Personal Value

Personal Value This page intentionally left blank Personal Value Toni Rønnow-Rasmussen 1 3 Great Clarendon Stre...

Author: Toni Ronnow-Rasmussen

74 downloads 1334 Views 1MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Personal Value

This page intentionally left blank

Personal Value

Toni Rønnow-Rasmussen

1

3 Great Clarendon Street, Oxford ox2 6dp Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With ofﬁces in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York # Toni Rønnow-Rasmussen 2011 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2011 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, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by MPG Books Group, Bodmin and King’s Lynn ISBN 978–0–19–960378–7 10 9 8 7 6 5 4 3 2 1

To Elly, Anna, and Soﬁa

Preface Values accrue to objects. They are not free-ﬂoating entities that exist independently of other things: necessarily, they are ‘carried’ by objects in a wide sense of this term. Clearly, not all objects carry the same kind of value; and some (perhaps most) have more than one kind of value. Some objects have only minor value, while others seem worth ﬁghting hard for. Moral values such as justice or equality belong to this latter category. Some valuable objects—whether material in nature, such as the concert ticket we have kept all these years because something extraordinary happened that evening, or something less tangible, like having a certain memory whenever you hear a particular song on the radio—are distinctly connected with one person. They carry special value for that person. Valuable objects like these are often historical in character, tying us to some particular event in our past, but they need not do so. What is valuable might rather be a future state of affairs. It may be something you strive hard to realize (say, fame or clearing your name in a scandal); or again, it may be neither a material thing nor an abstract entity (such as a state of affairs) but rather involve some person— your children, parents, spouse, or friends. Some of these persons and objects are obviously very important to you; others carry only minor value for you. The discovery that your dog has eaten the concert ticket will probably have no real, or at least no enduring, impact on your life. Of course, people tend to react differently, and it is probably true of many objects that most of us consider to have triﬂing value that their loss or damage would make a major difference in a person’s life. Here we differ, and there is obviously room for disagreement about how much we should value such objects. In fact, we might even think that something can be of great value to a person even if it plays no actual role at all in that person’s life. I am inclined to believe that it is at least an open question what to think about a case like the following: unknown to you, you have a living child (either because you never realized you were a parent, or because a child you thought had died is in fact alive). I imagine that, in many people’s eyes, an unknown child would nevertheless be of value to its oblivious parent. There are certain patterns of behaviour, and certain attitudes, that we would expect the parent to display as a matter of what is ‘ﬁtting’, or of what one has reason to do, when one learns of the existence of a child of one’s own. So the impact of an object, or person, in one’s life is not necessarily (or so we may suspect) a measure of the value the object actually has for someone. The role of values is different: they function, generally speaking, as reason-grounded incentives or disincentives to action and the formation of attitudes. Objects like those mentioned set more or less negotiable boundaries dictating what we must, should, may, or have reason to, do or not do with them. In some cases we ought perhaps to help another person to achieve fame, or see to it that

P R E FAC E

vii

someone’s name is cleared, or simply respect a person’s wish to live life as she thinks ﬁt. It may be that we should have overcome our fear, stuck ﬁngers into the dog’s mouth, and saved the ticket. Of course, we should not do so for the sake of the ticket but rather for the sake of the person (and perhaps for the sake of the dog, too) whose value we think is at stake here. It may be that the thing to do for the sake of your friend is to buy her record, or to stop people talking behind her back. Personal values, as I will refer to these sorts of value, are what this book is about. Any attempt to analyse personal value will at some point have to face the following two questions: In what sense is personal value a kind of value? And (if it is a value) in what sense is this value personal? I will suggest that personal value is in fact a kind of value. One way to elaborate this suggestion is to say that in ascribing personal value to something we are, as I will claim, actually evaluating. That is, we are employing a particular notion of value, and one that we standardly express with the help of the following bona ﬁde locutions: value-for, good-for. Now, I might be mistaken about this in more ways than I care to think. For instance, I may be making an evaluative blunder in saying that the examples I will invoke are in fact examples of value-for. Since I will not be getting involved in substantive value theory, I will have to choose my examples carefully so that they do at least appear to be plausible bearers of personal value—that is, bearers of something that most of us would be ready to say is good for or valuable for themselves or someone else. Here I will have to trust my intuitions. (Needless to say, this will also be the case when it comes to impersonal values.) Equally, I might make another kind of mistake. I want to work with the assumption that there is a kind of evaluation that is conducted in terms of notions such as good-for, but it might be objected that there is no such notion as good-for, and that therefore genuine evaluation cannot make use of that notion. Whenever we speak about what is good for person a we are merely reporting on what is good according to a, or from a’s perspective. We are not committing ourselves evaluatively, in the sense of agreeing or disagreeing with what we take to be the view of a. Sometimes, certainly, good-for judgements are indeed like this; and when they are they should perhaps be enclosed in scare quotes. But I am convinced that there are other times when we are genuinely evaluating, genuinely attributing value-for. And the two kinds of case must not be confused. The occasions on which we genuinely evaluate what is good-for someone are not rare. The best way to show this, as far as I know, is to present a plausible analysis of the more questionable notion. A caveat is now required regarding the verb ‘to value’ and my use of it. What precisely we are doing when we value something will in this book be left open. This is, of course, in some ways regrettable, but it would be even less satisfactory to set out from something that is not yet settled.1 1 This work is a contribution to mainly formal value theory. Traditionally such theory admits of two relatively independent approaches. On the one hand, we have a number of semantic theories about the

viii

PR EFAC E

What I will be proposing in this work, then, is that we should employ a certain pattern of value analysis, one which has recently attracted much attention and which has roots in the writings of the Austrian philosopher Franz Brentano. If we ﬁne-tune this kind of value analysis it will come to cover personal values, too. I believe these goals can be achieved. A fortiori such an achievement supports the idea that we are dealing with a genuine value (because, in important ways, the ﬁnely tuned analysis structurally resembles the way in which values have been analysed according to a plausible pattern). Moreover, I think this approach has many advantages. Since my proposal is novel, my concern is less with evaluating rival proposals and more with articulating and defending my own positive view. In trying to understand value-for it will be helpful, as I mentioned, to have a better grasp of value, period. But anyone who has followed value theory in recent years will know only too well that there are a number of more or less interrelated value concepts. Therefore, it will be necessary at the outset to outline the more important and fundamental ones. So before elaborating my views on the particular pattern of value analysis I endorse, I will introduce, and distinguish between, various kinds of value. The presentation of a value taxonomy often leads to discussion of the nature of things that bear value. That there must be an object of some sort that value accrues to is something over which value theorists, with very few exceptions, agree. Most philosophers would also say that many different things can be ascribed value. Thus, in everyday talk we speak of physical objects such as paintings bearing value (e.g. beauty), but we also refer to events, facts (e.g. that it is more valuable that two persons feel well than that one person feels well), places, times, and so on, as having value. The consensus that exists today on this matter—i.e. the wide agreement that there is a relation between value and something that bears this value—does not, however, dissipate when we ask questions about the nature of these value bearers. While some philosophers argue that just about everything can be a bearer of value, others have taken a much more restrictive view, and, in effect, allowed only one kind of thing to bear value. Investigation of the metaphysical nature of value bearers may at ﬁrst glance seem rather inconsequential. However, I do think there is reason to more carefully look at this issue, and not only to satisfy a philosopher’s curiosity. In fact, one of the things I will argue in Chapter 10 is that, by excluding certain types of value bearer, we limit the substantive views available. Equally, by endorsing certain types of substantive theory of meaning of our value terms (e.g. versions of cognitivism, non-cognitivism, expressivism). But not every formal value theory makes it a priority to improve our understanding of the ‘semantic meaning’ of value terms. What I will be doing in this work falls, for instance, beyond the scope of these semantic theories. Rather, the general aim of work of the kind undertaken here is to reach an understanding of value (in particular, ﬁnal personal value). A good way of doing this is to exhibit structural connections that exist in the ﬁeld of value, and in particular those holding between our central evaluative concepts. In other words, the focus is the logic of value: the relevant conceptual distinctions, the relations between different types of value, and the relations between value concepts and deontic (normative) notions such as ‘ought’, ‘reasons’, and ‘ﬁttingness’.

PR EFAC E

ix

value, we will quite naturally be led to certain positions regarding the nature of value bearers. I suspect both of these claims will be looked at with some scepticism. This work easily divides into three parts. First, there is the scene setting of Chapter 1. This is followed by a much longer part, Chapter 2 to Chapter 8, which contains the analysis, together with discussions more directly concerned with my proposal. The third part, consisting of Chapter 9 and Chapter 10, discusses some rather different ideas which have had a signiﬁcant inﬂuence on my analysis. The notion of personal value plays only a minor, indirect role in these last two chapters. Since I believe this book to be somewhat dense, I shall now brief the reader on the main structure. In Chapter 1, I introduce some contrasting notions of value. G. E. Moore’s inﬂuential views on intrinsic value, as well as his much employed, so-called isolation test, are discussed in some detail. This test, as I shall explain, might have done more bad than good to value theory in general. At least, it may well have blinded value theorists to a value notion that is central to both the chapter and the book—namely, the notion of ﬁnal value, or what is valuable for its own sake. This kind of value plays an important role throughout the book. Chapter 1 also discusses the common claim that values are supervenient features: the notion that, whatever values are, they are dependent on something else, some subjacent set of properties (which set is generally taken to contain natural properties). In particular, I disentangle two intuitions involved in the idea that value is a supervenient feature which, regrettably, are often lumped together. These are what I call the ‘consistency intuition’ and the ‘dependence intuition’. Although closely related, these intuitions seem to be different in nature. My discussion leads me, among other things, to identify what is missing in Jaegwon Kim’s formulation of supervenience. In this work, supervenience plays a special role, in that it is one of the key elements I employ in order to get clear about the often misconstrued distinction between subjectivism and objectivism about values. Since I take it to be an advantage of my analysis of personal value that it is essentially neutral on this vexed issue, I want to make clear at the outset just how I conceive of the distinction between subjectivism and objectivism. So much has been written on this topic that it would require a very substantial work or two to do it justice. I have decided, therefore, to set out, in as much detail as is possible in the present discussion, just how I conceive of the debate, and to leave the writing of the substantial books to someone else. This is not a very nice thing to do to my forerunners, and I can only hope that they will bear with me. The two core ideas at play in this discussion are those of supervenience and value constitution. The notion of value constitution is a particularly tricky one, and I am painfully aware that much work remains to be done here. The ‘Invariance Thesis’, the claim that what is valuable for its own sake is invariant over possible worlds, is often thought to serve as a line of demarcation between subjectivists and objectivists. However, I argue that the Invariance Thesis, which often divides value theorist into two camps, is only reasonable given certain assumptions about the nature of the bearer of value. Once we take into consideration the fact that the supervenience base might contain non-essential properties it becomes clear that the thesis should be rejected.

x

P RE FAC E

Chapter 2 opens with a discussion of the ﬁtting-attitude analysis, or buck-passing account (as it is also called), that I will be employing.2 Besides commenting on two of its more important architects, Franz Brentano and A. C. Ewing, I outline the central features of the analysis and consider some possible objections to it. One objection stands out as being particularly troublesome to this pattern of value analysis. If we want to distinguish the right kind of reason from the wrong kind (where only the former justify value ascriptions), which it seems we must do, we ought to be able to say what makes a reason right or wrong. Unfortunately this has proven to be a rather stubborn problem. Chapter 3 is devoted to it. My suspicion is that ﬁtting-attitude analysts face a further, related problem—one that concerns whether we should regard value as a reason-providing property on its own. At the end of Chapter 3 I explain why I think this might, in effect, be the case. In the next chapter I formulate my analysis of valuefor. I also consider some inﬂuential views on personal value. The ﬁrst, and most important, is G. E. Moore’s famous attack on the very notion of a personal good. Another approach that I examine is a competitor to my own position. In this alternative, value-for is understood in terms of the de facto attitudes of its beneﬁciary (i.e. the person whose value it is). In discussing these views I take the opportunity to clarify my own analysis on several important points. The normative has primacy, according to my analysis, over the evaluative: value is explained in terms of what we have reasons to favour. The reason element in the analysis carries much of the weight in this view (but not all of it, as we shall see). My suggestion is that the analysis of personal values requires a special kind of attitude type—what I refer to as a for a person’s sake-attitude. Such attitudes are both very common and important; they have received less attention than they merit. In Chapter 5 I describe what I believe is a deﬁning feature of these peculiar attitudes. To do this, I introduce a distinction between two roles that properties may play in the intentional content of such attitudes. The general aim of Chapter 6 is to deepen the analysis, and I seek to do this primarily by discussing a range of issues which, I believe, my analysis accentuates. The ﬁrst concerns the possibility that more than one sense, or notion, is associated with the expression ‘good for’. One notion in particular interests me here, namely that whatever is welfare-constitutive is good for the agent. A second issue concerns the extent to which the analysis permits us to identify intrinsic personal values. I also ask to what extent personal values can be treated as ‘private’ values once they are understood in terms of agent-relative reasons for action. Finally, I consider a more complex matter. This is the question whether universalizability seriously undermines an analysis like mine. This alleged problem has nothing to do with the fact that the notion of reason employed in the analysis is likely to be agent-relative. It is generally recognized that 2

There might be something in the idea that these two analyses differ in some respects. For instance, it might be that, as a matter of fact, they have been applied to different things. However, in this work I will not seek to separate the two notions.

PR EFAC E

xi

agent-relative reasons are as universalizable as supposedly neutral ones. The obstacle seems to arise elsewhere, and in fact it turns out to be connected with the peculiar nature of the for a person’s sake-attitudes. Since at least some of these take non-fungible (i.e. non-replaceable) persons as their intentional objects, they are not obviously universalizable. There is no cause for alarm, though. The chapter concludes by highlighting some plausible responses to this challenge. Chapter 7 is devoted to discussion of the views of Thomas Hurka, Michael J. Zimmerman, Donald Regan, and Connie Rosati. Hurka and Regan have launched separate attacks on what they take to be a number of confused ideas about good-for. I try to respond to these. I also comment on Rosati’s and Zimmerman’s recent analyses of good-for. In particular, I explain why I wish to distance my views from Rosati’s suggestion as to how we should understand personal value. Moreover, I explain why I ﬁnd Zimmerman’s account of good for in terms of beneﬁt not promising. In Chapter 8, I consider remaining objections to my analysis. I also look at Stephen Darwall’s theory of welfare, and at an objection to ﬁtting-attitude analyses of good-for recently raised by Chris Heathwood. It will probably prove useful in what follows to keep in mind that, throughout the present book, I understand reasons as what are sometimes called ‘good’, ‘justifying’, or ‘normative’ reasons for action or the adoption of an attitude. The contrast here is with explanatory reasons for action and attitudes (and also with motivating reasons for action). The precise nature of these normative reasons will be left open in this work. However, I will take them to be facts favouring action or the adoption of an attitude. In other words, I am inclined to share the views of the large number of philosophers these days who think the notion of a normative reason deﬁes non-circular analysis, and who therefore treat it as a primitive. Of course, this does not mean that we cannot try to improve clarity when it comes to normative reasons; and in fact on one particular issue I hope to do just this. Thus in Chapter 9 I argue—against what I myself held for a long time—that, given plausible assumptions, all normative reasons are agent-relative. If I am right about this, a number of things follow about the dichotomy between personal and impersonal value. For one thing, obviously this distinction cannot now be accounted for in terms of the distinction between agent-relative and agent-neutral reasons. We will have to look elsewhere if we want to understand these two kinds of value. Hence my suggestion that the analysis of personal values requires the special kind of for a person’s sake-attitude. Prior to Chapter 10 I assume that very different kinds of thing can be valuable. I think this assumption is reasonable, but some value theorists have expressed reservations about such pluralism. In Chapter 10 I therefore present an argument that has implications for the question whether we should be monists or pluralists when it comes to the bearers of ﬁnal non-derivative value. The idea is quite simple. The argument sets out from the notion that a formal analysis of value should be accessible to as many plausible, substantive value theories as possible. A pair of commonly held, substantive views about value—preferentialism and hedonism—will be considered. Much speaks in their favour: pleasure seems quite obviously to be something of value and many

xii

PR EFAC E

people would agree that the satisfaction of preferences is valuable. I argue that, given certain plausible assumptions about the nature of pleasure and preference, the value accruing to pleasure (from a hedonist point of view) cannot be accounted for by preferentialists. This will probably strike many observers as a rather incredible idea. However, my reasoning is quite straightforward: the two positions are most plausibly regarded as ascribing non-derived ﬁnal value to different kinds of value bearer (even if they ascribe derived ﬁnal value to the same sort of bearer). Since these positions do not ascribe non-derivative value to the same kind of value bearer, we have reason to endorse a formal view of value that is consistent with value-bearer dualism. The focus of this book, to repeat, is not value in general but personal values. I should stress that by ‘values’ here I do not have in mind that which is valued (a new car perhaps, friendship, or medical care). I have in mind rather that which is ascribed to such things, namely value and value-for. Personal values accrue to objects in virtue of the fact that those objects have value-for, or are valuable to, someone—not in virtue of the fact that those objects have value simpliciter, or are valuable, period. The objects are considered not just valuable but valuable for someone. This suggests that the expression ‘to value’ is subject to an ambiguity. As well as thinking that objects are good or bad, we may also think they are good-for or bad-for someone. Sometimes we value in terms of the impersonal ‘to value’ (examples include judging that something is good and judging that something is bad). Sometimes—at least, this is the contention I shall try to render credible in this work—we value in terms of what is ‘good-for’ or ‘bad-for’ somebody.3 In the latter cases we are valuing something with someone in mind; in the former we are valuing something with an eye to nothing but the object of value.

3 Valuing something need not involve so-called thin value concepts such as ‘good (for)’ or ‘bad (for)’; often it will involve only thick value concepts such as ‘admirable (for)’, ‘loveable (for)’.

Acknowledgements A number of people made the work on this book a much more pleasant, albeit not necessarily easier, endeavour than I had anticipated. One person in particular who should be mentioned is Wlodek Rabinowicz. For about a decade he has been my regular lunch companion, and the generosity with which, over the years, he has served me with new ideas, arguments, and philosophical know-how simply cannot be overestimated. (My weight gain is the most tangible proof of who has been doing the talking and who has been doing the learning.) So entangled are his views with my own that I am quite certain they do not appear explicitly in enough places in this book. Were it not for my bad memory, and my equally bad custom of transforming clear ideas into less clear ones, I would have mentioned his name a good deal more often. I have been fortunate in other ways, too. David Alm, Johan Brännmark, and Michael Zimmerman, who read and criticized earlier drafts of this book, provided me with invaluable help and encouragement. They forced me to clarify the exposition and sharpen many of my arguments. I also wish to thank Erik Carlson, Finn Collin, ¨ sterberg, who were subjected to readings of a very early version. Their kind and Jan O comments were greatly appreciated. I have had a privileged working situation. My ofﬁce at the Department of Philosophy, Lund University, lies between the ofﬁces of my two good friends Bjo¨rn Petersson and Dan Egonsson. Over the years I have spent a tremendous amount of time in Dan’s room, discussing more or less every aspect of life. His original and down to earth way of approaching any subject in philosophy, and especially my metaphysical problems, has been a constant reminder that I now and then ought to put on the brakes and reﬂect on what I am actually doing. Over the years I have also learned much from my discussions with Bjo¨rn about many of the book’s themes. His gentle way of managing the department makes Kungshuset such a splendid working place. Core ideas of the book were presented at various places during a hectic 2004: thus I want to thank the members of a workshop I attended at the Swedish Collegium for Advanced Study in the Social Sciences, Uppsala, and in particular David Brax, Johan Brännmark, John Broome, Krister Bykvist, Erik Carlson, Sven Danielsson, Margaret Gilbert, Jonas Olson, Wlodek Rabinowicz, the late Howard Sobel, and Daniel Svensson; thanks are also due to Michael Stocker and David Chan, and the other audience members who attended my talk at the University of Wisconsin, Stevens Point. (David did a great job in organizing such a splendid conference.) Kind invitations from the University of Stirling and King’s College, University of Aberdeen, were followed by stimulating discussions, and for this I am especially grateful to Peter

xiv

A C K N OW L E D G E M E N T S

Baumann, Michael Brady, Rowan Cruft, Adrian Haddock, Alan Miller, and other members of these audiences. In Krakow, in 2008, I attended the European Congress of Analytic Philosophy and a workshop arranged by Kevin Mulligan and Wlodek Rabinowicz. I received many acute comments on topics dealt with in the book from members of the audience, some of which I have already mentioned. In particular, however, Jonas Olson, Christian Piller, and Andrew Reisner gave me much to think about. Later that year I was invited to give a talk at the meeting Metaphysics of Value, (the ﬁrst Eidos-Thumos-Episteme meeting, convened in Geneva), and I am grateful both to the organizers and audience at that event. The latter gave me valuable feedback on a couple of arguments that ended up in the book; I owe particular thanks to Graham Oddie, Olivier Massin, Kevin Mulligan, and Nishi Shah. I was also grateful for the responses I received from Albert Musschenga, Robert Heeger, Simon Kirshin, Michael Quante, and Andreas Vieth during a visit to Münster in 2010. Lunches and dinners have played a most nourishing role in the work on this book. I had some very rewarding ones in Lisboa, and later on, with Jonathan Dancy, in Skano¨r. I am also grateful for the advice and encouragement I received from Kevin Mulligan during our occasional meetings—would that I possessed his knowledge of the Austrian–German value theorists! I learned much from correspondence with John Skorupski. I was honoured by his talk in Lund, and our discussion about reasons over lunch in Copenhagen’s Glyptotek the day after it helped me steer clear of some bad ideas. Åsa Wikforss and I had some not-so-good sushi in Lund, but fortunately the dialogue was all the better; she gave me insights into areas of philosophy of which I am quite ignorant. Michael Zimmerman merits a special place in this list. My friendship with Michael goes back many years now, and I am really very thankful for his great support and excellent advice, and for the arguments we have had on a wide range of topics over the years. I have learnt much from the people at my department, especially those who attend the weekly Higher Seminar in Practical Philosophy. I have also beneﬁted from occasional comments and email correspondence with a number of people over the years. I cannot thank all who merit my gratitude, in part because of their number, but chieﬂy because my memory does not allow me to recollect all of the many times that I have beneﬁted in this way. But I can at least mention some: Gunnar Bjo¨rnsson, Fred Feldman, Fritz-Anton Fritzson, Alan Goldman, Christopher Grau, Ish Hadji, Chris Heathwood, Noah M. Lemos, Ia Maurin, Derek Parﬁt, Ingmar Persson, Douglas W. Portmore, Connie Rosati, Tim Scanlon, Mark Schroeder, Caj Strandberg, Folke Tersman, Alex Voorhoeve, and Heath W. White. Comments sent by two anonymous readers from Oxford University Press were very helpful in the later stages of revision. Paul Robinson polished the English in this work, for which I am very grateful. My work was also supported by generous grants from The Bank of Sweden Tercentenary Foundation and the Swedish Research Council. This work includes, with the kind permission of Springer Science and Business Media, revised portions of my articles: ‘On For Someone’s Sake Attitudes’ (2009b),

A C K N OW L E D G E M E N T S

xv

‘Normative Reasons and the Agent-Neutral/Relative Dichotomy’ (2009a), ‘Love, Value and Supervenience’ (2008b), ‘Buck-Passing Personal Values’ (2008a), ‘Analysing Personal Values’ (2007a), ‘Instrumental Values—Strong and Weak’ (2002b), and ‘Hedonism, Preferentialism, and Value Bearers’ (2002a). I dedicate this book to my wife and our two daughters.

This page intentionally left blank

Contents Preface Acknowledgements 1. Types of Value 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9

Preparing the ground—different value concepts Isolating intrinsic value Isolating extrinsic ﬁnal value Competing notions of intrinsic value The source of value Supervenience The Invariance Thesis Constitution Subjectivism and objectivism

2. Fitting-attitude Analysis 2.1 2.2 2.3 2.4 2.5 2.6

The origins A. C. Ewing The primacy of the normative over the evaluative Different value idioms The FA analysis of ﬁnal positive values W. D. Ross’s objection

3. The Wrong Kinds of Reason 3.1 3.2 3.3 3.4 3.5

The wrong kind of reason problem Dual-role attitudes Solving the WKR problem? Danielsson and Olson on the WKR problem The biconditional buck-passing account

4. Mistaken Value Analyses 4.1 Moore’s objection to good-for 4.2 Instrumental value 4.3 A bad approach to personal value

5. For Someone’s Sake 5.1 5.2 5.3 5.4 5.5 5.6

Types of attitude Sake, end, and objective Discerning attitudes Identiﬁers and justiﬁers Two kinds of FSS attitude Summing up

vi xiii 1 2 3 5 7 9 10 12 13 14

19 20 22 24 25 27 30

33 33 37 38 40 42

46 49 51 53

55 57 60 63 68 73 75

xviii

CONTENTS

6. Examining the Analysis 6.1 6.2 6.3 6.4 6.5 6.6 6.7

Dropping an FSS attitude Non-fungible persons and identity Good-for and welfare Two sorts of personal value Personal intrinsic value Agent-relativity and private values Personal value and universalizability

7. Mo(o)re Objections 7.1 7.2 7.3 7.4 7.5

Hurka’s views on good for Regan’s views on good for Rosati and good occurring in a life Personal and impersonal value Thin-thin and thin-thick conceptions

8. Problems and Possibilities 8.1 8.2 8.3 8.4 8.5 8.6

Janus values Some counter-examples Darwall on welfare Different strategies Heathwood’s objection The good life and the argument from fetishism

9. One Reason Dichotomy Less? 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8

Introductory notes The essentialist sense The number approach Reason-for Meeting the challenge An overcrowded boat More replies to the challenge A positive argument

10. Value Bearers and Value Pluralism 10.1 10.2 10.3 10.4 10.5 10.6 10.7 10.8 10.9

Extrinsic ﬁnal values Hedonism and preferentialism Preferentialism Value bearers Value-bearer monism Simplicity Some reductions Separating the concrete from the abstract Recapitulating

Bibliography Index

77 78 80 83 86 88 91 92

95 96 98 99 104 105

109 110 111 115 118 121 122

126 127 128 131 132 137 140 141 144

152 153 156 158 159 160 161 163 166 168

172 181

1 Types of Value

Certain things we value are best described as possessors of impersonal value. That the world is just, and that there are more happy people than there are unhappy people, are states of affairs that are, in a clear enough sense, impersonally valuable: such states of affairs are valued, when they are, in themselves, without reference to anything else. Other things appear to carry a different kind of value: they are seemingly personal rather than impersonal in nature. They are examples of value-for rather than value, period. For instance, some time ago, tidying my desk at home, I found a poem in a drawer that Anna, my oldest daughter, wrote for me when she was just a young girl. It was a typical child’s poem. Despite the fact that my room was cluttered with paper, I could not throw it away. The piece of paper, with its few lines of Swedish, had some value for me, i.e. it was a value that related in some way to me as a person in a way other valuable things do not. It was clear to me that although this poem carried value, it was not the sort of value that people are ready to ascribe to, say, Baudelaire’s Les Fleurs du Mal or a sonnet by Shakespeare. Many people (whether they are right or wrong in this is another question) take poems, television series, rock bands, special parts of our planet, to be valuable. Now it might, of course, be the case that what remains of the Amazon jungle carries personal value for someone, as my daughter’s poem does for me, but whether or not this is the case, it is also quite common to value nature that is largely untouched by man in an impersonal way. Wild nature, untouched by human beings, is just valuable. So is the fact that more people feel well than feel bad, or that justice rather than injustice prevails. In contrast with my daughter’s poem, these facts do not carry person-relative values—they are just valuable, period. This distinction between value, period, and value-for, or—as I shall also be referring to it—between impersonal and personal value, is just one among many distinctions in the area of value that philosophers have focused on. However, at this point it is probably not wise to proceed directly to an analysis of personal value. To set the stage something needs to be said about the various types of value that have played an important role in modern value theory; personal values will be subsumable under at least some of these value types.

