Often, the threat is diagnosed simply in terms of a conflict between the ‘libertarian’ conception of the will and the supposed implication of the scientific world view that, in the natural world, everything happens as a result of some mixture of causality and chance. Such a diagnosis makes libertarianism seem an incoherent position, impossibly seeking some middle ground between lawful causal determination and mere randomness. The problem may be posed in the form of a dilemma. All of our actions, including our supposedly free ones, are (it may be said) just events, and all events are either causally determined by prior events or else are chance occurrences (though their chances may be fixed by prior events). Either way, there is no room for the notion that we are the authors of our actions in any sense that would suit the libertarian. For the libertarian wants to say that we are the authors of our actions in a way which renders those actions neither mere chance occurrences nor events that are wholly causally determined by prior events. The libertarian wants to say that we sometimes make a difference to the historical course of events – that is, make the actual course of events different from what it would have been without our interventions. But this requires our ‘interventions’ not to be seen merely as certain events amongst others in the natural course of events. Once we see our own ‘interventions’ in the latter way, they cease to qualify as ‘interventions’ in the appropriate sense, but become instead just part of the natural order which, as free agents, we conceive ourselves as being able to affect through our free decisions. This is part of a phenomenon that is sometimes described, quite aptly, as the ‘disappearance of the agent’.1 We begin to see ourselves not as independent authors of our own actions who by acting freely can intervene to affect the course of nature, but just as part of the natural stream of events, swept up by 338

E. J. Lowe

339

and even constituted by that stream, whose flow is consequential upon factors independent of us, in the shape of causal laws and chance. Physicalists may say that the conception of ourselves that we thereby give up is one that is well lost, because it is a pre-scientific and ultimately incoherent one. There is simply no room between chance and causal necessity, they may say, for such a conception of ourselves to find a foothold. The rejected conception belongs to an animistic world view which sees all change in the natural world as brought about by spirits, demons or gods. According to that world view, nothing ever happens either by chance or by mindless causal necessity. What the libertarian hankers after is an inconsistent mixture of this prescientific conception and the modern naturalistic world view, so that animism is preserved in the exclusive domain of human action while chance and causal laws reign everywhere else. But, so the charge goes, there is no consistent way of combining these views because they clash at the interface between human agency and the natural world. Either the natural world must be reabsorbed into the world of free agents, as in animism, or else human agency must be absorbed into the realm of chance and causal necessity. The latter path is the only one that is realistically now open to us and it means abandoning our pre-scientific self-conception. If we want to retain anything that was valuable in that self-conception, we must reinterpret ourselves in new ways by, for example, developing a convincing compatibilist conception of freedom and a naturalistic theory of reasons and rationality. We must accept, for instance, the dictum that ‘reasons are causes’, abandoning the idea that there is a fundamental distinction to be drawn between causal and rational explanation. Likewise, we must abandon ‘final’ causes and teleological explanation, except in some suitably naturalized form. The agents that will ‘disappear’ under this reconceptualization are merely mythical in any case. Human kind will finally come of age when it recognizes itself as being merely part of the natural order, not set over and against it as a species of demi-gods. However, matters are not as simple as this stark picture represents them as being. Our conceptions of reason and rationality are not so easily negotiable as the picture suggests. In fact, the image that comes to mind when contemplating the naturalistic philosopher attempting to construct that picture is that of the hapless home-decorator who ‘paints himself into a corner’ or the inept forester who saws off the branch that he is sitting upon. Or, again, that of the artist who, when asked to paint a portrait of the world in his vicinity (‘The World As I See It’), inevitably omits something of himself from the scene. I am not going to attempt to prove here that the naturalistic world view, as I have been calling it, is rationally unacceptable, although I do believe that this is so: that is, that one could never be confronted with empirical evidence which made it rational for one to accept that that picture is correct. Rather, I want to challenge the idea that the libertarian conception of ourselves is incapable of being accommodated within a world view

340 Free Agency, Causation and Action Explanation

which incorporates a fundamental distinction between free agency and natural causality. In other words, I want to argue that we are not in fact faced with a stark choice between a return to ‘animism’ and a complete acceptance of the ‘scientific’ world picture as painted by most physicalists, in which libertarian free will has no place. Other philosophers have also attempted to do this, of course, and have generally failed to convince the physicalists. One reason for the failure of some previous attempts, I believe, is that they already accepted too much of the ‘scientific’ world view before trying to rein it in where human agency is concerned. The key, I think, lies in the very notion of causality itself and the presumption that all causation is fundamentally event causation. Once it is assumed that all causation is a matter of one or more events causing another, it is difficult to resist the thought that human decisions and actions are, after all, just events and as such woven into the universal web of event causation, the only alternative being that they are random outcomes of chance. I should make it clear that my object now is not to defend the distinction between ‘agent causation’ and ‘event causation’, as this would classically be understood. Classic agent causalists hold that an irreducibly distinct species of causation comes into play uniquely in the sphere of intentional agency, while event causation reigns elsewhere.2 According to the classic agent causalist, when an intelligent agent, A, intentionally performs a so-called ‘basic’ action, such as raising his arm, A is literally the cause of a certain occurrence – in this case, a rising of his arm – in a distinctive sense in which no event is or could be its cause and in a way which never happens in the world of inanimate objects.3 In particular, the classic agent causalist will deny that A’s causing the arm-rising consists in, or even in any way involves, the causation of the arm-rising by some action or event in which A participates or of which A is the subject. So, for example, it will be denied that the arm-rising is caused by a volition or act of will of A’s, or by the onset of some combination of belief and desire attributable to A. This is supposed to rule out as fundamentally misconceived any question of the sort ‘How, or by doing what, did A cause the arm-rising?’. We are supposed to accept that the agent, A, was the (or at least a) cause of the arm-rising in a sense which precludes there being anything about A in virtue of which A caused that occurrence. Moreover, we are to suppose that this sort of thing is uniquely distinctive of intelligent agency, so that whenever, in the realm of inanimate objects, we speak of such an object being the ‘cause’ of an event, we can take this as being elliptical for some statement to the effect that some event ‘involving’ that object caused the effect in question, or that the object caused that effect by behaving in a certain way. Classic agent causation is regarded by its naturalistic opponents as being utterly mysterious – and I sympathize with this attitude to a large extent. How can a person simply as such be a ‘cause’ of some event? How, for instance, would this explain why the event occurred when it did – for the

E. J. Lowe

341

same person could cause different events at different times and yet in each case we are supposed to believe that the event just has the same cause, namely, that person. There are ways in which the classic agent causalist can attempt to answer this sort of objection – for example, by adverting to a time-reference in the agent’s reasons for acting, or by appealing to other events as contributory causes, or causal preconditions, of the given effect. But such strategies are not, I think, particularly convincing. My own view is that classic agent causalism is an unstable half-way house. It rightly sees that the notion of the causation of an event by something in the category of substance – in this case, a person – is an important key to understanding human agency, but mistakenly treats this as something special and distinct from the sort of causation that goes on in the inanimate world, where event causation supposedly reigns supreme. My challenge to the event causalist presumptions of modern physicalism go far deeper. I think that all causation is fundamentally substance causation – the causation of events by substances – even in the world of inanimate objects. This is not a revival of ‘animism’, of course. I am not proposing that everything that happens does so as a consequence of intentional agency. Rather, my point is that substances, not events, are the engines of all change in the world. It is substances, not events, that have causal powers and liabilities – the powers to act and be acted upon in various ways – and causation is fundamentally a matter of substances exercising their causal powers to act upon other substances possessing suitable causal liabilities. So, substances are causes, in the fundamental sense of ‘cause’: they are the entities that do all the causing in the world. Causing is acting, and it is substances that act. More particularly, causing is acting upon some substance so as to make that substance in turn act in some way. This happens all the time in the inanimate world. A magnetized piece of iron acts upon some nearby iron filings to make them move towards it: it ‘attracts’ them. Some sulphuric acid acts upon some bits of copper to make them dissolve and transform their constituent atoms into copper ions: it ‘oxidizes’ them. In all such cases, a substance brings about its effects by acting in some distinctive way. The same substance may act in different ways in different circumstances to bring about different effects. For instance, the very same piece of magnetized iron which acts upon the iron filings to make them move towards it may also act in a different way to extend a steel spring from which it is suspended. In the one case, the piece of iron exerts its power of magnetic attraction, while in the other it exerts its weight – this being a power that it has in virtue of the force of gravitational attraction that it and the earth exert upon each other, owing to their respective masses and the distance between their centres of mass. To say, as I do, that all causation is fundamentally substance causation is not necessarily to rule out all talk of event causation as entirely erroneous or misconceived: it is just to say that such talk is never ontologically fundamental.

342

Free Agency, Causation and Action Explanation

I am not, after all, denying the existence of events. Events occur when substances act or are acted upon in certain ways: indeed, events just consist in the doings of substances.4 A death is the dying of a living substance, such as an animal or plant. A marriage is a legal transaction between two human beings. A collision is the coming into contact of two material substances. We can make true statements of event causation, as when we say that a certain collision caused a certain death – for instance, that the collision of the car and the pedestrian caused the pedestrian’s death. What this means, however, is just that the car, by coming forcefully into contact with the pedestrian, caused the pedestrian to suffer lethal injury. The source of all change in the world lies in substances, in virtue of their causal powers and liabilities, which they characteristically exercise or manifest when they enter into suitable relationships with one another. But philosophers may be misled by event-causation talk into supposing that causation is fundamentally a relation between events – that, ultimately, it is events that bring about changes (i.e. other events). This is to invest events themselves with causal powers. But events as such are utterly powerless. They are mere changes in things and not the source of those changes. This is why a pure event ontology would be an ontology without causation in any serious sense. It is unsurprising, indeed, that idealists and phenomenalists often explicitly acknowledge this – that, for instance, Berkeley thinks that ideas really cause nothing, but merely succeed and are succeeded by other ideas in certain regular patterns of occurrence (although, of course, he also thinks that spiritual substances – ourselves and God – are truly causes of ideas). This also helps to explain why it is that attempts to ‘analyse’ event causation – for example, in terms of counterfactual dependence amongst events – have so signally failed, generating epicycles upon epicycles in order to overcome counterexamples to proposed analyses.5 Precisely because causation is not, fundamentally, a relation amongst events, but a family of transactions between substances, attempts to analyse it as such unsurprisingly fail to accommodate all of our pre-theoretical intuitions concerning causality. I have spoken of substance causation as being fundamental, and this requires me to say something about the notion of a substance. By a ‘substance’ – or, to be more precise, an ‘individual substance’ – I mean only what philosophers have traditionally meant by this term: a substance is an ontologically independent entity that bears properties, stands in relations to other substances, persists through time and undergoes qualitative change over time. Most importantly, for our purposes, substances possess causal powers and liabilities. These are species of disposition. I shall not offer here a general theory of dispositions.6 It does not much matter to me for present purposes whether we categorize dispositions and thus causal powers as being ‘properties’ of substances, or whether we say, as some philosophers do, that properties ‘confer’ causal powers upon substances which possess those properties. It suffices that we all instantly recognize many clear examples of causal powers

E. J. Lowe

343

and liabilities, such as magnetism and solubility. The causal powers and liabilities of substances are evidently intimately connected with the natures of those substances, that is, with what kinds of substances they are and how they are constituted. The magnetic power of a piece of iron arises from the way in which the atoms making it up are aligned. The oxidizing power of a quantity of sulphuric acid arises from the orbital electron structure of its constituent ions. But some powers may well be fundamental, such as the power of an electron to repel another electron or to attract a proton. This is because an electron may well be a simple substance, with no constituent structure. But, of course, almost all of the macroscopic substances with which we are empirically acquainted are certainly not simple substances, but highly complex ones. Living organisms provide an excellent example. These have characteristic causal powers and liabilities which arise from their inner natures and constitutions. For example, many plants have the power to turn their leaves towards the sun, to convert sunlight into stored energy, and to draw water from the soil. I make these obvious points just to remind us that talk of causal powers and liabilities is ubiquitous in scientific descriptions of the workings of the natural world. It is at best a philosopher’s myth to suppose that the causal language of science is the language of event causation – a myth born, perhaps, of positivism, with its presumption that science is only concerned with recording observable events and noticing their patterns of recurrence, in order to facilitate reliable prediction. Real science is concerned, rather, with revealing the causal mechanisms underlying and explaining recurrent patterns of events. As I remarked earlier, I see human agency as involving substance causation in the foregoing sense. This is because I see human agents – that is, human persons – as being, themselves, substances of a distinctive kind. One view that I am emphatically opposing here is the ‘Humean’ view of persons or selves as being mere ‘bundles of perceptions’, which is to regard persons as being constituted by mental events. If a person were no more than a collection or series of mental events – thoughts, feelings, urges and so forth – then, because events are, as I have remarked, utterly powerless, so too would persons be. But we certainly conceive of ourselves as being ‘substances’, in the sense explained earlier: we regard ourselves as being bearers of properties, as persisting through change, and, most importantly, as possessing distinctive causal powers and liabilities. Different kinds of substance are distinguished not least by their characteristic existence and identity conditions (under which we may subsume their so-called persistence conditions) – and persons constitute a distinctive kind of substance for this reason, or so I maintain. There are some philosophers, who characterize their position as ‘animalism’, according to whom human persons are just human animals.7 (This position, of course, is quite distinct from the ‘animism’ discussed earlier.) More precisely, they take ‘person’ to be a term which characterizes a human animal only during certain phases of its

344

Free Agency, Causation and Action Explanation

existence – namely, those during which the animal is capable of conscious thought, reason and self-reflection. According to this view, I existed before I was a person and may well continue to exist after I cease to be one. I reject this view as being incompatible with the characteristic existence and identity conditions of persons. Very plausibly, persons can, in principle, persist through changes which cannot be persisted through by animals – and vice versa. Animals are essentially biological beings, but persons are not: persons are essentially psychological beings. In this respect, at least, I agree with the ‘Humean’ view and, indeed, with Locke’s conception of persons. But against Locke and Hume I insist, with Descartes, that the person who I am is a psychological substance, not a mere collection of psychological events happening in or to an animal body. On the other hand, I do not agree with Descartes that I am an essentially immaterial substance – that I occupy no space and have no spatial dimensions and no mass. Rather, I hold that I am a psychological substance coinciding with, but numerically distinct from, my animal body. A position quite close to mine holds that I am constituted by my body, rather as a bronze statue is constituted by, but not identical with, the lump of bronze whose particles compose both the statue and the lump at the time of their coincidence.8 This is not my view, because I agree with Descartes that I am a simple substance, composed of no proper parts. In fact, I think that persons and perhaps some other species of conscious, embodied beings, are the only macroscopic simple substances. The relationship of embodiment in which a person stands to his or her body I regard as sui generis, being neither identity nor constitution and yet implying, as those relations also do, spatial coincidence. I do not expect many philosophers to be immediately convinced by all of these claims, although I think that strong arguments can be marshalled directly in their favour.9 I also think that this conception of the ontology of persons enables us to make more sense of human agency – and to demonstrate this will be part of my object in what follows. If persons are psychological substances, distinct if not separable from their biological bodies, then one should expect them to have causal powers and liabilities that are also distinct from those of their bodies. And indeed I think that this is so. Central amongst our distinctive psychological powers are our powers of perception, thought, reason and will. It is I who perceive, think, reason and will, not my body nor any distinguished part of it, such as my brain or central nervous system. Of course, animalists will vigorously deny this. They will charge my sort of substance dualism – for that is what it is – with harbouring absurdities and contradictions. For instance, they will urge that all it takes to create a thinking being such as myself is the creation of a biological body organized in the right sort of way, with a suitably functioning central nervous system. Hence, they will say, all of my psychological features must supervene upon physical features of my biological body, whence it is absurd to deny that my body itself possesses

E. J. Lowe

345

those very psychological features: it perceives, thinks, reasons and wills, whenever I do. But then, if we deny that I am identical with my body, we must say that two distinct beings are simultaneously the subjects of the same or exactly similar perceptions, thoughts, reasonings and volitions. And then how do I know which of these beings I am? This is the notorious ‘multiple thinkers’ problem.10 However, this sort of objection is entirely spurious, in my view. What it ignores is the quite general point that one can only attribute dispositions, including causal powers and liabilities, to substances which possess suitable identity conditions. And biological substances, I would claim, do not possess identity conditions suitable for the attribution to them of psychological powers. The general point can be brought out in the following way. When we attribute a disposition to a substance, we imply something about how it will behave in certain hypothetical or future circumstances, and this requires that the attribution be sensitive to the substance’s identity conditions, including its persistence conditions – for it is these that determine which substance, if any, can be identified with the subject of the attribution in the hypothetical or future circumstances in question. For instance, when I say of a rubber ball that it is ‘elastic’, I imply, amongst other things, that it will bounce if dropped from a height on to a hard floor – and this presupposes an account of the identity conditions of such a ball, licensing the assumption that this very entity could persist through a change in which it first falls through the earth’s gravitational field and then undergoes a change of shape and motion upon impact with a solid object. Similarly, then, when I attribute a power of thought or reason to a substance, I must attribute it to a substance whose identity conditions are such that, of its very nature, it can persist identically through the sorts of changes that are constitutively involved in processes of thinking and reasoning. Thinking and reasoning take time, and it makes no sense to suppose that a single train of thought or reasoning might have different substances as its subject at different times – as though one subject could ‘begin’ a thought and a quite different subject ‘end’ it. Equally, it makes no sense to suppose that different ‘parts’ of a single thought might have different substances as their subject at one and the same time. For this sort of reason, biological substances are ill-suited by their identity conditions to qualify as psychological subjects. Largely, this is because of their highly composite and mutable nature, as I shall now try to make clear. (Recall that, on my own view, persons are simple substances.) No one doubts, of course, that the brain and central nervous system of an embodied human person are the sites of numerous biological processes which are somehow intimately connected with the person’s psychological processes of thought and reasoning. But the biological processes in question cannot simply be, or in any way constitute, those psychological processes of thought and reasoning. Why not? Because neither the organism as a whole nor any biologically distinguishable part of it, such as the brain or any

346

Free Agency, Causation and Action Explanation

specific region of the brain, can be singled out at any given time, or over time, as being uniquely related in an appropriate way to any particular psychological state or process, much less to a multiplicity of such states and processes all of which are assignable to a single subject. Suppose, for instance, that I am now wondering about the state of the weather in Vienna. I could engage in that very same train of thought whether or not my body lacked some minor organic parts here or there – for instance, whether or not it lacked my right ear or my left little finger. So my body as a whole is no better qualified than my body minus my right ear or my body minus my left little finger as being, in virtue of the biological processes that it is undergoing, the subject of my train of thought about the state of the weather in Vienna. But we cannot say that all three of these biological entities (plus indefinitely many more) are having those same thoughts, on pain of falling into the very ‘multiple thinkers’ problem of which my sort of substance dualism itself stands falsely accused. We should conclude, then, that none of these biological entities qualifies as the subject of my train of thought in virtue of the biological processes that it is undergoing. And hence we should further conclude that the psychological process in question – my train of thought about the state of the weather in Vienna – is neither identical with nor constituted by any of the biological processes being undergone by those biological entities: for if it were, then, contrary to what we have already just concluded, one or more of those biological entities would, after all, qualify as being the subject of my train of thought about the state of the weather in Vienna. Exactly the same form of argument can be run with my brain, or any biologically distinguishable region of it, being substituted for my whole body in the foregoing version of the argument. In place of minor parts of my body, such as my right ear or left little finger, we need merely substitute minor parts of my brain or of the selected region of it, such as a few cells here or there. No doubt, of course, I couldn’t have my train of thought about the state of the weather in Vienna whether or not my brain was lacking some quite large part. But that is quite irrelevant and obviously does not imply that precisely that part of my brain qualifies just as well as I do as being the subject of the train of thought in question. Indeed, various different parts of my brain will doubtless be necessary for my capacity to think various different thoughts, whether at the same or at different times, but these thoughts will nonetheless all be thoughts of a single subject – me. Clearly, there can be no identifiable subset of organic parts of my brain of which it can be said that that subset is uniquely necessary for the having of all and only my thoughts. The lesson of all this is, as I suggested earlier, that the unity of thought and reasoning requires a unity of their psychological subject which is inconsistent with the identity conditions of biological substances, given the enormous compositional complexity and mutability of the latter.

E. J. Lowe

347

Suppose we accept, then, for the foregoing reasons, the claim that psychological substances – that is, conscious, thinking beings, including human persons – are distinct if not separable from their physical bodies and possessed of distinctive and irreducible psychological powers, including the central powers of perception, thought, reason and will. Some philosophers will persist in purporting to find this view mysterious, but that is because the only sort of account of persons and their psychological powers that they would not find ‘mysterious’ is one that is couched in reductive biological terms – and such an account, it seems clear, is simply not to be had. We just have to accept that there are aspects of nature which defy ‘explanation’ in the reductive terms that alone seem to make such philosophers feel comfortable. A datum that is not to be denied is that we ourselves exist and have the psychological powers that are definitive of ourselves as persons. This is simply an immovable framework assumption of rational debate in this or any other area of philosophy or science, for to deny one’s own existence or rationality is automatically to exclude oneself from any such debate. What remains is to see how, precisely, the conclusions that we have so far reached bear upon the issues in the philosophy of action with which we began. For this purpose, the psychological power that we must focus on is the power of will, or volition. Talk of volition in the philosophy of action is still somewhat unfashionable.11 This is partly because there is a broad consensus that the mental causes of human action are combinations of belief and desire or, more precisely – since such items are taken to be mental states rather than mental events – the ‘onsets’ of combinations of belief and desire.12 Adherents of this view are, of course, event causalists and hostile to classic agent causalism. Volitionism has remained a minority view in the philosophy of action because it is opposed both by adherents of the belief/desire approach and by classic agent causalists. One thing that is wrong with the belief/desire approach, quite apart from its assumption of the sovereignty of event causation, is that it makes no room at all for the executive element in intentional action. It effectively treats all intentional actions as mere happenings with distinctive mental causes. On this approach, the only difference between my arm’s rising as a result of a muscular spasm and its rising because I raised it intentionally is that, in the latter case, the arm rising has amongst its causes certain mental events – the onsets of certain beliefs and desires – where the beliefs and desires in question supposedly also constitute my ‘reasons’ for raising my arm on that occasion. But this analysis precisely leaves out what is distinctively active about my raising my arm and ‘passive’ about my arm’s rising as a result of a muscular spasm. In the former case, I did something to cause my arm to rise, but not so in the latter. What did I do? The classic agent causalist, while agreeing that there is something distinctively ‘active’ about my raising my arm, will decline to answer this question: such a theorist will contend that, because the action is a

348

Free Agency, Causation and Action Explanation

so-called ‘basic’ one, there is nothing more to be said than that I simply caused my arm to rise, as a primitive case of agent causation. However, not only does this make little sense metaphysically, it is false to the phenomenology of intentional action. When I raise my arm, I feel that I am doing something to cause it to rise – something that I could characterize as trying to raise it. I could do this same thing on another occasion and fail to raise my arm – for instance, if my arm were strapped down, or if I were suddenly to be struck with paralysis in my arm. What is this thing that I do? It is a mental act – an exercise of a distinctive mental power of an ‘executive’ character. The power in question is what philosophers have traditionally called ‘the will’ and our exercises of that power are our ‘volitions’ or ‘acts of will’.13 Acts of will are events, for they are the doings of substances, namely, of persons. But they are not mere happenings, in the sense that the falling of a stone is a mere happening. We can describe what the stone is ‘doing’ by saying that it is falling, but a ‘doing’ in this sense is not distinctively ‘active’ in the way that exercises of the will are. What is it, then, about exercises of the will that requires us to say that they are ‘active’ in a way that the mere falling of a stone is not? When it falls, a stone is manifesting – and in this sense exercising – one of its causal powers: the power in question is its weight, which, as we noted earlier, enables a stone to extend a steel spring from which it is suspended. Such a power is ‘passive’, however, because the stone has no option but to manifest the power in some appropriate way whenever suitable circumstances arise. When the stone is suspended from a steel spring, it inevitably manifests its weight by extending the spring, unless some other object prevents it from doing so by exercising a contrary power – as, for example, a strong magnet might do by restraining the spring from extending and so counteracting the weight of the stone. The will, however, is not like this. The way in which it is exercised is not determined by external objects and circumstances. Of course, some philosophers will try to persuade us otherwise. They will say, perhaps, that volitions, or exercises of the will, assuming them to occur, must all be caused to occur by prior events, such as the very onsets of belief and desire of which we have already spoken. That, indeed, is one reason why such philosophers may urge that it is unnecessary to invoke volitions in the analysis of intentional action, because (so they will say) volitions are mere idle wheels between the real causal antecedents of intentional action – the onsets of belief and desire – and the ensuing actions. But to suppose that exercises of the will are causally determined, even by the onsets of belief and desire, is precisely to deny their essentially ‘active’ character. Volitions so conceived would indeed be mere idle wheels in the aetiology of action. What is distinctive of the will as a power is that its exercises are precisely not exogenously determined by exterior causes. In willing we initiate trains of action, whereas the ‘passive’ characterization of the will now under discussion represents willing as being at best a mere passing on

E. J. Lowe

349

of causal influence from prior causes, rather as one billiard ball inherits its momentum from another that collides with it. Why, then, do so many philosophers find the belief/desire analysis convincing? Partly, no doubt, because they recognize that it fits in so well with the presumed sovereignty of event causation. But another important factor is that it may seem that unless we accord beliefs and desires the sort of causal role just described, volition and action will appear to be blind and unmotivated. When we act intentionally, our beliefs and desires must certainly play some role in explaining why we act, for to deny this would be to represent all of our actions as being done merely arbitrarily, or on a whim. Sometimes, indeed, I do act merely on a whim: I kick a stone, perhaps, as I am walking along, and do this ‘for no reason at all’, although not merely unintentionally or accidentally. Explaining my action as one that I did ‘just because I felt like doing it’ scarcely seems to be much of an explanation at all, and certainly not one that we should wish to extend to all of our intentional actions. Generally, the contents of one’s beliefs and desires constitute one’s reasons for acting intentionally in the ways one does. When one acts rationally, one acts ‘in the light of’ one’s beliefs and desires. From this it may seem but a small step to say that those beliefs and desires – or their ‘onsets’ – are causes of one’s actions and explain one’s actions causally, while at the same time ‘rationalizing’ one’s actions. But this seemingly small step is, I believe, an erroneous one of enormous magnitude. I was careful to say, a moment ago, only that the contents of one’s beliefs and desires constitute one’s reasons for acting intentionally in the ways one does: I emphatically did not say that one’s beliefs and desires themselves constitute those reasons. Suppose that I decide to raise my arm so as to catch the speaker’s eye at a lecture, with the aim of asking a question, and exercise my will accordingly by endeavouring, successfully, to raise my arm. Asked to explain why I raised my arm I might respond as follows: ‘My reason for so acting was that I wanted to ask a question and believed that by raising my arm I would catch the speaker’s eye.’ However, there are two ways of interpreting this response, only one of which, I think, is really intelligible. On the correct interpretation, it was the contents of my belief and desire – what it was that I believed and desired – that constituted my reason for acting as I did. On the incorrect interpretation, it was my belief and desire themselves – those very mental states of mine – that constituted my reason for acting as I did. This, I believe, makes no sense. A reason for acting is a ground which would justify the performance of a certain action. The contents of belief and desire can constitute such grounds, but a belief or desire as such cannot, because these are not even items in the right ontological category to constitute grounds. Now, of course, a ground may exist which constitutes a reason for me to act in a certain way in certain circumstances, even though I am unaware of that ground. For instance, if a dislodged roof-slate is about to fall on my

