Values and the Reflective Point of View: On Expressivism, Self-knowledge and Agency

Citation preview

VALUES AND THE REFLECTIVE POINT OF VIEW

This page intentionally left blank

Values and the Reflective Point of View On Expressivism, Self-Knowledge and Agency

ROBERT DUNN The University of Sydney, Australia

© Robert Dunn 2006 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior permission of the publisher. Robert Dunn name has asserted his moral right under the Copyright, Designs and Patents Act, 1988, to be identified as the author of this work. Published by Ashgate Publishing Limited Gower House Croft Road Aldershot Hampshire GU11 3HR England

Ashgate Publishing Company Suite 420 101 Cherry Street Burlington, VT 05401-4405 USA

Ashgate website: http://www.ashgate.com British Library Cataloguing in Publication Data Dunn, Robert Values and the reflective point of view : on expressivism, self-knowledge and agency 1.Values 2.Reflection (Philosophy) 3.Agent (Philosophy) I.Title 121.8 Library of Congress Cataloging-in-Publication Data Dunn, Robert, 1949Values and the reflective point of view : on expressivism, self-knowledge, and agency / Robert Dunn. p. cm. Includes bibliographical references (p. ) and index. ISBN 0-7546-5412-5 (hardback : alk. paper) 1. Values. 2. Practical reason. 3. Expressivism (Ethics) I. Title. BD232.D798 2006 121’.8–dc22 2005034563 ISBN-13: 978-0-7546-5412-4 ISBN-10: 0-7546-5412-5 Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall.

Contents Preface Acknowledgments

vii ix

1

Values and Reflection

1

2

Moral Psychology and Expressivism

7

3

Self-knowledge, Truth and Value

37

4

Perverse Agency

61

5

Two Mistakes about Practical Reasoning

83

6

What’s Wrong with the Sensible Knave?

107

Appendix: Making Sense and Mental Partitions

133

Bibliography Index

143 149

This page intentionally left blank

Preface I owe a special debt of thanks to two philosopher friends. The first is André Gallois. Over the years I have had the privilege of participating in many enjoyable and searching discussions with André about the themes that I take up in the present essay. This is especially so in the case of self-knowledge, the topic under consideration in Chapter 3. I had the opportunity to talk over my ideas on self-knowledge and the first-person point of view with André as he was developing his own account – now published as The World Without, The Mind Within. These discussions sharpened my thinking about the issues, and my debt, in the present essay, to André’s way of conceiving of the modes of self-knowledge will be evident. The second philosopher and friend to whom I owe a special debt of thanks is David Simpson. In innumerable conversations David has engaged with me about almost every issue that I discuss in the present essay – as well as about many others. These conversations have been invaluable and have kept the philosophy alive and fun to pursue. I would also like to pay tribute to the various authors whom I discuss. Their ideas, and the arguments they advance for them, have largely shaped my preoccupations in this discussion. I would especially like to note, in this connection, Simon Blackburn’s work. Several years ago I wrote a book on weakness of will in which I defended cognitivism about values – that is, the thesis that values are simply beliefs. The present essay marks a change in view. It includes a defense of a version of expressivism about values – with a focus on the moral psychology of the matter. Blackburn in his writings, more than anyone, has persuaded me that expressivism is the best naturalistic approach within the theory of value. This is something of a heresy in some parts of the philosophical world in Australia. In the period 2000–2004, Michael Smith invited me to attend several Moral Psychology Workshops at the Research School of Social Sciences, Australian National University. I would like to thank Michael for the invitations and to register the benefit that I derived from these pleasant and stimulating occasions, which coincided with the period during which I was writing the book. Of particular benefit was a Workshop where Michael Bratman conducted some seminars on his work on identification. The debt I owe to Bratman’s contribution to the theory of identification will be evident from Chapter 4. Robert Dunn Sydney, February 2006

This page intentionally left blank

Acknowledgments Permission to quote extracts from the following published material is gratefully acknowledged: Simon Blackburn, Ruling Passions (Oxford: Oxford University Press, 1998); Christine M. Korsgaard, The Sources of Normativity, ed. Onora O’Neill (Cambridge: Cambridge University Press, 1996); Michael Smith, The Moral Problem (Oxford: Blackwell, 1994), and J. David Velleman, The Possibility of Practical Reason (Oxford: Oxford University Press, 2000). Material from the following published articles by the author has been incorporated into the present essay and is reproduced here with permission from the original publishers: ‘Two Theories of Mental Division,’ Australasian Journal of Philosophy 72 (1994): 302–16; ‘Knowing What I’m About To Do Without Evidence,’ International Journal of Philosophical Studies 6 (1998): 231–52; ‘Is Satan a Lover of the Good?,’ Ratio 13 (2000): 13–27; ‘Moral Psychology and Expressivism,’ European Journal of Philosophy 12:2 (2004): 178–98.

This page intentionally left blank

Chapter 1

Values and Reflection Values are central in our lives. We call our values beliefs, but it’s a question how best to categorize them. These values pervade and shape our psychology, our agency, and our lives as reflective and self-knowing subjects. A philosopher like Kant holds that reason itself forces certain values on us. We can be true to our values or we can betray them. Some philosophers hold that we must be divided in our minds when we betray our values. We can even betray our values to the point of perversity, opting for what we see as entirely bad. What makes this possible is that, despite the way in which many theories of agency cast us, we are not always lovers of the good in so far as we are pursuers of goals. Sometimes we are just lovers of success in action, focussed on how to realize a current end-inview, be it cheery or bleak. The topics of practical reflection are plural, as perverse agency makes vivid. One mistake about deliberation about what to do is to suppose that it never foregrounds our desires, except incidentally – that it is exclusively focussed on features of our choice situations that those desires make salient for us. This mistake typically goes along with a collapse of desires into values, a move which ignores the vicissitudes of valuing and which erases the difference between purposive and reasoned activity. A different kind of mistake about practical deliberation is to suppose that it always foregrounds our desires, a move which ignores the plurality of considerations that provide reasons within reflection and which contributes to various distortions within the theory of agency. These themes set the agenda for the present essay. They mark out for discussion various aspects of our values as these values figure within reflection, especially reflection that relates to our desires, wills, deeds and passions. They also provide a context for indicating how an expressivist approach to values can avoid an impoverished moral psychology. Indeed, the familiar assumption that a commitment to expressivism about values incurs an impoverished moral psychology turns out to be far from the truth. There is a ‘natural fit’ between the roles that values play in the lives of reflective and self-knowing agents and an expressivist interpretation of these values that is sensitive to the complexities of the relations which such agents can have towards their own values. Let’s briefly consider the arguments of Chapters 2–6. We number our values among our beliefs. But it’s never enough simply to leave it at that. Consider our beliefs about what it would be good to do, or about how it would be admirable to be. These beliefs seem to be psychologically very different from our non-evaluative beliefs about what the world around us is like. The values

2

Values and the Reflective Point of View

we set on courses of action or ways of being seem to be somehow closely related to our desires or emotions in ways that set them apart from our simple, worldly beliefs. The problem is how to spell out the contrast. I think that we can do this by looking to the way in which, within deliberation, certain questions take others as discursive transforms – in particular, to the way in which questions about what to desire, will, do and feel reflectively ‘give way’ to evaluative questions as the issues in focus, with answers to the latter dictating answers to the former. In Chapter 2, I give an account of discursive transforms. The central idea is that some questions are reflectively indistinguishable from others, and there’s an order of dominance among them. So, for example, the question of whether to desire p will ‘give way’, upon consideration, to the question of whether p (or the situation p represents) is somehow good, with the answer to this evaluative question dictating an answer to the question of what to desire. The phenomenon of discursive transforms, to my mind, reveals what is distinctive about our practical and passional values: typically, within reflection, we put ourselves in a position to form desires, make choices, take up affective reactions, and the like, by reasoning about evaluative issues; and the evaluative conclusions we come to are, or involve, dispositions to acquire matching desires, attitudes of will or feelings. On this reading, the phenomenon of discursive transforms points to a cautious expressivist account of values: expressivist, because the values are taken to be, or to involve, dispositions that relate to practical and passional aspects of our natures; and cautious, because the story allows for the vicissitudes of valuing – for how pathologies can sever the link between our values and our desires, wills or emotions.1 There is, however, a serious challenge to this line of thinking. It comes from the kind of ‘internalist’ position defended by Michael Smith. Smith, like the cautious expressivist, holds that there is a necessary (albeit defeasible) connection between our values and our desires. However, Smith’s defense of internalism turns on an analysis of the content of our evaluative beliefs. Thus, our beliefs about what it’s desirable to do, he claims, are beliefs about what we would want to do if we were fully rational; and so, anyone with a belief with that content will either have a matching desire or be irrational (as a matter of a failure of coherence). In cases where we are rational, says Smith, what causes us to desire the good (as we see it) is our evaluative belief, combined with the tendency towards coherence that characterizes us to the extent we are rational. Now, Smith looks well placed to give an account of why consideration of what it’s desirable to do is especially suited to function as a focus for practical deliberation, without construing any evaluative belief we form as intrinsically practical: the content of such an evaluative preoccupation, as revealed by analysis, secures its 1 As I indicate in Chapter 2, the way into expressivism through an account of discursive transforms has an obvious affinity with Simon Blackburn’s recommendation that we can read off what is distinctive about valuing by considering our evaluative practices, in which evaluative propositions serve as the focus for practical thought and discussion.

Values and Reflection

3

unique relevance to deliberation about what to desire/do. In Chapter 2, I argue that we shouldn’t accept Smith’s brand of internalism. On the one hand, Smith seems to accept the very plausible idea that propositional mental states are (in core part) rational in their make-up in a way that reflects their direction of fit; that beliefs (for example) are (in core part) constituted by coherence-preserving dispositions to infer, like the modus ponens habit of inference, which reflect their mind-to-world direction of fit.2 On the other hand, Smith treats the tendency towards coherence at work where we match desires to values about what to do as external to those values. And yet, we can make perfectly good sense of the idea that such values are dual in their direction of fit, having a mind-to-world direction of fit with respect to their evaluative contents (which marks them out as passive thoughts) and a world-to-mind direction of fit with respect to their embedded contents (which marks them out as active thoughts).3 Moreover, we can do all this without incurring the cost that Smith thinks vitiates any such flexibility about direction of fit – namely, that we ignore the vicissitudes of valuing, by treating values as besires,4 or as states that are, or include, desires. For the cautious expressivist will hold that it is in terms of the very tendency towards coherence that operates where we match desires to values that we can make good functional sense of the notion that values have a world-to-mind direction of fit in relation to their sub-contents. And this approach yields a further gain. It lets us keep the account of values within a simple, unified theory about how our minds tend towards coherence; namely, they have this feature because mental states are rational in their constitution in a way that reflects their direction(s) of fit. By these lights, we should think of the psychology of values in the following way, if we share the commonsense idea that mental states have a constitution that can be mapped onto their direction of fit. Values are complex in their make-up because they are dual in their direction of fit: they include coherence-preserving dispositions that relate to them as passive thoughts (such as habits of modus ponens inference)5 and coherence-preserving dispositions that relate to them as active thoughts (such as the disposition to desire what’s seen as good, choose what’s regarded as fit to be chosen,

2 I take it that Smith accepts this idea as part of his acceptance of a version of commonsense functionalism about mental states. 3 This applies to our beliefs about what we should believe, as well. If we believe that we should believe p when we shouldn’t, we fault our belief as false; and, if we fail to believe p when we think that we should, we fault our pattern of belief. Thus, it’s misleading to say that difference in direction of fit marks out the difference between (say) theory and practice. Hence, I introduce the contrast between passive and active propositional thoughts. I discuss values about what to believe in Chapter 2, pages 28–9. 4 So-called besires are held to be single, unitary kinds of mental state that are belieflike/desire-like amalgams. I discuss these in Chapter 2, pages 14–19. 5 Even granting that our practical and passional values have a passive aspect, an expressivist treatment will want to go on to theorize in a distinctive way about the dispositions to infer that they therein incorporate. See Chapter 2, note 66.

Values and the Reflective Point of View

4 6

and so on). But then, this brings us back to expressivist internalism and to a cautious expressivist reading of the role of evaluative questions as the discursive transforms of questions of what to desire, will, do and feel.7 The role of some questions as discursive transforms of others within deliberation helps us to understand more than just the nature of valuing. It also (perhaps surprisingly) helps us to understand a certain kind of deliberatively available selfknowledge and the way in which valuing enters into its acquisition. This is the theme I develop in Chapter 3. I argue there that there’s a non-observational way of taking up questions about what we desire, will, feel or are going to do, such that the topic in focus becomes what’s good, or fit to be willed or fit to be felt about in a certain way, just as there’s a non-observational way of taking up questions about what we believe such that the topic in focus becomes what’s true. By reasoning about truth and values, we have a unique way of coming to know our own minds and hearts, and what we are going to make happen in the world. The values that we form within reflection enable us, where all goes well, both to shape and to come to know about aspects of our mental lives and our future agency. However, our practices and passions do not always match the values we arrive at within reflection: our values have their vicissitudes. Sometimes even, at the extreme, we are perverse and opt for what we reckon to be evil, without qualification. Perhaps, filled with pride or despair, we seek to destroy everything we value. There are, however, conceptions of agency that make perversity look impossible: on these stories, we are always pursuers of the good in so far as we are pursuers of goals; and so, apparently perverse agents turn out to be closet lovers of the good. In Chapter 4, I argue that we can save the possibility of perversity (and save it while remaining expressivist about values) by not conflating purposive and reasoned activity, and by recognizing how, as agents of reasoned activity, we are sometime lovers of the good and sometime mere lovers of success in action. Perverse agents are the latter, bent on success with respect to their bleak aims – aims with which they may even identify. Lovers of success in action foreground a current desire, or set of desires, in their practical reasoning. However, some philosophers hold that we are always self-absorbed in our reasonings about what to do, invariably foregrounding our current desires. By these lights, the point of view we occupy in practical reflection is characterized by self-consciousness. I discuss this issue in Chapter 5, arguing that it’s a distortion of practical reflection to hold either that we always foreground our current desires or that we never do (except incidentally). Practical reflection is plural in its content. Sometimes we reason about our desires, as when we are merely after success with respect to an end-in-view, or when we take the satisfaction of a desire, or set of desires, to matter to the realization of some serious or substantive 6 It’s then an interesting question whether we should count the dispositions that our values include as active propositional thoughts as dispositions to infer. I consider this issue in Chapter 2, pages 30–36. 7 In Chapter 2, pages 14–19, I argue against the competing view that our values are intrinsically practical and so-called besires.

Values and Reflection

5

good. However, very often, perhaps even typically, the topics of our reflection don’t mention our desires at all: we reason about features of the situation of choice that attract our attention, but not about the desires that might make those features salient. The view that practical reflection always takes a current desire, or set of desires, as its (focal) topic has been called ‘the objectifying mistake’. In Chapter 5, I indicate ways in which this mistake has played a role in various distortions in the theory of agency, especially in relation to themes like identification with our desires and autonomy. I also discuss how some significant attempts to avoid the objectifying mistake about the contents of reflection have involved the substitution of a different kind of error – namely, ‘the desire-as-valuing mistake.’ This is a mistake that mislocates desires, assimilated to values, among the mental states that enter into practical reasoning as constituents. Expressivism about values might be supposed to be especially vulnerable to this second mistake; however, it is avoided by cautious expressivism. Finally, in Chapter 6, I turn to the perennial issue of whether there are any values that reason itself forces on us within the reflective point of view. In particular, I focus on a recent attempt by Christine Korsgaard to argue that this is indeed the case. Korsgaard takes her cue from Kant’s idea that willing evil (in formal terms) implicates us in a contradiction and (in material terms) violates an end that is inescapable or necessary within practical reflection. Korsgaard argues that evil values (like those of Hume’s sensible knave) are ‘reflectively unstable’ because they are inconsistent with a deep value to which we are all committed, if we are to value anything at all – namely, our common humanity or rational nature (Kant’s necessary end). It’s a fascinating idea. However, so far as I can see, Korsgaard fails to make it good. Her argument, in its detail, relies on an unsuccessful defense of the Kantian theme that reasons for one are reasons for all – in her terms, that reasons are public. For all that Korsgaard says, the ambition of showing that the cure for evil values (like those of the sensible knave) lies in being freed of a contradiction remains among the illusions of philosophy.

This page intentionally left blank

Chapter 2

Moral Psychology and Expressivism It’s a common thought that our beliefs about what’s good, or fit to be chosen, or obligatory or admirable are different from our beliefs about what the world around us is like. We call our evaluations beliefs, but they seem to be somehow intimately tied up with our desires, decisions, intentions, actions and emotions – with our practical and passional lives – in a way that marks them out from our non-evaluative, worldly beliefs. The question is how to get at this difference and articulate it. Expressivists try to get at and articulate this difference by looking to our evaluative practices. Simon Blackburn characterizes the expressivist approach in the following terms. Evaluations that take our desires, wills, deeds and feelings as their topics are beliefs, but it doesn’t help us to identify what’s distinctive about them to call them beliefs, or their linguistic expressions statements: The smooth clothing of statements proposed as true or denied as false disguises the living body beneath. The expressivist task is to reveal that clothing for what it is – but that is not to say that we should always try to do without it.1

The way to uncover what’s distinctive about our practical and passional values, says Blackburn, is to attend to the activities that surround them. These include: … valuing, grading, forbidding, permitting, forming resolves, backing off, communicating emotion such as anger or resentment, embarrassment or shame, voicing attitudes such as admiration, or disdain or contempt, or even disgust, querying conduct, pressing attack, warding it off.2

1 Ruling Passions (Oxford, 1998), p. 51. Blackburn’s view is that to hold something to be good or right is to have a belief or make a judgment, but that we need a sorted notion of belief or judgment. See his remarks in ‘Realism, Quasi, or Queasy,’ in John Haldane and Crispin Wright (eds), Reality, Representation and Projection (Oxford, 1993), pp. 366–8. See also his sympathetic remarks about the way in which Ramsey and Wittgenstein need to work with a sorted notion of a proposition or truth-aptitude (rather than truth) in ‘Wittgenstein, Wright, Rorty and Minimalism,’ Mind, 107 (1998): 166–8. Of course, as Blackburn notes, many expressivists have insisted on ‘a sharp separation of spheres,’ and so on a sharp contrast between our evaluations and our beliefs, treating only the latter as expressing propositions or having truth-apt contents. And indeed, Blackburn himself has at times seemed closer to this contrastive position. See his remarks in Spreading the Word (Oxford, 1984), pp. 187–8. 2 Ruling Passions, p. 51.

8

Values and the Reflective Point of View

What we learn, according to Blackburn, is that the role of the evaluative proposition is to serve as the focus for practical thought and discussion: what to do, what to admire, whom to badger, when to repent, and so on.3 Evaluative thought and talk just is thought and talk about what to do, what to admire, and so on. Thus, in typical cases, wondering whether X is good or right is wondering what to do/what to admire; being undecided about whether X is good or right is not knowing what to do/what to admire; knowing X is good is knowing to choose X/admire X, and so on.4 In this way, attention to our evaluative practices locates what is distinctive about the values that underlie them: these values connect with our practical and passional natures, and comprise clusters of dispositions that relate to our desires, decisions, intentions, actions and emotions.5

Evaluative Thought and Talk: Its Distinctive Context Let’s consider this theme in more detail. It can be conceded from the outset that evaluative sentences (evaluatives) are not, or do not entail, imperatives. Suppose I say something like the following: ‘You ought to leave at once. However, I’m not telling you, or advising you or even suggesting to you, what to do.’ It’s true that such a remark is odd; but it contrasts with remarks like ‘Snow is white, but I’m not saying that this is so’ or ‘Leave at once, but I’m not saying to you that you do this.’ The latter two remarks are paradoxical in a way the first remark isn’t. Presumably, in the latter cases, the paradox has to do with mood meaning: the indicative form of the main verb of a sentence tells us that, in the absence of contextual counter-signs, literal utterances of that sentence will be cases of stating that something is so (or the like), just as the imperative form of the main verb of a sentence tells us that literal utterances of that sentence typically will be instances of telling someone to make something so (or the like). In short, signs of mood are signs of illocutionary-force potential.6 The contrast between the paradoxical examples and the evaluative example shows that we go wrong if, like R.M. Hare, we construe evaluative sentences as entailing imperatives.7 3 Blackburn (like many) often uses ‘practical’ as a cover-term for thought and discussion that relates to our desires, wills, actions and feelings. I shall regularly do the same for convenience. 4 Ruling Passions, pp. 51, 68–9. See too Think (Oxford, 1999), p. 286. It’s appropriate to talk of knowledge, says Blackburn, in cases where we’re certain that an improvement in our practical stance to the world is ruled out (‘I know that happiness is better than pain, that promises deserve some respect, and so on.’): Ruling Passions, pp. 85, 306–7, 318–19. 5 Blackburn ‘fleshes in’ his picture of the dispositions involved in valuing in ibid., pp. 66–8. 6 See my discussion, The Possibility of Weakness of Will (Indianapolis, IN, 1987), pp. 30–39. 7 I set out the development of Hare’s thesis that categorical affirmatives that include evaluative terms have, as a result, an imperative-mood element in their meaning, in ibid.,

Moral Psychology and Expressivism

9

However, a suitably crafted expressivism can acknowledge this. Let’s grant that evaluative sentences are as they seem syntactically: evaluatives are indicatives. And let’s concede that valuers can differ in their preparedness to give commands, offer advice, issue exhortations, or even make suggestions, about what to do. When the moralistic form their values, they will be only too ready to lay down the law; busybodies will be forever cued to tell people what to do; those who tend to mind their own business will be more sparing, as will those who respect the independence of others, and so on. What the expressivist should claim is as follows. We understand what is distinctive about our evaluative states of mind by attending to the way in which the evaluatives that express their contents are especially apt as answers to questions of what to choose or do, what to desire or how to feel. What makes evaluatives apt as answers to such questions – which are the interrogative transforms of imperatives (or optatives) – is that these issues become tractable in thought and discussion by turning into questions of what is right, or good, or desirable, or admirable, and so on. Evaluatives are distinctively apt to serve as answers to questions about what to choose, do, desire or feel because they are answers to questions which typically serve as the discursive transforms of these imperatival (or optative) questions. This point can be put in terms of the indistinguishability, or inseparability, of certain questions. Consider the question of whether to desire p: of whether p is to be desired. Suppose I take up this question in an object-focussed way. In that case, I consider the question of whether to desire p in such a way that any reason for desiring p recommends desiring p by virtue of the way in which it recommends p (or the situation that p represents) as good, and hence is a first-order reason for desire.8 And when I take up the question of whether to desire p in this object-focussed way,

Chapter 2. Hare holds that, in so far as such evaluative sentences have an imperative-mood element in their meaning, the evaluative thoughts that they express are revealed as (partially) motivational. 8 I can take up the issue of whether to desire p in an object-focussed way or in an attitude-focussed way. If I consider the question of whether to desire p in an attitude-focussed way, then any reason for desiring p recommends desiring p by virtue of the way in which it recommends having the desire that (being a desirer of) p, and hence is a second-order reason for desire, which effectively recommends that I should desire to desire p. Our desires are representations of how things are to be. It is this direction of fit-related feature of desire that explains why object-related reasons for desiring p are considerations that support p as good (in some way). Just because a desire that p dictates that the world be a certain way, there’s always an issue of whether it’s a well-directed dictate: of whether it’s a desire that things be a way that it would be good for them to be. The dictum ‘desire aims at the good’ points to the fact that any desire that p is successful as a dictate on the world just to the extent that it’s a dictate that the world be a way that’s of value. And this locates the logical space of object-focussed reasons for desiring p: they are considerations that recommend the desire that p as a successful dictate just to the extent that they support the situation that p represents as a good way for things to be. I say more about the world-to-mind direction of fit of desires in the text below.

10

Values and the Reflective Point of View

it will be indistinguishable, or inseparable, from the question of whether p is good in the following sense: any consideration (or set of considerations) that I endorse and count as decisive in favor of p’s being good (in some way) will equally count for me as decisive in favor of desiring p (in so far as it has the relevant feature), and the converse.9 Indistinguishability is symmetric. Even so, within the deliberative perspective, there’s an order of dominance among indistinguishable questions. Thus, when I consider the (object-focussed) question of whether to desire p, that question will be perspectivally ‘transparent’ with respect to the question of whether p is good (in some way): the first question will ‘give way’ to the second as the issue in focus, and the answer to the second question (whether p is good) will dictate an answer to the first (whether to desire p). In this way, the question of whether p is good (in some respect) will come to do duty, in such a case, as the deliberative or discursive transform of the question of whether to desire p. The expressivist claims that we understand what is distinctive about beliefs to the effect that p is good by attending to the distinctive functional role of the evaluatives that express such contents as answers to questions of what to desire. The lesson to draw is this: typically, we deliberatively put ourselves in a position to form desires by reasoning about what’s good; and the evaluative conclusions we come to in such cases are, or involve, commitments (dispositions) to form matching desires. Note that the careful expressivist won’t say that the evaluative conclusions we draw, when we consider questions of what to desire, are themselves desires. For we need to be mindful of the ways in which pathologies of the heart can break the links between our values and our desires. In healthy cases, the values that we form in deliberatively taking up issues of what to desire will find expression, or manifest themselves, in our desires. However, the story has to allow for the vicissitudes of valuing under the impact of pathologies that cause our values and desires to come apart. Of special interest to expressivist accounts is the role of evaluatives as answers to questions of what to do. This has been the classic focus of such accounts. Evaluatives are especially apt as answers to questions of what to do because they are answers to evaluative questions which typically serve as the discursive transforms of these questions. Two steps get us to the evaluative questions which discursively substitute for questions of what to do: questions of what to do ‘give way’ in deliberation to questions of what to choose (decide or intend); and the latter, in turn, transform into questions of which option is fit to be chosen (decided on or intended).10 9 To count a consideration k that one accepts as decisive in favor of p’s being good (or, as settling the question of whether p is good in the affirmative) is to have an attitude towards k whose expression would be judging that p is good on the basis of k. Similarly, to count a consideration k that one accepts as decisive in favor of desiring p (or, as settling the question of whether to desire p in favor of so desiring) is to have an attitude towards k whose expression would be forming a desire that p on the basis of k. 10 The issue of what to do perspectivally ‘gives way’ to the (inseparable) issue of what to choose (decide or intend): the first issue counts as settled by the answer to the second. And

Moral Psychology and Expressivism

11

The expressivist then claims that the lesson to draw from the way in which evaluatives come to serve as answers to questions of what to do is this: typically, we deliberatively put ourselves in a position to make choices or decisions, or to form intentions, by reasoning about what course of action is fit to be chosen, decided on or intended; and the evaluations we make in such cases are, or involve, dispositions to make choices or decisions, or to form intentions, that match them. Once again, the careful expressivist won’t say that the evaluative conclusions we arrive at, when we consider questions of what to do, are themselves choices, or decisions or intentions. For we need to allow for the ways in which pathologies of the heart can disrupt the connection between our values and our wills. Rather, the point will be that these evaluations are, or involve, dispositions which, in felicitous cases, will find expression in acts or attitudes of the will that match them. We can draw the picture this way. In cases where we judge that a particular course of action is best, everything considered, and no pathology intervenes, our judgment of which action is fit to be chosen (decided on or intended) will match our all-in best judgment. And, in cases where we deem a certain action as fit to be willed in one of these ways, and no pathology intervenes at that point, the outright evaluation will be matched by a relevant choice, decision or intention.11 The distinctive role of evaluatives as answers to questions of what to desire or do enables us to identify what is special about the evaluations whose contents they express: these evaluations are, or include, practical dispositions. This strategy for identifying what’s special about evaluations extends to evaluative indicatives in their distinctive role as answers to questions of how to feel. Thus, consider evaluative sentences like ‘It’s regrettable that I did A’ or ‘Freddy is admirable.’ By expressivist lights, we understand what’s distinctive about the evaluative states of mind whose contents are expressible by these evaluative sentences, by attending to the way in which the sentences are apt for answering questions which are the discursive transforms of (object-focussed) questions of what to regret or whom to admire. What this functional aptitude reveals is that the relevant evaluations are, or incorporate, dispositions to regret or dispositions to admire.

the issue of what to choose (and so on) ‘gives way’ to the (inseparable) issue of what option is fit to be chosen (and so on): the issue of what to choose (and so on) counts as settled by the answer to the relevant evaluative question. 11 Here I am adopting a distinction introduced by Donald Davidson between all-in evaluations and all-out or outright evaluations. However, Davidson identifies the former with desires and the latter with intentions, thus ignoring the way in which our values can come apart from our desires and intentions. See my discussion in Chapters 4–5. Davidson summarizes his use of the distinction between all-in and all-out evaluations in ‘Intending,’ Essays on Actions and Events (New York, 1980), pp. 97ff., and in ‘Reply to Christopher Peacocke,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford, 1985) pp. 209–11. Note Davidson’s coining of the term ‘intention-worthy’ for deployment in all-out evaluations (in his reply to Peacocke, p. 211). I am using ‘fit to be chosen (and so on)’ in a similar way.

12

Values and the Reflective Point of View

Of course, in so far as evaluative sentences that recommend desires, choices, decisions, intentions, actions and sentiments are indicatives, they are also apt as answers to questions of what to believe, because they are apt as answers to questions of what’s true.12 The expressivist can acknowledge this. The expressivist shouldn’t deny that evaluations are beliefs whose contents are expressible by indicatives; rather, the expressivist point should be that this, however, doesn’t help us to uncover what’s distinctive about the psychological states whose contents are expressible by evaluative indicatives. To get at that – to get at what’s special about evaluative beliefs – we need to attend to the distinctive role of evaluative indicatives as answers to questions of what to desire, choose, decide, intend, do and feel. It’s instructive to consider more closely the general point about indicatives and belief. Suppose I consider whether to believe p: of whether p is to be believed; and suppose I take up this question in an object-focussed way. In that case, I consider the question of whether to believe p in such a way that any reason to believe p recommends believing p by virtue of the way in which it supports p’s being true, and hence is a first-order reason for belief.13 And when I take up the question of whether to believe p in this object-focussed way, it will be indistinguishable from the question of whether p is true: any consideration (or set of considerations) that I accept and count as decisive in favor of p’s truth will equally count for me as decisive in favor of believing p, and the converse.14 Moreover, the question of whether p is true will be the dominant one: the question of whether to believe p will ‘give way’ to the issue of whether p is true, and the answer to this second question will dictate an answer to the first. In this way, the question of whether p is true will come to serve, in such a case, as the discursive transform of the question of whether to believe p. There’s an important difference between this kind of case and the earlier practical examples, where evaluative questions do duty as the discursive transforms of ‘whether to V+’ questions. The considerations just remarked on show that we can deliberatively put ourselves in a position to form beliefs by reasoning about what’s 12 The question of whether p is true is the discursive transform of the (object-focussed) question of whether to believe p. See the discussion that follows. 13 By contrast, if I ask whether to believe p in an attitude-focussed way, any reason for believing p recommends believing p by virtue of the way in which it supports having the belief that (being a believer of) p, and so is a second-order reason for belief, or reason for wanting to believe p. The direction of fit of belief explains why object-focussed reasons for believing p are truth-related considerations. Our beliefs are our representations of how things are. The dictum ‘belief aims at the truth’ signifies that our beliefs are successful as ways of taking things to be just to the extent that the situations that their propositional objects represent obtain. And this locates the logical space of object-focussed reasons for believing p: they are considerations that recommend the belief that p as a successful model of how things are just to the extent that they support p as true. I say more about the mind-to-world direction of fit of belief in the text below. 14 To say that one counts a consideration k that one accepts as decisive in favor of p’s truth, or as decisive in favor of believing p, is to say that one has a disposition in respect of k whose expression would be coming to believe p on the basis of k.

Moral Psychology and Expressivism

13

true. However, in such cases, the conclusions we come to already are beliefs (rather than just dispositions to believe): in deciding that p is true, I form the belief that p. It’s not possible for pathologies to prize apart the conclusion that p from the belief that p. This, of course, doesn’t mean that our beliefs are immune from pathologies, including pathologies of the heart. It just means that those pathologies have to find openings elsewhere to do their work, as in the gap between seeing a proposition as the one with the best truth credentials and actually believing it.15

Internalism without Intrinsic Practicality? I have been setting out an expressivist attempt to get at the practical difference between our beliefs about what’s good, or fit to be willed, or worthy of being felt about in a certain way, and our non-evaluative, worldly beliefs. The story has been that our evaluative talk indicates that the former are intrinsically practical in so far as they are, or encompass, dispositions to desire or will something, or to have certain sentiments. Unfortunately, there’s a problem with this line of argument. Consider the kind of ‘internalism’ espoused by a philosopher like Michael Smith. Smith defends what he calls the ‘practicality requirement,’ according to which: (P) If someone believes that it is desirable to do A in circumstances C then, other things being equal, they will be motivated (desire) to do A in C.16

Smith defends (P) as a truth about a necessary, defeasible connection between valuing and desiring; but the defense ensures that evaluative belief is not construed as intrinsically practical. Put briefly: Smith’s defense of (P) relies on an analysis of the content of evaluative beliefs. The analysis of evaluative content (in responsedependent terms) yields that anyone’s belief that it is desirable to do A in C can be represented as the belief that they would want to do A in C if they were fully rational. Smith means the content of this belief to be read in a certain way: facts about what it’s desirable to do are facts about what our fully rational selves would advise (want) our everyday, situated, actual selves to do.17 The claim then is that anyone with a 15 Perhaps I should add a qualification to this point. Even where we form beliefs, they can fail to wholly ‘take’ in our psychological economies. So, even though pathologies of the heart cannot prize apart a decision that p is true from coming to believe p, they can interfere with the degree to which that belief becomes integrated into our network of attitudes, and so our mental lives. Sometimes we ourselves realize that we don’t fully believe what we are prepared to avow as true, by reflecting (say) on how we behave. See my remarks on Moore’s paradox, Chapter 3, note 17. 16 See The Moral Problem (Oxford, 1994), pp. 7–9, pp. 61ff. 17 Ibid., pp. 151–2. ‘Full rationality’ is taken to consist in having a maximally informed and coherent and unified set of desires. See The Moral Problem, pp. 155ff.; ‘In Defense of The Moral Problem: A Reply to Brink, Copp, and Sayre-McCord,’ Ethics, 108 (1997): 88ff., and ‘A Theory of Freedom and Responsibility,’ in Garrett Cullity and Berys Gaut (eds), Ethics and Practical Reason (Oxford, 1997), pp. 301–7.

14

Values and the Reflective Point of View

belief with that content will either desire to do A in C, or be irrational (by virtue of a failure of coherence).18 In cases where all goes well, what causes the agent to have a desire that matches their evaluative belief is the belief, combined with their ‘tendency to have a coherent psychology’(which characterizes them in so far as they are rational).19 It’s central to the account that the tendency here isn’t a desire, and that evaluative belief isn’t practical in and of itself.20 Now, suppose we are impressed by the claim that evaluative thought, typically and distinctively, is apt to serve as a focus for deliberation about what to do. Someone like Smith seems well placed to accept and explain this, while denying that evaluative beliefs are intrinsically practical. The explanation will be that the content of a belief about what it’s desirable to do (as shown by analysis) secures its unique relevance as an answer to a question of what to do: it presents as a piece of advice from my fully rational self to my situated, actual self, and so, at least in rational cases, will serve as a prompt to desire – and so, to a disposition to (choose to) act. Such an evaluative belief isn’t, in and of itself, a disposition to desire; but, because of its content, it will get me to desire, at least where I’m rational. So, it looks as if Smith can explain the intimate tie between our evaluations and our practical lives, without treating the former as intrinsically practical. Why is Smith anxious to ensure that we can be internalists without treating evaluative beliefs as intrinsically practical? A special quarry here is the proposal that we treat evaluations as besires – that is, as single, unitary kinds of state that are at once belief-like and desire-like. Smith finds this proposal anathema because it violates what he takes to be the (correct) Humean conception of belief and desire, according to which these mental states are ‘distinct existences,’ in the sense that ‘the two can always be pulled apart, at least modally.’21 Moreover, he thinks that there’s an evident flaw in the proposal that our evaluations are besires: it ignores the way in which pathologies can defeat the connection between our (still intact) values and what we are motivated to do. The proposal that our evaluations are besires supports, not (P), but a stronger, false requirement, which interprets the relation between values and desires (or desire-like states) as indefeasible.22 Smith’s concern here is well placed. Any plausible version of internalism needs to allow, as (P) does, for the ways in which pathologies can impact on the motivational hold of our values without destroying their content. Smith discusses a well-known example of John McDowell’s.23 McDowell holds that the virtuous 18 See The Moral Problem, p. 177, and ‘In Defense of The Moral Problem: A Reply to Brink, Copp, and Sayre-McCord’: 99ff. 19 This is how Smith puts it in ‘In Defense of The Moral Problem: A Reply to Brink, Copp, and Sayre-McCord’: 100. 20 See ibid.: 111ff. 21 The Moral Problem, p. 119. 22 Ibid., pp. 119–21. See too ‘Minimalism, Truth-aptitude and Belief,’ Analysis, 54 (1994): 22ff. 23 See ‘Are Moral Requirements Hypothetical Imperatives?,’ Proceedings of the Aristotelian Society, supp. vol. 52 (1978): 13–29.

Moral Psychology and Expressivism

15

person’s conception of someone as shy and sensitive is distinctive. It’s a conception that entails a certain disposition of the possessor’s will – namely, a disposition to make them feel comfortable when in company, to protect them from those who would embarrass them, and so on. Smith objects that this way of individuating the content of the virtuous person’s evaluative conception, when they see someone as shy and sensitive, doesn’t fit with moral phenomenology. He comments: For if the virtuous person’s appreciation of the fact that someone is shy and sensitive really were a besire, it would be impossible to break the belief-like part of that state apart from the desire-like part; it would have to be impossible, for example, for the virtuous person to retain her appreciation of the fact that someone is shy and sensitive and yet not be in a desire-like state to the effect that that person be made to feel comfortable when he is in company. This is in effect to insist that those who were once virtuous who, through some mishap or other, suffer from weakness of the will on some occasion, mustn’t any longer really appreciate the fact that someone is shy and sensitive; that they must have forgotten something that they used to know all too well … But that is surely incredible. … It is a commonplace, a fact of ordinary moral experience, that when agents suffer from weakness of the will they may stare the facts that used to move them square in the face, appreciate them in all their glory, and yet still not be moved by them … That this is so is crucial for an understanding of how horrible it can be for people who are weak. McDowell’s description of the virtuous person is simply inconsistent with this commonplace.24

The point is surely well taken. Any form of internalism that’s true to the phenomenology of valuing needs to allow for cases where we ‘fall from grace,’ and yet our values survive with their content intact, despite distortion of their motivational grip on us. The view that our evaluative commitments are belief-like/desire-like amalgams fails to do this.25 Expressivists differ in their moral psychology from friends of besires, and are usually counted as friends of Hume. They hold that our evaluations are practical reactions to our beliefs about natural features of the world around us, and that input and output here are modally separable.26 Some contemporary expressivists are prepared to say that the distinctively practical output states that are our evaluations are also beliefs; but that just means that our evaluations are mental states that enjoy dual classification as practical states and as beliefs, with the first classification being the most revealing. By these lights, there are different sorts of belief, some on the 24 The Moral Problem, p. 123. 25 Of course, pathologies can destroy values; and a cautious expressivist account of values can distinguish between the two kinds of case. In the one kind of scenario, our value survives the impact of the pathology as (or as a mental state that involves) a disposition to desire: it’s just that the pathology blocks the expression of that disposition. In the second kind of scenario, the pathology robs us of the disposition itself: we aren’t even inclined to desire to do what (as it were) we take to be good. In such a case, the pathology destroys the value: even if we continue to mouth the value, we no longer are entertaining the same proposition as the one we did before the evaluative words ‘went dead’ in our mouths. 26 See Blackburn’s discussion of besires, Ruling Passions, pp. 97–104.

16

Values and the Reflective Point of View

side of input (like beliefs about natural features) and others on the side of output (like evaluative reactions to those features). Now, a cautious expressivist, who is prepared to allow that evaluations are after all a kind of belief, can agree that (P) is true (without any drawing away from talk of evaluative belief). However, by these lights, (P) is true because evaluative beliefs are, or incorporate, dispositions to desire. And, of course, Smith would oppose that reading of (P), though he never actually discusses (P) under an expressivist construal that treats evaluations as a kind of belief that is intrinsically practical.27 For he takes it to be the distinctive claim of expressivists that evaluations are not, strictly speaking, beliefs (and so are not truth-apt), despite our ordinary talk about them as beliefs.28 And it has to be conceded that this is hardly a surprising interpretation of expressivism, given the way expressivists (including Blackburn) historically have represented their own metaethic.29 In any event, Smith wouldn’t welcome the suggestion that our evaluative beliefs are among our intrinsically practical reactions to the world around us, even given caution about allowing for pathologies to disrupt the connection between our values and our desires. For Smith holds that our evaluative beliefs are mental states with truth conditions that are realizable in the natural world. They have a necessary connection with desire, to be sure, because of the meaning of their contents. However, they are the wrong kind of mental state to be among our intrinsically practical reactions to the world around us.30 On the face of it, the idea that our evaluative beliefs are not practical in and of themselves seems odd. We can put the point in terms of respect for the platitudes that surround belief and desire.31 These platitudes saliently include the ‘practicality requirement’ and the platitudes that make up belief-desire psychology. A very natural 27 Smith associates Blackburn with (P), but reads the latter as denying that evaluative commitments are (real) beliefs. See The Moral Problem, pp. 12, 61. Actually, Blackburn seems to vacillate in how he reads the relation between values and desires. On the one hand, he is sensitive to the way in which the connection is defeasible. See Spreading the Word, pp. 187–9, and Ruling Passions, pp. 59–66. On the other hand, he often treats our values as (or as including) desires (in a ‘thin’ sense) or concerns. Our deep values are, he says, our ‘stable concerns.’ See Ruling Passions, pp. 66–8, 123–4, and Think, pp. 282ff. 28 The Moral Problem, p. 12. See also: ‘Why Expressivists about Value Should love Minimalism about Truth,’ Analysis, (1994): 1–12; ‘Minimalism, Truth-aptitude and Belief’: 21–6; and (co-authored with Frank Jackson and Graham Oppy) ‘Minimalism and Truth Aptness,’ Mind, 103 (1994): 287–302. 29 And Blackburn, even now, shifts around. Thus, in some recent contexts, he says that moral thought and talk is truth-apt in a way that’s perfectly genuine. See Ruling Passions, Questions 14–18, pp. 317–19, and ‘Wittgenstein, Wright, Rorty and Minimalism’: 164–8, 174–5. However, contrast his remarks in Think, p. 283. 30 Smith discusses the relation of his account to naturalism in The Moral Problem, pp. 184–7. 31 Smith is a friend of platitudes. He subscribes to the school of thought that holds that when we do conceptual analysis, we have to respect the network of platitudes that surround a concept and to provide an analysis that is faithful to those central platitudes that are robust

Moral Psychology and Expressivism

17

reading of this assembly of platitudes is that they mark out a difference, in relation to practice, between evaluative beliefs and beliefs about means – namely, that whereas the former are intrinsically practical, the latter aren’t, given the way in which (other things being equal) evaluations directly yield matching desires, whereas beliefs about means yield desires to take those means, only given prior ends.32 Needless to say, Smith doesn’t agree. He holds that the asymmetry in the relation between evaluative beliefs and desires, and means–end beliefs and desires, as marked out by the relevant platitudes, is best interpreted as an asymmetry that involves beliefs all of which have truth conditions that are realizable in the natural world.33 It might be thought that the idea that evaluative beliefs are practical in and of themselves can be disarmed by appealing to the so-called contrast in ‘direction of fit’ between beliefs and desires. In the terms of this contrast: beliefs have a mindto-world direction of fit, whereas desires have a world-to-mind direction of fit. Smith gives the metaphor a functional gloss. On this reading, the difference in the functional roles of belief and desire amounts (roughly): … to a difference in the counterfactual dependence of a belief that p and a desire that p on a perception with the content that not p: a belief that p tends to go out of existence in the presence of a perception with the content that not p, whereas a desire that p tends to endure, disposing the subject in that state to bring it about that p.34

It’s true that difference in direction of fit (so construed) would preclude the belief that p (a state that tends to cease to exist, given the perception that not p) from being, or including, the desire that p (a state that tends to endure, given the perception that not p, disposing the subject to make it the case that p). However, difference in direction of fit wouldn’t preclude (say) the belief that p ought to be, from being, or incorporating, the desire that p, if we were to allow that a belief with a complex content like that could be dual in its direction of fit. For then we could say that one and the same state – the belief that p ought to be – has a mind-to-world direction of fit with respect to the content that p ought to be (marking out its status as a belief) and a world-to-mind direction of fit with respect to the embedded content that p (marking out its status as, or as a state that involves, the desire that p).

under scrutiny. See ‘Minimalism and Truth Aptness’: 291–7, and The Moral Problem, Chapter 2. 32 Compare Blackburn’s comments, Ruling Passions, pp. 113–14. 33 Presumably, the central reason Smith contends that our evaluations are beliefs with truth conditions that are naturalistically realizable (conditions having to do with convergence of desire under conditions of full rationality) is that he takes this interpretation to best fit the platitudes that support the idea that moral beliefs aspire to be objective. See his remarks, The Moral Problem, pp. 4ff., 39–40. For Blackburn’s discussion of how moral beliefs, construed as practical reactions to natural features, can aspire to be objective, and related matters, see Ruling Passions, Chapters 7–9. Note the comments, pp. 307–8. 34 The Moral Problem, p. 115.

18

Values and the Reflective Point of View

Now, friends of besires have exploited just such a possibility of duality in direction of fit. Thus, David McNaughton recommends that we think of the awareness of a moral requirement as ‘Janus-like, as having directions of fit facing both ways.’35 And Smith, for his part, has responded to the appeal to dual directions of fit as a move by friends of besires. He lodges a moral psychological objection to their attempt to establish the credentials of a single, unitary kind of state that is at once belief-like and desire-like, as we have already seen. The crediting of such a state to valuers distorts the psychology of pathologies like weakness of will, construing them as defects of belief or judgment, rather than as failures of desire despite continuing intactness of belief or judgment.36 I have already agreed that this moral psychological point against friends of besires is well taken. However, the idea that evaluations are versatile in their direction of fit isn’t tied to the defense of besires. For all that versatility of direction of fit would tell us is that evaluations are mental states with alternative classifications as beliefs and practical states.37 It certainly would be a mistake to say that such dually classifiable states are a kind of state that is, in a way that can’t be disentangled, belief-like and desire-like, because that would impoverish our moral psychological theory. Cases of weakness of will provide one vivid kind of illustration of this, and there are others, like viciousness. Consider Blackburn’s instructive example of the school bullies who victimize another student. Suppose that, by all ordinary tests, the bullies can be reckoned to know that a newcomer is shy and sensitive. We might then suppose that the bullies go on to taunt the newcomer just because of that; that’s where their cruelty lies. However, friends of besires will have to call it differently. They will have to say that, despite the ordinary tests, the bullies don’t see the same situation as the kind students do: that the former lack a unitary belief-like/desire-like state that characterizes the latter. But that just makes what’s wrong with the bullies indetectable – namely, that they are moved to aggression by the recognition of certain features in another (like shyness and sensitivity), whereas the recognition of those very same features moves kind people to be protective.38 Blackburn comments:

35 Moral Vision (Oxford, 1998), p. 109. 36 See The Moral Problem, pp. 118–19, and ‘Minimalism, Truth-aptitude and Belief’: 22ff. 37 Huw Price defends the possibility of there being a state that has both the direction of fit of belief and the direction of fit of desire, in ‘Defending Desire-as-belief’, Mind, 98 (1989): 119–21. Smith numbers Price among the friends of besires. See ‘Why Expressivists about Value Should Love Minimalism about Truth’: 9. However, it’s not clear to me that Price is defending besires. His chief focus seems to be defending the possibility of a state that, since it has the directions of fit of both belief and desire, is (on this way of drawing the boundaries) a state that is classifiable as both a belief and a desire. As I go on to say below, I don’t think we should regard evaluations as states that are at once beliefs and desires anymore than we should regard them as besires. 38 See the discussion, Ruling Passions, pp. 97–104.