2

T Y P E S O F VA L U E

I will turn to this in a moment. First of all, however, I must enter a minor but not unimportant caveat: in what follows, I shall take the idioms value for and good for to refer to personal values. Likewise, ‘value’ and ‘good’ will be terms for impersonal values. ‘Good for’ will be taken to refer to a positive value for, ‘good’ to a positive value, ‘bad’ and ‘bad for’ to negative values. (There are other, so-called ‘thick’ evaluative terms, such as ‘desirable’ and ‘admirable’. We shall consider these later, but for the moment I wish to focus on ‘value’ and ‘goodness’.) Often the context in an example being discussed makes it quite clear that ‘value’ refers there to ‘positive value’, but sometimes I shall not specify the sense in which I am using the word ‘value’. The beneﬁts of making it clear, on each occasion, what kind of value—positive, negative, or neutral— I have in mind are for the most part extremely limited. Little clarity is gained, I have found, from such an approach. ‘Value for’ and ‘good for’ (and their negations) are expressions that can reasonably be expected to carry more than one sense, and it is quite likely that the analysis I present in Chapter 2 and on will fail to suit all the different senses of these expressions. Much later on I will discuss a couple of these, as it were, errant senses; although at least some of them are, from a value-theoretical perspective, less interesting, they need to be commented on.

1.1 Preparing the ground—different value concepts There are many ways to sort values. In fact, philosophers of value in the nineteenth century concerned themselves a great deal with sorting tasks. Values were organized into more or less important sets of subject matter, such as moral, aesthetic, cultural, and natural values; they were ranked and placed along one scale or another. For instance, higher values would typically be those which (in contrast to lower values) play, or ought to play, an important role in our lives. However, none of these tasks will be mine in this work. My concern is a different one. The value notions that I want to draw attention to next are in some respects philosophers’ phrases of art. Notwithstanding this, many of these notions do play an important role in people’s everyday reasoning about what is valuable. Parents (and grown-up children, for that matter), are likely to recognize, for instance, the perhaps most common of all value distinctions—namely, that between something being valuable as a means and something having ﬁnal value. The child who asks why she must brush her teeth, and who is not satisﬁed with ‘It is what every good boy or girl ought to do!’ will soon turn the conversation into a search for the ﬁnal value, or end, that makes it no longer necessary to ask, once again, ‘But why is that good?’ Why is having clean teeth good? Why is fresh breath good? Why is it valuable to be liked by one’s fellow men? Typically (I know I did), people end up appealing to the pain connected with toothaches, or the pleasures that come from having healthy teeth, or both. The history of the distinction between something’s being valuable as a means to something—what is commonly referred to as ‘instrumental value’—and something’s

1.2

I S O L AT I N G I N T R I N S I C VA L U E

3

being valuable, not as a means, but for its own sake, is impressive.1 Nonetheless, it is a distinction which, over the years, has been expressed in a number of ways. Consider again the earlier example. Pleasure is one of the most common items invoked to put an end to iterated ‘But why is so-and-so valuable?’ questioning. At least many people do seem to think it is unnecessary to continue asking why experiencing pleasure is something good. It seems pointless to search for some consequence of one’s feeling pleasure that would somehow justify our conviction that pleasure is valuable.2 In the light of this axiologists3 have said that when we value pleasure in this sense we value it ‘for its own sake’ or ‘for the properties it has in itself ’, or ‘as an end’, or ‘in its own right’, instead of valuing it as a means to something else or for its consequences. In fact, recently philosophers have seriously pondered the possibility that these distinctions should be taken more seriously than has historically been the case. But it is not only ‘end’ values that have been described differently. Instrumental values, also, though to a lesser extent, have received different descriptions, as we shall see later on.

1.2 Isolating intrinsic value One major issue in recent value theory is whether something’s being valuable for its own sake (that is, what I will refer to as ﬁnal value) and something’s being valuable for the properties it has in itself (what I shall call intrinsically valuable) are coextensive. Up until recently, the distinction barely attracted attention. One major reason for this is, perhaps, that G. E. Moore, in his seminal Principia Ethica (1903a), regarded this division as being of little importance. To value something in virtue of its internal properties is to value it for its own sake and, vice versa, to value something intrinsically is to value it for its own sake, and not, say, for the sake of its consequences.4 To understand why such an approach seemed perfectly legitimate for a long while, we need to say something more about G. E. Moore’s views on value. He went to quite some pains in the Principia to determine the nature of intrinsic goodness, but arrived somewhat paradoxically at the position that the concept deﬁes analysis. And by ‘analysis’ he had in mind a quite particular activity. To analyse a concept was to 1

The distinction appears in the Timaeus, where Socrates asks what is good as such (pleasure). For more historical examples, see the introduction to Recent Work on Intrinsic Value (2005), by myself and Michael Zimmerman. 2 Although people generally have fairly strong hedonist instincts (an odd exception to this being one of Plato’s nephews, who considered both pain and pleasure to be evils), it is easy to imagine examples that create problems for this kind of intuition. Think of someone who ﬁnds pleasure in torturing another person. Again, what should we say about a case where someone ﬁnds it pleasurable when he falsely thinks he is making someone suffer (unaware that his victim is in fact not suffering)? 3 The term ‘axiology’ is used here as a synonym of ‘value theory’; apparently it was given this sense by Paul Lapie in his doctoral thesis Logique de la Volonte´ (1902). Two other early references are W. Urban, Valuation: Its Nature and Laws (1909) and E. von Hartmann, Grundriss der Axiologie oder Wertwägungslehre (1908). 4 Moore does not actually express himself in this way. He rarely applies the phrase ‘for its own sake’ to goodness. There are many works in which ﬁnal and intrinsic value are treated as the same sort of value. See, for instance, Gregory Vlastos, ‘Human Worth, Merit and Equality’ (1969).

4

T Y P E S O F VA L U E

break it into simpler concepts, but, as he argued, not all concepts are capable of such a treatment. We cannot, he claimed, do it with a colour concept such as yellow, but we can do it with complex notions such as, say, a mule or a mare. Goodness is like yellow in the sense, he thought, that it cannot be broken into parts. What we can say, he thought, about what is intrinsically valuable is that it is a value which depends on the internal properties of its bearer, i.e. the thing which is valuable.5 In due course we shall return to the idea that value is a dependent, in fact, supervenient, notion. Meanwhile, it must be stressed that Moore did provide his readers with guidelines to improve their grasp of intrinsic value, and, in particular, of what things carry (or possibly would carry) this kind of value. One of Moore’s more discussed ideas is the method he proposed for ﬁnding out what things are intrinsically valuable, by which he wanted to ensure that we did not confuse what is valuable in itself with other kinds of value (in particular, instrumental values). Thus, Moore maintained that: it is necessary to consider what things are such that, if they existed by themselves, in absolute isolation, we should yet judge their existence to be good.6

The instruction is quite clear and simple: think about the alleged value bearer as if it existed all by itself. Doing this we should be able, he thought, to determine whether the existence of the thing is good or not. This so-called ‘isolation-test’ was supposed to enable us to sort what is valuable in itself from what is either valuable for something else’s sake or has no value at all.7 But Moore’s thought-experiment has not been generally accepted. In part this is not that difﬁcult to understand. As Noah Lemos points out: ontological isolationism is not very helpful since there are certain sorts of things that are intrinsically good but simply could not be the only things that exist. Consider the fact of Smith’s being happy and let’s suppose that it is intrinsically good. If there are certain abstract entities such as numbers or properties or states of affairs that necessarily exist, it would be impossible for Smith’s being happy to be the only thing that exists. More important, though, is the fact that Smith’s being happy could not exist without Smith’s existing, as well as, I suppose, Smith’s having certain pleasures and certain desires satisﬁed and his having certain beliefs to the effect that he had those pleasures and that his desires were satisﬁed. Since it is necessarily false that Smith’s being happy could be the only existing thing, this sort of ontological isolationism is not very clear or very helpful. (Lemos, 1994, p. 37; cf. Zimmerman, 2001a, p. 132)

5 Moore did not use the term ‘supervenience’ for the dependence relation between evaluative properties and non-evaluative ones. Johan Brännmark informs me (personal communication) that the term was used among British emergentists. See Samuel Alexander, Space, Time and Deity (1920), C. D. Broad, The Mind and Its Place in Nature (1925), and C. Lloyd Morgan, Emergent Evolution (1923). R. M Hare’s The Language of Morals (1952) introduced the notion to many moral philosophers. 6 Principia Ethica (1903, p. 187). 7 See Principia Ethica (1903, p. 187). As is pointed out by Lemos in Intrinsic Value (1994), Moore is not the only one to propose such an isolation test. David Ross did so in The Right and the Good (1930, pp. 68–9). According to Zimmerman in The Nature of Intrinsic Value (2001a), Richard Price made a similar claim. See the latter’s ‘A Review of the Principal Questions in Morals’ (1897, p. 148).

1.3

I S O L AT I N G E X T R I N S I C F I N A L VA L U E

5

The shortcomings of Moore’s isolation test led Lemos, and, later, Zimmerman to propose an amendment to it.8 The novel proposal is that we should reﬂect on the supposedly valuable object, setting aside the consequences of it without paying attention to the circumstances. The demand that we contemplate the object in this way is not incoherent. It might be hard to comply with sometimes, but it is not asking the impossible of us. This new proposed test tells us, then, to consider the thing without reference to context and what it brings in its train, and then to evaluate it. Lemos refers to it as the ‘intentional isolation test’ as opposed to Moore’s ontological one.9 If this kind of contemplation ‘leads’ you to believe that the thing is valuable, or good, and granted that you have in fact managed to keep circumstances and consequences out of your contemplation, you will have arrived at what is valuable, or good, in itself or for its own sake. Or so the argument goes.

1.3 Isolating extrinsic ﬁnal value Moore’s isolation test has presumably played some role in the story of why, for a long time, value theory treated the distinction between being valuable in itself and for its own sake as one of little importance. By focusing so ﬁrmly on things in isolation, not admitting external, relational properties to be considered as ‘valuable in itself ’-making properties, the perspective of value deﬁnitely came to be, for quite some time, too narrow. Not everything that we value for its own sake carries a value that supervenes only on the internal features of the value bearer. The intentional version is an improvement.10 However, once things are able to carry extrinsic ﬁnal value, i.e. be valuable for their own sake in virtue of at least some relational property that is not internal to the value bearer, matters become much more complicated. Obviously, the foolproof air of Moore’s original test has evaporated; what is left is a more modest mode of determining what has ﬁnal value. The test continues to be a catalyst that helps us trigger substantive evaluative reactions. However, we must now be on our guard; even the improved isolation test may be doing more harm than good. For instance, we cannot conclude, just because we are not ready to ascribe goodness to something when we contemplate it in isolation, that the object is not valuable for its own sake. We can only say that it is not a ﬁnal intrinsic value. It might be retorted that we can always extend the intentional isolation test to cases of extrinsic ﬁnal value; after all, there is nothing particularly difﬁcult about 8

Zimmerman shows that the ontological version is not only incoherent, but also misleading in certain cases. It can lead us to identify “certain states as parts of other states when they are in fact not parts of them at all” (2001a, p. 141). 9 Lemos introduces these expressions in (1994, pp. 10–11); cf. Roderick M. Chisholm, ‘Deﬁning Intrinsic Value’ (1981). 10 However, I think it is a mistake to suggest, as Chisholm appears to have done, that by saying that something is intrinsically good we mean by this that it is what we should evaluate as good if we applied the intentional isolation test. It is one thing to regard the test as a method for learning what things are intrinsically valuable; it is quite another to maintain that this test is part of the semantic analysis of value.

6

T Y P E S O F VA L U E

contemplating a thing being related to something else. True, I do think the improved isolation test can help us determine whether some object has or does not have extrinsic ﬁnal value. But it is important to steer clear of some complications here. For one thing, the kind of contemplation that we are now considering cannot merely be the contemplation of a state of affairs involving relata; focusing our attention on that would tell us at most whether this state of affairs has an intrinsic value. The relata would be internal to the state, which is quite consistent with the state’s being an intrinsically valuable object. Rather, the contemplation must concern an object that is related to some other object. If we manage such contemplation, we should at least in principle be able to arrive at more or less well-founded judgements about the extrinsic ﬁnal value of objects. Again, this does not amount to the 100 per cent test of Moore’s original method. Nor, however, does it leave us empty-handed. Naturally, we cannot be certain that taking more relational properties into the picture will not tilt our all-things-considered judgement of a particular object’s value. But this kind of tilting is just to be expected.11 Moore’s view that goodness is a simple, non-natural property has probably intrigued as many people as it has offended. The critics form a rather heterogeneous group. Without going into details that soon would take us far astray, we can at least identify the following camps. Some object to what they deem mystical, namely that goodness is non-natural. Others react to what they believe is only an expression of a false view on what we do when we evaluate. It is not difﬁcult to ﬁnd representatives of these views today. Whether goodness and value are natural properties or not, is something that I will leave open in this book. An answer to this question would, eventually, emerge if we knew the answer to various questions on the semantic side of formal value theory. For example, can evaluations be true or false? Is their main function to express beliefs or desires—to express attitudes that should ﬁt how the world in fact is, or attitudes that are such that the world should change to ﬁt the attitudes? However, today’s metaethical consensus leaves questions about the semantic nature of value judgements very much open, and so will I.12 (Much of interest can still be said about value and goodness that does not depend on taking a decisive stand on these semantic issues.) Some critics have gone a step further and argued that it does not make sense to speak of goodness in the way Moore apparently thought we are doing. This strain of thought goes back to Peter Geach’s inﬂuential paper ‘Good and Evil’.13 Geach argued there that every meaningful use (what he referred to as attributive use) of ‘good’ presupposes a certain domain of things to which the use applies. Things are never plain good. There Moore is not alone in coming up with a test for what has value. Another towering ﬁgure in value theory, Immanuel Kant, suggested a different method. Instead of contemplating things in isolation, Kant wanted us to place the presumed bearers of value in a different context (my own suggestion resembles this idea). For a discussion of Kant’s view, see Brännmark, Morality and the Pursuit of Happiness: A Study in Kantian Ethics (2002, sec. 2.2.). 12 In short, utterances of sentences containing value terms express the speaker’s value judgement. The nature of these judgements throughout the present work, then, is regarded as being consistent with both cognitivist and non-cognitivist metaethics. 13 Geach, ‘Good and Evil’ (1956). 11

1.4

C O M P E T I N G N O T I O N S O F I N T R I N S I C VA L U E

7

is no such thing as the predicative—as he called it—use of ‘good’. This discussion is important, and much can be learned from it. Despite this, I will sidestep further discussion of it. However, although I think this is an interesting approach, and clearly worth pursuing, perhaps I should add that I also think it faces a couple of problems. Let me brieﬂy state one. In maintaining that there are only attributive uses of ‘good’, we seem to assume the truth of a certain metaethical position—namely, one according to which we can understand the goodness of, say, a car purely in terms of its nonevaluative properties. But, of course, this is not an obviously correct position. We might argue, for instance, that whatever the constitutive properties of the car are, in the end it takes a genuine evaluation to say which car is a good car. Now, since I believe that ‘goodness’ does have a genuine evaluative meaning (although it is not something that I argue here), the attributive sense seems to me to be at most a non-interesting, but probably common, way of expressing what is not a genuine evaluative judgement. The ‘Geach camp’ could certainly deny this; they might argue that there are only attributive senses. However, so far no one has managed to come up with an altogether convincing metaethical view, and hence it is perhaps best to leave this issue open. (The analysis that I will be proposing does just this.) Moreover, although Geach and others may have a point as regards the use of “plain goodness/value”, it would be a mistake to confuse this ‘goodness’ with a different sort of goodness that is used interchangeably with ‘intrinsic (ﬁnal) value’. The phrase ‘object x is intrinsically (ﬁnally) valuable’ easily splits into ‘x is an object and x is intrinsically (ﬁnally) valuable’, and there seems to be nothing necessarily wrong or odd about this.14

1.4 Competing notions of intrinsic value So far I have suggested that objects may exist that are valuable for their own sake but whose value nonetheless does not depend exclusively on their internal features. It is, as mentioned, the task of substantive value theory to determine whether there is this kind of value. Still, we have to at least make the idea plausible. One reason the idea of an extrinsic ﬁnal value has met with suspicion has been an assumption among some inﬂuential value theorists concerning the sorts of thing that can bear ﬁnal value. The main reason for limiting oneself to just one kind of value bearer has been simplicity. This is a very understandable motive, but it has not convinced me to give up a pluralistic position when it comes to the nature of value bearers. I will return to this issue later on in Chapter 10, where I present an argument for value pluralism. (Since a pluralist position seems intuitively plausible, I prefer the more direct way to the analysis of personal value.)

14 This response to the Geach position comes originally from Zimmerman. In his (2001a) he rejects the main arguments put by Peter Geach, Bernard Williams, Philippa Foot, and Judith Jarvis Thomson. See Jonas Olson, Axiological Investigations (2005, pp. 32–44), who also defends the notion of ﬁnal (intrinsic) value. Recently Sven Danielsson has come to Geach’s defence in ‘On Geach on Good’ (2007). See also the recent article by Brännmark ‘Excellence and Means: On the Limits of Buck-Passing’ (2008).

8

T Y P E S O F VA L U E

It is not difﬁcult to ﬁnd examples that plainly appear to be inconsistent with the traditional Moorean view that all ﬁnal values, without exception, are intrinsic. I will present some examples in a moment. Meanwhile, another possibility that has so far not received any attention here must be addressed. The fact that there is seemingly no agreement on what sort of things have, say, intrinsic value, may not necessarily mark a difference in substantive views. It might just be the case that there is more than one concept of value circulating. In fact, it has recently been argued by Ben Bradley15 that there is not a single concept of intrinsic value but two: Mooreans accept the three tenets that intrinsic value: (a) supervenes only on the value bearer’s intrinsic properties; (b) accrues to the value bearer by necessity; and (c) is such that something that has it would continue to have this value “even if it were alone in the universe” (2006, p. 113). Kantians reject these principles. On their view, intrinsic value is “the sort of value that requires respect”. Bradley also argues, in many ways convincingly, that on certain points Mooreans and Kantians have been talking past each other. Whereas the Mooreans regard intrinsic value as something that ought to be promoted, or something that makes the world a better place, Kantians look upon it as something that should be respected or honoured. The position defended in this work is obviously not Moorean. It does recognize something like Moore’s intrinsic value, though. However, it also claims that since it can be subsumed under the notion of ﬁnal value, this is not the fundamental value notion. Nor is the position Kantian, however, although it bears some resemblance to what Bradley characterizes as such. Thus, on my view, certain valuable objects ought to be respected or honoured; but equally there are a great number of different attitudes that one ought to take with regard to objects that are valuable. Bradley’s idea that there are just two value notions is not that intuitive from this perspective. If there are indeed valuable things that we should cherish, protect, love, admire, and so on, it might well be suspected that there are not two value concepts but many—one each, say, for objects that ought to be cherished, promoted, loved, desired, and so on. This is, of course, a possibility. However, we might also try to look for a theory of value that could bring all of these values under one and the same roof. Later on, in Chapter 2, I will outline a pattern of value analysis which has received considerable attention lately and purports to do precisely this.16 Let us consider some examples that seem inconsistent with Moore’s views on intrinsic value. Shelly Kagan, for instance, discusses the value of a unique or rare stamp; John O’Neill refers to a wilderness that may be considered valuable precisely because it is untouched by humans,17 and Wlodek Rabinowicz and I have suggested the following:

15 16 17

Bradley, ‘Two Concepts of Intrinsic Value’ (2006). Bradley recognizes this. See his discussion of the ﬁtting attitude analysis (2006, pp. 127–8). O’Neill, ‘The Varieties of Intrinsic Value’ (1992).

1.5

T H E S O U R C E O F VA L U E

9

Another class of examples involves cases in which a thing is valued for its own sake in virtue of its special relationship to a particular object, event, or person. An original, say, an original work of art, may be valued for its own sake precisely because it has the relational property of being an original rather than a copy. Its ﬁnal value supervenes, in part, on its special causal relation to the artist. Princess Diana’s dress may be another case in point. The dress is valuable just because it has belonged to Diana. This is what we value it for. (Rabinowicz and Rønnow-Rasmussen, 2000, p. 41)

Being unique or rare, untouched by humans, original, or related to someone of importance are all external relational properties, so they would not be admitted in the supervenience base of an intrinsic value. Examples of this kind can easily be multiplied, which makes one wonder why the Moorean dogma has not been very widely abandoned once and for all.

1.5 The source of value Recently a number of philosophers have argued that there can be ﬁnal values that are not intrinsic. One argument in particular, by Christine Korsgaard, has been widely discussed. Korsgaard calls an object intrinsically valuable if the “source” of its goodness lies in the object itself. But, in her view, not all things valuable as ends are intrinsically valuable. Separating the two distinctions in goodness [intrinsic/extrinsic and ﬁnal/instrumental], however, opens up another possibility: that of something which is extrinsically good yet valued as an end. An example of this would be something that was good as an end because of the interest that someone took in it, or the desire that someone had for it, for its own sake. (1983, p. 252)18

The source of some end values is located outside the value bearer. Thus, instead of being internal the source may well lie, say, in our interests, or in desires directed onto the objects themselves, in which case the source will be external (Korsgaard, 1996 [1983], p. 252). Now, the notion of a value source is not straightforward. As Rabinowicz and I have argued elsewhere, it is important to separate two possible interpretations of this notion; but when this is done it becomes clear that Korsgaard’s argument is misdirected. To identify the source of value may be either to detect the supervenience base of the value or to discover the constitutive grounds of an object’s ﬁnal value. The distinction between supervenience and constitution is important if you want to understand why value theorists understand value in so many different ways. But both notions are far from transparent, and a more detailed discussion of the distinction is required. Therefore, in the remainder of this chapter, I will outline how I think these notions should be understood. Let us begin with supervenience. 18 See her ‘Two Distinctions in Goodness’ (1983), which is also published in her Creating the Kingdom of Ends (1996), from which the quotation comes.

10

T Y P E S O F VA L U E

1.6 Supervenience Supervenient properties are ‘dependent properties’.19 To insist that value always accrues to an object because it has certain other, non-evaluative properties is to express a belief, by implication, that there is a relation of dependence between value and those other, subjacent (i.e. underlying) properties. This does not seem obviously true in the case of mental predicates. At least, it is not obviously implausible to suggest that we, as language users, were mental ‘predicaters’ long before we became aware of the body/ mind problem. The second job the supervenience claim is customarily taken to do is that of capturing what we might call the consistency intuition. This is the conviction that if you assert that x is valuable, and if you agree that y is relevantly similar to x in its natural properties, you must be prepared to assert that y, too, is valuable. To deny this would be to leave the listener in a kind of perplexity similar to that which follows the detection of a contradiction. I take it that if we were to believe that values are, so to speak, free-ﬂoating ontological entities that just land on objects (either haphazardly or because, say, they were bestowed in accordance with God’s will), we would not be logically committed to ascribing value to y here. It is at least not a requirement of consistency to draw the conclusion that y is valuable, for x and y may differ precisely in that value has ‘landed’ on one but not the other. It might be suggested that the fact that we can imagine such ‘value butterﬂies’ shows the supervenience thesis to be, not a conceptual thesis, but an expression of our deeply rooted beliefs about what and how things are. However, this reasoning is ﬂawed. From the fact that we seem to be able to imagine, without contradiction, ‘free-ﬂoating entities’, it does not follow that we are conceptually able to imagine that values are such entities. The discovery that this ‘butterﬂy’ reasoning seems to be fundamentally confused should not obscure the fact that it is hard to come up with a knock-down argument showing that the supervenience thesis is not, at bottom, based on ontological convictions. I am not sure what such an argument would look like. Whether this mistrust warrants serious disbelief in supervenience is another matter. For my own part, I do not think so. At any rate, I am certainly inclined to say that value terms are, for purely conceptual reasons, supervenient terms. Hence, in what follows I will treat the supervenience thesis as a conceptual thesis. Here I merely wish to make it clear that I am unable to see how to settle beyond question whether or not my intuitions are purely linguistic.

19 In other words, when, henceforth, I talk about value-makers, I have in mind those properties on which value depends, leaving it for discussion how ‘depends’ should be read. Perhaps dependence here is one involving “makers”, or “enablers”, or both. See Jonathan Dancy, Ethics Without Principles (2004). Cf. Caj Strandberg, Moral Reality: A Defense of Moral Realism (2004, pp. 229–39).