350

Free Agency, Causation and Action Explanation

head, this fact constitutes a good ground for me to take avoiding action by jumping out of the way. Unless I am aware of the fact, however, I will be unlikely to take the appropriate action. I need to be aware of how the world is and of how it had better be to serve my interests, if I am to act rationally in the furtherance of those interests. But it is not my believing that the world is thus and so and my desiring to further those interests that constitute a reason for me to act in a certain way, just its being thus and so and my having those interests.14 I must obviously be aware of that reason if I am to act ‘in the light of’ it, but we should not confuse the reason itself with the mental states involved in my awareness of that reason. Most importantly, moreover, to act ‘in the light of’ a certain reason is not to be caused to act in a certain way by the mental states involved in one’s awareness of the reason. In fact, to be caused to act by certain of one’s beliefs and desires is precisely not to act freely and rationally, even if the ensuing behaviour is not contrary to reason. This is because an action which is so caused lacks the crucial element of choice, which is precisely the contribution of the will.15 To act rationally one must freely choose to act on a reason of which one is aware, by exercising one’s will in an appropriate way. (I shall say more in explication and defence of this view shortly.) I have been making these points in the course of trying to explain in what sense it is that the will is an ‘active’ power. Part of the upshot is that we have to conceive of the will as a power whose exercises are not characteristically brought about by the operation of prior causes. To this extent we may call the will a ‘spontaneous power’. We have every reason to believe that such powers exist, even in the inanimate world. The power of a radium atom to undergo radioactive decay is, apparently, just such a power. Where it differs from the will is in that its manifestations are not responsive to the demands of reason. When a radium atom decays, its decay is not caused by prior events, that is, by the causal influence of other objects acting upon it. And yet, of course, the atom does not decay ‘for a reason’, or ‘in the light of’ its beliefs and desires, for it has none. Indeed, there is no ‘reason’ for it to decay or not to decay. What is special about the will is that it is a rational power, that is, a power whose exercises are not only spontaneous but also responsive to the demands of reason. This, at least, is how I think we implicitly do, and plausibly must, regard the will if we are to regard ourselves as rational beings. But it may be wondered whether the ontology of powers, substance causation and free will being proposed here is compatible with what physical science reveals about the causal structure of the natural world. I believe that it is. Certainly, we must abandon any very strong principle of the closure of the domain of physical events under the relation of event causation: it cannot be the case that every physical event has a sufficient cause consisting of prior physical events. But we know from the example of the radium atom that such a principle is untenable in any case. It is not even the case that the

E. J. Lowe

351

decay of such an atom has its chances of occurrence fixed by prior physical events, for whether or not it decays at a given time is completely insensitive to such prior conditions. Certainly, radium has a distinctive ‘half-life’, meaning that, given a large number of radium atoms, there is a certain objective probability that half of them will have decayed in a certain period of time. But this does not mean that the chances of any particular atom’s decaying at any given time, or during any given period of time, are fixed in any way by prior events. When such an atom decays, it does so without any cause and, of course, for no reason at all. The difference with the will is just that, although its exercise has no cause, it is characteristically exercised for a reason, which the person in question is aware of and normally able to articulate. And the will, recall, is a power of the person and not of the person’s body nor of any distinctive part of it, so we are not required to suppose that purely physical substances have rational powers. Even so, it may be wondered how exercises of the will can be causally efficacious in bringing about bodily effects: that is to say, how a person, by exercising his or her will, can cause part of his or her body to move. There would be a problem here if we had to suppose that all bodily effects, because they are physical events, must have sufficient causes in the form of prior physical events. But we have already rejected any strong principle of physical causal closure that would have this requirement. Again, it may be objected that our proposal requires some mysterious form of ‘downward causation’ of mental acts upon the behaviour of physical objects, in violation of the causal autonomy of the physical world.16 But what can justify a belief in such autonomy, other than a discredited form of causal closure principle of the sort that we have just rejected? I am not suggesting that the will somehow exploits ‘gaps’ in physical causation left open by the indeterminacy of the quantum world, as though my will becomes causally efficacious by bringing about such events as the radioactive decay of atoms. For events such as the latter really are uncaused and so, a fortiori, not even caused by exercises of the will. Moreover, if they were so caused at some times and not at others, quantum physical statistical data such as those relating to atomic half-lives would be inexplicable. My appeal to the case of radioactive decay was just intended to establish the fact that spontaneous powers exist in nature, so that not all physical events have sufficient physical causes, contrary to the strong causal closure principle. Physical events brought about by the exercise of the will are, of course, caused, at least in part by those exercises, that is, by our volitions. Indeed, we may suppose those physical events to have sufficient causes (albeit not wholly physical ones), at least at the times at which the volitions in question occur. However, the volitions themselves, I have said, are uncaused – and they, by my account, are not physical events, since they are exercises of a non-physical mental power. The only significant implication of all this as far as physical events are concerned is just that some of them are such that, at certain times prior to their occurrence, they do not

352 Free Agency, Causation and Action Explanation

possess sufficient physical causes – for instance, a certain willed movement of my arm does not possess a sufficient physical cause at any time prior to my volition to move my arm in that way. But this is consistent with what is currently empirically known about the neurophysiology of voluntary action.17 And, as we have just seen, no acceptable principle of physical causal closure stands in the way of its truth. So, while I would not pretend to have explained how a person can, by willing, effect physical changes in his or her body – and maybe this is something that is destined forever to remain a mystery to us – I see no philosophical or scientific ground for denying that when we do so we exercise a non-physical power which is at once spontaneous and rational, in the sense that I have proposed. Let me conclude now by expanding a little on what I said earlier concerning reasons for action and action explanation. I am happy to acknowledge that sometimes the fact that an agent has a certain belief constitutes a reason for that agent to act in a certain way. Here I can do no better than to adapt an example of Jonathan Dancy’s.18 The fact that someone has the paranoid belief that he is being pursued by alien beings from Mars is a good reason for that person to visit a psychiatrist – although, of course, it is unlikely that such a person will be able to act upon this reason, because it is unlikely that he will be aware that his belief is paranoid. By contrast, in our earlier example of the falling roof-slate, it is not the fact that I believe that the slate is falling that constitutes a reason for me to jump out of the way: rather, it is the fact that the slate is falling – although, once more, I must be aware of this fact and so believe, or at least suspect, that the slate is falling if I am to be in a position to act upon that reason. The mistake that is characteristically made by those who favour an ‘internalist’ view of reasons for action is to regard such psychological states themselves, rather than the contents of such states, as constituting reasons for an agent to act – where those ‘contents’ are the facts or states of affairs whose existence is revealed to the agent in virtue of his being in such psychological states.19 The mistake is twofold in character. First, it represents a reason for action as being a state of the agent, when in fact it is a state of affairs. Second, it represents a reason for action as always being psychological in nature, when in fact this is only the case in exceptional circumstances, such as that of the paranoid believer. Even in the latter sort of case, the relevant reason for action is a psychological state of affairs – the fact that the agent has the paranoid belief – rather than the psychological state of the agent, his paranoid belief. I am also happy to acknowledge that there are many cases in which an agent’s beliefs do cause, or help to cause, some behaviour of the agent’s. The example of the paranoid believer in persecuting Martians will again serve us well. An agent with such a paranoid belief may well be caused by that belief to act in various seemingly bizarre ways, such as hiding behind bushes when approached by strangers in the park. Now, of course, the fact – were it to obtain – that one is being pursued by Martians might well constitute a good

E. J. Lowe

353

reason to hide behind bushes when approached by strangers; and so we are in a better position to understand this agent’s seemingly bizarre behaviour once we know that he has the paranoid belief. But we would be wrong to judge that the agent is acting rationally, or for a reason, just because we know that he has this belief and that its content would, if it were true, provide him with a reason for acting in the way that he does. If his paranoid belief is causing him so to act, then he is not choosing to act in this way in the light of that belief: he is simply being impelled by the belief and as such is not acting rationally. It is already well known that there is a problem of ‘deviant causal chains’ for those theorists who maintain that to act rationally is to be caused to act in a certain way by beliefs and desires which represent such an action as being in the agent’s interests in the circumstances in which he finds himself. To use Davidson’s own famous example, a climber may be caused to let go of the rope supporting his companion by his strong desire to save himself and his belief that by letting go he would increase his own chances of survival – and yet the climber’s behaviour may be entirely unintentional, as it was only because the onset of this belief and desire so unnerved him that his grip slackened.20 It has proved notoriously difficult to say, quite generally, how beliefs and desires should be supposed to cause behaviour if the behaviour is to qualify as rational action performed in the light of those beliefs and desires. What I am now contending, however, is that there can be no correct account of rational action in such causal terms, because an action’s being caused by the agent’s beliefs and desires is incompatible with its being an action that the agent chose to perform in the light of those beliefs and desires – and no action is rational if the agent does not freely choose to perform it. An action can be in accordance with reason, or reasonable, without being rational. To be rational, it must be done for a reason which the agent freely chooses to act upon – or so I claim. For example, jumping out of the way of a falling roof-slate is a reasonable thing to do, but it is not done for a reason and so rational unless the agent choses to jump out of the way in the light of his belief that the slate is falling. This may seem to turn on its head a famous argument of Davidson’s in defence of the doctrine that ‘reasons are causes’.21 And so it does. The argument appeals to the fact that it is not enough, to explain why an agent acted in the way he did, just to point out that the agent was aware of certain reasons for so acting. For the agent might well have been aware of various different reasons for so acting and, indeed, of various reasons for acting in other ways instead. To explain why the agent acted as he did, we need to know which of these reasons was the reason for which he actually did so act. We are supposed to conclude that the reason for which he actually acted was the one that actually moved him so to act, that is, the one that caused him so to act – which would have to be a psychological state of the agent, such as a certain combination of belief and desire. However, I have just maintained that being caused to act by certain of one’s beliefs and desires is in fact incompatible

354

Free Agency, Causation and Action Explanation

with one’s acting rationally, even though it may be compatible with one’s acting ‘reasonably’, or ‘in accordance with reason’. And the problem of deviant causal chains is grist to my mill here. Of course, this still leaves me needing to answer the question that Davidson’s argument poses: which of those reasons for acting of which the agent may have been aware at the time of acting was the reason for which he actually acted, if there was one? But my answer to this question is straightforward and obvious: the reason for which the agent acted is the reason which the agent freely chose to act upon – an answer that is unavailable to Davidson simply because he has no room for the notion of free choice in his philosophy of action. Normally, the agent himself will be able to tell us which reason this was. In situations in which an agent is aware of a variety of reasons for acting in a certain way, and of other reasons for acting in other ways, the agent is confronted with a choice as to how to act and on which reason to act. By choosing to act in a certain way in the light of a given reason, the agent makes that reason the reason for which he acted on that occasion.22 This is something that the agent himself brings about through his mental act of free choice. I conclude that we can indeed turn Davidson’s argument on its head. We can exploit the very considerations to which it appeals in support of a position diametrically opposed to the Davidsonian claim that explanations of action couched in terms of the agent’s reasons must be causal explanations.

Notes 1. Compare what Gideon Yaffe calls the ‘Where’s the Agent Problem’, in his Liberty Worth the Name: Locke on Free Agency (Princeton, NJ: Princeton University Press 2000), pp. 121ff. 2. See, for example, Richard Taylor, Action and Purpose (Englewood Cliffs, NJ: PrenticeHall, 1966), pp. 111–112. For a more recent defence of agent causalism, see Timothy O’Connor, Persons and Causes: The Metaphysics of Free Will (New York: Oxford University Press, 2000). 3. For the notion of a ‘basic’ action in play here, see Arthur C. Danto, Analytical Philosophy of Action (Cambridge: Cambridge University Press, 1973), p. 28. 4. For a defence of such a view of events, see my A Survey of Metaphysics (Oxford: Oxford University Press, 2002) ch. 14. 5. For an examination of some of these attempts, see my A Survey of Metaphysics, ch. 10. 6. For the sort of theory I favour, see my The Four-Category Ontology: A Metaphysical Foundation for Natural Science (Oxford: Clarendon Press, 2006), Part III. 7. See, for example, Eric T. Olson, The Human Animal: Personal Identity without Psychology (New York: Oxford University Press, 1997). 8. See, for example, Lynne Rudder Baker, Persons and Bodies: A Constitution View (Cambridge: Cambridge University Press, 2000). 9. For more details, see my Subjects of Experience (Cambridge: Cambridge University Press, 1996) and also my ‘Identity, Composition and the Simplicity of the Self’, in K. J. Corcoran (ed.), Soul, Body and Survival: Essays on the Metaphysics of Human Persons (Ithaca, NY: Cornell University Press, 2001).

E. J. Lowe

355

10. The problem has several different versions: see, for example, Trenton Merricks, Objects and Persons (Oxford: Clarendon Press, 2001), pp. 47–53. 11. But for a relatively recent defence, see Carl Ginet, On Action (Cambridge: Cambridge University Press, 1990). I defend volitionism myself in my Subjects of Experience, ch. 5. 12. The locus classicus for this view is, of course, the work of Donald Davidson: see his Essays on Actions and Events (Oxford: Clarendon Press, 1980). 13. Locke famously held precisely this view: for discussion, see my Locke on Human Understanding (London: Routledge, 1995), ch. 6. 14. In saying this I am advocating a so-called ‘externalist’ account of reasons for action. Compare, for example, Jonathan Dancy, Practical Reality (Oxford: Oxford University Press, 2000), pp. 15–17. 15. See further my An Introduction to the Philosophy of Mind (Cambridge: Cambridge University Press, 2000), pp. 252–262, and compare John R. Searle, Rationality in Action (Cambridge, MA: MIT Press, 2001), ch. 3. 16. See, for example, Jaegwon Kim, ‘ “Downward Causation” in Emergentism and Nonreductive Physicalism’, in A. Beckermann, H. Flohr, and J. Kim (eds), Emergence or Reduction? Essays on the Prospects of Nonreductive Physicalism (Berlin: de Gruyter, 1992). 17. See Benjamin Libet, Neurophysiology of Consciousness: Selected Papers and New Essays (Boston: Birkhaüser, 1993) and B. Libet, A. Freeman, and K. Sutherland (eds), The Volitional Brain: Towards a Neuroscience of Free Will (Thorverton: Imprint Academic, 1999). 18. See Dancy, Practical Reality (Oxford: Oxford University Press, 2000), p. 125. 19. Of course, the content of a belief, say, is a fact only if the belief is true—and this is sometimes thought to present a fatal problem for philosophers espousing ‘externalist’ theories of reasons for action. For such philosophers, it is alleged, are constrained either to say, most implausibly, that an agent cannot act rationally in the light of a false belief or else to compromise their externalism by saying that an agent who acts in the light of a false belief has only an ‘internal’ reason for action, in the form of that belief. However, the proper thing for an externalist to say about this matter, it seems to me, is that while it is indeed obviously true that an agent cannot be said to act ‘for a reason’ which does not exist, it does not follow, even on a purely externalist theory of reasons for action, that an agent who acts in the light of a false belief necessarily acts irrationally. ‘Rational’ and ‘irrational’ are contraries but not contradictories. Strictly speaking – the externalist should say – an agent who acts in the light of a false belief has no reason for acting as he does, but this by no means implies that such an agent has not properly exercised his powers of reasoning in deciding how to act and so is in that sense acting irrationally. For it cannot be a requirement of rational agency that rational agents be infallible in their exercise of such powers. By that standard, only God could qualify as a rational agent. 20. See Donald Davidson, ‘Freedom to Act’, in his Essays on Actions and Events (Oxford: Clarendon Press, 1980). The example appears on p. 79. 21. See Donald Davidson, ‘Actions, Reasons, and Causes’, in his Essays on Actions and Events (Oxford: Clarendon Press, 1980). 22. Compare Searle, Rationality in Action, p. 16.

19 Gods and Mental States: The Causation of Action in Ancient Tragedy and Modern Philosophy of Mind1 Constantine Sandis

It would be better to follow the mythology about gods than be a slave to the ‘fate’ of the natural philosophers: the former at least hints at the hope of begging the gods off by means of worship, whereas the latter involves an inexorable necessity. Epicurus, Letter to Menoeceus 133–4

1

Double causation in ancient tragedy

R.P. Winnington-Ingram opens his famous study of causation in Hippolytus with the following words: Why did the events happen as they did? This is no problem to Aphrodite or to Artemis. Bitter enemies though they may be, on one point they agreed – that what takes place is the work of a god; and the responsibility which Aphrodite claims in the Prologue is endorsed by Artemis in the closing scene. Yet the human characters seem to choose their courses and to work out their disasters on the plane of human circumstance and motive, so that Wilamowitz could say: ‘Aphrodite is not necessary for our understanding of the action’.2 The suggestion is that both gods and mortals3 are responsible for the thoughts and actions of the latter. In the above example, Aphrodite causes the tragedy’s events to unfold as they do, yet these are also said to include actions chosen by mortals: What does Aphrodite say? Her preparations, she tells us, are far advanced (22f.). If I may put it rather grotesquely, those preparations turn out to 356

Constantine Sandis

357

have been very elaborate indeed. She has caused Phaedra to fall in love with Hippolytus: well and good, that is within her province. But, if she is to be responsible for the whole action, she must also have placed Phaedra in the fatal environment of the palace and (more important still) provided her, through the wider social environment, with a set of moral ideas which proved inadequate to the situation. For all these things played a part in her downfall. Aphrodite goes on to say that she will make the truth known to Theseus, and that Theseus will curse his son and kill him. But this involves the Nurse, her single-minded devotion to her mistress, and her moral limitations. It involves Theseus being what Theseus was – and a relationship (or rather a complete lack of relationship) between Theseus and Hippolytus, who himself has more aspects than the scorn of Aphrodite for which he is so cruelly punished.4 Double causation, for Winnington-Ingram, is a form of over- determination: while tragic actions are caused by both gods and mortals, either set of agents would have alone been sufficient not only for our understanding of all the behaviour involved but for its actual production too. He is here following a mainstream tradition championed by E.R. Dodds in his landmark publication The Greeks and the Irrational.5 Dodds distinguishes between normal actions and actions performed in a state of ate viz. from the ‘experience of divine temptation or infatuation’6, claiming that the latter are ‘doubly determined, on the natural and on the supernatural plane’.7 The origins of this view lie with Martin P. Nilsson who claimed that an important characteristic of the Homeric hero was that his actions seemed alien to him once he had performed them, even if they did not seem so at the time of action: Men turn to the gods [ ... ] to lay upon them the blame for that which has happened contrary to desire or intention. Even in the divine apparatus this idea finds expression in the common phrase: ‘Now this would have happened ... had not a god ... ’. The Homeric man is absolutely under the dominion of the emotion of the moment. When passion has subsided and the unhappy consequences begin to appear he says: ‘I did not desire this; hence I did not do it.’ His own behaviour has become something foreign to him, it seems to be something which has penetrated into him from without. He lays the blame on some daimon or god or Ate or on Zeus, Moira and the Erinyes, as Agamemnon does in regard to his treatment of Achilles.8 It would take another essay to properly examine the stirring account of the psychology of the Homeric man which follows the passage quoted above.9 For present purposes, however, what matters is not how the Greek notion of

358 Gods and Mental States

divine intervention originated but how it relates to various notions of double causation. For Dodds, this is essentially a question about agential control: When Theognis calls hope and fear ‘dangerous daemons’ [Theognis, 637f.], or when Sophocles speaks of Eros as a power that ‘warps to wrong the righteous mind, for its destruction’, [Antigone 791f.] we should not dismiss this as “personification”: behind it lies the old Homeric feeling that these things are not truly part of the self, since they are not within man’s conscious control; they are endowed with a life and energy of their own, and so can force a man, as it were from the outside, into conduct foreign to him.10 Aristotle would later rationalize tragic action as the result of a kind of error or fallibility (amartia) brought about by kakodaimonia, the opposite of eudaimonia (literally ‘good spirit’ but typically translated as ‘well-being’, ‘human flourishing’, ‘the good life’, or even just ‘happiness’). The word ‘daimon’ has been largely stripped of its previous supernatural status, still Aristotle understands the ultimate origins of good and bad fortune (including ethical luck) to be external to the agent’s control. Aristotle distinguishes between internal and external moving principles, yet what is external to the agent’s control need not literally lie outside the agent.11 One need not be in bad faith (as is Nilsson’s Homeric man) to experience such motivational forces as alien. Aristotle serves as a nice reminder of the fact that we need not take sides in the debate over whether gods are best characterized as psychological powers (J.-P. Vernant12), persons (Walter Burkert13), or both (Jan. M. Bremmer14) to agree with Dodds’ suggestion that they serve as a constant reminder of ‘man’s helpless dependence on an arbitrary Power’.15 Describing this state of affairs as a case of over-determination, however, is highly problematic. For how can we take seriously the idea that mortals are dependent on the gods if the latter are not required for either the occurrence or the explanation of human action? As Dodds portrays things, the Greeks would seem to want to have their divine intervention and eat it too; it comes as no surprise, then, that he wished to attribute irrational modes of thought to them. The picture he offers is that of a form of collective – though by no means always congruous – action that would have still occurred had only one agent been involved. It is not dissimilar to that of one or more strangers joining a person in pushing a vehicle which she would have managed to move regardless: the strangers might be said to interfere (in the sense that they might be intruding) and perhaps, on occasion, to help (if the pushing consequently required less effort), but they cannot be said to intervene since they do not alter the natural course of human action in any way. Such agents would lack the kind of impact on human conduct that Nilsson’s Homeric man attributes to the gods, their efforts are all but epiphenomenal. It is tempting, in the light of all this, to abandon the notion that double causation necessarily implies over-determination. There are two obvious

Constantine Sandis

359

ways of doing this. One retains the collective model of tragic action but rejects the idea that a single agent would have been sufficient. The other rejects not only the collective model of tragic action but the very idea that double-causation requires two ontologically distinct causes. I begin with the latter. Using Sophocles as a study case, H.D.F. Kitto famously argued that divine causation is reducible or otherwise identical to human or otherwise natural causation: the world of the gods and the world of men are not two worlds but one ... the gods are not superior and remote powers who kick us around for our good or for their own amusement ... the course of things in this human universe is not fixed by arbitrary decrees or a blind fate ... the divine activity in the plays, provides them, so to speak, with a system of coordinates against which we can read off the permanent significance of the particular action.16 On such a view the divine is nothing more than a redescription of the natural, the term ‘double causation’ indicating only that truths regarding the causation of tragic action may be captured in two strikingly different ways. Consequently the question of sufficient agency does not properly arise, which is not to say that we cannot talk of sufficient explanation (the latter being sensitive to description). Kitto’s interpretation promotes the notion (particularly popular among continental classicists such as Jean-Pierre Vernant, Louise Bruit Zaidman, Pauline Schmitt Pantel, T.C.W. Oudemans, and A.P.M.H. Lardinois17) that the gods are little more than representations of human powers or motivations. While happy to allow that gods ‘really are implicated in the action’18 indeed ‘are active’19 Kitto denies that divine occurrences exist over and above natural ones: these gods, collectively, are the natural order of events; they are not ‘supernatural’ at all, except that they are immortal and omnipresent [ ... ] Sophocles could combine in one play human and divine drama, in such a way that each is a sufficient explanation of what happens, yet neither obscures the other’.20 Accordingly, the sole purpose of divine motivation is said to be that: it makes us see the particular action that we are watching not as an isolated, a causal, unique event; we see it rather in its relation to the moral and philosophical framework of the universe ... the divine background ... means ultimately that particular actions are at the same time unique and universal.21

360 Gods and Mental States

This ‘universalist’ interpretation, which originates in Hegel, 22 allows Kitto to reject ‘the grim old doctrine that human beings are only puppets in the hands of omnipotent and often malignant gods’ and accept that ancient tragedy is ‘full of entirely responsible and convincing human character and action’. 23 In an earlier book he portrays Homeric events in much the same way, arguing that the modern reader would be mistaken to think that mortals are being portrayed there as ‘nothing but pawns pushed about on a chessboard by a set of capricious and irresponsible deities’ and that Homer paints a ‘picture of autonomous, responsible human agents’. 24 The sentiment would later be echoed in Bernard Williams, who in Shame and Necessity declares that ‘when the gods do intervene, they do not standardly do so by making people do things – winding them up, so to speak, and pointing them in a certain direction’.25 We must not forget, however, that that Nilsson, Dodds, and Winnington-Ingram are all equally keen to reject the ‘mere puppet’ view of ancient tragedy as psychologically bankrupt. Kitto’s particular achievement was to do so without invoking over-determination. But at what cost? Kitto downplays the interventionist nature of the gods to a dangerous extent, effectively denying that there are any genuinely alien (or ‘other’) forces at play at all in ancient tragedy, a move which even Aristotle’s agentcentred ethics takes care to avoid. He thus does away with what is perhaps the most central idea in ancient tragedy, namely that certain things which lie beyond human control may seriously affect the course of human lives. In so doing, he also rejects the corollary that mortals are not always entirely responsible for their actions.26 It is these precise ideas that motivate the other noted way of denying over-determination viz. that which retains a robust sense of double causation as a particular form of collective action. This route results in the kind of picture presented by Adrian Poole in his recent Tragedy: A Very Short Introduction. Under the heading ‘In league with the gods’ he writes: Greek tragedies do not just expound upon beliefs about the gods; they debate them. And the most important aspects of this debate concern the points at which human beings and gods act together. This is the image used by Darius and his queen in Persians, when they try to explain how their son has committed this terrible hamartia: a god or daimon must have ‘joined in’ with him (724, 742). It takes two to make the tragic act happen, both the human agent and the divine or non-human. This is a helpful clue to the treatment of blame in tragedy, that there is never a singular cause. Indeed, two agencies represent the absolute minimum; there may turn out to be far more, a whole complex chain or sequence of causation.27

Constantine Sandis

361

The point about debate is an important one, for there is no reason to suppose that the tragic poets and their audiences all had firmly fixed common views about such things.28 It is somewhat surprising, then, to find Poole limiting the scope of the debate to the points at which gods and mortals act together. That they act collectively and, to a certain extent, even the way in which they do so, is simply taken for granted. As Poole presents things, a person whose behaviour is doubly caused is not unlike a vehicle that only moves if at least two agents are pushing it together (with or without shared goals and/or methods). Each agent participating in such collective action is thereby always open to partial – but never to complete – praise or blame. The truth is in the details which each tragedy, nowadays also each staging (and perhaps even each performance) of any given tragedy, disputes. Pace Kitto, blame and responsibility do not, in ancient tragedy, rest absolutely with mortals any more than they do with the gods. But is this really because, as Poole puts it, these agencies perform each action together? Poole assumes this because, like Nilsson, Dodds, Kitto, Winnington-Ingram, and many other commentators before him, he is working with a paradigm of double (or even multiple) causation of tragic action according to which double causation is the causation of one particular action (or one particular series of actions) by two or more agents. That is to say, there is one thing (or one set of things) with two (or more) causes. But how true to ancient tragedy is this modern outlook? One of the aims of this essay is to argue that it seriously misinterprets the role of the divine: it is simply false that the gods of ancient tragedy typically, if ever, causally determine human action, or produce it by any other means. In thinking otherwise, I shall maintain, Dodds et al. are conflating the motivation and/ or enablement of action with its actual production, an error shared by many modern philosophers of mind. Indeed, the most popular current accounts of action causation are strikingly similar to those which classical scholars have mistakenly ascribed to the ancients, or so I shall argue in what follows. I replace this vision with a radically different interpretation of what is assumed and what is debated in ancient tragedy, one which ascribes to the ancients a considerably more plausible understanding of agency. Predictably, yet perhaps also provocatively, I suggest that modern philosophers would do well to adopt an analogous outlook.