Moral Psychology and Expressivism

19

We need to be able to say what is good about the virtuous and bad about the vicious. If our last word were to be that the virtuous see things one way, and the vicious another, we lose the normative space that opens up when we say that the virtuous react to things in the right way, whereas the vicious do not. The boys at the school were not locked into some barely interpretable amalgam of ‘besire’. They saw that someone was shy and sensitive, and were moved to humiliate him. That is what is wrong with them.39

If we suppose that our values are versatile in their direction of fit, this will mark them out as mental states that have alternative characterizations as both beliefs and practical states. On the one hand, as just emphasized, we shouldn’t then suppose that this means that they are besires. We can recognize that our values have the direction of fit of belief and the direction of fit of practical states without departing from what Smith has called an ‘austere psychological theory.’40 On the other hand, we also shouldn’t suppose that dual direction of fit would mark out our values as mental states that can be represented as both beliefs and desires. For we always need to respect the way in which, under the impact of pathologies, we can fail to desire what we value. The cautious expressivist, who wants to exploit the idea that direction of fit provides a way of sorting mental states, and that values are Janus-like in their direction of fit, would say the following: evaluations, such as our judgments about what it would be good or right to do, or about what ought to be done, have a mind-to-world direction of fit with respect to their whole contents, which certifies them as beliefs, and a world-to-mind direction of fit with respect to their embedded contents, which certifies them as dispositions to desire, or as practical states that involve dispositions to desire.41 To be sure, if we thus interpreted the world-to-mind direction of fit of valuing in a way that was sensitive to its vicissitudes, we wouldn’t accept Smith’s bald claim that ‘being in a state with which the world must fit is desiring.’42 On the other hand, we would be able to remain true Humeans (in Smith’s terms) about what happens 39 Ibid., p. 102. 40 In early discussion, Smith called such belief-like/desire-like amalgams ‘quasi-beliefs,’ whose recognition in explaining moral practice required a revision of ‘austere psychological theory.’ See ‘The Humean Theory of Motivation,’ Mind, 96 (1987): 55–8. 41 Actually, later on, on pages 28–30, I shall make the point that difference in direction of fit really marks out (more broadly) the contrast between passive and active propositional thoughts. Practical states (thought of as states that pertain to desire, the will or the emotions) are a kind of active propositional thought. However, so too (for example) are beliefs about which propositions are most credible in light of the evidence: they have a world-to-mind direction of fit with respect to our patterns of belief or credence. (They also have a passive, mind-to-world direction of fit in relation to their whole contents.) 42 The Moral Problem, p. 116. Smith recognizes that this claim is in trouble if there are mental states with more than one direction of fit (pp. 117–18). However, he interprets the potentially troublesome suggestion as the proposal that there are besires, which he rejects. For early comment by him, see ‘The Humean Theory of Motivation’: 55–8, and ‘On Humeans, Anti-Humeans, and Motivation: A Reply to Pettit,’ Mind, 97 (1988): 589–95. The latter is a response to Pettit’s ‘Humeans, Anti-Humeans, and Motivation,’ Mind, 96 (1987): 530–33.

20

Values and the Reflective Point of View

when people act on their values. For the natural construal of cases where we enact our values, according to cautious expressivism, would be that those values find expression in distinct states of desire, which we act on, combined with relevant beliefs about our situation.43 But is it really plausible to suppose that our values have both the direction of fit of belief and the direction of fit of practice? Well, talk of direction of fit involves a metaphor. On standard readings, the metaphor can be given a normative or functional gloss, or both. Consider the notion that our beliefs have a mind-to-world direction of fit. The normative gloss on this idea typically goes (roughly) like this. Our beliefs are the ways we take things to be. As such, there’s an onus on them to match the world: we fault our beliefs as false, as mistakes, to the extent that they fail to represent things as being the way they in fact are. This onus or responsibility on our beliefs to match the world constitutes their mind-to-world direction of fit.44 We can also interpret the claim that our beliefs have a mind-to-world direction of fit in functional terms. I gave Smith’s proposal above: a belief that p is a mental state that tends to cease to exist in the presence of a perception with the content that not p. And variations in the spirit of this theme are on offer.45 Consider next the idea that our desires have a world-to-mind direction of fit. Here too there’s both a normative and functional gloss available. In normative terms, we can say the following. Our desires are representations of how things are to be. As such, they set an onus of match on the world: relative to any desire that p, the way things are is satisfactory to the extent that p, and unsatisfactory to the extent that not p. This onus or responsibility on the world to conform to our desires constitutes their world-to-mind direction of fit. We can also characterize the direction of fit of desires functionally, or at least point in the direction of how that should go. Earlier, I set out Smith’s proposal: a desire that p is a mental state that, in the presence of a perception with the content that not p, tends to endure, disposing the person who has it to bring it about that p. Smith acknowledges that this is only a rough and simplified functional account. Obviously, it best suits desires to bring about change, considered as mental

43 See The Moral Problem, Chapter 4. The crucial point for Smith is that agents are motivated to act in a certain way just in case they have an appropriate desire-belief pair, ‘where belief and desire are, in Hume’s terms, distinct existences’ (Proposition 3, p. 126). See also p. 119. Smith is aware that someone might interpret the claim that a mental state enjoys dual direction of fit not as the claim that it is a besire, but as the claim that it is a belief that is also a desire. He remarks that, in that case, the Humean will insist that agents are motivated to act just in case, inter alia, they have a desire-that-is-not-a-belief. See note 10, pp. 210–11. The cautious expressivist can agree that the desires on which we act, when we enact our values, are desires-that-are-not-beliefs. 44 As I say in note 13 above, the dictum that ‘belief aims at the truth’ signifies that our beliefs are successful as ways of taking things to be just to the extent that they represent situations as being the way they are. 45 These variations typically highlight the way in which beliefs are psychological states that evolve in ways that are sensitive to the perceived weight of evidence, and the like.

Moral Psychology and Expressivism

21

states that dispose us to make the world conform to the way that they would have it be, for as long as we perceive that the (achievable) change hasn’t occurred.46 Let’s now consider values like our beliefs about what it would be good or right to do, or about what we ought to do. Is it plausible to hold that such values have a worldto-mind direction of fit? It’s clear that they have the direction of fit of belief with respect to their whole contents: (in normative terms) we fault any such evaluations as false where they hold that actions which aren’t good or right to do, or aren’t what ought to be done, are; and (functionally speaking) we tend to abandon them as their falsity becomes apparent. But is it also plausible to hold that these evaluations have a world-to-mind direction of fit with respect to their embedded contents? It certainly looks as if we can make good this claim in normative terms. Consider the case of desire again. Our desires about what to do set an onus of match on the world by being representations of how things are to be – of what is to be done or made true; and, relative to any desire to act, what happens in the world is satisfactory to the extent that we do as we desire and unsatisfactory to the extent that we don’t. Now, there’s a clear, though different, way in which values like our beliefs about what it would be good or right to do, or about what ought to be done, also have a direction of fit that makes the world answerable to them. These values set an onus of match on the world in so far as they represent our acting in certain ways as good or right, or as what is owing; and, relative to any such value, what happens in the world is satisfactory if we enact that value, and unsatisfactory if we don’t. Our desires about what to do have a world-to-mind direction of fit in relation to their whole contents because, being representations of what is to be made true, they mandate that we act in certain ways. Our values about what to do have a similar direction of fit in relation to their embedded contents because they represent certain potential candidates for mandation – certain courses of action – as good or right, or as what ought to be done, and so on. In these two different though intimately related ways, both our values about what to do (whose linguistic expressions are indicatives) and our desires about what to do (whose linguistic expressions are imperatives) set an onus on us to conform our deeds to them.47 Can we also provide a functional reading of the world-to-mind direction of fit of our values about what to do with respect to their embedded contents? Surely we can; but we need to do this in a way that respects the difference between values and desires, out of deference to the vicissitudes of valuing. So, we shouldn’t simply say 46 For Smith’s comments on the limitations of his own proposal, see The Moral Problem, note 7, pp. 208–9. He holds that we should tell the more general story in terms of betting behavior. I think that a natural functional account of direction of fit in the case of desires or wishes that things be a certain way, which aren’t desires about action by us, will include reference to reactive dispositions to be pleased or pained, given beliefs about how things are or become. 47 The two ways in which our values and desires about what to do have a world-tomind direction of fit are intimately related because the way in which the former acquire that direction of fit (through their evaluative content) equips them to serve as justifiers for our desires, considered as representations of what is to be made true.

22

Values and the Reflective Point of View

(for example) that a belief on my part that it would be good to do A has the direction of fit of practice with respect to its embedded proposition about my own future action, in so far as it disposes me to do A.48 We should instead say something along the following lines: my belief that it is good to do A has a world-to-mind direction of fit with respect to its embedded content, in so far as it disposes me to do A in a higher-order way, by disposing me to desire to do A. It’s an important result that we can in these ways provide a normative gloss on the idea that our values about what to do have a world-to-mind direction of fit, together with a functional one that is sensitive to the demands of an adequate moral psychology. For this enables us to insist that whatever there is to the idea of direction of fit can be made good in crediting our values about what to do with the direction of fit of practice in relation to their embedded contents, just as well as it can in crediting those values with a mind-to-world direction of fit in relation to their whole contents.49

Expressivism about Values I have said that it looks as if Smith is well placed to explain the distinctive tie between our values and our practical lives, without treating the former as intrinsically practical. However, I don’t think this impression survives scrutiny. Let’s consider the structure of Smith’s strategy. Perhaps the clearest summary of that strategy occurs in the following passages: Armed with the analysis of normative reasons I argue that it is easy to see why agents who believe that they have a normative reason to, say, keep a promise in certain circumstances C are motivated to keep their promise in C and the like: that is, absent practical irrationality …

48 This is the way Smith suggests that a friend of besires might offer a functional gloss on the claim that our values about what to do have a world-to-mind direction of fit. See The Moral Problem, pp. 118–19. 49 I have been defending the claim that our values are dual in their direction of fit, and are thus marked out as both beliefs and practical states. However, I have been working with examples that obviously contain embedded contents, with a focus on values about what to do. Yet the point about duality in direction of fit is meant to be general. So what of evaluations like ‘Hitler is evil’ or ‘St. Francis is good’? How are we to construe the thesis that evaluations like this have the direction of fit of practice in addition to a mind-to-world direction of fit? The answer is that we need to look to revealing paraphrases. For example: Hitler is evil if and only if we ought to abhor him and his works; St. Francis is good if and only if we ought to admire him and his works, and so on. Thus, as these paraphrases indicate, my judgment that Hitler is evil, or that St. Francis is good, sets an onus of match on the world with regard to the reactive stance I take towards the relevant individual and his life: things are aright, relative to the relevant judgment, just where I match my estimation of Hitler as evil with abhorrence, and my estimation of St. Francis as good with admiration.

Moral Psychology and Expressivism

23

Consider those who believe that they would desire that they keep a promise in circumstances C if they had a maximally informed and coherent and unified set of desires, and who desire that they keep a promise in C, and compare them with another group of people who have the belief but lack the desire. It seems to me that the first group plainly has a psychology that, in this respect at any rate, exhibits more in the way of coherence than the latter. … Rationality, in the sense of this sort of coherence, is thus on the side of agents whose desires match their beliefs about the normative reasons they have. Exhibiting this sort of coherence is what practical rationality consists in, and failing to exhibit it is what practical irrationality consists in. If this is right, however, then note that we not only have an explanation of why people who fail to desire what they believe they have a normative reason to do are practically irrational, but that we also have an explanation of the mechanism by which practically rational agents can come to desire to do what they believe they have a normative reason to do. Agents whose psychologies exhibit this sort of coherence – that is, whose psychologies include a tendency towards such coherence – will have desires that match their beliefs about their normative reasons, whereas those whose psychologies do not exhibit this sort of coherence may fail to have such desires. Beliefs about normative reasons, when combined with an agent’s tendency to have a coherent psychology, can thus cause agents to have matching desires.50

So, the picture is this. Where we are rational, our psychologies are characterized by a tendency to match our desires to our beliefs about what it’s desirable to do (or what we have normative reason to do). The contents of these beliefs (as revealed by analysis) are about what we would desire to do in our circumstances if we had a maximally informed and coherent and unified set of desires. And, in cases where we are rational, our evaluative beliefs (characterized as they are by such contents) unite with our tendency to match desires to them to produce the matching desires. It’s important to the story that the tendencies towards coherence that do part of the causal work, in rational cases, are not themselves to be thought of as desires: our evaluative beliefs are assisted, in such cases, to cause matching desires, but they do their causal work without aid from prior desires. It’s a striking feature of this picture that it treats the tendencies towards coherence that contribute to our practical rationality as external to our beliefs about what it’s desirable to do. And Smith seems committed to running a similar line about the tendencies towards coherence that contribute to our rationality where we match a belief that p to a belief that p is (most) credible. In one place, he discusses how we should defend the principle: If an agent believes she has (most) reason to believe p then she rationally should believe p.

And he argues that we should defend it in the same way we should defend a corresponding practical principle, namely: 50 ‘In Defense of The Moral Problem: A Reply to Brink, Copp, and Sayre-McCord’: 99–100.

24

Values and the Reflective Point of View If an agent believes that she has normative reason to do A then she rationally should desire to do A.

In brief: someone who fails to believe p, where they see p as what they have (most) first-order reason to believe – as what is (most) credible – is irrational by their own lights because they fail to match what they believe to what (as platituderespecting analysis reveals) they take it that they would believe if they were fully rational.51 Presumably, Smith holds that where we are rational in such cases, there’s a mechanism at work that is like the one at work where we match our desires to our beliefs about what’s desirable: our belief about what is (most) credible combines with a tendency towards coherence to produce a belief in the proposition we see as (most) credible. Smith, however, doesn’t hold that all tendencies towards psychological coherence are external to our mental states. Consider the issue between him and Philip Pettit about beliefs and habits of inference. Pettit treats habits of inference as external to beliefs, whereas Smith treats them as internal. Smith thinks that treating inferential dispositions as constitutive of beliefs helps with a problem about difference in direction of fit, as a marker of the difference between belief and desire, that seems to be generated by Pettit’s approach. He writes: Philip Pettit distinguishes between two kinds of state which are such that the world must fit with their content: desires and habits of inference … Thus, for example, he argues that the modus ponens habit of inferring is a state that produces the output of believing that q when the input is a belief that p and a belief that if p then q, and, he claims, this is just as much a state with which the world must fit as a desire; for it requires that our psychology be a certain way, and our psychology’s not being that way is not a reason for changing the habit of inference, but rather a reason for changing our psychology. Pettit claims that we therefore need to distinguish desires from habits of inference, and that considerations of direction of fit are inadequate to this task. His positive suggestion about how we might make this distinction is that desires, unlike habits of inference, are ‘belief-channelled’ … I am, however, unconvinced by Pettit’s claim that we need to distinguish desires from habits of inference in the way he suggests. What the example of habits of inference seems to me to show is just how much slack there is in the phrase ‘direction of fit’, and that, for certain purposes, we do better to talk directly in terms of patterns of dispositions, where these get spelled out in functionalist terms. For example, it seems plain enough that, for example, the modus ponens habit of inference must itself be internal to the dispositions of believing that p, believing that if p then q, and the like, not, as Pettit’s way of setting things up seems to suggest, a separate disposition that combines with the dispositions constitutive of these beliefs to produce other beliefs. After all, what would the dispositions constitutive of believing p be if they did not already include dispositions like the disposition to believe q when you believe if p then q, and other similar dispositions? Indeed, it seems plain enough that habits of inference must be thought of in this way, because they are themselves criticizable if they do not fit the world. …

51 The Moral Problem, pp. 178–9. The second, practical principle is given on p. 148.

Moral Psychology and Expressivism

25

When I talk of states having directions of fit, I therefore have in mind whole packages of dispositions constitutive of desiring and believing …52

Now, I agree with Smith that we do best to treat inferential dispositions as constitutive of beliefs and that it’s puzzling to suppose otherwise. And we can give a very good defense of this constitutive approach: it’s the natural way to think of beliefs if we accept the fruitful idea that the identity of a mental state as a belief is fixed by its functional role, and that the functional role definitive of belief includes, in crucial part, the way in which beliefs typically combine with other beliefs within coherent patterns of inference, such as modus ponens inference.53 Let’s examine, however, Smith’s picture of how our minds tend towards coherence. The view seems to be that there are two different ways in which this happens. In some cases, the tendency towards coherence is constitutive of mental states. We are to think of the mental states as rational in their make-up, in the sense that they (in core part) comprise dispositions whose manifestations keep us coherent in our patterns of thought, or in our patterns of thought and action. In the case of belief, these dispositions are dispositions to infer.54 In other instances where we tend towards coherence, however, the tendency is separate from the mental states and combines with them to promote patterns of thought, or patterns of thought and action, that make sense. This is how it is when our beliefs about what’s desirable or credible cause us to have corresponding desires or beliefs: the evaluation combines with an external tendency towards coherence to yield the relevant desire or belief. What is the principle of location here? Take the belief case. The principle of location obviously has to do with the direction of fit of a mental state with respect to its whole content. Various tendencies towards coherence are reckoned to be internal 52 Ibid., pp. 209–10. Smith gives as the Pettit reference The Common Mind (Oxford, 1993), pp. 18–19. Alternatively, see Pettit’s entry on desire in the Routledge Encyclopedia of Philosophy, ed. Edward Craig (10 vols, London and New York, 1998), vol. 3, p. 31. 53 I assume that Smith is sympathetic to this way of defending the position he puts forward. Consider the remarks of fellow (Canberra) commonsense functionalists David Braddon-Mitchell and Frank Jackson on how commonsense functionalism should accept that ‘part of what makes a state belief is that it evolves by and large rationally and so should include this in the functional role definitive of belief’ (Philosophy of Mind and Cognition (Oxford, 1996), p. 149). (See too their surrounding comment, pp. 145–9.) This is an application, in functional terms, of the principle of charity, according to which creatures with beliefs mostly have the patterns of belief that they ought to have. By these lights, it’s a core part of a functional account of belief that beliefs include clusters of inferential dispositions, in relation to both input and output. Thus, the dispositions that go to make up a belief that p will crucially include a disposition to be responsive to changes of mind about p’s truth credentials, a disposition not to credit that not p for as long as one continues to believe p, a disposition to infer that q in case one also comes to believe that if p then q, and doesn’t have a change of mind about either of the premises, and so on. 54 In the case of desires to act, presumably, the core disposition is the disposition to act in ways believed to satisfy the desire. See Smith’s comments on desires as sets of dispositions, The Moral Problem, pp. 113ff.

26

Values and the Reflective Point of View

to beliefs because they can be mapped on to their mind-to-world direction of fit in relation to their whole contents.55 Now, I agree with Smith that we should think of mental states with propositional contents as having a rational constitution that reflects their direction of fit. However, I don’t think there is any good reason to suppose that the coherence that our psychologies manifest when we match desires to values is the expression of a tendency that isn’t part of the rational constitution of those values. For, as we saw in the previous section, we can make good normative and functional sense of the idea that our values about what to do have the direction of fit of practice with respect to their sub-contents; and moreover, we can make good functional sense of this idea in terms of the very tendency towards coherence that Smith treats as external to them when we match our desires to our values. Furthermore, this strategy enables us to credit our values about what to do as intrinsically practical without any distortion to the psychology of valuers of the kind that Smith thinks fatally attends such a move. For we are not then supposing that the direction of fit of practice that characterizes such values in respect of their sub-contents indicates that they are besires, or even that they are, or include, desires. We are instead supposing that that direction of fit marks them out as mental states that include relevant (defeasible) dispositions to desire. The picture that emerges preserves a simple, unified account of how we tend to make sense in our patterns of propositional thought, or in our patterns of such thought and action. Our propositional thoughts are rational in their constitution in ways that reflect their direction(s) of fit. Of special interest are those thoughts with complex propositional contents like our values about what to do. These thoughts are dual in their direction of fit; and accordingly, they are to be regarded as having a complex rational constitution, with differing clusters of dispositions that reflect their dual identity as beliefs and practical states.56 Smith holds that we can have internalism without treating our values about what to do as intrinsically practical. However, he largely ignores the possibility that the tendency that we manifest when we match desires to values might be internal to those values, because he is focussed on the role of the claim that our values have the direction of fit of practice with respect to their sub-contents as part of the defense of treating values as besires. Yet it’s a mistake to ignore this possibility.57 For it turns out that it is in terms of the very tendency that we express when we match desires 55 I take it that Smith holds the same for desires: desires to act are, in core part, dispositions to act in ways believed to satisfy them qua mental states designed to get the world to fit them. 56 To say that a mental state is rational in its constitution is to say that it is (in core part) made up of dispositions whose expressions keep us making sense in our mental lives, or in our mental lives and deeds. 57 Of course, as I remarked earlier, Smith wouldn’t welcome the suggestion that our values incorporate tendencies to match desires to them. For he wants to hold, in the interests of a particular construal of objectivity, that our values are no more practical in and of themselves than are the beliefs about how the world works that serve our desires. See note 33.

Moral Psychology and Expressivism

27

to values that we can make good functional sense of the notion that values have a world-to-mind direction of fit in relation to the contents that they embed – and in a way that avoids the pitfalls of treating them as either besires or as states that are, or involve, desires.58 All of which, in turn, enables us to keep the theory of values within a simple, unified account of how our minds tend towards coherence.59 It’s time to take stock. Expressivists draw a lesson from the way in which propositions about what’s good, right, and so on, provide a focus for practical thought and talk. They hold that it indicates that evaluative thought is intrinsically practical. As I remarked earlier, construing matters in this way squares naturally with the platitudes that surround belief (including evaluative belief) and desire. For these indicate an asymmetry between the way in which our values about what to do and our beliefs about how the world works connect with desire – an asymmetry which is readily read as reflecting how these values are practical in and of themselves. We can now strengthen the case for (cautious) expressivism about values. It’s the theory 58 What if Smith objects that the cautious expressivist still distorts the psychology of valuing because we need to allow for the possibility that someone mightn’t be even in the slightest disposed to desire to do what they value? I think the answer is that, in such a case, we lose our grasp of what it means to count them as genuine valuers. See note 25. It’s interesting that, in his appeal to moral phenomenology in criticizing the way in which friends of besires can’t allow that our values can survive weakness intact (quoted in the main text on page 15), Smith comments that to understand ‘how horrible it can be for people who are weak,’ we need to allow, as ‘a fact of ordinary moral experience,’ that the weak may retain their appreciation of how good or bad something is and yet still not be moved by it (The Moral Problem, p. 123). The point is, though, that this isn’t the picture of someone for whom a value has totally ‘gone dead,’ and we can capture this (in part) by supposing that, even though they don’t desire the good, or aren’t averse to the bad, they retain some inclination to have those reactions. The question then is where to locate the inclination. 59 There’s also the issue of Smith’s analysis of the content of our moral beliefs. I do not intend to discuss this in any detail; however, it seems to me that the analysis is fatally flawed. Smith analyzes the content of our moral beliefs about what it’s desirable (or right) to do in terms of what our fully rational selves would want for our situated, actual selves. And he holds that the platitudes that surround rationality tell us that if we were fully rational, we would be fully informed and have a maximally coherent and unified set of desires. However, while it’s true that our moral beliefs about what it’s desirable to do are beliefs about what we would want for our situated, actual selves if we had ideal psychologies, what’s relevant here is ethical perfection, and it isn’t at all obvious that ethical perfection coincides with, or reduces to, rational perfection, considered as being fully informed and having a maximally coherent and unified set of desires. See Blackburn’s incisive discussion in Ruling Passions, pp. 104ff., 116–19, 261–8. In various places, Blackburn emphasizes that, while coherence is important in ethics, it isn’t everything. What we need, he thinks, are the virtues that keep us ‘both flexible and firm, principled yet sensitive’ in our responses – namely, coherence, along with maturity, imagination, sympathy and culture (pp. 309–10). See too Question 7, p. 313, and Question 15, pp. 317–18. Blackburn presses these themes again in his response (in ‘Replies,’ Philosophy and Phenomenological Research, 65 (2002): 174–6) to Smith’s criticisms in ‘Which Passions Rule?’, Philosophy and Phenomenological Research, 65 (2002): 157–63. The last two papers form part of a symposium on Ruling Passions.

28

Values and the Reflective Point of View

about values that best fits with the economical idea that our minds tend towards coherence because mental states are rational in their constitution in a way that corresponds to their direction(s) of fit.

Valuing: Its Active and Passive Aspects Let’s take the story of values a bit further. It’s a quite general feature of valuing that it is versatile in relation to direction of fit. Consider my belief that p is (most) credible – that is, that p is the proposition I ought to believe. Such a belief enjoys a mindto-world direction of fit with respect to its evaluative content (the proposition that I ought to believe p) and a world-to-mind direction of fit with respect to its embedded content (the proposition that I believe p). We can put this in normative terms: if I hold that p is what I ought to believe (in light of its truth credentials) when it isn’t, then my belief is faultable as false; and if I fail to believe what I think I ought to believe, my failure to believe is faultable, in the light of that value, as unsatisfactory in so far as it’s a failure to believe what’s owing to be believed. And we can make functional sense of the idea that my value about what to believe has both a mind-toworld and a world-to-mind direction of fit: it’s a state that will tend to disappear as its falsity becomes apparent, and it’s a state that will dispose me to believe p.60 We are therefore able (pace Smith) to credit my value about what to believe with a complex rational constitution: it contains differing sets of coherence-preserving dispositions to infer, which reflect the way in which it is at once a state with the direction of fit that it conform to the world (as regards its whole content) and a state with the direction of fit that the world conform to it (as regards its embedded content).61 The dual direction of fit of values about what to believe points to the need for an amendment in the story as developed so far. For it shows that it’s actually not quite correct to use difference in direction of fit as setting the boundary between theory and practice/passion. Rather, difference in direction of fit demarcates something more general. Let’s call this the difference between passive and active propositional thought.62 We can then say, in terms of direction of fit, that our values about what to

60 This way of functionally characterizing the world-to-mind direction of fit of the value respects how we can fail to believe what we think we should. 61 Consider a belief to the effect that p is the proposition with the best truth credentials (for example, a belief that p is the best evidenced proposition). It seems plausible to suppose that any such belief will be characterized by a duality of inferential dispositions: on the one hand, a disposition not to credit some other proposition as best evidenced for as long as one continues to hold it, and the like, and on the other hand, a disposition to conclude that p. We can defend attributing such a duality of dispositions to infer in the case of such a belief by an appeal to its revealing paraphrase as a belief to the effect that p is the proposition that ought to be believed on all the evidence. 62 Robert Noggle uses the terms ‘active’ and ‘passive’ in relation to world-to-mind and mind-to-world direction of fit in ‘The Nature of Motivation (and Why It Matters Less to Ethics

Moral Psychology and Expressivism

29

believe are passive with respect to their evaluative contents, and active with respect to their sub-contents (and so, with respect to our patterns of belief).63 Reflecting on values about what to believe is instructive in another way. So far, in defending the idea that values have an active direction of fit, I have been discussing, in the main, values about what to do, like beliefs about what it would be good or right to do, or about what ought to be done. The claim has been that such values have a world-to-mind direction of fit in relation to the propositions about future action that their whole contents contain. For simplicity’s sake, I have been supposing that the range of values about what to do in question equate (roughly speaking) to desirability judgments: values about what to desire.64 So considered, there’s a complexity that now needs to be registered: such values will have a world-to-mind direction of fit with respect to what we desire, just as our values about what to believe have a worldto-mind direction of fit in relation to what we believe. In which case, our values about what to do, considered as values about what to desire to do, will have an active direction of fit twice over: once in relation to what we do, and once in relation to what we desire. On the present story, we should cast this twice over active direction of fit in functional terms simply as a disposition to desire to act, which characterizes such values as part of their rational constitution. Now, the story about values being sketched here is meant to apply to our values quite generally. So consider again our values about what to do. These values include values about what to will (choose, decide, intend) to do, as well as values about what to desire to do.65 In the case of a judgment that a certain course of action is fit to be Than One Might Think),’ Philosophical Studies, 87 (1997): 99ff. Noggle is there discussing duality in direction of fit. 63 There are some complexities to note here. Consider, first, the dispositions that are inherent in our values about what to believe in virtue of their world-to-mind direction of fit. Suppose I believe that I ought to believe p. Then that belief incorporates the disposition to believe p. Now, suppose that ‘p’ is an evaluative proposition like ‘It would be good to help Freddy.’ Then, by expressivist lights, the illumining interpretation of the disposition inherent in my value about what to believe, in so far as it has the direction of fit of an active mental state, is that it’s a disposition to form a disposition to desire to help Freddy. Consider, next, the inferential dispositions that are inherent in our values about what to believe in so far as they have a mind-to-world direction of fit. Suppose I believe that I ought to believe p. Then that belief incorporates the disposition to believe q, in case I also believe that if I ought to believe p then q, and I don’t have any change of mind. Now, by expressivist lights, what’s special about my belief that I ought to believe p is that it’s a disposition to believe p. So, by expressivist’s lights, the revealing gloss on the relevant inferential disposition is this: it’s a disposition to combine a disposition to believe p with the belief that q, in the absence of any change of mind about the premises. Now, suppose that ‘p’ is an evaluative proposition like ‘It would be good to help Freddy.’ Then, by expressivist lights, the further revealing gloss on the inferential disposition is this: it’s a disposition to combine a disposition to form a desire to help Freddy with the belief that q, given no relevant change of mind. 64 That is to say, object-focussed values about what to desire. 65 Thus, outright evaluations about what to do are equivalent to judgments about what’s fit to be willed. See note 11.

30

Values and the Reflective Point of View

willed, similar considerations as just noted apply: the judgment that some deed is fit to be willed will have an active direction of fit twice over: once in relation to what we do, and once in relation to what we will. And this double active direction of fit will correspond, in functional terms, to a disposition to will (choose, decide, intend), which forms part of the rational make-up of our values about what to will. Finally, we can extend the story of twice-over active direction of fit to values about what to desire or wish more widely, and to values about what to feel. Suppose I think I ought to be pleased that an old friend has arrived. Then we can say that this value about how to feel has the direction of fit of having the world conform to it twice over: once as regards my friend’s arrival (which is a satisfactory way for things to be relative to that value), and once as regards how I feel about it (which will be satisfactory relative to that value just in case I feel pleased). This twice-over active direction of fit will correspond functionally to a disposition to feel pleased at my friend’s arrival, which is part of the rational constitution of my belief that that is how I ought to feel.66

Rational Processes and Sequences of Reasoning Let’s return, for a moment, to our values about what to believe. These values include dispositions that relate to each aspect of their dual direction of fit. It seems plausible to count these dispositions as inferential. What are we to say in the case of our values about what to desire, and the like? On the face of it, they include inferential dispositions that relate to their mind-to-world direction of fit, like the disposition to infer in accordance with modus ponens. But what about the dispositions they include that relate to their direction of fit as active propositional states, like the dispositions to desire, choose, feel a certain way, and so on? It’s not easy to say when a rational process is properly regarded as a sequence of reasoning. Philosophers have shown a tendency to assimilate all rational processes to 66 A full account of values about what to desire, will and feel will credit them with a rational constitution that reflects both their passive and active directions of fit. To the extent that such values have the passive direction of fit of belief, they can be reckoned to incorporate dispositions to infer like the modus ponens habit of inference. Of course, an expressivist will then want to theorize about such dispositions to infer as are internal to our practical and passional values in a way that reflects how they are propositional states with a self-revealing place as states with contents that dictate answers to questions of what to desire, will and feel – which a cautious expressivist reads as showing that they are dispositions to desire, will and feel. Thus, suppose I believe that it’s desirable to help Freddy. Then such a belief will include (say) a disposition to believe that it’s also desirable to help Jim, given a belief that if it’s desirable to help Freddy, it’s also desirable to help Jim, and no revision of belief. By expressivist lights, such a disposition will be most instructively theorized about as a commitment, in the absence of any relevant change of mind, to a combination of mental states whose salient feature is that they are, or include, dispositions to desire. Compare Blackburn’s discussion of crossing ‘Frege’s abyss’ in Ruling Passions, pp. 70ff. For recent, further comment, see his reply to Bob Hale in ‘Replies’: 166–72.

Moral Psychology and Expressivism

31

sequences of reasoning. In the theory of action, for example, there’s a tradition since Aristotle of counting actions done for reasons as conclusions of reasoning about what to do. Now, that seems very odd. Actions typically do not have propositional contents; yet surely, reasoning is a process that occurs among mental states or events with such contents.67 Moreover, the assimilation of rational processes to sequences of reasoning ignores the plurality of ways in which we can make sense of what people think or do. Thus, sometimes we make sense of what people do by fitting their behavior into a teleological pattern, construing it as activity in the pursuit of a goal; while, at other times, we make sense of what people do by treating their behavior as the outcome of a pattern of reasoning, construing it as activity in the pursuit of a (formal or substantive) good. I discuss these importantly different, though often confused, ways of understanding what people do in Chapter 4.68 For the present, let’s consider examples where we form desires on the basis of our belief that what’s desired is good or desirable in some way. It’s clear that there’s a rational sequence of thought in such cases: our psychology stays intact or coherent to the extent we match a belief that something is good in a certain way with a desire for it in that respect. But is such a rational matching of desire to evaluation a sequence of thought that ought to be treated as a piece of reasoning? It doesn’t seem so. For a sequence of thought to count as a sequence of reasoning, it has to be a sequence that can be represented as one wherein we move from the affirmation or endorsement of propositional contents of thought, as premises, to the affirmation or endorsement of propositional contents of thought, as conclusions. This isn’t so where we come to desire p (with respect to some feature) on the basis of a positive evaluation of p (in the relevant respect). In such a case, in so far as we come to desire p, we form a truth-favoring attitude towards the proposition that p (as a whole content); but that attitude can’t be represented as a kind of assent to p.69 For the logical 67 For the following seems correct: reasoning is a process that occurs among mental states or events with propositional contents, and is a process that is made correct by the formal relations that hold among the relevant contents (where ‘formal’ covers more than just ‘logical’). John Broome lists these features among the markers of which rational processes are sequences of reasoning in ‘Normative Practical Reasoning,’ Proceedings of the Aristotelian Society, supp. vol. 75 (2001): 175–93. I note that Davidson is sometimes prepared to count actions done for reasons as conclusions of practical reasoning, though the move seems to rely on the (peculiar) assimilation of actions to intentions (regarded as a kind of value judgment) in cases where agents straightaway act upon a piece of reasoning. See his comments in ‘Intending,’ Essays on Actions and Events, p. 99. 68 See also my discussion in the Appendix. There, in criticizing Davidson’s theory of the divided mind, I exploit the theme that reasoning well is just one kind of rational process, and that there are ways and ways of sensemaking. 69 We can usefully group together attitudes with a world-to-mind direction of fit like desires and intentions as attitudes wherein we favor certain propositions (as whole contents) to be true. These truth-favoring attitudes contrast with belief, considered as an attitude wherein we take a proposition to be true. Broome contrasts belief, considered as the attitude of taking as true, with intention, considered as the attitude of being set to make true. See ‘Normative

32

Values and the Reflective Point of View

space of assent to the proposition that p belongs to the belief that p. Any sentence that formulates the proposition that p will have the syntactic form of an indicative, and that form indicates that bare literal utterances of the sentence typically will be assertions, or the like, and hence expressions of assent in the form of belief, given that belief is the sincerity condition on illocutions of that kind. Philosophers who have tried to cast the formation of desires based on reasons as the last step in a piece of reasoning have typically looked for contents or objects of desire that can be treated as affirmable contents or objects. Thus, Davidson assimilates desires to a kind of value judgment, and treats the evaluative contents of these judgments as contents of desire. Desire, by this reckoning, turns out to be a kind of assent to a proposition after all: it’s assent to an evaluative proposition.70 I criticize Davidson’s treatment of desires as valuings in Chapters 4 and 5. Briefly, the complaint is that his way of dealing with desires makes a moral psychological mistake: it ignores the pathologies of valuing, and it blurs the contrast between purposive and reasoned activity. Another way philosophers have tried to construe desires as mental states wherein we affirm contents has been to credit desires with contents that are expressible in the imperative or optative mood. This, for example, is Hare’s approach. Hare holds that, when I desire p, I affirm or assent to a content of thought expressible in words as ‘Let it be that p.’ In this way, desire compares with belief, which is assent to a content of thought expressible by a sentence (simply) in the indicative mood.71 However, it’s implausible to suppose that, whereas belief is assent to an indicatively expressible content, desire is assent to a content that can be expressed in the imperative (or optative) mood. For our beliefs and desires can share one and the same propositional content (or object). Cases of wishful thinking and sour grapes rely on this: in the Practical Reasoning’: 175–93. We can acknowledge this more specific contrast and hold that to be set to make a proposition true is one way of having a truth-favoring attitude towards it (as a whole content). We can then map talk of truth-taking and truth-favoring attitudes onto (say) our values about what to do. These values are truth-taking attitudes with respect to their whole contents, and truth-favoring attitudes with respect to their sub-contents. To the extent that values about what to do are the latter, they include dispositions to form attitudes (like desires and intentions) that are truth-favoring with respect to their whole contents. 70 Davidson relies on a distinction between the contents and objects of desires: their contents are evaluative propositions, while their objects are the propositions embedded in those evaluative contents. See his ‘Reply to Christopher Peacocke,’ pp. 209–10. However, a distinction along these lines only becomes available, in the case of desiring p, given the assimilation of desires to a kind of evaluation. And then, following standard usage, it’s a distinction between the whole propositional content or object of the evaluation and the propositional content or object that that whole content or object embeds as its topic. 71 See Hare’s extended defense of ‘the imperative theory of the will’ in ‘Wanting: Some Pitfalls,’ Practical Inferences (London, 1971), pp. 44–58. I set out its chief elements in The Possibility of Weakness of Will, pp. 58–9. Hare thinks imperatives are suitable sentences in which to express the contents of desires of which (in Elizabeth Anscombe’s words) ‘the primitive sign is trying to get,’ and that optatives are suitable for putting into words ‘idle wishes.’

Moral Psychology and Expressivism

33

one case, we come to believe the very proposition that we want to be true; while, in the other, we come to want, with respect to the very proposition that we believe, that it be true.72 The difference between our beliefs and desires, in such examples, isn’t a difference in content of thought: it’s a difference between our (functionally discriminable) orientations towards one and the same proposition.73 The view that our desires have contents (or objects) that are expressible by imperatives or optatives rests on a mistake about the function of the relevant subjunctive mood signs. An imperative (or optative) sentence of the form ‘Let it be that p’ (which contains an operation on an embedded indicative sentence) is not itself apt to express a discrete content of thought; rather, it’s a sentence in a form that makes it apt to present a content of thought – the proposition that p – syntactically marked out as the content of an attitude (like a desire) that has a world-to-mind direction of fit with respect to its whole content.74 To be sure, we can (sincerely) say things like ‘p is good in so far as p is F; so, let p be so’; but it’s a mistake to suppose that such a remark expresses the affirmation of an imperatival (or optative) conclusion on the basis of a piece of reasoning. What we have, in such a case, is simply a stretch of discourse that puts the proposition that p forward as the whole content of an attitude with a world-to-mind direction of fit, which is formed on the

72 It’s true that, in English, we can’t say that we believe what (= that which) we want or that we want what (= that which) we believe. That’s because the dependent that-clauses in belief contexts are indicatives, whereas the dependent that-clauses in desire contexts are subjunctives of the will (imperatives or optatives). We shouldn’t be misled by this, though. The subjunctive mood sign that figures in that-clauses in desire contexts is serving as a marker of desire’s direction of fit, and shouldn’t be read into the content or object of that attitude (which is expressed by the indicative from which the subjunctive sub-sentence is formed). I pursue this point in the text that follows. With a bit of help from Velleman (and some artifice), we can summarize the situation in cases of wishful thinking and sour grapes in these terms: as wishful thinkers, we come to be credent towards the proposition we are already desirous towards, whereas, in cases of sour grapes, we come to be desirous towards the proposition that we are already credent towards. See Velleman, ‘The Guise of the Good,’ in The Possibility of Practical Reason (Oxford, 2000), p. 100. 73 I borrow talk of ‘orientation’ here from Alfred Mele. Mele contrasts two theories of propositional attitudes: one (which embraces ‘content constancy’) individuates different kinds of attitude in terms of ‘their differing intrinsic orientation toward content,’ while the other (which embraces ‘orientation constancy’ – namely, assent – across the range of types of attitude) individuates different kinds of attitude by ‘distinct types of representational content.’ See ‘Internalist Moral Cognitivism and Listlessness,’ Ethics, 106 (1996): 740–41. 74 I have myself made the mistake of supposing that attitudes like desire have contents that contrast with contents of belief by being expressible by subjunctives of the will rather than by indicatives. See: The Possibility of Weakness of Will, Chapter 3; and, more recently, ‘Knowing What I’m About To Do Without Evidence,’ International Journal of Philosophical Studies, 6 (1998): 239–40.

34

Values and the Reflective Point of View

basis of a reason. The direction of fit marker isn’t itself to be read into the content of the attitude that ‘let p be so’ expresses.75 We make the question of whether to desire p deliberatively accessible by focussing on the evaluative issue of whether p is somehow good; which is the interrogative transform of an indicative. And the conclusion of our reasoning, in such a case, is the value judgment that p is good in so far as it is F. When we go on to form a matching desire, our sequence of thought makes sense or is kept rational. However, the formation of the desire isn’t itself a further step in our reasoning. Inference is a rational process of thought wherein we move from affirmable contents (as premises) to affirmable contents (as conclusions); and there’s no content available for affirmation in the case of the desire. Rather, the formation of the desire is simply the manifestation of the evaluative conclusion we have reasoned to: that conclusion recommends a topic to desire as a result of our reasoning and, in a rational movement of thought that transcends inference, the value that it contains finds expression in a matching desire. What about our beliefs about what’s regrettable, admirable, and so on? Passional values like this include dispositions to feel that relate to the world-to-mind direction of fit that they enjoy with respect to their embedded contents. Are these dispositions also ones whose manifestations mark rational processes of thought that transcend inference? The detailed answer here depends on your favored theory of emotion. Take the example of regret. Suppose I come to regret p in the belief that it should be regretted. You might hold (as I do) that my regret is a complex psychological state that consists in my wish that not p, given my belief that p. In that case, it’s open for my regret to involve, in part, the drawing of an inference – namely, an inference to the affirmable content that p. However, it would be an error to suppose that there was (at least in rational cases) another inference involved in the formation of my regret at the point where I form the wish that not p. The formation of that wish, in rational cases, will manifest a negative value that I put on p’s being the case. The latter evaluation may well be the conclusion of an inference; but, in that case, the formation of the wish (whose optative context marks out not p as its content) will represent a coherence-preserving transition in thought that goes beyond inference. There are, however, other ways of understanding what an emotion like regret involves. For example, you might hold that my regret that p just is the wish that not p, locating my belief that p as part of its cause. In that case, the right thing to say will be this: even where it’s based on a negative value that I set on p’s being so, my coming to regret p won’t be an inference; rather, it will be a rational movement in thought that transcends inference. The important point, however exactly we construe the make-up of my regret that p, is that, to the extent that its formation includes my coming to wish that not p on the basis of my assessment that it’s a bad thing that p, my coming to regret p is properly reckoned to involve a rational sequence of thought 75 For related, helpful discussion, see Velleman’s remarks, ‘The Guise of the Good,’ in The Possibility of Practical Reason, pp. 105ff. I do not endorse Velleman’s account of direction of fit. See my discussion, Chapter 3, pages 55–7.

Moral Psychology and Expressivism

35

that goes beyond reasoning or inference. What’s crucial here is that, in forming the wish that not p on the basis of a reason, I come to have an attitude towards a content, without there being any content that I therein affirm, assent to or endorse. Even if I lament ‘Would that it were the case that not p!’ I do not thereby express assent to an optative content; rather, I simply put forward the proposition that not p, optatively marked out as the content (or object) of a wish. Consider, finally, cases where we match our beliefs about what’s fit to be willed with choices or intentions to act. Are these rational sequences of thought that instance sequences of reasoning? It seems not. Just as in the case of desires, we can’t represent choices or intentions as a kind of assent to their propositional contents (or objects). Thus, where I intend that p, the attitude that I form towards the (indicatively expressible) proposition that p can’t be represented as a kind of assent towards p, since that would collapse intention (the attitude of being set to make a proposition true) into belief (the attitude of taking a proposition to be true).76 Attempts to represent choices or intentions as a kind of assent to a content go hand in hand with attempts to cast desires in this way. Thus, Davidson treats choices or intentions as also a kind of value judgment and counts the evaluative contents of these judgments as contents of choice or intention; while Hare supposes that choices or intentions also involve assent to an imperativally expressible content. Neither of these approaches is any more satisfactory in the case of choices or intentions than is its analog in the case of desires. Davidson’s proposal neglects the way in which pathologies of the heart can defeat a match between our values and our wills. And a proposal like Hare’s makes the mistake of reading imperativity into the content of our choices or intentions; whereas, in fact, the imperative mood form just provides us with a syntactic facility for putting forward an indicatively expressible proposition, marked out as the content of an act or attitude that has a world-to-mind direction of fit with respect to its whole content, like a choice or intention.77 The true picture instead is this. We make questions of what to do tractable in deliberation by focussing on questions of what’s fit to be chosen or intended; and we conclude our reasoning, in such cases, with judgments about what we should choose or intend to do. When we then proceed to match our wills to our beliefs about what’s fit to be chosen or intended, we engage in a sequence of thought that preserves our coherence, but which is not itself a further step within reasoning. However, there’s a complexity in such cases. We put ourselves in a position to settle questions of what to do by settling questions of what’s fit to be willed. We also typically settle in our own minds questions of what we are going to do – of what we will do – by making up our minds about what to do. One issue here is the 76 See note 69 above. 77 I set out Davidson’s proposal in more detail in Chapter 5. Hare gives his proposal for intentions in ‘Wanting: Some Pitfalls,’ in Practical Inferences, pp. 52ff. To distinguish wants from intentions, Hare relies on the notion of difference in degree of assent to an imperative. I trace Hare’s views in this regard (as set out in The Language of Morals (Oxford, 1952) and Freedom and Reason (Oxford, 1963)) in The Possibility of Weakness of Will, pp. 6–8.