1.6

S U P E RV E N I E N C E

11

As mentioned earlier, there is, or so it has been argued, more to supervenience than the claim that if two things, x and y, are similar in their natural properties, they must— and ‘must’ here is conceptual—have similar value properties. In recognizing that the supervenience thesis expresses a consistency intuition, we must not lose sight of a further important relation that is said to hold between natural properties and value properties—namely, that the former necessitate the latter. This necessity relation might be moral or metaphysical, or it might be some other kind of necessity. It might mean, as Sven Danielsson has expressed it, “something like ‘fact-independently’. If Socrates is good because he is brave and wise, then, whatever the facts that may obtain, anyone brave and wise would be good” (2001, p. 95).20 The remedy for butterﬂy properties seems rather obvious. If our aim is to dislodge butterﬂies from value supervenience we need to formulate a relation between value properties and natural ones such that if the latter are present, then, in every possible world, the former are too. We need, in other words, to go from intra-world to interworld supervenience. Taking ‘V’ to refer to value, ‘N’ to refer to natural properties, and ‘Nec’ as a necessity operator, we need the relation captured by the following formula, due to Jaegwon Kim:21 Strong Supervenience: Nec 8x (Vx ! ∃N (Nx and Nec 8y (Ny ! Vy))) If this strong relation is what the supervenience relation is all about, properties that behave on the analogy with the butterﬂy are blocked; such properties cannot be supervenient. There is a necessity relation (the second one above) of some sort (moral necessity, metaphysical necessity?) between N-properties and V-properties that excludes this possibility. Strong Supervenience captures something essential about value supervenience. However, the question is whether it catches all there is to this dependence relation. One problem with it needs to be resolved. As far as I can see, it leaves out an important aspect, namely what we might call the ‘direction-feature’ of the ‘because of ’ or ‘in virtue of ’ relation. These relations suggest that moral and natural properties are connected in a special sense—the latter give rise to (or support) the former, and not vice versa. Consider the relation between ‘x is coloured’ and ‘x is extended’. This relation is not, in my view, a ‘because of’ or ‘in virtue of’ relation. But suppose we reformulate Strong Supervenience in terms of properties of colour and extension. Strong Supervenience would then seem to ﬁt the relation that exists between colour and extension. In other words, if we do not want to say that colour and extension exemplify the kind of relation found between moral properties and natural ones, we need to qualify Strong Supervenience in some way. At present I do not know how to do this.

20 21

Danielsson, ‘The Supervenience of Intrinsic Value’ (2001). See Kim, Supervenience and Mind (1993).

12

T Y P E S O F VA L U E

1.7 The Invariance Thesis The idea that values are constituted by subjects is at heart of subjectivism. Now, subjectivists are vulnerable to an objection that we need to examine. The attitudes we have appear to depend heavily on arbitrary factors. Hence, if we tie values to such capricious entities as attitudes, we make the question of what values there are highly contingent; and this is not how we normally conceive of values. The suspicion, in other words, is that if we relate value to attitudes, value becomes a relative notion, and value relativism is hard to combine with the following: The Invariance Thesis: The ﬁnal value of an object is invariant over possible worlds. It follows from this thesis that if an object is valuable, the object (and any object exactly similar to it) will carry precisely this value in every possible world in which it is present. Or, more accurately: if something is a supervenience base in the actual world, it will be such a base in every possible world in which it is present. It is a standard objection to subjectivist accounts that they have to reject the Invariance Thesis. If we ﬁx value to desires and other kinds of attitude, the complaint runs, a valuable object x will not be valuable (i.e. its value will not have been constituted) in possible worlds in which x is not the object of a desire. If the world were different, with an alternative geography of desires, the set of valuable objects would be quite other than the actual one. On this matter we may expect various subjectivist responses. I have discussed which ones elsewhere (Rønnow-Rasmussen, 2003), and I will therefore not repeat them here. In this context it is paramount to appreciate that the idea that value is invariant is contentious. Although a number of inﬂuential axiologists, including G. E. Moore, have defended the Invariance Thesis over the years, it has come under considerable ﬁre more recently.22 Its plausibility depends on the nature of the value bearer x or, to be more precise, on what kinds of property are included in the supervenience base. If x is valuable in virtue of its non-essential properties, the Invariance Thesis looks too encompassing. If La Gioconda had been painted with, say, different colour properties, it would not have been valuable (that is, assuming that the colours are not essential properties of La Gioconda). However, things are different if x is a value-bearing object like an abstract entity which is valuable in virtue of its essential properties. Such an entity will be the same in all possible worlds, and hence its value will remain the same in all worlds. If x is not the concrete La Gioconda itself but rather some abstract entity such as the fact (the obtaining of the state of affairs)23 that La Gioconda presents such and such features, then we cannot ‘tamper’ with the features making this very fact what it is. To do so would be to create a distinct fact whose possible value or disvalue would have no bearing on the issue under discussion.

22 23

For a recent defence of this thesis, see Zimmerman (2001a). See, here, Lemos, Intrinsic Value: Concept and Warrant (1994, pp. 25–6) and Zimmerman (2001a, p. 50).

1.8

CONSTITUTION

13

In an unqualiﬁed form, then, the Invariance Thesis is ruled out as a test of demarcation between objectivists and subjectivists. Advocates of either position can agree that if an object supervenes on non-essential properties, then, in a world in which the object lacks some of these properties, it will no longer be valuable. It should be noted, however, that the objectivist’s view may be quite a different one, for they (like subjectivists) can adopt a certain kind of value-bearer monism. On this view it is only entities that have whatever features they have essentially (e.g. states of affairs) that carry (ﬁnal) value. If objectivists and subjectivists share this view, their responses to the Invariance Thesis will inevitably differ. Subjectivists will naturally reject the Invariance Thesis, but they have quite a lot of leeway as to how they go about discarding it. Some ways, for instance, will be more ‘objectivistic’, or attitude-independent, than others. Suppose there is a state of affairs that I want to see realized for its own sake—say, that people look at, and appreciate, La Gioconda. Putting aside the substantial question whether this is a good example of a value bearer, the nature of my attitude here is taken to be universal. In other words, I have this attitude not only for the actual work of art, people, and world. Rather, the scope of my attitude ranges over all the possible worlds in which people appreciate a painting that is exactly similar to the actual La Giaconda in universal features; in any such world I prefer this sort of appreciation to be a fact.24 Given my actual preference, the value of the object x will, in other words, be invariant. One salient aspect of this example is that, in it, the constitutive ground of value does not coincide with the supervenience base. Actually, there is nothing particularly odd about this; we need just remind ourselves that some objects of our attitudes (say, objects of preferences) do not coincide in time with our preference. I may, for instance, now prefer that, when I am old, no one will seek to ensure that I go on living at any price. I may prefer this although I have good evidence that most people do in fact want to go on living, even if the cost is very high, when they grow old. Again, some of my preferences are conditional in form. For instance, in the bath I can prefer that, were I on the train, I would have clothes on.

1.8 Constitution The idea that we should differentiate between the constitutive grounds and supervenience base of value has appeared in several places in recent years.25 I ﬁrst became 24

Here I sidestep a complication, namely whether value can supervene only on universal features. In Chapter 6 I shall return to this matter. 25 However, it is also a distinction which, over the years, has been overlooked. Thus Henry D. Aiken (1950) objects to views that have deﬁned intrinsic value in terms of “the object in which interest, satisfaction or pleasure is taken” (p. 11). On his view, it is rather satisfaction in itself that is valuable (a position which he refers to as subjectivism). In comparison with the former views, his own avoids the paradox of ascribing intrinsic value to something on the basis of a relational feature (e.g. that it is an object of someone’s desires). The separation of the supervenience base of the value and its constitutive grounds would clear away this ‘paradox’, though.

14

T Y P E S O F VA L U E

¨ sterberg.26 Their familiar with it through a work by Wlodek Rabinowicz and Jan O distinction illuminates the division between subjectivists and objectivists, i.e. between those who believe value is necessarily related to subjects and those who deny this.27 Rabinowicz further clariﬁes the distinction between supervenience and constitution ¨ sterberg, 1996, by looking at the role conventions play in chess (Rabinowicz and O p. 21). Thus, in a chess game certain moves will be permissible and others will not. The permissibility of a particular move depends on how the board is set at the time of the move. For instance, a white rook at square a1 can take the black queen on a7 because the features of the game now make it possible to move the rook all the way up to a7 (while before there was, say, another piece blocking that row of squares). Now, the permissibility of the move a1xa7 does depend (supervene, if you like) on the (internal) features of that particular game, at that particular time. However, the constitutive ground for this move is to be found in the social convention that maintains the rules of chess, and which, among other things, specify how the rook and other pieces move. Rabinowicz’s analogy is illuminating. Of course it does raise important issues that need to be commented on. However, since I have done so in Rønnow-Rasmussen (2003), and since it would lead us too far astray from our main concern, I prefer to consider a problem for this subjectivist approach.

1.9 Subjectivism and objectivism There is, then, a possible objection to my approach here that I need to address. This will lead to yet another way of illuminating the difference between supervenience and constitution. The objection, which would most likely ﬁgure in an objectivist response to the subjectivist approach, is this. What the subjectivist refers to as the constitution of value is nothing but itself a supervenient feature. When the subjectivist claims that an object x is valuable, and that x’s value is constituted by the subject’s attitude, the constitutive attitude in question is really part of something else that is valuable— namely, the subject’s direction of a certain attitude on x. It is this obtaining of a state of affairs rather than x that is valuable. Since the attitude is in a sense part of the value bearer, it is more appropriate to regard it as a value-making, subjacent feature. Hence there is no need to distinguish in the original way between supervenience and constitution. Accordingly, subjectivism should be regarded as a position that ascribes value only to certain kinds of object: it ascribes value only to those facts that involve the subject directing an attitude on some object. 26 ¨ sterberg, ‘Value Based on Preferences’ A detailed elucidation of it can be found in Rabinowicz and O (1996). See also Rabinowicz and Rønnow-Rasmussen, ‘A Distinction in Value: Intrinsic and For Its Own Sake’ (2000) and Rønnow-Rasmussen, ‘Subjectivism and Objectivism: An Outline’ (2003b). I owe Wlodek Rabinowicz a great deal for having so generously discussed the distinction with me over the years. However, he is not to be blamed for anything that I say here. 27 David Gauthier’s Morals by Agreement (1986) is a good example of a subjectivist position. I owe FritzAnton Fritzson for pointing this out to me.

1.9

SUBJECTIVISM AND OBJECTIVISM

15

There is some truth in this reply. Thus, it seems correct to say that if the obtaining of the state of affairs of a favouring x is valuable, then it is reasonable to regard the attitude as an element of the supervenience base. The problem with the reply lies elsewhere. Why should a subjectivist conﬁne herself to the analysis of value judgements concerning facts about the subject directing some attitude on something? Of course, we might have taken a stand on what things are, and what things are not, in fact valuable; and we might have made the restriction in accordance with our evaluative views. But if subjectivism and objectivism are positions designed to answer questions about value and valuing, there is something fundamentally wrong when either subjectivism or objectivism is depicted as a position that says, necessarily, that certain value judgements cannot be analysed according to their own pattern of analysis (i.e. in the case of subjectivism, patterns of the kind ‘Object x is valuable’). This would be to describe the two positions from a standpoint that is both evaluative and, in this case, biased in favour of objectivism. Of course, objectivism might be true. If this is shown, subjectivists will have to give up their idea that attitudes have constitutive powers. But subjectivism might also be true, and if it is objectivists will have to face up to the fact that there is more to value than supervenient features. The above brings me to another way of elucidating the difference between supervenience and constitution. This will further support the idea that we should not include a subject’s attitude too readily in the supervenience base. Properties (or facts about these) belonging to the supervenience base are customarily invoked as reasons for or against various actions taken for the sake of the object. To appeal to the subject’s attitudes—say, those of his preferences or desires directed on the object—as a reason for acting is not always plausible. At least, it is not if what you have in mind here is not what caused you to act, but rather what you take to be your normative reasons for acting.28 In case we have in mind actions performed for the object’s own sake (rather than for the sake of being desired by us) this possibility is ruled out. The attitudes should therefore remain in the background, not the foreground, of our motivational set. They can be the constitutive grounds of ﬁnal value, but this does not make them a part of its supervenience base. Let us pick up the threads and return to Korsgaard’s suggestion that we should understand the difference between intrinsic and extrinsic ﬁnal values in terms of the location of the source of value. Recall her point: if the source is found in the object itself, we are dealing with intrinsic value; if it lies in the interests or desires that are directed on the object, the value is extrinsic (1996 [1983], p. 252).29 28 We shall return in Chapter 5 to the difference between what is part of the intentional content of an attitude (say, some features that call for you to act) and what caused the subject to have the attitude. 29 Thus Korsgaard writes: “Separating the two distinctions in goodness [intrinsic/extrinsic and ﬁnal/ instrumental], however, opens up another possibility: that of something which is extrinsically good yet valued as an end. An example of this would be something that was good as an end because of the interest that someone took in it, or the desire that someone had for it, for its own sake” (1996, p. 252). As noted by Bengt Brülde in The Human Good (1998, p. 390), in her discussion of ﬁnal value Korsgaard does not clearly

16

T Y P E S O F VA L U E

That this is not a compelling way of understanding the distinction between intrinsic and extrinsic values should be clear. ‘Source’ may refer either to the properties of an object on which its ﬁnal value supervenes or to the constitutive grounds of the object’s ﬁnal value. The latter may well lie outside the object itself—that is at least what subjectivists want us to believe—even though the former are internal to the object. For instance, the value of an object is, on a preferentialist conception, constituted by the preferences or desires directed on the object. But notice, where the object is desired only for features that are internal to it, then, on this subjectivist position, this externally constituted value is nonetheless intrinsic. The object’s good-making properties are all internal properties of the object: they are those for whose sake the object is preferred.30 To relate value to the desires of a subject under the heading ‘value source’ may be to conﬂate the supervenience base with the constitutive grounds of value. To establish that values can be externally constituted is in itself not an argument for extrinsic ﬁnal values. Final values may all still be intrinsic values if what makes them ﬁnal is their supervening only on features internal to the value bearer.31 Before we proceed, there is something I would like to emphasize. An anonymous referee for Oxford University Press pointed out that even if we grant that subjectivism, as I have presented it, is compatible with supervenience, it will remain a contingent fact that the values constituted by attitudes actually deliver that supervenience. This should certainly be underlined. That is, subjectivists cannot guarantee supervenience whenever an objectivist claims there is value supervenience. As I pointed out in Section 1.7, subjectivists cannot agree with all that objectivists want to say about value supervenience. However, they need not agree with everything objectivists say about value. Supervenience, according to the subjectivist, is a relation that holds between value and an object in virtue of the value bearer’s non-essential properties. Subjectivists recognize the logical requirement that if you say that x is valuable, you must agree that something about x makes it valuable. Since they do so, unless they change their minds about what is valuable, they are therefore forced to apply their value predicates in a consistent way. Of course, objectivists will claim there is more to “force” than “consistency”.

distinguish between being valued and being valuable (as an end, or for its own sake). In what follows, however, I focus on the latter. 30 When, in ‘The Limits of Well-being”, Kagan maintains that on “a radically subjectivist conception of value, [ . . . ] nothing would be valuable as an end in the absence of there being some creature who values it” (1992, p. 184), he (too) conﬂates the constitutive grounds of value with its supervenience base. Subjectivists will locate the constitutive ground in the subject, and not the supervenience base; and it is the nature of the latter which determines whether or not the value is intrinsic. See Rabinowicz and Rønnow-Rasmussen (2000). 31 See also Korsgaard’s later work, ‘Motivation, Metaphysics, and the Value of the Self: A Reply to Ginsborg, Guyer and Schneewind’ (1998), in which she claims that our particular ends “have only extrinsic value, since their value depends on our own desires and interests in them and is conferred on them by our own rational choices” (p. 63). Here, Korsgaard conﬂates supervenience with the constitutive grounds of value.

1.9

SUBJECTIVISM AND OBJECTIVISM

17

Obviously, subjectivists cannot constitute “objectivist” value—at least, not the sort of value that supervenes on the essential properties of the value bearer. But this is quite consistent with the idea that (subjectivist) value is constituted by a subject’s preferences and other attitudes. Supervenience tells us something about “where” a value is located. The difference between objectivists and subjectivists is that whereas the former think that if we know where a value is located we also know (or, chances are, we also know) what grounded (constituted) the value, the latter deny this, believing that it is one thing to identify a value location, but quite another to settle what constitutes the value. There are several serious problems with subjectivism, and these incline me to reject it. The most important is the fact that the very notion of value constitution remains somewhat obscure. However, objectivism is also problematic. But, my goal here has not been to settle the subjectivism/objectivism issue, but rather to outline how the debate should be understood. My contention is that the disagreement does not, in the ﬁrst place, concern the location of value—both parties can agree on what is valuable; it concerns the basis (constitution) on which the value is located where it is located. G. E. Moore was a stout defender of the idea that intrinsic value is in a sense the property of objectivists; only objectivists can speak of it coherently. I think Moore was wrong about this—just as I think those (mostly subjectivists) who argue that the notion of intrinsic or ﬁnal value is incoherent are wrong.32 There is nothing to prevent subjectivists from recognizing the ﬁnal or intrinsic value of something. There is an important difference, of course, between subjectivism and objectivism, but it is not one that prevents subjectivists from ascribing intrinsic or ﬁnal value to things. The distinction between supervenience and constitution is not the only tool needed to separate subjectivists from objectivists. Nonetheless, its being overlooked continues to give rise to unnecessary confusion. It is also true that proper appreciation of its signiﬁcance allows one to move beyond certain casual, one-dimensional assumptions—for instance, the idea that the subjectivist only locates value in subjective states, or the idea that objectivists cannot share the evaluations of a subjectivist. On the approach defended here, subjectivists and objectivists may well share each other’s evaluations, and they need not disagree about what objects are the appropriate bearers of value.33

32 Moore found at least two faults with subjectivism—it makes the objectivity of good disappear, and it makes it impossible for things to be intrinsically valuable. The following passage is often taken to express, by implication, the latter point: “To say that a kind of value is ‘intrinsic’ means merely that the question whether a thing possesses it, and in what degree it possesses it, depends solely on the intrinsic nature of the thing in question” (‘The Conception of Intrinsic Value’ (1922)). The subjectivist, according to Moore, makes value dependent on something external, the subject, and the subjectivist’s alleged value is therefore not an intrinsic value. For another early attempt to argue that subjectivists cannot coherently speak of intrinsic value, see the German value theorist Erich Heyde’s Wert: Eine philosophische Grundlegung (1926). 33 The distinction has a clear bearing, for instance, on the way we should understand a central claim in L. W. Sumner’s Welfare, Happiness & Ethics (1996) that only subjectivists make welfare logically dependent on the attitudes of the welfare subject.

18

T Y P E S O F VA L U E

Sometimes positions in this area (in particular subjectivism) are depicted as ﬁrst-order normative claims. For example, it has been said that you are a subjectivist if you think that each individual ought to do whatever she happens to think she should.34 Of course, there is no real point in quibbling about names. But in my view, since subjectivism and objectivism are naturally seen as arising from questions about how to understand the nature of value (and ultimately evaluative and normative-judgements like the one mentioned above), a great deal would be gained if we could retain ‘subjectivism’ and ‘objectivism’ for second-order views on values.

34 See, for example, J. L. Mackie, in Ethics: Inventing Right and Wrong (1977), who refers to this as a kind of ﬁrst-order subjectivism. Moreover, when Josiah Royce refers to the enquiry ‘Why should I obey the moral law?’ as “the old question of subjectivism” he is obviously not thinking of subjectivism in the same way (‘Discussions: The Outlook of Ethics’ (1891, p.109).

2 Fitting-attitude Analysis

To think that something possesses personal value is to evaluate it as having value for someone; in cases of positive or negative personal value we think the thing is good or bad for this someone. In my view, evaluations, whether they involve personal or impersonal values, may be more or less well grounded. Perhaps they may even be true or false. But whether or not they carry truth-value, the fact that it is my evaluation does not in itself make these evaluations true or invulnerable to criticism. And this, I take it, holds for impersonal as well as personal evaluations. I am not going to argue for this here. Though it is tempting to do so, I will spare the reader such a digression from my main task. The literature on well-being (philosophical, as well as non-philosophical) is inundated with suggestions as to what is good, and what is bad, for you. But these more or less wellgrounded pieces of advice and suggestions are seldom accompanied by analysis of the very notion they employ. Of course, there are some important exceptions to this, and we shall consider these in due course. Naturally, the literature is rich in suggestions about the kinds of thing that contribute to human welfare or well-being; and in all likelihood welfare or well-being are bona ﬁde examples of things that have personal value to us. Philosophical works on what makes us feel well, or what makes our lives go well, devote much space to the discovery of the constitutive parts—or the necessary and sufﬁcient conditions—of a good life or good mental state. For many years this has been, and still is, a stimulating and vivid area of philosophical analysis. However, it is also a ﬁeld which, perhaps too readily, has tended to monopolize the notions good-for and value-for. The present monograph tries to put up a bit of a ﬁght and resist this trend. That is, if am right there is more to value-for than well-being or welfare. The latter may well be typical cases of what is good for a person, but they do not exhaust the category of things that are good for us. A word on the very nature of this resistance: philosophical analysis is mainly reductive—it typically splits concepts into smaller conceptual elements wherever it makes sense to do so. This has proven to be a most fruitful intellectual enterprise, gradually affording us more and more ﬁne-grained conceptual tools. I will not depart from this approach. However, what I will initially be proposing is that the reader should take a step back and be open-minded enough to entertain, as at least a

20

F I T T I N G - AT T I T U D E A N A LY S I S

possibility, the idea that value-for is a more complex notion than ‘welfarists’ have assumed. Doing so, we shall eventually be led down another reductive avenue. The kind of analysandum that I have in mind is one typically accruing to certain mental states, but also to things like poems, tombstones, and the like, and to the obtainings of (abstract) states of affairs such as, say, the fact that my favourite football team wins. My way of arguing for this more complex notion of value-for is, then, this: I shall simply try to outline an analysis that is sufﬁciently cogent and attractive, and that is such that it becomes evident that the list of things just mentioned are all, at least potentially, more or less plausible carriers of value-for, as my analysis understands this notion. Now in recent years a certain pattern of value analysis—what I will refer to as the ﬁtting attitude (FA) analysis—has had something of a renaissance. This pattern leads, not necessarily, but very naturally, to a position that admits many different kinds of value bearer; and it is a promising pattern, since it shows what all values—thin ones (like goodness) as well as thick ones (like admirability)—have in common. Later on, in Chapter 10, I will present an argument for why I think we should recognize different kinds of value bearer. I shall argue that a well-known substantive view of value, hedonism, is best understood as ascribing value to sensations with a particular quality—a hedonic property. Hedonism alone is a rather imperfect value theory; things other than mental states carry value—things indeed that do not belong to the same metaphysical category as mental states. This initial contention leads quite naturally to a search for a pluralist value theory. The argument requires some unravelling, though. It seems therefore wise not to let it get in the way of the analysis’s main steps. In this and the next chapter I will outline some of the more important features of the FA analysis. I shall also highlight some of its more problematic elements. Unfortunately, the FA analysis does not only have much that speaks in favour of it; it also introduces a considerable number of more or less serious problems, some of which presently appear to lack any obvious solution. My aim is not to solve, or even to discuss, all of these problems, but to argue that if we ﬁne-tune this kind of analysis, we may add yet another positive feature to it—that is to say, it will become apparent that it is capable of giving us an account of personal value. Of course, the possibility that values are more heterogeneous than traditional FA advocates have so far realized cannot be regarded as a positive feature of the FA approach if, eventually, it is shown that we must give up on the pattern anyway (because, say, we cannot solve the problems mentioned above). However, my hope is that, by pointing at a fairly easy way of extending the analysis, we shall keep the focus on what is a most intriguing, and for years too-much neglected, view of value.

2.1 The origins A detailed history of the FA analysis has yet to be written. Here I shall conﬁne myself to saying something about two major exponents of it,1 although in fact the list of 1 For discussion of some of the leading proponents of FA analysis, I refer the reader to Jonathan Dancy’s excellent ‘Should we pass the buck?’ (2000a). See also Wlodek Rabinowicz and Toni Rønnow-Rasmussen (2004a). This section of the present text rests heavily on this work.

2.1

THE ORIGINS

21

prominent ﬁgures who, at one point or other, have advocated it is arguably quite long: it includes C. D. Broad, A. C. Ewing, and more recently Richard Brandt, Sven Danielsson, Alan Gibbard, Jonas Olson, Thomas Scanlon, John Skorupski, and Michael J. Zimmerman.2 The Austrian philosopher Franz Brentano (1838–1917) appears to have been the founding father of the analysis.3 However, his insights left no lasting, salient impact on his contemporary value theorists in Austria and Germany. There is at least no clear trace of the FA approach in the mature work of prominent Brentanists like Alexius Meinong and Christian Von Ehrenfels; and again, if we consider works on the history of ethics produced during, and for many years after, Brentano’s life, we ﬁnd that, strangely, the FA view is often neglected or misunderstood. The core idea of Brentano’s view can be expressed in a few words. Here is an example: . . . the good is that which is worthy of love, that which can be loved with a love that is correct.4

Notice that he is saying not that the good is worthy of love, but rather that what is worthy of love is good. The attitude is, in other words, directed not on the goodness, but on something else. Brentano had a quite speciﬁc idea in mind when he talked about love. ‘Love’ refers to what he regards as a “higher mode” of taking pleasure in something.5 Moreover, this feeling of pleasure has a further feature that distinguishes it from what he thought of as mere compulsive and instinctive feeling, namely that it is “experienced as correct [als richtig charakterisiert]”;6 and because the experience of correctness is conceived of as a part of the very attitude itself, Brentano is able to argue that such attitudes function as the basis of our knowledge of the good. Some further points about Brentano’s views ought to be made. He distinguished between what was good in itself (that which is “pleasing in itself ”) and what might be referred to as good as a means (that which is “pleasing in virtue of what it brings about or preserves or makes probable”).7 But what is more interesting is that he also applied the analysis to value comparisons. As a consequence, his analysis accommodated the idea that 2

C. D. Broad, Five Types of Ethical Theory (1930); and A. C. Ewing, The Deﬁnition of Good (1947). See also Alan Gibbard, Wise Choices, Apt Feelings (1990); Richard Brandt, ‘Moral Valuation’ (1946); Thomas Scanlon, What We Owe to Each Other (1998); Michael J. Zimmerman, The Nature of Intrinsic Value (2001a); John Skorupski, ‘Buck-Passing about Goodness’ (2007); and Sven Danielsson and Jonas Olson, ‘Brentano and the Buck-passers’ (2007). 3 Franz Brentano, The Origin of Our Knowledge of Right and Wrong (1969 [1889]). Other names sometimes mentioned in this context include Adam Smith (1759/1790) and Henry Sidgwick (1907 [1874]). 4 Brentano (1969 [1889], p. 18). 5 Brentano conceived of this higher mode as an emotive state. But taking pleasure in something is not necessarily being in an emotive state. In fact in recent years Fred Feldman has objected to what he refers to as the “distinctive feeling view”: the idea that pleasures are phenomenologically uniform; see his Pleasure and the Good Life: Concerning the Nature, Varieties, and Plausibility of Hedonism (2004). 6 Brentano, (1969 [1889], pp. 21, 22). It should be mentioned that Roderick M. Chisholm’s translation deviates from the German original but may still be faithful to it in spirit. On this point, see J. N. Findlay, Axiological Ethics (1970, pp. 21ff.). Cf. Rabinowicz and Rønnow-Rasmussen, ‘The Strike of the Demon: On Fitting Pro-Attitudes and Value’ (2004a). 7 Brentano (1969 [1889], p. 18).