2

Double causation in modern philosophy of mind

Whereas in ancient tragedy it is simply false that the gods intervene in every single case,29 in modern philosophy of mind it is commonplace to hold that all our thoughts and actions (and not just those which are particularly tragic,

362 Gods and Mental States

heroic, delude, akratic etc.) are to be explained by both causally efficacious mental phenomena as well as the reasons in the light of which we perform actions and form beliefs and desires etc. Alvin I. Goldman, for instance, writes: If I ask you, for example, ‘Why did Jones go to the concert tonight?’ you might rely, ‘Because Smith told him they were going to play the Trout Quintet’. The only event explicitly mentioned in this reply is Smith’s telling Jones something. How does this explain Jone’s [sic] act? Obviously, the reply implies that Smith’s telling Jones that they were going to play the Trout Quintet caused Jones to believe that they were going to play the Trout Quintet. And this belief, presumably in conjunction with Jones’ desire to hear the Trout Quintet, caused Jones to go to the concert. Thus, Jones’ act is explained by implicitly indicating certain beliefs and desires and explicitly indicating the cause of the relevant belief. 30 This popular causalist view has been attributed to philosophers as diverse as Aristotle, Aquinas, and Hume.31 It was revived in the mid-1960s by Donald Davidson, whose influential paper ‘Actions, Reasons, and Causes’ begins by setting an agenda which has ever since prevailed in modern philosophy of mind: What is the relation between a reason and an action when the reason explains the action by giving the agent’s reason for doing what he did? We may call such explanations rationalizations, and say that the reason rationalizes the action. I want to defend the ancient – and common sense – position that rationalization is a species of causal explanation ... Whenever someone does something for a reason ... he can be characterized as (a) having some sort of pro attitude towards actions of a certain kind and (b) believing (or knowing, perceiving, noticing, remembering) that his action is of that kind. Under (a) are to be included desires, wantings, urges, promptings, and a great variety of moral views, aesthetic principles, economic prejudices, social conventions, and public and private goals and values in so far as these can be interpreted as attitudes of an agent directed toward actions of a certain kind ... .Giving the reason why an agent did something is often a matter of naming the pro attitude (a) or the related belief (b) or both. Let me call this pair the primary reason why the agent performed the action ... rationalizations are causal explanations ... the primary reason for an action is its cause.32 Yet if all of our actions are produced by our so-called ‘mental states’ (from here onwards simply ‘mental states’33), it seems no more possible for us to perform voluntary and intentional actions at will than if they were caused

Constantine Sandis

363

by the gods.34 For Davidson, the solution to such problems lies with a distinction between mental states causing us to act (problematic), and mental states causing our actions (unproblematic): There can be a great difference between ‘The heat caused Samantha to return to Patma’ and ‘The heat caused Samantha’s return to Patma’. The former implies, or strongly suggests, a limitation on Samantha’s freedom of action; the latter does not. ... An action may be caused without an agent being caused to perform it. Even when the cause is internal, we sense a difference. ‘Desire caused him to do it’ suggests a lack of control that might be called on to excuse the action because it makes the act less than voluntary; ‘Desire was the cause of his doing it’ leaves the question of freedom open.35 The same solution had previously also been proposed by Wilfred Sellars: [T]here may be a sense of ‘cause’ which is stronger than that of mere predictability, and in which it is true to say that actions are caused, but not that persons are caused to do them ... in this new sense volitions are, at least on occasion, the causes of action.36 For Sellars, there is no harm in saying that ‘the volition is the cause of the action’ only so long as one is sufficiently sensitive to the varieties of relationships that can be called causal, and avoids construing this case in terms of a paradigm that would lead one to say that volition causes the person to do the action. In voluntary action, we are not caused to act.37 The Sellars–Davidson proposal has been largely ignored by causalist philosophers, perhaps with good reason, for the distinction between X causing an agent to act and X causing an action to occur is highly counter-intuitive. While we might easily agree with the claim that ‘an action may be caused without an agent being caused to perform it’ in cases where the action is caused (solely) by the agent, it is unclear why we should think this is so when the action is (also) caused by some other thing, be it internal or external to the agent.38 If anything, it is the latter kind of statement that places the greater emphasis on the limitations in question.39 The phrase ‘the heat caused Samantha to return to Patma’, for example, seems to involve the agent in a way in which the phrase ‘the heat caused Samantha’s return to Patma’ (ungrammatical as it may be) fails to. The former explicitly states that there was something Samantha did viz. return to Patma (possibly even bring about the event of her returning to Patma), even if she was caused to

364 Gods and Mental States

do it. The latter, by contrast, might be thought to downplay Samantha’s agency by quantifying over an event caused by the heat without specifying whether the return was something Samantha did or merely something that happened to her.40 Either way, it is far from obvious how mental states might cause the occurrence of an intentional action without causing its agent to perform it. Consider the protagonist of Vermeer’s Woman Writing a Letter, with her Maid. We might try to imagine her as composing the letter in a frantic or delirious state, with absolutely no control over her anxiety, nervousness, fear, or depression or any other psychological state which her action(s) might be a causal result of. On such a scenario the woman’s mental states could easily be said to cause her action by causing her to perform it. But we may also imagine her as being more cool, calm, and collected – to use the popular phrase which Michael Smith adopted as shorthand for those idealized conditions of reflection which form the naturalistic axis of his internalist realism viz. a maximally informed, coherent, and unified motivational set.41 It would be absurd to insist that in this case her mental states literally made or caused her to do it. So what would it be for the woman’s mental states to cause her action of writing the letter without making her to write it? Davidson’s answer is that in the latter sort of case – but not the former – the protagonist acts ‘in the light of some imagined good’42 in that ‘the beliefs and attitudes in the light of which the action is reasonable’43 reflect ‘the all-important process of weighing considerations’, be it conscious or otherwise.44 There are at least two distinct ways for the causalist to conceive of this process. According to the first, rationalist conception of human agency, we always intentionally do what we judge to be best, our actions caused by the reasons for which we judge it to be best.45 In its most popular form, this conception claims that: 1. We most desire to perform whichever action(s) we judge to be best. 2. We always intentionally do what we most desire to (and believe we can) do.46 On such a picture, all practical reasoning inevitably concludes in a corresponding action or omission; hence the age-old difficulty such theories face when it comes to providing an account of how it is that we can ever act against our better judgement. The rationalist conception also seems to render epiphenomenal the light through which reasons guide agents, as opposed to states and events.47 Some follow Freud in biting the bullet and asserting that agents act in the darkness of their reasons and not in the light of them. Others, such as the Churchlands, follow the later Nietzsche in combining causalism with a physicalism that denies that there are any mental causes at all.48 Either way, it would seem that a central aspect of human agency has been lost.

Constantine Sandis

365

The aspect in question is the focus of what Joseph Raz calls the classical conception of human agency (because he holds that it may be found in Plato and Aristotle49), contrasting it with the rationalist conception, favoured by Davidson, Goldman, and Sellars: In broad outline, the rationalist holds that paradigmatic human action is action taken because, of all the options open to the agent, it was, in the agent’s view, supported by the strongest reason. The classical conception holds that the paradigmatic human action is one taken because, of all the options the agent considers rationally eligible, he chooses to perform it.50 On this model, we do not necessarily will to do what we take to have the most reason to do; practical reasoning never concludes in action or intention but in belief about ‘the thing to do’ or ‘the thing to intend doing’ (just as theoretical reasoning of an analogous might be said to conclude in a belief about ‘the thing to believe’) which the agent may or may not decide to act upon.51 At the very most, it might be allowed that it can conclude in (first-order) desire or volition, so long as the will is thought to include second-order desires about the first-order ones. Accordingly, weakness of will is said to occur when we do not sufficiently desire or intend to act upon a rational belief about what we ought to do and consequently fail to act upon it. All this allows for greater agential space between practical reasoning and the action(s) it is sometimes said to conclude in, space filled by the will which plays a necessary role in the production of intentional action that is distinct from that of reason. It does so by insisting that is one thing to have rational beliefs and desires, and another to act upon them: ‘we can appropriately respond to reason because we have a will’52 but it is equally true that ‘we follow our will when we act against our better judgement’.53 Philosophers who follow the classical conception frequently use the terms ‘will’, ‘volition’, and ‘desire’ interchangeably.54 We should not let this mislead us into thinking that in their view the difference between action within and action beyond our control is not that between action caused by so-called ‘volitions’ and action caused in some other way, which bypasses the will. After all, such a stance may be easily attributed to rational causalists such as Goldman, Sellars, and Peter Railton who all defend the ‘standard’ Davidsonian view that an action is intentional if it is caused (in a non-deviant way) by an intention which was in turn caused by the agent’s beliefs and desires55 while allowing conceptual space for volitions, so long as they are construed ‘to include desires or intentions’.56 Railton, for example, uses the term ‘rational will’ to describe the desire of a rational organism in which ‘intentions spontaneously emerge ... through the chemistry of belief and desire’.57 Sellars similarly defines volitions as ‘mental episodes that cause actions’,58 specifying that they are ‘one variety of occurrent intention (state)’, that ‘it is proper to think of the fact that a person has certain beliefs and/or desires which serve

366 Gods and Mental States

as premises in his drawing of a conclusion, as the cause of his accepting this conclusion’59 and that ‘practical conclusions, in the form of volitions, cause actions’. In so doing, he leaves no room for the agent to will to act either upon or against the conclusion(s) of his practical reason. In taking mental states or episodes to play a sufficient role in the production of all intentional action causalists such as Sellars and Railton fail to explain how it is that we can be in control of actions that are the causal product of our will any more than we would be in control of actions that were the causal products of divine will. While it is true that in the former case (but not the latter) the will in question is that of the human agent (who is an implicit part of any determinist story that refers to mental states60), it nevertheless remains a mere effect of other mental states and episodes whose force is ultimately alien to the agent. Railton responds to such objections by stating that a human organism ‘is no mere bundle of states. It is a structured, functional whole’, a ‘self-organised complex that exhibits the proper functioning of thoughts, beliefs, emotions, and desires’.61 Quite so, but the fact that the mental episodes in question are parts of human beings and may rightly be thought to belong to them at the very most establishes only that humans cause whatever their mental sates cause.62 While there are obvious differences between psychological determinism and external force (whose respective contrasts Hume named liberty of spontaneity and liberty of indifference63), the difficult question for causalism – which Railton bypasses – is not that of how we might identify ourselves as the causes of actions that can emerge spontaneously, but a question of identifying criteria for being in control of and/or responsible for the things which we undoubtedly cause. One way of approaching this issue, favoured by Hume as much as by Nilsson’s Homeric man, is to make a distinction between those psychological causes that lie within our character and those that do not.64 This raises the vexed question of what it is for volition to stem from one’s character. In allowing that it is up to the will to choose whether or not to act upon the conclusions of our deliberations, the classical approach seems, at least prima facie, better equipped to answer this question. Indeed, they offer elaborate, often ingenious, accounts of what Harry Frankfurt has termed the ‘identification’ between agents and their motivation. These all view the will as the agent’s representative,65 appealing to various mechanisms thought to explain how it is that we can identify with actions that are the causal product of the will and the various alien (ultimately external) things that shape it. Examples of such mechanisms include reflexivity, decision, whole-heartedness, and second-order desires (Frankfurt), self-governing policies, self-reflection, and satisfied plans and intentions (Bratman), the semblance of responsiveness to reason (Raz), ‘normative competence’ or the capacity for critical evaluation (Watson), self-awareness and the desire to act in accordance with reasons (Velleman), and guidance control (Fischer).66

Constantine Sandis

367

Most of the above view the will as a mechanism or event which can cause behaviour. Michael Bratman, for example, thinks of the phenomena of planning, intending, and valuing as kinds of willing that may act as causes of action.67 Similarly, Frankfurt defines the will as ‘an effective desire – one that moves (or will or would move) a person all the way to action’.68 Such proposals may nonetheless be distinguished from those of rationalist causalism in virtue of their commitment to the claim that ‘it is no part of the nature of an action to have a prior causal history of any particular kind’.69 The negative emphasis here is on the words ‘prior’ and ‘history’ rather than ‘causal’. Frankfurt writes: Explaining purposive behavior in terms of causal mechanisms is not tantamount to propounding a causal theory of action. For one thing, the pertinent activity of these mechanisms is not prior to but concurrent with the movements they guide. But in any case it is not essential to the purposiveness of a movement that it actually be causally affected by the mechanism under whose guidance the movement proceeds. ... the causal movements which stand ready to affect the course of a bodily movement may never have occasion to do so; for no negative feedback of the sort that would trigger their compensatory activity may occur. The behavior is purposive not because it results from causes of a certain kind, but because it would be affected by certain causes if the accomplishment of its course were to be jeopardized.70 While this is by no means the behaviour of what Frankfurt calls a ‘wanton’,71 it remains true (on the above picture) that the will is itself caused by earlier, mental and/or neurological, events. Indeed Frankfurt is responsible for introducing the notion of volitional necessity viz. a constraint on the will that prevents agents from making use of their capacities to act (or refrain from acting) by making all apparent alternatives unthinkable.72 Such thoughts awaken anxieties regarding accountability, responses to which cannot sensibly appeal to a hierarchy of volitions without bounding themselves to an infinite regress of the kind made famous by Gilbert Ryle.73 Frankfurt concedes that an agent ‘will not be morally responsible for what he has done if he did it only because he could not have done otherwise, even if it was something he really wanted to do’,74 yet he also maintains that this does not mean that determinism necessarily constrains responsibility, on the grounds that determined actions may be performed wholeheartedly that is, for some reason other than the fact they were determined. This outlook seems to commit Frankfurt to the possibility of the kind of over-determination found in Winnington-Ingram’s account of double causation. Yet volitional necessity surely implies not only that we have no choice but to do certain things but that we have no choice but to do them because of certain reasons.75 So what we should be asking when applying Frankfurt’s revised

368 Gods and Mental States

principle of alternate possibility is not ‘why did A do x?’ but ‘why did A do x for reason y?’ (or, if one is interested in events, ‘why did the event of A’s doing x for reason y occur?’76). On a determinist worldview the answer sooner or later boils down to some aspect(s) of our facticity, some basic fact – or combination of facts – regarding the agent’s psychology, history, current situation, or anything else that lies beyond her control.77 By contrast, this need not be so on a libertarian understanding of intentional agency as a two-way causal power to decide (or refrain from deciding) as well as act (or refrain from acting) in accordance with or against not only one’s better judgement but, more importantly, also one’s will ?78 Here the argument reaches a new standstill but one need not take sides to appreciate (in light of all of the above) that the question of agency as understood in modern philosophy of mind and action is a question about how, if at all, we can be said to have agential control within a universe that appears to be highly deterministic (with certain micro-pockets of randomness). Even in its classical guise, this project remains the result of a highly mechanistic outlook not dissimilar to that projected onto ancient tragedy by Winnington-Ingram et al.79 In § 3 below I argue, pace these classical scholars, that the backdrop of ancient tragedy is the mirror opposite of that against which the modern search for agency is conducted. What is contested and debated there are not conditions for identification and control but rather the typically extraordinary conditions of alienation and constraint that present themselves in a world where agency and answerability are otherwise taken for granted. Accordingly, the difficult task of knowing thyself is not a matter of identifying oneself with one’s mental states but of disassociating oneself from all alien forces. Deterministic threats are present, but only in a pre-theoretical sense not unlike that which McCann ascribes to modern-day undergraduates upon their first encounter with compatabilism: ‘Most students fail to grasp compatabilism: they insist on taking it as a theory on which causal factors limit one’s options but allow ‘free will’ to operate among those that remain.’80 From this ‘naïve’ viewpoint, tragic or heroic decisions and actions are variously enabled, facilitated, motivated, and constrained by alien forces, but not causally produced by them. Mortal and divine causation exist side by side, with no threat of reduction or elimination.

3 Divine intervention: motivation not causation of action At the outset of the Odyssey, we find the following account of a divine assembly: [27] ... the other gods were gathered together in the halls of Olympian Zeus. Among them the father of gods and men was first to speak, for in

Constantine Sandis

369

his heart he thought of noble Aegisthus [30] whom far-famed Orestes, Agamemnon’s son, had slain. Thinking on him he spoke among the immortals, and said: “Look you now, how ready mortals are to blame the gods. It is from us, they say, that evils come, but they even of themselves, through their own blind folly, have sorrows beyond that which is ordained. [35] Even as now Aegisthus beyond that which was ordained, took to himself the wedded wife of the son of Atreus, and slew him on his return, though well he knew of sheer destruction, seeing that we spake to him before, sending Hermes the keen-sighted Argeiphontes that he should neither slay the man nor woo his wife; [40] for from Orestes shall come vengeance for the son of Atreus when once he has come to manhood and longs for his own land. So Hermes spoke, but for all his good intent he prevailed not upon the heart of Aegisthus; and now he has paid the full price of all.81 Zeus is explaining that Aegisthus acted as he did in full knowledge of the consequences. We are not given sufficient information to determine whether or not the action is akratic but what is certain is that Aegisthus’ choice is being described as a case of passion reigning over reason. It is only natural that a poet – or indeed anyone else seeking to explain action in an age predating modern psychology and neuroscience – would in such cases attribute the source of human motivation to gods. Among other things they exhibit basic motivational forces such as desire (Aphrodite), lust (Eros, who in some myths is Aphrodites’ son), ecstasy (Dionysus), memory (The Titan goddess Mnemosyne), fear (Phobos), dread (Phobos’ twin brother Deimos), faith (Pistis), need (Penia), obsessive folly (Ate), modesty (Aidos), thoughtfulness (the Titan goddess Metis), mad rage (Lyssa), and battle fury (Ares and Enyo), as well as action-guiding values relating to justice (Zeus), wisdom (Athena), beauty and self-restraint (Apollo), protection (Artemis), marriage (Hera), and the home (Hestia), often in complicated scenarios involving divine greed, lust, jealousy, and deceit. This is well trodden, comfortable, ground. A recent instance of this stance may be found in David West’s treatment of Dido in his introduction to the Aeneid: The quarrel between the goddesses [Venus and Juno] can be seen as a dramatization of her emotions, the internal turmoil between love for Aeneas, longing for marriage, loyalty to her dead husband and duty to the city of which she is queen.82 So far, so good. But a little later he then says of Aeneas that The plan to attack an undefended city is not in origin his own: ‘At that moment Aeneas’ mother, loveliest of the goddesses, put it into his mind ... to lead his army’ (554–5) against the walls of the city. We have

370 Gods and Mental States

already seen double motivation in action. For example when Dido fell in love as a woman, while at the same time Venus and Cupid manoeuvred her into the madness of love. There the double motivation made the event more complex and more profound. Here it is put to ingenious use. When the hero thinks of a course of action which does him little credit, any stain on his character is lessened by a narrative which attributes the motive force to a god, who by definition cannot be resisted.83 Yet, to the extent that we can talk about double motivation at all, the two motivations in question are those of (1) the gods motivating the mortals and (2) mental states motivating the mortals. There is no conflict here of the sort that Winnington-Ingram was trying to capture, at least not given the commonly accepted view (which, as we previously saw, West himself adheres to) that part of the gods dramatize the psychology of the mortals. And while human beings cannot directly resist being motivated (by either gods or mental states), we should not conclude from this, as West seems to, that the motivation itself cannot be resisted, that is, it inevitably leads to action because human beings are powerless before it. It is one thing to produce and/or explain motivation, and quite another for this motivation to produce and/or explain action. As far as human action is concerned, the gods are only responsible for its motivation, not its execution.84 It is simply not true that gods cannot ever be resisted, let alone true by definition. Hence Zeus can truthfully claim that the mortals should not blame the gods so readily; this would be as absurd as attributing our actions to our beliefs and desires rather than to ourselves. Motivational states, like the ancient gods, can typically also be resisted. In claiming that we always do what we are most motivated to do, causalists neglect to place a sufficiently large wedge between the motivation and the causation of action, thus leaving no space for a two-way agential power to either perform or refrain from performing that action which we find ourselves most motivated to perform.85 It is only within such a closed framework that the presence of a motivating god (or mental state) leads to the puzzle of double causation. The distinction between the motivation of action and the causation of action carries with it a parallel distinction between internal motivating forces and the (external) reasons for which we act (regardless of whether or not reasons are being thought of as causes here). The mere presence of motivation is insufficient to produce and/or explain voluntary and intentional action. This is not to say that gods never function (dramatically) as reasons but only that in so far as they do so their role is analogous to that of the socalled contents of mental states (e.g. the things believed, and indeed at times desired by mortals), rather than the psychological entities themselves (e.g. the beliefs and desires of mortals).86 There is, after all, a difference between a god guiding someone towards (or away from) reason by affecting that

Constantine Sandis

371

person’s psychology (thereby enabling them to see a reason, blinding them from doing so, or merely clouding their judgement) and doing so by appearing before them (typically in disguise) and simply presenting them with a reason. Indeed, the former kind of situation might even be used to explain why an agent behaved a certain way in the latter kind of situation (e.g. by offering an account of why she took something to be a reason for acting in a certain way).87 Either way, both gods and mental states merely motivate – rather than produce – actions which we perform for reasons. It is worth emphasizing out here that the dramatizations of the agent’s psychology may work, no matter what we (or indeed the poet in question) think about free will and determinism. They work as explanations not because gods and mental states causally produce human behaviour,88 but because they can enable, structure, motivate, or constrain it. In Aeschylus’ The Libation Bearers, for example, mortals such as Agamemnon and Aegisthus are agents of their own fates, even if their family history has ensured that certain outcomes are far more likely than others. Clytemnestra tries to justify her action as ordained by the gods (because it is a righteous retribution for the slaughter of her daughter), portraying herself as an instrument of divine causality and fate, but the Chorus will not hear of it and her guilt is announced with the appearance of Aegisthus, her lover. Likewise Aegisthus himself may give revenge as his motive, but he is also motivated by power and ambition. More importantly, he does not overcome this motivation but gives into it. A different example of the limits of divine intervention may be found in Euripides’ Hippolytus, where Phaedra’s own perception of what is happening to her is focused on her motivation not her action: Where have I wandered from the path of good sense? I was mad, I fell by the stroke of some divinity. Oh, how unhappy I am. Nurse, cover my head up again. For I am ashamed of my words. Go on, cover it: the tears stream down from my eyes and my gaze is turned to shame. For to be right in my mind is grievous pain, while this madness is an ill thing. Best is to perish in unconsciousness.89 And again a little later, ‘A friend destroys me. Neither of us wills it.’90 Falling in love is not something we can do at will (voluntarily and intentionally, following deliberation etc.), but something that happens to us, as with belief.91 There will typically be reasons we respond to but we find ourselves falling in love, sometimes against our will, a picture affirmed by Phaedra’s nurse: ‘So, my daughter, leave off these wicked thoughts, leave off this pride. It is pride, nothing else, to try to best the gods. Bear up under your love: it was a god that willed it.’ 92

372 Gods and Mental States

This is no cause for defeatism, however. Aphrodite may have clouded Phaedra’s judgement by causing her to fall in love with Hippolytus (against her will), but there are also things that she can do to cure herself. Hence the nurse’s advice: ‘And if you are ill with it, use some good measures to subdue it. There are incantations, and words that charm: something will turn up to cure this love. Men will be slow to invent such contrivances if we women do not find them.’93 Gods and mental states need not win out every time, for while mortals may find themselves driven by the gods, they often also possess the ability to put an end to it. Tragedy strikes when, for whatever reason, we fail to overcome internal and external obstacles, and a god may succeed in manipulating us in much the same why that external factors and/or our drives and appetites may win out. Whether or not we should be held accountable is in part a matter of degree, and in part a matter of circumstance. Often we are excused rather than absolved: You have done dreadful deeds, but for all that it is still possible for you to win pardon for these things. Aphrodite willed that things should happen thus, sating her anger. Among the gods the custom is this: no god contrives to cross the will of another, but we all stand aside. For be in no doubt, if it were not that I feared Zeus, I would never have come to such a pitch of shame as to allow the death of the man I love most among mortals. Ignorance acquits your misdoings of baseness, and further the death of your wife made impossible the testing of her words, and thus she persuaded your mind. Chiefly upon you do these misfortunes break, but I too feel grief. For the gods do not rejoice at the death of the godly, but the wicked we destroy children, house, and all.94 But mitigating circumstances come by degrees, and are frequently too slight to excuse us from our ‘dreadful deeds’ (e.g. when they are the result of delusion and listlessness as opposed to disability or duress95). We may of course find ourselves in denial about this, or indeed deceived in any number of other self-induced ways.96 Like mortals, the gods do not always get what they want either (think of Artemis in Hippolytus, or for that matter Juno in the Aeneid). They may lose out to other gods (just as one drive or instinct may be overridden by another), or to mortal heroes who can and do succeed in their struggles against gods (just as it is possible for us to fight against our own instincts and overcome our moral weaknesses). The fact that this is always done with the help of another god or motivational drive does not entail that we play no (necessary) role in the production of our actions. Gods and mental states cause mortals to be motivated to act in certain ways, but they do not produce the human actions themselves. Divine motivation typically occurs in the realm of mental antecedents (and the

Constantine Sandis

373

internal and external factors that give rise to them), not in the realm of bodily movement. Such intervention also remains compatible with Aristotelian insights regarding the ways in which we can train ourselves to develop certain habits and dispositions over time, not as an immediate result of volition but through a far slower and less conscious process.97 It is we humans – not gods or mental states – who produce both our dreadful and our heroic deeds; in the latter case we live up to our better selves by evaluating our own drives and instincts through second-order beliefs and desires as when a mortal favours one god over another thus setting off a chain of divine and human events, often tragic.98 Sometimes motivational power and rationality come hand in hand, other times it takes real effort; a powerful god is not necessarily wise, and a weak one not necessarily imprudent.99 If the gods can be said to cause actions at all, it is only in a very indirect sense of meddling and interfering in their affairs. They might do this on both a psychological level (e.g. by endowing them with certain psychological attitudes, putting ideas in their heads, providing them with reasons, giving them courage, causing them to fall in love, lose their nerve, etc.) and a nonpsychological level, for example, by causing natural events which may have the positive effect of enabling, facilitating, or expediting action and – perhaps more typically – the negative effect of impeding, delaying, encumbering, or otherwise hindering it (see § 4 below for examples of nature deities). Such ‘acts of gods(s)’ do not determine behaviour but, at most, help to increase or decrease the likelihood of its occurrence. As noted earlier a god may well, on occasion, cause certain human bodily movements to occur, but in such cases the behaviour in question is never portrayed as mortal action, let alone action that is intentional and/or voluntary (cf. note 84). Divine causation may help to either foster or suppress human ability to initiate rational action at will (and for reasons), but it does not eliminate, supervene upon, join forces with, or characterize it.