36

Values and the Reflective Point of View

relation between deciding what to do and coming to know what we are going to do. I offer an account of this relation in Chapter 3. Very briefly, I argue that, within our perspectives as agents, the question of what we are going to do confronts us as one to be settled by an answer to the question of what to do; and hence, our choices or intentions are typically matched by perspectivally conditioned beliefs about our own future deeds.78 Let’s suppose this is correct. The story is that, as agents, we typically settle in our minds questions about what we are going to do by settling questions of what to do. This means, in turn, given the rest of the story, that we, as agents, typically make up our minds about what we are going to do by deciding what’s fit to be chosen or intended. What then are we to make of a transition in thought like: ‘I ought (to choose) to do A; so, I’ll do A’? Is this an inference? It looks like a plausible candidate for an inference. In the terms of the present account, there’s a movement in thought from one affirmable content (of belief) to another.79 Moreover, as I develop that account in Chapter 3, the claim is that, in such cases, we match affirmable contents under a (perspectival) constraint of intelligibility. So, the case for counting this movement in thought as a kind of inference is quite strong. By these lights, when we are deliberating about what to do, there typically is a step within inference beyond the belief we form about what’s fit to be willed. This isn’t at the point where we choose or intend to do what we take to be fit to be willed; rather, it’s the step wherein we then go on to form a belief about what we are going to do.80

78 In Chapter 3, pages 55–9, I dispute Velleman’s view that our intentions just are a kind of expectation. 79 It’s been a standing puzzle how to understand utterances like ‘I will (shall) do A,’ as made to express our intentions. The sentences involved look like future indicatives; and yet, in the relevant range of uses, we use them to express choices or intentions. All manner of claims have been made. Hare, for example, theorizes that sentences like ‘I will do A,’ as used to express intentions, are really imperatives (expressing contents of intention). I too have previously claimed that, under use to express intentions, such sentences are not the simple future indicatives they seem. (For a summary of a part of the debate over this issue, and for my own earlier view, see The Possibility of Weakness of Will, pp. 66ff.) I now hold that sentences like ‘I will do A,’ as used to express intentions, are as they seem. They are future indicatives, which formulate the contents of the expectations that we typically form when we make up our minds about what to do. As a matter of contextual pragmatics, we manage to express our choices or intentions by voicing our matching expectations. For further relevant discussion, see Chapter 3, pages 47–55. 80 Broome proposes that sequences of thought like ‘I ought to do A; so, I shall do A’ are (or may be) examples of ‘correct normative reasoning.’ However, he seems untroubled by the indicative garb of sentences of the form ‘I shall do A,’ as used to express intentions. Yet there is a puzzle. The indicative garb of the relevant sentences semantically connects their bare literal utterances with belief, and Broome isn’t running the line that intentions are a kind of belief. Indeed, his account works with a contrast (see note 69) between belief as the attitude of taking as true, and intention as the attitude of being set to make true.

Chapter 3

Self-knowledge, Truth and Value The role of some questions as discursive transforms of others within deliberation reveals the active aspect of evaluative belief. It also enables us to grasp the difference between deliberative and observational self-knowledge and to appreciate the way in which valuing crucially enters into the acquisition of the former. Let me call the deliberative perspective, as applied to questions of self-knowledge, the purely first-person point of view. By self-knowledge I mean knowledge of our own mental states, and also knowledge of what we are going to do. I shall be arguing that we can appeal to our facility for considering certain questions in a way that renders them indistinguishable to make sense of the difference between purely first-person and observational self-knowledge. It turns out that questions of value provide us with a focus when we inquire, in a purely first-person way, into what we desire, will, feel or are going to do, just as issues of truth provide us with a focus when we similarly inquire into what we believe.

The Purely First-person Point of View What does it mean to say that a perspective is purely first-personal? This is meant to mark the feature that the perspective, while first-personal, is not in any way observational. Some first-person self-knowledge is observational or partly observational, and some is not observational at all. A clear case of first-person selfknowledge that is observational would be a case of consciously being in pain at a certain time.1 Conscious pain has a distinctive phenomenology, in the sense that it is not possible for someone consciously to be in pain at a certain time without there being something that it is like to be in pain at that time.2 The distinctive phenomenology which attaches to conscious pain shapes its firstperson epistemology. I am able to self-ascribe conscious pain because of what it is like for me to be in pain. I know that I am in pain on the basis of salient, phenomenological evidence uniquely available to me. Such first-person self-knowledge is evidential, and so, in a broad sense, observational;3 but it contrasts with third-person knowledge 1 I am here concerned with non-intentional pain. 2 For an account of what it is for a conscious mental state to have a distinctive phenomenology, and the relation of this to first-person self-knowledge, see André Gallois, The World Without, The Mind Within (Cambridge, 1996), Chapter 1. 3 In these terms, my first-person knowledge that I am in pain, in such a case, is observational just in that it is evidential. It is left open whether such phenomenological self-

38

Values and the Reflective Point of View

that I am in pain. Third-person knowledge of this kind is also evidential, but the two cases differ in the kinds of evidence that are relevant. When you know that I am in pain, or when I third-personally know that I’m in pain (supposing I can be in pain without feeling it), the evidence is behavioral or neural. Observational first-person self-knowledge is not restricted to non-intentional sensations. Sometimes first-person knowledge of our own present, conscious propositional attitudes is crucially observational. For example, sometimes I selfascribe the desire that p or the fear that p because, in crucial part, of how it feels to have the attitude. Indeed, some desires, like sexual desires, and some fears, like phobic fears, are characteristically self-ascribed, in important part, on the basis of what it’s like to be the subject of those desires or fears. The phenomenological evidence, which is uniquely available to the subject, provides the basis for identifying the attitude as being of a certain kind – as being lust or fear, for example.4 Such firstperson self-knowledge contrasts with third-person cases where you or I know (say) that I desire p, or that I fear p, on the basis of watching what I do. Some first-person self-knowledge is wholly observational, as when I am consciously in pain. Some first-person self-knowledge is crucially observational, as when I self-ascribe a present, conscious propositional attitude because, in crucial part, of what it feels like to have it. Of special interest, in the present context, is a third kind of case of first-person self-knowledge: first-person knowledge of one’s own current, conscious propositional attitudes, which is wholly non-observational. This is the kind of first-person self-knowledge that I call purely first-personal. It is the kind of self-knowledge that I typically have of my own current, conscious beliefs and intentions and that I sometimes have of my own current, conscious desires and emotions. It is also, I contend, the kind of knowledge I have, as agent, of my own future action.5

Purely First-person Knowledge of One’s Own Beliefs Consider my purely first-person knowledge that I believe p. How is such nonobservational self-knowledge possible? Gareth Evans goes to the heart of the matter in the following comments:

knowledge (which may be said to be based on introspective evidence) is perceptual or quasiperceptual. Compare ibid., pp. 3, 21. 4 There is a problem with supposing that self-ascriptions of our own current, propositional attitudes can be wholly phenomenological. The problem is how to explain our knowledge of the content of our attitude in such cases. 5 For an early development of the view that there is a generalizable story about nonobservational self-knowledge that applies both to current, conscious propositional attitudes and to future practical self-knowledge, see Stuart Hampshire’s discussion, Freedom of the Individual (London, 1975).

Self-knowledge, Truth and Value

39

… in making a self-ascription of belief, one’s eyes are, so to speak, or occasionally literally, directed outward – upon the world. If someone asks me ‘Do you think there is going to be a third world war?’, I must attend, in answering him, to precisely the same outward phenomena as I would attend to if I were answering the question ‘Will there be a third world war?’. I get myself into a position to answer the question of whether I believe that p by putting into operation whatever procedure I have for answering the question whether p.6

The central, constitutive feature of knowing that I believe p, in a purely first-person way, is that, from the perspective in question, I can’t distinguish or separate the issue of whether I believe p from the issue of whether p is true. You ask me whether I think that there’s a snake on the track in the front of us and I take a closer look at the track. In case I thereby come to know that I do think there’s a snake in front of us, I selfascribe that belief in the light of my affirmative answer to the question of whether there’s a snake ahead. What I don’t do is make up my mind about what I think by taking a ‘look within.’7 The epistemic situation can be characterized like this. Consider the question ‘Do I believe p?’ (Q1). Suppose I take up this question in a purely first-person way. What’s distinctive about taking up Q1 in this way is that Q1 will thereby be indistinguishable for me from the question ‘Is p true?’ (Q2). That is: any consideration (or set of considerations) that I accept and count as settling the answer to Q2 in the affirmative will equally count for me as settling the answer to Q1 in the affirmative (as reason to self-ascribe the belief that p), and the converse.8 Moreover, the autobiographical question Q1 will be ‘transparent’ with respect to Q2: Q1 will ‘give way’ to Q2 as the issue in focus and the answer to Q2 will dictate an answer to Q1. In this way, Q2 will come to serve, in such a case, as the deliberative transform of Q1. Purely first-person self-ascriptions of belief reflect this perspectival displacement: when we self-ascribe a belief that p in a purely first-person way, we also, and dominantly, affirm that p.9 Indeed, it’s the defining feature of purely first-person knowledge of our own beliefs that the self-reports that express its content are also, and saliently, avowals of the contents of the reported beliefs. In similar vein, purely 6 The Varieties of Reference (Oxford, 1982), p. 225. 7 Compare Hampshire, Freedom of the Individual, Chapters 3–4, and Roy Edgley, Reason in Theory and Practice (London, 1969), pp. 89ff. See too Gallois, The World Without, The Mind Within, Chapter 3. 8 To say that Q1 and Q2 are indistinguishable questions within a purely first-person perspective isn’t meant to suggest that, in so far as we take up those questions purely firstpersonally, we must cease to recognize that they relate to different matters of fact, namely, one’s believing p and the fact that p itself. The point, rather, is that anyone who takes up Q1 in a purely first-person way considers the question of what they believe from a point of view that won’t allow them to differentiate between having a reason that settles Q2 in the affirmative and having a reason that settles Q1 in the affirmative. 9 By some lights, purely first-person claims of the form ‘I believe p’ aren’t selfascriptions at all. They are just (tentative) claims to the effect that p. I argue against this view on pages 53–4 below.

40

Values and the Reflective Point of View

first-person explanations of belief mark the way in which, within the relevant point of view, questions about what we believe transmute into questions about what’s true. Such explanations are internal to object-focussed, truth-related justifications – like ‘(I believe) p because (I believe) e, which (as I believe) makes p most likely.’10 Indeed, the way in which we must regard our beliefs as responses to the ‘evidences of truth,’ from within a purely first-person point of view, sets a constraint on how we can sustain that perspective in cases of openly self-deceptive belief. For selfdeception does not always hide itself and, where it doesn’t, the fruits of deception, no doubt precariously enjoyed, will require just such a feat. Consider cases of openly wishful self-deception, where someone ends up believing what they want to be true against the better evidence available to them. In examples like this, the person will be aware that their thinking, and the way they treat the evidence, is being skewed by desire. If they are to sustain a purely first-person perspective on the question of what they believe, in such cases, and so continue, in the main, at least for a time, to be reassured by what they say they believe, there will have to be division within their field of attention, such that their recognition of the bias behind the belief that they end up at once self-ascribing and avowing is excluded from their focus of attention.11 What explains the feature that, within a purely first-person point of view, Q1 (the question of whether I believe p) will be indistinguishable from, and indeed ‘transparent’ to, Q2 (the question of whether p is true)? We can think of this phenomenon in terms of a two-stage process. When I take up Q1 in a purely first-person way, I am asking a question about what I believe in a way that is focussed on the proposition that p as a candidate for belief. The contrast here is with contexts where I ask the question of what I believe in a way that is focussed on the proposition that I believe p as 10 I am not supposing that we must always be able to cite evidence or truth-related support for what we believe when we ascribe our beliefs in a purely first-person way. To be sure, central and straightforward examples of such ascriptions are cases where, asked whether we believe p, we make up our minds about whether p on the basis of relevant truth-related considerations, and ascribe a belief that p in the light of deciding that p. However, someone may be able to immediately ascribe a belief that p in a purely first-person way, without considering or having in mind any truth-related support for p. Thus, someone (relying on memory) may non-observationally self-ascribe a belief on a topic which their minds have been made up about for some time, without calling to mind any supporting evidence for their answer to the question of what’s true – say, in responding to a survey about what people believe about the topic. (I borrow this example from Gallois, The World Without, The Mind Within, pp. 114–15.) It remains true, in such cases, that we self-ascribe beliefs from a perspective where the question of what we believe is epistemically indistinguishable from the question of what’s true. The self-ascriptions we make carry a commitment to the truth of the propositions we say we believe, and this indicates that the point of view from which we ascribe these beliefs to ourselves is such that anything that we did accept and take to settle the truth issue would equally settle the belief question for us, and the converse. It’s just that we don’t actually have in mind any case for our answer to either question. Examples like this show that we can ascribe already held beliefs to ourselves in a purely first-person way. 11 For related discussion, see my paper ‘Motivated Irrationality and Divided Attention,’ Australasian Journal of Philosophy, 73 (1995): 325–36.

Self-knowledge, Truth and Value

41

a candidate for empirical discovery, including introspective cases, where I attend to ‘inner evidence’ that is uniquely available to me.12 Because the focus is on the proposition that p as a candidate for belief when I ask Q1 in a purely first-person way, Q1 will be indistinguishable, in such a case, from the question of whether to believe p (Q3), considered in an object-focussed way. And (object-focussed) Q3, in turn, will be indistinguishable from, and indeed, take as its discursive transform, the question ‘Is p true?’ (Q2). For I have the concept of belief and I know that ‘belief aims at the truth,’ and so I know that, in the case at hand, there is an identity between considerations that support p’s truth and reasons to believe p.13 We can summarize all this in the following way. When I ask the question of what I believe in an observational, empirical way – even in introspective cases – I ask it in a way that brackets the issue of what’s to be believed, and so, what’s true. Self-ascriptions of belief, from this perspective, will be just that. However, when I ask the question of what I believe in a purely first-person way, I ask it in a way that closes the gap between what I believe and what’s to be believed, and so, between what my belief is and what’s true. Self-ascriptions of belief, from this perspective, will involve a commitment to the truth of the beliefs ascribed: they will go beyond mere reports on those beliefs and express our ‘actual point of view on the world.’14 One way to characterize the feature that questions of what we believe yield, within a purely first-person perspective, to questions about the world that is the topic of those beliefs is to say that, in such cases, we relate to facts about our own mental lives or states in a ‘transcendental’ way.15

12 I am not supposing that we never attribute beliefs to ourselves because, in crucial part, of what it is like to have them. Beliefs can have a phenomenology, as in the case of felt suspicions. 13 See note 13, Chapter 2. 14 Richard Moran’s phrase: ‘Self-Knowledge: Discovery, Resolution, and Undoing,’ European Journal of Philosophy, 5 (1997): 156. See also his discussion, Authority and Estrangement (Princeton, NJ, 2001), Chapters 2 and 3. Suppose I consider whether I believe p by considering whether other beliefs I hold entail that belief or make it likely – that is to say, by considering my belief set. Am I, in such a case, considering whether I believe p in a purely first-person way; that is to say, in a way that is world-directed? Yes, I am; and just because I am pervasively directing my attention to my beliefs in a way such that what’s true of the world presents to me as given by what I believe about it – in a way such that there is no distance for me between the facts and the facts as I take them to be. In other words, I am considering whether (I believe) p by considering p’s truth credentials, as given to me by other beliefs I continue to hold – by considering whether, for example, p is entailed by the fact (as I believe it to be) that q, or made likely by the fact (as I believe it to be) that r. For related discussion, see Moran’s ‘Interpretation Theory and the First Person,’ The Philosophical Quarterly, 44 (1994): 165–8. 15 Moran’s phrase again: Authority and Estrangement, p. 89. Moran connects the contrast between empirical and transcendental relations to the self with Sartre’s view of persons as divided in their natures between the self-as-facticity and the self-as-transcendence. See his remarks in ibid., pp. 77–83.

42

Values and the Reflective Point of View

Even conceding that we have uniquely available to us a non-observational procedure for answering autobiographical questions like Q1, how can such a procedure yield self-knowledge, or at least a justified belief about oneself? Provenance of belief is one thing, justification another. The issue here is perspectival justification; for purely first-person self-knowledge, if there be such, is knowledge acquired from within a point of view. I suggest that the relevant perspectival justification works like this. When I consider whether I believe p (Q1), in a purely first-person way, that question takes the question of whether p is true (Q2) as its deliberative transform; and I am under a Moore-paradox constraint to match an affirmative answer to Q2 with an affirmative answer to Q1. I am under such a perspectival constraint on pain of otherwise affirming something that I could not make sense of, could not be in a position to affirm, from within the perspective. For, within a purely first-person perspective, I simply could not have reason to credit conjunctive claims like ‘p, but I don’t believe p’ or ‘p, but I believe not p.’16 So, when I take up Q1 in a purely first-person way, and match an affirmative answer to Q2 with an affirmative answer to Q1, I conform to the Moore-paradox constraint that holds for the perspective, a constraint that I may be reckoned to appreciate (at least implicitly) in so far as I consider Q1 in such a way that it takes Q2 as its deliberative transform.17 And this means that, when I purely first-personally 16 Of course, within a purely first-person perspective, I am also under a Moore-paradox constraint to match an affirmative answer to Q1 with an affirmative answer to Q2; and I could not have reason to affirm conjunctive thoughts or remarks like ‘I believe p, but not p.’ 17 A tenacious presumption of paradox attaches to thoughts or comments like ‘p, but I don’t believe p’ or ‘p, but I believe not p.’ Why is this? It might be thought that the presumption of paradox is itself odd. After all, in each case, both conjuncts could be true and, in each case, the conjunction is something there could be reason to believe on all the available evidence. On some accounts, the paradox is pragmatic. Remarks like ‘p, but I don’t believe p’ or ‘p, but I believe not p,’ it is claimed, are self-defeating: in the second conjunct we represent ourselves as not believing, or as disbelieving, what in the first conjunct we assert, and so put ourselves forward as believing. The problem with pragmatic accounts of the paradox is that the paradox is not confined to speech: it also characterizes unspoken thoughts like ‘p, but I don’t believe p’ or ‘p, but I believe not p.’ Gallois notes this in The World Without, The Mind Within, p. 46, as does Moran in Authority and Estrangement, p. 70. I do not here propose to offer an account of why such thoughts or remarks are presumptively paradoxical. What fundamentally needs to be explained is why there is a very strong presumption, in such cases, that the thinker or (sincere) asserter of ‘p’ is already adopting a point of view from within which the question of what they believe will be settled for them by their answer to the question of what’s true. For it is on this interpretative presumption that thoughts or remarks like ‘p, but I don’t believe p’ or ‘p, but I believe not p’ are paradoxical. In the first case, instead of matching an ascription of a belief that p to an avowal of the proposition that p, the thinker or sayer ascribes a lack of belief in the proposition avowed – something they could have no reason to do on the presumption in place. In the second case, instead of matching an ascription of a belief that p to an affirmation of p, the thinker or sayer ascribes disbelief in the proposition affirmed. This is doubly puzzling on the presumption in

Self-knowledge, Truth and Value

43

self-ascribe a belief that p in such cases, I ascribe to myself the belief that I ought to self-ascribe, from within that perspective. For, internally to a purely first-person perspective, the belief I ought to attribute to myself is the belief that p, if that’s the self-ascription whereby I conform to the Moore-paradox constraint in my pattern of answers to Q2 and Q1, given an affirmative answer to Q2.

Purely First-person Knowledge of Other Mental States The way in which I can non-observationally know what I believe has a generalizable pattern. For example, sometimes I know non-observationally what I desire or what I intend or what emotion I have, and such purely first-person self-knowledge can be explained in terms of the way in which certain questions are perspectivally indistinguishable. Non-observational Knowledge of Our Own Desires Current, conscious desires have a plural or versatile first-person epistemology. Often they are felt, and their self-ascription is crucially phenomenological. Indeed, there are some kinds of desire, like those associated with hunger, thirst or lust, that we typically attribute to ourselves, in crucial part, because of what it’s like to feel them. place. The thinker or sayer affirms a conjunction that they couldn’t have any reason to affirm. And moreover, assuming that the perspective of the thinker or sayer remains constant across the conjunctive thought or remark, the conjunction they affirm involves a perspective-relative flat self-contradiction, since it effectively conjoins an avowal of p with an avowal of not p. Given enough contextual detail, the paradox associated with thoughts or remarks like ‘p, but I don’t believe p’ or ‘p, but I believe not p’ can be more or less attenuated, as where it becomes obvious that there is a shift in perspective across the thought or remark and that the thinker or sayer is explicitly considering the question of what they believe in an observational way. For, in that case, someone could well have reason to credit the relevant conjunctions, and could think or say something of the form ‘p, but I believe not p’ without flat self-contradiction. For example, I think or say: ‘My hands are clean, but I (or, a part of me) can’t help believing that they aren’t.’ Here I affirm that p (in the light of what I take to be my evidence about the state of my hands) and, switching perspective, self-ascribe a felt phobic belief that not p (in part because of what it feels like to be beset by that belief). In such a case, I openly admit to being a ‘doublethinker’ about the cleanliness of my hands. Consider also cases where the epistemology of the self-ascription of belief is third-person. I think or say ‘He’s innocent, though I don’t as yet quite believe he is,’ avowing the proposition that I take the evidence of innocence to support, while self-ascribing a lack of full belief in the fellow’s innocence, drawing (say) on behavioral evidence indicating that belief in what I avow hasn’t fully ‘taken.’ Or I think or say (without flat self-contradiction) ‘He’s innocent, though I (or, a part of me) still harbor(s) a belief that he isn’t,’ affirming that the fellow is innocent on the basis of my evidence, while (say) self-ascribing a conflicting belief by way of making sense of how I continue to behave towards the poor fellow when I have contact with him. In such a case, I again openly admit to being a ‘doublethinker.’

44

Values and the Reflective Point of View

However, often our desires are not felt and we know that we have them in a purely first-person way. How is such non-observational self-knowledge possible? As in the case of belief, what is central is that certain questions can’t be perspectivally distinguished or separated. You ask me whether I want to leave early. I consider what’s to be said for an early departure and decide that I do want to leave early, having judged that it would be a good move. In such a case, I take up the issue of what I want to do in such a way that it can’t be teased apart from question of what it would be good to do, and I answer the autobiographical question by matching my response to it to my answer to the evaluative one. What I don’t do is make up my mind about whether I want to leave early with the help of an ‘inward glance.’ We can represent the epistemic situation in examples of self-knowledge like this in the following way. In cases where I know, in a purely first-person way, that I desire p, I do so by virtue of the way in which, within that perspective, I can’t distinguish, or separate, the question of whether I desire p (Q4) from the question of whether p is good (in some respect) (Q5). From the relevant perspective, any consideration (or set of considerations) that I accept and count as settling the issue of whether p is good (in some way) will equally count for me as reason to self-ascribe the desire that p, and the converse. Moreover, as in the case of belief, there will be an order of dominance here: Q4 will be ‘transparent’ to Q5, taking Q5 as its discursive transform.18 What explains the feature that I can’t distinguish Q4 (a question about what I desire) from Q5 (a question about what is good), from within a purely first-person point of view? The situation is analogous to the belief case. When I consider Q4 in a purely first-person way, I am asking a question about what I desire in a way that is centered on the proposition that p as a candidate for desire. By contrast, where I consider Q4 in an observational way (including cases where I rely on ‘inner evidence’), I am asking a question about what I desire in a way that is centered 18 Again, the point isn’t that anyone who considers Q4 in a purely first-person way must somehow lose their grasp of the fact that there are two different facts at issue: one that has to do with their state of mind, and one that has to do with the value attaching to a state of affairs. Rather, the point is that, within a purely first-person perspective, the epistemologies of these two facts coincide: we can’t differentiate between being having a reason that establishes that p is good (in a certain respect) and having a reason that establishes that we want p (in that respect). It’s important to emphasize that, on the present story, the first-person epistemology of current, conscious desire is plural. Sometimes (and, in the case of some kinds of desire, typically) we take up a question of whether we desire p in a way that takes account of ‘inner evidence,’ and sometimes we consider the question of whether we desire p in a way that focusses on whether it would be (somehow) good if p. It’s only in cases of the latter kind, where we consider the issue of what we desire in a purely first-person way, that the question of whether we desire p (in a certain respect) will be settled in the affirmative by a judgment that p is good (in that respect). Otherwise – by taking account of how we feel – we may well be in a first-person position to think or say (for example) that we don’t desire p, even though p is good. See my remarks below, at note 20.

Self-knowledge, Truth and Value

45

on the proposition that I desire p as a candidate for empirical detection. Because the focus is on p as a candidate for desire, in cases where I ask Q4 in a purely first-person way, Q4 will be indistinguishable, in such examples, from the (objectfocussed) issue of whether to desire p (Q6). And (object-focussed) Q6, in turn, will be indistinguishable from, and indeed, take as its discursive transform, the issue of whether p is (somehow) good (Q5). For I have the concept of desire and I know that ‘desire aims at the good,’ and so I know that, in the present case, there is an identity between considerations that support p’s being good (in some way) and reasons to desire p.19 We can summarize the general situation in these terms. When I take up the question of what I desire as a self-spectator, including cases where I rely on phenomenological evidence, I raise that question in a way that is closed off from the issue of what is to be desired, and so, what’s good. And correspondingly, when I self-ascribe a desire, based on phenomenological or behavioral evidence, I will be simply ascribing a desire to myself. However, in sharp contrast, when I consider the question of what I desire in a purely first-person way, I raise that question in a way such that there isn’t any distance for me between the issues of what I desire and what is to be desired, and so, between what my desire is and what’s good. And correspondingly, when I selfascribe a desire in a purely first-person way, the ascription will go beyond a mere report on the desire and involve a commitment to the goodness of what’s desired. Thus, we can also say that, when I ask whether I desire p in a purely first-person way, I am relating to my own mental life in a ‘transcendental’ way. A question about one of my mental states transforms into a question about a candidate object for that state, and any ascription of the state goes beyond mere self-description and involves a commitment about its object. What kind of commitment is involved depends on the kind of state being ascribed. In the case of a belief that I self-ascribe in a purely first-person way, the further commitment is to the truth of the proposition believed. In the case of a desire that I similarly self-ascribe, the further commitment is to its being (somehow) good that the proposition desired be true. So, it seems that we have uniquely available to us a non-observational procedure for answering questions about whether we desire p. But how can such a procedure yield knowledge of our own desires; or, at least, justified beliefs about them? The epistemic situation is like that for purely-first person knowledge of our own beliefs. Such self-knowledge is perspectival. When I take up the question of whether I desire p (Q4) in a purely first-person way, that issue ‘gives way’ to the issue of whether p is (somehow) good (Q5); and I am subject to a conative analog of the Moore-paradox constraint that governs the perspective. Internally to a purely first19 When I say that, as someone who has a command of the concept of desire, I know that ‘desire aims at the good,’ I mean I know that any desire that p is a success (in some measure) as a dictate on the world just in so far as it dictates that the world be a way that’s (somehow) good, and that, along with knowing this, I know that object-focussed reasons to desire p will be considerations that support it’s being true that it would be good if p. See note 8, Chapter 2.

46

Values and the Reflective Point of View

person point of view, I simply could not have reason to credit conjunctive claims like ‘p is good in so far as it is F, but I don’t desire p in that regard’ or ‘p is good to the extent that it is F, but I desire not p in that respect.’ So, when I consider Q4 in a purely first-person way, and match an affirmative answer to Q5 with an affirmative answer to Q4, I abide by a constraint that holds for the perspective, a constraint that I may be held to appreciate (at least implicitly) inasmuch as I consider Q4 in such a way that it takes Q5 as its discursive transform.20 And this means that, in making the self-ascriptions of desire that I do in such cases, I do what I ought to do, internally to a purely first-person point of view. For, within that point of view, what I ought to do is attribute the desire that p to myself, if that’s the self-ascription whereby I conform to the conative analog of the Moore-paradox constraint in my pattern of responses to Q5 and Q4, given an affirmative answer to Q5. Non-observational Knowledge of Our Own Emotions Sometimes the first-person epistemology of emotions is also non-observational.21 Here, what is central (at least in terms of the account of the emotions that I favor) is the way in which emotions contain other propositional attitudes (like beliefs and desires) as constitutive elements. For we come to know that we have an emotion in a purely first-person way by virtue of having the concept of that emotion and of having purely first-person access to the propositional attitudes that comprise it. Consider an example like regret. Suppose I regret that I lashed out in anger. In that case, some combination of propositional attitudes like the following characterizes 20 I say that a conative analog of the Moore-paradox constraint holds within a purely first-person point of view. It’s noteworthy, however, that while this is true, thoughts or remarks like ‘p is good, but I don’t desire p’ or ‘p is good, but I desire not p’ are not presumptively paradoxical in the way that thoughts or remarks like ‘p, but I don’t believe p’ or ‘p, but I believe not p’ are. Why is there this difference? The answer (which needs elucidation) must be that, whereas there is a standing presumption in the belief examples that the thinker or sayer is adopting a purely first-person perspective on what they believe, there isn’t a similar presumption in conative examples with respect to their perspective on what they desire. For paradox to attach to thoughts or remarks like ‘p is good, but I don’t desire p’ or ‘p is good, but I desire not p,’ we need some contextual indication that a thinker or sayer is considering what they desire in a purely first-person or deliberative way. Contrast thoughts or comments like ‘p is good, but I don’t think it is’ or ‘p is good, but I think that it isn’t.’ Such thoughts or comments are presumptively paradoxical. There is a standing presumption that thinkers or sayers who affirm that p is good are already taking a purely first-person perspective on whether they value p. 21 No doubt the availability of purely first-person knowledge of our own emotions is more common in the case of some emotions rather than others. For example, perhaps fears are more commonly phenomenologically self-attributed (at least in part) than regrets. In any event, some first-person emotional self-knowledge is non-observational, and the issue is how we can come to know what emotion we have in such cases.

Self-knowledge, Truth and Value

47

me: (i) I believe that I lashed out in anger and (ii) I wish that I hadn’t. How can I non-observationally know that I regret that I lashed out in anger? I can know this simply by virtue of having the concept of regret and of non-observationally knowing the autobiographical facts (i) and (ii). The epistemic procedure that is uniquely available to me, in such a case, works like this. I have the concept of regret. So, when I ask myself whether I regret that I lashed out in anger, that question will resolve into two issues: whether I believe it’s really correct that I lashed out in anger, and whether I wish I hadn’t. My epistemic perspective is purely first-personal. So, I take up each of these issues in a purely first-person way. This means that my twin (object-focussed) concerns will become whether to believe that I lashed out in anger, and whether to wish that I hadn’t. Suppose I reckon, in turn, that I did lash out in anger, and that it was a bad thing. I will then be in a position (and indeed, under a perspectival constraint) to self-ascribe purely first-personally both the belief that I lashed out in anger and the wish that I hadn’t. And hence, given that I have mastery of the concept of regret, I will be placed (indeed constrained) to self-ascribe regret for having lashed out in anger, non-observationally.

Purely First-person Knowledge within Agency The kind of knowledge that we, as agents, have of our own future actions is of a piece with our non-observational knowledge of our own current, conscious thoughts. Suppose I consider what I will do next and decide that I will sample a particular shiraz since it is by far the best drop. In such a case, I apparently come to a view about what I’m about to do in the light of an hedonic consideration. Isn’t this just incoherent?22 It isn’t, if we have regard for the fact that, in such a case, I make up my mind about what I’m about to do from a special epistemic perspective – namely, my purely first-person point of view as agent. This point of view contrasts with my epistemic perspective when I observationally predict what I will do next. It’s constitutive of my observational perspective as selfpredictor that my reasons for predicting will consist in evidence that bears on my future behavior. Thus, I might well anticipate what I will do next on the basis of my knowledge of how I tend to respond, given pressures or enticements of the kind before me. By contrast, it’s constitutive of my non-observational perspective as agent that I won’t be able to distinguish or separate the issue of what I will do next from the issue of what to do next, and that the former informative issue will ‘give way’ to the latter directive one. In this way, any consideration (or set of considerations) which I accept and count as decisive in favor of my acting next in a certain way will perspectivally double for me as reason to conclude that that is what I’m about to do. In the example above: the consideration that I accept and see as settling the issue of 22 Compare H.P. Grice, ‘Intention and Uncertainty,’ Proceedings of the British Academy, 57 (1971): 266–8.

48

Values and the Reflective Point of View

what to do next in favor of my tasting the shiraz – namely, that it’s by far the most enjoyable drop – will perspectivally double for me as reason to self-ascribe that action as my next move. In cases where I non-observationally come to know what I’m about to do, I do so in virtue of conforming to a perspectival constraint to match a decision about what to do next with an expectation about what I will do next that has the same content. Thus, in the working example, I non-observationally come to know (or have a justified belief) that I’m about to taste the shiraz because, in making that self-ascription, I conform to a perspectival constraint to match a decision to taste the shiraz next with an affirmative answer to the question of whether that is what I’m about to do. This constraint holds for the perspective because I simply could not make any sense, in terms of any available reasons within it, of a failure to match such a decision with a corresponding expectation of what I’m about to do. We can track the story here a bit further. We saw in Chapter 2 that the issue of what to do becomes deliberatively tractable because it is deliberatively indistinguishable from the (object-focussed) issue of what to choose, decide or intend to do, which issue, in turn, takes as its discursive transform the question of what is fit to be chosen, decided on or intended. So, for anyone as agent, the question of what they are going to do will become answerable, at base, by converting into the question of which option is fit to be chosen, decided on or intended. In the working example: the issue of whether I’m about to try the shiraz will become answerable for me, in the end, by my answering the question of whether that’s the option that’s fit to be willed. Moreover, we can see how, from within a purely first-person perspective as agent, no one will be able to distinguish or separate the question of what they are going to do from the issue of what they intend to do, and hence, how, within such a perspective, knowledge of what we are going to do goes ‘hand in hand’ with knowledge of what we intend to do.23 In the case of desire, questions about what we desire, considered in a purely first-person way, take questions about what’s good as their discursive transforms, because they can’t be distinguished from (object-focussed) questions about what to desire. Likewise, questions about what we intend, when considered in a purely first-person way, will take questions about what’s fit to be intended as their discursive transforms, because they won’t be able to be distinguished from (object-focussed) questions of what to intend. This enables us to explain why, within a purely first-person point of view, we won’t be able to distinguish or separate the issue of what we are going to do from the issue of what we intend to do: each issue will be indistinguishable for us from the (object-related) question of what to intend to do, and so, in turn, from the issue of which option is fit to be intended. Of course, in pessimistic moments, we can anticipate a (probable) mismatch between our own future action and a current intention for that future time that we self-ascribe

23 For the sake of simplicity, I cast the discussion at this point in terms of intentions because we can think of intentions as the states we enter into whenever we form a will to act in a particular way.

Self-knowledge, Truth and Value

49

in a purely first-person way. However, in such cases, we must be considering the prospect of what we will do as observational self-predictors. We have seen that when anyone takes up a question about whether they believe p, or desire p, in a purely first-person way, they are relating to their own psychological states in a ‘transcendental’ way. They ask a question about one of their own mental states in such a way that the focus becomes a candidate object for that state, and any attribution of the state to themselves goes beyond mere self-representation and involves a commitment about its object. Where a belief that p is self-ascribed in a purely first-person way, there’s a commitment to p’s truth. Where a desire that p is self-ascribed in that way, there’s a commitment to its being (somehow) good that p be true. In similar vein, where someone takes up the question of whether they intend p, in a purely first-person way, they ask that question in such a way that there isn’t any epistemic gap between it and the issue of what to intend, and so, what’s fit to be intended. And so, when they say of themselves, in a purely first-person way, that they intend p, that ascription surpasses a mere report on their intention and involves a commitment to a particular course of action’s being fit to be intended. The contrast is with cases where people, as self-spectators, ascribe intentions to themselves in the light of what their behavior reveals, or in the light of some phenomenological cue.24 In such examples, the self-ascriptions of intention are just that and involve no object-related commitment. Indeed, in cases where we purely first-personally self-attribute intentions, our ‘transcendental’ way of relating to our own mental lives is double-aspected. The selfascription of an intention that p carries a commitment about the propositional object of that intention twice over – once in relation it qua object of intention (namely, that it’s intention-worthy that p be true), and once with respect to it qua object of belief about the future (namely, that p is true). This is because, when we consider whether we intend p in a purely first-person way, we consider that question in such a way that we can’t distinguish or separate it from either the issue of what’s fit to be intended or the issue of what (we believe) we will do.25 This sharply contrasts with cases where we consider whether we intend p in a way that takes that fact as a candidate 24 As in the case of beliefs, there typically isn’t any tell-tale phenomenology associated with intentions, but there can be. Thus, we can know that we are set upon a course of action because of a felt resolve. 25 The two kinds of object-related commitment that attach to our self-ascriptions of intentions as agents show up in two different object-related ways that others might try to counter us when we announce that we intend to act in a certain way. I announce as agent: ‘I intend to leave early.’ You might respond: ‘That’s a bad idea’ – thus inviting me to reconsider whether I’ve made the right decision. Or you might respond: ‘You won’t leave early, you’ll hang around’ – thus denying that the future will be as I indicate. Of course, another way you might try to counter what I say when I announce that I intend to act in a certain way is by challenging whether I really have the intention. The possibility of this kind of rejoinder underscores the point that when we announce, as agents, that we intend to do something, we are, in addition to issuing object-related commitments, issuing a self-description.

50

Values and the Reflective Point of View

for detection, either from within or without. In such cases, these are all separable issues: we can reckon that we don’t intend to do, or intend not to do, what’s fit to be intended,26 and we can on occasion come to expect, either in a practical or inductive way, that we won’t enact intentions that we ascribe to ourselves as discoveries. To sum up: we are self-knowing subjects who can take a purely first-person perspective on our own present thoughts and future deeds. We can non-observationally come to know what our own current, conscious beliefs, desires, emotions and wills are in virtue of having made up our minds about what to believe, what to desire, what to feel and what to will. Similarly, we can non-observationally come to know what we are going to do in virtue of having settled on what to do. And, because of the role of these intermediary questions about what to think and what to do, in such cases, we end up answering questions about ourselves by answering questions about what’s true or of value. Moreover, where we consider both what we will to do and what we are going to do in a purely first-person way, we won’t be able to separate knowledge of the one fact about ourselves from knowledge of the other since, within the perspective, the epistemology of each of these facts will coincide with that of settling the issue of what deed is fit to be willed.

Decision and Discovery It is sometimes claimed that the difference between practical and empirical inquiry into what we are going to do is this: whereas the latter ends uncertainty with a prediction, the former ends uncertainty with a decision. This way of drawing the contrast between our perspective as agents on what we are going to do and our perspective as self-spectators highlights an important truth: as agents, we come to know what we are going to do by making up our minds about what to do, whereas, as self-spectators, we come to know what we are going to do by considering the evidence. However, this way of drawing the relevant contrast also tends to mask an important truth: whether we come to know about our own future action in a practical or empirical way, we therein form an expectation about what we will do. Consider what Richard Moran has to say on the matter. On the one hand, he apparently agrees that we do come to have expectations about what we are going to do, whether we end inquiry about our own future action from within a practical or empirical perspective. Thus, he emphasizes that someone who declares their intention ‘is committed both to the practical endorsement of the action and to the expectation of a future event’;27 that even though what they say isn’t based on evidence, it’s ‘something that may be true or false,’ which remains ‘answerable to the evidence,’ 26 In phenomenological examples, we might attribute a failure of intention, or a perverse intention, to ourselves on the basis of a lack of any felt resolve to do what (as we see it) merits being intended, or on the basis of a felt determination to do the opposite. 27 By ‘practical endorsement’ of the action, I take Moran to mean endorsement of the action as fit to be willed (for example, chosen). See his remarks, Authority and Estrangement, pp. 105–6, 126–7.

Self-knowledge, Truth and Value

51

and that practical and empirical forms of self-inquiry are ‘two routes to knowledge of the same facts.’28 On the other hand, sometimes, when he is focussed on the contrast between the practical and empirical perspectives in terms of decision and discovery, Moran writes as if he thinks that it’s distinctive of us that we don’t form expectations about what we are going to do when we end uncertainty about the future from within a practical perspective. Instead, we end uncertainty simply with a decision or an intention had by virtue of making a decision. For example, Moran characterizes the situation of Sartre’s akratic gambler in the following way. The gambler vainly seeks confidence about what he will do by substituting an empirical perspective on his own future behavior for a practical one. From within each of these perspectives, he relates to his decision not to gamble differently. Empirically, he relates to his decision as a psychological fact about himself with a certain degree of strength, which provides evidence for any expectation that he will avoid gambling. Practically, however, the gambler doesn’t relate to his decision as evidence about what he will do; rather, he relates to it as involving a commitment to its being true that he will avoid the gambling tables. Moran puts it like this: For the gambler to have made such a decision is to be committed to avoiding the gambling tables. He is committed to this truth categorically, as the content of his decision; that is, insofar as he actually has made such a decision, this is what it commits him to. For him his decision is not just (empirical) evidence about what he will do, but a resolution of which he is the author and which he is responsible for carrying through.29

This way of representing the gambler’s commitment, as agent, to its being true that he won’t gamble seems wrong. It ignores the role of expectations within the practical perspective. To be sure, in so far as the gambler decides not to gamble, he becomes committed to avoiding gambling: in making the decision, he acquires the disposition of being set to make it true that he doesn’t gamble in the future. That’s what it is to come to have an intention in virtue of deciding to act.30 However, being committed to making a proposition true is different from being committed to its being true. This is not to deny that the second kind of commitment is also relevant to the case. It is. For, from within his perspective as agent, the gambler, in ascribing the decision not to gamble to himself, will express a commitment to its being true that he won’t gamble. However, it isn’t true that, in such a case, he will express a commitment to the truth of that proposition qua object of his decision or intention.31 Rather, he will express a commitment to the truth of that proposition qua object of the belief or expectation that he forms in virtue of making up his mind not to gamble. To be sure, from within the gambler’s perspective as agent, the fact of his decision won’t serve as evidence 28 Ibid., pp. 87–8. 29 Ibid., p. 79. 30 See note 69, Chapter 2. 31 The object-related commitment that he expresses, qua object of decision or intention, is that not gambling in the future is what’s fit to be willed.

52

Values and the Reflective Point of View

for what he will do. For as long as he sustains that perspective, the relation for the gambler between his question of how he will behave in the future and his decision not to gamble will be this: the decision will settle in his own mind the question of what he will do, since he won’t be able to separate the fact that he has decided not to gamble from the fact that he won’t. Consider next some other remarks by Moran: Ending my uncertainty about, for example, what I will wear is indeed coming to know something; and, like other things I know, I can tell it to someone else so that they know it, too. But in other ways, although we rightly speak of knowledge here, it is not purely a theoretical or epistemic matter. My knowing what I will do next is not based on evidence or other reasons to believe something, so much as it is based on what I see as reasons to do something. Hence, a person’s statement of intention is not to be challenged by asking for his evidence. When I make up my mind about what to do, and tell someone else, I do indeed provide him with a reason to expect something, a very good reason if I am not vacillating, or a liar; but what I possess myself is not an expectation, based on evidence, but an intention, based on decision.32

These remarks are odd. It’s true that when I, as agent, end my uncertainty about what I’m going to do, I typically come to know something that I can communicate to you. And it’s also correct that, though you and I may now share a piece of knowledge about the future, our epistemic situations are very different. What you know about what I’m going to do is based on evidence, including the evidence of my statement of intention. By contrast, what I know about what I’m going to do is based on my deciding what to do, and so on practical considerations. Thus, it’s true that whereas you have an expectation about my future action based on evidence, I don’t. However, it surely isn’t a difference between us, as Moran seems to be suggesting in the passage above, that whereas you have an expectation about what I’m going to do, I don’t. I, instead, just have ‘an intention, based on decision.’ How could that be if both of us share the same piece of knowledge? The difference between us here, rather, has to do with how each of came by the belief about the future that we currently share. In the passage just quoted, Moran also says that ‘a person’s statement of intention is not to be challenged by asking for his evidence’ (my italics). That’s certainly correct. Where I was speaking as agent, you would show that you didn’t understand my perspective on the question of what I was going to do, if you challenged me for my evidence about my future behavior. Even so, you could still challenge what I say by appealing to the available evidence. Whereas you might object to my statement of intention on the grounds that it reveals a misguided decision, you might, instead, contest it with the claim that the evidence is against my doing as I say I am going to. The first way in which we can criticize statements of intention marks the fact that such statements are under use in context to manifest intentions. The second way in which we can criticize statements of intention – which, as noted towards the beginning of this section, Moran accepts – attests to the fact that agents match their 32 Moran, Authority and Estrangement, p. 56.

Self-knowledge, Truth and Value

53

decisions or intentions with expectations where they end uncertainty about what they are going to do by deciding what to do. For the way in which statements of intention are answerable to the evidence, even though they aren’t based on it, underscores the point that, while such statements are under use in context to manifest intentions, they are none the less statements that retain their identity as expressions (in the indicative mood) of corresponding beliefs about the future (attitudes about the future with a mind-to-world direction of fit). As agents, we end our uncertainty about what we are going to do, not by consulting the evidence, but by deciding what to do. But it would be a mistake to suppose that when we announce what we are going to do, in such cases, we are not issuing selfdescriptions that may be true or false. For, in practical examples, such descriptions – expressions of self-knowledge, where felicitous – become available to us in virtue of our deciding what to do. Similarly, in purely first-person cases, we end uncertainty about what our mental attitude is, not by referring to evidence for the attitude, but by deciding what to think, want, will or feel. And equally, it would be a mistake to suppose, as some philosophers have, that when we declare what we think, want, will or feel, in such cases, we are not really putting forward self-ascriptions that may be true or false. For, in purely first-person cases, such ascriptions – expressions of self-knowledge, where felicitous – become available to us by virtue of our making a decision about what to think, and the like. Consider the claim that apparent self-ascriptions like ‘I believe p’ are not really ascriptions at all, but just vehicles, in context, for expressing various object-related commitments. For example, in the belief case, it has been held that someone who says ‘I believe p’ typically isn’t describing themselves at all; rather, they are simply asserting that p in a guarded way.33 Consider that central range of examples where we use such first-person, present-tense, psychological sentences to express our conclusions about what we think, want, and so on, after reviewing the relevant evidence, desirability-considerations, and the like. It’s true that in cases like this, we use such sentences to issue object-related commitments. That is their ‘transcendental’ dimension in context. However, it’s entirely implausible to suppose that these sentences are not also being used, in such cases, to issue genuine self-ascriptions and often to describe known facts about our mental lives. There are various ways to emphasize that the ‘no self-ascription’ view is untenable.34 Suppose you ask me whether I think that p is true, I consider what’s to be said for and against p and reply with ‘Yes, I think p is true.’ Suppose further that you report on my reply to a third party, saying ‘Robert thinks that p is true.’ On the view in question, incredibly, it seems that what you say in your report on my reply 33 See ibid., pp. 70ff. Moran calls this ‘the Presentational View’ because the central thesis ‘is that in the first-person present-tense use, the verb-phrase “I believe” does not in fact have any psychological reference, but is instead a mode of presenting the relevant proposition’ (p. 71). 34 In what follows, I draw on Gallois’s discussion, The World Without, The Mind Within, pp. 112– 14.