22

F I T T I N G - AT T I T U D E A N A LY S I S

the sphere of value should be understood in terms of more than one kind of attitude or emotion. This is an important move. It suggests that value might need even more attitudes to be fully understood. Moreover, if these attitudes take different sorts of object, then this manoeuvre can be seen to carry the seeds of a pluralistic position on value. Brentano understood value comparisons of the kind reported when we say that something is better than something else in terms of the former being “worthy of greater love”.8 He carefully stressed how this greater love should be understood. It was not a question of more intense love, or of a more strongly felt pleasure. Instead ‘greater’ refers to: . . . the peculiar phenomenon to be found within the sphere of emotions—namely, to the phenomenon of preferring . . . When we call one good ‘better’ than another, we mean that the one good is preferable to the other. In other words, it is correct to prefer the one good, for its own sake, to the other.9

Evidently Brentano regarded preferences as emotive, rather than purely connative, states (the latter being a common view today).10 That apart, the twist, or turn, which value analysis takes via this expansion is of great importance. “Worthy of greater love” can apparently be read as meaning ﬁtting to be preferred. The value analyst’s toolbox is enriched by one more item besides love.

2.2 A. C. Ewing If Brentano founded the FA analysis, Ewing is its foremost cultivator. During the period in which Ewing presented his views on value in a series of works, formal as well as substantive value theory was rapidly becoming marginalized in certain inﬂuential English-speaking philosophy circles. This explains in part why his views were somewhat neglected for several decades. Ewing objected to naturalism as well as to the idea, widespread at the time, that goodness was some indeﬁnable quality. Actually, his dismissal of this Moorean idea (and the corollary that we somehow discern such a quality) becomes the springboard from which he launches his own deﬁnition (Ewing, 1959, p. 81). While he thought that we are unable to obtain a clear idea of goodness, he maintained that this is not the case with obligations. In his view it was: . . . quite clear in fact, that we are aware of obligations and that the concept of obligation is quite distinct from an empirical concept. (p. 82)11 8

Ibid., p. 25. Ibid., p. 26. 10 Brentano would not, in other words, regard preferences as attitudes that aim (at least, directly) at the realization of some state of affairs. 11 Just how much trust Ewing had in introspective knowledge is not clear to me. There are passages in Second Thoughts in Moral Philosophy suggesting that he regarded introspective awareness (e.g. of obligations) with some suspicion, as something that does not necessarily disclose something in the real world (see 1959, 9

2.2 A . C .

EWING

23

Originally, in the early work The Deﬁnition of Good (1947), Ewing distinguished between two sorts of obligation and two corresponding oughts: he understood the ﬁrst, and primary, notion as “signifying ‘ﬁttingness’” (1947, p. 151), which he took to be an “irreducible ethical concept”; and the second as the “‘ought’ of moral obligation” (see 1959, p. 90). It is not obvious why he refers to ‘ﬁttingness’ as an ethical concept. The passage would make more sense had he characterized it as a normative concept. He then deﬁned goodness, in terms of the ﬁrst kind of ought, as a “ﬁtting object of a pro attitude”.12 However, later, in Second Thoughts in Moral Philosophy, he modiﬁed this view: I still recognize these two meanings of ‘ought’, but now I wish to lay more stress on the moral ‘ought’ and to reduce the ‘ought’ of ﬁttingness to reasonableness. (1959, p. 90)

He still deﬁnes ‘good’ as “what ought to be the object of a pro-attitude”, but his new position stresses that both the moral ‘ought’ and the ‘ought of reasonableness’ are unique in the sense that they cannot be reduced to other concepts. Certain things may be morally good, and we morally ought to adopt a favourable attitude to these objects. Other things are good in a non-moral sense: they justify, or provide reasons for, the attitude. For example, my own pleasure is such that I ought to adopt a positive attitude to it, but the ‘ought’ here is hardly the moral ‘ought’ but rather the ‘ought of reasonableness’: pleasure provides a reason to adopt a pro-attitude towards it, but not necessarily a moral reason.13 Brentano worked with a limited set of pro-attitudes. Ewing’s notion of a proattitude extends to a much wider range of attitudes.14 In fact, he takes it to cover “any favourable attitude to something”: “choice, desire, liking, pursuit, approval, admiration”.15 He also underlines the notion that different pro-attitudes, or even conglomerates of such attitudes, ﬁt different kinds of valuable object; and he takes this to show that ‘good’ can have different senses.

p. 54). For a most interesting essay on Ewing’s metaethical views, see the forthcoming ‘A. C. Ewing’s First and Second Thoughts on Metaethics’, by Jonas Olson and Mark Timmons. At the end of his career Ewing developed a peculiar metaethical hybrid theory combining elements of both non-cognitivism and cognitivism. Olson and Timmons argue convincingly that this theory is ﬂawed in several respects. 12

Ewing (1947, p. 152). As mentioned earlier, Ewing’s metaethical views changed considerably over the years, and in ways that are not always entirely transparent. Perhaps the most noteworthy change was that he gave up on the idea he once had (in 1947) that evaluative language is primarily concerned with describing (i.e. ascribing properties), replacing this with a more prescriptivist (expressivist) account (in 1959). He then tried to combine the latter with the idea that value judgements have truth-values. 14 His account of preferences is interesting and quite different from Brentano’s: “‘Prefer’ is such a general word that I think it may be used to cover any case where a more favourable attitude is adopted to something A than to something else B, whatever the pro-attitude in question” (Ewing, 1959, p. 85). This seems to suggest that, of two attitudes, x and y, if the former is more favourable than the latter, and I apply x to object A, and y to object B, then I have a preference for A over B. Needless to say, such an account raises a number of difﬁcult issues. For instance, what makes us say that one attitude is more favourable than another? Is admiration of A a more favourable attitude than the attitude of respecting B? 15 Ibid., p. 149. 13

24

F I T T I N G - AT T I T U D E A N A LY S I S

2.3 The primacy of the normative over the evaluative Brentano’s and Ewing’s decision to give primacy to normative concepts (such as ﬁttingness, correctness, ought, should, and reason) over evaluative concepts (principally, the good) is, in my view, highly persuasive. Thus, the FA theory that will be proposed here reduces evaluative claims to deontic claims about attitudes, to the valuable objects in question, that it is ﬁtting to have, or that one ought to have, or that we have reason to adopt. The theory developed here departs from Ewing’s later view, however. Where he distinguished between two sorts of ought—the moral ought and the ought of ﬁttingness, interpreted as an “ought of reasonableness”—I will be more cautious. True, it is a central feature of values that they are primarily normative in character: we think there is a reason why a valuable object demands, requires, or calls for, certain attitudes from us, and that this reason derives from the object. But whether or not the reason is moral in character seems to me a secondary matter.16 At least, I have not found any compelling reason to follow Ewing here. This is not to deny that there are two oughts. It is merely to leave the matter open, and to assume that what is essential to an analysis of value is that it brings out the normative character of value. Here it might be a good idea to point out, as Joshua Gert17 recently did, that normativity has two dimensions: There is the strong level of requirement—what we ought to do—and there is the weaker level of permission—what we may do. Although I will not pay any particular attention to these two levels, it is an important point to keep in mind when I give examples of what there might be reasons to favour. These reasons may not all be about what we are required to do. A noteworthy feature of the FA analysis is that, in it, it is properties other than the value property that provide reasons to respond to the valuable thing by taking up an attitude to it.18 The reason we ought to take a certain responsive stance is to be looked for among the subjacent properties and not in what supervenes on those properties.19 It

16 Zimmerman, in ‘Understanding What’s Good for Us’ (2009), does not seem to agree with this. The fact that I leave it unspeciﬁed whether the reason I have in mind is prudential or moral is a problem for my view, according to Zimmerman. Perhaps. However, the concept of personal value I have in mind does not imply anything other than that something is a reason—what this reason is will become clear, eventually, when we take into consideration the supervenience base of the value. However, since there is no mention of any subjacent properties in the analysis, it cannot be determined whether the reason is moral, prudential, or some other kind of reason. There is nothing peculiar about this; in my view we often identify something as a reason without having a clue as to whether the reason is moral, prudential, or of some other kind. 17 See Gert, ‘Value and Parity’ (2004). 18 Fittingness (worthiness, correctness) might be regarded as a primitive notion that cannot be subsumed under standard generic deontic notions, including notions such as ought, must, and should. On this view, ﬁttingness constitutes a special kind of deontic notion. Ewing took this approach in his early work The Deﬁnition of Good (1947). See also Danielsson and Olson (2007). 19 Cf. Dancy, ‘Should We Pass the Buck?’ (2000a, p. 161): “Ewing seems . . . to be in a position to say that goodness is not a distinctive evaluative and intrinsic property in objects, one whose presence we can discern and to which we do or at least should respond with approval and admiration. The goodness of the object just is the relational fact that we should respond to it with approval, admiration or other pro-attitude. The evaluative ‘good’ has been deﬁned in terms of the deontic ‘should’. And with this result, the intuitionists

2.4

D I F F E R E N T VA L U E I D I O M S

25

is the set of value-making properties that provides us with a reason to take a proattitude to the valuable. Following Scanlon (1998), we could say that what carries the buck is not value but what value supervenes on! In other words, to be valuable is to have the property of having a value-making base that gives us reason to have a proresponse to the valuable object. A caveat is in place here, however: although I referred a moment ago to the supervenience base of the valuable object in order to explain what the analysis is all about, it must not be assumed that the analysis somehow directly refers to the relevant subjacent properties.20 The FA analysis is, in effect, a very simple analysis; it neither mentions nor refers to the properties in virtue of which the object is valuable. It just says that to be valuable is to be an object there is reason to favour. The FA analysis defended in this work is pluralistic in the sense that it acknowledges that ﬁnal value accrues to different kinds of metaphysical entity, including abstract states of affairs and concrete objects such as persons.21 Great importance is attached to the idea that different kinds of valuable object invite different kinds of pro-response. For instance, a brave person ought to be admired, a precious artefact calls for care and protection, and a desirable state should be desired. In fact, in many cases we may describe the response that is ﬁtting more accurately as a conglomerate of different kinds of attitude—or even as a conglomerate of various kinds of attitude-cum-behaviour. Some of these attitudes are especially thing- or person-oriented: we care for our children; we cherish and preserve objects of historical importance like the Magna Carta; we protect a part of the Brazilian rainforest that is as yet untouched by humans. Other attitudes are instead state of affairs-orientated. The most obvious cases are desires and preferences, both of which typically take states of affairs (e.g. that Denmark will qualify for the World Cup, or that people do not say bad things about one behind one’s back) as their objects rather than (concrete) things and persons.

2.4 Different value idioms The FA analysis has typically been employed to analyse what is valuable ‘for its own sake’ or ‘as an end’ or has value ‘in its own right’. However, none of these descriptors is

reversed Moore’s position . . . Moore deﬁned the right, that which we ought to do or should do, in terms of the good. Ewing deﬁned the good in terms of how we should respond.” 20 This is important when it comes to the question whether evaluative properties may themselves be reason-providing. As outlined here, FA analysis leaves this issue open, and as far as I can see there are good reasons to allow that value may supervene on other kinds of value. Scanlon, for instance, was originally of a different opinion, believing that the subjacent properties were all natural properties. However, he changed his mind in response to R. Jay Wallace, ‘Scanlon’s Contractualism’ (2002, pp. 447–9): see Scanlon, ‘Reasons, Responsibility, and Reliance: Replies to Wallace, Dworkin, and Deigh’ (2002, p. 513). 21 See, for instance, Elisabeth Anderson, Value in Ethics and Economics (1993); Christine Swanton, ‘Proﬁles of the Virtues’ (1995); Marcia W. Baron, ‘Kantian Ethics’ (1997); Wlodek Rabinowicz and Toni RønnowRasmussen (2003a). For monist approaches see Noah M. Lemos (1994); Michael J. Zimmerman (2001a) and his ‘Intrinsic Value and Individual Worth’ (2001b); and Jonas Olson, ‘Revisiting the Tropic of Value: Reply to Rabinowicz and Rønnow-Rasmussen’ (2003).

26

F I T T I N G - AT T I T U D E A N A LY S I S

quite satisfactory. The notion that something is valuable as an end, for instance, appears not to be consistent with the possibility that the value bearer is a non-propositional object, such as a thing (a painting, a stamp, or the like). It seems right, at least, to agree with W. D. Ross that ends are objectives and never things. In this sense, the other two expressions seem to represent an improvement. Both ‘in its own right’ and ‘for its own sake’ imply something about whether or not the value is dependent on something other than the valuable object. Of the two expressions ‘in its own right’ and ‘for its own sake’, I prefer the second. The notion that something is valuable in its own right might lead one to wonder whether the object has value in virtue of a capacity of the object that has been realized. If that were so, this kind of value would also exclude, it seems, certain objects from the ﬁeld of value bearers. (Would a state of affairs, say, have value in its own right in this sense?) ‘In its own right’ also seems to suggest that the value accruing to an object does so without the help of a subject. This suggests that the expression would lend itself to some extent to an objectivist analysis of value. Admittedly, the advantage of the phrase ‘for its own sake’ may not be immediately obvious, and certainly it has a somewhat counter-intuitive ring in expressions such as ‘x is valuable for its own sake’ (compare ‘x is desired for its own sake’). However, I prefer it. Following others, I will say that something that is valuable for its own sake has ﬁnal value. My reason is that the value in question here is a natural counterpart, in a sense in which the value picked out by ‘in its own right’ is not, to two other familiar kinds of value: contributory value (or what is valuable for the sake of some whole) and instrumental value (or what is valuable for something else’s sake rather than for its own). Moreover, ‘valuable for its own sake’ does not impose bias on the value-bearer issue; it rules out neither abstract states nor concrete things, unlike the notion of something being valuable as an end. Therefore, in what follows I will speak of ﬁnal values in terms of what is valuable for its own sake (and, accordingly, of what is good/bad for its own sake and desirable for its own sake). If the value is intrinsic (one kind of ﬁnal value), the object is valuable for its own sake in virtue of its internal features; if the object has extrinsic ﬁnal value, it is valuable for its own sake in virtue of at least some external relational feature. Rae Langton has objected to this sort of characterization: “to talk of something having ‘ﬁnal value’ that is ‘extrinsic’ is to talk of valuing for the thing’s own sake, something that has extrinsic value. We keep the talk of ﬁnality and ends to a distinction, not in the way things have value, but in the way we value things” (2007, p. 165). Although Langton’s proposal has its merits, ultimately I think we should resist it. What determines whether a value is intrinsic or extrinsic is, in the end, the nature and scope of the supervenience base. Moreover, her suggestion is not obviously consistent with the pattern of value analysis that I employ. It should be stressed, though, that there is a sense in which my aim here concerns only indirectly ﬁnal values. My main concern in this book is to discuss the extent to which some version of the FA analysis, if true, is helpful when it comes to understanding personal values.

2.5

T H E F A A N A LY S I S O F F I N A L P O S I T I V E VA L U E S

27

The caveat ‘if true’ here is important, and it will occupy us in a moment. Meanwhile, some further features of the FA analysis need to be highlighted.

2.5 The FA analysis of ﬁnal positive values For reasons of simplicity let ‘favour/disfavour’ be schematic place-holders for a range of positive and negative responses called for by various kinds of valuable object. Applied to ﬁnal positive values, the analysis then says the following: The FA analysis of ﬁnal positive values: The ﬁnal value of an object x consists in the existence of normative reasons for favouring x for its own sake. This pattern of analysis can be extended to other kinds of value, too. For instance, an object’s instrumental value might be said to consist in the existence of reasons to favour it for its effects’ sake. That is, given that instrumental value is in fact a kind of value. That this is not obviously the case is something that I will comment on in Section 4.2. The FA approach also applies, and equally well, to both thick and thin varieties of ﬁnal and instrumental value.22 Thus an admirable person is a person whom there is reason to admire, a respectable person should be respected, and so on. Of course, the match between the value and the ﬁtting attitude is not always as obvious as it is in these cases. For instance, to be brave is also, for the FA analyst, to be the object of a normatively required attitude. The question, of course, is what kind. That the FA pattern of analysis might encounter a problem with certain values has recently been pointed out by Roger Crisp: A painting may be beautiful because of its sublimity, delicacy, profundity, boldness, imagination, vitality, grace, honesty . . . According to the version of BPA [i.e. buck-passing analysis23] we are considering, since these properties are evaluative, we might assume that buck-passing accounts of them also ought to be available. How will BPA distinguish them? Reference to the appropriate response may be appropriate in certain cases, especially those in which reference to a response is analytic to the concept in question. ‘Being awesome’, then, may be said to be the higher-order property of having lower-order, natural properties that provide a reason for awe. But the BPA is unlikely to be able, by reference to different responses, to distinguish properties with no such analytic reference to responses themselves. Consider, say, grace and delicacy. To be sure, they call for certain responses, but the responses themselves are too similar to enable us to distinguish the properties.24

22

For the distinction, see Bernard Williams Ethics and the Limits of Philosophy (1985). The term “buck-passing analysis” is due to Scanlon (1998, pp. 95–100). 24 Crisp, ‘Value, Reasons and the Structure of Justiﬁcation: How to Avoid Passing the Buck’ (2005). See also Andrew Reisner, ‘Abandoning the Buck Passing Analysis of Final Value’ (2009). That it is often not possible to pair up emotions and values in any neat way was also argued by Kevin Mulligan in ‘From Appropriate Emotions to Values’ (1998, see pp. 174–5). 23

28

F I T T I N G - AT T I T U D E A N A LY S I S

Is this pattern of analysis too blunt an instrument, then, to deal with values like grace and delicacy? Could it discriminate between these kinds of value? Crisp offers two reasons for thinking that the buck-passing, or FA, analysis fails. First, thick values do not always connect with a unique attitudinal response. For instance, being delicate does not have an attitude or emotional response that is appropriate to it. Second, the properties in virtue of which the object is delicate—what Crisp refers to as the lower-order natural properties—are too coarse-grained to enable us to distinguish between values such as delicacy and gracefulness. Crisp’s ﬁrst observation is correct. As to his second point, I am not sure whether it really is impossible to detect a difference between delicacy and gracefulness in terms of their supervenience bases. Crisp thinks that an account of these values in purely naturalistic terms will not do. With this I fully agree, but that is because I am inclined to deny that naturalism is a viable theory in the ﬁrst place. However, this denial does not commit me to holding that we cannot differentiate between values in terms of their lower-order properties, or what I would refer to as their supervenience bases. We can, even if doing so is often a difﬁcult business; it is just that there might be more to value than their supervenience bases. Thus, in response to Crisp, I would suggest that it is the combination of lower-order properties and their reason-making nature that enables us, in principle, to differentiate between these two values. In the case of a delicate object there will be a set of properties such that they provide us with a reason to feel, say, delight. In the case of grace, another set of properties will give us reason to have another attitude—say, aesthetic enjoyment (as has been suggested by Jonas Olson).25 In fact, since the two values are very similar, they might well provide us with a reason to adopt the same attitude to the object. The difference would then be accounted for only in terms of the supervenience bases. For instance, the properties that make a vase into a delicate vase do not seem to be the properties that make it into a graceful vase; among the latter we would ﬁnd, say, its ﬂowing shape, whereas a delicate vase (in at least one sense of ‘delicate’) would have properties that have something to do with how easily it is damaged—say, that it is made of porcelain. Of course, things are often both delicate and graceful, and so the supervenience base will contain common elements. Again, this is not in any way to assert, or imply, that these values can be captured in a purely naturalistic analysis. A rejoinder might be to say that, if these claims are correct, we will have to admit that we are really dealing not with two different values at all, but with one kind of value that we refer to differently because it has two distinct supervenience bases. But supposing that this were the case, would that be a serious problem? I do not think so. Crisp also suggests that the buck-passing, or FA, analysis fails in another aspect. The experience of beauty is not merely the experience of the natural properties in virtue of which the object is beautiful. The reason we admire the artefact is that it is beautiful:

25

Olson, ‘How to Pass the Buck’ (2004).

2.5

T H E F A A N A LY S I S O F F I N A L P O S I T I V E VA L U E S

29

. . . the reason for admiration lies not in the natural properties—these could be understood by someone with no aesthetic sense—but in the beauty.26

The issue Crisp raises here is both complex and deserving of detailed discussion. However, I will conﬁne myself to a short reply. Even a buck-passer might go along with the possibility that two persons do not reach an agreement on whether something is beautiful or not despite perceiving the same natural properties. But instead of invoking the ability of one, and the disability of the other, to perceive non-natural properties, the explanation would be that one grasped and the other failed to grasp that the natural properties are reason-providing properties. Moreover, I am not conﬁdent that my own ‘value phenomenology’ rules out the latter explanation; I cannot be certain that for phenomenological reasons the ﬁrst explanation is better (or worse) than the second. Given this, and given that the second explanation does not require us to posit a nonnatural property, the FA pattern of analysis appears to be the more plausible one.27 There is another recent objection to the FA analysis that deserves a comment.28 It was ﬁrst pointed out by Wlodek Rabinowicz in ‘Value Relations’,29 and later Mark Alfano has expressed a similar point.30 Alfano’s argument sets out from the idea that it is conceptually true that ‘good’ and ‘bad’ are polar predicates, i.e. they are contraries. And so an analysis of good and bad should keep this polarity in its analysans, but the FA analysis runs into problems here since it understands value in terms of attitudes that are not necessarily contraries. There is nothing conceptually problematic about being in an ambivalent state of favouring and disfavouring one and the same object. So if we suppose that it is actually ﬁtting to be in such a state, then on FA, we need to say that the object is both good and bad. Now, a modiﬁed FA need not be endangered by this. Suppose we take the FA to say something to the following effect: x is good if and only if it is ﬁtting to favour x and it is not ﬁtting to disfavour x. This should handle this sort of challenge.31 Rabinowicz’s point (2008, p. 40) is more general, though; FA reduces conceptual truths concerning value notions to formal constraints on appropriate attitudes. For instance, consider the ‘betterness’ relation. As Rabinowicz suggests this seems to be as a matter of conceptual truth, a transitive relation. However, it is not obviously a conceptual truth that preference in all permissible preference orderings is transitive (e.g. I may prefer x to y, and I may prefer y to z, but it is not permissible that I, at the same time, do not prefer x to z). So if the FA analysis forces us to treat a permissible preference ordering as transitive, (which is not unreasonable but nonetheless not conceptually true), it obviously has non-tautological implications. Of course, 26

Crisp (2005, p. 82). Crisp raises other objections as well. For discussion of these, see Olson (2004). Recently Crisp has returned with a new objection to the buck-passing account. I will return to this objection at the end of the next chapter. 29 Rabinowicz, ‘Value Relations’ (2008). 30 Alfano, ‘A Danger of Deﬁnition: Polar Predicates in Moral Theory’ (2009). 31 The point is not mine. Unfortunately I cannot remember its source, or the person who originally suggested this solution to Alfano’s challenge. 27 28

30

F I T T I N G - AT T I T U D E A N A LY S I S

an FA analyst should perhaps bite the bullet here. After all, as I have pointed out earlier, FA is to some extent a revisionary analysis. However, there is another, albeit more radical, possibility, namely that we need not understand ‘better’ in terms of dyadic attitudes (such as preferring x to y). Perhaps we could understand ‘better’ in terms of it being ﬁtting to have certain monadic attitudes rather than other monadic attitudes. We could then account for the ‘betterness’ of x over y in terms of the difference of strength between the two monadic attitudes. For example x is better than y, if and only if, it is ﬁtting to, say, like x to a degree n+1, and it is ﬁtting to like y to a degree n. Needless to say, whether such an analysis can be worked out in a plausible way is an interesting issue. However, I will not pursue this matter here.32 Instead, at this point it is convenient to draw attention to one of the earliest attacks on FA analysis.