4

Fate and facticity

In being motivated by gods or mental states, mortals are not choosing to do (let alone doing) anything, be it voluntarily, intentionally or otherwise. Rather, they find themselves motivated. Simon Blackburn captures the truth of such situations as follows: What is true is that we ‘find ourselves’ with this or that concern, and if the experience is unpleasant, as when we find ourselves shamefully addicted to something, or obsessed by something or someone, or for that matter guilty ceasing to care for things or people we once did care for, then it can feel as if we were assailed from outside, a victim of forces not of our own making. But, after all, it is always forces not of our own

374 Gods and Mental States

making that are responsible for who we are in the first place: in my case they have made me into a middle-aged white male of a certain education and class ... nothing on this earth that makes deliberations is free of his or her natural and acquired dispositions as they do so. You, when you deliberate, are whatever you are: a person of tangled desires, conflicting attitudes to your parents, inchoate ambitions, preferences, and ideals, with an inherited ragbag of attitudes to different actions, situations, and characters. You do not manage, ever, to stand apart from all that.100 These truths may all be captured through accounts of divine intervention, the gods dramatizing not just features of our psychology but also environmental causes, be they man-related – such as our upbringing, education, and particular situation (including the aforementioned values of home, marriage, etc. as well as the actions of other mortals101) – or purely natural as portrayed by nature deities such as the sun (Helios), the ocean (Okeanos), the moon (Selene), the night (Nyx) various winds (Boreas and Euros), and so on.102 A boastful statement of such assorted meddling may be found in Aphrodite’s prologue in Hippolytus: Yet for his sins against me I shall punish Hippolytus this day. I have already come a long way with my plans and I need little further effort. One day when he came from Pitheus’ house to the land of Pandion to see and celebrate the holy mysteries of Demeter, his father’s high-born wife Phaedra saw him, and her heart was seized with a dreadful longing by my design. ... But since Theseus has left the land of Cecrops fleeing the bloodguilt he incurred for the murder of the Pallantidae, and sailed with his wife to this land, consenting to a year-long exile from his home, from this point on the poor woman, groaning and struck senseless by the god of love, means to die in silence, and none of her household knows of her malady. But that is not the way this passion is fated to end. I shall reveal the matter to Theseus and it will come to light, and the young man who wars against me shall be killed by his father with the curses the sealord ... Phaedra noble though she is, shall nonetheless die. I do not set such store by her misfortune as to let my enemies off from such penalty as will satisfy my heart.103 We might, in an existentialist moment, say that Greek and Roman gods represent what Heidegger calls our facticity: the ‘that it is’ of human existence, into which we are ‘thrown’104 (or, according to existentialists such as Paul Tillich, have ‘fallen’ into105). While we all share a common facticity, groups and individuals may also be said to have their own unique facticities, comprised of facts regarding such things as their sex, genes, income, appearance, upbringing, neurology, socio-political surroundings, and so on.

Constantine Sandis

375

These and other aspects of our factual frameworks appear to place serious limitations upon all of our thoughts and actions. To misquote Wittgenstein, a dog may intend to greet its master, but it cannot intend to greet its master the day after tomorrow (it lacks the appropriate conceptual facticity).106 Sartre appropriates Heidegger’s term with the aim of distinguishing it from determinism.107 To conflate the two, he tells us, is to be in bad faith, while to underestimate the constraints of facticity (as one does when attempting telekinesis) is to be in good faith (in the negative sense which this contrasts with the self-recovery that is authenticity).108 The latter begins with the realization that without facticity of any kind, action is impossible. To put it more colloquially: we can only choose to play from the cards we are dealt, but we need to be dealt some cards in order to play at all, cards whose ultimate origins precede our existence. Sartre famously describes this all-too-human desire to deal ones own cards from out of nothingness as ‘the desire to be God’,109 a characterization that owes much to Nietzsche’s causa sui argument: [T]he desire for ‘freedom of will’ in that metaphysical superlative sense which is unfortunately still dominant in the minds of the half-educated, the desire to bear the whole and sole responsibility for one’s actions oneself, and to absolve God, world, ancestors, chance, society from responsibility for them, is nothing less than the desire to be precisely that causa sui and, with more than Münchhausen temerity, to pull oneself up into existence out of the swamp of nothingness by one’s own hair.110 Even ancient gods fall short of this ideal, with the possible exception of Zeus (who is not always clearly separable from fate). One would, however, be in bad faith to infer some kind of hard determinism from this (be it biological, psychological, or any other kind). As Nietzsche proceeds to say: ‘unfree will’ too is a ‘mythology’ which ‘amounts to an abuse of cause and effect’ for ‘in real life it is only a question of strong and weak wills’.111 While we may debate whether or not Nietzsche’s paradoxical ‘become what you are’ outlook presumes the existence of necessitating constraints on the potential strength level of one’s will, there is no doubt that he would have agreed with Sartre’s claim that humans have the ability to transcend their facticity (of which freedom is the nihilation112). Sartre warns further that there is no clear dividing line between facticity and its transcendence, each new situation we find ourselves in being the result of both. There are parallels here with the more radical Buddhist doctrine of Paticca Samuppada (usually translated as ‘Dependent Arising’) according to which all phenomena are causally – and therefore also ontologically – interdependent. The same thought arguably also motivated Nietzsche’s amour fati and the desire for eternal recurrence which follows from it (cf. The Gay Science, § 341) since, given such interdependence, the

376 Gods and Mental States

wish for any aspect of the past to be otherwise is a (de re-) desire for selfnegation. Only tragedy could call for such regret. This interdependence feeds our propensity to mistake effects for causes.113 This adds to the complexity of the ancient debate on moral luck, it not always being clear to what extent any given factor – and the moral possibilities it may enable to or constrain – results from luck, let alone luck that is obviously either good or bad.114 Such tension between what is in and what is out of our hands is a central theme of Sartre’s own dramas, particularly The Flies, his version of the Elektra myth. In ancient tragedy the conflict is played out by Zeus in his two guises: as the head of a divine family and as Fate. We must not, however, fall into the trap of conflating fatalism with causalism. The necessity alluded to in the Greek claim that ‘there is nothing here that is not Zeus’ is that of facticity, not of scientific determinism. People’s fates are typically shaped rather than produced by their facticity.115 Even in its strongest form, fatalism remains nothing more (or less) than a belief in extended forms of facticity, albeit of a kind no less mystical that the opposite belief of thinking that one can control external events through sheer volition. A person who takes astrological predictions seriously is no more committed to determinism than one who uses expressions such as ‘it was meant to be’, ‘knowing my luck’, ‘he’s a jinx’, and the ‘the devil made me do it’. This remains true even in cases where, for whatever reason, we choose to interpret fatalistic language literally instead of metaphorically, thus treating it as mystical rather than merely poetic (which is not to say that none of us are superstitious or religious116). There is no obvious reason why we should judge the pronouncements of the ancients any differently.117 Indeed we have already seen that, despite its focus on tragic or otherwise abnormal circumstances, the divine theatre of the ancients is considerably less causalist than that of most current philosophy of mind, the latter marked by a tendency to conflate the motivation of action with its production. ‘A people gets the gods which it deserves’ C.M. Bowra famously remarked,118 this is certainly true of the majority of modern scholars who model their philosophies of action on the natural sciences.119

Notes 1. As this essay is primarily written for philosophers who may not necessarily also be classicists, all Greek quotations shall be provided in translation (where appropriate always the ones used by the critics I discuss). Throughout the essay I use ‘ancient tragedy’ to refer to both Greek drama and Greek and Roman epic, the term ‘gods’ to refer to gods, goddesses, and daemons (be they nature deities or Olympians) and – despite strong personal reservations – the term ‘mental states’ to refer to psychological phenomena such as beliefs, desires, emotions, and so on. The terminology is intended to emphasize common ground, without implying that there are no important differences (such as that Sophocles is a far more religious writer than Euripides). For the view that epic is not a form of tragedy (where ‘tragedy’ is understood in a rather technical sense) see H.D.F. Kitto, The

Constantine Sandis

2. 3.

4. 5.

6. 7. 8.

9.

10. 11.

12. 13. 14. 15. 16. 17.

18. 19. 20. 21. 22.

377

Greeks, (Pelican, 1951), p. 59f. For a ‘non-statist’ account of most mental phenomena that I am in great sympathy with see Arthur Collins’ The Nature of Mental Things (Notre Dame: University of Notre Dame Press, 1987). R.P. Winnington-Ingram, ‘Hippolytus: A Study in Causation’, in (ed. Olivier Reverdin) Euripide, Entretiens sur l’antiquite classique, VI ( Fondation Hardt, 1960). I shall also use the terms ‘mortal’ and ‘human’ interchangeably even though mortal demi-gods are, strictly speaking, not purely human. I remain neutral, here, on whether we should view this semi-divine status as a cause or effect of divine intervention. Winnington-Ingram (op. cit.). E.R. Dodds, The Greeks and the Irrational (University of California Press, 1951), p. 30ff , referred to in R.P. Winnington-Ingram, ‘Zeus in the Persae’ (The Journal of Hellenic Studies, Vol. 93, 1973) p. 213, note 14. Dodds 1951, p. 2. Ibid., p. 31. Cf. p. 7 and p. 16 where he talks of such actions as overdetermined events. Nilsson, A History of Greek Religion, 2nd edn. Trans. by F.J. Fielden (OUP, 1952) pp. 163–5. Originally published as Den Grekiska Religionens Historia (Svenska Kyrkans Diakonistyrelses Bokförlag, 1921). But compare the above quotation to the following aphorism from Nietzsche: ‘ “I have done that” says my memory. “I cannot have done that” – says my pride, and remains adamant. At last – memory yields.’ Beyond Good and Evil (1886), § 68 (trans. by R.J. Hollingdale, 1973). Dodds, op. cit., p. 41. Cf. Harry Frankfurt’s ‘Identification and Externality’, reprinted in The Importance of What We Care About (CUP, 1988), pp. 58–68. Frankfurt rightly further distinguishes the internal/external and the identification/alienation distinctions from the active/passive distinction, all of which he takes to somehow relate to the action/happening distinction. Les origines de la pensée grecque (Paris, 1962) & Mythe et pensée chez les Grecs: Etudes de psychologie historique (Paris, 1965). W. Burkert, Greek Religion (Oxford: Blackwell, 1985), ch. 3. Cf. Jan. N. Bremmer, Greek Religion (Cambridge University Press, 1999), pp. 22–3. Dodds, op. cit. p. 30. H.D.F. Kitto, Sophocles-Dramatist and Philosopher (Oxford University Press, 1958), p. 57. Vernant (Op.cit. 1962 and 1965); Zaidman & Pantel, Religion in the Ancient Greek City (CUP, 1992), trans. P. Cartledge, p. 177; Oudemans & Lardinois, Tragic Ambiguity: Anthropology, Philosophy and Sophocles’ “Antigone”, (Leiden: Brill, 1987), p. 94. Op. cit. p. 31. Ibid. p. 38. Ibid. pp. 40–2. 1951, p. 57 (cf. Kitto 1958, p. 53 and p. 57). Hegel, Aesthetics – Lectures on Fine Art (OUP, trans. T. M. Knox, 1975), p. 220ff. Cf. his Introductory Lectures on Aesthetics, XIV. For related exegesis see Kenneth R. Westphal’s Hegel’s Epistemology: A Philosophical Introduction to the Phenomenology of Spirit (Indianapolis: Hackett, 2003, § 5) and Michael Inwood’s Commentary to the Introductory Lectures (London: Penguin, 1993), p. xxii and p. 103 (note 4 on XIV).

378 Gods and Mental States

23. 24. 25. 26.

27. 28.

29. 30. 31. 32.

33. 34.

35.

36. 37. 38. 39.

40.

Contrast this with M. Nussbaum’s anti-Hegelian (particularist) interpretation of Antigone in The Fragility of Goodness (CUP, rev. ed. 2001), ch. 3, IV (pp. 67–79). 1958 (op. cit.), p. 43. 1951, p. 53. B. Williams, Shame and Necessity (University of California Press, 1993) p. 29. Kitto concedes that the divine may intervene in human affairs (e.g. he allows that Athena makes Ajax mad at the last moment) but claims that such intervention is ‘confined to the plot’ and ‘does not affect the significance of the play, and is no part of Sophocles’ philosophy or religion’ (1958, p. 56). A. Poole, Tragedy: A Very Short Introduction (Oxford University Press, 2005), p. 50. I am thinking here primarily of issues relating to action theory (agency, control, free will, responsibility etc.) and to ethical conflicts such as that between individual and social good. Cf. Nicholas White, Individual and Coflict in Greek Ethics (OUP, 2002). For a wonderfully informative examination of ‘group identity’ at drama festivals see Edith Hall’s ‘The Sociology of Athenian Tragedy’ in (ed. O.E. Easterling), The Cambridge Companion to Greek Tragedy (Cambridge: Cambridge University Press, 1997), pp. 93–126. Cf. B. Williams, op. cit., p. 29. See also Williams’ ‘Understanding Homer’ in his The Sense of the Past (New Jersey: Princeton, 2006). A.I. Goldman, A Theory of Human Action (Princeton University Press, 1970), p. 138. For relevant discussions see the papers by Baier, Boulter, and Price in this volume. Donald Davidson, ‘Actions, Reasons, Causes’, Journal of Philosophy 60 (1963). Reprinted in Essays on Actions and Events (London: Clarendon Press, 1980) pp. 3–19. Julia Tanney’s essay in this volume neatly places the paper in its historical context, and urges for a resurrection of some of the views which Davidson rejected. But see note 1 above. Since gods and mental states are frequently not only thought to be determinants of our behaviour, but also taken to be equally capable of explaining them, the worry operates at both a causal and an explanatory level. Essays on Actions and Events, p. 65; the quotation is from the paper ‘Freedom to Act’, originally published in (ed.) T. Honderich, Essays on Freedom of Action (London: Routledge and Kegan Paul, 1973). See p. 129 of Essays on Actions and Events for Davidson’s related pessimism regarding the possibility of an analysis of the concept of agency in terms of other concepts. ‘Fatalism and Determinism’ in Keith Lehrer (ed.) Freedom And Determinism (New York: Random House, 1966), p. 144. Emphasis in the original. Ibid., p. 156. Of course talk of ‘what caused her to do x’ need not imply determination of any kind (cf. G.E.M. Anscombe, Intention (Oxford: Blackwell, 1957), §§ 9–12). Davidson unhelpfully oscillates between talk of what is free, what is voluntary, and what is in our control. My interest in this essay is confined to the last of these three (to what we can do at will), regardless of whether we do so voluntarily, under duress, and so on. Linguistic nuances will vary, of course. The phrase ‘X caused her to act’, for example, typically seems to allow the agent more control over her action than the phrase ‘X caused her to perform the action’, presumably because in the former case the term ‘cause’ can be used in a figurative sense more readily. But the subtleties of ordinary speech should, in any case, be of no concern to causalists like Davidson who are working with a fixed analysis of the logical form of singular causal statements such as ‘the heat caused Samantha’s return to Patma’, according

Constantine Sandis

41.

42.

43. 44. 45. 46. 47.

48. 49.

50. 51.

52. 53.

54. 55. 56.

57. 58. 59. 60.

379

to which the term ‘caused’ means something like ‘causally explains’ where, crucially, the success of the explanation relies at the very least on its being true that ‘if A causes B, there must be descriptions of A and B which show that A and B fall under a law’ (Cf. Essays on Actions and Events, p. 162 and p. 262). Michael Smith ‘Realism’ in Peter Singer (ed.) A Companion to Ethics (Oxford: Blackwell, 1991), p. 399–410. Cf. his ‘Moral Realism’ in H. LaFollette (ed.) The Blackwell Guide to Ethical Theory (Oxford: Blackwell, 2000), p. 33ff. ‘How is weakness of the will possible?’, originally published in (ed.) J. Feinberg, Moral Concepts (Oxford: OUP, 1970). Reprinted in Essays on Actions and Events, p. 22. ‘Actions, Reasons, Causes’, op. cit. p. 9. Essays on Actions and Events, pp. 35–6.; the quotation is from ‘How is weakness of the will possible?’. Ibid., p. 23. Ibid., Davidson’s claim that this states ‘a mild form of internalism’ (p. 26) is somewhat modest. This is not to say that it renders epiphenomenal the reasons themselves for these could still structurally explain why an action was triggered by a certain event. Cf. Fred Dretske’s essay in this volume, as well as his Explaining Behavior (Cambridge, MA; MIT Press, 1988). Cf. Nietzsche, Twilight of the Idols (1888), The Four Great Errors – 3 (‘The Error of False Causality’). But see A.W. Price’s ‘Aristotle’s Conception of Practical Thinking’ in this volume. Cf. Michael Stocker’s ‘Raz on the Intelligibility of Bad Acts’ in (eds) R.J.Wallace, Philip Pettit, S. Scheffler, & M. Smith, Reason and Value: Themes from the Moral Philosophy of Joseph Raz (OUP, 2004) pp. 303–4 and Peter Railton’s ‘How to Engage Reason’ (same volume, pp. 178–9). J. Raz, Engaging Reason (OUP, 1999) p. 47. See J. Raz, ‘Comment on Dancy on Practical Reasoning and Inference’ (2nd On-line Philosophy Conference, May 2007), pp. 4–5. This is a response to Jonathan Dancy’s paper at OPC2, both available, along with Dancy’s reply, at http:// experimentalphilosophy.typepad.com/2nd_annual_online_philoso/. See also Raz’s ‘Reason, Reasons & Normativity’ (forthcoming). Ibid., p. 115, my italics. For Davidson’s characterization of the classical conception see his Essays on Actions and Events, p. 35. Engaging Reason, p. 116; here Raz also suggests that to follow ones will is to choose to act upon a reason (though not necessarily the one which you judge to be strongest). CF. H. Frankfurt, The Importance of What We Care About (e.g. p. 164) and G. Watson, ‘Free Agency’, The Journal of Philosophy, 1975 (e.g. p. 218). A Theory of Human Action, pp. 72ff. E.g. A.I. Goldman, ‘The Volition Theory Revisited’ in Brand and Walton (op. cit.), p. 81. Cf. pp. 68–9 for Goldman’s helpful overview of the philosophical history of the concept. See also references to Sellars, below. P. Railton, ‘How to Engage Reason: The Problem of Regress’ in (eds) R. J. Wallace, Philip Pettit, S. Scheffler, & M. Smith (op. cit.) pp. 180 & 199. ‘Fatalism and Determinisn’, p. 156. ‘Volitions Re-affirmed’ in M. Brand and W. Walton (eds) Action Theory (Reidel Publishing, 1976) pp. 47, 53, and 63; emphasis in the original. Cf. Goldman (op. cit.) p. 83.

380

Gods and Mental States

61. Op. cit. p. 200, emphasis in the original. 62. Even this is contentious. If a part of me (say my heart) causes something (e.g. blood circulation) it does not obviously follow that I have caused my blood to circulate. 63. D. Hume, A Treatise of Human Nature (2nd edn Selby-Bigge, OUP), 407–8. Cf. Gary Watson ‘Free Action and Free Will’, in Agency and Answerability (OUP, 2004), pp. 161–2. 64. See Annette Baier’s essay ‘Acting in Character’, in this volume. 65. Davidson, op. cit., p. 36. 66. Cf. Essays in Harry Frankfurt’s The Importance of What We Care About (CUP, 1988), Joseph Raz’s Engaging Reason (OUP, 1999), Michael Bratman’s Faces of Intention (CUP, 1999), and Structures of Agency (OUP, 2007), J. David Velleman’s The Possibility of Practical Reason (OUP, 2000) and Self to Self (CUP, 2006), Gary Watson’s Agency and Answerability (OUP, 2004), and John Martin Fischer’s, My Way: Essays on Moral Responsibility (OUP, 2006). These volumes chronicle philosophical progress and thereby lack internal (as well as external) consistency, but they all propose forms of volitional identification as a source of accountability. This project should not be conflated with the parallel one of appealing to volitional identification as a source of normativity, most notably undertaken by Christine Korsgaard in The Sources of Normativity (CUP, 1996). 67. M. Bratman, ‘Shared Valuing and Practical Reasoning’ in Reason and Value (op. cit.), p. 29. 68. H. Frankfurt, ‘Freedom of the Will and Concept of a Person’ Journal of Philosophy, 68 (1971). Reprinted in The Importance of What We Care About, pp. 14–15. 69. H. Frankfurt, ‘The Problem of Action’, American Philosophical Quarterly, 15 (1978). Reprinted in The Importance of What We Care About, p. 69. 70. Ibid., p. 75. 71. Freedom of the Will and Concept of a Person’ (op. cit.) p. 16. 72. The Importance of What We Care About (title essay), p. 86. For various interpretations see Gary Watson ‘Volitional Necessities’ § IV, reprinted in his Agency and Answerability, pp. 100–122. 73. G. Ryle, The Concept of Mind (London: Penguin, 1942), ch. 3. For related arguments see Railton (op. cit., pp. 196–9) and Davidson (‘Freedom to Act’, op. cit., pp. 63–4). E.J. Lowe outlines an anti-causalist response to such worries in his Locke on Human Understanding (Routledge, 1995) pp. 124–8. 74. H. Frankfurt ‘Alternate Possibilities and Moral Responsibility’ Journal of Philosophy 66 (1969) (reprinted in The Importance of What We Care About), p. 10, my emphasis. For the plausible view that Frankfurt’s rejection of the weaker principle that ‘a person is not morally responsible for what he has done if he did it because he could not have done otherwise’ begs the question against libertarians and/or irreducible agent causation see D. Widerker ‘Libertarianism and Frankfurt’s Attack on the Principle of Alternative Possibilities’ Philosophical Review 104, April 1995, pp. 247–261 & ‘Libertarian Freedom and the Avoidability of Decisions’, Faith and Philosophy 12, 1995, pp. 113–8, H.J. McCann, ‘Agency, Control, and Causation’ in The Works of Agency (Cornell University Press, 1998) pp. 173–9, and Timothy O’ Connor Persons & Causes (OUP, 2000), pp. 81–4. Theoretically unmotivated criticisms may be found in M. Alvarez’s excellent ‘Actions, Thought-Experiments, and the “Principle of Alternate Possibilities”‘, The Australasian Journal of Philosophy, 2008.

Constantine Sandis

381

75. It is tempting to write ‘for certain reasons’ here, but it is unclear to me whether an agent whose will is both necessitated and necessitating could sensibly be said to act in the light of reasons. 76. Cf. my ‘The Explanation of Action in History’, Essays in Philosophy, Vol. 7, No. 2, June 2006. 77. See § 4 below. 78. Cf. T. Reid, Of Active Power in General (1788, Ch. 1) and A. Kenny, Will, Freedom, and Power (Blackwell, 1975), R. Kane, The Significance of Free Will (OUP, 1996), p. 33, and H.J. McCann (op. cit.). For the view that one can endorse a compatabilist interpretation of the principle of alternate possibilities see Watson (op. cit.) pp. 300–301. 79. For detailed criticism of the mechanistic aspect of all causalist accounts see Norman Malcolm’s ‘The Conceivability of Mechanism’, The Philosophical Review, Vol. 77. No.1 (January 1968), pp. 45–72. 80. Op. cit. p. 178. One difference between the ‘student’ viewpoint and the one I shall be ascribing to the ancients is that the latter allow that alien forces may enable choice as much as they might prevent it. 81. Homer, Odyssey, trans. by Samuel Butler (Cape, 1922). 82. Virgil, The Aeneid, trns. David West (Penguin Classics, 2003), introduction. See also J-P Vernant’s Myth and Thought Among the Greeks (New York: Zone Books, 2006, trans. J. Lloyd and J. Fort), p. 116. 83. Ibid. Cf. Albin Lesky’s Göttliche und menschliche Motivation im homerischen Epos (Heidelburg: Academie Press, 1961). 84. Notwithstanding rare incidents of single divine causation in the heat of battle (where the lines between intentional action and non-intentional reaction begin to blur) such as that of Athene forcing a spear in Ares’ belly in Book 5 of The Iliad. 85. This lack of space is variously understood by competing causalist views as being conceptual, metaphysical, and/or physical. 86. For an account of the role of desire in reason-giving explanations of action see Maria Alvarez’s paper in this volume. For a complementing view of role of beliefs see her ‘Reasons and the Ambiguity of “Belief” ’, Philosophical Explorations, 2008. 87. Cf. Shame and Necessity, pp. 29–32 for examples and elaboration. 88. It is neither here nor there whether the causality in question here is determinist or indeterminsist. 89. Euripides, Hippolytus, trans. by David Kovacs (Harvard University Press, 1996), pp. 240–49. 90. Ibid., p. 319. 91. For Presocratic interpretations of Aphrodite’s role see Jonathan Barnes, The Presocratic Philosophers, (Routledge, 1979), pp. 117–8. 92. Ibid., pp. 473–5. 93. Ibid., pp. 476–80. 94. Ibid., pp. 1328–40: For the view that the category of ‘the will’ did not fully emerge until Aristotle see J-P Vernant’s ‘Intimations of the Will in Greek Tragedy’ in his Myth and Tragedy in Ancient Greece (New York: Zone Books, 1988, trans. J. Lloyd). My overall argument would not be affected if we chose to translate ‘boulesis’ as ‘wish’ instead of ‘will’ (cf. André Rivier’s work on necessity in Aeschylus). 95. Cf. Shame and Necessity (Ch. 3) for some interesting observations on the relation between responsibility and temporary losses of selfhood and/or control.

382

Gods and Mental States

96. For a stock of examples from Euripides cf. William Allan, The Andromache and Euripidean Tragedy (OUP, 2000), p. 246. See § 3 below for issues relating to bad faith. 97. Cf. Annette Baier’s essay in this volume which gives a clear account of Hume’s understanding of such processes. 98. Ibid., p. 239ff. 99. Ibid., p. 265, fn. 125. 100. Cf. Simon Blackburn, Ruling Passions (Clarendon Press, 1998), p. 252. Blackburn further embraces the view that attitudes have a standard motivational role. For a critique of this position see Nick Zangwill’s paper in this volume. 101. Think, for example, of the impact of Agamemnon’s actions at Aulis on Clytaemestra. 102. Cf. Burkert (op. cit., ch. III). As William Allan notes in relation to Euripides they do so ‘both as characters in the plays and as unseen, enigmatic agents shaping the action from outside’ (Op. cit. p. 233). Allan also discusses W. Burkert and J-P Vernant’s contrasting approaches to how the gods are meant to be perceived before agreeing with J.M. Bremmer’s suggestion that we should view them as both persons and powers, without characterising their existence as being either ‘fictive’ or ‘actual’ (a point he takes from C. Sourvinou-Inwood): ‘we might state that the gods, without necessarily being the gods that the audience believe in, are to be taken seriously as agents within the play, acting with both power and personality’ (p. 236). 103. Op. cit. 23–51. 104. Cf. M. Heidegger, Being and Time (1927; UK trans. by J. Macquarrie & E. Robinson, Blackwell, 1962), especially p. 56 – where ‘Dasein’s facticity’ is defined in relation to, but also distinguished from ‘factuality’– and p. 135 – where ‘throwness’ is defined in relation to ‘facticity’. 105. P. Tillich, Systematic Theology Vol. 2. (1957, University of Chicago Press), Part B. 106. Cf. Philosophical Investigations (Blackwell, 1953, § 174). 107. In Hazel E. Barnes’ ‘key to special terminology’ which accompanies her landmark translation of Sartre’s Being and Nothingness (Routledge, 1958) factticity is defined as ‘The For-itself’s necessary connection with the In-Itself, hence with the world and its own past. It is what allows us to say that the For-itself is or exists. The facticity of freedom is the fact that freedom is not able to be free’ (p. 631, cf. p. 439ff). 108. Cf. op. cit. p. 70, inc. fn 9: ‘Good faith seeks to flee the inner disintegration of my being in the direction of the in-itself which it should be and is not. Bad faith seeks to flee the in-itself by means of the inner disintegration of my being . ... If it is indifferent whether one is in good or bad faith, because bad faith reapprehends good faith and sides to the very origin of the project of good faith that does not mean that we cannot radically escape bad faith. But this supposes a self-recovery of being which was previously corrupted. This self-recovery we shall call authenticity. The description of which has no place here’. Joseph S. Catalano capture’s Sartre’s thought here well when he writes that ‘bad faith is an attempt to flee from our freedom, whereas good faith is an attempt to face our freedom’ (A Commentary on Jean-Paul Sartre’s Being and Nothingness, University of Chicago Press, 1974, p. 89; see also p. 87 for how good and bad faith compare in their attitudes towards factual evidence). 109. Being and Nothingness, p. 566. 110. Beyond Good and Evil (1886) § 21.