54

Values and the Reflective Point of View

isn’t true just in case what I say in my reply to you is true. For, whereas I am simply asserting that p in a cautious way, you (presumably) are talking about my thoughts on the matter and attributing the belief that p to me. Next, suppose that you report on my reply to a third party with ‘Robert doesn’t think that p is true.’ On the ‘no self-ascription’ view, I can’t complain that you are contradicting what I explicitly said. For, on that view, after all, I just said in a guarded way that p is true. And you, for your part, aren’t saying anything about whether p is true. You are saying that I don’t believe it is, something that I (incredibly enough) made no comment on. Or, suppose that p isn’t true. On the ‘no self-ascription’ view, what I say when I utter ‘I think p is true’ turns out to be false even if it’s true that I do believe p. Surely the correct alternative here is that, in such a case, I speak truly about what my own mind is on the matter, but a further proposition to whose truth I commit myself in declaring what it is that I think – namely, the proposition that p – is false. Lastly, there’s a point about answerability to evidence. You ask me what I think about p, I consider the evidence for and against its truth, decide what to believe, and announce ‘I think p is true.’ There are two ways you might challenge me. You might counter that I’m wrong about p, or unlikely, on the evidence, to be right about it. Or you might query whether I really have the belief in question. Moran gives the following example of the latter kind of challenge: If I am known to vacillate on some question, where the interests of other people are dependent on what I really believe, my avowal of my belief may well be met not by denial of its truth, but by the demand that I take in the full empirical background of my own inconstancy before I announce my belief so confidently.35

Just as the first way of challenging what I say when I utter ‘I think p is true’ marks the way in which I thereby give a commitment to p’s truth, so the second way of challenging my statement in such a case marks its aspect as a bona fide autobiographical remark about what I believe.36 To recap: we need to be wary of drawing the contrast between the different ways in which we can end uncertainty about ourselves in terms of a decision–discovery contrast. The contrast, so drawn, speaks to how, in the empirical case, the issue in focus is the truth about ourselves, and so we consult the evidence, whereas in the purely first-person case, attention turns to the issue of what to believe, want, will, feel or do, and so we consult relevant object-related considerations. However, the contrast should not be read as meaning that, whereas we end uncertainty in an empirical way with genuine self-descriptions that may be true or false, this isn’t so where we end uncertainty in a purely first-person way. In both cases, we end uncertainty with bona fide self-ascriptions of mental attitudes or future deeds that

35 Moran, Authority and Estrangement, p. 88. 36 Compare ibid., pp. 66–77, 87–8, 100–107.

Self-knowledge, Truth and Value

55

may be true or false. In successful cases, these self-ascriptions will be expressions of different modes of self-knowledge.37

Velleman on Self-knowing Agency Let me finish with some remarks about a competing account of self-knowing agency. J. David Velleman holds that our choices (decisions or intentions) just are the expectations that we form about what we are going to do when we make up our minds about what to do. In outline, Velleman’s theory of choice as expectation goes like this.38 We have various first-order motives on any particular occasion of action. But our choice of behavior is regulated by the aim of knowing what we are doing, which is the constitutive aim of action.39 This aim inhibits us from doing anything until we have formed an expectation of what we are going to do; and then, once we have formed any such expectation, ‘the balance of forces will be decisively tilted in favour of doing the expected thing.’40 Such a self-fulfilling expectation, says Velleman, functions in a peculiar way: it settles in our own minds the question of what we are going to do, and it thereby settles the same question in fact. In short, it has exactly the function of choice.41 In its detail, this state that constitutes choice: … [1] represents as true that we are going to do something … [2] aims therein to represent something that really is true … and [3] causes the truth of what it represents.42

By Velleman’s lights, this means that choice (1) has the direction of fit of cognition, (2) the constitutive aim of belief, and (3) the same practical ‘direction of guidance’ as desire.43 Some preliminary comments are in order. Velleman has reconfigured the notion of direction of fit. The basic story, as he retells it, is this. There’s a contrast between 37 Within empirical or observational self-knowledge, there’s an important distinction. Knowledge of one’s own mental states that is phenomenologically based will be a kind of first-person self-knowledge (self-consciousness). Knowledge of one’s own mental states that is based on behavioral (or neural) evidence will be third-person self-knowledge. 38 I revisit Velleman’s theory of agency in Chapter 5, pages 96–9. 39 This constitutive aim of action, according to Velleman, is sub-agential: it is not represented in our practical reasoning as an end-in-view. It also might be sub-personal. That is to say, rather than being among a person’s desires, it might just be an aim that is implicit in a sub-personal psychological mechanism that characterizes them. See note 46, Chapter 5. 40 ‘Introduction,’ in The Possibility of Practical Reason (Oxford, 2000), p. 22. 41 Ibid., pp. 23–4. Velleman first introduced his theory of choices or intentions as selffulfilling expectations in his book Practical Reflection (Princeton, NJ, 1989). I set out many of the salient details of that early presentation in my paper ‘Knowing What I’m About To Do Without Evidence,’ International Journal of Philosophical Studies, 6 (1998): 233–4. 42 ‘Introduction,’ in The Possibility of Practical Reason, p. 26. 43 See ibid., pp. 24–6.

56

Values and the Reflective Point of View

the directions of fit of desires and cognitions (in general): whereas our desires regard their propositional objects as to be made true (as representations of how things are to be arranged), our cognitions (including our beliefs) regard their propositional objects as true (as representations of how things are arranged). The difference among cognitions, given their common direction of fit, is then given by their different constitutive aims: … belief is regulated by the aim of regarding something as true only if it really is true; whereas imagining and assuming entail accepting a proposition fancifully or hypothetically, with some aim other than getting at the truth.44

Velleman exploits this reconfiguration of direction of fit to make space for a range of beliefs that have the same ‘direction of guidance’ as desires. He writes: … An attitude’s direction of guidance consists in whether the attitude causes or is caused by what it represents. Now, there is a temptation to assume that if a mental state is cognitive, representing how things are, then it must be caused by how they are; whereas if it causes what it represents, then the state must be conative, representing how things are to be. This assumption implies that a cognitive direction of fit entails a passive direction of guidance; and, conversely, that only states with a conative direction of fit can be active or practical. But when I make a choice, a question is resolved in the world by being resolved in my mind. That I am going to do something is made true by my representing it as true. So choice has the same direction of fit as belief but the same direction of guidance as desire: it is a case of practical cognition.45

Some obscurities attach to this reconfiguring of direction of fit and introduction of the notion of direction of guidance. For one thing, it’s hard to know how to psychologically ‘place’ the direction of fit aspect of attitudes, as Velleman reworks it. By standard intuitions, what Velleman calls an attitude’s ‘direction of guidance’ looks like it belongs to an account of the functional upshot of the attitude’s direction of fit. Thus, it would be reckoned to be part of our concept of belief about the sensible world, given the topic and mind-to-world direction of fit of such mental states, that they are typically caused by perceptual input, just as it would be reckoned to be part of our concept of desires to act, given the world-to-mind direction of fit of such mental states, that they typically promote their own satisfaction, given true beliefs, and so on.46 44 Ibid., p. 25. See ibid., pp. 9–10, 24–5, and ‘The Guise of the Good,’ in The Possibility of Practical Reason, pp. 105ff. 45 ‘Introduction’, in The Possibility of Practical Reason, p. 25. 46 Velleman takes up this issue in relation to belief in ‘On the Aim of Belief,’ in The Possibility of Practical Reason, pp. 249ff. He argues that the motivational role normally assigned to belief as partner to desire in promoting action is really the motivational role that is distinctive of regarding-as-true, the sub-doxastic attitude (of acceptance) that beliefs have in common with all cognitions. See especially, pp. 255ff.

Self-knowledge, Truth and Value

57

Moreover, there’s a strange feature of Velleman’s story that gets blurred by the claim that choices are beliefs with the same ‘direction of guidance’ as desires. For, in the terms of the story, the beliefs about what we are going to do that are reckoned to be our choices come to be self-fulfilling courtesy of the aim of knowing what we are doing: unlike desires, these belief-choices are not, in and of themselves, dispositions to act. So, by these lights, it isn’t intrinsic to choices, decisions or intentions, for all that they are supposed to be ‘practical cognitions,’ that they tend to promote what they represent; rather, they get to be that way by virtue of an extrinsic feature of the practical psychological setting in which they are arrived at. And that seems wrong. Why is Velleman keen to identify our choices with the beliefs we form about what we are going to do when we make up our minds to act? Of course, the identification is very congenial to his account of agency as behavior regulated by the aim of knowing what we are doing. However, it’s also obvious that Velleman thinks that the identification is intuitively appealing. Let’s consider why he thinks this. I have already mentioned the features of choice that Velleman takes his account to square with: our choices involve answering in our own minds the question of what we are going to so, and they thereby settle the same question in fact. For Velleman, the first point means that our choices have to go beyond the direction of fit of desire and have the same the direction of fit as belief. He comments: Because choosing entails settling a question in one’s mind, it requires more than representing an answer as to be arranged. If it were still to be arranged that I was going to act, then it would not yet be settled that I was going to act; and insofar as I regarded it as to be arranged, I would not yet regard it as settled. Settling on a future action thus requires representing the action as arranged: my choice makes it true that I’m going to act, by representing it as true that I’m going to act. It therefore has the same direction of fit as a belief.

And for Velleman, the second point, that our choices aim to settle a question in fact by settling a question in our minds, means that our choices share the constitutive aim of belief. He writes: When I form a choice or decision … I aim to settle a question in my mind only insofar as I can thereby settle it in fact; I aim, that is, to avoid representing an arrangement that I am not thereby managing to make. In the face of compelling evidence that I’m not going to do something, my mind cannot remain made up that I’m going to do it: the decision to do something, like the belief that I’m going to, cannot withstand evidence to the contrary. It thus represents things as having been arranged in some way, with the aim of representing how they really have been arranged, albeit by this very representation. It has the cognitive direction of fit and the associated aim of being true, despite having a practical direction of guidance.47

Let’s take these considerations in turn. The first claim is that our choices wouldn’t settle in our own minds the question of what we were going to do if they had the 47 ‘Introduction’, in The Possibility of Practical Reason, p. 25.

58

Values and the Reflective Point of View

same direction of fit as desires – that is, if they were simply representations of how things are to be arranged. There are two points here. When we are deliberating about what to do, we have a complex target. We are trying to settle at once the question of what to do (transform of an imperative) and the question of what we are going to do (transform of an indicative). And we can settle the question of what to do by coming up with a representation of how things are to be arranged: of what is to be done. Moreover, when we are deliberating about what to do, we are trying to get to know what we are going to do in a special way: by deciding what to do. If I am right, the role of our purely first-person point of view as agents explains how we can work this: within that point of view, the two issues coalesce, with the issue of what to do taking the burden of inquiry. The two issues perspectivally coalesce, but we enter into two different mental states as a result of our inquiry: we form a belief about what we are going to do in virtue of deciding what to do. Velleman’s collapse of choices, decisions and intentions into future expectations ignores the special role of perspective in practical inquiry. It’s our purely first-person perspective as agents that enables us to go from how things are to be arranged (and so, from a state with a world-to-mind direction of fit) to how things in the future are arranged (and so, to a state with a mind-to-world direction of fit). The second way in which Velleman finds his account of choices as a kind of belief about the future intuitive is that our choices aim to settle a question in fact by settling in our own minds a question about what we are going to do. However, that way of characterizing the characteristic function of choices already smuggles in the theory of choices as beliefs that promote their own truth. The neutral alternative is to say that it’s the function of choice to settle what we are in fact going to do by settling in our own minds a question of what to do. It’s a further issue what the relation is between settling in our own minds a question of what to do and settling in our own minds a question of what we are going to do. On my view, the assimilation of answers here is implausible: it ignores the way in which issues can perspectivally coalesce in their epistemologies without the mental states which we enter into in answering them being one and the same state. A gloss on the function of choice which respects this would be: our choices aim to settle in our own minds a question about what to do, and thereby to settle in our own minds, and in fact, a question about what we are going to do. Velleman is impressed by the way in which our choices or decisions are sensitive to the ‘evidences of truth.’ He makes the point that we cannot remain decided on doing something in the face of compelling evidence that that isn’t what we are going to do. It’s true that decisions are typically destabilized or undone in such cases; but that’s explicable without supposing that decisions are beliefs. It’s a matter of what it takes to be able to occupy a point of view. In order to sustain a decision to act in a certain way as our answer to a question of what to do as agents, we need to be able to continue to treat the question of what to do as inseparable from the question of what we are going to do; and to do this, we need to be able to continue to see the future action that we have settled on as at least one that isn’t definitely ruled out as

Self-knowledge, Truth and Value

59

an option by our evidence (about our surroundings, abilities, dispositions, strength of motives, and so on). Finally, there’s a point about the phenomenology of agency. Velleman holds that his account of choice as expectation that manages to make true what it represents is faithful to what it’s like to be an agent. Speaking of creatures designed such that their antecedent motives are restrained until there’s something they expect themselves to do, whereupon they become decisively motivated to do the expected thing, he writes: Your creatures should also share our sense of having an open future or a metaphysically free will. … When they imagine various alternative futures for themselves, they would be correct to believe in any one of them, since the future they end up believing in will be the future they end up having. And being in a position to believe correctly in any one of several different imaginable futures feels like being in a position to have any one of several different imaginable futures. … In reality, your creatures aren’t in a position to have any one of several different futures, only to believe in them correctly. But the confusion is understandable.48

Now, it’s true that within the perspective of agency the future feels free. And it’s true that it’s our capacity to make choices that makes us feel free as agents. However, this can be readily explained without supposing that we feel free because we recognize, when we consider our options, that whichever one we come to expect, we’ll (typically) be right about what the future holds. Again, what’s crucial, and what Velleman neglects, is our distinctively first-person point of view as self-knowing agents. Quite simply: when we consider our future as agents, the future looks up to us because the question of what it holds presents as something that’s settled by however we choose or intend to make it.

48 Ibid., p. 23.

This page intentionally left blank

Chapter 4

Perverse Agency By reasoning about values, we reflectively put ourselves into a position to shape our desires, wills, actions and feelings; and, typically, the values we form do shape our practices and passions. By cautious expressivist lights, these values are dispositions that, in the normal course of events, find expression in our practices and passions. However, our values have their vicissitudes, and sometimes we betray them. At the extreme, we are perverse, taking what we see as entirely bad as our goal and pursuing it. Yet the possibility of perversity can seem problematic. There are two inherited stories of intentional action. On one story – let’s call it the motivational story – intentional agents are pursuers of goals or desired outcomes. On another story – let’s call it the evaluative story – intentional agents are pursuers of value, as they see it. On the motivational story, someone who acts intentionally is conceived of as manifesting the strength of their desire for an outcome. On the evaluative story, someone who acts intentionally is conceived of as responding to their perception of the good. Within the terms of the evaluative story, intentional agents look like practical reasoners, who act under the guidance of reason. Within the terms of the motivational story, intentional agents look like they just act at the behest of desire, guided by belief.1 In a spirit of unification, we might try to supplement the motivational story with the evaluative one (as Elizabeth Anscombe does), or even collapse the former into the latter (as Davidson does). By these lights, agents are pursuers of (perceived) value – and so clearly practical reasoners – in so far as they are pursuers of goals. Yet moves like this is cannot accommodate pathologies of agency like perversity. Agents who seem perverse – like Satan and self-haters – implausibly turn out to be closet lovers of the good, as indeed do some other pathological agents. We need a theory of agency that doesn’t turn us into lovers of the good just because we are pursuers of goals, and yet still leaves us intelligible, where we are perverse, as agents of reasoned activity. We get this, I argue, by allowing for the difference between purposive and reasoned activity, and by acknowledging how, as agents of reasoned activity, we are sometime lovers of the good and sometime mere lovers of success in action.

1 Compare Velleman, ‘The Guise of the Good,’ in The Possibility of Practical Reason (Oxford, 2000), pp. 99–102, and Michael Smith, The Moral Problem (Oxford, 1994), pp. 131–3.

62

Values and the Reflective Point of View

Is Satan a Lover of the Good? Anscombe reconstructs Satan as a lover of the good. Satan resolves: ‘Evil be thou my Good.’ According to Anscombe, this is intelligible only if we can say what, for Satan, is good about evil. Bonum est multiplex. Anscombe suggests an answer on Satan’s behalf: ‘the good of making evil my good is my intact liberty in the unsubmissiveness of my will.’2 Whatever we make of Satan’s project ethically, we can now at least understand why he makes evil his aim. He remains intelligible as an agent as a pursuer of value. Satan sees the pursuit of evil as good in so far as it preserves an unsubmissive will. However, if this is to be the picture of Satan, it’s problematic. It makes it look as though Satan is a closet lover of the good.3 Now, it is Milton’s Satan who resolves ‘Evil be thou my Good.’ So let’s take a closer look at Satan as Milton depicts him in Paradise Lost. Satan has lost Heaven through pride and ambition (Bks I, V, VI). Satan, a pre-eminent Archangel, refuses to accept God’s appointment of His Son as Commander of Heaven (Bk V, 658ff.); and, having rebelled, refuses to repent. With the loss of Heaven, and because he will not repent, Satan is, as he realizes in anger and despair, condemned to misery and pain. For this desperate plight, he seeks revenge on God; and he is consumed with envy of those who still enjoy happiness (Bks I, II, IV, IX). When Satan resolves ‘Evil be thou my Good,’ he, of course, is not revising his theory of moral good and evil. What makes him satanic is that, while holding his moral theory constant, he substitutes evil for good as his goal in action, seeing the pursuit of evil as a means to several ends. The pursuit of evil will preserve an unsubmissive will (Bk I, 160ff.). It will go some way to evening things with God in so far as it’s a spoiling exercise. It will afford Satan the pleasure of making companions in misery of those who currently enjoy the goods from which he is forever excluded (Bk IX, 129–30; 476–79). Finally, it will help Satan to win glory and power for himself (Bks IV, 79ff.; IX, 97ff.). The spirit of Anscombe’s remarks seems to be this. When Satan substitutes moral evil for moral good as his goal in action, he must nevertheless conceive of the new goal as good in some way. What can’t be the case is this: Satan pursues evil while seeing it as entirely bad. And Anscombe’s suggestion seems to be along these lines: Satan’s further goals in pursuing evil indicate the (perceived) goods he thereby seeks. For Satan, pursuing evil promotes goods like insubordination, getting even with God, the pleasure of making the blessed wretched, and power over what was once God’s. Yet Milton’s Satan doesn’t fit Anscombe’s mold. Milton’s Satan does not have a theory of the good under which these goals – or the actions that promote them – count as goods. This is crucial to the picture. When Satan says ‘Evil be thou my Good,’ he seems to be focussing on evil means rather than evil ends; but of course, 2 Intention (Oxford, 1976), p. 75. 3 Compare Velleman, ‘The Guise of the Good,’ in The Possibility of Practical Reason, pp. 118–20.

Perverse Agency

63

since he shares his moral theory with God, Satan sees these ends as as evil as the means that secure them. Moreover, Satan doesn’t see the achievement of his satanic goals as benefiting him. To the contrary! Satan realizes that achieving these goals will worsen his pain and wretchedness in Hell (Bk IV, 25–6; 91–2; Bk IX, 124ff.). So, what are we to say of Satan, so conceived? Satan is driven by a bevy of dark emotions – pride, ambition, envy and revenge. The temptation is to think of each of these as targeting (perceived) goods. However, this is potentially a very misleading way of understanding Satan’s psychology. Consider Satan’s pride. It’s true that Satan, in his pride, has an insubordinate will as his guiding goal. So, for proud Satan, an insubordinate will counts as good in just the sense that, being what pride aims at, it is that whose preservation is the criterion of success for any related action. However, this is merely a formal sense of ‘good,’ in which any aim whatsoever counts as good because it counts as that whose achievement makes any related action a success. Proud Satan can see a defiant will as good without seeing such a will as good in terms of any theory of the good (moral or otherwise) that he endorses.4 Similar remarks hold for ambitious Satan. Satan lusts for the power and glory that will be won by corrupting innocence. We might say that, internal to his overweening ambition, Satan sees power and glory as goods; but then we had better be careful. For this can’t mean that Satan sees power and glory of the sort he aspires to as good, under his own substantive conception of the good. For Satan – with an acute and angry sense of loss (BK IV, 24ff.; 502ff, IX, 119ff.) – construes his own predicament as one in which he is utterly alienated from all genuine goods. Just before he defiantly resolves ‘Evil be thou my Good,’ Satan, in despair, laments ‘all Good to me is lost’ (Bk IV, 109). If we say that Satan, in his ambition, sees power and glory as goods then all we can mean is something like the following. For ambitious Satan, power and glory – being what ambition seeks – are the things whose realization makes related actions successful. According to Colin McGinn, ‘Pain is the devil’s gift to the envious.’5 This is because the envious are pained at the envied’s enjoyment of life and pain is uniquely suited to lessen its sufferer’s attachment to life, even to make them hate it. Cruelty, motivated by envy that seeks relief from its own pain in the suffering of the envied, ‘makes sense in terms of the agent’s pursuit of his own well-being.’6 This isn’t how it is with Milton’s Satan. It’s plausible that envious Satan wants so to reduce Adam and Eve that they come to hate life; indeed, after the Fall, Eve suggests joint suicide to Adam (Bk X, 1000–1006). However, Satan definitely doesn’t believe that he can help himself by making Adam and Eve miserable. It’s true that he finds the prospect of the pleasure this will afford attractive (Bk IX, 129– 30; 477–9); but this doesn’t mean that, by his own lights, he counts such pleasure 4 Compare Velleman, ‘The Possibility of Practical Reason,’ in The Possibility of Practical Reason, p. 190. For further comment on the contrast between formal and substantive valuing, see pages 76–7 below. 5 Ethics, Evil and Fiction (Oxford, 1997), p. 81. 6 Ibid., p. 82.

64

Values and the Reflective Point of View

as a (substantive) good. In his envy, Satan merely counts winning the pleasure that comes from ruining the innocent as the mark of success in action. It may be that Milton’s Satan is therefore unlike some other antiheroes, often thought of as satanic. Consider Genet* – Genet as he presents himself in fiction. Genet* seeks to sever the bonds of intimacy by evil methods like betraying friendship in order to secure the ‘diamond’ of solitude – and with it, proud freedom.7 But does Genet* think that by perpetrating an evil like betrayal he will secure a good in the freedom of solitude? Or is the ‘diamond’ of solitude just a lure that fascinates him into its pursuit through evil? And does Genet* think that acts like selling out a friend to the cops are really evil or does he just think that that’s how they conventionally count? It’s hard to say. Sometimes Genet* seems to be a revisionist about morality, affirming ‘Evil be thou my Virtue’ in an attempt to forge a new ethic – an individualistic code with solitude as its central value.8 If Genet* is a pursuer of value in so far as he is a pursuer of solitude by acts like betrayal, he’s an ersatz Satan. Genet* is truly satanic only if he sees betrayal – and the isolation from human bonds that it secures – as unredeemably bad. We can understand a truly satanic resolve of ‘Evil be thou my Good’ without reconstructing its utterer into a lover of the good, in any serious sense. Satan sees his project of evildoing as furthering the goals of pride, ambition, envy and revenge. He is bitterly aware that he is excluded from all authentic or substantive goods – that is, goods that count as goods by some standard that is independent of what he is presently aiming at. In so far as these goals can be cast as perceived goods, this can only mean that, in his dark emotions, Satan sees them as the things whose attainment counts as the (ultimate) criterion of success for related actions. When Satan resolves ‘Evil be thou my Good,’ he simply sets up evildoing – especially the evil of corrupting innocence – as a (secondary) goal, and hence as that whose realization counts as the (intermediate) criterion of success for his action. Let me call the thesis that agents are lovers of the good in so far as they are pursuers of goals Perversity Lost. Perversity Lost rules out as impossible genuine perversity in action – that is, action like Satan’s that is directed at the entirely bad. However, we can make sense of perversity in action, at least up to a point, by exploiting the contrast between substantive and merely formal valuing. Satan is no lover of the good. Filled with certain dark passions, and treating as his ends-in-view the goals of pride, ambition, envy and revenge, Satan is just bent on the success in action that doing evil will bring. There are other modalities of perverse action that we can save by exploiting the idea that, as practical reasoners, we are sometime lovers of the good and sometime mere lovers of success in action. Consider cases of seeking the bad where agents act from emotions – like anger, hate, contempt, despair, and the like – in which what is

7 The Thief’s Journal (Harmondsworth, Middlesex, 1975), pp. 202–3; Funeral Rites (St. Albans, Hertfordshire, 1973), pp. 72–3. 8 Ibid., pp. 157–8, 198.

Perverse Agency

65 9

desired is simply and precisely harm to others or themselves. In these cases, what is directly wanted is what will be bad for the other or oneself and agents act on a belief about how to secure the bad. Examples of desiring the bad like this differ from Satan’s because they do not involve further ends, like (in the case of envy) replacing the pain of envy with delight in a downfall. To be sure, a standard move for converting such examples into ones wherein the serious good is pursued is to posit further goals – like pleasure for oneself, power, getting revenge, and so on. However, as the Satan case makes vivid, this move (which is not true to the psychology of such phenomena anyhow) is not enough to effect the conversion.10 We can save the hate that directly seeks the harm of the hated, just as we can save Satan’s satanism, by reckoning that such haters take the harm they seek to be good in just a formal sense: it’s a good in just the sense that its infliction (on themselves or others) has become the bleak mark of success in action for them. Perversity Lost excludes other pathologies of agency besides perverse action. Consider akrasia with respect to a single value. Michael Stocker has argued that monists about value can be akratic. For even within a single value like pleasure, what is seen as best and what attracts most can diverge: despite unity of value, there can be plural objects of attraction. So, for example, while Akrat (a hedonist) might judge it best to drink an unadulterated wine, since that would be most pleasurable, he might be fascinated to try a wine laced with pepper for the novelty of it, even though he thinks this would be less pleasurable to drink. In a case like this, says Stocker, the attractive feature – the novelty of tasting the wine laced with pepper – provides Akrat with a reason to act, though not, by his hedonist lights, a justifying one, since Akrat does not count the feature that fascinates a good-making one.11 Like the perverse agent, Akrat doesn’t act as a lover of the good. Akrat has an aim: to sample a novel taste; and, in so far as tasting the wine laced with pepper promises novelty, it presents as a way of realizing this aim. Perversity Lost would force us, in such a case, to reconstruct Akrat as a closet pluralist about the good, who counts at least pleasure and novelty as goods. However, we can avoid this distortion of yet another vicissitude of agency. Akrat is no lover of the good just because he tries for novelty; and, as an agent of reasoned activity, he is simply someone who is set on success with respect to a passing fancy.

9 Compare Michael Stocker, ‘Desiring the Bad,’ Journal of Philosophy, 76 (1979): 748–9. 10 Note Velleman’s memorable remarks: ‘Just as Satan would have to shed his satanism in order to value evil sub specie boni, so I would have to shed my despair in order to pursue a self-destructive course under that positive guise. If I were swayed toward an action because its badness made it seem like a good thing to do, then I’d be in the business of finding silver linings, a business that’s closed to me so long as I’m truly acting out of despair’ (‘The Guise of the Good,’ in The Possibility of Practical Reason, p. 122). 11 Plural and Conflicting Values (Oxford, 1990), Chapter 7.

66

Values and the Reflective Point of View

Perversity Lost and Ways of Sensemaking Let’s go back to the beginning. There seem to be two inherited stories of intentional action: the motivational and evaluative stories. One way of dealing with this dualism is to supplement the motivational story with the evaluative one. This seems to be Anscombe’s tactic. Another way to deal with the dualism is to collapse the motivational story into the evaluative one. This is Davidson’s tactic. Either of these tactics gives us Perversity Lost, the commitment that we are pursuers of the good to the extent that we are pursuers of goals. Consider the following explanation of an action: (Mot) Adrian adds sage to the stew because he wants to improve its taste and he thinks that adding sage will do this.

Davidson holds that we can perspicuously rewrite (Mot) as: (Eval) Adrian adds sage to the stew because he thinks any act is desirable in so far as it improves the stew’s taste and he thinks that adding sage will do this.

(Eval), on Davidson’s view, explains Adrian’s deed by ‘rationalizing’ it. That is, (Eval) enables us to see Adrian as trying to secure some (perceived) good in so far as he tries to improve the taste of the stew by adding sage to it.12 The collapse of explanations like (Mot) into explanations like (Eval) yields a commitment to Perversity Lost: wherever we try to get something we want, we are lovers of the good. By these lights, defiant Satan (for example) sees the evil act of corrupting Adam and Eve as good in so far as it preserves an unsubmissive will; self-haters see prospective self-destructive acts as good in so far as they harmful to themselves; and Akrat is revealed as a pluralist about the good, counting at least pleasure and novelty as goods. The crucial step in Davidson’s collapse of the motivational story into the evaluative one is his assimilation of desires to evaluative attitudes, with conditional contents; and he has never given any persuasive reason for it.13 It’s a move that we should be wary of in any case. For it doesn’t just banish perversity by turning us into lovers of the good in so far as we are desirers or havers of goals. Since the move equates desires with taking things to be good in some way, it also turns us into desirers or havers of goals in so far as we form such values, and thus rules out 12 ‘Action, Reasons and Causes,’ in Essays on Actions and Events (New York, 1980), pp. 3–19. 13 I set out Davidson’s developed view about desires, intentions and evaluations in Chapter 5, pages 99–100. I also criticize there his account for what I call the ‘desire-as-valuing mistake.’ For Davidson’s theory of desires as values with conditional contents, see ‘Intending,’ in Essays on Actions and Events, pp. 96–102, and ‘Reply to Christopher Peacocke,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Action and Events (Oxford, 1985), pp. 207ff.

Perverse Agency

67

a certain kind of weakness, where we fail to desire something in the respect that we value it.14 There is, however, another way to mark the cost of rewriting the motivational story in evaluative terms. What the rewrite actually does is erase the contrast between two different, compatible ways of making sense of agents as Rational Animals, and along with this, two different, compatible ways of conceiving of behavior as activity. This is sufficient reason to reject it. Let’s consider these contrasts in more detail. Consider (Mot). (Mot), as it stands, doesn’t ‘rationalize’ what Adrian does – that is, it doesn’t explain his behavior by letting us see what he takes to count in its favor. It’s easy to suppose that it does, perhaps; but to suppose that it does is to neglect what (Mot) actually says, as it stands. To adapt an example of Velleman’s: (Mot) represents Adrian as desirous toward the proposition ‘I improve the taste of the stew’ and credent toward the proposition ‘Adding sage will improve the stew’s taste.’ The contents of these attitudes don’t in themselves in any way jointly represent the act of adding sage to the stew in a favorable light, or as justified. So (Mot), as it stands, doesn’t explain what Adrian does by depicting him as responding to some property of the deed that he sees as justifying it. How then does (Mot) explain what he does? Well, (Mot), as it stands, is an attitudefocussed explanation. It explains Adrian’s adding of sage to the stew by representing him as ‘manifesting the valence’ of his desirous attitude toward the proposition ‘I improve the taste of the stew,’ under the impact of his belief that adding sage will improve the taste.15 In this way, (Mot) explains Adrian’s behavior, conceived of simply as a piece of goal-directed or motivated activity. It explains what he does by setting out his motives. Of special note, (Mot), as it stands, doesn’t explain what Adrian does by representing it as the outcome of a sequence of reasoning.16 Now contrast (Eval). (Eval) does (in Davidson’s sense) ‘rationalize’ what Adrian does. (Eval) is an object-focussed explanation. It explains what Adrian does in terms of attitudes of his whose contents or objects do together imply that adding sage to the stew is desirable in so far as it’s a way of improving its taste.17 Thus, (Eval), 14 I give an extended example of this kind of weakness in Chapter 5, The Possibility of Weakness of Will (Indianapolis, IN, 1987), pp. 112ff. The example comes from Simone de Beauvoir’s The Woman Destroyed (London, 1971). Michael Smith also draws attention to this kind of phenomenon in his critique of Davidson on values and desires, in The Moral Problem, pp. 135–6. 15 ‘The Guise of the Good,’ in The Possibility of Practical Reason, pp. 100–101. 16 We can use explanations of the form of (Mot) to make sense (in motivational or teleological terms) of agents like small children and non-human animals, who can’t plausibly be credited with evaluative competence, but who can plausibly be credited with aims. Compare Velleman, in ibid., p. 104. 17 The important contrast between (Mot) and (Eval) turns on the difference between values and desires, which the cautious expressivist acknowledges. These are two different (though closely related) ways of having a truth-favoring attitude towards a proposition p: in the case of a value, we favor p to be true as a sub-content; in the case of desire, we favor p

68

Values and the Reflective Point of View

unlike (Mot), does explain what Adrian does, conceived of as a piece of activity that is guided by his reasoning and grasp of a justification.18 It explains his behavior, conceived of as a piece of reasoned activity.19 So, when Davidson proposes that we rewrite explanations like (Mot) as explanations like (Eval), thus turning us all into lovers of the good to the extent that we are pursuers of ends, he ignores the contrast between two different though compatible kinds of explanation of behavior, requiring that the one conform to the pattern of the other; and he thereby collapses purposive activity (which may or may not also be reasoned) into reasoned activity.20 The way to resist this road to perdition is to refuse the very first step – the construal of desires as values. And resisting perdition here helps us to make sense of the pathologies of agency that concern us. We also need to acknowledge how, as agents of reasoned activity, we are sometime lovers of the good and sometime mere lovers of success in action. Suppose Adrian is a hedonist and, guided by his theory of the good, comes to believe that any act is good in so far as it improves the stew’s taste. In such a case, Adrian will count any relevant act as good in a serious or substantive way. If we explain that Adrian to be true as a whole content (see note 69, Chapter 2). Corresponding to these two different, compatible ways of having a truth-favoring attitude towards a proposition are two different, compatible ways of explaining what we do. The difference in kind of explanation is given by the difference in whole content of the two attitudes; the compatibility of the two kinds of explanation reflects the way in which the one kind of truth-favoring attitude towards a proposition (valuing) includes a disposition to form the other kind (desire). 18 Michael Smith agrees that Davidson ignores two different ways of making sense of agents as Rational Animals, and that one of these ways focusses on attitudes, while the other focusses on contents of attitudes. See The Moral Problem, pp. 130–41. In Smith’s preferred terms, an explanation like (Mot) is an explanation of what an agent does from the intentional perspective in terms of motivating reasons. In explanations from this perspective, ‘we are interested [in] which psychological states of the agent explain her actions’ (ibid., p. 132). By contrast, an explanation like (Eval) is an explanation of what an agent does from the deliberative perspective in terms of the pattern of reasoning (implicit or explicit) behind the action. In explanations from this perspective, ‘we are interested in which propositions, from the agent’s point of view, justify her actions’ (ibid., p. 132). 19 In my paper ‘Is Satan a Lover of the Good?,’ Ratio, 13 (2000): 22, I say that the motivational story is really a story about motivated behavior and not about intentional action. However, the truth is that the concept of intentional action unhappily straddles both motivated or purposive activity and reasoned activity. This equivocation is blurred in Davidson’s collapse of the two stories of action. 20 In characterizing the way in which Davidson retells the story of motivation as the evaluative story, Velleman remarks: ‘He demands that the outcome of motivation – the act of adding sage – be justified by some propositional content of the agent’s attitudes, as if it were a conclusion following from premises. And he obtains the required content by incorporating the valence of the agent’s attitude toward “I improve the taste” into a new proposition: “Improving the taste is desirable”. The story of motivation is thus transformed into the story of an inference, in which the agent is under genuinely rational guidance’ (‘The Guise of the Good,’ in The Possibility of Practical Reason, pp. 103–4).

Perverse Agency

69

adds sage to the stew because he sees this as a way to realize the hedonic good of improving its taste, we make sense of what Adrian does by representing him as a lover of the good. On the other hand, improving the taste of the stew may just be a whim of the moment for Adrian, and he may simply count any act as good in so far as it satisfies his present fancy, to the extent that it’s an act of improving the stew’s taste. If we then explain that Adrian adds sage to the stew because he sees this as a way to achieve the good of improving its taste, conceived of in this merely formal way, we make sense of what Adrian does by representing him as a lover of success in action, as a reflective pursuer of a desired end.21 In sum, we can make sense of agents like Satan and Akrat once we allow for the contrast between purposive and reasoned activity and accept that we are sometime lovers of the good and sometime mere lovers of success in action. Agents like Satan and Akrat don’t then turn up as already lovers of the good just because they are trying to bring about certain states of affairs; and they are left intact as agents of reasoned activity, at least to the extent that they are bent on success in action in a self-aware way. Let’s now consider in more detail how we should understand reasoned agency and its relation to purposive agency.

Saints in Action and Sensemaking Suppose that Genet* is an ersatz Satan – a saint in action, guided by his theory of the good. And suppose we explain Genet*’s act of betrayal thus: he sells out his friend to the cops because he thinks that any act is good in so far as it secures solitude, and that betrayal will secure solitude. In such a case, we give it to be understood that the evaluative explanation is matched by a motivational one: that Genet* is the pursuer of a goal because he is a lover of the good. In other words, we give it to be understood that, because he values solitude, Genet* is out to secure it by selling out his friend to the cops, given his belief that betrayal of intimacy is the path to solitude.

21 Smith holds (in The Moral Problem, pp. 131–2) that all evaluative (in his terms, deliberative) explanations represent an action as the product of normative reasoning. However, this is a mistake. Normative reasons figure where the action is being explained as a response by an agent to their serious, substantive values. Of course, an agent can be interpreted as, at one and the same time, a lover of the good and a lover of success in action in their reasoned activity. Suppose Adrian (a hedonist) acts after reasoning: I want to improve the taste of the stew because any act is good in so far as it does this; and so on. I note that explanations of the form ‘A Vs because A wants to Y and A thinks Ving is a way to Y’ have double lives. They typically are explanations of behavior, conceived of as purposive activity. However, they can serve as explanations of behavior, conceived of as reasoned activity. This is especially so in the first-person. Suppose whimsical Adrian says ‘I’m adding sage to the stew because I want to improve its taste and (I think) adding sage will do this.’ He’s presumably indicating the content of the reasoning behind what he’s up to. In full dress, his initial premise is: ‘Any act is good (a success) in terms of satisfying my current desire, to the extent that it’s an act of improving the taste of the stew.’

70

Values and the Reflective Point of View

The pragmatics of explanation here reflect the process whereby Genet* comes to act. This process can be characterized naturally in a way that reflects a cautious expressivist construal of values. In accordance with his theory of the good, Genet* judges that any act is good in so far as it secures solitude. This evaluation then finds expression in an end-setting desire to secure solitude, which combines with his belief about betrayal as the way to secure it, to produce his act of betrayal. To the extent that what Genet* does counts as the outcome of reasoning from the value he sets on being alone and his belief about how to ensure this, it counts as behavior that is reasoned activity. To the extent that what Genet* does counts as the outcome of his goal in action (which expresses his value but is not itself a topic of consideration) and his belief about how to achieve it, it counts as behavior that is purposive activity. Considered all up, Genet* counts as someone who seeks to realize solitude, because he values it, by betraying his friend to the cops, because he privileges the betrayal of intimacy as the way to solitude.

Lovers of Success in Action and Sensemaking Consider now, Akrat. The proposal is that we can make sense of pathological agents like Akrat, and perverse agents like Satan, simply as agents of reasoned activity who are lovers of success in action. So, suppose we explain Akrat’s act like this: Akrat tastes the peppered wine because he considers any act to be good (a success) in terms of satisfying his present fancy, in so far as it’s an act that affords a novel taste, and he thinks that tasting the wine laced with pepper will afford novelty. In a case like this too, we give it to be understood that the evaluative explanation is matched by a motivational one: we imply, in this example, that Akrat is acting to satisfy his present fancy for novelty because he is a lover of success in action with respect to that fancy. The pragmatics here again reflect the process whereby the agent comes to act. And once again, this process can be represented intuitively in a way that is congenial to cautious expressivism about values. Akrat is a monist about the good: he counts pleasure alone as good, and so his theory of the good doesn’t include mere novelty. However, falling in with a passing fancy, he judges that any act will be good (that is, a success) in relation to that fancy to the extent that it yields novelty of taste. This (merely formal) evaluation is then expressed in a desire to satisfy the fancy, which combines with his belief that tasting the peppered wine will afford a novel sensation, to prompt his tasting that wine. To the extent that what Akrat does can be construed as the outcome of reasoning about how to be successful in action with respect to the fancy that has currently taken his attention, it counts as behavior that is reasoned activity. To the extent that what Akrat does can be regarded as the result of his desire to satisfy his present fancy and his means–end belief, it counts as behavior that is

Perverse Agency

71

22

purposive activity. Putting these strands together, Akrat counts as someone who seeks to satisfy a present fancy for a novel taste, because he prizes any act that does satisfy that fancy as a success in relation to it, by tasting the wine laced with pepper, because he thinks that doing that will provide a novel taste. Let’s suppose (as seems plausible) that Akrat is right about his present fancy. In that case, there’s a third dimension to the explanation of what Akrat does. So far, we have it that Akrat’s tasting of the peppered wine is a piece of second-order purposive activity: he does the tasting in an attempt to satisfy his current fancy for novelty. However, if Akrat isn’t mistaken about what’s currently taking his fancy, we can also represent his tasting of the wine as a piece of first-order purposive activity: he does the tasting in an attempt to sample a novel taste. Of course, agents can be mistaken about their desires. Perhaps even Akrat is kidding himself in thinking that he is interested in novelty; maybe it’s all just part of a self-deceptive fantasy that he can break out of the mold of a life ruled by pleasure. In that case, we would have to withdraw the third explanation of his tasting the wine laced with pepper. Akrat would be just a misguided lover of success in action, who acts on behalf of a desire to satisfy an interest that he doesn’t really have.

Is Satan Irrational? Like Akrat, Satan and perverse agents are not, and are not to be understood as, lovers of the good. They are simply lovers of success in action who act on desires to satisfy the ends-in-view they take as their cues for action. Of course, there are important differences between Akrat and someone like Satan. When Akrat acts on the consideration of novelty, he acts on a consideration that he sees as not seriously supporting what he does. Whereas when Satan tries to corrupt the innocent because it’s evil, he acts on a consideration that, in terms of all that he seriously values, mandates that he not act in that way; he acts on a ‘dysjustfying’ consideration.23 But are agents like Akrat and Satan irrational? Blackburn doesn’t see the point of calling Satan irrational. He writes: Doing what you know to be bad is bad. We might describe it as irrational … But that is not, as it stands, a very interesting thing to say, for it is not at all obvious what further or specific kind of charge it makes. Satan’s pride got him where he is, but his pride is not inexplicable or unintelligible. It is not as if he insists that 7 + 5 = 13, or that he is a teapot. In fact, Milton’s poem is so great because Satan is not only intelligible, but not wholly unadmirable. His reaction to God’s suffocating superiority makes sense. … Rebelliousness takes courage. Similarly, jealousy may be a very bad emotion to feel, but it is not clear what is meant by saying that it is irrational to feel it. It makes sense, in human terms, as a response to fear of loss or betrayal, even if heavenly minds are clear of it. Satan

22 Akrat’s desire to satisfy his present fancy expresses the merely formal value he sets on any act that satisfies that fancy, but is not itself a topic of consideration. 23 Stocker’s term. See ‘Desiring the Bad,’ Journal of Philosophy, 76 (1979): 746.

72

Values and the Reflective Point of View is not irrational, but bad. And if that is true of Satan, it is still more so of the more mundane cases of human weakness: we are depressed, we succumb to temptation, we even make little attempts to break the moral mould. We are not irrational, but weak. Or, if ‘rational’ implies boundless strength of mind in pursuit of the good, say that we are irrational. But we are perfectly intelligible, and may deserve sympathy, and if our characters require improvement, it is not usually in the direction of better calculation or logic.24

Now, there’s a problem here. There’s no doubt that we can make sense of Satan’s action, up to a point. Thus, we, like Satan, can explain what he does as an agent of reasoned activity: he’s a lover of success in action with respect to a cluster of dark aims. And we can all make sense of what he does as a purposive agent, in so far as he has these aims and wants to realize them. Moreover, we can accept that there’s something admirable about Satan’s courage in defiance; and we can agree that there’s a psychological story to be told about how Satan came to have the practical and affective stance toward his situation that he does. Even further, we can set aside the idea that rationality is fortitudo moralis as unhelpful; and we can agree that, at least with regard to getting what we are after, there’s typically nothing wrong with our powers of ‘calculation or logic’ when we are being weak in more mundane, human ways. However, there’s still a limit to the intelligibility of what Satan does, as there is to what Akrat does; neither is ‘perfectly intelligible.’ For, after all, both Akrat and Satan act against their own better judgment; and that’s surely a failure of coherence of some kind. Moreover, when Akrat and Satan act against their better judgment, neither has a reason for what he does under that description. Each has a reason for what he does in the sense that (as he sees it) the act is a success with respect to achieving some current end-in-view; but neither has a reason for what he does qua act against his own better judgment. And finally, when Akrat and Satan act against their own best lights, neither has a reason for what he does in terms of anything that he seriously values; and Satan, indeed, as I have already remarked, acts on a consideration that, in terms of all that he seriously values, forbids the action. So, there is, at the heart of what each of Akrat and Satan does, ‘something essentially surd.’25 If we ask whether these agents are irrational, the answer surely is: well, yes; they are, to a degree, if the issue is intelligible patterns of thought and action.26

24 Ruling Passions, p. 65. 25 Davidson’s phrase. See ‘How is Weakness of the Will Possible?,’ in Essays on Actions and Events, p. 42. 26 It’s a further issue whether we need to invoke division in the mind to accommodate such lapses in intelligibility. Davidson holds that we do. I discuss his theory of mental division in the Appendix, arguing that it ignores the different ways in which we can make sense of what people think, or think and do.

Perverse Agency

73

Perverse Identification I hold that we can make sense of agents like Satan and Akrat as agents of reasoned activity, even though they are not lovers of the good. For we can make sense of them just as agents who are after success in action with respect to some current end(s)-inview. We can make local sense of them in this way, just as they can, even if, globally considered, there is ‘something essentially surd’ in what they do. But do Satan and Akrat ‘identify with’ their ends-in-view if they are just lovers of success in action? And if they don’t, are their actions really intelligible in the way suggested after all? Let’s consider Satan and Akrat in turn. Why might someone think that Satan can’t identify with the desires that he treats as ends in his reasoning as a lover of success in action? Here’s one suggestion: just because he supposedly doesn’t regard what those desires are for as good in any way. Joseph Raz seems committed to this kind of view. He casts the issue of whether we see (say) a desire as ‘ours’ in terms of whether we are active with respect to it; and the test of whether we are active with respect to a desire is whether, in our eyes, it is responsive to (and so made intelligible by) reasons.27 So, by these lights, if Satan doesn’t really see anything good about evildoing and what it furthers, then he can’t own the desires he treats as ends in his reasoning as a lover of success in action. There are two points to make about this. First, all anyone who supports this account of what it takes to own a desire can draw from the conclusion that Satan doesn’t identify with his ends-in-view is what they have already accepted in applying the test for whether he does – namely, that those ends aren’t intelligible to him as desires for goods, and so, his actions aren’t intelligible, to him or anyone else, as the pursuit of the good. It’s an altogether further issue whether Satan or anyone else grasps anything at all by conceiving of him as an agent who reasons from certain desires as ends, even though he doesn’t see those ends as backed by the good. And obviously something is grasped. To know that you are doing evil because (as you see it) that’s the way to succeed in satisfying your desire to pay God back is to have a better grasp of what you are up to than just knowing that you are doing evil but can’t think (even in this thin sense) why!28 The second point is this. The account of identification in terms of (believed) responsiveness to reasons or values doesn’t look promising anyway. For it seems that identification has its vicissitudes: that we can back our desires against our values.

27 Joseph Raz, ‘When We Are Ourselves: The Active and the Passive,’ Proceedings of the Aristotelian Society, supp. vol. 71 (1997): 224–7. 28 And knowing what desires set the standard of success for what you do might be able to be fitted into a larger story about how things are with you. Thus, Satan and the rest of us can place his ends-in-view in the context of his emotions: his desire for the pleasure of bringing down the innocent expresses his envy, and so on. And then, in turn, we can all place his emotions in terms of his character and circumstances: his envy is his reaction to an acute sense of loss of all genuine goods, and so on.