2.6 W. D. Ross’s objection Ross objected to the whole FA project.33 Attempts to reduce statements such as ‘x is good’ to ‘it is ﬁtting to admire x’ appeared to him not to offer a genuine analysis:34 in Ross’s view, FA ‘analysis’ got things backwards. However, this was not his only worry; FA seemed to him to also involve a circle: [A]dmiration is not a mere emotion; it is an emotion accompanied by the thought that that which is admired is good. And if we ask on what ground a thing is worthy of being thought to be good, only one answer is possible, namely that it is good. It would be absurd to say that a thing is good only in the sense that it is worthy of being thought to be good, for our deﬁnition of ‘good’ would then include the very word ‘good’ which we were seeking to deﬁne.35

To analyse goodness in terms of the attitude of admiration would be circular, if admiration involves a judgement to the effect that the object admired is good. This is an important objection.36 Given his other claim, then—that goodness is the very feature in virtue of which the object is, to use Ewing’s original phrase, worthy of admiration, and cannot therefore be understood in terms of being worthy of admiration—Ross appears to be challenging advocates of FA analysis on two fronts.37

32 Recently Rabinowicz has developed what a modelling of value relations would look like from a monadic-centred perspective; see his ‘Values Compared’ (2009). 33 Ross, Foundations of Ethics (1939, pp. 276, 278). His direct target was Ewing’s view. 34 At least, if what is valuable is an action. If it is instead an experience of pleasure, then, insofar as it is valuable, it is a ﬁtting object of satisfaction rather than admiration (ibid.). 35 Ibid. p. 279. 36 According to Mulligan, “Meinong [in (1968) Abhandlung zur Werttheorie, Gesamtausgabe, Vol. 3, Graz: Akademische Druck- u. Verlagsanstalt] noted that analyses of value in terms of justiﬁed or correct emotions avoid circularity only if justiﬁcation and correctness are deontic properties and if deontic properties and axiological properties are different types of property. He accepted both claims and took deontic properties and value-properties to be formal objects of desire and of emotions, respectively” (2009b, p. 490). 37 The circularity need not be vicious. It depends, I suppose, on what we hope to accomplish by the analysis. For instance, David Wiggins accepts the circularity. He argues that, owing to its “detour through

2.6

W. D. RO S S ’ S O B J E C T I O N

31

Ewing replied to Ross’s criticism, arguing: [T]he reason why it is proper to admire anything must be constituted by the qualities which make the object of admiration good, but it does not follow that the thought that it is good must, if the admiration is to be justiﬁable, intervene between the perception of the factual qualities admired and the feeling of admiration.38 [T]he ground [for a pro attitude] lies not in . . . goodness, but in the concrete, factual characteristics of what we pronounce good. Certain characteristics are such that the ﬁtting response to what possesses them is a pro attitude, and that is all there is to it.39

Thus (i) admiration is a response to the object’s ‘good-making’ qualities and not to its goodness. For this reason, (ii) the attitude of admiration need neither involve, nor presuppose, any judgement that the object admired is good. Do pro-attitudes have to involve any evaluations? In recent years Ross’s position has been questioned. Justin D’Arms and Daniel Jacobson,40 in particular, have mounted an attack on various forms of ‘judgementalism’ and ‘quasi-judgementalism’41 that purport to analyse emotions: the fact that evaluative language may be useful for describing our emotions should not lead us to believe that all emotions essentially involve some judgement of value. And likewise, it would be a mistake to think that to be in an emotive state the agent has to be entertaining evaluative judgements or concepts.42 The evaluative concepts we employ to depict emotions need not correspond to any ‘phenomenological’ reality in the emotion itself. D’Arms and Jacobson therefore join ranks with Ewing in rejecting the Ross-style circularity objection to the FA analysis. D’Arms and Jacobson have a point, as I see it. However, even if they were mistaken, and emotions somehow did necessarily involve evaluations or were deﬁned in terms of certain values, this would not in itself be an insuperable obstacle to the FA approach. It would probably present a problem for any attempt to provide a semantic reduction of value claims to claims about the attitudes and emotions we have reason to have. But the FA analysis is not (at least, primarily) an exercise in semantic reduction. Suppose that the attitudes FA employs can be characterized in ways other than by their relation to value. If so, there is a point to explicating value in terms of a normative component and these attitudes. At some level such an approach might well be circular, but the circle would be informative. sentiments”, the analysis remains informative to some extent; see David Wiggins, ‘A Sensible Subjectivism?’ (1987, p. 189). Cf. D’Arms and Jacobson, ‘Sentiment and Value’ (2000b, pp. 732ff.). 38

Ewing, The Deﬁnition of Good (1947, p. 158). Ibid., p. 172. 40 See D’Arms and Jacobson, ‘The Signiﬁcance of Recalcitrant Emotion (or, Anti-Quasi judgmentalism)’ (2003). Also relevant is their ‘Sentiment and Value’ (2000b). 41 Quasi-judgementalism is, according to D’Arms and Jacobson, a revision of judgementalism in that it “still type-identiﬁes the emotions by their deﬁning propositions, and claims that certain thoughts are partially constitutive of being in an emotional state, but loosens the requirement that these thoughts must be afﬁrmed by the agent” (2003, p.130). 42 For an overview of different positions on this matter, see Mulligan’s ‘Emotion, Value, and Morality’ (2009b). 39

32

F I T T I N G - AT T I T U D E A N A LY S I S

Skorupski has recently addressed this issue. Supposing the morally wrong is that towards which there is reason to feel blame, he considers the objection that this (FA) proposal: cannot yield a deﬁnition of ‘morally wrong’, since ‘blame’ in the relevant sense should be deﬁned as that sentiment which is appropriate—which there is reason to feel—towards the morally wrong. A good objection, but not the end of the story. Suppose that there is a distinctive sentiment of blame whose object is the morally wrong; suppose also that we can be given an independent characterization of this sentiment, say in terms of the actions to which it disposes, or by having it explained as the sentiment we feel when we consider some speciﬁc paradigm cases. Suppose, ﬁnally, that when we have been made familiar, in the ﬁrst-person way, with the sentiment, we ﬁnd ourselves able to go on spontaneously, and reasonably convergently, making conﬁdent new judgements about when there is reason to feel that speciﬁc sentiment. In that case we have everything we need to grasp the concept of moral wrongness. The morally wrong is that which is blame-sentiment-worthy in the absence of an excuse. The concept can in this sense be exhaustively captured in terms of the concept of a reason and the concept of a certain sentiment, even though ‘morally wrong’ cannot be deﬁned in terms of ‘reason’. Furthermore if we can individuate the sentiment by these methods we can also use them to introduce a term to refer to the sentiment, say ‘BS’. The morally wrong is that which is BS-worthy (which there is sufﬁcient reason to respond to in that way) in the absence of an excuse.43

Skorupski is right, I think. We can indeed say something about what it means to be in possession of the concept moral wrongness even if we cannot semantically reduce statements about what is morally wrong to statements about what there is reason to blame. Once we grasp the concept of a reason and the concept of blame (particularly in terms of a morally wrong independent characterization) we are in a position to grasp moral wrongness. Even if we allow, then, that emotions and attitudes necessarily involve evaluations or somehow relate to values, it remains to be shown that the sort of circularity involved in FA analysis, given such an allowance, would be devastating to the analysis. Some, but far from all, circles are vicious.44

43 Skorupski, ‘What is Normativity?’ (2007b, p. 249). That FA analysis need not involve a vicious circle was considered as a possibility in Rabinowicz and Rønnow-Rasmussen (2004a, 2006). 44 See I. L. Humberstone’s seminal ‘Two Types of Circles’ (1997). See also J. A. Burgess, ‘When is Circularity in Deﬁnitions Benign?’ (2009).

3 The Wrong Kinds of Reason

The ﬁtting attitude (FA) analysis of value has many attractions, but it also faces several more or less serious internal problems, and these will need to be acknowledged and examined. Its more salient positive features will emerge, eventually. But at this stage it is only fair to the reader to present what has been thought of as a major problem for the analysis. This is far from the only problem. Some of the problems concern the analysis in general;1 others relate to my speciﬁc application of it to personal values. The latter will be examined in due course. In this chapter I focus mainly on what Wlodek Rabinowicz and I have called the ‘wrong kind of reason’ problem (henceforth, the WKR problem). At the end of the chapter I will comment brieﬂy on another problem, one that concerns whether an FA analyst has to admit that goodness cannot be a reason-giving property.

3.1 The wrong kind of reason problem The FA analysis equates value with the existence of reasons to favour the object carrying the value. One obvious advantage of the FA analysis is that it satisﬁes a requirement that all formal axiological theories should meet: it is silent as to what objects carry value. Substantive axiological reasoning aims to tell us what kinds of thing are in fact valuable, but this is not the objective of formal axiology. This notwithstanding, if the analysis is too generous and implies, counter-intuitively, that certain kinds of

Although I do not share his conclusion, Krister Bykvist’ s recent article ‘No Good Fit: Why the Fitting Attitude Analysis of Value Fails’ (2009), contains an interesting discussion concerning ﬁtting attitude analyses that take non-obtaining states of affairs to be valuable. Bykvist considers whether we can favour a (nonobtaining) state of affairs such as “there being happy egrets but no past, present or future agents (i.e. beings who intentionally bring something about)” (p. 5). He considers several possible ways of understanding “to favour” something, but none of them, he thinks are such that we can reasonably say that it is ﬁtting to favour such a state. For example, it would not make sense to say that it is ﬁtting to promote such a state or to desire it. This certainly seems true. However, it might be ﬁtting to direct some optative attitude on this state of affairs. We can certainly wish that it obtains even if we know that it cannot be promoted. 1

34

T H E W RO N G K I N D S O F R E A S O N

apparently valueless objects are valuable, then there is obviously a reason to be sceptical about it. The problem—perhaps, indeed, the biggest problem—facing advocates of FA analysis is precisely this: the FA analysis seems to open the door to worthless objects, and to construe them as bearers of value. Equally, we might have reasons not to have pro-attitudes to objects that obviously are valuable.2 Discussions of the WKR problem have focused mainly on the ﬁrst kind of case. Most likely the explanation is that in cases of the second kind we do have reasons to have a pro-attitude, but these reasons are outweighed by reasons not to have such an attitude. So, these are not counterexamples to FA analysis. Here are two examples in which there seem to be perfectly good reasons to take pro-attitudes to objects of no value at all:3 (i) A demon threatens you with punishment if you do not admire a saucer of mud.4 (ii) A demon threatens you with punishment if you do not admire him.5 These examples appear to take the FA analyst straight into a cul-de-sac. Here is why: the saucer of mud has no value (or so we assume); yet, if the analysis is correct, we are forced to say that it does have value, since there are reasons to take up a positive attitude to it. The same goes for the demon in (ii). What is signiﬁcant about not only (i) and (ii) but WKR cases in general is that it is seemingly the pro-attitudes themselves, rather than the objects, that are valuable in these cases: it is the admiring of the saucer or demon, not the saucer or demon in themselves, that carries value. However, in the context of the FA analysis, the fact that we have reason to admire the object is just another way of saying that these objects are in fact valuable. We are apparently cornered. Objects which obviously are not valuable are depicted in the analysis as being of value.

2 Justin D’Arms and Daniel Jacobson have identiﬁed and discussed this difﬁculty, which they refer to as the “the Conﬂation Problem”, in their ‘Sentiment and Value,’ (2000b). See also D’Arms and Jacobson (2000a) and their ‘The Signiﬁcance of Recalcitrant Emotion (or, Anti-Quasi judgmenta-lism)’ (2003). 3 As Sven Danielsson and Jonas Olson have pointed out in their recent ‘Brentano and the Buck-passers’ (2007), G. E. Moore raised a kind of WKR objection that would be an example of the other type of case. In his review of the English translation of Brentano’s Vom Ursprung Sittlicher Erkenntnis, Moore argued that beautiful objects are such that we should favour them. However, this does not mean, he thought, that value accrues to the beautiful object; it is rather our appreciation of them that is valuable. Moore’s contention is not uncontroversial, though. It is not obvious that it is only our appreciation that is valuable, but not the beautiful object itself. 4 This example is from Roger Crisp’s review of ‘Value . . . and What Follows’ by Joel Kupperman (2000a). If you prefer a more realistic example, recall the scene from the movie Catch 22 in which the two corrupted colonels, Cathcart and Korn, offer to send Captain Yossarian home from the war (they want to hush up an ugly affair they are involved in), something the captain has wanted for a long time. There is a catch, though. Yossarian hates the colonels, and their offer comes with a price. Yossarian has to, as the colonels put it “like us”. The scene is at the end of the movie. 5 This example is discussed in Wlodek Rabinowicz and Toni Rønnow-Rasmussen (2004a); it was originally suggested by Folke Tersman (personal communication).

3.1

T H E W RO N G K I N D O F R E A S O N P RO B L E M

35

Several philosophers have tried to solve this problem. To do justice to all the attempts would require too much space, but I will say something about what is perhaps the most popular general approach. It sets out from Derek Parﬁt’s distinction between state-given and object-given reasons for attitudes:6 Of our reasons to have some desire, some are provided by facts about this desire’s object. These reasons we can call object-given. We can have such reasons to want some thing either for its own sake, or for the sake of its effects . . . Other reasons to want some thing are provided by facts, not about what we want, but about having this desire. These reasons we can call state-given. Such reasons can also be either intrinsic or instrumental.7

In other words, facts provided by the very attitude, and not the object of the attitude, introduce what Parﬁt calls state-given reasons; and facts that have to do with the object of the attitude introduce object-given reasons. How does this distinction offer a solution to the WKR problem? Suppose we have identiﬁed a reason for favouring an object. Given this, we can ask: Is our reason for favouring the object an object- or a state-given reason? The straightforward solution on offer at this point is to claim that the right kind of reason will always be an object-given reason, and the wrong kind of reason always a state-given reason.8 Recall the example of the saucer of mud. This is an object without value, as (we are supposing) is the demon. In cases (i) and (ii) we seem to have a reason to admire the object (the saucer or the demon), but the properties that provide these reasons are of the wrong kind: they belong to the attitude rather than to the object. If I do not admire the saucer or the demon, I will be punished. This is a fact about the attitudes involved, not about the saucer of mud or the demon. But the success of this approach is not quite as clear-cut as it seems to be. There is a complication, as Rabinowicz and I pointed out: It is easy to see that for any property P of the attitude there is a corresponding property P’ of the object: If a pro-attitude towards an object a would have a property P, then, ipso facto, a has (or would have, if it existed) the property P’ of being such that a pro-attitude towards it would have the property P. Consequently, to the attitude-given reason, provided by P, corresponds the object-given reason, which is provided by P’. In exactly the same way, of course, for any property P of the object of the attitude there is a corresponding property P’ of the attitude itself: the property of being such that its object has (or would have) property P. Thus, to each objectgiven reason corresponds an attitude-given reason, and vice versa.9

Parﬁt, ‘Rationality and Reasons’ (2001). Ibid., pp. 21f. Parﬁt returns to this distinction in his forthcoming ‘On What Matters’. See its Appendix B. 8 Parﬁt himself denies in ‘Climbing the Mountain’ (circulating ms) and in ‘On What Matters’ (circulating ms) that state-given reasons are reasons for favouring, and so he would deny that there is a WKR problem in the ﬁrst place. I will return to this approach in a moment. 9 Rabinowicz and Rønnow-Rasmussen (2004a, p. 406). 6

7

36

T H E W RO N G K I N D S O F R E A S O N

To differentiate between these two sorts of property—between “object” properties and “attitude” properties—will, then, not obviously serve to remove the present difﬁculty.10 Now, there might in fact be a rather easy way out of this problem, namely to exclude all relational properties from the supervenience base of value. At least, it might seem we can do this for ﬁnal value. If it is only among an object’s internal, nonrelational features that we ﬁnd value-making properties, and if, therefore, ﬁnal value is nothing but intrinsic value, which the object has in virtue of its internal features alone, then the WKR issue seemingly will never arise. And so it can be argued, then, that not every kind of FA analyst has this problem to contend with—only those, like me, who consider ﬁnal value to be extrinsic in some cases.11 There is a rejoinder, however, that questions the idea that the WKR problem only arises for ﬁnal extrinsic value. All we need to do is to ﬁnd an object whose internal features are such that they provide us with a reason to favour the object despite the fact that it is obviously not valuable. If there were such an object we would have to be able, as FA analysts, to distinguish the sort of reason involved in such a case from the kind employed in an FA value analysis. Perhaps there are such objects. Rabinowicz recently suggested to me (personal communication) that a fact of the following kind might be an example: I will be punished if I do not favour this fact (that I will be punished . . . ). This is a peculiar fact—a self-reﬂexive one. It does seem to give me a reason to favour the threat. So if we are ready to accept this particular kind of fact, then it does seem as the nature of this sort of reﬂexive fact is such we have a reason to favour a valueless object in virtue of In ‘How to Deal with Evil Demons: Comment on Rabinowicz and Rønnow-Rasmussen’ (2005), Phillip Stratton-Lake is worried that if we tolerate such properties as that of being an object whose favouring would have property P, we will end up with an objectionable ontology. This might be true if all such ‘mirror properties’ had an ‘imaginary’ touch to them, but this is not the case. For examples of more robust object properties that reﬂect the properties of the attitudes, see Rabinowicz and Rønnow-Rasmussen (2004a, p. 407). For example, take the case of the demon who demands that we admire him rather than the saucer of mud, and who makes it clear that he will strike us if we do not. We seem to have a reason to admire the demon which is of the wrong kind. That the demon is disposed to strike is not something that makes him admirable or valuable in a positive sense. Nonetheless, his disposition to hurt us provides us with a reason for admiring him. But this feature of the demon certainly is a most robust feature. 11 Even if we were to allow relational properties among the value-makers, could we not get rid of the WKR problem by qualifying the supervenience base in some way? For instance, I discussed the following example with Heath W. White (personal communication): consider a thick value such as praiseworthiness. We might want to say that if, and only if, a property contributes to the overall purpose for which we have the general practice of praising can it be a praiseworthy-making property. So if we were to ﬁnd some object that does not contribute, or even runs counter, to this overall purpose, we could be sure that it is not valuable. But whether or not we believe that there is such an overall purpose, this is not the way to handle WKR problems. Suppose we did argue that a certain alleged valuable object cannot be valuable, because the properties in virtue of which it is claimed to be valuable are not, or are not consistent with, the properties belonging to ‘real’ valuable objects. But is it possible to determine the purpose of the practice of praising without taking a stand on what are the right and wrong reasons for praising? I suspect it will be quite hard to do so. In other words, if what we do here is merely set out from a substantive view of the right kinds of reason, we have not achieved very much. In this particular case, I think it would be a mistake to restrict the properties on which praiseworthiness depends in the manner suggested above. Why should we be excluded from praising someone just because he or she, say, tries to change the practice of praising? 10

3.2

D UA L - RO L E AT T I T U D E S

37

its internal features alone. Rabinowicz’s suggestion therefore indicates that the WKR issue has a more general relevance than what Rabinowicz and I originally believed (Rabinowicz and Rønnow-Rasmussen, 2004a). At least certain objects give rise to the WKR issue in virtue of only their internal features.

3.2 Dual-role attitudes Several attempts to deal with this sort of problem have been made. The following suggestion is one that Rabinowicz and I have considered: On the FA-analysis, the value of an object consists in there being reasons to favour the object in question. But favouring is not supposed to be a non-discerning attitude. Rather, the idea is that we are to favour the object on account of some of its properties. These properties, then, appear in the intentional content of the pro-attitude. At the same time, they are supposed to make the object valuable. Consequently, they also provide reasons for favouring the object. Thus, they have a dual role. On the one hand, (i) they appear in the intentional content of favouring as the features on account of which the object is favoured. On the other hand, (ii) they justify favouring the object in that way; i.e., provide reasons for the pro-attitude in question.12

In cases where the right kinds of reason are in play, properties of the valuable object justify our favouring of the object, but these properties also appear in the intentional content of the attitude as those for the sake of which we favour that object. For instance, someone I am fond of, whose friendship I value very much, will call for various pro-attitudes, and these attitudes will have in their intentional content such properties of the friend as his kindness, wit, generosity, and so on. These are the qualities for which I like him. They are part of the intentional content of the attitude. But they also provide reasons for my liking him. Consider now the saucer of mud example again. Here the idea is that we should desire it for its own sake, since doing so would protect us from the fury of the demon. In this case the property that justiﬁes our favouring the saucer of mud for its own sake (i.e. the property of being such that favouring will protect us from the fury of the demon) cannot appear in the intentional content of that favouring as the property on account of which we favour it in this way. The requirement that reasons play a dual role appears to explain why certain cases involve the wrong kind of reason. Unfortunately, however, it cannot handle every kind of case. There are examples in which the reason for favouring an object does play a dual role but is still the wrong kind of reason. To see this, consider the demon case, (ii), in more detail: the demon wants us to admire him for his own sake precisely on account of his determination to punish us if we do not do so (i.e. he wants to be admired as an individual making that threat). Under these circumstances, the dual-role requirement seems to be satisﬁed: the demon’s determination to punish us if we do not comply with his wish 12

Rabinowicz and Rønnow-Rasmussen (2004a, p. 414).

38

T H E W RO N G K I N D S O F R E A S O N

provides the reason for our admiration and at the same time appears in the intentional content of that attitude as the feature for which its object is being admired. Certainly, the demon’s determination to inﬂict punishment if we do not admire him provides us with a reason. But it is the wrong kind of reason, since it relates to a feature that makes him worse, not better.

3.3 Solving the WKR problem? Should we then consider the WKR problem unsolved? Perhaps not. Let us return to the saucer of mud example, and consider what might provide us with a way of tackling WKR cases. Let us call the attitude involved in admiring a saucer of mud W. As we have seen, two sorts of reasons seem to be involved here. That is, certain features of the mud give us a reason not to admire it. (These are, in Parﬁt’s terms, the object-given reasons for the attitude.) On the other hand, in cases involving a threatening demon, intuitively, we also seem to have reason to admire the saucer of mud, and this reason has to do with the positive effects of having this attitude. Admiring the mud would prevent us from being punished. Now, Parﬁt argues that the way I have just described this case is not ideal.13 It is a mistake to think that the property of the attitude (of seeing to it that the demon does not carry out his threat) gives us a reason for the attitude W. Rather, what we are given, according to Parﬁt, is a reason for a second-order attitude: a pro-attitude favouring possession of the ﬁrst-order attitude of admiring the mud or bringing it about that we admire the saucer of mud.14 John Skorupski has taken a similar approach to the WKR problem:15 Suppose the violin performance is not good, but the evil demon will punish me with eternal torture if I fail to admire it. Is that not sufﬁcient reason for me to admire it, even though it is not good?16

Skorupski then suggests that: . . . we can apply the distinction between reasons to believe or feel on the one hand and reasons to bring it about that one believes or feels on the other. Thus: there is no sufﬁcient reason for me to admire the performance, though there certainly is sufﬁcient reason for me to bring it about

13

Personal communication. Personal communication. For clariﬁcation of Parﬁt’s case for saying that we should describe facts about attitudes as giving us a reason to have a second-order attitude towards the attitude, see his ‘Climbing the Mountain’, Chapter X. 15 See also John Skorupski, ‘Buck-Passing about Goodness’ (2007a), and Ingmar Persson, ‘Primary and Secondary Reasons’ (2007). Though the distinctions presented in these works are not identical to Parﬁt’s, all three works take the same broad approach. 16 Skorupski, ‘What is Normativity?’ (2007b, p. 258). 14

3.3

S O LV I N G T H E W K R P RO B L E M ?

39

that I admire the performance, if I can. In other words, in this case there is reason for me to bring it about that I admire something which there is no reason for me to admire. (p. 258)

Skorupski’s solution is then quite straightforward: the wrong kinds of reasons are reasons to cause or bring it about that we have attitudes (say, that you intend to do some action); these should not be confused with the right kind of reasons to favour (say, admire the demon). But Parﬁt and Skorupski’s contention is surely somewhat contentious. It rests on the far from obviously correct idea that having a reason to adopt an attitude can never give us a reason to actually favour the object of the latter; it is always a reason to bring it about that we favour some object (while not bringing about our actually favouring it). But it is hard to understand on what ground we make this categorical exclusion. What difference between ﬁrst- and second-order attitudes ensures that only the latter, not the former, can be called for by certain kinds of reason-giving property? (Moreover, it might be suspected that we have smuggled in a substantive conception of reasons in order to shut out the conceptual possibility that a reason to have an attitude provides the subject, in effect, with a fully-ﬂedged reason to favour the object rather than giving her merely a reason to bring about our favouring the object. However, since I am not sure whether this is in fact what is going on here, I will not press this point.17) Suppose we go along with the idea that facts about an attitude do not give us reasons to have the attitude; they only call for us to favour having it. If we are ready to ignore the ad hoc character of this move, we have a way of tackling WKR cases. However, this still does not clear the way for the FA analysis, for, as Rabinowicz and I have previously pointed out,18 a problem remains: we need next to distinguish what makes something a reason for admiring, wanting, or desiring an object from what gives us a reason to a have a second-order attitude to these ‘direct’ attitudes. The challenge is to do this without involving our notion of what is valuable. At the moment I am not sure this can be achieved. It might be replied that we could just say that something X is a reason for an attitude F if and only if X makes F ﬁtting or correct. This would, in effect, be to follow in Franz Brentano’s and A. C. Ewing’s footsteps and regard notions such as ‘correctness’ or ‘ﬁttingness’ as primitive notions in the analysis. Of course, it is never ideal to rest one’s analysis on an unexplained notion. For one thing, it seems quite pointless to replace one obscure notion (value) with another one (correctness, ﬁttingness). If we do not know how to distinguish between considerations that make ﬁrstorder attitudes ﬁtting or correct and considerations that make second-order attitudes ﬁtting or correct, nothing seems to be gained by taking such a step. 17 See Andrew Reisner, ‘Abandoning the Buck Passing Analysis of Final Value’ (2009), who proposes a counter-example to Skorupski’s solution: “One difﬁculty with Skorupski’s view is what I have called elsewhere blocked ascent. In blocked ascent, the evil demon will cause trouble for you if you cause yourself to have the relevant pro-attitude. The only way to get the prize, so to speak, is just to have the pro-attitude. If you are lucky enough to already have it, then necessarily you can have it. So, there can be at least some cases in which the Skorupski solution will fail to provide any reason at all” (p. 388). 18 Rabinowicz and Rønnow-Rasmussen (2004a).

40

T H E W RO N G K I N D S O F R E A S O N

I would like to consider yet another approach, recently suggested by Sven Danielsson and Jonas Olson (2007). What they are proposing is not, according to themselves, identical to the Skorupski (and, we might add, the Parﬁt) solution. This might, strictly taken, be true. Nonetheless, as I shall be arguing, it is quite difﬁcult to spot any real difference.