Constantine Sandis

383

111. Ibid. 112. Op. cit., p. 444. 113. Cf. F. Nietzsche, Twilight of the Idols, The Four Great Errors – 1 (‘The error of confusing cause and effect’). Sartre’s own radical view was that motives were parts of actions rather than causes of them and that ‘under no circumstances can the past in any way by itself produce an [intentional] act’ (op. cit. p. 436). 114. Cf. Martha Nussbaum, The Fragility of Goodness (Cambridge University Press, 2nd updated edition, 2001) esp. pp. xii–xvi, 1–22, & 378–394; Bernard Williams, Moral Luck (Cambridge University Press, 1981), esp. Ch. 2; D. Statman, Introduction to Moral Luck, (State University of New York Press, 1993) & (ed.) Moral Luck (State University of New York Press,1993); N. Athanassoulis, Morality, Moral Luck and Responsibility: Fortune’s Web (Palgrave Macmillan, 2005). 115. Cf. Being and Time (op. cit.) p. 384; here Heidegger relates facticity to what he calls ‘fateful destiny’, noting that ‘our fates have already been guided in advance, in our Being with one another in the same world and in our resoluteness for definite possibilities’. 116. Cf. D.Z. Philips, The Problem of Evil (Minneapolis: Fortress Press, 2004), p. 175. 117. As William Allan writes (op. cit., p 237) ‘we simply do not know how far ordinary people thought in terms of gods rather than of human motives when interpreting the behaviour of others and of themselves’, a point he takes from Winnington-Ingram’s Euripides and Dionysus: An Interpretation of the Bacchae (Cambridge University Press, 1948). 118. The Greek Experience (Weidenfeld and Nicholson, 1957), p. 41. 119. Many thanks to Elizabeth Sandis for many stimulating comments and discussions, as well as for reading an earlier draft and making helpful suggestions. This paper was written with financial help from the “Aspectos Modales del Realismo Materialista” project, HUM2007–61108, MCYT-Spanish Government. A previous version was presented (in absentia) at the Australasian Society for Classical Studies XXVI, University of Otago, New Zealand, 31 Jan–2 Feb 2005.

20 Aristotle’s Conception of Practical Thinking A. W. Price

I One thing that philosophy owes to Aristotle is a recognition of a kind of thinking that is practical by nature. In De Anima [DA] III 10, Aristotle spells this out as follows: Both of these can produce movement in respect of place, intellect and desire, but intellect which reasons for the sake of something and is practical; and it differs from the contemplative intellect in respect of the end. Every desire too is for the sake of something; for that of which there is a desire is the starting-point of the practical intellect, and the final step is the starting-point of action. Hence it is reasonable that these two appear the sources of movement, desire and practical thought. For the object of desire produces movement, and, because of this, thought produces movement, because the object of desire is its starting-point (433a13–20, after Hamlyn, 1993). ‘Practical’ (praktikos) connotes at least productive of action (praxis). The Eudemian Ethics [EE] calls virtue of character a state that makes us ‘doers of what is best’ (praktikoi tôn beltistôn, II 5, 1222a6–7, tr. Woods, 1992). It is doubtless in the same sense that a man is said to be practically wise (phronimos) not by knowing only, but by being praktikos (AE III 10, 1152a8–9). (I follow Kenny, 1979, in referring to the three ‘common books’, which constitute at once Nicomachean Ethics [EN] V to VII and EE IV to VI, as AE, i.e. Aristotelian Ethics, A to C.) But how are we to understand the preceding ‘for the sake of something’ (heneka tou, a14)? And what is the ‘end’ (telos) of practical intellect? There would seem to be two alternatives. The clearest contrast with theoretical intellect is made if (A) the end of practical intellect, that for whose sake it reasons, is action. If so, the ‘and’ that links ‘reasons for the sake of something’ to ‘is practical’ is epexegetical, clarificatory rather than conjunctive, as is common in Greek. One may then compare Metaphysics II 1, 384

A. W. Price

385

993b20–24: ‘The end of theoretical knowledge is truth, while that of practical knowledge is action (ergon) (for even if they consider how things are, practical men do not study what is eternal but what stands in some relation at some time).’ However, one might take it instead that (B) the end of practical reasoning is not action, but the end of action, the further goal for the sake of which the agent both reasons and acts. Thus, in the De Motu Animalium [DMA] (6, 700b23–5), ‘the end in the sphere of action’ (to tôn praktôn telos) is at once the object of desire (to orekton), and the first cause of movement (kinei prôton); hence it cannot be action itself, but must rather be the goal of thought and action. Thus in (A) the end of practical reasoning is action, whereas in (B) it is the end of action. (A) was argued by Richard Loening (1967), with a further purpose. He was reacting against a tendency to read a Humean division of labour into such remarks as this: ‘Virtue [sc., of character] makes the target (skopos) right, phronêsis the things that lead to it’ (AE B 12, 1144a7–9). Loening wished to avoid any suggestion that it is the role of phronêsis, as a slave of the passions, to devise means towards the ends set by the virtue of character that is the product of the training of desire. Hence he preferred to construe the intellect ‘which reasons for the sake of something’ as the intellect which aims at action, and not the intellect which reasons on behalf of some goal set by character. He is quite right to observe that the words ‘Every desire too is for the sake of something’ (433a15) cannot mean that every desire is at the service of some preordained end. They might just be an awkward way of expressing the truism that every desire has an object. Yet it seems best to read them as an ‘Aristotelian categorical’ (see Foot, 2001, pp. 29–35): though not every idle desire is defective, it remains the general function of desire to generate action, and more particularly, in the case of human beings, rational action. The sentence continues, ‘For the object of desire is the starting-point for the practical intellect, and the final step is the starting-point for action’ (a15–17). All this hangs well together if Aristotle’s point is that desire is for the sake of action, and the ensuing clauses explain how desire leads to action through the operation of the practical intellect. The DA passage may still seem problematic for Loening. For it contains the sentences, ‘That of which there is a desire is the starting-point for the practical intellect’ (433a15–16), and ‘Thought produces movement, because the object of desire is its starting-point’ (a19–20). Which might suggest a Humean view: an end of action is presented to intellect by desire. Yet Loening has a point to make against this (1903, pp. 34–5). Aristotle later remarks that the object of desire ‘produces movement without being moved, by being thought of or imagined’ (433b11–12). Thus the object of desire, to orekton, is simultaneously an object of thought, a dianoêton (DMA 6, 700b24). Any object of desire, whether it be actual or proper, has to be presented to desire as an object of thought or imagination (cf. 7, 702a18–19): a man cannot desire anything of which he does not already have a conception. And that

386

Aristotle’s Conception of Practical Thinking

conception might be provided by the intellect that has a mastery of concepts. However, it may still be true that distinctively practical thinking, thinking for the sake of an end, only ensues once there is not only a conception, but a desire, for an end. So (B) remains an option. Intellect may have a prior role to play in the selection of an end; and yet thi nking that counts as practical may still be thinking in the service of an already desired end. The other crucial paper for Loening is AE B 2. This states, ‘Thought itself moves nothing, but that which is for the sake of something and practical’ (1139a35–6). Again the question arises: is ‘practical’ straightforwardly explanatory of ‘for the sake of something’? If so, we have reading (A): practical thought occurs for the sake of action. Alternatively, practical thought may be for the sake of the end of action, which is reading (B). What is new in the AE context is reference to ‘acting well’ (eupraxia) as the end: it is what is to be done (to prakton) that is an end absolutely, ‘for acting well is the end and desire is for this’ (b2–4). And here we find a way, I think, of reconciling (A) and (B), so that Aristotle’s ambiguity becomes perfectly intelligible – possibly, even, motivated. A proper articulation of Aristotle’s conception of deliberation requires a distinction between three objects of desire that lie at different points of the spectrum from the maximally general to the maximally specific. Somewhere in the middle falls the ‘starting-point’ (archê) of a process of deliberation. Thus the EE states, ‘Those who have no target before them are not in a position to deliberate’ (EE II 10, 1226b29–30, cf. 1227a6–7) immediately after offering the specific example ‘The carrying of goods is a cause of walking if it is for the sake of that that a man walks’ (1226b28–9). And we read this in the common books: ‘Syllogisms of things to be done have a starting-point, namely, since the end, or what is best, is such-and-such, whatever that may be (let it for the sake of argument be whatever you like). This is only apparent to the good man; for wickedness perverts us and causes us to be deceived about the starting-points of action’ (AE B 12, 1144a31–6). Here the ‘starting-points of action’ are explicitly plural. It is indicated that they vary not only between good and vicious agents, but also, for any given virtuous agent, from occasion to occasion. So we should think of goods such as wealth and pleasure that are often but not always properly pursuable: the vicious agent pursues them too often, the virtuous agent only on occasions when they can be pursued commendably. Through deliberation, the partly specific starting-point is translated into the choice of an act that the agent can perform without more ado. So Sarah Broadie states, ‘The agent’s view of the fact is sufficiently determinate (and the rational choice is made) as soon as he perceives a means immediately within his power and at the same time knows that how precisely he handles it is of no negative evaluative significance’ (1991, p. 247). His goal is not only to achieve whatever partly specific goal set his deliberation in train, but to do so in a manner that has minimal ‘negative evaluative significance’.

A. W. Price

387

We must take into account statements of a more open-ended goal of deliberation: ‘It is thought to be a mark of a man of practical wisdom to be able to deliberate well about what is good and beneficial for himself, not in some particular respect, for example, about what sorts of thing conduce to health or to strength, but about what sorts of thing conduce to the good life in general’ (AE B 5, 1140a25–8). The practically wise man is the man who is ‘without qualification good at deliberating’, being ‘capable of aiming in accordance with calculation at the best for man of things attainable by action’, presumably living well (II 7, 1141b13–14). Some have inferred from these statements that the ultimate starting-point of deliberation, lying behind the fairly specific goals or rules that appear in Aristotle’s examples, is nothing short of the determinable end of all action, which is eudaimonia itself; and that this requires a different style of deliberation that brings to bear an inarticulable conception of what eudaimonia comes to, either in general or in the present circumstances of action. I argue elsewhere (Price, forthcoming b) that this conception is unsupported in our texts. And yet there is evidence that Aristotle does associate deliberation with the wise selection of a goal of action: ‘If, then, it is characteristic of phronimoi to have deliberated well, excellence in deliberation will be correctness with regard to that which conduces to the end of which phronêsis is the true supposition’ (AE B 9, 1142b31–3). As I read the implications of this, it must mean that it is through deliberation that phronêsis tests the acceptability of a provisional goal. The deliberator has to decide not merely whether some proposed end is achievable in his situation, but whether it is acceptably achievable; if it is, achieving it can count, in context, as acting well. This may only become clear to him when he has a pretty specific conception of what it would be to achieve the end here and now. (Illuminating here remains Woods, 1986, 163–4.) As we might put it, desire proposes, but intellect disposes. Deliberation may complicate the initial target, or even require that it be discarded or replaced. Rational desire selects a target with a reservation, that is, subject to concrete confirmation, through deliberation, of its acceptability in context. Let us take an example from Aristotle himself: ‘With what purpose did he come? In order to get the money. And that in order to pay back what he owed; and that in order not to act unjustly’ (Posterior Analytics I 24, 85b30–32). Here there are three acts: of coming, of getting the money and of paying back what he owed. It cannot be inferred that the agent actually started his deliberation by wondering how, in context, to avoid injustice. Yet let us suppose that something in his situation initially oriented him towards justice as an appropriate end. He then calculates that, in order to act justly, he must now repay a particular debt. And all this is in the service of his acting well. When he pays the debt, he will be performing an action that falls under a sequence of descriptions: handing over certain coins, repaying a certain debt, acting justly, acting well. When he reflects how to act, wondering how

388

Aristotle’s Conception of Practical Thinking

to act both justly and well, this will be thinking for the sake of action. More precisely – since this is a case of full deliberation and full action, unlike the misplaced calculations of the acratic (AE B 9, 1142b18–20) – it will be thought for the sake of acting well. Its starting-point will be a provisional specification of a general way of acting that plausibly counts, in context, as acting well. Thus the agent deliberates at once for the sake of action, for the sake of achieving his initial goal (if it stands up to scrutiny), and for the sake of acting well. Aristotle does not have to choose between these, for they coincide.

II So much can be said to flesh out more concretely what Aristotle has in mind by ‘intellect which reasons for the sake of something and is practical’ (DA III 10, 433a14). There is also a more general question to be raised. Should we think of deliberation as a stretch of reasoning that is not itself distinctively practical, but which happens to be conducted with a practical purpose? Or should we think of it as a distinctive kind of reasoning that is inherently practical? This question has been explored by Anselm Müller. He noticed that it is not enough to say that thinking is practical if it is pursued with a certain purpose. Suppose I want to think intelligently: then I could do some thinking in order thereby to be thinking intelligently. That would not be practical thinking. Müller gives this example (1982, pp. 238–9): ‘I can, e.g., have the intention when I do mental arithmetic of practising my understanding, or showing off the speed of my replies, or passing the time.’ ‘Such intentions’, he remarks, ‘in no way make my thinking practical.’ It isn’t even enough if the thinking takes a purpose as its theme: ‘So could one say: to be practical, thinking must proceed in its content from a goal for whose sake, moreover, it occurs? Even this isn’t enough. If a prisoner doesn’t do mental arithmetic in order to pass the time, but reflects about how to pass the time, this thinking is necessarily theoretical: it indeed occurs to pass the time – but not with the intention to apply the results of the thinking.’ Instead, he offers the following formulation: ‘One gets clear through what means (of one’s own) B can be realized, with the intention of realizing B through such means’ (1982, p. 239). Yet even this, he says, is inadequate. It fails to capture that a piece of practical thinking ‘involves consciousness of its own practical function as well as a judgement relating means to that end’ (1979, p. 98). The idea is this: just as a sequence of acts (in the example I cited, coming, getting the money, paying back what one owes) is all directed towards a goal (there, acting justly), so the sequence of thoughts that constitute a piece of deliberation is also all directed towards a practical goal. Yet, in their case, the purposiveness has to be, Müller argues, ‘unreasoned’. For it would be incoherent to reason as follows about what form the deliberation should take: ‘I will w; a (or the) way to find out how to do this is to reflect that my

A. W. Price

389

intention to w will be best realized by c-ing; so let me reflect that my intention to w will be best realized by c-ing.’ As Müller comments, ‘This is a problem when the first order reflection is represented by a conception of its very content’ (1992, 166). Rather, the deliberation is inherently purposive, in a manner of which the agent is conscious, but not in a way which involves direction by some higher order process of thinking. We find no such elaboration in Aristotle; and yet we can view it as, broadly, Aristotelian in that it articulates a conception that is evidenced in Aristotle. There is, first, the very way in which he writes of practical intellect in DA III 10: here, ‘intellect which reasons for the sake of something and is practical’ is presented as a special kind of intellect, differing in its end from theoretical intellect (433a14–15). It is true that it is through its connection with desire that practical intellect causes movement: ‘There is one thing which produces movement, the faculty of desire. For it there were two things which produce movement, intellect and desire, they would do so in virtue of some common form; but, as things are, the intellect does not appear to produce movement without desire’ (a21–3). How Aristotle places rational desire in relation to reason, on the one hand, and the irrational soul, on the other, is a difficult question, to which there may not be an unequivocal answer. Statements in the Topics (IV 5, 126a13) and DA (III 9, 432b5) that wish (boulêsis) occurs in the rational part (to logistikon) may represent inferences from Plato rather than Aristotle’s mature view, of which there are conflicting indications. (For a discussion of the difficulties, but what may well not be the best solution, see Price, 1995, ch. 3 § 3.) However, he has good reason to attach some desires, and also pleasures, closely to the operations of theoretical reason: for, as Brentano urged, that has its own goals and pleasures. (See Price, 1995, p. 192 n. 11.) And his claims in DA III 10 that practical intellect is a special kind of intellect that causes movement, but only in association with desire, make best sense if practical reason and rational desire are tightly connected within the soul’s rational faculty. In AE B 2, Aristotle ascribes a distinctive goal to practical thinking: What affirmation and negation are in thinking, pursuit and avoidance are in desire; so that since moral virtue is a state of character concerned with choice, and choice is deliberate desire, therefore both the reasoning must be true and the desire right, if the choice is to be good, and the latter must pursue just the things that the former affirms. Now this kind of intellect and truth is practical; of the intellect which is contemplative, not practical nor productive, the good and the bad state are truth and falsity respectively (for this is the work of everything intellectual); while of the part which is practical and intellectual the good state is truth in agreement with right desire (1139a21–31, after Ross, 1925).

390

Aristotle’s Conception of Practical Thinking

This passage would appear to lend itself to two readings, both of which bring out the distinctive role of practical intellect. It is perhaps most natural to interpret an instance of ‘pursuit’ as a case of desiring to w, and an instance of ‘avoidance’ as a case of desiring not to w. The reasoning (logos) is ‘true’ if it is sound, moving from true premises to a true practical conclusion by a route that counts as valid. (As Kenny observes of this passage, ‘The true logos is not the phasis [affirmation] itself, but the argument leading to it.’; 1979, p. 93 n. 1.) That choice ‘must pursue’ what reasoning ‘affirms’ (a25–6) may seem an odd condition of the choice’s being good, since there is no choice unless reasoning and desire coincide in their object. However, we should stress the ‘asserts’: what is being emphasized is the need not just for rationality, but for decisiveness; and the ‘so that’ (hôste, a22) is only satisfied by a conclusion which, in requiring that reason affirms precisely that which desire pursues, confirms the relation between affirmation and pursuit, negation and avoidance (a21–2), by revealing not just an analogy, but an interlinking of reason and desire. It is no doubt consciously that Aristotle requires that reason should assert the things (plural) that desire pursues: desire initiates the reasoning by proposing a goal of boulêsis as a starting-point, and it concludes it by generating an object of choice. As the De Anima has it, the former is ‘the starting-point of practical reason’, whereas the latter is ‘the starting-point of action’ (III 10, 433a16–17). However, there is an alternative, which we owe to Elizabeth Anscombe (1981, 2005). On her reading of AE B 2, ‘Aristotle is not comparing attraction and contrary affects of the psychic faculty of desire to affirmation and negation: no, he compares pursuit and flight, which are possible actions, to positive predication and negation’ (2005, p. 151). We have then, I take it, to reinterpret this sentence: ‘Both the reasoning must be true and the desire right, if the choice is to be good, and the latter must pursue just the things that the former affirms’ (1139a23–6). For the last clause cannot then be a condition of the choice’s being good: rather, the reasoning’s being true and the desire right makes the choice right, and this in turn is a precondition of desire’s pursuing, in action, what the reasoning affirms. And Aristotle’s gloss upon practical truth as ‘truth in agreement with right desire’ turns out to mean ‘things (i.e. whatever is in question to bring about in action) being as rightly desired’ (2005, p. 153). Practical truth is ‘brought about ... by action (since the description of what [the agent] does is made true by his doing it), provided that [he] forms and executes a good “choice” ’ (1981, p. 77). If his choice is ethically flawed, so that he acts unjustly while purporting to act justly, ‘his description “justice performed” of what he has done will be a lie’, so that he ‘will have produced practical falsehood’ (ibid.). This reading well prepares the later statement, ‘Acting well is the end, and desire is for this’ (1139b3–4). Practical thinking remains inherently incomplete until it achieves its natural end, which is action fully in accordance with right reasoning and choice. We capture much of this if we recall the charac-

A. W. Price

391

terization of choice that is explicit in the Eudemian Ethics: it is at once ‘of something’ and ‘for the sake of something’ (II 11, 1227b37; cf. II 10, 1226a11–13). Hence the agent’s reasoning must proceed from what is, in context, a good goal to a good means. However, that permits the following mishap: ‘It is possible to attain even good by a false syllogism, and to attain what one ought to do but not by the right route, the middle term being false’ (AE B 9, 1142b22–4). The full achievement of practical truth involves acting in a way that makes true, without any admixture of falsity, a sequence of act-descriptions running from the most general ‘acting well’ to the most specific. Whether or not this is the right interpretation of ‘practical truth’ in AE II 2, it coheres well with Aristotle’s tendency to identify the conclusion of practical thinking that actually leads to action with nothing less than the action itself. Consider a passage in the De Motu Animalium (7, 701a10–16): There [i.e. ‘in thinking and inferring about immovable things’, a9] the end is a piece of theoretical knowledge (for, whenever one thinks the two premises, one thinks and puts together the conclusion), but here the two propositions result in a conclusion which is the action. E.g., whenever one thinks that every man is to walk and one is a man, immediately one walks; or if one thinks that, in this case, no man is to walk and one is a man, immediately one remains at rest. And one so acts in the two cases, if nothing holds one back or compels one. (Also relevant is AE B 3, 1147a25–31, on which see Price, 2006, pp. 239–40, p. 250 nn. 16–17.) Aristotle may then appear to take this back (a17–23): I need a covering; a cloak is a covering. I need a cloak. What I need, I am to make; I need a cloak. I am to make a cloak. And the conclusion, the ‘I am to make a cloak’, is an action. And he acts from a starting-point. If there is to be a cloak, it is necessary that there be this first, and if this, this. And this he does at once. Now, that the action is the conclusion is clear. Even as Aristotle identifies the conclusion with an action, he states it propositionally (‘I am to make a cloak’), and describes how it sets further thought in train. Yet we read before the final piece of thinking, ‘He acts from a starting-point’; and Aristotle concludes that it is clear that the action is the conclusion. What is going on? It would seem that, when the execution of a choice is not wholly mechanical, what Aristotle counts as ‘action’ includes thought. (See Nussbaum, 1978, pp. 194–5, 344.) I don’t, for instance, cease thinking when I start to write a letter, or indeed (I hope) a paper: writing a letter or paper requires a continuing exercise of intelligence. As evidence of a final stretch of purely executive thinking, simultaneous with what counts as acting, Anscombe (1981, p. 67)

392 Aristotle’s Conception of Practical Thinking

cites this: ‘If a thing appears possible, we try to do it. By “possible’ things I mean things that might be brought about by our own efforts’ (EN III 3, 1112b26–7). Trying to do something amounts, if all goes well, to actually doing it; and yet the term ‘try’ conveys that thought and care are still involved. When mere habit or dexterity cannot fully take over, the upshot of deliberation is an agent who thinks as he acts, and acts as he thinks. What counts as the beginning of action need not ensue upon the end of thought, but may coincide with the point at which there is no uncertainty about whether now to pursue the goal. We can count the thinker as an agent before he has stopped thinking, and as soon as he is embarked upon action, confident of his ability to proceed and the acceptability of his proceeding (which cannot, of course, exclude the possibility of accident or interruption). His action will then itself constitute thought in action. That the proper conclusion of a piece of practical thinking cannot be less than action is intelligible, and even inevitable, in a way: as practical thinking is inherently for the sake of action, it does not even achieve its goal as thinking until it is enacted. Yet how can an act stand to practical thinking not just as its closure, but as its conclusion? Consider again, ‘The conclusion, the “I am to make a cloak”, is an action’ (DMA 7, 701a19–20). Entailment is clearly a relation that holds between propositions: what (= all that) I need, I am to make; I need a cloak; so, I am to make a cloak. The third of these propositions could only, itself, be an action in the sense of specifying one. However, it is not likely that all that Aristotle has in mind by ‘conclusion’ (sumperasma) is an entailed proposition. He fails to abstract propositions from statements when he presumes, near the start of the Prior Analytics, ‘A proposition, then, is a logos affirming or denying something of something’ (I 1, 24a16–17). He pronounces accordingly, ‘A syllogism is a logos in which, certain things being laid down, something other than what is stated follows of necessity from their being so’ (b18–20). So, when he calls the conclusion an action, it is more likely that he means that drawing the conclusion is enacting it. Even so, he adds a qualification: ‘[The agent] does both of these things [i.e., walks, or remains at rest], if nothing prevents or compels him’ (a15–16). While his meaning is debatable, he must recognize the possibility of an agent who sets himself to walk, and finds that he is paralysed (cf. NE I 13, 1102b18–20). Even where enactment is unimpeded, we might add, applying a characteristically Aristotelian form of words, that, though the conclusion and the action are the same, their being – or what it is for each of them to be what it is – is not the same (e.g. DA III 2, 426a15–17). This is surely what he would say of practical judgement and choice, which cannot diverge, but differ in their logical grammar (EN III 2, 1112a3–5). Aristotle’s line of thought may be the following. Even when p entails q, ‘A thinks that p’ does not straightforwardly entail ‘A thinks that q.’ And yet, if the inference is relevant, and not obscure, then, in the absence of a special explanation, we shall expect A to think that q if he thinks that p. Analogously,

A. W. Price

393

within practical thinking, if A accepts that he has to make all that he needs (no doubt within a contextually delimited range), and that (within that range) he needs a cloak, then, again in the absence of a special explanation, we shall expect – given background conditions of ability and opportunity and awareness – to find A making a cloak. (So, post Wittgenstein via Malcolm, Nussbaum, 1978, pp. 175, 179.)

III I conclude by a summary statement of how Aristotle understands the relation of thought to action, and a brief expression of one reservation that we may have. (For a fuller treatment, see Price, forthcoming a & b.) A mature agent brings to a situation a pattern of habituated responses, affective and cognitive, that constitute his character. They shed a distinctive light upon the situation that makes salient for him features that invite the pursuit of some fairly specific end. (He does not select this end by reasoning, EE II 1, 1227b22–5. Reasoning follows.) He now deliberates about whether and how to achieve this end, in his situation, in a way that does justice to his other ends and concerns. (It is this wider perspective that makes the deliberating an exercise of nothing less than phronêsis; AE B 5, 1140a25–8.) The kernel of his deliberation can be set out in a piece of reasoning from an end (such as retaining his health) to a way or means (such as taking a walk), which prefigures the action that may ensue (such as taking a walk in order to retain one’s health). If nothing gets in the way, either by telling against so acting for other reasons, or by preventing him from so acting, he may be expected to act accordingly. Calculating means to an end can be idle rather than practical (as it is when Aristotle’s offers simplified examples). It becomes practical when it is pursued for the sake of an end, and so can serve intelligibly to yield a choice out of a wish. (As we would put it, it serves the generation of a sequence of intentions.) And it becomes an exercise of phronêsis when the agent is open to reflection about how and whether he can achieve his provisional end acceptably, that is, in a way that constitutes acting well. If there is room for disappointment, it is that Aristotle fails to do justice to a feature of practical thinking that is sometimes evidenced in his texts. If the inherent purpose of practical thinking is to achieve an end acceptably, it is no objection to an inference to one choice that another choice would have served as well. (Thus Anscombe has remarked, ‘What is important is surely that the end will be attained by the means arrived at, not whether it is the only means’; 1995, p. 9.) Occasional examples bear this out (e.g. DMA 7, 701a16–17), but Aristotle typically sets out reasoning to some necessary way or means. He appears to have thought that action is only possible if, within any set of alternative ways and means, one stands out as best (cf. EN III 3, 1112b16–17, De Caelo II 13, 295b30–4). Which is not so. Yet

394 Aristotle’s Conception of Practical Thinking

there would be more reason to regret what he got wrong if philosophers had not often lost sight of what he got right.