74

Values and the Reflective Point of View

Gary Watson raised this possibility in his contribution to the debate on identification. In his 1975 paper ‘Free Agency,’ Watson takes issue with Harry Frankfurt’s proposal that identification should be understood in terms of hierarchies of desire: (roughly) that we identify with desires that we want to control what we do.29 The problem, Watson argues, is that no mere appeal to levels of desire can secure ownership of a desire at the first order.30 The solution, he suggests, is to treat our system of values as constituting our standpoint, and to say that we identify with our desires to act where they concern actions that we see as favored by our values. However, in his 1987 paper ‘Free Action and Free Will,’ Watson withdraws this account of identification because he concedes that we can be alienated from our values, identifying, in ‘perverse’ cases, with desires to act in ways that we don’t regard as best.31 Various writers on identification have agreed that we need an account that allows for its vicissitudes.32 However, it might be thought that these concerns about the vicissitudes of identification don’t quite get us to Satan. For they, it might be claimed, are about how we can identify with desires to act in ways that we see are at odds with what we think best.33 In the case of Satan, however, the issue is whether, considered as a mere lover of success in action, Satan can identify with his ends-in-view, given that these are desires to act in ways that he doesn’t regard as good in any respect whatsoever.34 There is, however, a fairly persuasive story to be told about how we can allow that Satan does identify with his perverse ends-in-view. The story draws on Michael Bratman’s account of identification. In a seminal discussion, Bratman argues that to identify with a desire is (roughly) to decide to treat it as setting an end, where that decision doesn’t conflict with other decisions, intentions or polices about which desires are to be treated as end-setting.35 This account apparently allows for agents 29 Watson is responding to Frankfurt’s proposal in ‘Freedom of the Will and the Concept of a Person,’ reprinted in The Importance of What We Care About (Cambridge, 1988), pp. 11– 25. Actually, Frankfurt’s account allows for the iteration of levels of desire: we identify with first-order desires to act that a relevant, highest-order desire favors as our will (pp. 21–2). 30 See ‘Free Agency,’ Journal of Philosophy, 72 (1975): 218–19. 31 ‘Free Action and Free Will,’ Mind, 96 (1987): 149–51. 32 Michael Bratman is a notable example. See his comments, ‘Identification, Decision, and Treating as a Reason,’ in Faces of Intention (Cambridge, 1999), pp. 189–90, 205–6. 33 Raz, for his part, however, shouldn’t welcome any concession towards our being able to identify with desires that we see are at odds with what we think best. For it’s not enough, by Raz’s lights, for a desire to belong to our active side (and so, for us to identify with it) that we see it as backed by the good. It needs to be, in our eyes, ‘proportionate to reason’ (‘When We Are Ourselves: The Active and The Passive’: 223–7). 34 Watson himself doesn’t absolutely rule out the possibility of being ‘fully behind’ a desire to act or action where we don’t see the action as good in any way. He says that perhaps in such cases we must see the action as good (‘Free Action and Free Will’: 150). 35 ‘Identification, Decision, and Treating as a Reason,’ in Faces of Intention, pp. 195– 204.

Perverse Agency

75

to endorse desires, regardless of whether they see the actions those desires are about as good or best, in any serious or substantive way. All that matters is that the agents decide to treat the desires as ends-in-view in their practical reasoning, in the absence of any general policy to the contrary.36 If we apply this account of identification to the Satan case, it becomes plausible to hold that Satan does identify with the bleak desires that characterize him as a lover of success in action. We know how it is with Satan. Blackburn vividly characterizes his state of mind this way: Satan is racked with all the ‘foul distempers’ going. … It is his perpetual curse that he remembers how things could have been otherwise (he and God, representing the Good, could have still been together, as they once were); Milton’s whole poem describes how his resolution to do evil is a response to this nightmarish predicament. It is a reaction against an acknowledged, internalized, active set of concerns which align with those of the good. This is why his predicament is one of rebellion not only against the good, but against a part of himself.37

Satan is sorely in conflict with himself. And, in his torment, it helps him to segregate that ‘part of himself’ that remains susceptible to the good by deciding to treat as ends his desire to do evil and the desires that evildoing serves – namely, the desires that mark him out in his pride, envy, ambition, and so on. And Satan can complete his identification with these ends because, even though still under the influence of the good in some measure, having lost it, he hates it, and his hate has displaced any commitments he once might have had not to treat these desires as ends. And what of Akrat? Well, just suppose you accept that Akrat can’t own his desire for a novel taste because he doesn’t take novelty seriously as a good. The claim that Akrat can’t own his desire for novelty doesn’t then add any weight to the charge that what Akrat does in going for the wine laced with pepper isn’t intelligible beyond what’s already been assumed in deciding that Akrat can’t endorse his desire for novelty; namely, that he identifies with that desire only if he finds it intelligible as a desire for a good. It’s a further question whether Akrat or anyone else understands anything at all by construing him as an agent who treats his desire for novelty as an end in his reasoning, even though he doesn’t regard novelty as a good. And, once again, it’s clear that this does afford some (even if thin) insight into Akrat’s agency. For it’s one thing to know that you are trying something out for the sake of a particular passing fancy (even if you don’t see what the fancy is for as a serious good) and quite another not even to know what the fancy is that has prompted you to act. Can Akrat identify with his passing fancy? On Bratman’s model, we don’t have to suppose that he doesn’t. Akrat counts as identifying with the fancy he treats as an 36 In what follows, I apply Bratman’s early account of identification to the cases of Satan and Akrat. I comment on some later developments within Bratman’s account on pages 76–82 below, relating these back to Satan and Akrat. 37 Ruling Passions, pp. 62–3.

76

Values and the Reflective Point of View

end wherever he decides to treat it as such and doesn’t have a relevant counter-policy, like one that tells him to avoid treating whims as ends, given the risk that following them can pose to the pleasant life. Here’s a world in which those conditions hold: Akrat values pleasure alone. However, he is fascinated by his own whimsicality and this narcissism modifies the way he holds pleasure as a value: the value he sets on pleasure isn’t matched by a policy of avoiding whims as ends, given the risk they pose to leading the good life. In such a world, when Akrat decides to treat his fancy for novelty as an end he will count, on the model being applied, as identifying with it. So, we can think up examples where Akrat owns his passing fancy. However, it makes no difference to whether Akrat is intelligible as an agent of reasoned activity, in so far as he is a lover of success in action, whether or not he does identify with his passing fancy in these terms. Either way, we all have this much insight into his agency: he is drinking the wine laced with pepper because he sees that as the way to succeed with respect to his desire for novelty.

Ends-in-view and Identification It has to be conceded that, in his initial account of identification, Bratman is less than fully clear about how we are to understand the notion of treating a desire as setting an end. He proposes to give an account of a central case of treating a desire as reason-giving, in a modest sense of reason-giving, in terms of the notion of treating the desire as end-setting – ‘where to treat it as end-setting is, in part, to treat it as potentially justifying, at least to some extent, [the] performance of relevant means and/or relevant preliminary steps.’38 It is notable that, as Bratman conceives matters, an addict who is critical of his desire for a drug, might none the less treat, or decide to treat, that desire as reason-giving, in the modest sense of treating it as end-setting; and, if the addict does decide to treat his desire as reason-giving in this modest sense, proceeds to do so, and has no policy to the contrary (because, say, he is resigned or depressed), then, in the terms of this account, he does identify with the desire, even granting that ‘he still sees [it] as criticizable.’39 This leaves matters unclear. In what sense does the addict, in such a case, decide to treat his desire for the drug ‘as potentially justifying, at least to some extent’ his going about satisfying it? In my view, the best way to construe the sense of justification relevant to decisions to treat desires as end-setting is in terms of the notion of merely formal value. Let’s briefly revisit the relevant distinction. To count an act as good in a merely formal way is to count it as good in so far as it realizes (or conduces to the realization of) a current end-in-view. Goodness in action, in this sense, is grounded in success in action (or the promotion thereof) with respect to a present desire. On the other 38 ‘Identification, Decision, and Treating as a Reason,’ in Faces of Intention, p. 198. 39 Ibid., pp. 199–200, 205.

Perverse Agency

77

hand, to count an act as good in a serious or substantive way is to count the act as good in terms that are independent of its being (or conducing to) the fulfilment of any current end-in-view.40 Goodness in action in a serious or substantive sense is end-justifying: we justify our ends-in-view by reference to their (in this sense) goodmaking features. The suggestion I am making is this: where agents decide to treat desires as endsetting, we should interpret them as deciding to treat desires as value-conferring in a merely formal sense. On this way of understanding matters, to decide to treat a desire to do A as end-setting is (roughly) to form a commitment to treat, in one’s relevant practical reasoning, doing A (or its promotion) as good in so far as that satisfies (or helps to satisfy) the desire to do A. So construed, the sense in which we settle on our desires as justifiers when we treat them as setting ends in our practical reasoning never as such introduces substantive or serious valuing into the picture. In the case of the addict, this interpretation of what’s involved in the addict’s decision to treat his desire as end-setting leaves open a moral psychological possibility that we surely need to leave open – namely, that he regards his addiction as wholly bad (even if, resigned or depressed, he identifies with it).41 Similarly, the approach enables us to continue to regard Satan as a perverse agent, even where he decides to treat his dark desires as end-setting, and to continue to regard Akrat as a monist about the good, even where he decides to treat his whim for novelty as end-setting. And in this way, if we apply Bratman’s account of identification, the construal of these ends-in-view as figuring within the endorsed reasoning as justifiers in a merely formal sense enables us to allow, as seems very plausible, that Satan and Akrat, without shedding their values, might identify with motives that oppose those values. Bratman’s early account of identification with a desire appeared in the paper ‘Identification, Decision, and Treating as a Reason.’ In several later papers he has developed the account and situated it in a broader context of issues, such as self-

40 It’s worth noting that we can ascribe value in a serious or substantive way to acts that we see as satisfying our current desires. Suppose I judge that any act is good to the extent that it satisfies my present desire to do A. Suppose that I take the ground of goodness in action here to be just the feature that the act represents the satisfaction of my current desire. In that case, my judgment is an instance of merely formal valuing. But suppose I judge that any act is good in so far as it promotes my own welfare, to the extent that it satisfies my present desire to do A. Suppose, further, that the focus in this case is the promotion of my own welfare, considered without regard to whether it is itself among my present objects of desire. In that case, my judgment expresses a serious or substantive value. 41 Compare Smith’s comments on what he calls ‘the Ayer-Frankfurt-Watson cases,’ in The Moral Problem, pp. 133–4.

78

Values and the Reflective Point of View

governance as a temporally extended agent and autonomous action.42 In what follows I wish to focus on one central aspect of the developed account.43 As he extends his account of identification, Bratman seeks to specify more exactly the content of the higher-order commitment or policy that figures in identification with a desire. He favors a specification of content that combines two elements: (i) endorsement of the desire as a motive, and (ii), as he puts it, ‘endorsement of forms of deliberation and practical reasoning for which the desire sets an end, an end which is treated as justifying appropriate action.’44 With respect to (ii), Bratman wishes to allow that a desire might motivate by being end-setting for practical reasoning in either of two ways: it might set an end for reasoning that ‘foregrounds’ the desire and treats what the desire is for as desirable relative to the desire; or it might set an end for reasoning that doesn’t foreground the desire and which effectively represents

42 See ‘Reflection, Planning, and Temporally Extended Agency,’ The Philosophical Review, 109 (2000): 35–64, and ‘Autonomy and Hierarchy,’ Social Philosophy and Policy, 20 (2003): 156–76. 43 Some other significant aspects of the developed account are the following: (1) In the initial discussion, Bratman emphasizes the role of higher-order decisions in identification with a desire. However, in later discussion, what is reckoned to be essential to identification with a desire is not a decision, but a relevant commitment or policy to treat the desire as reasonproviding, where such a commitment or policy is a plan-like attitude which we may or may come to have through a decision (‘A Desire of One’s Own,’ Journal of Philosophy, 100 (2003): 233). (2) Bratman makes it clear that, in his view, the commitments relevant to identification can be singular as well as general (‘Hierarchy, Circularity, and Double Reduction,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt, pp. 78–9). (3) Bratman holds that, in addition to self-governing policies about the shape of our practical lives, there are self-governing quasi-polices. These are higher-order concerns about the shape of our practical lives which are not subject to strong demands for consistency in the way that polices are (‘Reflection, Planning, and Temporally Extended Agency’: 57–60). (4) Bratman changes his interpretation of the so-called ‘satisfaction’ or ‘wholeheartedness’ condition. In the early discussion, his claim is that, to identify with a desire, we must not be divided in our wills in a certain way: our decision to treat a desire as reason-giving must not conflict with other decisions about which desires to treat as reason-giving (‘Identification, Decision, and Treating as a Reason,’ in Faces of Intention, pp. 200–201). In later discussion, Bratman shifts the emphasis from mere division in our wills with respect to a commitment about how to treat a desire in our practical lives to division which challenges that commitment. A higher-order commitment C is said to be challenged by another higher-order commitment C* where C* conflicts with C and, as a result, tends to undermine the role of C as a selfgoverning policy across time (‘Reflection, Planning, and Temporally Extended Agency’: 49–50). Bratman further refines the notion of satisfaction to take account of both policies and quasi-policies (ibid.: 58–60). 44 ‘Reflection, Planning, and Temporally Extended Agency’: 52. Bratman also favors a content that is self-referential: the policy supports the desire’s functioning, by way of that very policy, as end-setting for practical reasoning. See ‘Hierarchy, Circularity, and Double Reduction,’ p. 77, and ‘Autonomy and Hierarchy’: 173–5. This is a fifth significant aspect of Bratman’s developed account of identification.

Perverse Agency

79 45

what the desire is for as desirable whether or not it is desired. In my terms, Bratman wishes to allow that a desire might motivate by providing a justifying end either for reasoning that in its premises ascribes merely formal value to an end or for reasoning that in its premises ascribes serious or substantive normative value to an end. It’s clear that we can readily enlist the idea that identification with a desire involves having a commitment or policy in support of the role of the desire as providing a justifying end in motivationally effective practical reasoning in a merely formal way in order to finesse an account of how Satan and Akrat might identify with their motives.46 Such a move is therefore robust against a complaint by Frankfurt against Bratman’s approach to identification with a desire. Bratman, Frankfurt charges, wrongly supposes that identifying with a desire involves treating the desire as providing a justification for action, whereas ‘to identify with a desire means merely that – for whatever reason, or for no reason whatever – [someone] joins himself to the desire and accepts it as his own.’47 Frankfurt’s chief concern here seems to be to preserve the possibility that ‘people may identify with desires that they do not regard as meritorious and even with desires of which they disapprove.’48 And, as we have seen, there is a way to preserve this possibility while allowing that, in identifying with a desire, a person comes to be committed to treating the desire as reason-providing in their motivationally effective reasoning. We do this if we suppose that the person becomes committed to treating the desire as valueconferring in such practical reasoning in a merely formal sense.49 45 See Bratman’s discussion of motivation by desire by way of Model 1 or Model 2 reasoning in ‘Autonomy and Hierarchy’: 161–3. The idea is that, in identifying with a desire, we have a commitment or policy in support of the role of that desire, or what the desire is for, as providing a defeasible justifying end in motivationally effective practical reasoning. This preserves a point Bratman made in the early paper, ‘Identification, Decision, and Treating as a Reason’ (in Faces of Intention, pp. 195–8): we can identify with a desire even though, on some particular occasion, we decide not to act on it. 46 I leave aside the further detail here. A full account of identification with a desire, taking Bratman’s contribution as its inspiration and point of departure, would need to consider other aspects of his mature story, as noted in notes 43 and 44 above. It also, in my view, would need to draw a distinction that Bratman does not make. This is the distinction between identifying with a desire, considered as reason-providing (such as we have in the cases of Satan and Akrat), and identifying with a desire, considered as reason-tracking. I elaborate on this distinction in the main text below. 47 ‘Reply to Michael E. Bratman,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt, p. 89. Frankfurt is replying to Bratman’s paper ‘Hierarchy, Circularity, and Double Reduction’ in the same volume. 48 ‘Reply to Michael E. Bratman,’ p. 89. 49 It’s true that, in his reply to Bratman, Frankfurt also seems to be criticizing the idea that we should regard identification with a desire as committing us to any mode of practical reasoning. See ‘Reply to Michael E. Bratman,’ pp. 89–90. However, in his ‘Reply to Richard Moran’ (same volume), Frankfurt, speaking of identification, comments that ‘[t]he creation of reasons is, indeed, close to being the whole meaning of it’ (p. 219). Frankfurt clearly means this in the present context in the following way: we can, in virtue of identifying with a desire,

80

Values and the Reflective Point of View

In his developed account of identification with a desire, Bratman wishes to allow that a desire might motivate by providing a justifying end in practical reasoning in either of two ways. On the one hand, a desire might motivate by providing a justifying end in reasoning that treats what the desire is for as desirable with respect to (the satisfaction of) that desire. On the other hand, a desire might motivate by providing a justifying end in reasoning that represents what the desire is for as desirable anyway (that is, considered independently of whether it is desired). Specifying these different paths along which a desire might motivate by way of practical reasoning respects the fact that we may, but need not, foreground our desires when making our minds up about what to do. In this way, Bratman frees his account of identification with a desire from a mistake about practical reasoning that has bedevilled such accounts. This is the so-called ‘objectifying mistake’: the mistake of supposing that, as practical reasoners, we invariably take our desires as our topics in practical reflection. I discuss this mistake in Chapter 5. There is, however, a curious aspect to Bratman’s developed account of identification. It has to do with the idea that one way a desire might motivate is by providing a justifying end in practical reasoning that casts what the desire is for as desirable anyway. Bratman puts it like this. In such cases, ‘my reasoning has as its major premise an appropriate expression of my desire, or of a thought involved in my having that desire’; and, to illustrate, he cites reasoning that has as its major premise a premise like ‘Revenge is a justifying consideration.’ It’s evident that Bratman would also accept, as an example of a relevant major premise, one like ‘Revenge is desirable,’ or some fine-tuned variation on that.50 The inspiration behind Bratman’s story, at this point, is Davidson’s theory of practical reasoning and its construal of the desires that combine with beliefs to yield action as attitudes that have evaluative premises as their ‘natural propositional expression.’51 However, Bratman, unlike Davidson, does not assimilate desires to values, and so commit the ‘desire-as-valuing mistake,’ a mistake already considered on pages 66–8 above and which I discuss further in Chapter 5. Bratman, for his part, concedes that the connection between a desire and the action it motivates ‘may or may not involve a conception, on the part of the actor, of the desired end as justifying.’52 Even so, Bratman’s developed account of identification still seems to be tainted by the desire-as-valuing mistake. In cases where a desire motivates by way of our reflection on it as value-conferring in a purely formal sense, it seems correct to say that the desire motivates by way of providing a justifying end for practical thinking. come to treat it as a reason even though we don’t regard what the desire is for as desirable in any way and may even regard it as undesirable (see pp. 223–4). I do not know whether Frankfurt would be sympathetic to the suggestion that the sense in which a desire as such may be counted as a reason is best thought of as purely formal. 50 ‘Autonomy and Hierarchy’: 162. 51 See ‘Autonomy and Hierarchy’: 163, note 21, and ‘Hierarchy, Circularity, and Double Reduction,’ pp. 66–7. 52 Ibid., pp. 67–8.

Perverse Agency

81

Consider, by contrast, cases where a desire that functions as an effective motive involves, on the part of the agent, an associated thought that what the desire is for is desirable in some way, conceived independently of the presence of the desire. A central example here would be a situation in which the desire on which the agent acts is a value-based desire. In such a case, it seems confused to suppose that the desire motivates by way of providing a justifying end in practical reasoning. What is surely correct, rather, is this: the substantive value on which the desire is based (which represents what is desire is for as fit to be an end) motivates (in part) by way of the desire that it (subjectively) justifies and sustains. Of course, within the terms of Davidson’s theory of practical reasoning, it is natural to say that, wherever desires combine with beliefs to yield actions, they motivate by providing justifying ends for practical reasoning. However, that’s because the theory misconstrues desires as values. What does this mean for a satisfactory account of identification with a desire? To my mind, it means that we need to allow for a contrast between identifying with a desire, considered as a reason-providing motive, and identifying with a desire, considered as a reason-tracking motive. In what follows I sketch a proposal for doing this which draws on the insights assembled so far. We identify with desires, considered as reason-providing motives, just in case we identify with their role as effective motives by way of reasoning that treats them as value-conferring in a purely formal sense. This is the mode of identification that enables us to explain how Akrat and Satan might identify with their motives without abandoning their conception of the good. On the other hand, we identify with desires, considered as reason-tracking motives, just in case (i) we identify, in a reason-providing way, with the values on which those desires are based, and (ii) we identify with the role of the desires as effective motives, considered as motives that mediate the influence of the values that sustain them. If this is along the right lines, it shows that it’s identification in a reason-providing way that is basic in the story of identification with a desire. This is the mode of identification relevant where, as mere lovers of success in action, we identify with the motives that we treat as sources of (merely formal) value. And, where we identify with desires, considered as reason-following motives, we must first identify, in a reason-providing way, with the values on which those motives are based (which those motives express). That is, we must first identify with the role, in shaping our practical lives, of reasoning about what to desire (and do) which treats those values as reason-providing.53 This prior form of identification with values is essential in cases of identifying with desires in a reason-tracking way because, while it’s surely correct that identifying with value-based desires must involve identification with the values on 53 I have introduced the notion of identifying with a desire in a reason-following way by focussing on desires that are based on substantive values. However, the full story here will need to accommodate the complication that we can identify in this way with desires that track merely formal values.

82

Values and the Reflective Point of View

which those desires are based as reason-providers, it can’t be assumed that, where our desires do stem from values, that we identify with those values in this way. For sometimes people form desires or preferences ‘in the grip’ of values that serve as reason-providing in their motivationally effective practical reasoning without their having any higher-order commitment or policy that endorses them in this role.54 This last point bears on the relation between identification and self-knowledge. In the terms of the present discussion, it turns out to be possible for someone to selfascribe a desire in a purely first-person way without identifying with it, considered as a reason-tracking motive, because they do not identify, in a reason-providing way, with the value(s) which they rely on to make the self-ascription.55 Indeed, someone might not identify with a desire that they first-personally self-attribute because, in making the ascription, they rely on reasoning about what to desire which invokes values that they actually eschew as reason-providing, as a matter of policy. For example, in a case of sour grapes, someone might conclude that they want a situation to be the way they believe it has turned out, appealing to values which they have a (conveniently concealed or less salient) policy of not giving any weight to in their motivationally effective practical reflections.56

54 Bratman discusses such cases, referring to Alan Gibbard’s distinction between ‘accepting a norm’ and ‘being in the grip of a norm.’ See ‘Hierarchy, Circularity, and Double Reduction,’ pp. 74–5, and ‘Autonomy and Hierarchy’: 172. 55 Where someone does identify with the value(s) which they we rely on to self-ascribe a desire in a purely first-person way, they will be under a constraint of coherence to identify with the desire, considered as a reason-tracking motive. 56 Moran defends a different account of identification and of the relation between it and purely first-person self-knowledge. In the book Authority and Estrangement, he construes identification with a desire (in line with a more general account of identification with a mental state) as, in a certain sense, reason-tracking and he restricts identification to desires which we know about in a ‘transcendental’ – in my terms, purely first-person – way. Some desires – so-called ‘brute’ desires, like the desires of hunger, thirst and lust, which are never based on reasons – are not even candidates for identification in the story (see Chapters 3–4). However, a concession that Moran makes to Frankfurt elsewhere – in ‘Frankfurt on Identification: Ambiguities of Activity in Mental Life,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt, pp. 212–14 – is in tension with this account of identification with a desire. For Moran there concedes to Frankfurt that we typically do identify with the desires that comprise love and care – presumably, desires that we can consciously have – even though we typically don’t reason ourselves into them. We identify with such unreasoned desires (which we are in no position to self-ascribe transcendentally) in a reason-providing way.

Chapter 5

Two Mistakes about Practical Reasoning Questions about value provide us with a focus in thought and talk about what to desire, will, feel and do, as well as with a focus in thought and talk about what we do desire, will, feel and shall do. But what kinds of consideration do we bring to bear in trying to answer questions about value? What are the topics of our reflection? In particular, what are the topics of our reflection when we try to settle a question of value in an attempt to make up our minds about what to do? We are sometime lovers of the good and sometime mere seekers of success in action. As mere seekers of success in action, we reason about our desires and how to satisfy them, thus foregrounding our desires in our deliberations as ends-in-view. As lovers of the good, we may also foreground our desires. Thus, we may count a course of action as good in a serious or substantive way – it will help us to flourish (say) – because it satisfies a desire or ambition that is central to our sense of who we are.1 However, on some accounts of the deliberative setting, our desires are always salient among the topics that we consider when we reflect about what to do. Our practical reflections, in other words, are invariably self-preoccupied in this way. Desires as Topics of Reflection Christine Korsgaard holds that we have normative problems because our minds are self-conscious (which she calls ‘reflective’) in structure. She writes: A lower animal’s attention is fixed on the world. Its perceptions are its beliefs and its desires are its will. … But we human animals turn our attention on to our perceptions and desires themselves, on to our own mental activities, and we are conscious of them. That is why we can think about them. And this sets us a problem no other animal has. It is the problem of the normative. For our capacity to turn our attention on to our own mental activities is also a capacity to distance ourselves from them, and to call them into question. I perceive and I find myself with a powerful impulse to believe. But I back up and bring that impulse into view and then I have a certain distance. Now the impulse doesn’t dominate me and now I have a problem. Shall I believe? Is this perception really a reason to believe? I desire and I find myself with a powerful impulse to act. But I back up and bring that impulse into view and then I have a certain distance. Now the impulse doesn’t dominate me and now I have a problem. Shall I act? Is this desire really a reason to act? The reflective mind cannot settle 1 See note 40, Chapter 4.

Values and the Reflective Point of View

84

for perception and desire, not just as such. It needs a reason. Otherwise, at least as long as it reflects, it cannot commit itself or go forward. … We need reasons because our impulses must be able to withstand reflective scrutiny. We have reasons if they do. The normative word ‘reason’ refers to a kind of reflective success.2

And Korsgaard has an account of when our desires enjoy reflective success: they become normative reasons where we (as active wills) endorse acting on them under some self-given law or principle of choice. Needless to say, this account of normativity is Kantian in its inspiration.3 Korsgaard’s account is striking in the way it pictures the deliberative setting: when we deliberate about what to believe or to do, our topics of consideration are our own psychological states – our perceptions or our desires. Now, this is an odd idea. It doesn’t seem right at all. Even if some of our evidence is about how the world presents to us, most of it is about how the world is; and, even if we sometimes explicitly take account of our desires when deciding what to do, very often we focus on features of the world around us that matter to us in some way. Korsgaard herself can seem ambivalent about the topics of reflection. She contrasts her own moral psychology with Kant’s in that, whereas Kant thinks that there are two principles of choice that may govern our actions (self-love and the moral law), her own account allows for a plurality of principles. However, she marks her agreement with Kant on what counts as the reason for the action. She comments: Neither the incentive nor the principle of choice is, by itself, ‘the reason’ for the action; rather, the reason is the incentive as seen from the perspective of the principle of choice. That you desire something is a reason for doing it from the perspective of the principle of self-love. From the perspective of the moral principle, however, it is only a reason for doing it if the maxim of doing it passes the categorical imperative test. In my own account, the principle of self-love is replaced by the various principles associated with our contingent practical identities. That Susan is in trouble is a reason for action from the perspective of Susan’s friend; that the law requires it is a reason for action from the perspective of a citizen, and so forth.4

There’s an evident shift in this passage in relation to incentives. At the outset, incentives are desires. By the end, incentives have transmuted into features of the world that matter. This ambivalence persists in Korsgaard’s attempt to elucidate Kant’s view about the difference between the moral person who chooses to help someone in need from duty and the sympathetic person who chooses to help because he wants to. Korsgaard remarks:

2 3 4

The Sources of Normativity, ed. Onora O’Neill (Cambridge, 1996), p. 92. See ibid., pp. 97ff., 238–42. Ibid., p. 243. My italics.

Two Mistakes about Practical Reasoning

85

Kant condemns the naturally sympathetic person not for the content of his incentive, but rather for making an insufficiently reflective choice. … for both of these characters the very fact that someone is in need is an incentive and to that extent is a reason to help. But it is a different sort of reason for one who sees the needs of another as the source of a claim on him, as the source of duty, than it is for one who sees helping another merely as something he would like to do.5

Once again, it isn’t desires that are reckoned to be incentives; rather, it’s features of the world that attract or matter. Even so, Korsgaard’s central image is that our desires and impulses provide reason with its materials in practical deliberation. This is especially vivid in the way Korsgaard introduces her own account of normativity as reflective endorsement. She sees that account, with its Kantian affinities, as transposing to particular contexts of deliberation the kind of test of the normativity of our moral dispositions and sentiments that we find in Hume’s moral theory, and in Bernard Williams’s virtue theory.6 Korsgaard comments: If the reflective endorsement of our dispositions is what establishes the normativity of those dispositions, then what we need in order to establish the normativity of our more particular motives and inclinations is the reflective endorsement of those. That after all is the whole point of using the reflective endorsement method to justify morality: we are supposing that when we reflect on the things we find ourselves inclined to do, we can accept or reject the authority those inclinations claim over our conduct, and act accordingly. But what I have just described is exactly the process of thought that, according to Kant, characterizes the deliberations of the autonomous moral agent. According to Kant, as each impulse to act presents itself to us, we should subject it to the test of reflection, to see whether it really is a reason to act.7

Even in the case of desires we are in part responsible for, Korsgaard sticks to her favored picture of practical reflection, where the will or practical reason is confronted by passive desires that it must endorse or reject as reasons if a decision is to be made. According to Korsgaard, our contingent practical identities are in part given to us by what we happen to find important, given our biology, psychology and history; but, as mature agents, we also enter into their construction, actively embracing what we find important, by being laws to ourselves. In some cases, she says, the practical identities we make for ourselves will generate new desires; and so, some of the desires that confront us as decision-makers will bear the mark of our own handiwork and not just be supplied by instinct. These, she remarks, are like the cultural desires Kant discusses in his historical essays. And yet, says Korsgaard, at the moment of action,

5 6 7

Ibid., p. 244. My italics. See ibid., pp. 49–89. Ibid., p. 89.

Values and the Reflective Point of View

86

these desires or impulses are ‘the incentives, the passively confronted material upon which the active will operates, and not the agent or active will itself.’ 8

Features that Matter as Topics of Reflection Blackburn charges that Korsgaard shares Kant’s fundamental mistake about practical deliberation in supposing that the deliberative standpoint is ‘a standpoint which surveys or takes account of desire, and then weighs it at … the [independent] tribunal of reason.’9 The idea, he complains: … crucially mistakes the process of deliberation. Typically, in deliberation what I do pay attention to are the relevant features of the external world: the cost of the alternatives, the quality of food, the durability of the cloth, the fact that I made a promise. I don’t also pay attention to my own desires … My own concerns and dispositions determine which features I notice and how I react to them. If I am a miser, the cost takes my attention; if I am a gourmet, the quality of the food does; if I am prudent, the durability of the cloth; if I am not a knave, the fact of the promise. If I am extravagant, or a glutton, or concerned only with my appearance today, or if I am a knave, none of these features presents itself as important. … Deliberation is an active engagement with the world, not a process of introspecting our own consciousness of it.10

The point, according to Blackburn, can be put in terms of incentives: Kant … is habitually presented as referring to our desires and concerns as ‘incentives’, and this is exactly what they are not. An incentive is a further, external aspect of a situation that, by being the object of a further desire, helps to bend the will in some direction. A desire is not an incentive. Desire is that through which we see things as incentives.11

Korsgaard, for her part, is impressed by the way in which our very reflectiveness gives us a distance from our desires: this, she contends, is how we come to have normative problems as agents. Blackburn counters that the claim that reflectiveness puts us at a distance from our desires: … is ambiguous, and right only in a sense which does not help the Kantian. In the sense in which it is right, it means only that one can stand back from a particular desire or impulse, and accept or reject its pressure on one. Certainly we can do this, in the light of other desires and concerns. What is not thereby given is that we can do it from a standpoint independent of any desire or concern: independent of a desire for our own good, or for the happiness of humanity, or respect for this or that, or the myriad other passions that make up our individual profiles of concern and care.12

8 9 10 11 12

Ibid., pp. 240–41. Ruling Passions (Oxford, 1998), p. 253. Ibid., pp. 253–4. Ibid., p. 256. Ibid., p. 252.

Two Mistakes about Practical Reasoning

87

Blackburn calls the Kantian image which Korsgaard embraces ‘the objectifying mistake’ about practical reflection, since it supposes that we take our desires as the topics or objects of consideration when deliberating about what to do.13 The objection, as he puts it, isn’t that we never foreground our desires when deliberating; it’s that, when we do so, it’s only incidental. Blackburn writes: Of course, we do scrutinize our desires, but consider what really happens when we do. I find myself wanting to go to a smart restaurant, suppose, and then confront an imagined or real critic. We can describe the critic as raising an issue about the desire, if we wish, for, kept on a tight rein, this could be harmless – a mere translation of the fact that the critic is querying something about the object of the desire. What it really means is that the critic forces further thought about the object of the desire. What is it about going to a smart restaurant that attracts me? Have I thought of what I or my family must forgo if I go there; have I thought about the reactions of others; have I really imagined what likely pleasures or pains it will bring; etc., etc.? It is the world that we contemplate, not our own psychologies.14

By these lights, the topics which we reflect on when deliberating about what to do – the topics which ‘present themselves as reasons for or against action’ – are not our desires or impulses, but various aspects of our situation that those desires make salient and which ‘come to mind, perhaps unbidden.’15 On Blackburn’s story, passion rules all: both reason and the will are the slaves of passion. We make choices about what to do on the basis of these affectively marked representations of our situation of choice;16 and the choices that we make are the creatures of our strongest desires (or concerns).17 13 Blackburn casts Kant’s version of the objectifying mistake in terms of the figure of the Kantian Captain who is ‘the will, yourself as an embodiment of pure practical reason, detached from all desire’ (Ruling Passions, p. 246, 243–50). 14 Ibid., p. 255. 15 Ibid., pp. 122–3. 16 Ibid., pp. 125–32. 17 Ibid., pp. 132–3. Blackburn here has in mind what he calls a ‘thin’ sense of desire. In the relevant sense of desire or concern, he claims, ‘[t]he rule that the strongest concern wins is a piece of book-keeping’ (ibid., p. 125). This ‘piece of book-keeping’ matches Blackburn’s favored reading of the principle of maximizing expected utility as ‘definitional: a grid imposed upon the process of interpreting others’ (ibid., p. 161). So construed, the principle issues no empirical predictions (nothing an agent eligible for interpretation does is inconsistent with it) and it’s not a rational rule of conduct (you can’t disobey it, if you are eligible) (ibid., pp. 161–8). Blackburn doesn’t agree with Hobbes that the will is the last appetite in deliberating; but he does see the will in any particular case as hostage to our strongest desire or concern at the time (ibid., pp. 132–3). The enemy of the truth here is that image of the will encapsulated in the figure of the Kantian Captain (see note 13 above). Blackburn’s (decision-theoretic) picture of the will is, in its own way, too uncompromising. For one thing, there’s the role of decisions as tie-breakers between competing options that attract us equally. In examples like this, we decide on an option from among a set of alternatives of which there is no one alternative that we prefer to any other; and it looks as if we can then

88

Values and the Reflective Point of View

The Objectifying Mistake It surely is a mistake to suppose that our desires ‘fill the foreground’ of practical deliberation. For one thing, the phenomenology of reflection doesn’t lend any sometimes enact our decisions, without any change in the motivational tie (compare Mele, Springs of Action (New York, 1992), pp. 67ff., and Thomas Pink, The Psychology of Freedom (Cambridge, 1996), pp. 123ff.). In such examples, the will isn’t the slave of any of the desires that favor the options that it must decide between; and what makes it plausible to suppose that we can sometimes enact decisions in cases of a motivational tie without any change in the tie is just the idea that the will plays a unique role in the causation of actions. It doesn’t function, when decisions or intentions are formed, by adding to motivational strength. Rather, it settles the question of what to do in any particular case, disposing us to treat the matter as closed, with all that that involves (see Mele, Springs of Action, pp. 76–7, 130ff., 166–70, and Bratman, Intention, Plans, and Practical Reason (Cambridge, MA, 1987), Chapters 2–3). Of course, there’s the question of how we get to make decisions in cases where there’s a motivational tie. To be sure, we aren’t transmuted into Kantian Captains, in such cases, transcending all our desires. The vying options might equally engage us; but we might be under pressure to make a choice, and so prefer to choose one or other of the options, rather than leave things undecided. And this preference might prompt us to settle on a particular course of action, even though there’s something else that we want to do just as much. In such a case, we don’t transcend all our desires; yet it’s also true that the will isn’t the creature of the last appetite, in the sense of a strongest desire that favors one of the competing options. Other kinds of case challenge the idea that the will is always hostage to strongest desire. For we can surely sometimes decide to do something and then find that we are pitted against ourselves in trying to stick to the decision. What such cases typically call for are exercises of self-control in an effort to change the balance of motivation in favor of our decision. Often, what drives resistance to the suggestion that choice or decision can cut across the balance of motivation is the idea that we are bereft of any basis for ascribing strongest motivation other than choice (compare Blackburn, Ruling Passions, p. 163). However, there clearly are other indicators of desire-strength besides choice. Thus, the more we want to do something, the more we tend to notice or pay attention to opportunities to do it, to think about ways of doing it, to imagine ourselves doing it, and so on (compare Mele’s comments, Irrationality (New York, 1987), pp. 15, 83). Indeed, Blackburn actually shows a certain ambivalence in the way he treats the relation between decision and strength of desire. On the one hand, he endorses the picture we have been discussing: our decisions and intentions are the creatures of our strongest desires. On the other hand, he sometimes allows for other possibilities. Thus, he writes of examples where we learn about ourselves – about what we care about most – because of the way in which what we end up doing differs from the deed we decided on (see Ruling Passions, p. 191). What’s important here is the line between decision and deed. For, while it seems that we can sometimes decide on an option other than the one that concerns us most, it doesn’t seem that we can actually go ahead and fully execute such a decision. It seems that, in the end, cases of ties aside, our actions do reveal what we want most. This explains the point of (apparently paradoxical) exercises of self-control where our decisions are out of kilter with our balance of desire: we need to change that balance if we are to execute the decision (compare Mele, Springs of Action, pp. 72, 166ff.; for related discussion, see Mele, Irrationality, Chapter 5, and Autonomous Agents (New York, 1995), Chapters 3–4).

Two Mistakes about Practical Reasoning

89

support to the idea that we are self-preoccupied like this as practical deliberators. The situation is at least very often as Blackburn depicts it: we consider this and that, where this and that are the various features that weigh with us as we try to decide what to do. Moreover, the suggestion that we always foreground our desires, in deliberating about what to do, ignores the way in which many of the reasoned desires that we form in such contexts are indifferent to other desires as justifying ends – both temporally and modally. These desires are different from the kind of desire we form when (say), in the course of considering how to react, we come to want to touch something because we are curious to know what it feels like to touch something with that kind of texture: in such a case, the desire we form is a desire to touch the thing in question at any time or world where touching it will satisfy our curiosity about the relevant sensation. Consider, by contrast, a scenario like the following. I deliberate about whether to finish a project early and, in the course of weighing up the pros and cons, form a desire to do so because I’ve given my word that I will – a desire, let’s say, that finds expression in a decision to finish the project early for that same reason. How are we to think of this reasoned desire? In standard cases, it seems that it will be a desire that is temporally and modally indifferent to the existence of any desire on my part to keep my word. That is to say, it will be a desire to finish the project at any time or world where to do so honors the commitment I made – irrespective of whether, at that time or world, I want to keep my word. This reflects the fact that, in standard cases, the desire to finish the project that I form in deliberation will not be tracking any such inclination on my part as a reason.18 Of course, as noted at the outset, sometimes we do foreground our desires in the premises of our practical deliberations. For example, sometimes we are seekers of success in action who reason about how to achieve ends that we self-ascribe. And, indeed, in some cases, as is clear from the examples of Akrat and Satan, that’s all that’s going on: we are mere lovers of success in action. The point is instructive. Blackburn supposes that, where we do foreground our desires when deciding what to do, we do so just incidentally. However, examples of whimsical and perverse agency show that this isn’t so. Sometimes our practical reasoning really is just about how to achieve something that we covet. There isn’t any further stretch of reflection to be had about what we covet. Moreover, as noted before, sometimes we also foreground our desires when we reason about what to do as lovers of the good, because we connect the satisfaction of a desire, or set of desires, with something we see as being of serious worth.

18 The point is an adaptation of one made by Pettit and Smith in their seminal critique of the objectifying mistake, in ‘Backgrounding Desire,’ The Philosophical Review, 99 (1990): 575–8. Pettit and Smith there argue that the view that we always foreground our desires in deliberation forces us to mischaracterize ‘the scope of many of our [reasoned] desires.’ However, they focus on reasoned desires that they suppose we form when we make decisions. I am supposing, instead, that the reasoned desires that we form as we deliberate are higherorder dispositions to act which, in some cases, find expression in decisions.

90

Values and the Reflective Point of View

Even so, the point remains that we make a mistake about the content of practical reflection if it’s supposed that we always, or even typically, therein objectify our desires, foregrounding them in the premises of our deliberations. Practical reflection, instead, is plural in its content. Let’s now trace some significant ways in which the objectifying mistake has set its mark on contemporary theories of agency, especially in relation to the themes of identification and autonomy.19 The Theory of Identification Korsgaard’s account of normative reasons for action in terms of the reflective endorsement of desires that confront us in particular contexts of deliberation is tied to her account of identification and autonomy: we (as active wills) make such psychic givens our own by endorsing them as reasons under some self-given principle of choice. In this way, Korsgaard situates the phenomenon of identification with a desire within a mistaken conception of the deliberative setting. On this conception, practical reflection always foregrounds desire and – since, as persons, we need a reason to act on – it becomes an issue ‘at every turn’ in deliberation whether we endorse or identify with the desire or desires posited before consciousness.20 Identification with a desire also plays a central role in Frankfurt’s account of freedom of the will. In the initial version of that account – in Frankfurt’s 1971 paper ‘Freedom of the Will and the Concept of a Person’ – the idea is that we enjoy freedom of the will where we make a first-order desire our own by forming a relevant highestorder volition (desire) to be moved to action by it.21 The picture seems to be that, to the extent that we are reflective agents, the task before us within deliberation is to come to a point where we in this way establish one of the first-order desires that confront us as our will (as the desire that is to move us to action). In this discussion, Frankfurt (unlike Korsgaard) allows that we can foreground our desires within deliberation, and reason about what to do, without reflectively considering whether we identify with those desires. Any of us who are adult humans may be ‘more or less wanton’ and not care about which of our desires is our will; and yet, in such cases, we may be ‘rational wantons,’ reasoning about how to do what we want to do.22 In two later papers – ‘Identification and Externality’ (1977) and ‘Identification and Wholeheartedness’ (1987) – Frankfurt moves towards a revised account of what it takes to identify with a desire, which emphasizes the role of a decision that

19 Pettit and Smith discuss some of the ways in which the objectifying mistake has shaped issues and solutions in the theory of agency, in ‘Backgrounding Desire’: 578-92. For ways in which Blackburn sees the objectifying mistake as ramifying through practical philosophy, see Ruling Passions, pp. 255–9. 20 Compare ‘Backgrounding Desire’: 586. 21 This paper is reprinted in Harry G. Frankfurt, The Importance of What We Care About (Cambridge, 1988), pp. 11–25. 22 ‘Freedom of the Will and the Concept of a Person,’ in ibid., p. 17.

Two Mistakes about Practical Reasoning

91

23

terminates a hierarchical sequence of desires. In the second of these two papers, Frankfurt acknowledges the force of Gary Watson’s criticism of his earlier approach in ‘Freedom of the Will and the Concept of a Person’: a mere appeal to hierarchies of desire can’t by itself explain identification. In response, he seeks to explicate a further element that he had obscurely indicated in the early account – namely, the notion of a ‘decisive commitment’ to a first-order desire – in terms of a decision, ‘made without reservation,’ that backs a relevant highest-order desire and its associated lower-order desires.24 Frankfurt’s notion of decision here is embedded in a conception of the deliberative setting as one in which our desires provide the materials for reflection; and, in ‘Identification and Wholeheartedness,’ as elsewhere, he reneges on the concession that wantons, who do not stand back from their desires and consider whether to endorse them, can be practical deliberators.25 Our decisions, he contends, have us as their immediate object and have to do with ‘the formation and maintenance of the self.’26 These decisions contribute to the resolution of ‘two quite different sorts of conflicts between desires’ that we face as practical deliberators. Frankfurt writes: In conflicts of the one sort, desires compete for priority or position in a preferential order; the issue is which desire to satisfy first. In conflicts of the other sort, the issue is whether a desire should be given any place in the order of preference at all – that is, whether it is to be endorsed as a legitimate candidate for satisfaction or whether it is to be rejected as entitled to no priority whatsoever. When a conflict of the first kind is resolved, the competing desires are integrated into a single ordering, within which each occupies a specific position. Resolving a conflict of the second kind involves a radical separation of the competing desires, one of which is not merely assigned a relatively less favoured position but extruded entirely as an outlaw. It is these acts of ordering and rejection – integration and separation – that create a self out of the raw materials of inner life. They define the intrapsychic constraints and boundaries with respect to which a person’s autonomy may be threatened even by his own desires.27

23 Both papers are reprinted in The Importance of What We Care About, pp. 58–68 and pp. 159–76, respectively. 24 ‘Identification and Wholeheartedness,’ in ibid., pp. 166ff. 25 Ibid., p. 176. See also Frankfurt’s remarks in ‘Reply to Richard Moran,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt, pp. 220–21. There he insists that there can be no practical reasoning without prior identification ‘with the thoughts of which the reasoning is comprised.’ 26 ‘Identification and Wholeheartedness,’ in The Importance of What We Care About, p. 172. Frankfurt (curiously) contrasts decisions with choices in this respect: in the case of choices, he remarks, ‘the immediate object … is not the chooser but whatever it is that he chooses.’ 27 ‘Identification and Wholeheartedness,’ in The Importance of What We Care About, p. 170. See also ‘Identification and Externality’, in The Importance of What We Care About, pp. 66–8. It’s noteworthy that, on this revised account, a person may identify with a desire on a particular occasion (one they include in their ordering of preferences) even though their relevant highest-order volition doesn’t favor it to be their will. See Bratman’s comments,

92

Values and the Reflective Point of View

In this move to an account of identification that explicitly appeals to the role of decisions, Frankfurt is sensitive to one of Watson’s criticisms of his earlier account of identification – namely, that hierarchies of desire alone can’t explain identification – but he does not heed another. For, in his 1975 paper ‘Free Agency,’ Watson also criticizes Frankfurt’s picture that agents initially ‘ask themselves which of their desires they want to be effective in action’. Rather, Watson says: ‘they ask themselves which course of action is most worth pursuing. The initial practical question is about courses of action and not about themselves.’28 In short, Watson effectively criticizes the role of the objectifying mistake in shaping Frankfurt’s thinking about practical deliberation and identification.29 This kind of criticism of Frankfurt-inspired approaches to the issue of identification is acknowledged, however, by Bratman, as he develops his own account of identification with a desire, taking Frankfurt’s work on the issue as his point of departure. As noted in Chapter 4, page 80, Bratman develops his account in a way that respects the fact that we do not invariably foreground our desires when we are deliberating about what to do.30 And, in his 2003 paper ‘Autonomy and Hierarchy,’ Bratman explicitly agrees with Watson that the ‘initial practical question’ that we typically consider in our practical reasoning is ‘about “courses of action” and not about ourselves’ – that is to say, our motivation.31 Frankfurt has claimed that the notion of identification is ‘fundamental to any philosophy of mind and of action.’32 However, while the notion of identification may well be important in a deep way in the philosophy of mind and action, the case for this will be compromised for as long as that notion is yoked to the objectifying mistake. That mistake generalizes from a special range of cases, where we do foreground our desires in our practical premises, across the whole field of practical deliberation; and it sets the stage for one or other of two views about identification with a desire, ‘Identification, Decision, and Treating as a Reason,’ in Faces of Intention (Cambridge, 1999), pp. 191–2. 28 ‘Free Agency,’ Journal of Philosophy, 72 (1975): 219. 29 Frankfurt, concerned about a regress problem, eventually abandons a decision-based account of identification for one in terms of reflective satisfaction with a higher-order desire that favors a first-order motive. This occurs in his 1992 paper ‘The Faintest Passion.’ Reflective satisfaction, says Frankfurt, ‘is a state of the entire psychic system’ and ‘is simply a matter of having no interest in making changes’ – all of which ‘develops and prevails as an unmanaged consequence of the person’s appreciation of his psychic condition’ (in Necessity, Volition, and Love (Cambridge, 1999), pp. 104–5). It’s my impression that, in this 1992 discussion, Frankfurt continues to think of the deliberative setting as one in which we foreground our desires. See, for example, his comments in ibid., p. 105. 30 Bratman seems pretty much to share Frankfurt’s conception of the deliberative setting in his early discussion, ‘Identification, Decision, and Treating as a Reason,’ in Faces of Intention, pp. 185–206. It’s in later papers that he registers sensitivity to the different profiles of practical reflection. 31 ‘Autonomy and Hierarchy,’ Social Philosophy and Policy, 20 (2003): 157, 160. 32 ‘The Faintest Passion,’ in Necessity, Volition, and Love, p. 103.