3.4 Danielsson and Olson on the WKR problem Danielsson and Olson’s approach is strongly inﬂuenced by Brentano. They distinguish to begin with between two sorts of reason for attitudes: a ‘holding-reason’, which they understand as a reason for having the attitude, and a ‘content-reason’, which is a reason for the correctness (or ﬁttingness) of the attitude. To give a holding-reason for an attitude is, then, to provide arguments supporting the idea that we ought to have the attitude. On the other hand, to give a content-reason for the attitude is to argue that the attitude is correct. (We shall return to this notion of ‘correctness’ in a moment.) Moreover, Danielsson and Olson believe that: a content-reason for some attitude, a, implies a defeasible holding-reason. But crucially, that there is a holding-reason for some attitude does not entail that there is a content-reason for that attitude. In other words, that there is a reason to have an attitude does not entail that the attitude is correct. (2007, p. 515)19

The distinction between these two kinds of reason furnishes Danielsson and Olson with a response to the WKR problem. Consider again, the demon case. Danielsson and Olson can acknowledge that the demon does give us a reason to favour the saucer of mud or the evil demon (when the demon insists, on pain of a severe punishment, that we admire him rather than the saucer of mud) but insist that it is the wrong kind of reason. We ought indeed to favour the saucer of mud (or admire the demon), but that is not the correct attitude to take towards the mud (or demon). The authors’ proposal, then, is that “‘x is good’ means that ‘x has properties that provide content-reasons to favour x’” (2007, p. 520). This is a neat suggestion. However, how to distinguish between holding- and content-reasons remains a problem. Danielsson and Olson suggest that we follow Brentano and take correctness as a primitive. As I mentioned earlier, this is obviously not an optimal way of dealing with these issues, but perhaps it is in the nature of things that we are forced to take such a step. So I will not repeat my reservations, but rather address the more interesting issue, namely whether their approach adds something substantially different to the Parﬁt–Skorupski solution. Danielsson and Olson are quite 19 Later the authors claim that “to say that there is a holding reason to favour the evil demon is to say that there is a content-reason to favour favourable attitudes towards the evil demon, or possibly a content-reason to disfavour the non-occurrence of such attitudes. This implies that favouring the demon is good, which is the right implication; favouring the demon is instrumentally good since it shields us from punishment” (2007, p. 519).

3.4

DA N I E L S S O N A N D O L S O N O N T H E W K R P RO B L E M

41

explicit about this matter: their distinction between holding- and content-reasons does not coincide with the one between reasons for attitudes and reasons to bring it about that one has an attitude (nor with Parﬁt’s distinction between state- and objectreasons). Notwithstanding this, I think it is hard to see that their approach is substantially different from the Parﬁt–Skorupski solution. To see this we need to understand in more detail how Danielsson and Olson conceive of the relation between holding- and content-reasons. Consider the following passage: . . . the notion of a holding-reason should be analysed in terms of the notion of a content-reason. To say that there is a holding-reason to have some attitude is to say that there is a content-reason to favour the occurrence of this attitude, or possibly that there is a content-reason to disfavour the non-occurrence of this attitude. (2007, p. 518)

Apparently, the talk about there being two kinds of reason is not entirely accurate. What Danielsson and Olson in effect are saying here is that holding-reasons are nothing but a certain kind of content-reason, namely content-reasons to have second-order attitudes. But if Danielsson and Olson’s approach is to reduce holding-reasons to contentreasons, what difference is there between their solution and the one provided by Parﬁt and Skorupski? Parﬁt, for instance, recognizes only object-given reasons for attitudes.20 On his account, there is no reason to admire the demon but a reason to favour admiring the demon. Danielsson and Olson would concur: there is no content-reason to admire the demon, but there is a content-reason for the second-order attitude (admiring the demon). Danielsson and Olson’s talk of “holding-reasons” becomes therefore empty if holding-reasons are nothing but content-reasons for second-order attitudes. Thus, the difference between the two positions seems purely verbal if holding-reasons are analysable in terms of content-reasons.21 Danielsson and Olson also maintain that what we ought to do and what we have the right kind of reason to do do not necessarily coincide. Thus, it might not be correct to admire the demon (in the sense that there is no right reason to do so, and hence no content-reason to do so) even though that is what we ought to do, where this ‘ought’ should be understood in terms of holding-reasons. Intuitively, we do seem to admit that there is a difference between what we ought to 20 Parﬁt’s ‘On What Matters’, Appendix B (circulating ms), ends in a way that I ﬁnd puzzling. After making it clear that “we do not, I suggest, have state-given reasons to have beliefs or desires”, he adds the following paragraph: “We may have state-given reasons to be in some other kinds of state. I might truly claim, for example, that I have a reason to be in Paris next April. But as I have argued, such reasons would have no importance. It would be enough to claim that I have reasons to want to be in Paris next April, and to go there, if I can” (p. 426, with reservation for changes). If there can be state-given reasons of this kind, why cannot there be state-given reasons for attitudes? I owe Rabinowicz for drawing my attention to the paragraph. 21 Cf. Gerald Lang, ‘The Right Kind of Solution to the Wrong Kind of Reason Problem’ (2008), who argues (and I believe for different reasons) that Danielsson and Olson’s solution is “uncomfortably close to John Skorupski’s favoured solution to the WKR problem”. That Lang’s view is based in part on a misunderstanding of their views is argued by Olson in ‘The Wrong Kind of Solution to the Wrong Kind of Reason Problem’ (2009).

42

T H E W RO N G K I N D S O F R E A S O N

favour and what it is correct to favour. That I ought to favour the demon seems to say something different from that it is correct to favour the favouring, say, of the demon. However, since holding-reasons are just content-reasons for second-order attitudes (of the kind exempliﬁed in the claim that we should admire the demon), then Danielsson and Olson might in effect be said to offer a solution that blurs this distinction. So unless Danielsson and Olson can explain in some other way why we should divide contentreasons in the way they suggest, little will have been gained. Resort to the claim that a certain attitude is correct (as Brentano would say) or ﬁtting (as Ewing would say), and to the further claim that correctness, or ﬁttingness, are primitive notions, is not quite the optimal solution to the WKR problem (cf. Reisner, 2009). The ‘primitive’ way out of the problems with the FA analysis might nonetheless be quite tempting. Let me therefore ﬁnally express yet another worry about the Danielsson and Olsson approach. The notion that something is correct is problematic in that there seem to be ways of understanding it that are of limited present interest. For example, it is incorrect in some circles to clean your dinner plate by licking it, but this does not show that we are dealing with a fully-ﬂedged normative notion of correctness here. Nor does the fact that it is correct to drive on the left side of the road in England but not in Sweden express something fully prescriptive. This suggests that we need to distinguish between at least two kinds of correctness: one that is somehow conditional, and one which is not. If we can achieve this without employing or presupposing any value judgement, the Danielsson and Olsson solution will be unaffected. However, if we need to separate the fully normative notion of correctness from the conditional one by invoking value, it will turn out that we are not well served by their approach. At any rate, little can be gained if we are obliged to understand value in terms of a notion of correctness that itself needs to be illuminated by invoking value.

3.5 The biconditional buck-passing account I want to close this chapter with a discussion of another feature of the FA analysis. To introduce this idea it will be convenient to consider what Crisp recently has said are the two key elements of a buck-passing account.22 Crisp thinks we should reject what he refers to as the biconditional buck-passing account on which “whenever something has lower-order properties such that there is reason to respond to it” by favouring it in some way, “that thing is valuable”. One reason this account should be abandoned is that there are what Crisp calls deontological reasons that do not ground value. Here is one example of such a reason: suppose I sincerely utter “I promise to ç”; the fact that I have said these words gives me, according to Crisp, a reason to choose to ç.23 On a biconditional account, “because ç-ing has the property that I have Crisp, ‘Goodness and Reasons: Accentuating the Negative’ (2008). Notice that Crisp formulates the buck-passing account in terms of responses. However, I will for the sake of simplicity and discussion simply assume that choosing to ç is in fact a form of favouring ç If this is to strain the meaning of ‘favouring’, just think of an example in which my promise to ç gives me a reason to favour to ç 22

23

3.5

T H E B I C O N D I T I O N A L B U C K - PA S S I N G A C C O U N T

43

promised to do it”, it follows that ç-ing is therefore valuable. But is it? Crisp quite rightly questions this; ç-ing does not seem to be valuable. Noteworthy here is the discovery that we are stuck, apparently, with a WKR problem whether or not we actually think there is a distinction between ‘object-given’ and ‘state-given’ reasons. As Crisp himself notes, this distinction “has no purchase” when it comes to this kind of example. Crisp goes on to argue that biconditional accounts are eliminativist theories of value; and this, he thinks, is too big a step to take. He formulates what he takes to be the two key elements in a buck-passing account: Negative component: Being good is not itself a reason-providing property Positive component: Being good is merely the higher-order property of having lower-order properties that provide reasons to respond in particular ways. (2008, pp. 263–4)

Of these, he thinks it is the positive component that we need to throw overboard. We should retain the idea that goodness, or value, is not itself a reason-providing property. But is this what we should do? Is it really clear that goodness, or value, is not a reason-provider on its own? At this point one might be tempted to interject that we should give up on the negative component but that buck-passers have no choice: they have to endorse the negative component, qua Buck-passers. This is a mistake, however. Let us therefore consider the matter in more detail, and let us do so from the perspective of the FA analysis. Suppose I am told that an object x is good: someone tells me that I should give money to a certain organization; I believe the organization is good; and this is pretty much all I know about the organization. On the FA approach, what I believe is that x has the property of providing reasons to favour x. Let us focus on this very property. Could it not be argued that it is itself reason-providing? Might not the very property of being a reason-provider provide us with a reason to favour O? In other words, is it not possible that the fact that something is such that it gives me a reason to (say) desire it furnishes me with a new reason to desire it? Now, although I am not quite sure what kind of reason we are talking about here, I am nonetheless inclined to think that we are dealing with some sort of reason. But if the property of providing-reasons-to-favour indeed provides us with a reason to favour,

PR property the property of providing reasons to favour

Figure 3.1

is in itself reasonproviding

44

T H E W RO N G K I N D S O F R E A S O N

then, obviously, we should give up what Crisp refers to as the negative component of the buck-passing account: goodness (the positive component) in the FA analysis is itself a reason-providing property. In passing, we should recall that the goodness of x provides us with a reason to favour x and not (at least not necessarily) with a reason to favour the goodness of x. This distinction is important not to overlook. We might otherwise end up believing that value accrues to goodness. It does not. It accrues to the bearer of goodness. Goodness plays a role, then: it tells us that certain properties are reason-providers. To shed some light on the issue of whether goodness also is in itself reason-providing, we need next to say a bit more about what a reason is, and in particular about what a normative (i.e. justifying) reason is. Letting ‘R’ refer to a fact, we might follow John Broome and suggest the following: (R1) R is a normative reason to ç iff R explains why we ought to ç24 Suppose this is correct. To believe there is a reason for me to ç (favour something) is to believe, by implication, that there is a fact that explains why I ought to ç But now, if this fact signals that there is an explanation of why we ought to ç we appear to have a reply to the question above. Neither goodness nor value is a reason-provider. This is easily shown on Broome’s account: the fact that there is a reason to ç is not itself a reason to ç since this fact does not itself explain why we ought to ç. Hence, value and goodness are not reason-providers. (R1) needs to be qualiﬁed. Explanation is a success notion; if a fact fails to explain something it is obviously not a reason according to (R1). However, as I do not think the notion of a normative reason is a success notion, I ﬁnd it hard to understand reasons in terms of explanations. But suppose we could ﬁnd a way of dealing with this complication.25 There is still, I think, a drawback with the Broomean response. Although Broome’s notion catches something important about normative reasons, it seems to fail to capture all there is to our notion of a normative reason. There is, at least, a different notion of normative reason in play, namely: (R2) R is a reason to ç iff R bears positively on the question of whether one ought to ç Pamela Hieronymi has argued that the fundamental relation in which a consideration becomes a reason is something like that adumbrated in (R2).26 (By ‘consideration’ she

24 See Broome’s ‘Reasons’ (2004). I am here simplifying matters. Broome summarizes his view in the following way: “A reason is either a perfect reason or a pro tanto reason. A perfect reason for you to ç is a fact that explains why you ought to ç. A pro tanto reason for you to ç is a fact that plays a characteristic role in a potential or actual weighing explanation of why you ought to ç or of why you ought not to ç or of why it is not the case that you ought to ç and not the case that you ought not to ç” (p. 55). Broome treats ‘explain’ as a primitive. 25 I am not quite sure what Broome would say about this. 26 See Hieronymi, ‘The Wrong Kind of Reason’ (2005).

3.5

T H E B I C O N D I T I O N A L B U C K - PA S S I N G A C C O U N T

45

has in mind a broad notion that includes, among other things, facts and propositions; this is how I, too, would regard this notion.) By differentiating the questions, she manages to separate different kinds of reason. For instance, that it is raining bears positively on the question of whether we ought to believe that it is raining. It should be pointed out that Hieronymi wants to deal with the WKR problem in cases of value, as she deals with reasons for belief. The fact that you would get beaten by the demon unless you believe that London is the capital of France obviously does not bear on the question of whether London is the capital of France. Hieronymi, then, goes on to argue that we can reason in an analogous way when it comes to cases of value. But, from the perspective of the FA analysis, this conclusion does not follow. The fact that the demon wants to hurt me unless, say, I favour a saucer of mud, clearly bears on the question of whether we have a reason to favour the saucer of mud (i.e. whether it is valuable). Now, depending on what kind of reason we have in mind (R1 or R2), the question ‘Is the buck-passer’s goodness reason-providing?’ will receive rather different answers. As we saw earlier, on Broome’s account the answer would be a clear ‘no’. On Hieromymi’s account, by contrast, the answer is ‘yes’: the fact that there is a reason to ç is itself a reason to ç, since this fact bears on the question of whether I ought to ç Thus, if you are a buck-passer who does not want to give up on the idea that goodness is a reason-provider, endorsing (R2) rather than (R1) will be of great help; (R2) explains why we should, as FA analysts, throw the negative component of the buck-passing account overboard. Embracing (R1) (with the proviso mentioned) on the other hand, paves the way for the negative component. FA analysts have, in other words, a choice: they may or may not throw the negative component of the buckpassing account overboard.27

27 Cf. also Mark Schroeder’s recent ‘Buck-Passers’ Negative Thesis’ (2009). Schroeder contests what he refers to as the “buck-passing inference” (p. 342)—that the negative thesis follows from the positive one.

4 Mistaken Value Analyses

The general aim of this work is to develop a novel way of understanding values that we regard as personal, i.e. as being in some way good or bad for us. Sometimes what has personal value will also have an impersonal value. This fact complicates matters in more ways than I care to think about. What if these values conﬂict with each other? How do we compare them? Which should we give priority to? These and many other questions are all important; but although I will have more to say about some of the issues later on, I have had to leave out the general question about the relation between personal values and other values. My own welfare and well-being are examples of something having both kinds of value.1 Certain of my mental states are good for me, but such states, regardless of whether they have value-for me, might also carry impersonal value. The link between good-for and welfare or well-being is tight, though. In fact, numerous writers treat this connection as analytic. This is not obviously the best way to regard the relation between these notions. Depending on what one takes it to be, well-being might be just one of the things that exempliﬁes value-for. Other ‘things’ might carry value-for as well. In what follows I will examine various possibilities of this sort. I will seldom discuss well-being. This does not reﬂect my view of its importance. Most examples I discuss will, most likely, not be especially important values—not by comparison with the value of well-being at any rate. But the less important values need to be understood if we want to obtain a complete picture of value. Moreover, in my experience it is often easier to discuss cases that appear to be rather trivial. It is easier in the sense that one can focus on what is really relevant. Here is an example of a putative personal value which I am inclined to say carries no impersonal value. My father made a bookshelf for me when I was a child. It was, perhaps, not an aesthetic revelation—more a piece of robust and heavy furniture that served its function well. As an adult I took it with me as I moved from place to place 1 I do not take well-being and welfare to be the same. However, in what follows I will not pay any attention to these differences, since they are not of crucial importance to what is discussed.

4.1

M I S TA K E N VA L U E A N A LY S E S

47

(a bookshelf always comes in handy). Each time it looked more and more shabby. At one point, it was in such a bad state that I decided to get rid of it, and ﬁnally I got around to throwing it away. Or so I thought. Now, many years later, bits of it still pop up in the most unexpected places. I ﬁnd this quite amusing, and I have begun to wonder whether, in my own life, I shall ever see the back of this item of furniture. These screws, shells, wood pieces, ornaments, and so on, hardly have impersonal value, but they do seem to carry some sort of value for me. However, it would be a gross simpliﬁcation to conclude that personal values, and in particular those that are not also of impersonal value, must all be found in objects relating to one’s past. I mention these examples since they appear to me as more obvious. All examples of personal value have this much in common: they are receptive to the FA pattern of value analysis. What we need to do is ﬁne-tune the relevant analysans so that it brings into the analysis a reference to the person whose personal value it is. But here a caveat is in place. This requirement on the analysis (that it should somehow bring in the person whose value it is) is ambiguous. It might refer to what is perhaps best described as an owner claim relation between the value and the person whose value it is. Like G. E. Moore, to whom I will return in a moment, I have a problem with this kind of view. The suggestion I endorse sets out from another idea, namely that personal values are just like other kinds of value but for the fact that they stand in a certain relation to some person (or perhaps group of persons). But this relation has nothing to do with the value being ‘his’ or ‘her’ private value. Here, then, is how I want to modify the standard analysis, where FAP stands for the Fitting Attitude analysis of Personal value: FAP: An object x’s value for a person a (i.e. x’s personal value), consists in the existence of normative reasons for favouring/disfavouring x for a’s sake FAP is not as original as I once thought when I proposed it in a series of talks in 2004; there are points of contact between it and the view Stephen Darwall takes on welfare. There are also some important differences, though; but more about this later, in Chapter 8. Also, as we shall see in Chapter 7, there is a chance that Henry Sidgwick understood good-for along the lines of FAP. The analysis is stated in terms of reasons and pro- and contra-attitudes; as may be recalled, “favour” is a notational alternative to “having a pro-attitude”. In the chapter that follows I will say something more about how I understand the reason component of FAP. The notion of a pro-attitude (and with the proper adjustments, a contraattitude) is here understood to cover quite a large set of different pro-responses that are called for by the value bearer. I will not try to specify what makes a certain mental state into a pro-attitude. I trust that this notion is sufﬁciently clear to render FAP intelligible. Some cases will clearly fall outside FAP. To feel thirsty or hungry would not, in my view, count as having an attitude. Other cases will be more difﬁcult. What should we say, for instance, about a speciﬁc kind of thirst—say, feeling thirsty for orange juice?

48

M I S TA K E N VA L U E A N A LY S E S

There are all sorts of problems involved here, but I will not pursue them.2 (FAP will eventually be modiﬁed somewhat in Chapter 5, where I seek to say something more detailed about one feature of the attitudinal element.) As to the variable x, it should be understood as gathering a wide variety of things, including living beings (persons and animals that carry the personal value to x). One advantage with this pattern is its wide range: very many different kinds of value are analysable in terms of the same general formula. But perhaps more important, by situating the distinguishing quality in the attitude, rather than reason, part, the analysis allows that personal value is recognizable as a value not only by the person for whom it has personal value but everyone else too. In this way we avoid introducing two completely different notions of value, one impersonal and the other personal. Of course, it is not immediately obvious why not having two completely different notions of value is a good thing. Why, for instance, should we not locate the distinguishing mark in the much-discussed agent-relative/neutral dichotomy? Just as personal values are well suited to be understood in terms of agent-relative reasons, so impersonal ones appear connected to agent-neutral ones. One can imagine that this might indeed come in handy. However, as will emerge in Chapter 9, there are powerful reasons to be wary of such a move. However, for the time being I will not discuss in any more detail just how we should understand the normative component of the analysis. Moore saw that there is something intuitively odd about the idea that value is relative. His scepticism is shared by many philosophers, though. A more recent opponent of good-for is Donald H. Regan, whose views I will return to in Chapter 7. Regan expresses well what is strange about this notion: The good for Abel must be peculiarly Abel’s—the goodness or the value must be peculiarly Abel’s—in a way that the mere occurrence of universal good in Abel’s life does not necessarily satisfy. But if the good for Abel is peculiarly Abel’s—if its value is somehow essentially a value for Abel—then why indeed should Cain care? We think there is a deep connection between value and reasons. That suggests precisely that if the good for Abel is a matter only of value for Abel, then it creates reasons for Abel, and for no one else.3

However, notice that even if you think that value-for is a coherent notion which does not depend on value period, it does not follow that it must be analysed in terms of agent-relative reason. Connie Rosati has argued in ‘Objectivism and Relational Good’ (2008) that these reasons may be agent-neutral. She suggests, contrary to Regan, that

2 See Thomas Scanlon (1998), where the issues are delineated in the following way: “The class of attitudes for which reasons . . . can sensibly be asked for or offered can be characterized, with apparent but I think innocent circularity, as the class of ‘judgment-sensitive attitudes.’ These are attitudes that an ideally rational person would come to have whenever that person judged there to be sufﬁcient reasons for them and that would, in an ideally rational person, ‘extinguish’ when that person judged them not to be supported by reasons of the appropriate kind” (p. 20). 3 See Regan, ‘Why am I My Brother’s Keeper?’ (2004, p. 211).

4.1

MOORE’S OBJECTION TO GOOD-FOR

49

the notion good-for does not involve a relativization of normativity or goodness. When I started writing on personal values in 2004 (later published as ‘Analysing Personal Values’ (2007a)), I had the same thoughts as Rosati. The way to avoid the problems that Moore pointed to was to understand these values in terms of agent-neutral reasons. I now believe that the real distinction between good and good-for is not to be found in the reason element, but rather in the kind of attitude that is called for. More about this in the coming chapters, though. Moore’s sound scepticism about ‘private’ values led him astray, to the mistaken conclusion that we should give up the notions of good-for and bad-for altogether (or at the very least be aware that these expressions are elliptical). His arguments are not convincing, as I shall argue in the next section. We need not give up these notions. We just have to ﬁnd a more appealing account of good/bad-for than the one Moore objected to.

4.1 Moore’s objection to good-for In his much-discussed attack on egoism in Principia Ethica,4 Moore expresses great scepticism about certain ways of understanding the expressions ‘my own good’ and ‘good for me’:5 What then is meant by ‘my own good’? In what sense can a thing be good for me? It is obvious, if we reﬂect, that the only thing which can belong to me, which can be mine, is something which is good, and not the fact that it is good.6

The passage is bewildering (perhaps especially as it comes from Moore). First, Moore is, of course, right that the fact that something is good is not something that can belong to me. Facts are not the kind of object that can be in someone’s possession. But this does not force us to say that ‘my own’ in ‘my own good’ cannot refer to anything but the object that is valuable and hence cannot refer to the value itself. It is quite surprising that Moore in the ﬁrst place takes our relation to facts to be an issue. It is hard to believe that any of Moore’s opponents would actually disagree with Moore on this matter.7 It is important to underline the fact that Moore cannot simply assume there are no facts about ‘my own good’. As yet he has given us no reason to suppose that there are 4 For a classic and quite devastating objection to Moore’s attempt to show that the doctrine of egoism is self-contradictory, see C. D. Broad’s ‘Moore’s Ethical Doctrines’ (1942). 5 As Rosati (2008) has recently put it: “Moore appears to have found the expression ‘good for me’ akin to the peculiar expression ‘true for me’” (p. 324). Rosati’s interesting paper contains detailed discussion of not only Moore’s, but also some modern objections to the notion good-for. Her idea that critics of this notion “need not be understood as advocating that we purge ordinary discourse of good for talk (as if we could do that), or even (more realistically) that we theorists refrain from such talk” is, I believe, correct. Of course, it remains possible that Moore was trying to make us stop using the expression ‘good for’. 6 Moore, Principia Ethica (1993 [1903a], p. 150). 7 It may be that, in concluding the passage quoted with the words “and not the fact that it is good” Moore intended no more than to assert that goodness is not something that can belong to a person. However, whether or not this is true, it does not change our assessment of what Moore accomplishes argumentatively.

50

M I S TA K E N VA L U E A N A LY S E S

not two different kinds of value fact: one, that something is good, period (for everyone), and another, that something is good for person a. The states of affairs involved in these two cases may well, when they obtain, constitute two different kinds of fact. Here is Moore’s own conclusion as to what we can mean by ‘my own good’: When I talk of a thing as ‘my own good’ all that I can mean is that something which will be exclusively mine . . . is also good absolutely; or rather that my possession of it is good absolutely. The good of it can in no possible sense be ‘private’ or belong to me; any more than a thing can exist privately.8

Again, if it is not merely that it is mine or that it exists that is valuable, but the thing itself, then Moore has not shown that the value of such an entity cannot, in an interesting sense, be private. If there is something like a peculiar ‘good-for fact’, then it will be a fact for everyone, since facts are not the kind of thing that can be in someone’s possession. But this should not prevent us from saying that the fact contains a ‘private’ element that makes it distinguishable from facts involving good, period. Moore’s objection to understanding ‘my own good’ as meaning anything but that which is exclusively mine carries an absolute good must be understood in the light of his criticism of egoism—a view he believes he can refute. Moore is not easy to follow here, but a core idea concerning his objection to ‘my own good’ seems to be a claim to the following effect: (1) To be a good at all, the good must be (a) good in itself, i.e. an intrinsic form of goodness, and (b) a ‘universal good’ (cf. pp. 150–1)9 Moore’s famous attack on egoism and the idea of something being ‘good for me’, which runs over several paragraphs, boils down to this very idea that to be a good at all the putative good must be a universal intrinsic good. However, he does not support the idea—at least, not in a persuasive way. Therefore, the attack is in effect not so much an argument as an assumption that Moore is asking us to join him in making, or work with. He has not given us any convincing reason why the notion good-for should be rejected. As mentioned, Moore is not alone in ﬁnding the notion of good-for objectionable. Thomas Hurka (1987) argues: As currently used, ‘good for’ is multiply ambiguous, and for none of its meanings is ‘good for’ the appropriate term. (p. 72)

For this reason Hurka thinks it should be “banished from moral philosophy” (p. 72). The senses he lists are: “‘good for’ means satisﬁes the desires of, or further the interests

8

Moore, Principia Ethica, p. 150. The notion of a ‘universal good’ is left undeﬁned. A natural interpretation (especially since he is discussing Sidgwick’s view on the matter) is that Moore has in mind everyone’s good. 9

4.2

I N S T RU M E N TA L VA L U E

51

of”; it may also mean “good from the point of view of ”; what is ‘good for’ a person may be what “she believes good”; and ﬁnally it can “mean that portion of the good (from whatever point of view) that falls within a person’s own life” (pp. 72–3). Hurka has misgivings about all of these suggestions. I share these, and do so for the reasons he gives. However, as I hope to make plausible by developing an analysis of the notion, this does not mean that ‘good for’ cannot be given, or associated with, a single sense of its own.10

4.2 Instrumental value FA analyses of ﬁnal impersonal value say that an object has ﬁnal value if there is reason to favour it for its own sake. Applied to ﬁnal personal values the FA approach adds that the attitude must be of a certain kind: it must be a favouring of x for a person’s sake. (This kind of attitude has not received much attention in the literature, so we shall need to examine it below in Chapter 5.) An object of impersonal value is then said to be instrumentally valuable if there is reason to favour it for the sake of its effects. In the case of instrumental personal value we have, on the analysis, a reason to favour, for a person’s sake, something for its effects’ sake. Here it is important to pause for a moment, though. For this analysans is, of course, only plausible to begin with if instrumental value is a genuine kind of value in the ﬁrst place. However, this is not obvious. Standard suggestions—such as ‘x has instrumental value’ means x is conducive to something that has ﬁnal value—are not very helpful if the aim is to ﬁnd out just what the speaker means by ‘instrumental value’. Such suggestions tend to leave us in the dark with regard to whether x is a bearer of something that belongs to the category of value or whether x is merely related in some way to something belonging to this category. In other words, to say of an x that it is an instrumental y does not yet determine whether or not x is a kind of y. What happens here is something that seems to occur with many expressions we employ to qualify something. Just as ‘quicksilver’ does not refer to a kind of silver, ‘instrumental value’ may not refer to a kind of value. In the literature it is possible to detect at least two distinct usages of ‘instrumental value’. To begin with, there is what I will refer to as the Strong evaluative sense: S: ‘x has instrumental value’ means ‘x bears a (certain particular) value, and it does so only if x is conducive to (the existence of something with) ﬁnal value.’