References Ancient Aristotle, De Anima [DA] __, De Caelo __, De Motu Animalium [DMA] __, Eudemian Ethics [EE] __, Metaphysics __, Nicomachean Ethics [EN] __, Posterior Analytics __, Prior Analytics __, Topics

Modern Anscombe, G. E. M. (1981), ‘Thought and Action in Aristotle: What is “Practical Truth”?’, in Collected Philosophical Papers, Volume One: From Parmenides to Wittgenstein (Oxford: Basil Blackwell). Anscombe, G. E. M. (1995), ‘Practical Inference’, in R. Hursthouse, G. Lawrence and W. Quinn (eds), Virtues and Reasons: Philippa Foot and Moral Theory (Oxford: Clarendon Press). Anscombe, G. E. M. (2005), ‘Practical Truth’, in M. Geach and L. Gormally (eds), Human Life, Action and Ethics: Essays by G. E. M. Anscombe (Exeter: Imprint Academic). Broadie, Sarah (1991), Ethics in Aristotle (New York: Oxford University Press). Foot, Philippa (2001), Natural Goodness (Oxford: Clarendon Press). Hamlyn, D. W. (1993), tr. and com., Aristotle: De Anima Books II and III, new edn. (Oxford: Clarendon Press). Kenny, Anthony (1979), Aristotle’s Theory of the Will (London: Duckworth). Loening, Richard (1967 reprint from 1903), Die Zurechnungslehre von Aristoteles (Hildesheim: Georg Olms). Müller, Anselm Winfried (1979), ‘How Theoretical is Practical Thinking?’, in C. Diamond and J. Teichman (eds), Intention and Intentionality: Essays in Honour of G. E. M. Anscombe (Brighton: Harvester Press). Müller, Anselm Winfried (1982), Praktisches Folgern und Selbstgestaltung nach Aristoteles (Freibürg/München: Karl Alber). Müller, Anselm Winfried (1992), ‘Mental Teleology’, Proceedings of the Aristotelian Society 92, 161–83. Nussbaum, Martha Craven (1978), ed., tr. and com., Aristotle’s De Motu Animalium (Princeton, NJ: Princeton University Press). Price, A. W. (1995), Mental Conflict (London: Routledge). Price, A. W. (2006), ‘Acrasia and Self-Control’, in R. Kraut (ed.), The Blackwell Guide to Aristotle’s Nicomachean Ethics (Oxford: Blackwell). Price, A. W. (forthcoming a), ‘The Practical Syllogism in Aristotle: A New Interpretation’, Philosophiegeschichte und logische Analyse/Logical Analysis and History of Philosophy, issue on ‘Practical Syllogism and Practical Deliberation’.

A. W. Price

395

Price, A. W. (forthcoming b), ‘Aristotle on the Ends of Deliberation’, in M. Pakaluk and G. Pearson (eds), Collection Of Essays On Aristotle’s Moral Psychology (Oxford: Clarendon Press). Ross, W. D. (1925), tr., Ethica Nicomachea (Oxford: Clarendon Press). Woods, Michael (1986), ‘Intuition and Perception in Aristotle’s Ethics’, Oxford Studies in Ancient Philosophy 4, 145–66. Woods, Michael (1992), tr. and com., Aristotle: Eudemian Ethics, Book I, II and VIII, 2nd edn (Oxford: Clarendon Press).

21 Action in Moral Metaphysics Jonathan Dancy

In this essay I will be trying to show that the linguistic philosophy of the 1960s still has something to teach us, despite changes in the concerns of moral philosophers. One can hear my title in two ways. In one way, it alludes to the great upsurge in the attention paid to metaphysical considerations in recent moral theory. In this sense, there has been a lot of action, or at least activity, in moral metaphysics. But what I really mean to allude to is the place of action in the metaphysical scheme of things, and here I want to suggest that we don’t need a metaphysics of action at all. Actions, I am going to suggest, should not be considered as independent elements in our metaphysics. There should be less of action in our moral metaphysics, not more.

I There are all sorts of ways in which we seem to contrast action and agent in moral and other practical evaluation. We say, for instance, that someone is not to be blamed, for though he did the wrong thing he did it for the right reasons, or from a good motive, or with good intentions. We say that what he did was right, but that he gets no credit for it because he did it for the wrong or for bad reasons, in the wrong way, from a bad motive or with bad intentions. We can say that he was right to do what he did, but that nonetheless he did the wrong thing. We can say that he acted wrongly, though what he did was itself right. What we seem to be using in these ways of speaking is a distinction between two independent objects of moral evaluation – or perhaps, so as not to exaggerate, I should say semiindependent, since the question how independent they are of each other has not yet been examined. And this is what I would like to understand better. In particular, I would like to understand better the notion of what is done, the it in done it. If I understood this better, I think I would have a better perspective from which to understand these contrasts between action and agent in moral evaluation.

396

Jonathan Dancy

397

II I start, however, with a different set of contrasts. A considerable amount of work was done by British moral philosophers in the first half of this century, especially in the 1930s, on the relations between action, intention, motive and consequences. The question they were asking was how to map these concepts relative to each other. The action stood somehow between the motive and intention, on the one side, and the consequences on the other. How was this to be understood? Was the motive distinct from the intention? Was the intention itself part of the action, in some sense? Or was it, like the motive, a non-essential cause (non-essential in the sense that the same action could have been done with different intentions, just as it could supposedly have been done from different motives). The question that has occupied theorists of action more recently, where the action stops and the consequences take over, was not so much discussed. To give you something of the flavour of the debate: Sir David Ross held that the motive was distinct from the action, since the action was chosen, and choice is not of a motive but for or from a motive. So for him the story went motive – choice – action.1 (I think that he is wrong about this. I think that it is possible to choose an action because it expresses a certain motive. If so, the motive chosen need not be the motive for choice.) H. W. B. Joseph had argued that if one distinguished both the motive and the consequences from the action, the action could be nothing more than a bodily movement, which in his view was a reductio ad absurdum, since a bodily movement was incapable of bearing moral properties, and so could not be right or wrong. He concluded that motive must be part of action, not a mere cause.2 The debate, then, was about how to characterize the relations between motive, intention, action and consequences (and we could add to this reason, aim, purpose and end) in such a way as to create a conceptual map of the situation. In this debate, the action was conceived as what was left once one had subtracted from the total situation everything that was not a proper part of the action – and what was left would probably be a bodily movement. If the intention was a proper part of the action, say, then it would not be subtracted, and the action as a whole could be a combination of intention and bodily movement. The methodology that underlay the debate, if one can call it that, was that we can contrast an action with the motive from which it is done, with the reasons for which it is done, with the intention with which it is done, with the consequences that flow from it, and even I suppose with the movements that are made in doing it. To locate the action itself, as distinct from its accompaniments, we have to determine the space that each occupies on the map. If we have determined the space occupied by all but one element, then we have finished our task. For the remaining element must occupy the remaining space.

398 Action in Moral Metaphysics

There is a pleasing inevitability about this approach, but one cannot fail to wonder if things are as simple as that. How can we be so sure that the amalgamation of contrasts will succeed in isolating each element in this kind of way? The isolation method depends upon the assumption that if we take all the relevant contrasts together, they will combine to yield an intelligible result. In fact, those who use this method are most likely to conclude that an action is a bodily movement, as I said; though for myself I think that in reaching this result they have forgotten that we can contrast the action with the movements made in doing it. If we add that contrast to the others, the action is likely to vanish entirely from view. Ross, under pressure from his attempt to distinguish the action from everything else, is driven to take seriously the suggestion that an act is a ‘setting oneself to bring about a change’.3 Perhaps correct application of the method should drive us to some such conclusion as this; but I think that this just shows that the isolation method is not always appropriate. Whether it is appropriate or not will depend on whether the area we are dealing with does in fact consist of isolable elements in the way necessary for the method to yield reliable results. And that will itself depend on whether the contrasts that make up our subject matter are in fact coincident, that is, whether they can be got to add up, as it were, in the way we want. If they cannot be got to add up in that way, the apparent inevitability of our method will be shown to be an illusion, since we are trying to amalgamate what one might call non-coincident contrasts, and it will not be surprising if the result is incoherent. It would be as well, in offering this tentative diagnosis of what was going on at that time, if one could provide a persuasive example of a set of non-coincident contrasts. The one that I want to suggest is that of a fact. We contrast fact with a wide range of other things: law, value, theory, hypotheses, prediction, belief, opinion, fiction, falsehood and so on. Suppose now that we are interested in what a fact is in itself, as it were. A fact must be something that is not theory, is not value, is not law and so on. Is the isolation method appropriate in this case? I think not. The reason why not is that what is fact in relation to value may not be fact in relation to law. Equally, what is fact in relation to theory may be value in relation to fact. The most striking thing is that theory is (probably) fact rather than law, while law is fact rather than theory, though both theory and law are themselves contrasted with fact. Further, some theories are facts and other theories are opinions (perhaps). So the idea that we can find out what a fact is by adding up all the contrasts in which the notion of a fact figures is in my view not a runner. And if we did add up all those contrasts at once we would be in danger of finding that there was nothing left for a fact to be. What I am suggesting here is that much the same thing is true of ‘fact’ as Austin thought true of ‘real’, introducing the unfortunate term ‘trouser word’. The notion of a fact is always used with some contrast in mind, but which contrast is in mind determines the sense of the word on the particular

Jonathan Dancy

399

occasion of use. If this is so, there is no hope of determining the sense of the word, a sense it has on every occasion of use, by adding up all the possible contrasts at once and expecting to find a suitable residue to stand as the meaning. ‘Fact’ in fact versus law is just not the same as ‘fact’ in fact versus value. Let us now return to action. Obviously I am going to try to say similar things about action. My view is that there is indeed a sort of ‘actionrelevant’ course of events or set of considerations, in terms of which action is contrasted with what is not action, but that since every element in that set of considerations is able to be contrasted with the action, it is not surprising that if we add up all the contrasts at once we find that nothing is left – or not enough, anyway. Perhaps the most telling point in support of this view is that we are going to have to add the contrasts between action and agent and action and inaction to those between action and motive, action and intention, action and movement, and action and consequences, in trying to isolate the action from other elements of the overall situation. If we think that in order to find room for actions on the metaphysical map, we should look for the space left when we have subtracted all those other things, it is surely very plausible to suggest that no space will be left, and so that there are no actions. Since there must be actions in some sense,4 the method that leads us to the conclusion that there are none must be mistaken, if not in general then at least in the present case. So the first thought is simply that if we apply to action the method of amalgamating contrasts, we will end up with the view that there are no actions at all. It would be pleasing if we could also show the kind of result I think one gets with facts, namely that what is on the fact side of one contrast is on the non-fact side of another, as value, though itself contrasted with fact, is fact rather than law or theory. And there is at least some chance that we can do this. For instance, if we contrast intention with action (he meant to do this but only succeeded in doing that) the action is conceived as separate from the intention, since what was meant is not what was done. To make that contrast, however, we have to put some aspects of outcomes, consequences, on the action side of things. He meant to upset her, but failed entirely in that, and only succeeded in making a fool of himself. Suppose now that I contrast action with consequences. To do that, I am likely to find myself introducing thoughts of intention on the action side. The action, that is, will be conceived as purposeful, in order to contrast the intended outcome with the actual outcome. So just as ‘fact’ in fact versus value is not the same as ‘fact’ in fact versus law, so ‘action’ in action versus consequences is not the same as ‘action’ in action versus intention. These sorts of considerations seem to show that the notion of an action is involved in various contrasts, and to cast doubt on the idea that the contrasts can eventually be made to coincide. They also cast doubt on a different sort

400

Action in Moral Metaphysics

of isolation test for actions. This test starts from the thoughts that there can be action without bodily movement, without motive, without intention, without consequences and without reason.5 If we add all these thoughts together, we might conclude that action is something distinct from all of these, since it can occur in the absence of each of them. Thinking in this way, we may end up with Ross’s notion of ‘setting oneself to initiate a change’ as all that is left for an action to be. But again, I suggest that this form of test is unsound. To sum up the present section: I have been suggesting that if we amalgamate all the possible contrasts in which the notion of an action figures, actions will vanish from view entirely.6 I take this to show that there is something wrong with a certain metaphysical approach to the notion of an action, an approach which if it were sound would serve to show the applicability of the isolation method.

III I want now to introduce a distinction that I have taken from the work of Jennifer Hornsby, between the action and the thing done. According to Hornsby, ‘[a]ctions are particulars – unrepeatable things, named by phrases like ‘Hyam’s setting light to the petrol at two o’clock on the fateful day’ and ‘my reading this paper now’. Something done, on the other hand, is not a particular: things done are named by phrases like ‘inflict damage’, or ‘eat an egg’, or ‘throw a brick’.7 Actions are ‘variably describable’; this means that there are different things that can be said about an action, and all are equally descriptions of it, for there is no ‘single truth’ about what an action is. What is done, however, is not a particular, and there is only one thing that can be said about a thing done. In that sense, its essence exhausts its nature. No doubt this is one perfectly good way to distinguish an action from a ‘what is done’. But there are others. For instance, one could distinguish an action as an acting, a doing, from an action as a thing done. If there are others, which one we adopt is probably a matter of our philosophical purposes. So what I want to say about Hornsby’s way of carving things up is that it does not suit the purposes that interest me. This is not merely selfabsorption. The purposes I am talking about are the purposes I mentioned at the beginning, namely the attempt to understand the complexities, subtleties and internal tensions of our evaluative practice. Perhaps I should start by confessing that anyway I don’t see the pressure to announce that things done are not particulars. Later in the same paper Hornsby speaks of ‘a distinction between actions (particulars) on the one hand, and things that may be done (not particulars) on the other’.8 But this seems to me a quite different point. It may be that ‘the children I may have’ are not particulars, any more than is the possible fat man in the doorway, in Quine’s famous example; but the children I do have are particular enough. Perhaps

Jonathan Dancy

401

it is like that with things done, that possible things done are not particulars, but actual things done are particulars (variously describable etc.). Second, and more relevant to my present concerns, to think of things done as universals is of no help if we are trying to understand the sense in which what is done can be wrong. If I say, in a given case, that what you did was wrong, am I predicating wrongness of a universal? I think that such predication is unintelligible. It might be that all instances of that universal are wrong, and wrong as such; but this would not make the universal itself wrong. A further question which I do not know how to answer is whether an action is to be thought of as a particular instance of a thing done – in fact, as an instance of many things done. If so, all that Hornsby’s distinction amounts to is a distinction between action-universals and particular actions. And though there is indeed that distinction, it is odd to say that the relation between a doing and a thing done is the relation between particular and universal – the same relation as that between an object and a property. For these reasons, then, Hornsby’s way of carving things up does not address the sorts of problems I was engaged with earlier. In fact, if we are to have any such distinction at all, I would prefer to Hornsby’s two-way distinction the three-way distinction I mentioned earlier between the agent, the doing, and the thing done – a distinction that is certainly not new. Of these three both the doing and the thing done can be called actions, and both in this version are particulars. We might try to show the need for both these ‘moments’ by arguing that, for instance, the doing can be cautious when the thing done is not. The manner of the doing is not itself an aspect of what is done, even if what is done is an action, and is done carefully (say). What is done may be unintended, when the doing of it was not. The doing may be exhausting when what is done is not. And so on. This way of carving things up at least makes it possible for us to think of attributing moral properties to all three moments – and different ones to each. The agent can be wrong to do something, the doing of it may be wrong, and the thing done may be wrong. Maybe the agent was not wrong to do it, but what he did was wrong. And so on. We are also enabled to say, as we do want to say, that in doing this I did many things. In such a case we have one doing and many things done. Hornsby’s distinction would only give us one action instantiating many universals, which is not at all the same thing. My things done are particulars, and there are many of them, though only one doing which is the doing of all at once. All I did was to press the button, and I fired the engines, launched the rocket, destroyed the city, justified my expensive training, and gave a good example to my men. So I did something that was a firing, something that was a launching, something that was a destroying and so on. It is left open for us either to assert or deny that these ‘somethings’ were all identical with each other. In the desire to say that they are non-identical, we might suggest that the pressure to think of them as identical is absorbed, as it were, by the uniqueness of the

402

Action in Moral Metaphysics

doing. This leaves the things done able to be distinct both from the doing and from each other. When I say that I prefer this way of carving things up, I do not mean by that to say that I agree with it. My view is that the distinctions it draws are grammatical distinctions rather than real ones. The difference between the acting and the thing done lies in their relation to the verb ‘do’. The first is a gerundive, and the second is an internal accusative. This will not help anyone unversed in Latin grammar; so I will have to try to say what I take these terms to mean. A gerundive is a verbal noun, or a verb converted into a noun. Most languages that I know have a way of converting verbs into nouns; in English, one adds the suffix ‘-ing’ to the verb, for example, ‘building’ from ‘build’. There is a complication, which is that in English the present participle, which functions as an adjective, is also formed from the verb in the same way. We can make the verb-noun ‘running’ from ‘run’ – his running was fast and elegant, while mine was slow and clumsy – as opposed to the verb-adjective ‘running’ in ‘running home from work, he slipped’. Equally, in ‘he is building’, ‘building’ is a participle, an adjective; in ‘building on this site would be irresponsible’, ‘building’ is a noun. (An alternative construction uses the infinitive: ‘to build on this site would be irresponsible’; and this is the way in which some other languages build verb-nouns.) An internal accusative is rather harder to explain. The idea is that with some verbs we can create a noun to attach adjectives to. A good example is the verb ‘to strike’; the internal accusative for this verb is ‘a blow’, as in ‘he struck a well-aimed blow’. This is to be contrasted with an external accusative, as in ‘he struck Roger’. One can even have external and internal accusatives in the same sentence, as in ‘he struck Roger a well-aimed blow’. The ‘blow’ here, the internal accusative, is equivalent to a ‘striking’; he struck a well-aimed striking. In the case of ‘act’ or ‘do’, we could say (in a similar intentionally awkward way) ‘he did a doing’ or ‘he acted an action’. But in fact we say ‘he did an action’; and we then characterize the behaviour of the agent using adjectives attached to this noun rather than in other ways (of which the most obvious is the adverb – indeed my overall approach to actions might be called the adverbial approach). So instead of saying ‘he acted cautiously’ we can say ‘he did a cautious thing/action’; instead of ‘he acted wrongly’ we can say ‘he did something wrong’. There need be no suggestion that we take this manoeuvre seriously, that is, that there really are such things as things done or actions in any sense beyond what one might call the grammatical. So in my view both the acting and the thing done are more like grammatical constructs than items for which we need to find an independent place in our metaphysics. How does the tripartite distinction between agent, acting and thing done relate to my earlier discussion of the attempt to place action in relation to motive, intention and consequences? Is ‘action’ in that debate to be thought of as action in the sense of acting, or in the sense of thing done? It seems to

Jonathan Dancy

403

me that it must be the first of these. Where we would place the thing done in the story is a moot point, which we need not, for present purposes, address.

IV I want now, allowing for the moment that there are such things as particular actions in the sense of things done, to ask about their modal qualities. Concentrating on the thing done itself, whatever that is, let us ask how many of its features are contingent and how many are essential. How many features of this thing done, that is, are such that an action might lack them and be this action? 1. 2. 3. 4. 5. 6.

Could it have been done by someone else? Could it have been done for another reason? Could it have been done with a different intention? Could it have been done from a different motive? Could it have been done at a different time or place? Could it have been done in a different way?

In answering these questions, we have to distinguish epistemological possibility from metaphysical possibility. Epistemologically speaking, all these matters are open; but metaphysically they seem all to be closed. Metaphysically speaking, that is, the answers to all these questions, if we are concentrating furiously on the thing done, seems to be no.9 Could someone else have scored just that goal? It wouldn’t have been that goal, but another one, scored in the same way at the same time. The driving force here is the same as that which makes us say, in Kripke’s famous example, that this desk could not have been made of ice; one might have got a desk that is similar in various ways, but not this one. One just can’t tell a story in which this action gets to be done in a different way, by a different person and so on. The position is like this, then. Actions have to be placed with respect to motive, intention, reason, location, agent and so on. We have modal intuitions about the extent to which this action might have been different. What explanation can we offer for these intuitions? One explanation is the metaphysical explanation. Actions include intentions, or reasons, or agents, or whatever. Such thoughts are thoughts about the metaphysical extent of an action. But if my suspicions about the amalgamation of non-coincident contrasts are well founded, it may be that actions have no metaphysical extent. In that case we need an alternative explanation of our modal intuitions. A different explanation might appeal to the ways in which we individuate actions. An action is individuated by its agent, by its time and place, and by other features of its location and the reasons for which it is done. We might suggest that since we individuate an action by appeal to these features, it is

404 Action in Moral Metaphysics

not surprising that if those features are changed, we find ourselves individuating another action. In response to this, we could reply that the same is true of physical things, but that they have contingently the features we use to individuate them. Individuating features are different from identity conditions, and as such may be either contingent or necessary. However, this might alert us to something odd about actions: their identity conditions seem to be identical (more or less) with their individuating properties. They have, as it were, too many identity conditions – too many essential features. How can it be that so many of their features are essential to them? The answer that suggests itself is that, even if actions are particulars, they are at best derived particulars – particulars which consist in features of other particulars, some of which are derived and some of which are underived. An action, that is, consists in a combination of, or relation between, features of: 1. 2. 3. 4.

the agent the surroundings other actions other events

Of these four, the first two are underived particulars, the second two derived. Agents and situations have non-essential features, and to the extent that actions have non-essential features, they get them from their agents or from the surroundings in which they are done. For instance, though this action could not have been done by anyone else, it could have been done by a bald man if its agent could have been a bald man. Equally, it could have been done in bright sunlight if the weather could have been sunny at the time and place where it was done. So, though actions do have contingent features, they get them from the contingent features of agents or of their surroundings. They seem to have a shortage of contingent features all their own, as it were. And this is to be explained, on the present hypothesis, by awarding them a secondary metaphysical status. Since they consist in a combination of features of other objects, those features are essential to them; it would not be this combination if we had different features combined. We now have two competing explanations of the modal nature of actions. Both are metaphysical. The first spoke of the extent of action: it includes various features, in ways that mean that one cannot have the same action with different parts. The second understands an action as a relation between various features of the overall situation, so that we cannot have the same action unless we have the same features related.10 My objection to both these explanations is the same. They turn what should be uncontentious truths into falsehoods. What we want to do is to retain the ability of statements like ‘I might have done it differently’, ‘anyone could have done it’, ‘we could do it tomorrow rather than today’, ‘another

Jonathan Dancy

405

person would have done the same thing for different reasons’, and so on, to express truths. We have to avoid awarding actions a place on the metaphysical map so distinctive and separate that the statements above are bound to express falsehoods rather than truths. But the two explanations above fail to do this. On their accounts of what an action is, the statements above are false rather than true. What we need to do, then, is somehow to legitimise the ways in which we talk of ‘the action’, ‘the thing done’ – that is, the ‘it’ in ‘done it’ –without getting embroiled in that consequence. It seems to me that we can do this by understanding expressions such as ‘the thing he did’ as internal accusatives. If we turn features of my acting into features of a thing done, and take what we have done seriously, we make all statements like ‘I might have done it differently’ false rather than true. But if we think of ‘the thing done’ as an internal accusative, we gain the advantages of this grammatical device without getting embroiled in questions of identity. In a way, there shouldn’t be a question whether it would have been the same action or not. Actions should not have identity criteria. To suppose otherwise is to misread a grammatical device as a piece of substantial metaphysics. This does not mean that there is something suspect about our action-talk. There are actions; they are what agents do, and they have properties. All this is perfectly in order.

V The previous section concerned things done, conceived of as particulars rather than as universals. But we should ask whether we can say the same about actings as about things done. It does seem that the modal questions yield equally uncompromising negative answers in the case of actings. Actings could not have been anyone else’s; they could not have been different in various ways, and so on. In the case of things done, I maintained that these modal answers raised problems, for they turned truths into falsehoods. Is the same true of actings? It seems to me that the problem here is less acute. We certainly do not want to say that this acting might have been someone else’s – that someone else might have done this doing. (Though Wittgenstein would want to persuade us that this should cast doubt on the idea that it was necessarily mine.) Could this acting have taken place at a different time? If not, have we turned some everyday truth into a philosophical falsehood? The difficulty we have in answering this question derives from the fact that though the thing done is a familiar item in everyday thought, the acting is not.11 Still, could the acting have been different in manner? Could it have been gentler? Could its agent have engaged in it for a different reason, or with different intentions? If the answer to these questions is no, does this create problems for actings in the way that it did for things done? Though it is much harder to be sure, I think that it does.

406 Action in Moral Metaphysics

Given this, I should be trying to argue that actings are as much grammatical fictions as are things done, and that accordingly they should not be counted, and do not have identity criteria. And this may seem ridiculous. Surely we are all familiar with detailed and lengthy discussion in the philosophy of action about the question whether the shooting was identical with the killing. That these questions are hard to answer does not show that there is something philosophically suspect about them. Further, the courts need and use questions which hinge on thoughts about the identity and non-identity of actions. The problem of double jeopardy is the problem whether it is reasonable for a person to be punished twice for the same action. And here by ‘the same action’ we mean ‘the same acting’. What is more, we naturally speak of whether there were two shootings rather than one. We don’t speak quite so naturally about whether a shooting was a killing. Are these perhaps the same doing, even though there are clearly two things done? This seems to be a question reserved for philosophers. As such, it is suspect (at least in my eyes). The situation seems to me to be like this. The verb ‘acting’ is a dummy action verb. As such, it is not what one might call a count-verb (by analogy with a count-noun). There are such things as count-verbs, however. Any verb V of which one can ask how often the agent V-ed is a count-verb. If the agent V-ed three times, he did three V-ings. Some verbs are not like this. It is not, for instance, clear whether the verb ‘think’ is a count-verb. But most, perhaps, are. If a verb is a countverb, there must be criteria of identity, or at least of enumeration. There must be some criteria for when one has several doings, and when one only has one multiple one. When I say that my son has only driven my new car once, we know what I mean, at least within limits, even if it was a multiple driving (he went from Oxford to Edinburgh). If the only occasion on which he drove it (his only driving of it) was a solo journey from Oxford to Greece, things have got out of hand. By this I mean, not that I should not have lent him my new car for a solo journey to Greece, but that the criteria of identity for drivings do not really apply to this case. So some verbs are count-verbs and some are not. Given a particular verb V, then, we can normally ask how often the agent V-ed. But this does nothing to substantiate a sense in which we should expect to be able to answer questions about how many actings the agent did in V-ing. The verbs ‘act’ and ‘do’ are dummy verbs, which do not carry criteria of identity with them. There is no generic count-verb that is capable of making it a real question whether a shooting is identical to a killing or not. If there were, we could ask whether in doing what he did the agent acted often or only once. There is a question whether there were two shootings rather than one, but no question whether a shooting was the same action/acting as a killing.12 We can, then, say that in pressing the button I fired the rocket and destroyed the city. But this should not lead us to take seriously questions about whether the pressing was identical with the firing, conceived as

Jonathan Dancy

407

questions about the identity of actings/doings rather than of things done. Perhaps the truth is only this, that I pressed, and I fired, and that I fired by pressing. I did two things, of course, since to fire is not to press. But I did not act twice. What about the principle of double jeopardy? I am afraid that my conclusion is not very hospitable to this principle. I find it hard to take seriously the question whether the criminal, in lying and cheating, did two doings/ actings rather than one. It certainly seems hard to suppose that substantial conclusions about praise and blame, culpability and punishment should hang on what we decide to say in such a case.