Two Mistakes about Practical Reasoning

93

neither of which is satisfactory. In the one case, it is supposed that, when we consider what to do, there will always be an issue for us, to the extent that we are not wantons about the shape of our practical lives, of whether we identify with the desires that we foreground in our practical reasoning. This seems to be Frankfurt’s view in the early paper ‘Freedom of the Will and the Concept of a Person.’ In the other case, it is supposed that, when we consider what to do, the question of whether we identify with the desires that confront us in our practical reasoning is inescapable: such reasoning forces reflective distance from these desires onto us and we can’t proceed to reason except in terms of desires we come to identify with. This is Korsgaard’s view and Frankfurt’s later view. Neither of these views about identification with a desire is satisfactory. Let’s suppose that we can be practical reasoners without being reflective in the sense of considering whether we identify with the attitudes associated with our practical premises. Frankfurt’s early view – shaped by the objectifying mistake – is that we are, in such cases, unreflective about the desires in terms of which we reason, and that, to the extent we are cured of our unreflectiveness, we will face questions of whether we identify with these desires. However, this isn’t so, since it’s only in some cases – as when we are mere lovers of success in action – that we have our desires in focus in our premises as we reason about what to do. More typically, to the extent that we reason about what to do with reflectiveness, the issues of identification raised will concern whether we identify with the values and beliefs associated (through their contents) with our premises, or with the desires and other attitudes that result from, and express, the conclusions of our reasoning from these premises. Consider next the view – held by Korsgaard and the later Frankfurt – that wantonness isn’t a possibility for practical reasoners: there’s always a relevant and significant issue of whether we identify with the desires that confront us when we try to make up our minds about what to do.33 Now, even if it were true that we can’t avoid reflectiveness as practical reasoners, it wouldn’t follow that we can’t reason about what to do without considering whether we identify with the desires that always occupy the foreground in such deliberations. To suppose that enforced reflectiveness 33 This is Frankfurt’s view in the 1987 paper, ‘Identification and Wholeheartedness.’ It’s not clear to me whether, in his 1992 discussion ‘The Faintest Passion’ (see note 29), Frankfurt is once again prepared to allow that wantons can reason about what to do. At one point he does allow that someone – who ‘is merely a wanton’ – might be satisfied with their first-order desires ‘without in any way considering whether to endorse them.’ Such an individual, he remarks, ‘is identified with those first-order desires.’ The idea seems to be that an individual like this does not identify with the relevant first-order desires – because they are unreflective about them; however, they are none the less identified with them – because they are satisfied with them (see ibid., pp. 105–6). Is Frankfurt also here supposing that such an individual could be a ‘rational wanton’ with respect to the realization of these first-order desires? It isn’t clear (see Bratman’s related comments in ‘Identification, Decision, and Treating as a Reason,’ in Faces of Intention, pp. 203–5). One thing that is clear, however, is that, in still later discussion, Frankfurt insists that a person can’t be engaged in practical reasoning without identifying ‘with the thoughts of which the reasoning is comprised.’ See note 25.

94

Values and the Reflective Point of View

in such cases means enforced reflective distance from the desires that compete for our endorsement as reason-providers is to make the objectifying mistake. All that would follow, if practical reasoning guaranteed reflectiveness, would be that, in a select range of deliberative contexts, as where we are mere lovers of success in action, we can’t avoid the issue of whether we identify with the desires that we foreground in our practical thinking. In other deliberative contexts, where our desires aren’t in focus, different issues of identification would inescapably arise. However, it doesn’t seem at all plausible to suppose that we can’t reason about what to do without being reflective in the relevant sense; that is to say, without taking account of issues that concern what Frankfurt calls ‘the formation and maintenance of the self.’ It’s difficult to imagine how our practical lives could go on if reasoning about what to do required constant self-interrogation in this way.34 Frankfurt’s response, I take it, is that only some sequences of ‘logically related thoughts’ that occur in our minds qualify as sequences of practical reasoning – namely, those that satisfy the prior identification condition.35 The thesis that informs this response seems to be that, without prior identification with the relevant premise-related attitudes, it’s a mistake to suppose that someone is engaged in a sequence of thought in which they are treating certain considerations as reasons.36 However, in terms of the functional 34 Compare Moran’s objection to Colin McGinn’s view that the rationality of attitudes requires self-conscious awareness of them, in Authority and Estrangement (Princeton, NJ, 2001), pp. 109–13. – What is the relation between reflectiveness (critical stepping back from, and endorsement or rejection of, a psychic given) and self-consciousness? For Korsgaard, reflectiveness defines the essential structure of the human mind as self-conscious. See The Sources of Normativity, pp. 92–4, and Moran’s related discussion in Authority and Estrangement, pp. 138ff. Frankfurt also links reflectiveness with self-consciousness: see ‘Identification and Wholeheartedness,’ in The Importance of What We Care About, pp. 162ff. However, he draws back from any necessary connection between identifying with or outlawing a mental attitude and being conscious of it in his ‘Reply to Jonathan Lear,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt, p. 297. 35 ‘Reply to Richard Moran,’ pp. 220–21. 36 Ibid., pp. 218–19. Korsgaard gives the following argument for denying that someone who might seem to be reasoning when they act to satisfy a desire, in the absence of reflective endorsement, is really engaging in practical reasoning. Such an individual, she claims, cannot be construed as thinking in a way that conforms to, or applies, the hypothetical imperative: they may well do whatever satisfies their desire of the moment, but they do not put themselves in a position to draw a conclusion about what in particular they ought (have normative reason) to do. This is because, since they fail to reflectively endorse the desire they act on as a normative reason – will it as an end – they fail to give themselves an initial premise which tells them that they have normative reason to take the means to the satisfaction of their desire (see The Sources of Normativity, pp. 160–64, and ‘The Normativity of Instrumental Reason,’ in Garrett Cullity and Berys Gaus (eds), Ethics and Practical Reason (Oxford, 1997), pp. 215–54). Korsgaard’s argument relies on two assumptions. The first – which Frankfurt doesn’t share with her – is that we construct desires as normative reasons when we reflectively endorse them. I don’t see any grounds for supposing that identifying with a desire invests it with normative force. The second assumption behind Korsgaard’s argument is a mistake about the

Two Mistakes about Practical Reasoning

95

role of such a sequence of ‘logically related thoughts’ in the psychological and practical life of an individual, this isn’t a very convincing thesis. It seems clearly more plausible and psychologically realistic to allow that we can – and more or less frequently do – treat certain considerations as reason-giving in our practical and other thinking without any higher-order scrutiny. The theory of identification with a desire seems doubly compromised by Korsgaard’s and the later Frankfurt’s view that reasoning about what to do requires us to be reflective with respect to the desires that invariably figure in the foreground of such thinking. In the first place, this view ties the theory of identification to the objectifying mistake. In the second place, the view connects the theory of identification with a further mistake about practical reasoning – which we might dub ‘the identification mistake’ – according to which such reasoning or reflection requires reflective distance, or reflectiveness.37 It’s only by detaching our account of identification with a desire from both these errors about practical reflection that we will be able to locate properly the role of that phenomenon in our practical lives. To be sure, it’s by identifying with or rejecting aspects of our psychic lives like our desires that we attend, in the practical sphere, to the task of ‘the formation and maintenance of the self.’ The question is how best to conceive of this task and its role in the practical lives of persons, especially in relation to their consideration of ‘initial questions’ of what to do.38 scope of ‘ought’ in the hypothetical imperative. This normative principle says: ‘If you will an end you ought to take the means to that end.’ Korsgaard reads the scope of ‘ought’ here as narrow and as governing only the consequent. When we apply such a principle, she supposes, we detach the normative consequent. She then supposes that for detachment of the normative consequent to work, the antecedent has to be read in a normative way, so that by willing an end we give ourselves a (normative) reason to pursue it. However, the hypothetical imperative is a normative principle that enjoins us to be coherent in our pattern of practical thought and action. As such, ‘ought’ should be read as having a wide scope, as displayed in the following rewrite: ‘It ought to be: that you take the means to an end if you will that end’ (compare: ‘It ought to be: that you believe q if you believe p and you also believe that if p then q’). There’s no problem about construing a mere lover of success in action, who is unreflective about their current end-in-view, and who treats it as a source of merely formal value, as thinking and acting in a way that conforms to this normative principle. John Broome discusses Korsgaard’s modal mistake in ‘Normative Requirements,’ in Jonathan Dancy (ed.), Normativity (Oxford, 2000), pp. 97–8. See also Blackburn’s comments in Ruling Passions, pp. 242–3. 37 In this essay, I am paying most attention to the objectifying and desire-as-valuing mistakes about practical reasoning. We can list the identification mistake as a third error about practical reasoning that figures in the contemporary theory of agency. 38 Bratman offers a fruitful way to think about identification with a desire, its relation to the task of ‘the formation and maintenance of the self,’ and its relevance to ‘initial questions’ of what to do. He locates this phenomenon in a more general theory of identification in which we solve different ‘framework’ problems of self-governance – problems that concern the contours of our practical lives, including, importantly, the contours of our motivationally effective practical reasoning – by forming relevant higher-order, controlling policies. In the case of desire, Bratman says, the problem is one of self-management. See especially

96

Values and the Reflective Point of View

Velleman’s Theory of Agency Velleman is another philosopher whose theory of agency bears the mark of the objectifying mistake. Indeed, the theory gives the mistake a new, expanded dimension. Velleman is especially interested in a difference between two exercises of our capacity for making things happen – namely, the difference between mere activity and action. The former, he says, is motivated or purposeful, and so distinct from mere happenings, and yet it it’s ungoverned. The latter is activity that is regulated in a way that marks it out by contrast as autonomous.39 Velleman’s point of departure is what he calls ‘the standard model’ of human action: we act where our behavior is caused by a desire-belief pair that also give us a reason for what we do. Davidson is the leading exponent of the standard model. There’s something right about the model, Velleman thinks. It encapsulates the idea that agency can be reduced to a process of event-causation. However, the challenge is to identify events whose causal role really does add up to someone’s role as agent rather than as the producer of mere activity. The standard model is right, Velleman holds, about what it would take for a chain of events to constitute someone’s role as the producer of an action: it would have to connect their behavior to reasons in a way that ensures that they do what they do for those reasons. However, the standard model fails to identify a chain of events that plausibly counts as constituting a person’s doing something for a reason. It holds that someone does something for a reason where their behavior is nonwaywardly caused by a desire-belief pair. This helps to isolate out motivated or purposeful activity, but it doesn’t necessarily take us to action done for a reason (autonomous action). This is because someone’s behavior can be guided by a desire and belief as motives without its thereby being guided by them as reasons (justifying considerations).40 ‘Autonomy and Hierarchy’: 156–76. Presumably, there are different ways in which problems of self-management can come to occupy us – being given, in one way or another, pause for thought about reasoning in terms of a certain desire-foregrounding premise, during the course of deciding what to do, is one possibility. In developing his account of identification, Bratman concentrates on aspects of identification that have the shape of our practical reasoning in focus. It is to be expected, however, that the full story of identification will be wider-reaching than this. Just consider our desires. These mental states play a role in our lives as reason-providers and reason-trackers; but they also influence our patterns of attention, the fantasies we entertain, and so on. A full account of identification with a desire will need to take account of the many-aspected and differently-aspected ways in which desires shape our psychological lives and our activities. I am reminded here of Frankfurt’s comment, in ‘The Faintest Passion,’ that a person’s identification (or wholeheartedness) with respect to a psychic element concerns how it, rather than another element that non-contingently conflicts with it, ‘should be among the causes and considerations that determine his cognitive, affective, attitudinal, and behavioral processes’ (in Necessity, Volition and Love, p. 103). 39 ‘Introduction,’ in The Possibility of Practical Reason (Oxford, 2000), pp. 1–5. 40 Ibid., pp. 6–9.

Two Mistakes about Practical Reasoning

97

The flaw in the standard model, according to Velleman, is concealed in versions (like Davidson’s) that (wrongly) construe desire as itself a desirability judgment, and so as involving the grasp of a justification for acting.41 So, what do we need to add to the standard model in order to ensure that we have behavior that is not only motivated, but also autonomous? Velleman poses the issue in terms of ‘creature design.’ How would you design an autonomous agent, given the design for a creature capable of motivated activity? The form of the answer, he says, is that: … you would add practical reason to the existing design for motivated creatures, and you would add it in the form of a mechanism modifying the motivational forces already at work. You would design practical reason to survey a creature’s motives, to block or inhibit some of them, and to reinforce others.42

Frankfurt’s hierarchical model of agency takes us a step in the right direction, Velleman thinks, because it requires autonomous agents to be aware of their motives. However, Frankfurt’s model won’t do as an account of autonomy because it’s indifferent to why or how subjects have a favorable response to their lower-order motives.43 This, says Velleman, brings us to the model we need for autonomous action: it’s one that represents agents as aware of their motives and engaged by them as reasons. The third and right model of agency, he contends, ‘would define action as behavior whose first-order motives are perceived as reasons and are consequently reinforced by higher-order motives of rationality.’44 So, the task is to identify these rational motives. What we are looking for, according to Velleman, is the content of a higher-order, rational motive that turns mere behavior into action by regulating how the behavior is caused by its lowerorder motives (desires and beliefs). We can call the content of such a rational motive the constitutive aim of action, and it would be the analog of truth as the constitutive

41 Ibid., pp. 5–10. Velleman holds that the mistake here is to suppose that, in so far as to desire p is to regard p as to be brought about, to desire p is to judge p desirable or good; whereas in fact, says Velleman, to say that desiring p entails regarding p as to be brought about is just to characterize desire’s world-to-mind direction of fit. See also his ‘The Guise of the Good,’ in The Possibility of Practical Reason, pp. 99–122. For related discussion, see my remarks in Chapter 2, pages 32–4. 42 “Introduction,’ in The Possibility of Practical Reason, pp. 11–12. My italics. 43 Ibid., pp. 12–14. 44 Ibid., p. 14. Velleman has changed his mind a couple of times about whether firstorder desires double as both motives and reasons for acting. This was his initial view in Practical Reflection (Princeton, NJ, 1989), pp. 199ff. However, in the later paper entitled ‘The Possibility of Practical Reason,’ he nominates as basic reasons one’s recognition of a first-order desire together with one’s belief about how to realize that desire. This paper is reprinted in the collection The Possibility of Practical Reason, pp. 170–199. See especially pp. 197–9.

Values and the Reflective Point of View

98 45

aim of belief. Velleman suggests that the constitutive aim of action is the secondorder aim of knowing what we’re doing.46 So, by these lights, autonomous action is behavior caused by a desire and belief under the regulation of the higher-order aim of knowing what we’re doing. Which adds up to the following: autonomous action is behavior caused by firstorder motives, perceived as reasons, and thereby reinforced by a rational aim. This is because: (a) the aim of knowing what we are doing is the aim of behaving in a way that we can explain (make sense of), and (b), since that aim is constitutive of action, the considerations that present as reasons for action are just those in the light of which we could make sense of what we were doing, if we were to adopt the expectation of acting because of them. Such considerations are our first-order motives and circumstances.47 Velleman calls the practical expectation that we fashion out of our reasons – and which decisively tilts the balance of forces within us in favor of doing the expected thing when we act autonomously – a rationale. Such a rationale is a storyline that serves as our reflective grasp of what we are up to as agents. The final contrast between merely purposeful activity and autonomous action is then this. The former is activity, considered as simply motivated by belief and desire. The latter, on the 45 ‘Introduction,’ in The Possibility of Practical Reason, pp. 15–19. The idea is that belief aims at the truth (has truth as its norm of correctness) because it is constitutively regulated by mechanisms designed to ensure that it’s true. This fixes indicators of truth as reasons for belief because these just are the considerations in response to which belief is designed to be regulated. Similarly, we will be able to identify reasons for action in relation to action’s constitutive aim: they will be considerations of a kind in response to which action is designed to be regulated. 46 Such an aim, says Velleman, is sub-agential: it is not itself represented in our practical reasoning as an end-in-view (ibid., p. 21). Moreover, there’s no need to suppose that constitutive aims – of belief or action – are the objects of higher-order, rational motives. They might just be aims that are implicit in sub-personal psychological mechanisms. Where constitutive aims are the objects of motives or desires, they are personal, sub-agential aims. Where they are just implicit in mechanisms of regulation, they are aims that are both sub-personal and sub-agential (ibid., pp. 19–21). It’s noteworthy that Velleman has variously characterized the motives or aims that constitute behavior as action. In Practical Reflection (pp. 27–37), he identifies them as the desire to know, at any moment, whatever it is that we are doing then and the desire to understand (be able to explain), at any moment, whatever it is that we are doing then. In the paper ‘The Possibility of Practical Reason,’ he identifies the constitutive aim of action as conscious control over our behavior, or autonomy, which he interprets as the aim of acting in, and out of, a knowledge of what we are doing – under an explanatorily revealing psychological description (in The Possibility of Practical Reason, pp. 188ff.). However, in recent discussion, Velleman rejects identifying autonomy itself as the constitutive aim of action (‘Introduction,’ in ibid., p. 30). He now prefers to identify the aim that constitutes behavior as action simply as knowledge of what we are doing under a description that sets our bodily movements in an explanatory context of motives and circumstances (ibid., pp. 26–7). 47 Ibid., pp. 26–8. The idea is that action’s constitutive aim sets an internal criterion of success for action, relative to which considerations qualify as reasons for acting.

Two Mistakes about Practical Reasoning

99

other hand, is activity, considered as regulated by a rationale, for which our motives and circumstances provide the materials as reasons.48 Velleman offers an intriguing account of action. However, as is obvious, the account is deeply shaped by the objectifying mistake. Indeed, Velleman is especially interesting in this respect. For he expands the scope of that mistake. This becomes clear in his remarks about how, on the account of action that he offers, we come to judge some reasons better than others. The idea is that reasons will count as better to the extent that they present as materials for a better storyline about what we are doing, and that, in coming up with more intelligible storylines, our further stretches of reflection will take into account our various motives as well as considerations like our customs, emotions and traits of character. We go beyond our motives in devising better storylines; but the aim of self-understanding that regulates action, and so the reasoning behind it, and which alone gives reasons their justificatory force, keeps us entirely self-preoccupied.49 By these lights, our reasoning about what to do always shapes around the question ‘Is this me?’50 The objectifying mistake has found its full expression. Velleman holds that the standard model of action fails because it isn’t an account of behavior guided by the grasp of a justification. It might seem as if someone like Davidson meets this constraint, but he doesn’t really. It just looks as if he does because he (wrongly) reconstructs us as lovers of the good in so far as we are desirers and pursuers of goals. Now, Velleman is right that purposive activity is one thing, and reasoned activity another, and that Davidson goes wrong when he blurs the difference. However, Velleman’s own account of action as behavior guided by the grasp of a justification distorts the contours of reasoned activity. That account forces the justificatory significance of the considerations that figure within practical reflection into a single, Procrustean mold: given the aim that constitutes behavior as action, these considerations must support a piece of behavior as making sense to an agent as part of their autobiography, including, saliently, their current mental lives. And that simply ignores the plurality of considerations (very often non selffocussed) that can serve as reasons for acting within deliberation.

The Desire-as-valuing Mistake Davidson, the foremost exponent of the standard model of action, explicitly eschews the idea that we foreground our desires as practical deliberators. By Davidson’s lights, 48 Ibid., pp. 28–31. On this story, the expectation, or rationale, or reason on which we act is our choice (decision, intention). So, autonomous action, in sum, is behavior regulated by choice, where choice constitutes our understanding of what we are doing in terms of our motives and circumstances. 49 Ibid., pp. 28–9. 50 Compare Practical Reflection, pp. 301ff. Velleman is sensitive to the criticism that his story makes autonomous agents ‘unduly self-absorbed.’ See his comments in ‘Introduction,’ in The Possibility of Practical Reason, pp. 30–31.

100

Values and the Reflective Point of View

we are to think of practical reasoning as a sequence of thought that contains a mix of cognitive and practical attitudes; and he proposes that we let ordinary indicatives express the contents of belief and indicatives with explicitly evaluative words express the contents of the practical attitudes. Davidson then draws a distinction among practical attitudes. Some (like desires as premise-states) are to be treated as conditional evaluations, with contents expressible in judgments like: It is desirable for me to do A in so far as doing A is F. Others (like intentions as conclusion-states) are to be treated as unconditional evaluations, with contents expressible in judgments like: It is desirable (or intention-worthy) for me to do A.51 By this reckoning, the premises of practical reasoning that express contents of desire are no more about those desires than are the premises of theoretical reasoning that express contents of belief premises about those beliefs. The proposal models cases where we act on a desire-belief pair of reasons as a sequence of practical reasoning wherein what we do issues (nonwayardly) from a combination of premisestates with suitable propositional contents: contents such that an evaluative content of desire combines with an informative content of belief to imply that the action is desirable in a certain respect.52 In this way, Davidson avoids the objectifying mistake. However, as we know, the proposal comes at its own cost. It ignores the pathologies of agency, turning us into lovers of the good in so far as we are desirers or havers of goals (thus ruling out perversity), and into desirers or havers of goals in so far as we take things to be good in some way (thus ruling out a certain kind of weakness). And, along with ignoring these pathologies, the proposal blurs the difference between purposive and reasoned activity. So far, in discussing Davidson’s account of action done for a reason, I have been falling in with a generic use of the label ‘desire.’ Actually, however, there’s an ambivalence in Davidson’s own talk of desire. Sometimes the story is this: desire is one of a variety of practical states – ‘pro attitudes,’ as he calls them – that may enter into the explanation of action done for a reason.53 Other times, ‘desire’ serves 51 See ‘Intending,’ in Essays on Actions and Events (New York, 1980), pp. 96-102, and ‘Reply to Christopher Peacocke,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford, 1985), pp. 207ff. I borrow talk of ‘premise-states’ and ‘conclusion-states’ from John Broome. See his ‘Normative Practical Reasoning,’ Proceedings of The Aristotelian Society, supp. vol. 75 (2001): 175–93. 52 Davidson has revised his model of the logic of practical reasoning, abandoning the idea that it is deductive for the view that it is an analog of inductive inference. The central issue here is how best to construe the logical form of the evaluative premises. See ‘How is Weakness of the Will Possible?’ and ‘Intending,’ in Essays on Actions and Events, pp. 31ff. and pp. 96ff., respectively. 53 Davidson introduces the term ‘pro attitude’ in his account of ‘a primary reason’ in ‘Action, Reasons, and Causes,’ in Essays on Actions and Events, pp. 3–4. There, he says it covers ‘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

Two Mistakes about Practical Reasoning

101

as a general duty word for the whole range of practical states (pro attitudes) that may combine with belief to produce action done for a reason.54 This ambivalence in Davidson’s talk about desire goes along with a tension about the evaluative content of desire in his theory of action. To the extent that he focusses on desire as just one pro attitude among others, Davidson tends to emphasize expressions of the evaluative contents of these pro attitudes which are sensitive to ordinary evaluative discriminations and nuances;55 whereas, to the extent that he treats all the pro attitudes that may figure as reasons for action as desires, Davidson tends (in decision-theoretic spirit) to regiment the expression of their evaluative content in a way that is nuance-blind.56 Where Davidson groups together all the practical states that may enter into the explanation of action as desires with evaluative contents, he makes a mistake in categorizing desires (havings of goals) as values, for the reasons indicated before.57 Let’s now consider what Davidson has to say about the assimilation of desires, considered as specific pro attitudes, to values. The first thing to remark on is that Davidson has not always recommended that we treat desires in this more narrow sense as values. In his 1982 Lindley Lecture, for example, he favors taking the basic contrast between believing that p and desiring (or preferring) that p to be a contrast between two different attitudes, directed towards the same propositional content, and he resists what he calls ‘the simplest view,’ which would be to treat desires as desirability judgments.58 However, Davidson’s prevailing view has been that we should credit desires (narrow sense) with evaluative content, the chief advantage being the way it enables us to represent

kind’ (p. 4). So, in this context, Davidson regards desires as just one kind of pro attitude among others. 54 See Davidson’s remarks about reason explanations in ‘Paradoxes of Irrationality,’ in Richard Wollheim and James Hopkins (eds), Philosophical Essays on Freud (Cambridge: 1985), p. 293. 55 See the example – from ‘Intending,’ in Essays on Actions and Events, p. 86, and quoted below in the main text, this section – of someone adding sage to a stew from, specifically, the desire to improve its taste. Davidson adds after the last sentence: ‘We may suppose different pro attitudes are expressed with other evaluative words in place of “desirable.”’ 56 See Davidson’s remarks about regimentation and evaluative nuance, as these relate to his account of practical reasoning, in ‘Reply to Paul Grice and Judith Baker,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Actions and Events, pp. 209– 11. Davidson outlines the decision-theoretic background to his account in ‘Reply to Michael Bratman,’ in Essays on Davidson, pp. 199–200. 57 It’s true, of course, that some of the states being called desires will have evaluative content. But that’s just because, by the magic of bad naming, the states being collected as desires now (misguidedly) include judgments about what’s good, right, obligatory, and so on. 58 ‘Expressing Evaluations,’ The 1982 Lindley Lecture, Pamphlet Series (Lawrence, KS, 1984), pp. 9ff. See especially pp. 19–20.

102

Values and the Reflective Point of View

a logical relation between an action and a desire-belief pair, where the latter give the reason on which the act was done.59 Davidson sets out this view in his comments on an example where someone adds sage to a stew in order to improve its taste. We know the premise that corresponds to the agent’s belief, he says. It’s: Adding sage to the stew will improve its taste. He then continues: The agent’s pro attitude is perhaps a desire or want; let’s suppose he wants to improve the taste of the stew. But what is the corresponding premise? If we were to look for the proposition toward which his desire is directed, the proposition he wants true, it would be something like: … He improves the taste of the stew. This cannot be his premise, however, for nothing interesting follows from the two premises: Adding sage to the stew will improve its taste, and the agent improves the taste of the stew. The trouble is that the attitude of approval which the agent has toward the second proposition has been left out. It cannot be put back in by making the premise ‘The agent wants to improve the taste of the stew’: we do not want a description of his desire, but an expression of it in a form in which he might use it to arrive at an action. The natural expression of his desire is, it seems to me, evaluative in form; for example, ‘It is desirable to improve the taste of the stew’, or, ‘I ought to improve the taste of the stew.60

The problem, however, is whether we are justified in taking the step that allows the action to be seen as the result of an inference – namely, the step of reading evaluative content into the desire. To recommend the step, Davidson appeals to the idea that desires and other pro attitudes figure in the sincerity conditions on evaluative illocutions. He remarks: There is no short proof that evaluative sentences express desires and other pro attitudes in the way that the sentence ‘Snow is white’ expresses the belief that snow is white. But the following consideration will perhaps show what is involved. If someone who knows English says honestly ‘Snow is white’, then he believes snow is white. If my thesis is correct, someone who says honestly ‘It is desirable that I stop smoking’ has some pro attitude towards his stopping smoking. He feels some inclination to do it; in fact he will do it if nothing stands in the way, he knows how, and he has no contrary values or 59 Needless to say, Davidson claims the same advantage for crediting desires, considered as pro attitudes in general, with evaluative content. Thus, in characterizing the structure of reason explanations of (intentional) deeds, he notes that there must be both a logical and a causal relation between the action being explained and desire-belief pair which give the reason. Of the logical relation, he remarks: ‘Beliefs and desires have a content, and these contents must be such as to imply that there is something valuable or desirable about the action. Thus a man who finds something desirable in health, and believes that exercise will make him healthy can conclude that there is something desirable in exercise, which may explain why he takes exercise’ (‘Paradoxes of Irrationality,’ p. 293). 60 ‘Intending,’ in Essays on Actions and Events, p. 86. See also Davidson’s remarks on the project of setting out schematically the form of practical reasoning, as he conceives it, in ‘Reply to Christopher Peacocke,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Actions and Events, pp. 209ff.

Two Mistakes about Practical Reasoning

103

desires. Given this assumption it is reasonable to generalize: if explicit value judgments represent pro attitudes, all pro attitudes may be expressed by value judgments that are at least implicit.61

The consideration invoked in this passage is hardly convincing. The clear sincerity condition on someone’s saying ‘It’s desirable that I stop smoking’ is that they hold it to be desirable that they give up smoking. It’s then a question to be decided whether holding desirable is a motivational or practical state, and then a further issue, if it is, what its relation is to other pro attitudes like desire. So, for example, we (as cautious expressivists) might decide this: holding desirable is, or involves, a practical disposition which typically grounds, and finds expression, in desire. In that case, we can agree that there’s a practical state whose presence is a sincerity condition on saying that it’s desirable to stop smoking, and which merits being credited with the same evaluative content as that illocution, and which is intimately related to desire, without supposing either that the presence of desire itself is a sincerity condition on the illocution or that desire has an evaluative content. The weak can be sincere in declaring their values, and their weakness consists in a failure to desire what their declared values favor and dispose them to desire. Davidson develops a theory of action that avoids the objectifying mistake about the content of practical reasoning. However, he substitutes for it a different mistake: a mistake about the states that can enter into our practical reasoning as constituents. For he supposes that, as agents of reasoned activity, we either always act on a sequence of reasoning that involves our desires (broad sense) as premise-states, or that we at least sometimes act on a sequence of reasoning that involves our desires (narrow sense) as premise-states. However, this picture of desires as constituent elements in the reasoning that leads to action relies on the mistake of treating our desires as among our values. Our desires are not among our values; they, at best, express values that ground them.

Blackburn: Concerns, Desires and Values Blackburn’s picture of practical reflection is this. Within the reflective setting, various considerations offer themselves as reasons for acting. In so far as this is so, such considerations ‘reflect’ the way in which they have a motivational impact on us. These considerations are a mixed bunch, but they are about the situation in which we find ourselves as agents. To the extent that we count them as reasons for action, we express our desires or concerns; but the considerations are not themselves about our desires or concerns. This representation of the reflective setting is like Davidson’s. Inasmuch as it supposes that we never foreground our desires in deliberation (except incidentally), it’s an exaggeration that makes certain deliberative possibilities invisible. For we can discriminate between projects within deliberation: we can see ourselves as pursuers 61 ‘Intending,’ in Essays on Actions and Events, p. 86.

104

Values and the Reflective Point of View

of the good, or as just pursuers of success in action with respect to our desires. In the one case, we bring what we do under our theory of substantive value. In the other, where we self-ascribe desires as ends or starting points, we bring what we do under the concept of merely formal value. Moreover, even where we are, and see ourselves as, pursuers of the good, we can foreground our desires in a non-incidental way; as when we see the satisfaction of a desire as an element in our well-being. Blackburn supposes that, to the extent we count considerations as reasons for acting, we express our desires or concerns. He collects the states that make up our structure of motivation under the rubric of ‘concerns’ in much the way that Davidson collects them under the head of ‘pro attitudes.’ Our concerns are our dispositions to act – our desires, in a ‘thin’ sense of desire – and (derivatively) the considerations that engage with those dispositions.62 Just as Davidson holds for pro attitudes, Blackburn holds that all our concerns or desires – be they passing fancies or stable values or principles – have contents that can be expressed by evaluative propositions or propositions about reasons for acting. Augustine had it right, in Blackburn’s view, when he wrote of ‘the pull of the will and of love, wherein appears the worth of everything to be sought, or to be avoided, to be esteemed of greater or less value.’63 However, as we saw when discussing a similar move by Davidson, there’s a problem with thinking of all the states that enter into our profile of motivations or practical states as values. The taxonomy ignores the vicissitudes of valuing, submerging the contrast between values and the motivational or practical states that express them, where no pathology intervenes.64 Indeed, the vicissitudes of valuing cast their shadow over the way in which topics can figure as reasons for acting within reflection. Blackburn calls the considerations that play this role our concerns (or ‘the things we care about’) because they are considerations that have an effect on our dispositions to act.65 Now, it’s true that such considerations will typically impact on our dispositions to act as we reflect; but we need to allow that they can do this, by having an effect on which actions we come to value, without unfailingly prompting us to form a desire to act. For we need to allow that practical reflection can itself bear the mark of the vicissitudes of valuing. In such cases, the values we set on actions will be expressed in ways other than what we desire to do, as in the

62 Ruling Passions, pp. 123–5. See also Think (Oxford, 1999), pp. 271–2. There, Blackburn characterizes desires in a ‘thick’ sense as desires ‘thought of as states of enthusiasm for an end – things that put a gleam in our eye.’ To desire (or want) to do something in this sense is to feel like doing it. 63 Think, p. 283. 64 It’s true that Blackburn often reads like a cautious expressivist, acknowledging the vicissitudes of valuing. See Chapter 2, note 27. However, he also often doesn’t read that way. Thus, in Think, pp. 282ff., he characterizes non-cognitivism as the theory of value that reads the following equation from left to right (as regards direction of explanation): ‘One of X’s concerns is to aim/promote/endorse I = X thinks I is good/thinks I is a reason for action.’ 65 Ruling Passions, p. 123.

Two Mistakes about Practical Reasoning

105

continuing affective salience of the topics that present as reasons, or in a wish for a change of heart, and so on.66 Blackburn, for his part, defends equating concerns or desires with values by appealing to the role of evaluative propositions in practical thought and talk. For example, he writes: When people have concerns, they express themselves by talking of reasons, and seeing the features that weigh with them as desirable or good. They do this in the ‘pull of the will and of love’. I believe we invent the normative propositions (‘This is good’; ‘That is a reason for action’; ‘You ought to do this’) in order to think about the concerns to demand of ourselves and others. We talk in these terms in order to clarify our motivational states, to lay them out for admiration or criticism and improvement.67

The general theme is familiar from Chapter 2. The function of evaluative propositions is to give our practical thought and talk a focus: thus, we make questions about what to desire or do deliberatively and discursively tractable by considering what it would be good to do or what there is reason to do. I agree that the role of evaluative propositions in practical thought and talk supports the claim that our values about what to do are (or involve) motivational or practical states. However, it’s another matter whether our concerns or desires are values about what to do. An obvious alternative is that some dispositions to act – like our concerns or desires, as these bear on actions – are indirectly responsive to reasoning about value via the values that we form as conclusions to such reasoning. This, of course, is the alternative I recommended in Chapter 2, along with an expressivist gloss. We can expect to have an impact on our desires (and more remotely, actions) by reasoning about what’s good or what there’s a reason to do, because any value we come to set on a course of action will ordinarily be expressed in a desire to pursue it. If we suppose, instead, that our desires are amenable to reasoning about value just because they are themselves values, we distort the relation between values and desires. Moreover, on the view that keeps that relation right, it’s easy to say why talk of values or reasons is an apt way to voice desires: in such cases, we voice our desires by voicing the values that they manifest and which ground them.68

66 Blackburn discusses examples of radical breakdown in deliberation where considerations lack any somatic or emotional marking (ibid., pp. 125–31). Clearly, I have in mind a different range of cases, where deliberation proceeds under conditions of evaluative stress. 67 Think, p. 286. 68 Compare Pettit and Smith (who are cognitivists) on the aptness of evaluative propositions for expressing desires, as their potential explainers and justifiers. See Pettit and Smith’s ‘Parfit’s P,’ in Jonathan Dancy (ed.), Reading Parfit (Cambridge, MA, 1999), pp. 72–5.

This page intentionally left blank

Chapter 6

What’s Wrong with the Sensible Knave? The values that we form within reflection typically enable us to mold aspects of our practical and passional lives. But are there any values that are reflectively unavoidable: are there any values that we must acknowledge in so far as we are practical reasoners? Famously, Kant thinks so: our very use of reason as setters of ends commits us to valuing the humanity or rational nature that we all share. Moreover, the view, or fascination with the view, that there are reflectively inescapable values is alive and well. Thus, Korsgaard seeks to vindicate Kant – with a proviso: our very use of reason commits us to valuing our shared humanity or rational nature if we are to value anything at all. And a philosopher like Michael Smith seems to hold that, if there are any moral values, they will be reflectively inescapable ones. By his reckoning: moral facts, if there are any, are facts we would all agree on under conditions of ideally conducted reflection; or, put another way, moral-making features of acts, if there are any, are features we would all want to characterize our acts after perfect reflection on the actual situations in which we find ourselves. Let’s now consider Kant’s view that humanity is an inescapable value within practical reflection, and Korsgaard’s defense of it. I center the discussion around the figure of the sensible knave. Defect of Passion or Defect of Reason? Hume’s sensible knave ‘observes the general rule, and takes advantage of all the exceptions.’ In his Enquiry Concerning the Principles of Morals, Hume comments: … a sensible knave, in particular incidents, may think that an act of iniquity or infidelity will make a considerable addition to his fortune, without causing any considerable breach in the social union and confederacy. That honesty is the best policy, may be a good general rule; but it is liable to many exceptions; and he, it may perhaps be thought, conducts himself with most wisdom, who observes the general rule, and takes advantage of all the exceptions.1

1 In Enquiries Concerning Human Understanding and Concerning the Principles of Morals, ed. L.A. Selby-Bigge, 3rd edn, revised by P.H. Niddich (Oxford, 1975), p. 282.

108

Values and the Reflective Point of View

By many lights, the knave isn’t as he ought to be. Some philosophers would say that he reasons poorly, others just that he’s a nasty piece of work. Blackburn draws the contrast between these two kinds of reaction to the sensible knave thus: What problem is posed by the knave? It is usually posed as that of finding a way of showing that the knave is irrational, and it is in this form that it preoccupied both Plato and Kant. The idea is to find a way of proving that if reason controlled him as it ought to, he would sacrifice self-interest out of his respect for others, or perhaps for truth, property, social order, or for the moral law, or for integrity and promises. For Hume and Smith, by contrast, … [the knave’s] defect is not one of rationality, but just one of lacking a normal desire or source of affect. We educate people to care that they share the desires we admire, but if our education has failed then it may be too late.2

What’s wrong with the knave, according to Hume and Adam Smith, is that his sentimental education has failed, and he lacks ‘the voice within.’3 And to the extent that this is so, he is beyond any susceptibility to argument. Blackburn, who agrees that the knave suffers from a defect of affect rather than a defect of reason, comments: We can exhort the knave to share our sentiments. We can try to turn up the volume of his feelings for those that he exploits. What we cannot do is argue the knave back into upright behaviour, for if his sentiments are only activated by what he perceives as his own advantage, then we can advance no effective consideration except by appealing to his own self-interest, and on the occasions that he exploits this is not enough.4

Of course, there will be knaves and knaves. The knave asks why he shouldn’t profit from wrongdoing in safe circumstances. Hume comments: I must confess that, if a man think that this reasoning much requires an answer, it will be a little difficult to find any which will to him appear satisfactory and convincing. If his heart rebel not against such pernicious maxims, if he feel no reluctance to the thoughts of villainy or baseness, he has indeed lost a considerable motive to virtue; and we may expect that his practice will be answerable to his speculation.5

However, Hume doesn’t think that this is all that can be said about the case. There will still be knaves who have an ‘interested obligation’ in being honest where they could safely not be. This is because of the way in which sympathy works: we live in the eyes of others when it comes to self-esteem. Even knaves will realize how others would regard them if they were found out; and so, unless they are hardened

2 Ruling Passions (Oxford, 1998), pp. 208–9. 3 See Blackburn’s discussion, in ibid., pp. 200ff. 4 Ibid., p. 209. 5 In Enquiries Concerning Human Understanding and Concerning the Principles of Morals, p. 283.

What’s Wrong with the Sensible Knave?

109

villains, they will retain an interest in acting honestly to avoid the pain of ruined self-esteem.6 Some philosophers think that what’s wrong with very hardened knaves is their hardness of heart: there would have to be a change of heart or character before they even became susceptible to arguments to change their ways. Other philosophers think that knaves display a defect of reason: they are susceptible to arguments to change their ways simply in virtue of their ability to reason, but their reasoning is faulty in some way; there’s a failure within practical reason itself. Kant is a clear example of a philosopher who holds the ambitious view that immoral conduct marks out a flaw in the agent’s very reasoning. Kant holds that what’s wrong with the knave is that, in failing to will in accordance with the categorical imperative, he flouts the constitutive principles of practical reason itself. The knave is the practical analog of someone who (allegedly) forfeits intelligibility by not conforming their thinking about what’s true to prohibitions like the law of non-contradiction.7

Arguing the Knave into Right Action Kant offers us a philosophically seductive construal of the knave: he wouldn’t be a knave if he were exercising his practical reason in a perfect or unimpaired way. However, it’s important to be clear that Kant can allow for pathologies of the heart to play a role in immorality. While Kant is committed to the view that it’s in principle open to argue the knave into changing his ways, simply in virtue of his rational nature, he can well accept that the knave’s immoral choice is a defect of reason that has its explanation in badness of heart or character.8 To say that the knave suffers from a defect of reason is simply to locate where the fault occurs, and not to offer a diagnosis of its cause. The point is familiar in theoretical cases. Some breakdowns within reasoning from beliefs to further beliefs are ‘cold,’ while others are ‘hot.’ Examples of the latter include cases where someone resists forming the belief that best fits with the rest of what they think, because it’s too hurtful. In examples like this, if the individual is to rid themselves of the defect of reason that characterizes them, they will need to come to terms with their pain.9 This familiar kind of point can become obscured in discussions of practical cases. Consider failing to take a (recognized) necessary means to an end. Kant counts such 6 See Korsgaard’s discussion, in The Sources of Normativity, ed. Onora O’Neill (Cambridge, 1996), pp. 55–60. 7 See Blackburn’s remarks, in Ruling Passions, pp. 214–16. 8 Compare Korsgaard’s remarks in ‘Morality as freedom,’ in Creating the Kingdom of Ends (New York, 1995), pp. 171–6. See too her discussion of the ‘internalism requirement’ on reasons for action and belief, in ‘Skepticism about practical reason,’ in ibid., pp. 315ff. 9 Davidson is sensitive to these possibilities in his account of self-deception as motivated irrationality. See ‘Deception and Division,’ in Jon Elster (ed.), The Multiple Self (Cambridge, 1985), pp. 79–92. For further discussion, see the Appendix.

110

Values and the Reflective Point of View

a failure as displaying a defect of reason: we fail to conform our actions to the hypothetical imperative, another constitutive principle of practical reason. However, Blackburn (following Hume) expresses reluctance to say that an agent suffers from a defect of reason where he fails to take the sole (recognized) means to a single professed end, commenting: ‘Perhaps he is just depressed, and then it is his mood or emotional state that needs a kick-start, and not his rationality.’10 By ‘rationality’ here, Blackburn means ‘reasoning.’ The example (as Blackburn acknowledges) is one where an agent fails to make sense in a certain way.11 Let’s set aside the issue of whether the lapse in sense is actually best thought of as involving a fault in reasoning. My present point is just that it wouldn’t be enough, as Blackburn seems to suppose, to point to the agent’s depression, in such a case, to rule out the lapse in sense as involving a defect of reason. For the agent could be reasoning poorly just because of that, and then the way to improve his reasoning would be to help him snap out of his glumness. To say that agents suffer from a defect of reason in failing to take, or in failing to intend to take, the means to a self-ascribed end is not to say that the origins of that defect lie within reason itself, rather than within affect or passion. The Knave’s Maxim and Universal Law Kant thinks that immoral maxims are contrary to practical reason: they do not pass the test of universal law, and they implicate our wills in a kind of contradiction. There are two different kinds of defect of practical reason in Kant’s account. In the one kind of case, we do something other than take the (recognized) necessary means to an end. The second kind of case is immoral conduct, where we act on a maxim that fails the test of universal law. The first kind of case involves an explicit contradiction in our pattern of willing.12 In the case of immoral conduct, there isn’t any explicit contradiction in what we will: our maxim isn’t somehow inconsistent within itself.13 However, we are inescapably tied to a contradiction, by the maxim and its universalization in accordance with the categorical imperative, the law we give ourselves in so far as we are free rational beings.14 This law tells us to: Act only on that maxim through which you can at the same time will that it should become a universal law.15 10 Ruling Passions, p. 320. 11 See ibid., p. 320; also pp. 239–40. 12 Compare Korsgaard, ‘Kant’s Formula of Universal Law,’ in Creating the Kingdom of Ends, pp. 93–4. 13 Compare Korsgaard, ‘Morality as freedom,’ in Creating the Kingdom of Ends, pp. 160–61. 14 See Groundwork of the Metaphysic of Morals, trans. H.J. Paton (London and New York, 1948), pp. 108–9. For comment, see Korsgaard, ‘Morality as freedom,’ in Creating the Kingdom of Ends, pp. 162–7. 15 Groundwork of the Metaphysic of Morals, p. 84.

What’s Wrong with the Sensible Knave?