10 Regan (2004) suggests that instead of speaking about good-for, we would do better to employ “goodness occurring in a person’s life”. Rosati has convincingly shown, I think, that this notion of a value occurring in a person’s life faces a number of difﬁculties of its own. More importantly, she shows, too, that what is “picked up” by good for is not covered by the notion of “goodness occurring in the life of ”. I will return to Regan’s views in Chapter 7.

52

M I S TA K E N VA L U E A N A LY S E S

The second conjunct here is open to more than one interpretation; but since I have considered these elsewhere, I will keep things simple and side-step the complications that ensue.11 For instance, George Dickie’s way of expressing the traditional distinction between intrinsic and instrumental goodness suggests that he takes the latter to be a kind of value: ‘‘Philosophers distinguish between intrinsic goodness, which is good all by itself (independent of its relation to anything else), and instrumental goodness, which is the goodness that something has because it is a means to something else which is good” (1979, p. 157; my emphasis). Here Dickie appears to assume that there are at least two kinds of goodness: intrinsic and instrumental.12 The point I wish to emphasize is that the value described in S should not be confused with instrumental value in the Weak evaluative sense: W: ‘x has instrumental value’ means ‘x is conducive to (the existence of something with) ﬁnal value’.13 In contrast with their stronger cousins, weak instrumental values do not belong to the category of value. But, it might be asked, is not that which leads to something of ﬁnal value a good example of something with a kind of value—namely, the kind of value accruing to objects that lead to what is of ﬁnal value? The suggestion implied here seems to set out from the idea that ‘being valuable’ is somehow part of the meaning of ‘being conducive to value’. But that is surely a mistake. Notice that the deﬁniens says, merely, that x is conducive to what is of value. There is no reason why ‘being valuable’ should somehow logically follow from the meaning of ‘being conducive to value’. To conclude that what is conducive to value must therefore be of value is to commit what we might dub the value-by-association mistake. Since the expression ‘instrumental value’ is equivocal and does not, in itself, clearly exclude either S or W, we will have to turn to metaethical theories of what constitutes value to settle what an instrumental value is. (That is to say, we need here something more than value’s supervenience base; we need some kind of account of how it is that the subjacent properties ground or constitute the value.) Meta-theories of value might imply, of course, that there is only one way of understanding ‘instrumental value’. In that case, we should expect a reason for this exclusion. For my own part, I believe that there are S-values. Moreover, I think that FA analysis gives us a good account of

11

Rønnow-Rasmussen (2002b). See also R. T. Allen, The Structure of Value (1993, p. 59), in which the author seems to assert that the instrumental value of objects is ‘realized’ when we use these objects. Cf. Robert Nozick’s claim “[ . . . ] there is something’s originative value which is a function of the value it newly introduces into the world, the new instrumental or intrinsic value it introduces that was not presaged by or already fully counted in previous instrumental value’’ (1981, p. 311). 13 Johan Brännmark has pointed out that there might be a structurally similar but nonetheless different sense of instrumental value—one relating the instrument to a goal rather than to a value (personal communication). I think he is probably right, but I will not discuss the suggestion further. 12

4.3

A B A D A P P RO A C H T O P E R S O N A L VA L U E

53

them as objects that we have a reason to favour for their effects’ sake.14 The important thing is to be aware of the two senses, S and W. I will not argue for this contention. Doing so would lead me astray. The point I wish to make clear, however, is that the analysis of instrumental value mentioned earlier is an analysis of an evaluative notion of instrumental value.

4.3 A bad approach to personal value I now want to open a discussion of the two key elements of FAP—the attitude and normative parts—by considering an approach which, in my view, is mistaken. This approach quite naturally suggests itself. It sets out from the notion of extrinsic ﬁnal values. What it suggests is that the road to an understanding of personal value is clear once we accept the idea of extrinsic ﬁnal value. Thus, according to what I will refer to as the De Facto Attitude analysis of Personal Value (or DEFA), personal values make up a subgroup of extrinsic values. As I say, DEFA is not a good approach to personal value. However, it might prove useful to glance at this erroneous alternative if that helps us to appreciate the merits of the more serious suggestion presented in the next section. DEFA is arrived at by the following path. If ﬁnal values can be extrinsic—i.e. accrue (at least, in part) to objects in virtue of their non-internal relational properties—we can expect some object of extrinsic value to carry this kind of value in virtue of its relation to a particular person. Once value is relativized to persons in this way, the stage is set for an analysis of personal values: we need, next, to specify the relata to which personal value accrues more carefully. This must be done in order to distinguish personal value from other kinds of extrinsic value accruing to objects merely in virtue of being related to a person. Intuitively, it seems reasonable to assume that a personal value is something more than a value that is related to a person. We need, that is, to analytically separate from each other cases like the following: Napoleon’s hat and a drawing made by my ﬁve-year-old daughter. It can certainly be argued that Napoleon’s hat has some kind of ﬁnal, impersonal extrinsic value.15 Even if Napoleon never endorsed any evaluative judgement whatsoever about the hat, it might still be of value because it belonged to, and is famously associated with, a historically important person. My daughter’s drawing, on the other hand, is unlikely to have impersonal value. It appears rather to be a carrier of personal value, if it has value at all. Suppose we accept these examples in the terms they are described. The question then is: How should we account for the difference? What makes the ﬁrst example a case of impersonal value and the other a case of personal value? 14 This will particularly be the case if we take into account the fact that “favour” refers not only to various different attitudes but also to attitudes-cum-behaviour and perhaps even to choices. That the latter possibility should be taken into account was pointed out to me by Brännmark (personal communication). 15 Cf. Shelly Kagan, ‘Rethinking Intrinsic Value’ (1998) and Rabinowicz and Rønnow-Rasmussen (2000).

54

M I S TA K E N VA L U E A N A LY S E S

The role that Napoleon played in the past, and my, and my daughter’s, historical insigniﬁcance, may explain why the hat has impersonal value while the drawing does not. But these facts do not account for the fact that the drawing, and not the hat, has personal value. Here we need to look for something else. A natural strategy would be to search for something special about the two salient objects involved in the drawing example, i.e. the drawing x and the person a for whom, ex hypothesi, the drawing has personal value. On the face of it, it is likely that a as a matter of fact favours the drawing. DEFA could, but need not, be cast in some kind of ﬁtting analysis mould. That is to say, it could say something to the following effect: x has personal value for a if and only if we ought to favour x in virtue of the extrinsic property that x is de facto favoured by a. Thus, my daughter’s drawing would have personal value only if the value accrued to it in virtue of being an object that is favoured (say, cherished or esteemed) by me, a. Whatever intuitive appeal it has, however, DEFA establishes at most that the value accruing to the drawing is relative; and that is not speciﬁc enough if it is personal value we are looking for. It might be objected that DEFA need only make an alternative, and quite reasonable, subjectivist claim, i.e. that we should regard the attitude of a as having value-constitutive powers. This might well be true. However, while the issue between subjectivism and objectivism is not settled in a convincing way, I will allow, as might be recalled, the analysis to reﬂect this by formulating a neutral analysans. Personal value is in a sense relative: since it is value-for, we need to mention, in analysing it, the person for whom it is valuable—but relative values are certainly not necessarily personal values. In other words, although the inference from “x is a personal value” to “x is a relative value” is valid, the reverse inference is invalid. There are two more reasons why we should regard DEFA with scepticism. First, it places personal values exclusively among extrinsic (ﬁnal) values. But this seems to be an arbitrary restriction. An account including intrinsic values would be more appealing. Of course, it might be that all personal values are in fact extrinsic ﬁnal values. A substantive evaluative argument might force us to take such a view. But the issue now is whether we could coherently think that some object x carries value-for in virtue of x’s internal properties and nothing else. I am inclined to think that examples of phenomena of this sort will be fairly controversial. (I will discuss some later on.) But in principle I do not see why the relation picked out by the expression ‘good-for x’ cannot supervene on the internal features of an object alone.16 Second, it is not obviously a necessary condition of something being a personal value that it is valued by the person for whom it is a value. Surely a convincing case could be made, albeit on substantive grounds, for the idea that there can be objects of personal value that are not valued (esteemed, cherished, and so on) by the relevant person.

16

I will return to certain problems with this idea in Chapter 6.

5 For Someone’s Sake

If one wants to understand values in terms of reasons, there seems to be a perfect dichotomy to use—namely, the dichotomy of agent-neutral and agent-relative reasons. After all, value-for is a relative notion, and what would be more logical than its connection with agent-relative reasons? However, as I shall argue in Chapter 9, there are good reasons to be cautious about this distinction; we should therefore be equally cautious about understanding the distinction between personal and impersonal values in terms of the dichotomy between agent-relative and agent-neutral reasons. But whether or not my scepticism concerning the reason dichotomy is sound, the sort of analysis that I am proposing stresses that the deﬁning feature of personal value is not to be found exclusively in its normative element, but is rather located in the analysans’s attitudinal element. Objects of personal value are such that there is reason to direct a certain type of attitude—i.e. what we might call a ‘for-someone’s-sake’ attitude—on them. I believe that these reasons are all agent-relative. Be that as it may, what is typical about objects that are good for, or bad for, persons is that a for-someone’s-sake attitude is called for. This is what distinguishes such values from impersonal ones.1 In this and the next chapter I attempt to add ﬂesh to the bones of these forsomeone’s-sake attitudes (henceforth FSS attitudes). They have not yet been properly analysed in the philosophical literature. How we should delineate attitudes in general, and how, in particular, we should isolate FSS attitudes in what follows from, on the one hand, pure non-propositional cravings such as hunger and thirst, and, on the other, purely cognitive beliefs, are matters I will bypass here. As might be recalled, what I have in mind by ‘attitudes’ are not only desires and preferences, but also ‘thicker’ attitudes such as admiration, respect, and love, to mention only a few examples. Moreover, I will assume that favouring (i.e.

1 Given the great variety of values, and given the ‘monotony’, to borrow Kevin Mulligan’s description (personal communication) of the normative, if we are to understand the former in terms of the latter, we had better come up with something more than a reduction of the valuable to the normative; this is precisely what my account purports to do here.

56

FOR SOMEONE’S SAKE

wanting, admiring, and so on) x is an attitude that differs from favouring y, and that they differ because their intentional contents are different: the former contains a propositional element involving x; the latter, one involving y. I am, then, setting out with the tentative assumption that there might be a special kind of attitude, namely the FFS attitude, which, intuitively, is the kind of attitude we have when we favour x with an eye to someone (else); and that this too is reﬂected in the intentional content of the attitude, which we describe as favouring x for someone’s sake. Here a caveat is in place. The idiom ‘to do something with an eye to’ is somewhat unclear. Often it means that we are doing something with regard to, in view of, or for the purpose of something. My way of using the expression here and throughout this book is not quite idiomatic, though. As will emerge later on, since I wish to stress that it is the person rather than the effects on or the implications of my favouring for that person that we have in mind when we ascribe personal value to something, I will be as bold as to strain the English and say that we can favour an object with an eye to a person. Unfortunately, there is no simple way to characterize ‘favour x for-a’s-sake’-attitudes. In fact, my own view is that the prospects of specifying the characteristic features of this sort of attitude in a fully satisfactory way are bleak. The fact that they seem resistant to analysis might suggest that they are not genuine attitudes. Perhaps this is so. However, since analysis of these attitudes has been largely ignored by philosophers, to brush them aside at this stage would be premature. Of course, although they remain something of a puzzle, this does not mean that we cannot say anything about them. FFS attitudes are discerning attitudes: like most attitudes, they are directed on objects on account of a particular property, or some properties, of the objects on which they are directed. However, as I shall be arguing, not all attitudes are discerning in the same way. Some discerning FSS attitudes are, as I shall put it, ‘identity-involving attitudes’ (for the sake of brevity I shall refer to these as Identity attitudes), while others are non-identityinvolving.2 To support this categorization, I will later consider two contrasting sorts of discerning attitude, namely love and admiration. Love, I shall claim, falls in the ﬁrst category; admiration in the second. Different conceptual barriers constrain the kinds of property that can be invoked to justify love and admiration. The kinds of property that typically justify love (on certain conceptions of it, at any rate) are not the kinds of property that justify admiration. In fact there is a good reason to believe that the properties justifying love could not justify admiration. In Rønnow-Rasmussen (2008b), I introduced the distinction between justiﬁers and identiﬁers, which refers to two kinds of property differing from each other in the role they play in the intentional content of the attitude (justiﬁers justify, and identiﬁers identify). I argued that when it comes to admiration, properties of the admired object play the role of justiﬁers rather than identiﬁers, but that in the case of love, properties of the beloved are identiﬁers. This distinction needs to be further clariﬁed, to be sure.

2

I owe thanks to Christian Piller for these expressions.

5.1

T Y P E S O F AT T I T U D E

57

However, once that is done, the distinction will help us to separate two kinds of FSS attitude, both of which are pertinent to the analysans of personal value. It will take some unravelling, in this and part of the next chapter, to make this identiﬁer/justiﬁer distinction comprehensible; and in the course of this unravelling I shall need to touch upon certain issues that have been much debated by contemporary philosophers. Thus I will say something about the intentional content of the sort of attitude that I will be discussing, and as I do so I will run into various problems concerning the sense and function of proper names in an attitude’s propositional content. Since I do not know how to settle these issues in a satisfying way, and since I am not familiar with any theory that provides all the right answers to the intricate problems that arise in this area, I have two options: to omit speculation in this area, or to enter into it regardless of the fact that it is bound to involve claims that are problematic on certain views on reference. With some qualms I have decided to take the latter option. Moreover, I have taken the decision to keep the discussion as simple as possible, and to ignore a number of views that have a bearing on much that I say even if they would not have settled the issues. Of course, I would very much have wanted a general theory about how to understand the propositional content of attitudes to back up some of the points I make in this chapter, but I do not have such a theory. Had I one, I would probably have written a book about this kind of theory in the ﬁrst place rather than one about personal values. In this and the next chapter I will also say something about the importance of people’s identity. From experience, I know there is a chance that some of my claims will suggest that I take a ﬁrm stance against certain views on what matters when it comes to ‘identity over time’, i.e. diachronic identity. But, as I hope will become clear in the next chapter, what I say here is, in my view, consistent with a number of positions regarding who we ought to care for in different possible scenarios.

5.1 Types of attitude Let us begin to ﬁll out the attitude required in the analysis in a more substantive way. To begin with, consider the following two attitudes, where a, as before, refers to some person, and x to some object (understood in a wide sense): (1) a favours x for its own sake (2) a favours x for some other object’s sake Here (1) describes what I shall call a ﬁnal attitude—the attitude I have, for instance, when I desire pleasure for its own sake. If we take the attitude in (2) to be about x’s effects on the object, then Claim (2) describes an instrumental attitude. I have this kind of attitude when I desire to drink a glass of water for the effects of drinking; drinking gives rise to, say, pleasure. Recall next the content of FAP: an object x’s value for a person b consists in the existence of normative reasons for favouring x for b’s sake.

58

FOR SOMEONE’S SAKE

What should we say about the kind of attitude referred to in FAP? (3) a favours x for b’s sake Although it may be hard to specify the details of this attitude, it is quite easy to think of examples in which it ﬁgures. Here is one: one treats a stranger amicably for a friend’s sake—the stranger being the friend of a friend, say. Interestingly, emotions also seem capable of this sort of construction. For instance, on learning that someone for whom we care has had good fortune, we are glad for his or her sake. Here is another example: I sometimes want to go to the theatre for my wife’s sake. She enjoys a good play, and although I frequently fail to see what is so good about these plays, I want nonetheless to go. Not for my sake, but for hers.3 To avoid yet another possible misunderstanding, it should be clear that FFS attitudes, and attitudes-cum-actions (and perhaps emotions) need not involve other people and frequently do not do so. I take it to be uncontroversial and familiar to most of us that we often favour things for our own sake. For instance, I may desire to know for my own sake whether or not I have a fatal illness, and I may have this desire whether or not I have the further desire to know this for my family’s sake. It is not especially hard to ﬁnd other examples. Here is a different example which suggests that we can believe that something has personal value without us knowing whose values they are. As a tourist I have visited a number of churches and I have often overheard comments from other tourists that one should not behave in certain ways for the sake of the religious believer. People are supposed to treat, say, the statues of Madonna as special objects for the sake of people attending worship. Two kinds of FFS attitude in particular are relevant to my analysis of personal values. One is needed to understand ﬁnal personal values, and the other to understand instrumental personal values—as is the case with impersonal values, personal values can be ﬁnal (as when something is ﬁnally good4 for us) or instrumental (as when something is good for us as a means). There may be other sorts of personal value too, such as personal contributory value. However, I will mostly be discussing ﬁnal personal value, which is the most interesting sort. I propose to explain ﬁnal personal value in terms of what I shall call ﬁnal FSS attitudes (FFSS attitudes for short): (4) a favours x for its own sake for b’s sake

3 So what is carrying the personal value in this particular case? The play? Perhaps, but it would be somewhat strained to say the play gives me a reason to want to accompany her to the theatre. More reasonably, what is valuable for her is rather seeing a good play together with someone whose company she enjoys. This is more likely to give me a reason to accompany her. I will address objections to the analysis in Chapter 8. 4 ‘Finally good’ should here and in what follows be understood as referring to what is good for its own sake, and not as what is good, at last, or something to that effect.

5.1

T Y P E S O F AT T I T U D E

59

In (4), x ranges over abstract as well as concrete objects; and, among the concrete objects, x may refer to things as well as to people. Things as well as persons have value for us. The idea I am advancing here, then, is that one way of understanding the attitude in (3) is to read it as a special case of (4). Of course, the very idea that we can favour something for its own sake for someone’s sake is counter-intuitive and linguistically awkward. Let us therefore consider in more detail what this sort of qualiﬁcation might mean. Admittedly, the inclusion of two for-the-sake-of qualiﬁcations appears at ﬁrst sight confusing. We might therefore be tempted to understand such descriptions as referring not to a single kind of attitude but rather to a complex consisting of at least two attitudes. This suggestion should not be rejected out of hand. There are, then, at least two salient ways to split an alleged FFSS attitude (such as the one we think is present when a favours x for its own sake for y’s sake) into more basic attitudes. The ﬁrst is to treat the example as one involving these attitudes: A1: a favours x for its own sake A2: a favours x for b’s sake This is a likely scenario. That is, a person we describe as having an FFSS attitude can, quite naturally, be taken to have attitudes like those in A1 and A2. Certainly, if a favours x for its own sake for b’s sake, then it is true that a favours x for b’s sake in some sense not yet speciﬁed and it is true that a favours x for its own sake in some sense not yet speciﬁed. However, from this we cannot simply conclude that an FFSS attitude is merely the conjunction of the attitudes in A1 and A2. This becomes evident once we realize that these attitudes are less speciﬁc than the FFSS attitude. As we shall see in a moment, they can be made more precise in two principal ways. Meanwhile, consider the following second suggestion as to how FFSS attitudes can be regarded as multipart attitudes. In connection with the example above, the idea is that we should take the FFSS attitude to be a combination of these attitudes: A1: a favours x for its own sake A3: a favours A1 (i.e. a favours favouring x for its own sake) for b’s sake Notice that A3 is a normal second-order attitude. There are at least two problems with this deployment of A1 and A3. I shall conﬁne myself to the more obvious of these.5 A1 and A3 cannot be the whole story behind FFSS. (The points I shall now make apply mutatis mutandis to the combined deployment of A1 and A2.) This becomes clear once we realize that A3 can be made more precise in two main ways: we can favour A1 either as a means (i.e. we can favour it for

5 The second problem (whose details would require considerable unravelling) has to do with the fact that if the only thing we favour when we are having an FSS attitude is another attitude, it would appear to follow, given the pattern of value analysis employed in this work, that the only kind of object that could have personal value would be other attitudes.

60

FOR SOMEONE’S SAKE

the effects it has on something else) or non-instrumentally (i.e. for its own sake). Suppose it is the latter we have in mind. In that case we are back where we started, namely with an FFSS attitude. For the attitude described in A3 has the same form as the typical FFSS attitude: a favours some object x (e.g. the attitude referred to in A1) for its own sake for b’s sake.

5.2 Sake, end, and objective FFSS attitudes seem to involve an ability to harbour attitudes for someone else’s sake—an ability, by the way, that appears to be an important part of empathy. To recognize that something is of personal value to a, on this line of reasoning, is in some way to echo, for a reason, a favouring that a has or ought to have vis-a`-vis the object of value for a’s sake. One might object that if we take this route we shall be left stranded with two notions of personal value: one employable only by the person for whom something is a personal value and another for those who recognize that something carries this person’s personal value. We would have to face, in other words, a ﬁrst-person value-for, and a second- and third-person value-for. However, this need not follow. The attitude that I am ready to take to, say, the stranger who is a friend of a friend of mine has something to do with my friend’s attitudes to the stranger. But it is hardly my friend’s de facto favouring that I will ‘take over’. Rather it is the kind of attitude that a has reason to take in those circumstances—a reason that has met with my approval. Given this, I cannot see why a’s claim that something has personal value should not be analysed in the same way. These speculations apart, the following response—with the focus on ﬁnal rather than instrumental pro-responses—can be anticipated: make up your mind! Either you favour x for its own sake or you favour it for your own sake. You cannot have it both ways. Actually, I am convinced that we can have our cake and eat it here. At any rate, I have failed to identify a single, precise content of ‘a’s sake’ that cannot be combined with ‘favouring x for its own sake’. This seems to suggest that the word ‘sake’ is ambiguous: it does not necessarily have the same content in sentences containing ‘for its own sake’ and those containing ‘for a’s sake’. Two things in particular complicate the evaluation of these FFSS attitudes. First, we should keep in mind that ‘favours’ is used here as a technical term that covers a wide range of attitudes. So what is possible with regard to one kind of attitude may be impossible with regard to some other kind of response. However, I have not found a clear example of an attitude that could not take the form of ‘a favours x for its own sake for b’s sake’ (or the instrumental attitude ‘a favours x as a means for b’s sake’). Of course, when we look at more speciﬁc attitudes various issues will arise. Still, since I obviously will not be examining all of the different attitudes covered by ‘favour’, we should watch for questionable generalizations on this matter. There might well be attitudes that cannot be transformed into the ‘for someone’s sake’ pattern. Evaluation of the FFSS attitude is problematic, secondly, simply because it involves awkward, unnatural-sounding descriptions. In general, it is surely a good policy to be

5.2

S A K E , E N D, A N D O B J E C T I V E

61

suspicious about descriptions of this sort. But, of course, the fact that we need an unnatural-sounding formulation to capture a complex mental state is in itself no reason to dismiss the associated attitudes on logical grounds. (Moreover, in this particular case the FFSS description is actually an attempt to specify an attitude rather than to give a more general description.) I am inclined to think that the perplexity created by the two for-the-sake-of qualiﬁcations derives to a great extent from a particular way of reading ‘sake’. Claims about a’s sake and x’s sake might suggest that I am ascribing, at least to the former, a sake or an end, in the sense of an objective (cf. the German ‘Sache’; see Rabinowicz and Rønnow-Rasmussen, 2000). However, generally, an analysis of attitudes does not essentially require (as a description of a psychological proﬁle might do) reference to an objective in the analysans. This needs to be kept in mind. Moreover, FFSS attitudes may be directed on a large number of different kinds of thing, only some of which can be understood, in a reasonable sense, as having objectives. Furthermore, the kind of ﬁnal personal value in which I am interested also accrues to objects that often cannot be understood as having objectives (cf. Rabinowicz and Rønnow-Rasmussen, 2000, p. 48). A poem that my daughter wrote to me, which plausibly carries personal value for me, is a case in point. However, as I hinted above, there is a simpler way of understanding ‘sake’. Suppose all attitudes are either ﬁnal or instrumental. Obviously, this simpliﬁes matters. I might, for instance, favour something for its part or contribution in a whole (organic unity). However, I am quite sure that the inclusion of a discussion of these attitudes would not change the general picture—other than by making the presentation more complicated. So in what follows I will set these attitudes aside. Supposing, then, that there are only ﬁnal and instrumental attitudes, I can mean one of only two things by saying that I favour x. First, I might favour x as a means. We might explain this further by saying that we favour x for its instrumental properties, these being the properties that we believe make x causally efﬁcient in some way. That the attitude is directed on a means will then be reﬂected in its intentional content; this occurs when the instrumental properties are part of the intentional content of the attitude, i.e. when a favours x only for its instrumental properties. As to the second possibility, sometimes we favour x not as a means. In that case x’s being causally efﬁcient is not that for which x is favoured. The features for which x is favoured do not, or at least do not only, include x’s instrumental properties. I propose that we take ‘for its own sake’ to indicate that we are favouring the object in a ﬁnal, non-instrumental way, as speciﬁed above. There need not, therefore, be anything perplexing about having two for-the-sake-of qualiﬁcations in our description of an FFSS attitude. There is further complication here. As has been argued in recent years, sometimes we value objects for their own sake in virtue of their instrumental properties (e.g. see Korsgaard, 1983; Brännmark, 2001; Rønnow-Rasmussen, 2002b). The attitudes involved in such cases would have to be distinguished from the two attitudes I am discussing here. I am not sure about the details, but I suspect the intentional content of an attitude might include an object’s instrumental properties, though it would not do

62

FOR SOMEONE’S SAKE

so only for the sake of their instrumentality or causal effectiveness. Whether this is possible remains to be shown. However, if it is we should be able to differentiate such attitudes from the two I discuss above. To advance the process of clariﬁcation, then, I shall take the ﬁrst and second ‘sake’ in ‘a favours x for its own sake for b’s sake’ to indicate what sorts of property of x and b are part of the intentional content of a’s attitude. If x as well as b are not viewed merely as means, the properties for which we favour x for b’s sake will not be their instrumental properties.6 As to the values understood in terms of these attitudes, the ﬁrst ‘sake’ in ‘favouring x for its own sake for b’s sake’ speciﬁes that the value accruing to x need not be conducive to something else that is valuable. Final personal value is best described, then, as an end-point value or an ultimate value; this sort of value is not determined by something else being valuable. Similarly, the second ‘sake’ indicates that the favouring attitude is not instrumental. Our attitude is to b, and it is ﬁnal in the sense that it does not depend on some other favouring. An attitude with the attainment of some objective as its object will only be relevant to the analysis of personal ﬁnal value if the attainment of the objective is favoured for the person’s sake. It might well be incoherent to interpret the attitude ‘favouring x for its own sake for a’s sake’ in terms of the ‘end/objective’ sense of ‘sake’. Certainly it would be reasonable to request more details about the ‘objective’ a person might be said to have. However, in the ‘end-point value’ sense, there is nothing strange about ‘sake’ referring to one set of properties in the case of x and another in b. Why a particular set of properties supplies us with a reason to ﬁnally favour something for a person’s sake is ultimately an evaluative issue; attempts to address this issue will require substantive justiﬁcation. (Think of the hedonist who only acknowledges that certain kinds of experience are ﬁnally valuable; such a person will reject alleged examples of non-derivative ﬁnal personal values that are said to accrue to things other than experiences of pleasure.) The interpretation of ‘sake’ in terms of b (rather than in terms of b’s objective) is perhaps, therefore, most plausibly understood as being about b’s (discerned) properties. However, at this point it is important to see that the idea that attitudes take as their intentional object properties can be understood in various ways. As I have recently argued (2008b), when it comes to at least one kind of love, and one kind of admiration, the former but not the latter is a bona ﬁde example of an attitude that can, in a certain sense, be said not to take properties as its object. Here, this claim needs to be ampliﬁed, so in a moment I shall say a little more about the way love and admiration differ from each other in an important respect. Meanwhile, in order to avoid a possible misunderstanding, let me say something about how I do not understand personal values. Suppose that it is the case that:

6

One more possibility ought to be mentioned. I might have the non-instrumental properties of x and the instrumental properties of b in mind. This would be a complicated (but, I believe, genuine) attitude involving yet another object or person.