VI There remains one tricky matter that needs to be addressed, even if only briefly. There is a question whether my deflationary attitude to actions extends, or should extend, to events. And there is a more general question about how I should understand the relation between actions and events. One thing is for sure, that events are not internal accusatives. To have an internal accusative, we need to start from a verb of a certain sort; and no such verb in available in the case of events. ‘Take place’, ‘happen’, ‘come to pass’; these are verbs of quite the wrong sort. Events seem to me much more like gerundives; an ‘event’ should be understood as a verbal noun, equivalent to a ‘happening’. There are no things that happen – or rather, though there is no harm in talking this way, again we would not take it too seriously. If this is the place that events occupy in the great scheme of things, it is a perfectly good place. There are events, and they have their own properties. (Just as there are actions, with their properties.) My suspicion of actions was directed really towards the suggestion that they can have evaluative properties that diverge in certain ways from those of their agents. This problem, which concerns actions in the sense of things done more than in the sense of the doing of them, does not affect events at all. They have no agents, and there is no analogue of the doing/thing done distinction that is relevant to them. But are actions events? And if they are, so what? One thought is that if actions are events, and their agents are not, we already have in place a considerable metaphysical distinction between action and agent – one that it is futile to try somehow to diminish. There are reasons, from the point of view of moral philosophy, for trying to keep actions and events apart. I have argued elsewhere that, unlike events, actions are done for reasons; and those reasons are not to be understood as causes, whether psychological or otherwise.13 This gives me a reason in advance, as it were, to be wary here. On the other side, however, is the enormous weight of Davidsonian theory of action, which understands actions as events caused in a certain way by beliefs and desires of the agent.

408

Action in Moral Metaphysics

And this is combined with, or placed within, the general Davidsonian theory of meaning, which relates that understanding of action to the apparatus of first-order quantification and identity, and to the logical form of particular causal statements. I have tried to suggest elsewhere that Davidson’s argument for his view that reasons are causes is less than fully persuasive.14 Here I speak only of the way in which actions, as I am trying to understand them, may yet be able to be assimilated into the general Davidsonian framework of first-order quantification with identity. In other places I have spoken of a perhaps unfamiliar object, the ‘agent-in-acting’.15 I view this object as a particular, amenable to quantification, and capable of standing as one relatum in a singular causal statement. So a statement like ‘His pushing her caused her to fall over’ asserts a causal relation between an agent-in-acting and an event. This is not what is sometimes known as ‘agent-causation’, the paradigm form for which would be ‘he, by pushing her, caused her to fall over’. There is no suggestion here that agents are causes, and, what is more, causes of a totally unique sort – so that agent-causation is quite different from eventcausation. Perhaps we should say instead that there are two sorts of happenings, events and agents-in-actings, each of which can be quantified over, and each of which can be causally related to the other. But still it seems to me that happenings are explained in one sort of way, and doings are explained in another. Happenings are not explained by showing that it is a good thing that they happened, but that is exactly the sort of way that we explain doings, actions. I don’t want to find myself saying that happenings of one sort, the agents-in-actings, occur for reasons rather than for causes; the idea of something that happens for reasons, as opposed to being done for reasons, is very unappealing.16 This is the place to say a little about the explanation of action, as I conceive it. To explain a thing done is to explain the doing of it – except in the unusual case in which, for instance, one explains a thing said in a way that does not explain the saying of it. Such things are best thought of as explications rather than as explanations, and the distinction applies as well to things done as to things said. I can explain the path he took without explaining his taking that path, that is, why he took it. Actions are of course particulars, but each action is an action of a type – of many types, in fact; which is just to say that each action has many features, aspects or qualities, is ‘variously describable’. When one explains an action, one does not explain a type; one explains someone’s doing some particular thing, acting in some particular way on some particular occasion. To explain such a thing is to explain its features, that is, to say why it had these features (perhaps instead of those ones), why it was as it was, why he acted as he did (again, perhaps instead of some other way). All this is compatible with the fact that when one sets oneself to act in a particular way, one does not thereby select a particular action to be the one that one will do. Deciding how to act is not selecting a particular from a

Jonathan Dancy

409

variety on offer, like choosing a pair of shoes in a shop, but deciding on some features that one’s behaviour will have. So I might decide that I will behave graciously, and even work out some appropriate phrases to use when the occasion arises, but the actions that I eventually do are not and cannot be ones that I have selected as particulars. In fact, I leave most of the features of my actions up to chance, or the spur of the moment. Nonetheless, when done, my action is a particular, or rather, I have acted in a particular way. H. A. Prichard was troubled by the question how an action can have any properties until it is done. He felt that until the action is done, there is nothing there to choose. This problem undercuts a picture of decision which sees the selection of a course of action as rather like the choice of different items on a tray. It also undercuts the idea that I am doing what I am doing because it is right to do it; instead, we have to say that when I have done it, it will be right. The approach to actions that I am recommending, which tries to turn everything into agents acting, should circumvent these difficulties, which anyway must derive from some misconception. We can describe the situation thus: what is it right that I should do? It is right for me to return the favour, rather than to refuse. We do not have to say that the as yet non-existent action of returning the favour is already right.

VII I tried in Sections II–IV to say why I am wary of the idea that we should think of actions as individuals, with identity conditions. Section V argued that a similar wariness should extend to actings. But from the point of view of moral evaluation, the crucial contrast is between agent and thing done, not between agent and acting. This is because there is much less pressure to ascribe to the acting evaluative properties other than those of the agent; it seems, rather, that even if there are actings their evaluative properties must shadow those of the agent pretty closely. For our purposes it is things done that we should be worrying about, for a deflationary attitude towards them already deprives us of the most obvious ground for a sharp action/agent distinction, whatever we think of actings. My overall aim is to understand claims like ‘he did the right thing for the wrong reason’. Suppose that the deflationary or adverbial approach to things done is correct. This claim now means something like ‘he acted rightly, but for the wrong reasons’. If this does not mean that the action was right but the agent somehow wrong to do it, or wrong in doing it, what does it mean? Without actions as independent bearers of evaluative properties, have we the resources to say what we want to say? Let us approach this question from the bottom up. ‘He acted rightly, but for the wrong reasons’ means here ‘he V-ed, and in the situation he was right to V, but the reasons why he V-ed were not the reasons why he was right to V.’ ‘He was right to V’ means something like ‘he V-ed rightly’ – but

410 Action in Moral Metaphysics

not quite that; the correct expression in English is ‘he rightly V-ed’ or ‘rightly, he V-ed’ or even ‘he V-ed, rightly’ rather than ‘he V-ed rightly’, since rightness is not a way of acting. But we cannot say ‘he rightly acted’, except in cases where what was right was action rather than inaction. So what does ‘he rightly V-ed’ mean? It means, say, ‘he gave the money back, as he ought to have done’. It does not mean that he did it as he should have done, that is, in the proper way, for the right reason, or whatever. Take now another sort of claim: ‘he did the wrong thing, but his intentions were good’. This claim introduces no new problems; we already have the resources to understand it without abandoning deflationism. Cutting a fairly long story short, it is beginning to seem that at least some potential act/agent contrasts can be understood perfectly well within the constraints of deflationism. What deflationism suggests is that when we evaluate an agent’s behaviour, there are many things that we can consider: reasons, intention, motive, consequences and so on. There need be no very uniform story to tell; the agent, like the curate’s egg, may be good in parts and bad in other parts. What can distract us as philosophers is that we are keen to reach an overall evaluation of the agent; this pressure to sum up can tempt us to ascribe disparate evaluations to different objects – to action rather than agent. In this way we resolve a tension in the evaluated object by carving it into two, one for each side of the tension. But the deflationary attitude suggests that this pressure should be resisted. This situation is structurally similar to the one that occurs when, having given a detailed account of a situation in terms of thick evaluative concepts, we attempt somehow to sum up what we have said – to come to an overall evaluation; we can be led to suppose that any use of thick concepts in a given case must eventually sum to the use of a thin concept, if overall evaluation is not to be stymied. But this is an unnecessary and purely philosophical prejudice. There is no need for our evaluation at the level of the thick to sum to an overall judgement expressed in thin terms. It may be that once we have achieved our thick evaluation, there is nothing more to be said. Often, of course, there is no difficulty about reaching a conclusion at the level of the thin. But there is no need to suppose that the nature of evaluation itself requires that some such conclusion be available. Are there, then, any forms of mixed evaluation that impose upon us a distinction between action and agent that goes beyond anything sanctioned by the deflationary attitude? The obvious place to look is for cases where the agent is faultless, somehow, though the action wrong; or where the action, the thing done, was right but the agent gets no moral credit at all for it. However, there is a complication to be sorted out, which is that as well as the act/agent distinction, there is that between objective and subjective evaluation of the agent (as well as of the act, perhaps). What is the relation between these two distinctions? They are both used when things are not perfect in every respect; but an occasion for one is not an occasion for the

Jonathan Dancy

411

other. There are two general styles of cases where things can be less than perfect: 1. ignorance, error, bad luck 2. bad intent, motive; selfish reasons etc. The distinction between objective and subjective evaluation concerns the first of these. It is normal to think of this as a contrast between how things would have been morally had the agent’s beliefs all been true, and been the only relevant ones, on the one hand, and how they are as things actually stand, on the other. But the whole purpose of the subjective/objective contrast demands that the agent’s error or ignorance be not itself culpable. We do not ask, for instance, how things would have been if it had, as the agent believed, been acceptable to torture young children. And this is not just because moral propositions are necessarily true, if true at all. It is because the purpose of establishing a subjective evaluation of the agent is to provide an input to moral evaluation that is itself independent of moral evaluation. First we ask whether, if things had been as the agent non-culpably believed, he would have been acting as he ought. Then we ask whether things were as he non-culpably believed them to be. If he has false culpable beliefs, as all false moral beliefs would be, or was culpably ignorant, such matters are held constant across this contrast. So the subjective/objective contrast is not complete, as it were. It only applies to certain aspects of the agent’s overall perspective. So when we are dealing with culpable versus non-culpable ignorance or error or bad luck, we are in the domain of the subjective/objective distinction. There is a slightly different distinction from the objective/subjective one, which is worth keeping separate in one’s mind, and this is between the action that the agent takes himself to be doing and the one that he is actually doing. The question which action the agent takes himself to be doing is identical with the question what action he would have been doing had all his beliefs been true (including his culpable and moral beliefs). One could use this distinction to say that agents are only to be held responsible for the actions they take themselves to be doing, not for the ones they actually do. But I would resist this. One could also attempt to combine the two distinctions by saying that agents are only to be held responsible for the actions that they reasonably take themselves to have been doing. But I would resist this too. That one thought one was doing something else is of course some mitigation or excuse – especially if one’s error was non-culpable – but not a complete passport to freedom. Some cases where the agent acts for the wrong reasons are of our first sort; for instance, the agent may be non-culpably unaware of the considerations that are reasons on this occasion. Others are of the second sort; if the agent

412 Action in Moral Metaphysics

is aware of those considerations, but not aware that they are reasons, this itself might be culpable. But it might still be non-culpable, as it would be if, for instance, he was aware of the consideration that is the reason but nonculpably unaware of a second feature, also present, and whose presence is required for the first to be the reason it is. Again, if the agent takes something to be a reason when it is not, or to be a stronger or weaker reason than it really is, this may be culpable or non-culpable. Cases where things turn out unexpectedly may also fall on either side, depending on whether the expectation or lack of it is culpable or not. This contrast between subjective and objective evaluation is orthogonal to the act/agent distinction, because it remains whatever stance we take on the latter; even if we are really left with just the agent, there is still subjective and objective to be considered. What is more, the objective/subjective distinction does not provide us with a recipe for easy sorting out of complex cases. Suppose that we have a subjective evaluation of the agent, and an objective evaluation as well, and we now want an overall evaluation. There appear to be no rules that will help us. What we should not do, however, is to say that we need no help, since subjective evaluation is of agent, while objective evaluation is of act. Or rather, we can say this, but we should not delude ourselves with the thought that we have more than a new label for an old contrast; we should not suppose, that is, that the act/agent distinction in some way explains, grounds, sustains or justifies the objective/ subjective one. Let us then return to the official question: are there any forms of mixed evaluation that impose on us a distinction between action and agent that goes beyond anything sanctioned by the deflationary approach? We said that there are two possible cases to look at. The first is where the agent is faultless, as it were, though the action wrong – in the sense that the wrong thing is done. Derek Parfit discusses this possibility under the name of ‘blameless wrongdoing’.17 But he has a special case in mind; for him, what he calls agent-evaluation hinges entirely on the question whether the agent’s motive was a member of a set of motives that the agent would be wrong to cause himself to lose. This clearly leaves aside a good number of other relevant considerations, such as the agent’s reasons or intentions, that are here being taken as relevant to the evaluation of act rather than of agent. But this way of carving up relevant considerations is merely one of convenience. To put it another way, the agent may be blameless in this sense without being faultless. The agent may have been careless or negligent, for example, even though her motives were unimpugnable. Blamelessness in Parfit’s sense is not the correct name for an overall thin evaluation of the agent. Is there any other sense in which the agent may be blameless though the wrong thing was done? First, the agent could be subjectively blameless, if, for instance he was non-culpably ignorant of relevant facts. However, this

Jonathan Dancy

413

will not establish that he was objectively blameless, for, as we have already seen, blamelessness is not identical with subjective blamelessness. Second, it might be that had the agent actually been doing the action he took himself to be doing, he would have been objectively blameless. (This is not the same as subjective blamelessness.) But this does not establish that he is blameless, given the action he actually did. Beyond these two possibilities, neither of which is what we are looking for, there seems to me to be no chance that the agent should escape blame if the circumstances demanded that he act in one way, but he acted in another. So I think that there is no room for the combination of blameless agent and wrong action that might force us towards some notion of an action as a distinct bearer of evaluative properties. What about the opposite case, in which the agent gets no credit at all for an action even though he did the right thing? His motives were quite the wrong ones, say, or his intentions impure, or his reasons infused with selfinterest or malice. He ought to have repaid the money, and he did repay the money. He did it purely in hope of getting a larger loan, with which he was intending to abscond to Spain. This is not a case for the objective/subjective distinction, because there is here no suggestion of non-culpable error. Can the deflationist describe it? What do we mean when we say that he did what he ought? We mean that, for example, he should have repaid the money and he did repay the money. The mystery seems to be how, in that case, he does not get the ordinary credit associated with doing what one ought. But that mystery is easily solved. He does not get the ordinary credit because the reasons for which he repaid the money were not (among) the reasons why he should have repaid it. Is it necessary to say anything more than this? One reason for thinking that more needs to be said is a form of atomism. We might suppose that if he does repay the money for reasons that are (among) the reasons why he should repay it, he gets credit not only for getting the reasons right, as it were, but also for repaying the money – or, we might say, for sound intention and for success. There are, it seems, two strikes in his favour. Take a different case, then, in which though he does repay the money he does it for the wrong reasons. There should be one strike in his favour, by simple arithmetic. But we are being told that there is none. So something has gone wrong. The response to this is that simple arithmetic does not apply in this area. This is a holistic domain, not an atomistic one.

Conclusion My conclusion is that the deflationary approach to actions is the only one that avoids turning truths into falsehoods, and that it is capable of capturing all the different distinctions that we need in evaluating actions and

414 Action in Moral Metaphysics

agents. There is no need to go beyond this and to think of actions in a non- deflationary way.

Notes This essay was originally written in 1996, just as I was leaving Keele. I am very grateful to David Bakhurst, John Gardner, Max de Gaynesford, David McNaughton, Derek Parfit, Joseph Raz, Constantine Sandis, J. O. Urmson, my colleagues at Reading and to those who attended the Warwick conference on Analytic Philosophy in March 1996, for helpful discussion and comments. 1. Ross wrote: ‘To count the being under the influence of a certain motive, which is the precondition of the choice, as part of the object of choice, is to commit an error which is, in a very distant way, analogous to the error of treating the principles upon which we reason, as additional premises from which we reason. What we choose is to do the act to which the already existing predominant motive points, and it is easy to fall into the error of supposing that what we choose is to act from that motive.’ This is in Foundations of Ethics (Oxford: Clarendon Press, 1939), p. 120. 2. See Joseph Some Problems in Ethics (Oxford: Clarendon Press, 1931), chs. 4 & 5. 3. Foundations of Ethics, p. 128: ‘So far I have only established that besides the motive and a physical change there is a third thing – the setting oneself to bring about a change, or in other words an intentional act – and that this is a possible subject of the predicate right or wrong; we have still to consider whether it is so in fact.’ 4. Note my use of the ‘in some sense’ operator here; it can be used to turn any falsehood into a truth. 5. This is a different test because there is no commitment in the contrast approach to any thought that there can be action without intention, for instance – even though on some occasions above I called the contrast approach the ‘isolation approach’. 6. We might think that we can stop short of that, and see actions as bodily movements. Joseph thought that this was ridiculous, on the grounds that bodily movements are not proper objects of moral predicates. In this he was both right and wrong. He was right, in that movements of the body cannot bear moral properties in their own right, as it were. He was wrong, in that we can allow such predicates to be ascribed to movements of the body, as transferred epithets. The epithets are transferred to the movements from their accompanying features such as intention, motive, reason and so on. These features, if we are thinking of the movement of the body as the action, have all been placed on the side of the agent. As such, then, it seems that the action (understood still as bodily movement) can be called right or wrong without metaphysical incongruity. There is no category mistake in such predication. But the action is going to be incapable of bearing moral properties that are at odds with those of its agent, since the relevant predicates have been transferred from the agent side to the action side of the action/agent distinction. We have therefore not been helped to make sense of the various moral contrasts between action and agent that I listed at the beginning. 7. J. Hornsby, ‘On What’s Intentionally Done’, in S. Shute et al. eds Action and Value in Criminal Law (Oxford: Oxford University Press, 1993), pp. 55–74, at p. 56. 8. Hornsby, op. cit., p. 61. 9. J. O. Urmson suggested to me that with some actions, the answer to the modal question is yes and with others no. Some actions, that is, are such that one cannot

Jonathan Dancy

10.

11.

12. 13. 14. 15. 16.

17.

415

be doing them unless one has a certain intention; pruning, for example. Others are different; one can be cutting a stem even if one has no relevant intention at all – cutting it inadvertently, for instance. One might conclude from this that some actions necessarily include an intention and others do not, on the grounds that he might have been cutting with a different intention, but could not have been pruning with a different intention. But I think that there would be two things wrong with such a move. First, there is a difference between the claim that a certain intention is a necessary condition for pruning and the claim that that intention is a part of pruning. Second, Urmson’s point seems really to be about action-types rather than action-tokens. For an action to be of the type ‘pruning’, it must be done with a certain intention; to be of the type ‘cutting’, it need not be. But my modal questions are supposed to concern particular actions. Given that this cutting was done with a certain intention, could it have been done with a different one? It is admittedly epistemologically possible that it should, but is it metaphysically possible? Michael Tye asked me whether I endorsed this inference, and I said that I did not. But in fact I need to endorse it, at least tentatively, if I am to offer this position as a possible explanation of the modal facts about actions. For if the inference is unsound, that position is not committed to endorsing the modal facts; and therefore that position would be no explanation of the modal facts. For if a view V is not committed to proposition P, we can hardly explain P by appeal to V; for it would not be true, as it would need to be, that if V were true P would be. Perhaps we should distinguish two families of modal questions. There are those that focus on intention, motive, reason, purpose, end and so on. And there are those that focus on circumstances, place, time, manner. The agent had better go in the first family, I think. I find a considerable coincidence between what I say here and what is argued in Jonathan Bennett’s The Act Itself (Oxford: Oxford University Press, 1995), §9. See my Practical Reality (Oxford: Clarendon Press, 2000). Practical Reality, pp. 161–3. See my Moral Reasons (Oxford: Blackwell, 1993) ch. 13. It leaves us, however, with the awkward question what the relation is between an agent, conceived of as a long-term particular, and an agent-in-acting, conceived as a short-term one. Should we say that the one is constituted by a series of the others? If so, which series? These old and thorny questions I leave for another place. In his Reasons and Persons (Oxford: Oxford University Press, 1984), ch. 1. For a discussion, see ch. 13 of my Moral Reasons or, more extensively, my ‘Parfit and Indirectly Self-defeating Theories’, in J. Dancy ed. Reading Parfit (Blackwell, 1997), pp. 1–23.

22 Non-cognitivism and Motivation Nick Zangwill

§1 It is usually thought, following Hume, that considerations to do with motivation favour non-cognitivism over cognitivism about moral judgements. Hume’s motivation argument was that non-cognitivism does better than cognitivism at respecting the common observation that there is an immediate or internal connection between moral judgements and motivation. Hume’s famously asked (roughly): how can cognitivism account for the fact that moral judgements are intrinsically motivating? Let us take cognitivism to be the view that moral judgements are beliefs with a certain conditional truth property. This property is that if they are true, they are so in virtue of moral facts or states of affairs. Non-cognitivism, by contrast, is the view that moral judgements are identical with or are expressive of ‘attitudes’, which are thought of as mental states, such as pleasures, which at least lack that conditional truth property. This is a minimal conception of an ‘attitude’. (As we shall see, there is a problem about saying more.) In my view, the most promising cognitivist reply to the motivation argument is to deny that there is an internal connection between moral judgement and motivation. And common sense does not after all support a close tie between moral judgement and motivation in the way Hume thinks. In fact, it seems that the connection of moral judgement with motivation is quite variable. For example, the thought that it is my moral and professional duty to grade papers has a different motivational effect on me before and after I drink my morning coffee. And mercenaries are typically less affected by the thought that what they do for a living is morally wrong than other people or than they used to be earlier in their lives. Thus it seems that common observation supports the idea that we are variably moved by similar moral judgements. That is, the same moral judgements move different people differently, and they move the same person differently on different occasions. Let us call this the phenomenon of variable motivation. Two negative comments may 416

Nick Zangwill

417

help to clarify the phenomenon. First – the point is not merely the obvious one that people often fail to act on their moral judgements about what they should do. The point is the stronger one that there is a varying connection between moral judgements and motivation. Second – and relatedly – the variable motivation point is not merely that the motivation to act on moral judgements is sometimes outweighed by other motivations. The point is one concerning moral judgements and the strength of the motivation to act in accordance with those judgements, irrespective of other motivations. This phenomenon of variable motivation seems to support ‘motivational externalism’, which is the view that the motivational upshot of moral judgements stems from desires that are distinct from them (and hence are ‘external’ to them). Motivational internalism – Hume’s view – is the view that there is an intrinsic or internal tie between moral judgements and motivation. Variable motivation seems to support externalism because the best explanation of the variation is that there is no intrinsic or internal tie between moral judgements and motivation, and the variation in motivational strength stems from the varying strengths of distinct desires. In principle one might hold that moral judgements necessarily motivate, but how much they motivate varies from case to case. That is, one might in principle hold that there is no standard strength of desire that moral judgements involve. But that would be highly mysterious, since the variation in motivation, which is nevertheless supposed to be internal to moral judgements, would be completely unexplained and taken as primitive. By contrast, the externalist can supply a ready and simple explanation of the varying strengths, which is that it stems from the different strengths of the distinct desires. (Philippa Foot famously argued along these lines in Foot 1972; see further Svavarsdóttir 1999b and Zangwill 2003, 2007, 2008.)

§2 If the phenomenon of variable motivation is real, as it seems to be, the philosophical dialectic does an abrupt U-turn. For the following is as good an argument as the Humean one, if not better: how can a non-cognitivist account for the fact that moral judgements are variably motivating? The argument is that the best explanation of this is that they are not intrinsically motivating; and if they are not intrinsically motivating, then it seems that they must be cognitive. This argument runs in the opposite direction from Hume’s: given variable motivation, it is non-cognitivism that has a problem with motivation. It seems reasonable to suppose that non-cognitive states, such as the attitudes of approval and disapproval, are intrinsically motivating. That’s what an attitude is. But it seems that we can be more or less motivated by our moral judgements. From a non-cognitivist point of view, this is puzzling. How can it be explained? The cognitivist explanation is simple: it is that the motivational upshot of a moral judgement depends on

418 Non-cognitivism and Motivation

a distinct desire. Such desires vary in strengths; so the motivation generated will tend to vary in line with the strengths of the distinct desires. The non-cognitivist lacks such an explanation. The non-cognitivist, therefore, has a problem with variable motivation; and, given the ease with which cognitivism can explain it, considerations to do with motivation seem – contrary to Hume – to favour cognitivism, not non-cognitivism.

§3 How might a non-cognitivist respond? One non-cognitivist who has addressed the problem of variable motivation is Simon Blackburn. In his book Spreading the Word, he concedes that people sometimes do not care about morality. His fix for the problem is to say that moral judgements have ‘indirect and variable influences on action’ (Blackburn 1984, p. 186). But this does not help remove the puzzle about how that is possible. There is a problem about how the non-cognitivist can explain those ‘indirect and variable influences’. Why should the motivational aspect of attitudes vary? While the most plausible way for the non-cognitivist to reply to the variable motivation argument is to say that even though moral judgements are non-cognitive, variable motivation is possible and actual, this variable motivation needs to be explained not just posited. In his essay, ‘The Flight to Reality’, Blackburn goes further and attempts not only to take the phenomenon of variable motivation on board but also to explain it (Blackburn 1995). However, all he does there is to emphasise how passions can be prevented from having their normal motivational effects on action by competing motivations and failures of rationality. But this does not meet the variable motivation problem. Of course, a non-cognitivist can easily account for a variable connection between moral judgement and action, given competing desires and failures of rationality. But that does not deal with the difficulty posed by the varying motivations of deprived coffee addicts and clear-eyed mercenaries. For in these cases, no motivation, or little motivation, is conjoined with the moral judgement, irrespective of other motivations in play in the mental economy, and also irrespective of the connection to action. In his book Ruling Passions, Blackburn goes still further and says that attitudes can fail to have their normal motivational influence, but only because it is necessary that they normally do have that influence (Blackburn 1998, p. 61). This seems a more promising account. He draws several interesting comparisons. One is with the way love may produce non-standard motivations, such as in jealous Othello’s case, where his love for Desdemona leads him to kill her. Another comparison is with the connection between pleasure in beauty and judgements of beauty, where we sometimes judge of beauty in the absence of pleasure; it is possible to be unmoved by judgements

Nick Zangwill

419

of beauty; but generally beauty pleases. A third example is Milton’s Satan, from Paradise Lost, who is motivationally attracted to evil; but Satan ‘fell’ from previously having a more usual motivational profile. All these cases are supposed to be examples of attitudes that fail to move a person on special occasions but in a way that is intelligible only because attitudes have a standard motivational role. On rare occasions, the standard motivational role is lacking or is perverted, when something about the agent has gone wrong (‘out of joint’ is Blackburn’s phrase). Blackburn’s idea is that there can be motivationally deviant cases, but only against a background of motivational normality. His slogan is necessarily mostly moral judgements motivate (Blackburn 1999, p. 31). This response to the problem of variable motivation is subtle and interesting, and it is a significant improvement on the account in Spreading the Word and ‘The Flight to Reality’. Nevertheless, as we shall, it is also problematic.