111

Kant gives four famous examples of maxims that supposedly fail the test of universal law. These are the maxims behind suicide, lying promises, refusing to cultivate one’s talents, and withholding assistance to others. It’s a matter of debate what kind or kinds of contradiction Kant thinks we are implicated in in so far as we act on immoral maxims.16 The maxim behind lying promises is of special interest in considering the sensible knave. The lying promisor has the maxim: ‘I will make promises that I don’t intend to keep in order to borrow money I need.’ Kant says that the maxim cannot be willed as a universal law without contradiction: For the universality of a law that every one believing himself to be in need can make any promise he pleases with the intention not to keep it would make promising, and the very purpose of promising, itself impossible, since no one would believe he was being promised anything, but would laugh at utterances of this kind as empty sham.17

What are we to make of Kant’s view that the maxim of the lying promisor fails the test of universal law? Korsgaard identifies three interpretations of the contradiction test: the practical, the logical and the teleological.18 Kant apparently has in mind either the logical or practical ways of regarding the test in his treatment of false promises. According to Korsgaard, the making of false promises is an exemplar of immoral conduct for Kant in that someone proposes to perform an action whose success depends on their making an exception of themselves.19

16 A further complication is that Kant distinguishes between contradictions in conception (as when we apply the test to the maxims behind suicide and lying promises) and contradictions in the will (as when we apply the test to the maxims behind refusing to cultivate one’s talents and withholding help to others). See ibid., pp. 86–7. I ignore this complication here. 17 Ibid., p. 85. 18 See ‘Kant’s Formula of Universal Law,’ in Creating the Kingdom of Ends, pp. 77– 105. Korsgaard herself favors the practical interpretation of the contradiction test. She thinks that it fits with Kant’s general notion of a practical contradiction as a pattern of willing that involves thwarted purposes (pp. 93ff., 102). However, she acknowledges that the practical reading of the test passes certain maxims that Kant would forbid us to adopt. Thus, the maxim of taking one’s own life in order to escape misery can be universalized without any practical contradiction: the suicide’s end will not be frustrated in a world where everyone shares the maxim (pp. 100–101). And there are others, like the maxim of lying to liars (including deceitful murderers): we can universalize this maxim because liars typically will believe the return lies they are told by those trying to resist their deception. See Korsgaard’s discussion in ‘The right to lie: Kant on dealing with evil,’ in Creating the Kingdom of Ends, pp. 136–7, 145. 19 See her comments in: ‘An introduction to the ethical, political, and religious thought of Kant,’ in Creating the Kingdom of Ends, pp. 14–15; ‘Kant’s analysis of obligation: the argument of Groundwork I,’ in Creating the Kingdom of Ends, pp. 63–4, and ‘Kant’s Formula of Universal Law,’ in Creating the Kingdom of Ends, pp. 92–3, 100–101. However, note Allen Wood’s criticism of this construal of the moral point of the test of universal law, in Kant’s Ethical Thought (Cambridge, 1999), p. 108.

112

Values and the Reflective Point of View

Let’s first consider the logical reading of the contradiction test. On this construal, in order to check whether it’s permissible to act on the maxim behind false promises, we are to try to conceive of a world in which the maxim is universalized and endures as a law. Kant’s claim (by these lights) is that the maxim fails the test because any such attempt will founder on the self-contradiction that’s embedded in the description of a world in which (as a persisting state of affairs) everyone gives false promises in order to borrow needed cash. Such a world would be one in which eventually everyone with the relevant aim makes lying promises even though no one does, since the practice of repayment promises would have broken down under its (acknowledged) universal abuse.20 Consider next the practical reading of the test. In applying the test in this way to a maxim, we don’t assume that the universalized maxim survives as a law, and we focus on worlds considered as the products of the operation of the maxim as a law. In the case of the maxim behind false promises, the question is whether anyone can, at one and the same time, will their maxim and its universalization, without thereby willing a state of affairs in which their own end, as contained within the maxim, is frustrated. Kant’s claim (by these lights) is that the maxim fails the test. If anyone wills that they act on the maxim that guides false promises in a world that is marked by the impact of its universalization, they unavoidably will a situation in which they fail to secure needed cash by trying to use a false promise as a means. This is because the universalization of the maxim behind false promises would yield a law that was self-undermining: under the stress of its (recognized) universal abuse, the practice of repayment promises would collapse, and so eventually nobody after ready cash would be able to achieve that end by using a false promise as a means. Now, the sensible knave isn’t promiscuous in deceit. The sensible knave is just that: he’s someone who’s sensible in his knavery. The knave’s underlying maxim is to better his own circumstances by safe deceit, and his maxim, in the case of lying promises, will reflect his cautious self-love. When it comes to borrowing, the knave, being sensible, will pick his targets with care and make promises to repay that he doesn’t intend to honor just in situations where he’s sure he can get away it. The cautious maxim on which the sensible knave acts when he makes false promises to pay back debts seems to be universalizable without contradiction. For we don’t need to suppose that the practice of repayment promises wouldn’t survive in a kingdom of sensible knaves. Such a kingdom could well be one in which the practice survived, surrounded by vigilance against fraud and a system of measures to deter it. We can allow that knaves could still safely secure loans under false pretences in such a social world, except that opportunities for safe deceit presumably would become limited. However, in that case, the maxim on which the knave acts when he gains cash by false promise passes the test of universal law. There’s no practical contradiction: 20 Compare Korsgaard, in ‘Kant’s Formula of Universal Law,’ in Creating the Kingdom of Ends, pp. 81–2, 84–5. See too Barbara Herman, The Practice of Moral Judgment (Cambridge, MA, 1993), p. 137.

What’s Wrong with the Sensible Knave?

113

knaves can will to act on their maxim in a world of sensible knaves without thereby willing the defeat of their own purposes in any such world. And there’s no logical contradiction: repayment promises could survive in a world of universal cautious abuse, and so we can conceive of a world of canny knaves safely (even if seldomly) securing loans under false pretences, without thereby having to imagine them per impossibile as promising to repay money at some time after the practice of repayment promises has died out. In short, the evil maxim on which the sensible knave acts when he borrows cash by deceit passes Kant’s contradiction test and turns out to be a so-called ‘false positive.’ Kant himself discusses a version of the sensible knave case. He sets it out thus: I have, for example, made it my maxim to increase my wealth by every safe means. Now I have a deposit in my hands, the owner of which has died and left no record of it. This is, naturally, a case for my maxim. Now I want only to know whether that maxim could also hold as a universal practical law. I therefore apply the maxim to the present case and ask whether it could indeed take the form of a law, and consequently whether I could through my maxim at the same time give such a law as this: that everyone may deny a deposit which no one can prove has been made. I at once become aware that such a principle, as a law, would annihilate itself since it would bring it about that there would be no deposits at all.21

Kant here seems to have in mind a practical reading of the test of universal law. If we apply the maxim of increasing our wealth by every safe means to the case at hand, we get a maxim that we cannot will along with its universalization, without thereby willing a world in which we fail to add safely to our wealth by trying to steal untraceable deposits, since there won’t be any deposits, the practice of making deposits having died out under its universal abuse. However, Kant exaggerates the impact of the endemic theft of untraceable deposits on the practice of making deposits. We don’t have to suppose that the practice itself would collapse in any such social world. It could well be a world where the practice remained intact, surrounded by vigilance against the lodging of deposits that can’t be tracked down to their owners and a system of measures to deter their theft. We could expect that in such a world it typically would be hard for knaves to locate deposits that they could steal without detection; however, we don’t even have to suppose that no untraceable deposits would exist. So, once again, the knave’s immoral maxim passes the contradiction test.22

21 The Critique of Practical Reason, in The Practical Philosophy, trans. and ed. Mary J. Gregor (Cambridge, 1996), p. 161. 22 The maxims that characterize the sensible knave are false positives. The contradiction test also attracts so-called ‘false negatives’. The problem here seems to be that the test is blind to the difference between simply taking advantage of the habits or actions of others and taking advantage of others. For relevant discussion, see Blackburn, Ruling Passions, pp. 218ff. (the credit card example); Barbara Herman, The Practice of Moral Judgment, pp. 136–43, and Wood, Kant’s Ethical Thought, pp. 105–10.

114

Values and the Reflective Point of View

The Knave’s Maxim and Respect for Humanity Kant’s view that immoral action is contrary to practical reason has a formal aspect and a material one. The formal aspect is that wrongdoing is implicated in a contradiction, as revealed by the failure of its maxim to pass the test of universal law. The material aspect is that immoral conduct fails to express respect for an end that is inescapable for us in so far as we have pure practical reason. Kant specifies this end in the formula of humanity: Act in such a way that you always treat humanity, whether in your own person or in the person of any other, never simply as a means, but always at the same time as an end.23

Kant regards the formula of humanity as equivalent to the formula of universal law. Humanity is the value, the unconditional value, behind the test of law: it is the matter of which universal law is the form. By ‘humanity,’ Kant means ‘rational nature’ – that is, ‘our capacity to set ends through rational choice.’24 The formula of humanity tells us that we are to act only in ways that treat persons (including ourselves) as ends, in the sense that those actions express respect for rational nature, wherever it occurs, as the condition of all value.25 The formula obviously rules out false promises, and indeed, all forms of lying. Kant puts it this way: … the man who has a mind to make a false promise to others will see at once that he is intending to make use of another man merely as a means to an end he does not share. For the man whom I seek to use for my own purposes by such a promise cannot possibly agree with my way of behaving to him, and so cannot share the end of the action.26

Lying is always wrong as a means because those lied to cannot assent to it or hold its end as their own. So, when we lie to or deceive someone, as we do when we give false promises, we fail to express respect for their capacity to set ends for themselves.27 23 Groundwork of the Metaphysic of Morals, p. 91. 24 Compare Korsgaard, ‘Kant’s Formula of Humanity,’ in Creating the Kingdom of Ends, pp. 110–14, and Wood, in Kant’s Ethical Thought, pp. 130–32, 144–7. 25 See Wood’s comments on expressive reasons for action and respecting humanity, in ibid., pp. 141–7. 26 Groundwork of the Metaphysic of Morals, p. 92. 27 See Korsgaard: ‘The right to lie: Kant on dealing with evil,’ in Creating the Kingdom of Ends, pp. 137–43; ‘Creating the Kingdom of Ends: Reciprocity and responsibility in personal relations,’ in Creating the Kingdom of Ends, pp. 192–3, and ‘Two arguments against lying,’ in Creating the Kingdom of Ends, pp. 346–8. Also Wood, in Kant’s Ethical Thought, pp. 148–9. Korsgaard argues that Kant’s theory may allow us to lie (or use coercion) for certain benevolent or paternalistic purposes, as when manipulating a small child for its own safety, because those lied to are insufficiently autonomous. However, we are under pressure from the moral law ‘towards treating every human being as a free rational being, regardless of actual facts.’ See her discussion in ‘Two arguments against lying,’ in Creating the Kingdom of Ends, pp. 349–52.

What’s Wrong with the Sensible Knave?

115

The formula of humanity thus also rules out the sensible knave’s maxim of safe deceit. That immoral maxim passes the contradiction test because of its caution in deceit; however, as the maxim of someone who manipulates others as means to an end, it clearly fails the test of respect for humanity. This result shows that, despite Kant’s own view, the two tests are not equivalent.28 Let’s therefore consider the idea that our reasoning has gone wrong in the practical sphere in so far as we, like the sensible knave, act on maxims that fail to express respect for humanity.

Kant: Humanity as a Necessary End Kant holds that we always act for the sake of some end. Ends may provide us with reasons positively, as purposes to be achieved, or negatively, as things we must not act against. If there is a categorical imperative, then there must be an end in itself (an end of unconditional value) that is the end for the sake of which we all act (by way of expressing respect) when we act on the categorical imperative.29 Kant argues that only our humanity or rational nature is fit to be this end. So, if there is a categorical imperative, it is for the sake of our humanity or rational nature that we act when we conform our actions to it. In the formula of humanity, Kant recasts the categorical imperative in terms of this end in itself. Kant’s argument here is notoriously hard to grasp. It has a negative and a positive part. In the negative part, Kant argues that objects of inclination, inclinations, and natural beings or ‘things,’ are not fit to be considered ends in themselves. In the positive part of the argument, Kant turns to the presuppositions of our practice of setting ends: as rational choosers of ends, we are committed to acknowledging that humanity or rational nature is the only thing fit to be an end in itself, since it is the source or condition of all goodness. Kant puts the positive part of the argument in these terms: The ground of [the supreme practical principle] is: Rational nature exists as an end in itself. [1] This is the way in which a man necessarily conceives his own existence: it is therefore so far a subjective principle of human actions. [2] But it is also the way in which every other rational being conceives his existence on the same rational ground which is 28 Other interesting cases involving deception show that the two tests are not equivalent. Thus, while we can universalize the maxim of resisting lies with lies in order to save a life, we still fail to express respect for the murderer’s rational nature when we lie to him. See Korsgaard, ‘The right to lie: Kant on dealing with evil,’ in Creating the Kingdom of Ends, pp. 135–43. However, note her remarks, pp. 143–4. Korsgaard holds that, even though Kant’s two tests are not equivalent, any violation of the test of universal law is also a violation of the test of respect for humanity. False negatives (mentioned in note 22) show that she’s wrong about this. For further comment on the relation between Kant’s two tests, see Wood, in Kant’s Ethical Thought, pp. 97–102, 107–10, 139–40, 182–90. 29 Compare Korsgaard, ‘Kant’s Formula of Humanity,’ in Creating the Kingdom of Ends, pp. 106–8. See also Wood, in Kant’s Ethical Thought, pp. 112–18.

116

Values and the Reflective Point of View valid also for me; [3] hence it is at the same time an objective principle, from which, as a supreme practical ground, it must be possible to derive all laws for the will.30

The following seems a plausible reconstruction: 1a. Each of us necessarily conceives of our own ends as good in so far as they are the objects of our own rational choice; so, each of us necessarily takes our own rational nature as an end in itself, it being the source of the goodness of our ends. 2a. Each of us recognizes that every other rational being similarly regards their own rational nature as an end in itself, it being the source of the goodness of their ends, and that in doing so they are treating as an end in itself a capacity that we all share. 3a. In consistency, then, each of us must accord the same kind of value-conferring status to the rational nature of others, and hence must treat rational nature, whether in our own person or in the person of any other, as an end in itself, it being the source of the goodness of all good ends.31

On this construal of Kant’s argument, the first two premises make what we presuppose as rational end-setters explicit. The conclusion (3a) draws out what these presuppositions commit us to: we must treat humanity or rational nature as an end in itself, it being the source of all goodness. In this way, the ‘deep structure’ of rational choice reveals the only thing fit to be the ground of the categorical imperative (if there is one). Some commentators have taken Kant to task for supposing that we would be correct to infer that the source of goodness must itself be good. The objection is that, in order for x to have the power to confer some property F, x doesn’t have to have the property F itself. So, Kant owes us an account of why the authority which confers goodness must itself be good.32 I am especially interested in another aspect of Kant’s argument. Kant’s concern is to identify what is fit to be the end for which we act when we act on the categorical imperative, if it exists. Such an end, according to Kant, would be an end in itself – something of unconditional value that gives everyone reason to act, and so something of unconditional, agent-neutral value. Now, when Kant interprets our implicit commitments as rational setters of ends, he credits us with a conception of our ends that neatly suits his project. For he takes it that we conceive of our own ends as good in an agent-neutral sense, as good in the sense that they provide everyone with a reason to make them their ends as well. So, by these lights, our 30 Groundwork of the Metaphysic of Morals, p. 91. 31 For discussion of the argument, see Korsgaard: ‘Kant’s Formula of Humanity,’ in Creating the Kingdom of Ends, pp. 114–24, and ‘Two distinctions in goodness,’ in Creating the Kingdom of Ends, pp. 260–62, 267–8. See also Wood, in Kant’s Ethical Thought, pp. 122–32. 32 See Berys Gaut, ‘The Structure of Practical Reason,’ in Garrett Cullity and Berys Gaut (eds), Ethics and Practical Reason (Oxford, 1997), p. 174. Wood defends Kant against the objection in Kant’s Ethical Thought, p. 130. He argues that we should think of Kant’s focus as the authoritative, prescriptive source of goodness.

What’s Wrong with the Sensible Knave?

117

implicit commitment, as rational end-setters, to the goodness of our own rational nature as the source of goodness is a commitment to its unconditional, agent-neutral goodness; that is, to its existence as an end in itself.33 Let’s set aside the issue of whether, as rational setters of ends, we really do presuppose that we construct normative value. Why should we agree that, as rational setters of ends, we are committed to regarding our ends as goods for all? For Kant, presumably, we are to be credited with regarding our ends as goods for all, in so far as we regard them as good, because genuine ascriptions of goodness are ascriptions of objective goodness, and objectivity is a matter of neutrality with respect to the individual identities of agents. The only really good ends, and so the only ends that provide genuine reasons, are ends that are (or are wholly subsumable under) goods and reasons for everybody.34 But why should we accept that normative force is tied to objectivity in this sense? Suppose I set my own happiness as an end in an act of rational choice. Kant is committed to holding that I thereby must regard my own (permissible) happiness as a good for all in the sense that it provides everyone with a reason for concern. Why should we accept a universality constraint on goodness or reasons as grand as this? Why not just something more modest, such as that anyone else’s indexical end of happiness is similarly a reason for them?35 Let’s return our attention to the case of the sensible knave. Kant’s view is that an immoral maxim like the one the knave acts on is contrary to practical reason. Formally, the knave’s wrongdoing is implicated in a contradiction, as revealed by the failure of its maxim to pass the test of universal law. Materially, the knave’s action is wrong because it fails to express respect for an end that is inescapable for any one of us in so far as we are practical reasoners or have a will – namely, the end of humanity itself. The formal part of Kant’s story fails: the knave’s maxim survives the contradiction test. On the material side, Kant’s postulation of humanity or rational nature as a necessary end which demands our respect makes an assumption that needs to be vindicated. This is that normative force is public, in the sense of applying to all, regardless of anyone’s individual identity.

Korsgaard and the Sensible Knave Korsgaard defends Kant’s view that normative force is public. She has a special interest in the example of the sensible knave. In The Sources of Normativity, she discusses the challenge that the knave poses to Hume’s account of ‘interested

33 See Wood’s discussion, in ibid., pp. 127–30. 34 See again Wood’s discussion, in ibid., pp. 127–30. Refer also to Korsgaard’s comments on two interpretations of agent-neutral value, in ‘The reasons we can share: An attack on the distinction between agent-relative and agent-neutral values,’ in Creating the Kingdom of Ends, pp. 278ff. 35 Compare Blackburn’s comments in Ruling Passions, p. 231.

Values and the Reflective Point of View

118 36

obligation,’ and she considers ‘a slightly more attractive version’ of the knave in order to advance the case for a Kantian over a Humean approach in ethics. Korsgaard thinks that considering the more attractive version of Hume’s knave enables us to point up a problem about moral resolve in his account of the moral sentiments. The problem supposedly arises because Hume treats our moral sentiments as sympathetic responses according to ‘general rules.’ Korsgaard holds that Kantian agents, who rely on the authority of self-given laws, are better fortified. She asks us to consider the following scenario: Our knave is the lawyer for a rich client who has recently died, leaving his money to medical research. In going through the client’s papers the lawyer discovers a will of more recent date, made without the lawyer’s help but in due form, leaving the money instead to the client’s worthless nephew, who will spend it all on beer and comic books. The lawyer could easily suppress this new will, and she is tempted to do so. She is also a student of Hume, and believes the theory of the virtues that we find in A Treatise of Human Nature. So what does she say to herself?37

Well, says Korsgaard, the lawyer will tell herself that she would disapprove of herself if she concealed the new will. But being a good student of Hume, she also knows why she would disapprove of herself in that case. Our moral sentiments are responses from ‘a general point of view’ according to ‘general rules,’ and we disapprove of ourselves for acting unjustly because unjust actions tend to bring down the system of justice.38 However, the lawyer realizes that this case is the exception: suppressing the new will won’t damage justice and indeed promises to do a lot of good. Korsgaard thinks this combination of insights may well destabilize the knavish lawyer. She may come to see her feelings of disapproval as discredited on this occasion and be tempted to do what will be useful.39 Korsgaard proposes, in the spirit of Kant, to offer an account of normativity that leaves the knavish lawyer – or rather, her Kantian counterpart – better fortified in her resolve. On the approach she favors (she claims), 36 See The Sources of Normativity, pp. 55–60. 37 Ibid., p. 86. Compare Kant’s example of someone who is ‘benevolent and philanthropic’ and considering devoting an untraceable deposit to his own use in order to rescue his family from the misery of financial ruin, rather than surrendering it to heirs who ‘are rich, uncharitable, thoroughly extravagant and luxurious, so that it would make little difference if the aforesaid addition to their property were thrown into the sea’ (‘On the Common Saying: “This May be True in Theory, but it does not Apply in Practice,”’ in Hans Reiss (ed.), trans. H.B. Nisbet, Kant’s Political Writings (Cambridge, 1971), pp. 70–71). Kant argues that it would be clear ‘even to a child of around eight or nine’ that it would be wrong to appropriate the deposit. 38 On Hume’s account, sympathy explains our moral sentiments. We typically feel good about ourselves when we are good because, considering ourselves as others would regard us, we see ourselves with approval as acting in ways that generally have good effects all-round; and we typically feel bad about ourselves when we are bad because, again through the eyes of others, we see ourselves with disapproval as acting in ways that generally have bad effects all-round. 39 The Sources of Normativity, p. 87.

What’s Wrong with the Sensible Knave?

119

our sense of being justified in what we do is able to survive our understanding of how moral motivation works.40 However, Korsgaard’s story goes further. It supports Kant’s view that when knaves (even in more attractive cases) make exceptions of themselves, they fail to respect an end that is inescapable for all of us in so far as we are practical reasoners. This is the end of humanity. And in Korsgaard’s case, as in Kant’s, the identification of humanity as a necessary end of practical reason depends on the assumption that normative reasons are public, in the sense of being reasons for all. Korsgaard seeks to discharge this assumption. Korsgaard: Humanity as an Inescapable Value Korsgaard argues that if we take anything to be of value, we must acknowledge the value of humanity, and hence our moral obligations. As we saw in Chapter 5, she holds that we have normative problems because we are animals whose minds are reflective in structure. Within the deliberative setting, our perceptions and desires are objects of attention for us and, confronted by them, we must figure out what to believe and what to do. As she puts it: ‘The reflective mind cannot settle for perception and desire, not just as such. It needs a reason. Otherwise, at least as long as it reflects, it cannot commit itself or go forward.’41 Kant, says Korsgaard, describes our reflective need for a reason in terms of freedom. Within the deliberative context, because we have reflective distance from our desires, we must act under a conception of ourselves as free causalities – that is, as causalities that are entirely self-determining. However, as causalities, we must act according to some law, while as free causalities, we must act according to some self-given law; and the problem is how, as free wills or causalities, we can have any reason for giving ourselves one law rather than another. Kant’s answer to this problem is that the categorical imperative is the law of a free will. A free will will select maxims of action that conform to the categorical imperative, and all the categorical imperative tells us to do is to choose maxims that we can regard as laws.42 Korsgaard accepts Kant’s view that, as self-conscious agents, we must act under the idea of freedom. However, she introduces a distinction that Kant doesn’t make. Kant thinks that, in so far as a free will gives itself the categorical imperative, it gives itself the moral law. Korsgaard holds that Kant outreaches what he is entitled to in

40 Compare Blackburn’s gloss on the example: Ruling Passions, pp. 233–7, 40–47. According to Blackburn, we needn’t suppose that the Humean lawyer will see herself as having reason to regret her sensibilities. Moreover, says Blackburn, Korsgaard’s knavish lawyer isn’t caught up in a ‘real emergency’ – that is, an exceptional circumstance in which, by Humean lights, the conflict between the general rules and the public good is so great that the admirable thing to do is to break the rules, despite one’s reluctance to do so. 41 The Sources of Normativity, p. 93. Compare pp. 46–7. 42 Ibid., pp. 94–8.

120

Values and the Reflective Point of View

supposing this. The Kantian argument about what law a free will will give itself, she comments: … establishes that the categorical imperative is the law of a free will. But it does not establish that the moral law is the law of a free will. Any law is universal, but the argument … doesn’t settle the question of the domain over which the law of the free will must range. And there are various possibilities here. If the law is the law of acting on the desire of the moment, then the agent will treat each desire as a reason, and her conduct will be that of a wanton. If the law ranges over the agent’s whole life, then the agent will be some sort of egoist. It is only if the law ranges over every rational being that the resulting law will be the moral law.43

Korsgaard’s central proposal is that we solve the normative problem that our reflective minds force upon us by giving ourselves laws, whereby we construct practical identities for ourselves; conceptions of ourselves which keep us intact as agents or wills and under which we value ourselves. Thus, it is our autonomy that is the source of value, just as Kant said it was. Our reasons express the identities that we make for ourselves in acts of reflective endorsement, while our obligations stem from what those same self-legislated identities forbid.44 On this proposal, there are any number of practical identities that we might construct for ourselves. Korsgaard writes: An agent might think of herself as a Citizen of the Kingdom of Ends. Or she might think of herself as someone’s friend or lover, or as a member of a family or an ethnic group or a nation. She might think of herself as the steward of her own interests, and then she will be an egoist. Or she might think of herself as the slave of her passions, and then she will be a wanton. And how she thinks of herself will determine whether it is the law of the 43 Ibid., p. 99. Korsgaard borrows Frankfurt’s term ‘wanton’ here. The use of this term in such a context is potentially confusing, and Korsgaard herself isn’t consistent in its use. Frankfurtian wantons are unreflective agents. By contrast, someone who treats their desires of the moment as reasons, in accordance with their self-given law of acting on the desire of the moment, is reflective in their agency. Korsgaard is aware of this, and in the passage quoted is careful to characterize the behavior of such an agent as that of a wanton, rather than simply calling them a wanton. In other words, such an agent ends up acting like a wanton even though, being reflective, they aren’t wantons (see her remarks in note 8 on p. 99). However, later on, Korsgaard is less careful and does characterize agents like this as wantons, saying that they give themselves ‘the law of the wanton’ (see the next passage quoted in the main text). – Must the laws or principles that a free will or causality makes for itself be universal? Korsgaard argues (ibid., pp. 225–33) that, as a self-conscious causality, I need to will universally (endorse acting the same way on every relevantly similar occasion, including this one, regarded as possibly other) ‘in order to see my action as something which I do’ (p. 229). This issue of universality is important for the theory of identification. Is it really correct that identification with a psychic element like a desire can’t be effected under a singular commitment? For relevant discussion, see Bratman’s ‘Review of Korsgaard’s The Sources of Normativity,’ in Faces of Intention (Cambridge, 1999), pp. 269–70, 274–8. 44 The Sources of Normativity, pp. 100–102.

What’s Wrong with the Sensible Knave?

121

Kingdom of Ends, or the law of some smaller group, or the law of egoism, or the law of the wanton that will be the law that she is to herself.45

Thus far, this account of normativity is highly relativistic. Faced with any desire that bids us to act, we have to ask whether it’s a reason by considering whether, under some identity that we have made for ourselves, we can will the maxim of acting on that desire as a law. However, such identities are many and contingently variable. The further issue, for Korsgaard, is whether there are practical identities that we can’t avoid as reflective agents. By her lights, this is the issue of whether there are any objective values. Korsgaard argues that we cannot escape a moral identity within the deliberative perspective. Within that perspective, upon a full stretch of reflection, we will have to set a value on reflective consciousness itself, if we are to value anything at all. And that means, at least in Enlightenment terms, that we will have to give ourselves a moral identity. The argument for this, she says, ‘is just a fancy new model’ of Kant’s argument for the formula of humanity. The main details of this ‘fancy new model’ are as follows. Most of the time our local and contingent identities are the sources of our reasons for action. However, what is not contingent or avoidable is that we have some conception of our practical identity. As reflective animals, if we are to live and act at all, we need reasons, and the practical identities whereby we govern ourselves are the sources of such reasons. Therefore, if we are to continue to treat our particular practical identities as normative after further reflection, we need to endorse our identity as reflective animals with a need for other identities, as the deep identity which gives us reason to conform to others. Which is just to say that, if we are to value anything at all, we need to give ourselves a law whereby we value our own rational nature or humanity as an end in itself. Which is just to say that, if we are to value anything at all, we need to give ourselves a law whereby we constitute a moral identity for ourselves.46 Korsgaard does not suppose that our moral identity always has the last word in cases of conflict.In the terms of her story, our moral identity is reflectively inescapable, and governs our other kinds of identity in a special way: it requires us to give up identities that are inconsistent with it, and immoral identities will be ‘reflectively unstable.’47 This enables us to reaffirm an objection to sensible knaves, in the name 45 Ibid., p. 101. Korsgaard holds that all of our identities give rise to unconditional obligations, though only our fundamental identities give rise to deep obligations. All obligation is unconditional, she says, in the sense that it ‘takes the form of a reaction against a threat of a loss of identity’ (ibid., pp. 102–3). She insists that her appeal to normative identities is true to Kant: there the conception of oneself as active and rational plays a central role (ibid., pp. 237–8). Compare Wood’s comments on the centrality of self-conception in Kant’s practical theory, in Kant’s Ethical Thought, pp. 71–5. 46 The Sources of Normativity, pp. 120–23. There’s an assumption here, which I discuss in the next section – namely, that valuing our own rational nature or humanity commits us to valuing that feature in others. 47 Ibid., pp. 125–8.

122

Values and the Reflective Point of View

of reason and humanity. Knaves consciously treat themselves as important in ways that contradict the laws of an identity to which they are reflectively committed, if they are to have any values – namely, their identity as just one reflective animal among others.48 Korsgaard’s account also has the resources to say why, for the Kantian lawyer, moral commitment can survive, and indeed be reinforced, by reflective scrutiny. The Kantian lawyer will appreciate that, in doing the just rather than useful act, she conforms to the law of a deep identity that she makes for herself, and which, as a reflective agent, she needs to keep intact and cannot sacrifice, if she is to value and have reason for anything else. Korsgaard: The Public Nature of Reasons Korsgaard’s argument that we can’t reflectively escape a moral identity makes an assumption: it supposes that if we must value our own humanity, then we must value humanity in others. Korsgaard tries to make good this assumption by defending the Kantian theme that reasons are public, in the sense of being reasons for everybody. She hopes to show that, given a full stretch of reflection, our need to give ourselves universal laws does after all get us to morality because: (i) it gets us to the need to value our own humanity and (ii) we can’t value our own humanity without valuing the humanity of others, since reasons are inherently public.49 Korsgaard notes that familiar attempts to show that valuing one’s own humanity is inseparable from valuing the humanity of others ‘assume that an individual agent has private reasons, that is, reasons that have normative force for her, and then try to argue that those private reasons give the individual some reason to take the (private) reasons of other people into account.’50 She accepts that these arguments are flawed, and comments: … the kind of argument we need here is not one that shows that our private reasons somehow commit us to public ones, but one that acknowledges that our reasons were never more than incidentally private in the first place. To act on a reason is already, essentially, to act on a consideration whose normative force may be shared with others. Once that is in place, it will be easy to show how we can get someone who acknowledges the value of his own humanity to see that he has moral obligations.51

Korsgaard’s defense of the claim that reasons are essentially public relies on the idea that our shared language (and so our social nature) forces us to treat the reasons of anyone else as reasons for us. We can only not do so on pain of hearing the practical talk of others as mere noise, and this isn’t really a serious option. She writes:

48 Compare Korsgaard’s remarks on the parallel between Kant’s account of evil and hers, in ibid., pp. 249–51. 49 Ibid., pp. 219–22. 50 Ibid., p. 133. 51 Ibid., p. 136.

What’s Wrong with the Sensible Knave?

123

It is nearly impossible to hear the words of a language you know as mere noise. And this has implications for the supposed privacy of human consciousness. For it means that I can always intrude myself into your consciousness. All I have to do is talk to you in the words of a language you know, and in that way I can force you to think. The space of linguistic consciousness is essentially public, like a town square …52

Applying these considerations to reasons for action means that all it takes for me to present you with a reason is for me to voice one of mine. Korsgaard writes: If I call out your name, I make you stop in your tracks. … Now you cannot proceed as you did before. Oh, you can proceed, all right, but not just as you did before. For now if you walk on, you will be ignoring me and slighting me. It will probably be difficult for you, and you will have to muster a certain active resistance, a sense of rebellion. But why should you have to rebel against me? It is because I am a law to you. By calling out your name, I have obligated you. I have given you a reason to stop.53

Korsgaard acknowledges that there’s overstatement here. You may well have reasons to keep going that you reckon (rightly or wrongly) outweigh the reason I gave you. Still, I gave you a reason – as is clear from the fact that typically, if you decide not to stop, you will counter with a reason for continuing on – and all it took for me to give you a reason was to voice one of mine. For you not to acknowledge as much, you would have to make out that I was ‘just making noise.’54 This is a curious line of thought. Just suppose that our practical identities are sources of reasons for us. We might even suppose that we create, or largely create, these identities by being laws to ourselves, in the way Korsgaard suggests. Korsgaard is then committed to the view that your identity is a source of reasons for me, no matter how I regard that identity! Be it, in my eyes, silly, or bigoted, or self-crippling or even downright evil, it is still a source of reasons for me, just as it is for anyone. Moreover, Korsgaard holds that I can deny this only on pain of treating you as ‘just making noise’ when you voice (to my mind) your silly, or bigoted, or self-crippling or evil reasons. Now that, surely, is implausible. Perhaps it’s just because I hear and understand you only too well that I’m in a position to judge that the considerations you put forward make no demands on me.55 It’s worth noting that Korsgaard explicitly accepts that evil identities are sources of real reasons. Consider her remarks about the idealized Mafioso. G.A. Cohen presents the case as an objection to her idea that our obligations ‘always take the form of a reaction against a threat of loss of identity.’ Cohen complains that this phenomenological treatment of obligation casts the net too widely. He writes:

52 Ibid., p. 139–40. 53 Ibid., p. 140. 54 Ibid., pp. 139–45. 55 Compare Raymond Geuss’s criticisms, in ‘Morality and Identity,’ in The Sources of Normativity, pp. 197–9.

124

Values and the Reflective Point of View Consider an idealized Mafioso … This Mafioso does not believe in doing unto others as you would have them do unto you: in relieving suffering just because it is suffering, in keeping promises because they are promises, in telling the truth because it is the truth, and so on. Instead he lives by a code of strength and honour that matters as much to him as some of the principles I said he disbelieves in matter to most of us. And when he has to do some hideous thing that goes against his inclinations, and he is tempted to fly, he steels himself and we can say of him as much as of us, with the same exaggeration or lack of it, that he steels himself on pain of risking a loss of identity.56

Korsgaard discusses this example in a reply. She insists that, when he gives himself the law of being a member of the Mafiosi, the Mafioso places himself under a real obligation. She remarks: I could say that there’s no obligation here, only the sense of obligation: no normativity, only the psychic appearance of it. … But I am not comfortable with this easy way out. There is a sense in which these obligations are real – not just psychologically but normatively. And this is because it is the endorsement … that does the normative work.57

But if we allow that all it takes to make a value is to endorse it in a self-given law, and that real values are essentially public, then the Mafioso is a source of reasons for us all when he constructs his evil identity; and we can only deny this by writing him off as ‘just making noise’ when he voices considerations of strength and honor in defense of his crimes. And this is all crazy. Let’s allow the Mafioso his self-made reasons. If the Mafioso then expresses his reasons to me, it will be because I hear and understand him only too well that I will be able to tell him that his cruel values don’t make any claim on me. Korsgaard has come very far from Kant in being committed to the view that even evil identities give rise to public reasons. Kant holds that, in giving ourselves ends, we create reasons, including reasons for ourselves, only if those ends are permissible under the one and the same self-given moral law that we all share. Korsgaard, on the other hand, holds that all it takes to create a reason is to give oneself a law, which she detaches from the moral law.58 However, if we do grant this, it becomes wild to insist as well that reasons for one are reasons for all. Of course, for Korsgaard, it’s the publicity of reasons which ensures that, to the extent we value our own humanity, we must value humanity in general. She seconds an argument of Nagel’s to clinch the case:

56 In ‘Reason, Humanity and the Moral Law,’ in The Sources of Normativity, p. 183. 57 The Sources of Normativity, p. 257. Of course, Korsgaard thinks that the Mafioso should abandon his evil identity. The system of values which constitutes it could not survive ‘the full light of reflection,’ since those values contradict the value of humanity, which is the value behind all our other values. See ibid., pp. 254–8. 58 Compare Wood’s more Kantian position, in Kant’s Ethical Thought, note 30, pp. 354–5.

What’s Wrong with the Sensible Knave?

125

Suppose that we are strangers and that you are tormenting me, and suppose that I call upon you to stop. I say: ‘How would you like it if someone did that to you?’ And now you cannot proceed as you did before. Oh, you can proceed all right, but not just as you did before. For I have obligated you to stop. How does that obligation come about? Just the way Nagel says that it does. I invite you to consider how you would like it if someone did that to you. You realize that you would not merely dislike it, you would resent it. You would think that the other has a reason to stop, more, that he has an obligation to stop. And that obligation would spring from your own objection to what he does to you. You make yourself an end to others; you make yourself a law to them. But if you are a law to others in so far as you are just human, just someone, then the humanity of others is also a law to you. By making you think these thoughts, I force you to acknowledge the value of my humanity, and I obligate you to act in a way that respects it.59

The strategy of the argument is that I force you, my tormentor, to acknowledge that you can’t value your own humanity without valuing it as a good for others (like me), in so far as it’s someone’s humanity – and hence, without valuing it in a way that commits you to valuing anyone else’s humanity as a good for all (including you). But why can’t you resist the strategy here? Why can’t you simply deny that valuing your own humanity involves valuing it as someone’s humanity? Korsgaard comments: … Of course it’s true that, as Nagel observes, the argument would not go through if you failed to see yourself, to identify yourself, as just someone, a person, one person among others who are equally real. The argument invites you to change places with the other, and you could not do that if you failed to see what you and the other have in common. Suppose you could say ‘someone doing that to me, why that would be terrible! But then I am me, after all.’ Then the argument would fail of its effect, it would not find a foothold in you. But the argument never really fails in that way.

And why not? Because, says Korsgaard, in so far as you even listen to the argument, you’ve already conceded that each of us is someone. She writes: For [the argument] to fail in that way, I [your tormentor] would have to hear your words as mere noise, not as intelligible speech. And it is impossible to hear the words of a language you know as mere noise. In hearing your words as words, I acknowledge that you are someone. In acknowledging that I can hear them, I acknowledge that I am someone. If I listen to the argument at all, I’ve already admitted that each of us is someone.60

So, the claim is that there’s no available linguistic space for anyone to deny that, in valuing their own humanity, they value it in so far as it’s just someone’s. To try to occupy such space, as the case in point shows, would be to try to deny something already conceded in any linguistic interchange; namely, that one is just one person

59 The Sources of Normativity, pp. 142–3. 60 Ibid., pp. 142–3.

Values and the Reflective Point of View

126

among others. It would be a step into solipsism – ‘practical solipsism,’ as Nagel calls it – and ‘no form of solipsism is an option for us.’61 Korsgaard’s gloss on the reversed-roles argument is unconvincing. Let’s grant that, in heeding the argument, the tormentor has already conceded that he is not alone in the world. Why can’t he admit that without conceding that he values himself as just someone, in so far as he does value himself? Korsgaard badly needs an argument for the idea that we can’t keep intact our sense of ourselves as just someone without valuing ourselves in that way, in so far as we do set a value on ourselves. So, Korsgaard hasn’t, after all, shown that valuing our own humanity commits us to valuing it in others; and so, she hasn’t, after all, shown that a moral identity is necessary or inescapable within reflection. She holds that the publicity of reasons turns the trick here. But her argument fails in the detail. The turning of the trick turns out to rely on the way in which valuing our own humanity is inseparable from valuing it as a good for all, in so far as it is someone’s humanity. For then it turns out that our own humanity is a public good in our eyes, in so far as we value it, in just the way (in consistency) anyone else’s is. However, Korsgaard doesn’t offer any good defense of the idea that we can’t value our own humanity without therein (at least implicitly) valuing it as the humanity of just someone. She insists that there’s no intelligible space available for any of us to deny the connection. But that’s hardly obvious, and seems obviously false. Nagel: The Possibility of Practical Solipsism Nagel discusses the reversed-roles argument in The Possibility of Altruism. In that context, he too is defending the thesis that all reasons are (in Korsgaard’s terms) public. However, Nagel doesn’t contend that we can’t intelligibly resist the argument’s strategy – that reasons are public in a way that refuses denial. The argument gets its bite, in his view, because it appeals to a deep feature of our make-up – namely, our sense of ourselves as persons in a world of others like us. However, this selfconception is something with which we can selectively tamper. More fully, Nagel’s account of the argument goes like this. Where it works, the argument does so because the tormentor sees himself as just one person among others equally real and judges that his need for help in a reversed situation would provide a reason for anyone, simply because it is someone’s need.62 The latter impersonal evaluation is the practical expression of the tormentor’s recognition of himself as one person in a world of similar others.63 This is a conception of ourselves that runs deep: we couldn’t totally lack it without being ‘unrecognizably different.’64 However, it’s still possible, according to Nagel, to avoid setting a public (objective, agent-neutral) value on our own needs – that is, to avoid treating them as important, 61 62 63 64

Ibid., p. 143. The Possibility of Altruism (Oxford, 1970), pp. 82–4. Ibid., pp. 88–9, 116–24. Ibid., pp. 18–23.

What’s Wrong with the Sensible Knave?

127

in so far as they are just someone’s needs. We can be and ‘are often weak, cowardly, self-deceiving, and insensitive to the reality of others’; and it is, says Nagel, ‘often a struggle to maintain the clear sense of oneself as just a person among others.’65 Indeed, as Nagel sees things, we can even imagine an individual who never matches the relative value he sets on his own needs with neutral value.66 However: … when the impersonal standpoint is forced upon us it becomes difficult to resist the pressures towards congruency. When they are wronged, people suddenly understand objective reasons, for they require such concepts to express their resentment. That is why the primary form of moral argument is a request to imagine oneself in the situation of another.67

Korsgaard holds that there’s no available linguistic space for anyone to deny that, in valuing their own humanity, they value it in so far as it is just someone’s. This, she contends, would be a step into practical solipsism and ‘no form of solipsism is an option for us.’ We see this in the case in point: ‘If I listen to the argument at all, I have already admitted that each of us is someone.’ Nagel doesn’t agree that we can’t be practical solipsists. We can be, though the argument in question puts us under pressure to change. Moreover, Nagel thinks that practical solipsism leaves intact our general sense of ourselves as just one person in a world of others. It’s a selective solipsism.68 So, there’s no reason to suppose, on his account, that tormentors who listen to the argument can’t withhold valuing themselves as just someone, while readily conceding that they are one person in conversation with another like themselves. The evaluative refusal leaves intact their general ability to regard themselves as just someone. Nagel: Neutral and Relative Reasons In his book on altruism, Nagel argues that the demand of reason on us, in the practical sphere, to match our relative values with neutral (public) ones is a demand for a kind of congruence between two points of view, and is the practical expression of our deep conception of ourselves as one person among others equally real. It is, says Nagel, like the demand of reason on us to match dated reasons with timeless ones, as the practical expression of our basic conception of ourselves as individuals with a single life, equally real over time.69

65 Ibid., p. 124. See too his concluding remarks, pp. 145–6. 66 Ibid., pp. 123, 145. 67 Ibid., p. 145. 68 Ibid., pp. 123, 145. 69 Ibid., pp. 18–23, 58ff. The approach is reminiscent of Kant. For Kant, we give ourselves the categorical imperative as the practical expression of our conception of ourselves as free, a conception that forces itself upon us within the deliberative perspective. See Nagel’s comments, pp. 13–14.

128

Values and the Reflective Point of View

So, for Nagel, the way we can resist accepting the publicity or neutrality of reasons is an aspect of how reason can be defeated.70 By his lights, it marks a point of dissociation within our pattern of practical reasoning, at which our conception of ourselves as ‘I’ has come apart from our sense of ourselves as ‘someone.’ The point of dissociation is said to be a kind of solipsism – practical solipsism. This is because, as dissociated reasoners, we are unable to ascribe value to (say) our needs, considered as the needs of just someone, in exactly the same sense that we are able to ascribe value to them, identified as ours in particular. The informing idea here is that the normative content of any first-person judgment about practical reasons is able to be re-expressed as a judgment to the same effect about what anyone has reason to want or to do; and that any judgment of the former kind commits us to accepting one of the latter kind, on pain of solipsistic dissociation in the practical sphere. The addition of a perspective marker to a judgment about reasons, which locates us in the world, ‘makes a great difference in how that world is conceived, but no difference in what is conceived to be the case.’71 Of special focus is what Nagel calls motivational content. What he has to say about this is not especially perspicuous. Such content consists, he says, in ‘the acceptance of a justification for doing or wanting something.’ It’s a content that is distinctive of firstperson practical judgments, as judgments about what to do, and it means that they are sufficient by themselves, other things being equal, to explain relevant motivation or action.72 When we make judgments about what we have reason to do, we commit ourselves to judgments, with the same motivational content, about what anyone has reason to want or to do, on pain of a solipsistic breach in our sense of ourselves as just someone.73 Nagel has subsequently abandoned the thesis that all reasons are reducible to public ones. He now allows (i) that some agent-relative reasons have no public or neutral correlates at all, and (ii) that even where relative reasons can be (and are to be) matched with neutral ones, the former need not be totally subsumable under the latter.74 Examples of relative reasons without neutral correlates include so-called ‘reasons of autonomy’ (reasons stemming from idiosyncratic desires) and deontological reasons.75 Examples of relative reasons that can be matched with neutral ones include (on an optimistic count) reasons that relate to avoiding pain and

70 Ibid., pp. 20–22. As is the way we can resist their timelessness. For discussion of breaches in the sense of one’s equal reality across time, see pp. 60–62. 71 Ibid., p. 103. 72 Ibid., pp. 66ff., 109ff. 73 Ibid., Chapters 11 and 12. 74 For discussion of (i), see The View From Nowhere (New York, 1986), Chapter 9. In relation to (ii), see Nagel’s postscript to The Possibility of Altruism, and his remarks in The View From Nowhere, p. 159. The relevant postscript is quoted by Korsgaard in ‘The reasons we can share: An attack on the distinction between agent-relative and agent-neutral values,’ in Creating the Kingdom of Ends, p. 279. 75 The View From Nowhere, pp. 165–6.

What’s Wrong with the Sensible Knave?

129

having pleasure, and to the satisfaction of basic needs, as well as ones that relate to possessing freedom, self-respect, and access to opportunities and resources.76 What, then, is left of the ambitious idea that our practical reasoning is defective where we fail to match relative values with neutral or public ones? Consider relative reasons that lack neutral correlates. An example (a reason of autonomy) would be wanting to climb a high mountain or to play all the Beethoven piano sonatas. Nagel now holds that in cases like this, there’s a degree of dissociation which we can’t avoid, which doesn’t mark a point of solipsism, and which isn’t a defect of reason. We can acknowledge such relative reasons from the outside as considerations that would also be reasons for others in the same circumstances. However, we can’t acknowledge them from the outside as reasons for all. We can accord them relative objectification, but not neutral objectification.77 These claims mark an interesting development in Nagel’s theory of reasons. First, dissociation needn’t any longer be solipsistic. It isn’t solipsistic where it isn’t curable. Second, what it takes to avoid solipsism has changed: all it takes to avoid solipsism, where neutral correlates aren’t available, is to universalize relative reasons. The concern that we be able to make a normative judgment, upon objectification, to the same effect as its first-person correlate has dropped out.78 However, in that case, it’s not at all clear how the failure to universalize relative reasons could plausibly be reckoned to involve solipsism. And indeed, it was never even convincing that the failure to subsume relative reasons under neutral ones, on the assumption that the latter were always there to be had, would mark a step into solipsism. In the early account, the lack of congruence between the two evaluative perspectives meant that the dissociated agent wasn’t able to say in neutral terms what they could say in relative terms. It didn’t mean that such an agent saw the lack of congruence as irreparable in principle; and so, it didn’t mean that they thought there were practical ascriptions they could make of themselves in the first-person that they couldn’t in principle make of themselves, considered impersonally as just one person among others equally real. You might perhaps think that if the dissociated agent was committed to this, talk of solipsism could get a foothold. But the dissociated agent, in the early story, wasn’t committed to anything like this. Indeed, to the extent that such an agent was trying to be good, they would be working themselves round to matching their relative evaluations with neutral ones with the same content. So, it was never really clear how talk of solipsism got any serious and literal purchase in Nagel’s account.79

76 Ibid., pp. 165, 171. 77 Ibid., pp. 159, 167–71. 78 Compare Nagel’s remarks in ibid., pp. 152–4. 79 Compare Nicholas L. Sturgeon, ‘Altruism, Solipsism, and the Objectivity of Reasons,’ The Philosophical Review, (1974): 374–402. Sturgeon challenges the idea that solipsism threatens even if we accept that some things we can say about ourselves in the first-person aren’t available to be said impersonally (pp. 384–9).