§4 Sigrún Svavarsdóttir has advanced some interesting criticisms of Blackburn account in Ruling Passions in her critical notice of that book (Svavarsdóttir 1999a), to which Blackburn replied (Blackburn 1999); so let us begin by looking at this exchange. Svavarsdóttir objects that Blackburn’s cases of the jealous or disappointed lover, the jaded aesthetic judge, and of Satan as described by Milton do not exhaust an important range of motivational phenomena which are a greater threat to sentimentalist internalism (Svavarsdóttir 1999a, p. 22). The anger of Milton’s Satan or the jealous or disappointed lover do seem to derive from the positive motivation normally associated with love and with judgements of the good. They are only angry because of these more typical motivations. Satan is crucially a fallen angel. He was once a good angel, with the attendant motivations, just as a lover normally cares for his beloved and wants to be with her. Blackburn seems to be right about these cases, and this can be conceded. But some failures to be motivated by a moral judgement seem not to have their source in some prior attraction to the good (Svavarsdóttir 1999a, pp. 22–23). Svavarsdóttir may have in mind the moral cynic (Svavarsdóttir 1999b). My early morning grader and mercenary are also examples. In reply to Svavarsdóttir, Blackburn says that such a person ‘must be rueful’ (Blackburn 1999, pp. 31–32). But this seems to be an ad hoc assertion with little intrinsic plausibility; and it is put forward with no support. Maybe the cynic should be rueful. But many cynical people simply do not feel such emotions or if they do these emotions are weak. True cynics are cool and unemotional. They feel no regret, or little regret, for their lack of moral concern, although I suppose they might be intellectually curious about it. Many mercenaries regret rien or very little.

420

Non-cognitivism and Motivation

§5 Svavarsdóttir has another kind of example, which she puts to Blackburn. She asks: what about Barbara, who thinks that bird-watching is a valuable activity, yet who lacks any interest in bird-watching? Bird-watching does not motivate Barbara at all, despite her judgement that it is a valuable activity. Blackburn responds that although Barbara may have no motivation to engage in birdwatching, she is motivated to treat those who are motivated by bird-watching differently from those who collect paper napkins, which she regards as a worthless pastime (Blackburn 1999, p. 32). So there is after all some motivation consequential on the evaluation. On this account, the motivation generated by a moral judgement may be entirely second-order: if we judge that X is of value then although we may have absolutely no motivation to do X, we have motivational dispositions to those who do have motivations and attitudes to X. But this seems an unsatisfying response, although it is ingenious. Three comments: First, it is odd that there is no standard connection between attitudes and the motivations that are associated with them, so that at one time the judgement that X is valuable may involve a first-order motivation to do X and another time it may involve only a second-order motivation towards those who have first-order motivations and attitudes towards doing X. The resulting internalist view looks arbitrary. Second, on this suggestion, it seems that the necessary link between judgement and motivation is becoming ever more remote! The content of motivational internalism is becoming thinner and thinner. Third, the example needs delicate treatment; its interpretation is not uncontentious. Yes, the person who values bird-watching but has no interest in it might respect and want to talk to those who do have an interest in birdwatching, and they might want to avoid the company of those who collect paper napkins. But here different judgements are in question. The motivating judgements are not first-order judgements about bird-watching or papernapkin-collecting but second-order judgements about people who engage in these activities. It is understandable that one might make different secondorder judgements about people who engage in bird-watching and papernapkin-collecting despite lacking any interest in both activities. For example, it might be that the motivational difference consequential on the secondorder judgement derives from a desire to have intelligent conversation, and one might believe that people who are interested in valuable activities are likely to be more interesting to talk to than people who are interested in worthless activities. This is consistent with there being no motivational consequences of the first-order judgement about bird-watching. The fact that the second-order judgement may motivate does not show that the first-order judgement is somehow indirectly motivating.

Nick Zangwill

421

It is true that first- and second-order attitudes are often closely connected. Blackburn explores this in paper 1 of his Ruling Passions, where he describes what he calls the ‘sentimentalist ladder’: we may simply have sentiments towards things; but we may also ‘ascend the ladder’ and have attitudes to the attitudes of others or to ways of forming attitudes to things. Blackburn is right that we do sometimes have higher order moral sentiments. This ascent up the sentimentalist ladder makes most sense when one ascends to higher order sentiments given lower order sentiments. Suppose one disapproves of racism. Then one might be expected to approve of those who disapprove of racism or one might have attitudes to ways of forming attitudes to racism. Would one ascend the sentimentalist ladder and have such higher order sentiments if one lacks a first-order sentiment? Suppose one lacks any attitude to touching one’s left elbow with one’s right hand: might one nevertheless have a second-order attitude to that? It seems that one might. If someone has wildly positive or negative attitudes to touching one’s left elbow with one’s right hand, then one might well have a negative second-order attitude to his attitude – but only because one has a neutral first-order attitude to it. Thus first- and second-order attitudes are often intimately connected. Still, Svavarsdóttir’s point remains good that one may judge bird-watching to be much more valuable than collecting paper napkins despite lacking any motivation to engage in either of those activities. But for a non-cognitivist, the judgement that bird-watching has value should express a motivating attitude to my engaging in the activity of bird-watching, not to some something different.

§6 One way to interpret Blackburn’s necessarily-mostly idea would be that he takes being an attitude to be a cluster of dispositions, one of which could fail to obtain, so long as the others held. (For the general kind of strategy see Fodor 1991.) Thus an attitude could fail to have its usual motivational aspect (one element in the cluster) so long as the other attitude-making dispositions are in place. This idea would seem to help the non-cognitivist capture cases of variable motivation, not merely the outweighing of moral motivations by other motivations or the failure to act on moral motivations. However, this cluster suggestion is rather empty as it stands. What are these other attitude-making dispositions? Attitudes are typically characterised by Blackburn and other non-cognitivists in terms of a disjunction of either feeling or motivation, or perhaps dispositions to feel or be motivated (Blackburn 1998, chapter 1). The trouble is that moral attitudes are not necessarily feelings, since there surely are what Hume called ‘calm passions’ – so moral attitudes need not have a felt qualitative aspect. That leaves only motivation in the cluster. But then there is the problem of variable motivation. An attitude that is a calm passion might variably motivate. So it is hard to see how the

422

Non-cognitivism and Motivation

cluster idea can make good its promise to supply other moral-attitude-making dispositions in the cluster apart from feeling and motivation.

§7 Another interpretation of Blackburn’s necessarily-mostly idea is that he is subsuming deviant non-motivating moral attitudes into some wider class of moral attitudes, of a person, or of a group of people, which are more motivationally conventional. On such an account, no particular moral attitude has to have a disposition to motivate; but such moral attitudes are of the same sort as moral attitudes that do motivate in normal cases. However, the question that this suggestion faces is this: what makes a non-motivating moral attitude a mental state of the same sort as a motivating moral attitude? How can we say that a non-motivating attitude is a deviant moral attitude unless we have some independent grip on what it is to be a moral attitude? The point here is similar to one that Sydney Shoemaker levelled at David Lewis’ attempt to give a functionalist explanation of ‘mad pain’ (Shoemaker 1984, p. 269). Lewis was defending a functionalist account of pain according to which it is essential to pain to stand in various standard dispositional relations to other mental states. But ‘mad pain’, which lacks its usual casual role with respect to other mental states, seems actual and possible. Lewis tried to account for mad pain by saying that it is an instance of a ‘population’ of states that have a normal causal role. Shoemaker objected – decisively in my view – (1) that Lewis’ account makes what it is for something to be a pain too dependent how things are with things (the other mental states) entirely distinct from it. And (2) that Lewis’s account appeals to arbitrary and unprincipled groupings of ‘populations’ of states. Blackburn faces parallel difficulties: (1*) it is unsatisfactory to make being a moral attitude depend on mental states that are distinct from those moral attitudes; and (2*) it is unsatisfactory to make being a moral attitude turn on how things are in some arbitrary grouping or ‘population’ of mental states. Thus the problem for this attempt to take on board the phenomena of variable motivation is that he loses his grip on what a moral attitude is.

§8 Faced with the problem of variable motivation, and assuming that the phenomenon is embraced and not denied, a more drastic non-cognitivist strategy seems imaginable: instead of saying that they are variably motivating attitudes, simply saying that no moral attitudes have any motivational consequences by themselves. Thus the link to motivation is entirely severed. This non-cognitivist strategy is firmly to distinguish the attitude of approval from the desire that is necessary for action; so our non-cognitive moral attitudes motivate only in conjunction with distinct desires. This would be a kind of

Nick Zangwill

423

sentimentalist externalism, since any motivation to act has its source in some mental state distinct from the sentiment that is, or that is expressed in, the moral judgement. But the resulting theory of motivation would seem rather contrived. What would be the relation between approval and desire? Once the non-cognitivist says that moral attitudes are not at all intrinsically motivating, the noncognitivist’s theory of motivation becomes peculiar. Surely an attitude is exactly the sort of state that we expect to be intrinsically motivating. Furthermore, the idea of a non-motivating attitude seems odd. Consider pleasure, liking and fear, for example: these are just the sorts of mental states that motivate us, by themselves, or else they are intimately connected with motivation. Why are approval and disapproval not like this? What kind of states are these attitudes, which are non-cognitive, yet have no motivational upshot? It might be replied that Kant held a view like this in his Critique of Judgement (Kant 1928). He held that pleasure in the beautiful is divorced from desire: it is neither based on desire, nor does it produce desire. But this is no help for the non-cognitivist, for disinterested pleasure has a felt aspect; indeed for Kant it has a phenomenological quality that makes it a pleasure; it is just that unlike pleasure in the agreeable, and in moral or prudential goodness, it lacks a direct connection to desire. (It is nevertheless true that the pleasure does cause us to dwell in our perception of beautiful things.) But the externalist sentimentalist needs to claim that there are non-cognitive attitudes that are not connected with desire and also that lack phenomenological quality. So this strategy is not available. *

* *

In sum, the non-cognitivist account of motivation is far from unproblematic. The non-cognitivist has trouble telling us what moral attitudes are in a way that is consistent with the phenomenon of variable motivation. Given that the cognitivist has an easy explanation of variable motivation, it seems that cognitivism is preferable to non-cognitivism on the score of motivation, which is a reversal of the way the issue is usually perceived.*

Note * Many thanks for helpful comments to Sigrún Svavarsdóttir.

References Blackburn, Simon 1984: Spreading the Word, Oxford: Oxford University Press. Blackburn, Simon 1995: Simon Blackburn, ‘The Flight to Reality’, in R. Hursthouse et al. (eds), Virtues and Reasons: Philippa Foot and Moral Theory (Oxford: Clarendon Press, 1995).

424

Non-cognitivism and Motivation

Blackburn, Simon 1998: Ruling Passions, Oxford: Oxford University Press. Blackburn, Simon 1999: ‘Reply’, Philosophical Books 42. Fodor, Jerry 1991: ‘Ceteris Paribus Laws’, Mind. Foot, Philippa 1972: ‘Morality as a System of Hypothetical Imperatives’, Philosophical Review. Kant, Immanuel 1928: Critique of Judgement, Oxford: Oxford University Press. Shoemaker, Sydney 1984: ‘Some Varieties of Functionalism’, in Identity, Cause and Mind, Cambridge: Cambridge University Press. Svavarsdóttir, Sigrun 1999a: ‘On Simon Blackburn’s Ruling Passions’, Philosophical Books 42. Svavarsdóttir, Sigrun 1999b: ‘Moral Cognitivism and Motivation’, Philosophical Review, Vol. 108. Zangwill, Nick 2003: ‘Externalist Moral Motivation’, American Philosophical Quarterly. Zangwill, Nick 2007: ‘Besires and the Motivation Debate‘, Theoria, 2007. Zangwill, Nick 2008: ‘The Indifference Argument’, Philosophical Studies, 2007.

Index actions arational, 311 basic, 122, 130, 195, 340 basic vs. nonbasic, 15–19, 130 as bodily movements, 3, 13–16, 18, 20–1, 83, 259, 262, 303, 313–37 as causal processes, 2–3, 16–20 deflationary account, 7, 396–414 as events, 1–2, 126, 263, 276–8, 407 see also ‘Standard View’ individuation of, 1, 8, 175, 277, 285, 403–4, 409 intentional, 2, 4, 5, 6, 7, 76–8, 80, 107, 113, 122, 131, 132–68, 169–82, 184, 194–9, 203–19, 222, 229, 300–2, 313–37, 347–9, 362, 364–6, 370, 381 involuntary, 76, 78, 80, 89–90 non-intentional, 381 sub-intentional, 295–312 voluntary, 75–6, 84, 90, 149, 205, 209, 274, 308, 352, 362–3, 370, 373, 378 action trees, 285 adverbial approach to things done, 409 Aeneid, 369, 372 Aeschylus, 371 agency as an animal power, 6, 298 free, 7 over-intellectualization of, 6, 297 over-mentalization of, 295–311 akrasia, 65, 124, 191, 201, 219, 362, 369 alienation, 368, 377 Allan, W., 381, 383 Alvarez, M., 2, 5, 380, 381 Anscombe, G.E.M., 1, 2, 4–5, 9, 46–7, 95, 101, 104–5, 106, 107, 132–68, 169–83, 204–10, 214, 217, 298, 311, 316, 323, 325, 326, 330, 335, 337, 378, 390, 391, 393, 394 Antigone, 358, 377 Aphrodite, 356–7, 369, 371–2, 374 Aquinas, 6, 218, 219, 257–75, 362

arational animals, 267 motive force, 267 Aristotle, 7, 165, 213–15, 217, 218, 257, 259–60, 261, 262, 263, 267, 274, 275, 326, 336, 358, 360, 362, 365, 384–95 arm raising, 91, 306, 324–5, 340, 347–9, 352 Armstrong, D., 1, 110 attitudes, 80, 99, 133, 137, 143, 147, 168, 276, 281, 283, 290, 314, 362, 364, 373–4, 381, 416–23 see also pro-attitude Austin, J.L., 3, 23, 43, 129, 144, 398 bad faith, 358, 375, 381, 382 Baier, A., 2, 5–6, 379 behaviourism, 95, 102, 109, 139–40 Berkeley, 342 biological determinism, 375 Bittner, R., 2, 50–1, 61 Blackburn, S., 374, 381, 418–23 bodily movements, 313–37 see also actions Boulter, S., 2, 6 Bowra, C.M., 376 Brand, M., 1 Brandom, R., 200 Bratman, M., 1, 4, 108, 136, 138, 163, 165, 171, 366, 367 Bremmer, J.M., 358, 381 Broadie, S., 386 Buddhism, 375 Burkert, W., 358, 381 Cartwright, N., 332 causal powers, 20, 271, 272, 341–5, 348 causal processes, see actions; inference causation deviant causation, 6, 195, 268–70, 353–4 environmental, 374 mental, 107, 303, 306–08

425

426 Index causation – continued as a natural relation between events, 27, 29 physical, 351–2 substance causation, 7, 305, 342–3, 350 cause mental, 306–8, 347, 364 character, 5–6, 86, 91, 241–56, 360, 366, 370, 384–5, 389, 393 clarification, 2, 203 Clarke, R., 318, 335 coke machine example, 19–20 Collingwood, R.G., 8 Collins, A., 2, 93, 376 command, 133, 134, 149, 165, 209, 266, 270 conceptual analysis, 2, 233 con-reasons, see reasons criminal, 242, 407 Crisp, R., 48 Cromwell, O., 243, 245–7, 250–5 Cullity, G., 5, 227 Dancy, J., 2, 3, 5, 7, 21, 23, 29–30, 35, 43, 44, 45, 51, 52, 56–8, 60, 64, 202, 222–3, 227, 240, 271–3, 352, 379 Darwall, S., 24–5, 52 Davidson, D., 1–2, 4, 8, 13, 22, 24, 41–2, 60, 62–6, 71, 92–3, 95, 96, 97, 99, 105, 106, 108, 135–6, 145, 152, 163, 169–70, 175, 181, 195, 218, 243, 253, 268–9, 291, 316, 323, 326, 335, 353–4 deflationary, 7, 407, 409–14 deliberation, 6, 7, 24–5, 28, 30, 42, 64, 69–71, 76, 82, 86, 91, 193, 221, 268, 285–91, 371, 386–9, 392–3 Dennett, D., 95, 101–2, 282, 304 Descartes, R., 105, 303, 304, 344 desires, 3, 4, 5, 203–19, 220–40, 241, 243, 369 determinism, 7, 313–34, 366–7, 371, 374–5 see also psychological determinism disposition, 68, 72, 76, 80, 81, 86, 102–4, 109, 137, 155, 234–6, 242, 249, 288, 342, 345, 369, 373, 374, 420, 421–2 Dodds, E.R., 357–8, 360, 361

doing/doings, 78, 192, 400–2, 407 see also thing done dormitive virtues, 309 D’Oro, G., 2 Dray, W., 8 Dretske, F., 2–3, 6, 336 dualism animal-body, 304 of conceptual apparatus, 326 mind-body, 304 substance, 2, 344, 346 Ducasse, C.J., 1 Dummett, M., 45, 22 Enç, B., 1, 3, 13 Epicurus, 356 epiphenomenalism, 270 ethics meta-ethics, 6, 271 virtue ethics, 226 eudaimonia, 261, 358, 387 Eudemian Ethics, 384, 394 Euripides, 371, 376, 382 events external, 14, 19, 376 internal, 16, 19–21 mental, 13, 81, 85, 97, 99–100, 106, 108, 326, 343, 347, 367 neurological, 20, 270, 369 physical events, 13–14, 18, 68, 350–1 value and disvalue of, 30 Everson, S., 2, 3 evil, 90, 207, 218, 264, 369, 419 explanations belief-desire explanations, 6, 246, 251 causal, 4, 27–9, 46, 53, 57, 62, 79–81, 82, 87, 90, 93, 94, 95, 97, 99, 112, 164, 275, 297, 316, 317–19, 335, 354, 362 character-based, 6, 251 interpretative explanations, 4, 112–31 of moral and non-moral action, 7 motive, 81–2, 92 neuroscientific, 7, 18, 320, 323, 326, 333 nomic, 313–16, 319, 332 non-causal, 4 psychological, 6, 43, 297, 300 rational, 7, 28, 313–23, 325, 327–9, 331–3, 335, 339

Index 427 explanations – continued rational relation between facts, 27 reason-giving, 6, 297–9, 306, 311, 381 teleological, 91, 314, 339 in terms of agential reasons, 82–6 externalism with regard to moral motivation, 8, 417, 422–3 facticity, 167, 368, 373–6, 382 facts, 3, 5, 6, 13–14, 22–3, 25–46, 184–96, 200, 201 fate, 356, 359, 371, 373–6, 382 Fischer, J.M., 366 Fodor, J., 1, 95, 97, 421 Foot, P., 5, 224–7, 385, 417 Frankfurt, H., 7, 195, 366–7 freedom, 132, 265, 338–54, 363, 375, 382, 411 free will, 106, 340, 350, 368, 371, 375 Freud, S., 219, 364 Gaut, B., 5, 227 genes, 374 Ginet, C., 171, 295–6, 310 God, 157, 167, 238, 247, 265, 342, 355, 375 Goldman, A., 1, 169, 170, 175, 285, 362, 365, 378 good, 5, 22, 29 grammar, 94, 96, 99, 109, 143, 185, 363, 392, 402, 405–6 Hacker, P.M.S., 2, 3, 4 Hall, E., 378 Hamlyn, D.W., 384 Hegel, 360, 377 Heidegger, M., 374, 382 Hempel, C.G., 1, 8, 108 Hippolytus, 356–7, 371–2, 374 Homer, 357, 358, 360, 366 Hornsby, J., 2, 6, 299–300, 400–1 Hume, D., 5–6, 95, 97, 106, 203, 225, 241–56, 262, 264, 267, 271, 275, 310, 311, 343–4, 362, 366, 416–21 Hursthouse, R., 4, 169–70, 311 Hutto, D., 2 idealism, 342 Iliad, 381 immoral, 87, 181

individuation, 1, 8, 40, 41, 73, 175, 186, 277, 285, 329, 406–4, 409 see also actions, individuation of inference, 74 intellect, 7, 75–6, 220, 262, 265, 267–8, 273, 274, 289, 291, 384–90 Jedi Knights, 285 Joseph, H.W.B., 397, 414 Kant, I., 54, 223–4, 289, 291, 423 Kelly, S.D., 2, 4–5 Kenny, A., 381, 384, 390 Kim, J., 116, 320 Kitto, H.D.F., 359–61, 376, 377 Knobe, J., 2, 4–5 Korsgaard, C., 380 Kripke, S., 403 Lardinois, A.P.M.H., 359 laws of Newtonian mechanics, 333 LePore, E., 1 Lewis, D., 106, 214, 422 The Libation Bearers, 371 Locke, 344 Loening, R., 385–6 Lowe, E.J., 2, 6, 7, 380 Ludwig, K., 1 Malcolm, N., 381, 393 McCann, H., 310, 368, 380 McDowell, J., 5, 6, 58, 226–8 Melden, A.I., 95–101, 104 Mele, A., 1, 169, 170, 172 mind–body relation, 20 monism methodological, 92 ontological, 326 Moran, R., 4 motivation variable motivation, 8 motive, 77, 78, 81–2, 87, 92, 94–8, 100, 102, 104, 105, 209, 211, 241, 250, 262–3, 267, 272, 297, 305, 356, 370, 371, 382–3, 396–7, 399, 400, 402–3, 410–15 Müller, A., 388–9 myth, 339, 343, 356, 375–6 Nagel, T., 5, 220–40

428

Index

narrative, 252, 322, 370 necessity, 106, 204, 222, 230, 317, 319, 339, 356, 360, 367, 392 neurological defects, 329–1 disease, 324 events, see events, neurological internal cause, 20 neurophysiology, 128, 320–5, 327–9, 352 neuroscience, 369 see also explanations, neuroscientific Newton, I., see laws of Newtonian mechanics Nicomachean Ethics, 257, 259, 384, 394 Nietzsche, F., 364, 375 Nilsson, M.P., 357, 358, 360, 361, 366, 377 non-cognitivism, 8, 117, 416–23 North Carolina State v. Detter, 17 North Carolina State v. Williams, 17 Nussbaum, M., 377, 391, 393 O’ Connor, T., 2, 354, 380 ontology, 110, 323, 325–6, 342, 344, 350 O’Shaughnessy, B., 298–9, 302, 310 Oudemans, T.C.W., 359 Pantel, P.S., 359 paralysis, 348, 392 Parfit, D., 51, 55, 412 particulars, 37, 165, 400–1, 404–5, 408–9 Peacocke, C., 38–9 Persians, 360 phenomenalism, 342 Philips, D.Z., 382 philosophy experimental philosophy, 2, 4–5, 172, 173, 237 naturalist philosophy, 6, 7 phronêsis, 385, 387, 393 Pietroski, P., 2, 310 Pigden, C., 2, 5 Plato, 365, 389 Pollard, B., 2 Poole, A., 360–1, 378 Prades, J.L., 2 Price, A.W., 7

Prichard, H.A., 409 Principle of Un-charity, 178–9 pro-attitude, 60, 63, 99, 108, 218, 243, 253, 262–3, 267, 269–70, 362 psychological determinism, 366, 375 psychology, 2, 4, 7, 8, 18, 44, 53, 73, 104, 106, 136, 144, 147, 153, 161–2, 163, 229, 234, 241, 261, 300, 321–2, 357, 368, 369–71, 374 Putnam, H., 1 Quine, W.V.O., 103–5, 229, 323, 400 Quinn, W., 225 Railton, P., 365–6 randomness, 325, 338, 340, 368 rationalism, 22–4, 364–5, 367 Raz, J., 2, 3, 4, 5, 22–3, 29, 55, 365, 366 reasons agential, 4, 75, 80, 82–6, 91 as causes, 1, 3–4, 64–5, 87, 106, 339, 353, 408 concrete vs. abstract, 29, 37 con-reasons, 4, 62–74 count-notion and mass notion, 3, 42–3 explanatory, 3, 5, 28, 48, 55, 62, 184–9, 195, 201, 215–17 externalist account, 24, 202, 355 as facts, 22, 29–42 internalist account, 24–6, 54, 194, 202, 352 motivating, 5, 48, 50, 52, 55–6, 59, 61, 62, 71, 195, 204–5, 208, 211–12, 214–17, 219, 223, 226, 272–3 normative, 5, 25, 28–9, 43, 48, 50–2, 54–6, 62, 184, 186–97, 199–200, 201, 204, 216, 227–8, 263, 271–4 propositional account, 23–8, 31–42 psychologistic vs. rationalist accounts, 22–5, 29 as states of affairs, 23, 29–35, 38, 40 as valuable aspects of the world, 3, 22, 29 as what we believe and desire, 3, 21 reductio ad absurdum, 331, 397 Reid, T., 308, 380 relation causal, 31, 94–7, 106–7, 108, 117–18, 305, 306, 316, 408

Index

429

relation – continued temporal, 37 robust propositions, 42 Ross, W.D., 389, 397–8, 400, 414 Ruben, D-H., 2, 3–4 Rundle, B., 2 Ryle, G., 4, 95, 98, 100–4, 367, 380

thing done, 218, 400–5, 407–10 Thomism, 257, 275 Tillich, P., 374, 382 tragedy, 7, 356–76

Sandis, C., 2, 7 Sartre, J-P., 132, 158, 374–6 Schroeder, S., 2, 237 Schueler, G.F., 2, 4 Searle, J., 2, 26–7, 48 Sellars, W., 314, 363, 365–6 Setiya, K., 195 Shoemaker, S., 106, 110, 422 Smith, M., 1, 5, 48–9, 55–6, 364 Snowdon, P., 312 Sophocles, 358, 359, 376, 377 Spinoza, B., 326 ‘Standard View’, 2, 13–17, 276 state of affairs, 3, 23, 25, 28–35, 38, 40–3, 50, 76, 97, 149, 234, 264, 271–4, 287, 316, 352, 416 Steward, H., 2, 6 Stoecker, R., 2, 3, 6 Stone, M.J., 4 Stout, R., 2, 3 Stoutland, F., 7, 50 Strawson, P.F., 3, 23 sub-intentional, see action, sub-intentional success, 90, 92, 158, 159, 165, 199, 268, 291, 295, 349, 413 supervenience, 20, 115–24, 126, 130, 322, 331, 344, 373 Svavarsdóttir, S., 417, 419–21

Velleman, D., 2, 6, 108, 218, 243, 246, 251, 252–3, 366 Venus, 369 Vermeer, 8, 364 Vernant, J.P., 358, 359, 381 virtue, 220, 225–6, 250, 384–5, 389 volitionism, 347, 355 von Wright, G.H., 9, 43, 73, 92, 335

Tanney, J., 1, 2, 4 Telekinesis, 285, 375 Theognis, 358

Zaidman, L.B., 359 Zangwill, N., 8 Zeus, 357, 368–70, 372, 375–6

universals, 401, 405 Urmson, J.O., 414–15

Wallace, R.J., 1 Watson, G., 366, 380 weakness of will, 65, 70, 365, 375 West, D., 369–70 Westphal, K., 377 White, A.R., 92 Widerker, D., 380 the will, 65, 70, 249, 267, 274, 338, 348, 350–1, 365–7, 372, 375 Williams, B., 3, 25–7, 51, 52–9, 194–6, 360 Winnington-Ingram, R.P., 356–7, 360, 361, 367, 368, 370 Wittgenstein, L., 4, 8, 91, 94–105, 109, 114, 130, 132, 134, 141, 151–2, 154–5, 220, 374, 393, 405 Woods, M., 384, 387 Yaffe, G., 354