130

Values and the Reflective Point of View

Let’s now consider the situation, according to the amended story, in cases where relative reasons do have neutral correlates. In what sense, if any, does a failure to match reasons mark a defect? One problem here is knowing how to tell whether considerations we accord relative value also have neutral value. Nagel thinks that primitive pleasures and pains are exemplars of relative reasons for pursuit and avoidance that do have neutral correlates. This, he holds, is conceivably, but not credibly, deniable. When we reflect from the outside on pain that we are experiencing, the only credible judgment is: ‘This experience ought not to go on, whoever is having it.’80 Nagel still invokes notions of integration and dissociation in the present context. However, their role is entirely changed, with the abandonment of any general reduction of relative to neutral values. What’s wrong with failing to accept that pleasure is a neutral good and pain a neutral bad isn’t that we are thereby engaged in a form of reasoning where a lack of congruence between our inside and outside evaluations marks out a stretch of solipsism in our practical thinking. What’s wrong with not accepting these neutral goods and bads is that it’s a denial that’s too hard to credit. So, someone who treats their pain as a relative but not a neutral bad has a split attitude toward their own suffering, and the split marks a failure to grasp the evident truth that pain is a neutral bad. The failure to recognize this truth isn’t now to be faulted because it marks out a split. So, Nagel’s account has changed radically. The early story locates the fault in our reasoning, where we fail to match relative values with neutral ones, in a lack of intactness between two evaluative perspectives. But the more recent story locates the fault in our reasoning, where we fail to match relative with available neutral values, at the point at which we fail to accept a truth about neutral value that’s (virtually) irresistible. This is a defect of reason which the split marks out. By these lights, the reversed-roles argument doesn’t work in quite the same way. It still puts us under pressure to match any relative value we set on ourselves or our needs with a neutral one, but the pressure doesn’t come from a formal constraint of publicity or neutrality on reasons. It comes instead from the content of our relative value. It’s too hard to credit that what happens to any one of us doesn’t matter, except to ourselves and those who care about us; and the argument makes this truth vivid for us, by forcing each of us to concentrate on their own case.81 However, it’s not especially clear on this story why exactly there’s a defect in our reasoning if we bite the bullet and deny that our lives and fates matter, considered from the outside. Nagel comments that, in defending the idea that pleasure and pain matter neutrally, he is ‘somewhat hampered’ by the fact that he finds it self-evident.82 He then goes on to offer a defense in terms of our evaluative authority as individuals. Thus, not to acknowledge that pain is bad, from the outside, he says, would be ‘to overrule the clearest authority present in the situation’ – namely, sufferers reacting to 80 The View From Nowhere, pp. 156–62. 81 The Last Word (New York, 1997), pp. 120–22. 82 The View From Nowhere, pp. 159–60.

What’s Wrong with the Sensible Knave?

131

83

pain from the inside. The thought seems to be that failing to credit pain as a neutral bad is the practical analog of failing to be guided by one’s best evidence. It’s all very obscure. So, What’s Wrong with the Sensible Knave? We began with the question: what’s wrong with the sensible knave? Kant defends an ambitious version of the thesis that the knave illustrates a failure within reason itself. He holds that when immoral agents like knaves make exceptions of themselves, their wills are implicated in contradiction; or (equivalently, as he mistakenly thinks) they fail to respect an end or value to which we are all committed as practical reasoners. However, sensible knaves seem to escape Kant’s charge of contradiction, saved by the caution in their self-love; and whether knaves violate a necessary end by failing to respect the humanity of those they deceive depends on how secure is Kant’s idea that humanity or rational nature is such an end. Kant’s defense of this idea occurs in an argument where he assumes that normative force is public or neutral. Korsgaard agrees with Kant that humanity or rational nature is a necessary end within practical reflection. In her terms, the reflective structure of our minds makes a moral identity inescapable for us as the identity that stands behind all our other identities. What’s wrong with immoral agents like knaves is that their selfish identities are inconsistent with the moral values to which their own reflectiveness commits them. So Kant is right after all: knaves’ wills are implicated in contradiction and, in taking advantage of others, and so failing to respect their humanity, knaves violate a value which their own reflectiveness demands they respect. Korsgaard’s attempt to vindicate Kant is no less than an attempt to vindicate the claim that there are objective values. It holds out the promise of vindicating objective values without any reliance on realism. Objective values are not values that are somehow ‘out there, anyway,’ waiting to be discovered; they are values that we are forced to acknowledge within the deliberative perspective, if we are to keep any values at all. Korsgaard makes this point in criticizing Nagel’s conception of objectivity. She remarks: The ideal of objectivity is to approach as closely as possible to seeing the world from no perspective at all – in Nagel’s famous phrase, From Nowhere. In a more ‘constructionist’ view like mine, by contrast, that ideal is regarded as incoherent: the fact that we can never escape viewing the world from somewhere is not a regrettable limitation, since there is nothing that the world is like from nowhere. There may be, however, something that the world is like for knowers as such or for rational agents as such, and the quest for ‘objectivity’ – that is, the surmounting of more local and contingent perspectives – is understood as the quest to view the world from these more necessary and inescapable points of view. Practical reasons that can only be found in the perspective of rational

83 Ibid., pp. 161–2.

132

Values and the Reflective Point of View agents as such or human beings as such are ‘objective’ if we have no choice but to occupy those perspectives.84

Now, the claim that we can have objective values without realism is a claim that many would like to make good. It’s the aspiration, for example, of any expressivist who wants to find a place for objectivity of values within their metaethic. However, Korsgaard doesn’t show us a way to make good the claim: her strategy is to show that the moral point of view is one that we are committed to taking up as practical reasoners, whatever else we value, and the strategy fails. It fails, as we have seen, because she ties it to an assumption which she shares with Kant, but which she never manages to establish – namely, that reasons for one are reasons for all. Korsgaard claims the early Nagel as an ally in her defense of the idea that reasons are public; but, actually, Nagel never held (as she does) that reasons are public or neutral in a sense that refuses denial in conversation. And the later Nagel is even less of an ally, as Korsgaard realizes. For he no longer even holds that all reasons (not even all moral reasons) are public or neutral.85 The sensible knave has values that make him bad. However, as yet, we have not been given any good reason to suppose that the knave is already set against himself, because his selfish values are inconsistent with moral values to which he (like any one of us) is committed within the reflective point of view, if he is to value anything. Korsgaard, in her attempt to vindicate Kant, has not given us any good reason to suppose that the cure for his evil lies in the direction of ‘trying to get him to the place where he can see that he can’t see his way to this kind of life anymore.’86 It seems, rather, that the hardened knave will need a change of heart before he comes within the reach of reason.

84 The Sources of Normativity, pp. 245–6. 85 Of special interest, Nagel allows that reasons of the kind the knave ignores, as he lies and cheats to get on in the world where he calculates that he can get away with it – so-called deontological reasons – are relative rather than neutral. Nagel discusses what he now takes to be the essence of agent-relative wrongness in cases where deontological constraints are violated, in The View From Nowhere, pp. 175–85. Korsgaard criticizes Nagel’s backdown on the publicity of reasons in ‘The reasons we can share: an attack on the distinction between agent-relative and agent-neutral values,’ in Creating the Kingdom of Ends, pp. 275–310. 86 The Sources of Normativity, p. 257. This is how Korsgaard thinks we should respond to the Mafioso, with his evil values. I am applying her remarks to the sensible knave, with his evil values.

Appendix: Making Sense and Mental Partitions Some philosophers hold that we must be divided in our minds when we betray our values. A case in point is Donald Davidson. According to Davidson, we need to partition the mind in order to describe coherently and to explain phenomena like selfdeception and weakness of will. However, Davidson’s theory of the divided mind founders on its criterion of mental division, which relies on a conception of rational processes that is too restrictive, assimilating all rational processes of thought, or of thought and action, to coherent sequences of reasoning. Consider Davidson’s account of self-deceptive belief. Self-deceptive belief, says Davidson, includes so-called ‘weakness of the warrant’ – the cognitive analog of weakness of the will – but goes beyond it in that the weakness is self-induced. A believer who displays weakness of the warrant is irrational because such an individual violates a principle of rational belief that we all share in so far as we are interpretable. This is the requirement of total evidence, which tells us that we ought to believe what we judge to be most likely on all the available evidence. The selfdeceiver, by Davidson’s reckoning, not only displays such weakness of the warrant, they induce (or sustain) it. Often self-deceivers are also wishful thinkers, in which case, they want the proposition they deceive themselves into believing to be true. However, ‘[t]he thought bred by self-deception may be painful’ – as in cases of jealousy or pessimism.1 According to Davidson, all that is necessary for someone to be self-deceived in believing a proposition p is that there be self-induced weakness of the warrant, where (against a suitable background of desire) the originating motive for inducing (or sustaining) the belief that p is a contradictory belief or the belief that one’s evidence favors such a contradictory belief.2 In developing his account of the divided mind, Davidson considers an example of self-deception that does involve wishful thinking. The example exhibits the following pattern. Given a desire that p and fear that very likely not p, a self-deceiver wants to avoid believing what their principle of rational belief, the requirement of total evidence, tells them to believe. This desire to avoid a painful belief prompts the agent to act in ways that promote the comforting belief, which they come to hold. Davidson insists that, in core cases, such self-deception will involve explicit irrationality, for it is awareness that the evidence favors the painful belief which motivates the self-deceiver’s efforts at promoting the alternative, pleasing one

1

‘Deception and Division,’ in Jon Elster (ed.), The Multiple Self (Cambridge, 1985), p.

2

Ibid., pp. 81–9.

87.

Values and the Reflective Point of View

134

– efforts that they will need to sustain when this belief is threatened by reality or memory.3 According to Davidson, the hard thing to explain, in such a case, ‘is how a belief, or the perception that one has sufficient reasons for a belief, can sustain a contrary belief,’4 and he proposes his theory of division as rational breakdown as a solution to this conceptual problem. The theory proposes that, between any such (obviously) conflicting beliefs, there is a boundary between different, though strongly overlapping, parts of the mind, thus allowing (in the case of contradictory beliefs) even open self-deceivers to be doublethinkers without being believers in a flat selfcontradiction. The boundary occurs at that point in the intervening causal sequence where there is a mental cause that is not a reason for what it causes; and in this way, the boundary of a subdivision is defined by the ‘breakdown of reason-relations.’5 In wishful self-deception, the relevant point of nonrational mental causation is where the self-deceiver’s desire not to believe what their principle of rational belief recommends is effective without being a reason against the principle.6 This example of self-deceptive belief has a certain complexity. In other discussion, Davidson proposes a simple case of irrational wishful thinking to clarify his theory that division within the mind is marked by ‘the breakdown of reason-relations.’ The case is as follows: Suppose a young man very much wishes he had a well-turned calf and this leads him to believe he has a well-turned calf. He has a normal reason for wanting to have this belief – it gives him pleasure. But if the entire explanation of his holding the belief is that he wanted to believe it, then his holding the belief is irrational. For the wish to have a belief is not evidence for the truth of the belief, nor does it give it rational support in any other way.

Davidson comments:

3 Ibid., pp. 89–91. 4 Ibid., pp. 89, 79–80. 5 This is how Davidson summarizes his theory of division in ‘Paradoxes of Irrationality’, in Richard Wollheim and James Hopkins (eds), Philosophical Essays on Freud (Cambridge, 1985), p. 304. Davidson eschews the idea that the resultant parts are separate centers of agency. See pp. 291, 303–4. This is a major difference between Davidson’s theory of the divided mind and the related functional account of division that David Pears develops to explain motivated irrational belief. The Pears references are: Motivated Irrationality (Oxford, 1984); ‘Reply to Annette C. Baier: Rhyme and Reason,’ in Ernest LePore and Brian McLaughlin (eds), Actions and Events: Perspectives on the Philosophy of Donald Davidson (Oxford, 1985), pp. 130–37; ‘The Goals and Strategies of Self-deception,’ in Jon Elster (ed.), The Multiple Self, pp. 59–77, and ‘Self-deceptive Belief-formation,’ Synthese, 89 (1991): 393–405. I criticize Pears’ theory of mental division, and compare it with Davidson’s account, in ‘Two Theories of Mental Division,’ Australasian Journal of Philosophy, 72 (1994): 302–16. 6 ‘Deception and Division,’ pp. 91–2.

Appendix

135

In standard reason explanations … not only do the propositional contents of various beliefs and desires bear appropriate logical relations to one another and to the belief, attitude or intention they help explain; the actual states of belief and desire cause the explained state or event. In the case of irrationality, the causal relation remains, while the logical relation is missing or distorted. In the cases of irrationality we have been discussing, there is a mental cause that is not a reason for what it causes. So in wishful thinking, a desire causes a belief. But the judgment that a state of affairs is, or would be, desirable, is not a reason to believe it exists.7

Davidson also applies his theory of mental division to another more complex example, that of akratic action. The akrates, says Davidson, contravenes his own principle of rational action, the principle of continence, which tells him that he ought to do what he holds to be best, everything considered.8 The following is an illustration of what goes wrong in such a case. A man who judges that it would be best, all things considered, not to return to a park, nevertheless is tempted to do so, and so not to act on his principle of continence, because (on second thoughts) he wants to replace a branch (which he has already shifted as a precaution) to its original position in order to remove a danger. The desire to replace the branch eventually causes the fellow to return to the park, thus ignoring his principle. This desire, says Davidson, does and does not provide him with a reason for what he does. It gives him a reason not to act on his principle, without giving him a reason against the principle. It’s not entirely clear how to read this example as a case where there is a mental cause that is not a reason for its effect. Presumably, Davidson means something like the following: the desire to replace the branch gives the agent a reason to return to the park, and so to ignore his principle; but it fails to provide him with a principlerelated reason not to act on his principle by returning; and hence, there’s a point in its operation as a cause of what he does where the desire to replace the branch functions as a mental cause that is not a reason for what it causes.9 The mark of mental division, as Davidson construes it, is the presence of nonrational mental causes. This means that there can be division without irrationality, since mental causes that are not reasons for their effects are not restricted to irrational cases. Of special interest are attempts at self-reform. In speaking to the example of self-reform, Davidson writes: What I have in mind is a special kind of second-order desire or value, and the actions it can touch off. This happens when a person forms a positive or negative judgment of some of his own desires, and he acts to change these desires. From the point of view of the 7 ‘Paradoxes of Irrationality,’ pp. 297–8. 8 In ‘Paradoxes of Irrationality’ (p. 197), Davidson allows that an agent might not subscribe to the principle of continence, and so might not be internally irrational in acting against their own better judgment. However, in ‘Deception and Division’ (pp. 81, 83–4), he treats subscription to the principle of continence, like subscription to the principle of total evidence, as constitutive of interpretability. 9 For Davidson, of course, all actions, under description as intentional performances, are mental events. See ‘Paradoxes of Irrationality,’ pp. 292ff.

136

Values and the Reflective Point of View changed desire, there is no reason for the change – the reason comes from an independent source, and is based on further, and partly contrary, considerations. The agent has reasons for changing his own habits and character, but those reasons come from a domain of values necessarily extrinsic to the contents of the views or values to undergo change. The cause of the change, if it comes, can therefore not be a reason for what it causes. A theory that could not explain irrationality would be one that also could not explain our salutary efforts, and occasional successes, at self-criticism and self-improvement.10

The following, I take it, is a case in question. Suppose Jack reviews his life and judges that he would do better if he could take a genuinely altruistic interest in Jim, whom he likes. That way, he reckons, he could enjoy a real friendship with Jim. Suppose, further, that Jack acts on his desire to change himself and succeeds: he now enjoys a real friendship with Jim, whose welfare matters to him simply because it’s Jim’s welfare. The situation in such a case, as Davidson reads it, is this. Jack has, in the consideration that it would benefit him to be altruistic in his concern for Jim, a reason to want to change his pattern of desire. However, the consideration of benefit to oneself isn’t a reason for desire from the point of view of an altruistic pattern of concern; from that perspective, the only consideration that is a reason is the good of the other. So, in the process wherein Jack acquires his altruistic concern for Jim, his desire for change functions as a cause that is not a reason for what it causes. Let’s examine, first, Davidson’s simplest example of a case where there is supposedly a need to partition the mind because of the presence of a mental cause that is not a reason for its effect. This is the case of simple wishful belief. The guiding idea is this: the young man comes to believe that he has a well-turned calf (in the absence of any evidence that this is so) just because he wants to have this belief for the sake of the pleasure it gives him; and so, while he has a reason for the desire that causes him to have the belief, he does not have a reason for the belief itself. Davidson construes the young man’s desire as a positive evaluation of the state of affairs in which he believes he possesses a shapely calf; and he claims that, since the desire causes the belief, without having a content that supports its truth, the example is one where there is a mental cause that is not a reason for what it causes. Let’s leave aside, for the moment, Davidson’s assimilation of the young man’s desire to an evaluation. In his gloss on the case, Davidson assumes that the consideration of pleasure is not really a reason for the fellow’s belief about his body, since it isn’t a truth-related consideration. It’s really a reason for him to want to have this belief (for him to count as desirable the state of affairs in which he has this belief). This is the feature (along with the supposed evaluative content of the desire) that ensures that, when the desire causes the belief, the sequence is one of reasonless causation.11 However, Davidson’s construal of the example is skewed by his supposition that rational processes of thought just are coherent sequences of reasoning. For, while it’s 10 Ibid., p. 305. 11 Compare Davidson’s remarks on reasons for belief in ‘Paradoxes of Irrationality,’ pp. 297–8, and ‘Deception and Division,’ pp. 85–6.

Appendix

137

undeniable that, when the young man forms his belief because of his desire for it, he doesn’t do so as a matter of reasoning or inference, it’s also evident that we can, after all, make sense of his belief in terms of his desire (evaluation) and its ground: he’s in the mental state that makes sense in the light of his desire to be in it (positive evaluation of being in it) for the reason that it yields pleasure. It’s just that we can’t make sense of his belief in a certain way: we can’t see the belief as an inference from the desire (evaluation). Davidson is concerned with so-called ‘blind causes.’ Where there’s a mental cause that isn’t a reason for what it causes, the relevant cause, according to him, is revealed as operating in the mode of a physical force, rather than in the mode of a mental cause, and the breach in the pattern of reason-relations is the mark of division within the mind.12 However, in the example under discussion, at the point where the desire (value) causes the belief, there is a pattern of causation by reasons, even though there isn’t a pattern of inference. We can characterize the situation in these terms. The young man has an attitude-related reason for believing that he has a wellshaped calf – it would please him to believe this – which doubles as an object-related reason for him to want to believe as much.13 When his desire (positive evaluation) causes him to have the flattering belief, it mediates the influence of its object-related reason (the acknowledged consideration of pleasure) beyond itself to the lowerorder belief, which shares this reason as an attitude-related one. In such a case, the desire (value) is no brute force: it functions to transmit the influence of a reason that it shares with the belief across a difference in levels of attitude. The rational process involved is one of trans-level causation by a shared reason. Of course, it’s true that we can’t make sense of the young man’s belief as an inference from his desire; and it’s true that we can’t make sense of his belief in the light of anything else that he believes. However, it’s difficult to see how the way in which the desire (value) functions can be used, as Davidson recommends, as a mark of psychic division. For the desire does its normal causal work as part of a coherent movement of thought within the ‘space of reasons’ available to it. It simply isn’t open to the desire to be a cause of the belief within a sequence of inferential thought. So far, I have set aside Davidson’s collapse of desire into value. Once we separate out these attitudes, as we should, we are in a position to uncover the diversity of rational processes involved in the example. In getting from the value he sets on pleasure to the value he sets on the flattering belief as a source of pleasure, the young man reasons in a way that makes sense. In the step where the value he sets on the belief finds expression in the desire to have it, he forms a desire that makes sense in the light of his reasoning; but that step isn’t itself a further inference to an affirmable content.14 And finally, at the point where he forms the belief he wants to have, the fellow comes to have a belief that makes sense in the light of his desire 12 ‘Paradoxes of Irrationality,’ pp. 290–91, 299–301. 13 I introduce the distinction between object- and attitude-related reasons in Chapter 2, pages 9–10, 12. 14 For relevant discussion, see Chapter 2, pages 30–36.

138

Values and the Reflective Point of View

and the reasoning behind it; however, at that point too, he doesn’t draw any further inference. The desire extends the influence of his evaluative reasoning to the belief, but not as a step within an inference. It would seem gratuitous to say that the young man is divided in his mind at the point where his value causes his desire, and again at the point where his desire causes his belief, simply because these are not points within a sequence of reasoning. Rather, these are points of mental causation, within a mixed rational process of thought, which show how we can keep our psychological integrity even where there are no inferences to be had.15 Consider next Davidson’s treatment of mental division in the case of selfdeceptive belief. He sets out the following example: Carlos has good reason to believe he will not pass the test for a driving licence. He has failed the test twice before and his instructor has said discouraging things. On the other hand, he knows the examiner personally, and he has faith in his own charm. He is aware that the totality of evidence points to failure. Like the rest of us he normally reasons in accordance with the requirement of total evidence. But the thought of failing the test once again is painful to Carlos … So he has a perfectly natural motive for believing he will not fail the test, that is, a motive for making it the case that he is a person who believes that he will (probably) pass the test. His practical reasoning is straightforward: Other things being equal, it is better to avoid pain; believing he will fail the test is painful; therefore (other things being equal) it is better to avoid believing he will fail the test. Since it is a condition of his problem that he take the test, this means that it would be better to believe he will pass. He does things to promote the belief … Suppose Carlos succeeds in inducing in himself the belief that he will pass the test. He is then guilty of weakness of the warrant, for though he has supporting evidence for his belief, he knows, or anyway thinks, he has better reason to think he will fail. This is an irrational state; but at what point did irrationality enter? Where was there a mental cause that was not a reason for what it caused?16

Davidson’s answer to the last question is this. There was a mental cause that was not a reason for what it caused, and so the drawing of a boundary that walled off Carlos’s commitment to the requirement of total evidence from the rest of his mind, at the point where his desire not to believe the requirement’s painful council caused him to believe its opposite, without providing him with a reason against the requirement.17 So, Davidson’s key idea is the following. Carlos has, in the (believed) fact that it will help him to avoid pain, a reason for wanting (and getting himself) to believe that he will pass the test, which is the opposite of what the principle of total evidence tells him to believe; and when this desire (and the actions that it prompts) causes Carlos to form the belief that comforts, there’s a point in its causal history where the 15 Compare the role of a reasoned intention in producing a matching action: the intention shares its reasons with the (lower-level) action that it causes, and its function is to mediate the influence of such shared reasons beyond itself to the action within a coherent process of thought that is noninferential. 16 ‘Deception and Division,’ pp. 89–90. 17 Ibid., p. 92.

Appendix

139

desire functions as a blind force. This is at the point where the desire counteracts the influence of Carlos’s commitment to the principle of total evidence without being a step in an argument against it. Davidson’s reading of this case is unsatisfactory. For it isn’t just what Carlos wants to believe that we can make sense of in terms of the avoidance of pain; it’s also what Carlos ends up believing. And indeed, we can identify a pattern of rational causation in the causal history of that belief. Carlos has an attitude-related reason to believe that he will pass the test (and so to ignore the principle) – the avoidance of pain – which is also an object-related reason to want to believe that he will pass. When the desire makes its contribution to the formation of the lower-order belief, it transmits the influence of its (recognized) object-related reason beyond itself to the belief, which shares this reason as an attitude-related one. Far from being a blind force, in such a case, the desire plays its role in a process of causation by a shared reason across a difference in levels of attitude. Of course, it’s true that there’s a way in which we can’t make sense of what Carlos ends up believing: we can’t make sense of it in terms of the rest of what he believes. The case is, after all, one in which someone irrationally holds a preferred belief against their own better evidence, in violation of the principle of total evidence to which they subscribe. And Davidson is right that, when Carlos’s desire to believe the opposite of what the principle of total evidence tells him to defeats his commitment to that principle, the desire isn’t functioning in a chain of reasoning as a reason against the principle. However, as in the case of simple wishful thinking, it’s once again hard to see how the way in which the desire does its causal work can be used, as Davidson recommends, as a mark of mental division. For the desire, like the value it expresses, does its normal work as part of a coherent process of thought within the ‘space of reasons’ that’s open to it. Neither the desire nor the value it expresses is fit to figure in a sequence of reasoning as a reason against the principle. Moreover, if we are careful, in considering a case of self-deceptive belief like the present one, to separate out the desire that causes the belief from the value that grounds the desire, we can once again do justice to the variety of rational processes that characterize the example. There is, of course, the sequence of reasoning whereby Carlos comes to value believing that he will pass, and in terms of which we can make sense of that value. However, the case is marked by points or patterns of mental causation that yet again display how making sense within the sphere of the psychological can go beyond making sense in terms of reasoning. Thus, there’s the noninferential step wherein the value Carlos sets on believing that he will pass finds expression in his desire to have that belief. At this point, Carlos forms the desire that makes sense in the light of his evaluative reasoning. Then, there’s the rational process in which this desire contributes to the formation of the pleasing belief, thus mediating the influence of the reasoning behind it to that belief. Moreover, as part of this process, there are, as Davidson indicates, the purposive activities (like selectively attending to the favorable evidence) which are motivated by the desire. These activities – like the desire that motivates them – figure in a pattern of rational causation, helping to transmit the influence of the recognized hedonic consideration

140

Values and the Reflective Point of View

to the belief, which shares that consideration as a reason with both these causal ancestors. And finally, the activities that assist the desire in producing the pleasing belief themselves make sense as outcomes of goal-directed or teleological processes of thought in a way that differs from how we make sense of reasoned activities as outcomes of inferences.18 Let’s now turn to Davidson’s construal of the akrates as a divided mind. The akrates in the example cited, Davidson claims, both does and does not have a reason for what he does. The desire to replace the branch gives him a reason to return to the park, and so to ignore his principle, without giving him a reason against the principle; and hence, there’s a point in the causal history of what he does where the desire to replace the branch functions as a blind cause, marking a division within his mind. At this point, that desire foils the akrates’s commitment to the principle of continence, without figuring as a step in a sequence of reasoning that defeats the principle. Thus, while the akrates has, in the desire to replace the branch, a reason to ignore his principle by returning to the park, he lacks a principle-related reason for not acting on his principle. By these lights, as seems correct, we can make sense of the akrates’s action up to a point, even if there is something ‘essentially surd’ in it. The problem is whether to follow Davidson in supposing that, where the desire to replace the branch overrides commitment to the principle of continence, it’s working as a blind cause across a boundary between parts of the mind, just because it isn’t operating in a sequence of reasoning as a reason against the principle. And again, the proposal seems misguided. For neither the desire nor the reasoned value it expresses are suitable to play such a normative role. Each of these mental elements does its normal causal work in such intelligible ways as are available to it – including the transmission of the influence of reasons. Finally, let’s consider Davidson’s view that self-reformers are divided in their minds. Earlier, I gave the example of Jack’s change for the better. Davidson, it seems, would hold that Jack is divided because the desire to become altruistic towards Jim, which prompts the change in him, works on behalf of a consideration – the acknowledged benefit to himself of changing in this way – that isn’t relevant as a reason for desire from the point of view of an altruistic heart; and so, the case is one where a desire (or value) operates as a mental cause that is not a reason for what it causes. Again, the proposal for division is odd. Here we can make sense of changed Jack’s desire to help Jim. Jack is now altruistic: he wants to help Jim just because it’s Jim, his friend. And we can make sense of why Jack wants to be altruistic in this way towards Jim: he values it as enrichening his life. The case is one where there’s a pattern of trans-level causation by a reason. Jack has an attitude-related reason to be altruistic in his concern for Jim – it will make his own life go better – which is also an object-related reason for him to want to be concerned like this for Jim. When Jack’s desire to be altruistic in his concern for Jim causes him (with 18 For relevant discussion, see Chapter 4, pages 66–8.

Appendix

141

the help of relevant action) to change in the desired way, it mediates the influence of its object-related reason (the acknowledged consideration of self-benefit) beyond itself to the new pattern of altruistic concern, which shares that consideration as an attitude-related reason.19 So, Jack’s desire for self-change is no blind force: it does its causal work at the behest of a reason within a process of causation by a shared reason across a difference in levels of attitude. It surely is ill-conceived to suppose that the way in which Jack’s desire works is a mark of division within his mind just because the coherent process of thought (and action) to which it contributes isn’t a sequence of reasoning.

19 Jack has, in the consideration of self-benefit, an attitude-related reason to want to help Jim simply for the sake of doing so.

This page intentionally left blank

Bibliography Anscombe, G.E.M., Intention, 2nd edn (Oxford: Basil Blackwell, 1976). Blackburn, Simon, Spreading the Word (Oxford: Oxford University Press, 1984). ——, ‘Realism, Quasi, or Queasy,’ in John Haldane and Crispin Wright (eds), Reality, Representation, and Projection (Oxford: Oxford University Press, 1993), pp. 365–83. ——, Ruling Passions (Oxford: Oxford University Press, 1998). ——, ‘Wittgenstein, Wright, Rorty and Minimalism,’ Mind, 107 (1998): 157–81. ——, Think (Oxford: Oxford University Press, 1999). ——, ‘Replies,’ Philosophy and Phenomenological Research, 65 (2002): 164–76. Braddon-Mitchell, David and Jackson, Frank, Philosophy of Mind and Cognition (Oxford: Blackwell, 1996). Bratman, Michael E., Intention, Plans, and Practical Reason (Cambridge, MA: Harvard University Press, 1987). ——, Faces of Intention (Cambridge: Cambridge University Press, 1999). ——, ‘Reflection, Planning, and Temporally Extended Agency,’ The Philosophical Review, 109 (2000): 35–64. ——, ‘Hierarchy, Circularity, and Double Reduction,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt (Cambridge, MA: MIT Press, 2002), pp. 65–85. ——, ‘A Desire of One’s Own,’ Journal of Philosophy, 100 (2003): 221–42. ——, ‘Autonomy and Hierarchy,’ Social Philosophy and Policy, 20 (2003): 156– 76. Broome, John, ‘Normative Requirements,’ in Jonathan Dancy (ed.), Normativity (Oxford: Blackwell, 2000), pp. 78–99. ——, ‘Normative Practical Reasoning,’ Proceedings of the Aristotelian Society, supp. vol. 75 (2001): 175–93. Cohen, G.A., ‘Reason, Humanity, and the Moral Law,’ in Christine M. Korsgaard, The Sources of Normativity, ed. Onora O’Neill (Cambridge: Cambridge University Press, 1996), pp. 167–88. Davidson, Donald, Essays on Actions and Events (New York: Oxford University Press, 1980). ——, ‘Expressing Evaluations,’ The 1982 Lindley Lecture, Pamphlet Series (Lawrence, KS: Department of Philosophy, University of Kansas, 1984), pp. 3–20. ——, ‘Deception and Division,’ in Jon Elster (ed.), The Multiple Self (Cambridge: Cambridge University Press, 1985), pp. 79–92.

144

Values and the Reflective Point of View

——, ‘Paradoxes of Irrationality,’ in Richard Wollheim and James Hopkins (eds), Philosophical Essays on Freud (Cambridge: Cambridge University Press, 1985), pp. 289–305. ——, ‘Reply to Christopher Peacocke,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford: Oxford University Press, 1985), pp. 207–11. ——, ‘Reply to Michael Bratman,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford: Oxford University Press, 1985), pp. 195–201. ——, ‘Reply to Paul Grice and Judith Baker,’ in Bruce Vermazen and Merrill B. Hintikka (eds), Essays on Davidson: Actions and Events (Oxford: Oxford University Press, 1985), pp. 201–6. de Beauvoir, Simone, The Woman Destroyed, trans. Patrick O’Brien (London: William Collins, Fontana Books, 1971). Dunn, Robert, The Possibility of Weakness of Will (Indianapolis, IN: Hackett Publishing, 1987). ——, ‘Reasons, Attitudes and the Breakdown of Reasons,’ Philosophia, 21 (1991): 1–16. ——, ‘Two Theories of Mental Division,’ Australasian Journal of Philosophy, 72 (1994): 302–16. ——, ‘Motivated Irrationality and Divided Attention,’ Australasian Journal of Philosophy, 73 (1995): 325–36. ——, ‘Knowing What I’m About To Do Without Evidence,’ International Journal of Philosophical Studies, 6 (1998): 231–52. ——, ‘Is Satan a Lover of the Good?’, Ratio, 13 (2000): 13–27. ——, ‘Moral Psychology and Expressivism,’ European Journal of Philosophy, 12:2 (2004): 178–98. Edgley, Roy, Reason in Theory and Practice (London: Hutchinson University Library, 1969). Evans, Gareth, The Varieties of Reference (Oxford: Oxford University Press, 1982). Frankfurt, Harry G., The Importance of What We Care About (Cambridge: Cambridge University Press, 1988). ——, Necessity, Volition, and Love (Cambridge: Cambridge University Press, 1999). ——, ‘Reply to Jonathan Lear,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt (Cambridge, MA: MIT Press, 2002), pp. 293–7. ——, ‘Reply to Michael E. Bratman,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt (Cambridge, MA: MIT Press, 2002), pp. 86–90. ——, ‘Reply to Richard Moran,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt (Cambridge, MA: MIT Press, 2002), pp. 218–225.

Bibliography

145

Gaut, Berys, ‘The Structure of Practical Reason,’ in Garrett Cullity and Berys Gaut (eds), Ethics and Practical Reason (Oxford: Oxford University Press, 1997), pp. 161–88. Gallois, André, The World Without, The Mind Within (Cambridge: Cambridge University Press, 1996). Genet, Jean, Funeral Rites (St. Albans, Hertfordshire: Panther Books, 1973). ——, The Thief’s Journal (Harmondsworth, Middlesex: Penguin Books, 1975). Geuss, Raymond, ‘Morality and Identity,’ in Christine M. Korsgaard, The Sources of Normativity, ed. Onora O’Neill (Cambridge: Cambridge University Press, 1996), pp. 189–99. Grice, H.P., ‘Intention and Uncertainty,’ Proceedings of the British Academy, 57 (1971): 263–79. Hampshire, Stuart, Freedom of the Individual (London: Chatto and Windus, 1975). Hare, R.M., The Language of Morals (Oxford: Oxford University Press, 1952). ——, Freedom and Reason (Oxford: Oxford University Press, 1963). ——, Practical Inferences (London: Macmillan, 1971). Herman, Barbara, The Practice of Moral Judgment (Cambridge, MA: Harvard University Press, 1993). Hume, David, Enquiries Concerning Human Understanding and Concerning the Principles of Morals, ed. L.A. Selby-Bigge, 3rd edn, revised by P.H. Nidditch (Oxford: Oxford University Press, 1975). Jackson, Frank, Oppy, Graham and Smith, Michael, ‘Minimalism and Truth Aptness,’ Mind, 103 (1994): 287–302. Kant, Immanuel, Groundwork of the Metaphysic of Morals, trans. H.J. Paton (London and New York: Routledge, 1948). ——, ‘On the Common Saying: “This May be True in Theory, but it does not Apply in Practice,”’ in Kant’s Political Writings, ed. Hans Reiss, trans. H.B. Nisbet (Cambridge: Cambridge University Press, 1971), pp. 61–92. ——, The Critique of Practical Reason, in Practical Philosophy, trans. and ed. Mary J. Gregor, with Introduction by Allen Wood (Cambridge: Cambridge University Press, 1996). Korsgaard, Christine M., Creating the Kingdom of Ends (New York: Cambridge University Press, 1995). ——, The Sources of Normativity, ed. Onora O’Neill (Cambridge: Cambridge University Press, 1996). ——, ‘The Normativity of Instrumental Reason,’ in Garrett Cullity and Berys Gaut (eds), Ethics and Practical Reason (Oxford: Oxford University Press, 1997), pp. 215–54. McDowell, John, ‘Are Moral Requirements Hypothetical Imperatives?’, Proceedings of the Aristotelian Society, supp. vol. 52 (1978): 13–29. McGinn, Colin, Ethics, Evil and Fiction (Oxford: Oxford University Press, 1997). McNaughton, David, Moral Vision (Oxford: Blackwell, 1988). Mele, Alfred R., Irrationality (New York: Oxford University Press, 1987). ——, Springs of Action (New York: Oxford University Press, 1992.).

146

Values and the Reflective Point of View

——, Autonomous Agents (New York: Oxford University Press, 1995). ——, ‘Internalist Moral Cognitivism and Listlessness,’ Ethics, 106 (1996): 727–53. Milton, John, Paradise Lost, ed. Christopher Ricks (Harmondsworth, Middlesex: Penguin Books, 1968). Moran, Richard, ‘Interpretation Theory and the First Person,’ The Philosophical Quarterly, 44 (1994): 154–73. ——, ‘Self-knowledge: Discovery, Resolution, and Undoing,’ European Journal of Philosophy, 5:2 (1997): 141–61. ——, Authority and Estrangement: An Essay on Self-Knowledge (Princeton, NJ: Princeton University Press, 2001). ——, ‘Frankfurt on Identification: Ambiguities of Activity in Mental Life,’ in Sarah Buss and Lee Overton (eds), Contours of Agency: Essays on Themes from Harry Frankfurt (Cambridge, MA: MIT Press, 2002), pp. 189–217. Nagel, Thomas, The Possibility of Altruism (Oxford: Oxford University Press, 1970). ——, The View From Nowhere (New York: Oxford University Press, 1986). ——, The Last Word (New York: Oxford University Press, 1997). Noggle, Robert, ‘The Nature of Motivation (and Why It Matters Less to Ethics than One Might Think),’ Philosophical Studies, 87 (1997): 87–111. Pears, David, Motivated Irrationality (Oxford: Oxford University Press, 1984). ——, ‘Reply to Annette C. Baier: Rhyme and Reason,’ in Ernest LePore and Brian McLaughlin (eds), Actions and Events: Perspectives on the Philosophy of Donald Davidson (Oxford: Blackwell, 1985), pp. 130–37. ——, ‘The Goals and Strategies of Self-deception,’ in Jon Elster (ed.), The Multiple Self (Cambridge: Cambridge University Press, 1985), pp. 59–77. ——, ‘Self-deceptive Belief-formation’, Synthese, 89 (1991): 393–405. Pettit, Philip, ‘Humeans, Anti-Humeans and Motivation,’ Mind, 96 (1987): 530–33. ——, The Common Mind (Oxford: Oxford University Press, 1993). ——, ‘Desire,’ Routledge Encyclopedia of Philosophy, ed. Edward Craig (10 vols, London and New York: Routledge, 1998), vol. 3, pp. 30–33. —— and Smith, Michael, ‘Backgrounding Desire,’ The Philosophical Review, 99 (1990): 565–92. —— and Smith, Michael, ‘Parfit’s P,’ in Jonathan Dancy (ed.), Reading Parfit (Cambridge, MA: Blackwell, 1997), pp. 71–95. Pink, Thomas, The Psychology of Freedom (Cambridge: Cambridge University Press, 1996). Price, Huw, ‘Defending Desire-as-belief,’ Mind, 98 (1989): 119–27. Raz, Joseph, ‘When We Are Ourselves: The Active and the Passive,’ Proceedings of the Aristotelian Society, supp. vol. 71 (1997): 211–27. Smith, Michael, ‘The Humean Theory of Motivation,’ Mind, 96 (1987): 36–61. ——, ‘On Humeans, Anti-Humeans and Motivation: A Reply to Pettit,’ Mind, 97 (1988): 589–95. ——, ‘Minimalism, Truth-aptitude and Belief,’ Analysis, 54 (1994): 21–6. ——, The Moral Problem (Oxford: Blackwell, 1994).

Bibliography

147

——, ‘Why Expressivists about Value Should Love Minimalism about Truth,’ Analysis, 54 (1994): 1–11. ——, ‘A Theory of Freedom and Responsibility,’ in Garrett Cullity and Berys Gaut (eds), Ethics and Practical Reason (Oxford: Oxford University Press, 1997), pp. 293–319. ——, ‘In Defense of The Moral Problem: A Reply to Brink, Copp, and SayreMcCord,’ Ethics, 108 (1997): 84–119. ——, ‘Which Passions Rule?’, Philosophy and Phenomenological Research, 65 (2002):157–63. Stocker, Michael, ‘Desiring the Bad: An Essay in Moral Psychology,’ Journal of Philosophy, 76 (1979): 738–53. ——, Plural and Conflicting Values (Oxford: Oxford University Press, 1990). Sturgeon, Nicholas L., ‘Altruism, Solipsism, and the Objectivity of Reasons,’ The Philosophical Review, 83 (1974): 374–402. Velleman, J. David, Practical Reflection (Princeton, NJ: Princeton University Press, 1989). ——, The Possibility of Practical Reason (Oxford: Oxford University Press, 2000). Watson, Gary, ‘Free Agency,’ Journal of Philosophy, 72 (1975): 205–20. ——, ‘Free Action and Free Will,’ Mind, 96 (1987): 145–72. Wood, Allen W., Kant’s Ethical Thought (Cambridge: Cambridge University Press, 1999).

This page intentionally left blank

Index

action, theory of 31, 101, 103 active direction of fit 28−30 Adam 63, 66 affirmable contents of belief 31–6 agency and purely first-person knowledge 47−50 theory of 1, 4−5, 50, 59, 61, 90, 96−9 akrasia and akratic action 65−6, 69−81, 89, 135, 140 altruism 136, 140 Anscombe, Elizabeth 61−2, 66 Aristotle 31 Augustine, St 104 autonomous action 96 besires 14−15, 19, 26 Blackburn, Simon 7−8, 16−19, 71, 75, 86−9, 100, 103−5, 108 ‘blind causes’ (Davidson) 137 Bratman, Michael 74−80, 92 bullying at school 18 categorical imperative 84, 109−10, 115−16, 119−20 choice as expectation, theory of (Velleman) 55−9 Cohen, G.A. 123−4 coherence, tendency towards 2−3, 14, 23−8 constitutive aim of action 97−8 continence, principle of 135, 140 contradiction test 111−17 Davidson, Donald 32, 35, 61, 66−8, 80−81, 96−104, 133−40 decision and discovery 50−55 defects of reason 109−10 deliberative perspective 37, 42, 84−90, 103, 121, 131 desire, hierarchy of 74, 90−91 desire-as-valuing mistake 5, 80, 99−103 desires as topics of reflection 83−6

directions of fit of desires and cognitions 55−6; see also active direction of fit; mind-to-world direction of fit; world-to-mind direction of fit ‘direction of guidance’ (Velleman) 56−7 discursive transforms 2, 4, 9−12, 37, 44−8 divided mind, theory of (Davidson) 133−40 drug addicts 76−7 embedded content of values and beliefs 21−2, 28 emotions 46 ends in themselves 115−17, 121 ends-in-view 76−83 epistemology 37, 43, 46, 50 evaluative thought and talk 7−17, 21, 23, 27−8, 32, 105 Evans, Gareth 38−9 Eve 63, 66 ‘evidences of truth’ 40, 58 evil values and identities 5, 62, 124 expressivism 1−16, 19−28, 70, 132 about values 22−8 false positives 113 formula of humanity114−15, 121 Frankfurt, Harry 74, 79, 90−94, 97 free will (Kant) 119−20 freedom of the will, theory of (Frankfurt) 90 Genet, Jean 64, 69−70 goodness in action 76−7 habits of inference 24 Hare, R.M. 8, 32, 35 Hume, David (and Humean thought) 5, 14−15, 19−20, 85, 107−10, 117−18 identification and ends-in-view 76−82 and self-knowledge 82 theory of 73−6, 92−5

150

Values and the Reflective Point of View

‘identification mistake’ 95 imperatives 8, 32−5 intentional action, theory of 61, 66 internalism 2−4, 13−15, 26 irrationality 135−9 Kant, Immanuel (and Kantian-ism) 1, 5, 84−7, 107−24, 131−2 Korsgaard, Christine 5, 83−7, 90, 93, 107, 111, 117−26, 131−2 on sensible knaves 117−19 lying promises 111−14 McDowell, John 14−15 McGinn, Colin 63 McNaughton, David 18 Mafiosi 123−4 mental partitions 133−41 Milton, John 62−4, 71, 75 mind-to-world direction of fit 3, 17−20, 25−6, 30, 56, 58 modus ponens 3, 24−5, 30 Moore-paradox constraint 42−6 moral identity 121−2, 126, 131 Moran, Richard 50−54 motivational content (Nagel) 128 Nagel, Thomas 124−32 neutral goods and bads 130−31 ‘objectifying mistake’ 5, 80, 87−100, 103 objectivity and objective values 121, 131−2 optatives 32−3 pain 37−8, 130−31 passional values 34 perspectival justification 42 perspectival self-knowledge 45 perversity 4, 61, 73−4, 89, 100 Perversity Lost thesis 64−6 Pettit, Philip 24 phenomenology 15, 37−8, 43, 45, 49, 59, 88−9 Plato 108 practical deliberation 1−3, 86−9; see also delib-erative perspective

practical reasoning 80−81, 89, 93−5, 99−100, 103, 110, 114, 119, 128−32 practical reflection 4−5, 83−5, 90, 95, 103−4, 107 practical solipsism 125−8 ‘practicality requirement’ (Smith) 13, 16 ‘pro attitudes’ (Davidson) 100−104 propositional thought 31, 35 passive and active 28−9 ‘purely first-person’ knowledge 37−8, 54, 58−9, 82 of one’s own beliefs 38−43 of other mental states 43−7 within agency 47−50 purposive/purposeful activity 51, 68−71, 96, 98, 100 rationales 98−9 rationality 2−3, 14, 23−8, 35, 97; see also irrationality Raz, Joseph 73 realism 131−2 reason-providing and reason-tracking motives 81−2 reasoned activity 51, 68−70, 73, 100 reasoning, sequences of 31−2, 35 reasons, theory of (Nagel)128−30 ‘reasons of autonomy’128−9 ‘reflective endorsement’ (Korsgaard) 85 regret 34−5, 46−7 revenge 80 reversed-role argument 126, 130 Sartre, Jean-Paul 51 Satan 61−6, 69−81, 89 irrationality of 71−2 self-ascription 53−5, 104 self-deception 40, 133−4,138−9 self-esteem 108−9 self-fulfilling expectations55 self-knowledge 37−8, 82 self-reform 135−6, 140 self-spectators, evidence of 50 sensible knaves 107−22, 131−2 shy and sensitive people 14−15, 18−19 Smith, Adam 108 Smith, Michael 2−3, 13−28 passim, 107 solipsism 125−30

Index sour grapes 32−3, 82 ‘standard model’ of human action (Velleman) 96−9 Stocker, Michael 65 success in action, seeking of 70−76, 81, 83, 93−4 transcendental thinking 49 universal law 110−14, 122 values and expressivism 22−8 theory of 27−8

151

Velleman, J. David 55−9, 67 theory of agency 96−9

virtue theory 85 Watson, Gary 74, 91−2 weakness of the warrant 133, 138 Williams, Bernard 85 wishful thinking 32−3, 40, 133−6, 139 world-to-mind direction of fit 3, 17, 20−22, 27−9, 33−5, 56, 58