5.3

D I S C E R N I N G AT T I T U D E S

63

(1) x has ﬁnal personal value for b. And that the following therefore is true: (2) There are normative reasons to favour x in a ﬁnal way for the sake of b. That we can have normative reasons to favour something because we favour something else is plausible. But in that case, the question is whether we should understand (2) and ﬁnal personal value along the lines of (3)? (3) There are normative reasons to favour x in a ﬁnal way because b is favoured. As one referee from Oxford University Press suggested, this would seem to be confused: “If one has a normative reason to favor x because one favors b, then how can that be a normative reason to favor x in a ﬁnal way? To have a reason to favor x in a ﬁnal way is to have a reason to favor x that is not dependent on one’s favoring some other thing.” This is a fair worry. However, my analysis is not assuming the truth of (3). I also want to emphasize that I am not taking (2) to mean (3), and hence, do not understand ﬁnal personal value along the lines of (3). What follows from (1) is that we should favour x, not as a means but as an end for the sake of b. This is not confused, at least not necessarily, if we understand ‘sake’ as I suggested above. It depends on what role the properties of x and b play in the intentional content of a’s attitude. If your attitude can be directed to two objects, x and b, then there is the possibility that the properties of these objects can play the role of “instrumental properties” or “end-properties” in the intentional content of such an attitude. Moreover, if there are no conceptual obstacles for thinking that the properties of x and b are not instrumental properties in the intentional content (i.e. they are not discerned with an eye to what they can lead to) then I do not see why it cannot be the case that there is facts that call for us to have such attitudes. In other words, if we can make sense of such an attitude, we should also be able to make sense of the claim that there is a reason to have such an attitude. When there is a reason for such an attitude there is an object that carries ﬁnal personal value.

5.3 Discerning attitudes Although we often invoke FSS (i.e. for someone’s sake) attitudes, it would obviously be desirable if we could describe them in a more precise way. Ideally, we do not want to rest our case on linguistic reports which, in my view, rarely if ever capture the full propositional content of our attitudes. For instance, it is not hugely clariﬁcatory to say that we are dealing with an attitude the propositional content of which is ‘that something is favoured for someone’s sake’. I think the attitudes at issue here display several quite different kinds of propositional content, and I suspect that these contents are seldom of the above-mentioned type. Still, although I am unable to give a full account here, I will at least separate two kinds of for-someone’s-sake attitudes, both of which are pertinent to the analysans of personal value; and I will do so by considering the intentional content of these attitudes.

64

FOR SOMEONE’S SAKE

I suggested above that it might be incoherent to understand FSS attitudes in terms of the ‘end/objective’ sense. I also said that there is another approach—one pressing ‘end-point value’ into service—and that this approach looks plausible. On that latter view there would be nothing odd in allowing the two instances of ‘sake’ in FAP refer to different sets of properties. In one sense (the precise content of which I will come to shortly) this talk about understanding ‘sake’ in terms of properties is quite understandable. However, this fact should not hide another fact, namely that not all attitudes take properties as their intentional content. Love, for instance, is a bona ﬁde example of an attitude that does not in any obvious way have properties as its object. Moreover, I believe that the most interesting group of FSS attitudes have a good deal in common with love, and so, I think, there is good reason to clarify what is so particular about love. At least, on one conception of it. The notion of love is highly complex. For instance, is it, as different authors have argued, an apprehension of something in the beloved, or an emotional response or perhaps a volitional state that has the beloved as its object? The latter suggestion is Harry Frankfurt’s, and my view, which I brieﬂy touch on in this chapter, is in several respects similar to his, and in particular his account of what he calls “active love”.7 People are inclined to ascribe value to persons they love. However, the relation between love and value is far from straightforward. This is particularly evident if love is depicted, as I think it should be, as an attitude that takes non-replaceable persons as its intentional objects. The extent to which people do in fact display the kind of love depicted here is a matter I leave open. The romantic literature is full of references to this sort of love, though. For instance, consider how W. B Yeats ends his poem ‘For Anne Gregory’: But I can get a hair-dye /And set such colour there, /Brown or black, or carrot,/ That young men in despair/ May love me for myself alone/ And not my yellow hair.

Love is an attitude typically directed on some person. I will not discuss here whether this happens in a more direct or indirect way. However, this much can be said right away: many of our attitudes require for their explanation mention of the fact that we love someone. This lends some support to the ‘indirect’ idea that love plays a ‘master’ role among our attitudes. We love persons, that is, and not, say, the properties of the persons or states of affairs involving the beloved.8 Of course, people can love things other than persons (animals, and perhaps even things, including their cars or football teams). But that apart, we 7 See Frankfurt, ‘Autonomy, Necessity, and Love’ (1999a). See also his description of love in ‘On Caring’ (1999b, p. 165). 8 For example, Max Scheler, as Mulligan pointed out to me (personal communication) rejects the property view of love. See Mulligan, ‘Scheler: Die Anatomie des Herzens oder was man alles fühlen kann’ (2008).

5.3

D I S C E R N I N G AT T I T U D E S

65

typically love persons but not properties. Now this claim has been contested. In fact, it is not uncommon to ﬁnd it argued that what we love, when we love a person, is the person’s properties. This idea can be found in Blaise Pascal’s Pense´es.9 He refutes the view that what we love is a person with the observation that when one’s beloved changes in some more or less substantive way one’s love also does. If the change is great enough, the love will eventually fade away. This shows, he concludes, that what one loves cannot be a person. Instead, love is love of qualities: A man who sits at the window to watch the passers-by; can I say that he sat there to see me if I pass by? No; for he is not thinking of me in particular. But someone who loves a person because of her beauty, does he love her? No; because smallpox, which will destroy beauty without destroying or killing the person, will ensure that he no longer loves her. And if someone loves me for my judgment, for my memory, is it me they love? No, because I can lose these qualities without losing myself. Where is the self, then, if it is neither in the body nor in the soul? And how can you love the body or the soul except for its qualities, which do not make up the self, since they are perishable? For would we love the substance of a person’s soul in the abstract, whatever qualities it contained? That is impossible, and would be unjust. Therefore we never love a person, only qualities. (Pense´es and Other Writings (1995 [1897], pp. 130–1)

It is not obvious why anyone has taken this argument seriously. If the target is the idea that the qualities of a person have absolutely nothing to do with love, he has a point. However, his observation is not sufﬁcient to establish something that he also concludes, namely that the love of persons is a love of qualities. So why the popularity? For sure, the position has a distinctly unromantic ring to it, which I am sure has some attraction to at least certain people. But this cannot be the whole story. Perhaps the argument beneﬁts from the confusion of this non sequitur with the plausible claim that to love someone is, minimally, to stand in a causal relation to some of his or her qualities. There is some truth to Pascal’s sinister claim, though. Nevertheless it is the expression of either a quite confused idea or, at best, an incomplete description of what it means to love someone. The kernel of truth is, of course, that we love persons in virtue of their properties (I will get back to the complex copula ‘in virtue of’ in a moment). But it is a fallacy to go from this to the conclusion that we love the beloved’s properties. To make this move is to overlook the familiar difference between the cause of an attitude and its intentional object. The object of love is a certain person; and we should not confuse an attitude’s causes with its intentional object. Perhaps these might, in some cases, coincide. The fact that Romeo loves Juliet in virtue of her black hair does not mean that what he loves is the blackness of her hair. Her hair might well have

9

For a critical survey of different positions on love, see Christopher Grau, The Irreplaceability of Persons (2002). This dissertation drew my attention to the fact (among many others) that we owe to Pascal the view that it is the qualities we love, not the person.

66

FOR SOMEONE’S SAKE

something to do with the fact that he loves her, but it is Juliet that he loves, not a feature of her hair. Let us dwell a moment on this idea that what causes us to have an attitude towards an object need not be found in the intentional content of the attitude. It plays an important role in what follows. Unfortunately, and surprisingly, there seems to be no suitable word or simple expression that indicates decisively whether we have the content or the causal source in mind. All the candidate expressions are ambiguous. Thus, to say that a favours y ‘on account of ’, ‘in virtue of ’, or even ‘for’—to mention some readily imagined possibilities here—is to say something that can be understood both as being about what caused a to have the attitude and as being about the content of a’s attitude. This is a peculiarity of the languages that I am familiar with. To avoid misunderstandings, it must therefore be kept in mind that I will be using these expressions, henceforth, with the attitude’s content (not cause) in mind, leaving it open whether what caused the agent to have the attitude is also what appears in the content. Although English leaves us in the lurch here, there is a further reason not to forget this distinction between attitudinal cause and content. That it is not a case of philosophical pedantry is clear, I think. Here is a down-to-earth case. (I will touch on an analogous matter in Chapter 9, Section 9.4, when discussing the person who believed God had told her to jump into the water.) Suppose you admire a person for her honesty. If asked why you do so, you might truthfully say something to the following effect: ‘because she is honest’. Now, it is quite likely that her honesty caused you to admire her, but it would also be possible, in a different case, for you to admire her for her honesty even though she is in fact a dishonest person. In short, you may be mistaken about her character, but this is consistent with your admiring her for her honesty. Honesty, here, appears in the intentional content of the attitude, but since the person is not honest, it cannot be what caused you to have the attitude. Again, suppose I like my boss because he or she is generous. It is, of course, possible that his or her generosity caused me to like him or her. But it is also possible that I am deluding myself, and that what makes me like the boss is something quite different—namely, that he or she is my boss, or that my positive attitude will, in effect, make it more likely that I will be favoured by the boss. Clearly the distinction between attitudinal content and cause illuminates these examples. It allows us to understand how people can be mistaken about what they favour, and this is important, since, as a consequence of favouring things for what we take them to be, we sometimes favour them for what they are not. Plainly, properties we wrongly take people to have cannot cause us to have corresponding attitudes. However, this is quite consistent with the fact that these properties appear in the intentional contents of the attitudes. Notice, too, that in distinguishing between cause and content I am not implying that non-veridical attitudinal content is somehow without cause. I am merely saying that, whatever the actual causal story happens to be, what caused the content to be what it is

5.3

D I S C E R N I N G AT T I T U D E S

67

need not be what caused the person to have the attitude in the ﬁrst place (which certainly is one of life’s mysteries). It is an interesting metaphysical question, albeit not one I will address here, just how to more precisely understand what it means to say that a property is contained, exempliﬁed, manifested, or more generally ﬁgures in a propositional attitude. These expressions, which are the ones I employ in this work, are, as far as I can see, consistent with classical metaphysical views on property instantiation. Let us now return to the attitude of love.10 I think it is clear that, though it is in many ways a special case, love is not the only attitude that takes persons rather than, say, the properties of the person as its intentional object. Attitudes in general seem to be capable of taking this shape. Suppose, therefore, that the identity of a person can in some way be a feature of the intentional content of the attitude. This would at least be one way of understanding the somewhat cryptic claim that love ‘takes a person as its object’. The supposition can be made, I think, in connection with the sort of love that I am interested in here. Naturally, this ‘essentialist’ claim is not easy to spell out in any more detail; but I would like to provide a further indication of what I am driving at. Love appears to consist in a complex set of attitudinal dispositions. Sometimes these dispositions are activated, and accordingly the person then has the attitude. Love, in such a case, is describable as an occurrent attitude. These dispositions often play no active part in our motivational set-up. When this is the case love can be said to be in a dormant phase, being ready to move out of the motivational background when, for instance, one thinks of one’s beloved, or when one actually sees him or her. Suppose we consider a case of occurrent love. One person is, say, directing an occurrent love attitude on a, the person he or she loves. How should we then express the fact that a’s identity is a part of the attitude? The gist of a suggestion I have previously made (in Rønnow-Rasmussen, 2008b) is this: the identity of the beloved plays an indispensable part in love, in the sense that the persistence of the attitude in question is conditional on the lover’s belief that the person on whom the attitude is directed is identical to the beloved. Unfortunately this idea is not as clear as it needs to be, so let me try to develop it further. What is characteristic of the conception of love I have in mind is that it pictures love as taking a non-fungible (i.e. non-replaceable) person as its object.11 Love, thus understood, is nonetheless quite consistent with what are generally taken to be facts about people’s psychology—namely, that a person can love more than one person at a time, or that one can fall in love with someone b despite already loving a. Love, like other attitudes, can be individuated. Hence, in principle, nothing prevents us from having many such attitudes. In his discussion of love’s constancy, ‘A Conceptual Investigation of Love’ (1973), W. Newton-Smith says much that I am in agreement with. However, what I miss in this otherwise insightful work is an awareness of the distinction between what causes us to love someone and the intentional content of love (e.g. see p. 122, where Newton-Smith uses the expression ‘on account of’ without specifying what he has in mind). 11 This conception of love raises philosophically interesting issues that I have had to set aside here. I discuss some of these in (2008b). 10

68

FOR SOMEONE’S SAKE

Once love is thought of as an attitude taking a non-fungible (i.e. non-replaceable) person as its object, some interesting but rather counter-intuitive features come to the surface. Perhaps the most vital of these is the following: How should we harmonize the idea that the identity of the beloved is a part of the relevant attitude with the idea that attitudes are in general discerning? By the latter I have in mind, as may be recalled, the idea that our attitudes are directed on objects on account of the properties of those objects. That is, at least some of the properties of the object are exempliﬁed in the intentional content of the attitude. To be more precise, what makes an attitude more or less discerning is that the intentional content of the attitude contains these (discerned) properties to a greater or lesser extent.12 Earlier, in Section 3.2, the dual role of properties in attitudes was discussed. This role adds an extra dimension to this matter. If we take into account the fact that, in a value context and according to a ﬁtting attitude account of value, attitudes are called for, the aforementioned properties will also be what makes the object valuable. That is, these properties will make up the supervenience base of the object (they are subjacent properties). Thus in the case of a unique painting, uniqueness will be what (in part) makes the artwork valuable, or at least more valuable. However, facts about this property will also provide a reason for favouring the object (e.g. for caring for it). We should favour the painting on account of its uniqueness. Can there be entirely non-discerning attitudes?13 Could love be characterized as such? Could properties of a beloved individual be wholly absent from the intentional content of the love borne to that individual? Is this what might be meant by the claim that love ‘takes a person rather than properties as its intentional object’? The old refrain that love is blind might suggest as much. But though we cannot just rule out this possibility, the idea of a non-discerning attitude is somewhat puzzling. Attitudes are discerning in degrees, but it is hard to imagine a wholly non-discerning attitude to an object x, if by this we mean an attitude the intentional content of which does not contain, or represent, any of a’s properties. However, there is in fact another approach to be considered, which looks more viable. This will also improve our understanding of the cryptic idea that FSS attitudes take non-fungible persons as their objects.

5.4 Identiﬁers and justiﬁers We need, to begin with, to introduce a general distinction that seems applicable to attitudes generally. This is the distinction between what I shall call justiﬁers and 12 En passant, recall here that by ‘on account of ’ I have in mind what appears in the intentional content of the attitude; whether that is also what caused the attitude is something I leave open. 13 On certain views on reference, when the content of an attitude only contains, say, a proper name, the attitude is not discerning: such an attitude contains an element that has reference but no sense. What I say in this work about discerning attitudes might, therefore, if these views are correct, concern the special case of attitudes that do contain more than a proper name—that is, they contain some of the properties of the object of the attitude.

5.4

IDENTIFIERS AND JUSTIFIERS

69

identiﬁers. We can separate two kinds of property which differ from each other in the role they play in the intentional content of the attitude: justiﬁers justify, identiﬁers identify. Later on, this division will help us to understand the two kinds of FSS attitudes. I will be claiming that the identity of the person for whose sake we have an attitude may be a part of the attitude’s intentional content, and that this is consistent with the fact that the attitude is discerning. The explanation is that properties can ﬁgure in the intentional content in more than one way. Various conceptual barriers limit the kinds of property we can invoke to justify attitudes. For instance, the properties that typically justify love are not the kind that justify admiration. In fact there is good reason to believe that the kind of property that justiﬁes love cannot be what justiﬁes admiration. This suggests that properties may have different functions, or play different roles, in an attitude. This is true of normal attitudes as well as being true of the more complex FSS attitudes. Thus, I shall argue that, when it comes to admiration, properties of the admired object play the role of justiﬁers rather than identiﬁers; but that, by contrast, in the case of love properties of the beloved are identiﬁers. Of course, this needs to be further elaborated. That elaboration will make it quite clear that there is nothing to suggest that FSS attitudes are not capable of being similarly distinguished. Let me therefore return to the case of admiration, which is in many ways a more typical discerning attitude. We admire a person a for his or her wit, or courage, say. Properties such as these are what make a admirable. They form the supervenience base of admirability, i.e. the set of admirability-makers. Now, these subjacent properties may well play some sort of value-making or attitude-justiﬁcatory role in associated intentional contents. Actually both possibilities seem likely. Thus, sometimes when we admire a for his or her wit and courage, the content of the attitude is such that these properties of a appear in it as value-makers, as what makes a admirable. At other times we are not quite sure what it is about a certain object that makes us admire it: we just do. When this happens we are often ready to argue our case, or to take measures to ﬁnd out what it is about the object that we admire. In such cases the role of the properties is likely to appear in the content as that which justiﬁes the attitude rather than as makers of value. There will also be, I imagine, times when attitudes combine both kinds of content. However, for simplicity’s sake let us just make a mental note of this possibility. Let us again consider love. Here things are different; what causes one to love a person a will to some degree appear in the intentional content, but not obviously as what makes a valuable or justiﬁes one’s love. 14 We might try to express the difference as follows: in cases of admiration we admire a, and a is made admirable by his courage and wit. In cases of love we love a, and a possesses such-and-such properties. That we love a person in the sense I have in mind here is quite compatible with the fact that we may 14

Sometimes we might think that our love is justiﬁed. I am just not assuming that to love someone is to regard one’s love as justiﬁed. There is ample evidence that people sometimes love persons they do not think deserving of their love.

70

FOR SOMEONE’S SAKE

have a number of other attitudes to the beloved. For instance, Elly, my wife, has a number of qualities; she is clever, unselﬁsh, and considerate. Her having these properties makes her a good person; her qualities are such that I have a reason to favour her in various ways. These and many other features she has make her desirable, someone that I ﬁnd worth caring for, and so on. If the qualities were somehow lost, I would stop desiring her for the sake of those properties. I would not necessarily stop loving her. In the admiration case the properties ﬁguring in the intentional content are what make a carry value in the ﬁrst place (ceteris paribus, they are together what justiﬁes us in admiring a). In the love case, the properties are rather identiﬁers of a; they do not play any value-making or attitude-justifying role.15 This latter role is instead played by the identity of the person. The properties discerned are what identify the person, albeit not necessarily successfully; and it is the identity of the person that justiﬁes our directing of the attitude on the person. It is important to realize that the properties that function in this way as identiﬁers are unlikely all to be part of the intentional content. This implies that in order for a property to qualify as an identiﬁer it is sufﬁcient that it contributes to identiﬁcation—whether veridical or not. (The notion of an identiﬁer is, in other words, not a success notion.) That the content of an attitude does not always encircle the identiﬁers (of a person) is a complication, and one to which I will return in Chapter 6. For the time being we can set it (and the question why it is important) aside. It might be objected that the distinction between identiﬁers and justiﬁers is improperly artiﬁcial; there is no real argument for ascribing different functions to properties. Admittedly, there is an artiﬁcial ﬂavour to this distinction. Despite this, however, the criticism can be deﬂected, as I shall show shortly. The matter is complicated, though. So let me in the meantime brieﬂy restate what I so far have claimed. Discerning attitudes are attitudes directed on some object on account of some of the object’s properties. In some cases, some of these properties provide a (right kind of) reason for directing the attitude on the object. Suppose I admire a person a for being W. We can then construct the following schema, where the subscript shows that the arrow signiﬁes the justiﬁcation relation: W !j admiration of a In other cases, our attitude may be of a quite different kind. Love (at least, a certain kind of love) is perhaps the most obvious example. Since it too is a discerning attitude, it will contain the properties of the beloved in its intentional content. However, these properties will not ﬁt into the above schema. It is not my beloved’s possession of the property W that justiﬁes my love for him or her. Rather, it is the identity of the person that appears to have that function. Despite being present, in the intentional content, the properties do not play the same role here as they do in the admiration case. It is

15 Cf. Robert C. Roberts’s point, made in Emotions: An Essay in Aid of Moral Psychology (2003), that “love, in the dispositional sense of attachment... is a construal that identiﬁes the object of concern” (p. 288).

5.4

IDENTIFIERS AND JUSTIFIERS

71

therefore necessary to interpret the arrow in a different way when it comes to love. To mark this need let us use the subscript ‘ij’ to signify ‘identity-justiﬁcation’: W !ij love of a16 Now, it is not obvious that ‘justiﬁes’ is the right term for what is going on in the case of love. Certainly, one reaction at this point would be to say that the person who loves a has no justiﬁcation for his love (on this picture of love). He or she just loves a, period. ‘I-justiﬁes’ would then mark the fact that something other than justiﬁcation is involved. For the time being, however, I will regard it as an open issue whether we should consider such attitudes properly justiﬁed attitudes, or whether we need to look at what is going on here in entirely different terms. When it comes to property-justiﬁcation, with one important exception all or some of the properties of the object on which the attitude is directed are invoked as justiﬁcation. The exception has to do with the fact, which attentive readers will have considered, that identity is, of course, a property too. But property-justiﬁcation does not cover the identity of the object. As to identity-justiﬁcation, only the identity of the object on which the attitude is (either directly or indirectly) directed is invoked. Having stressed this difference—and it is perhaps important to emphasize, in particular, that we should not merely assume that there are two sorts of justiﬁcation—I believe it is safe to continue without noting on each occasion whether we are discussing propertyor identity-justiﬁcation. Henceforth, then, I will stick with the term ‘justiﬁcation’. We will return later on to the question whether I-justiﬁcation is really a variety of justiﬁcation or merely a mark of something else. I am inclined to think the former is the case. For the time being I am more interested in establishing that properties can play various roles in attitudes. Thus, having said that the identity of the beloved is what justiﬁes one in loving this particular person, we still need to allow a role to be played by some of a’s other properties. We need to do this because we have agreed that love is a discerning attitude. Where love is claimed to ﬁt into the above schema—and we can leave it open for the moment whether love is held to be like this for conceptual reasons or merely on occasion—a’s properties function as identiﬁers of a. Let us therefore return to the complaint that the distinction between the two roles is artiﬁcial and of no substance. There is more than one reason to contest this reply. For a start, the distinction is backed up by the observation that we have agreed that certain of our attitudes cannot always be properly justiﬁed although they are directed on objects. That is, the relevant attitudes are not justiﬁed by invoking the subjacent properties of the object (the identity of the object not included). However, a caveat is in place. The problem is not epistemological. Sometimes we might be more or less unsure what the relevant justiﬁcation is, but this is not what I am discussing now. The point is rather that 16 I owe David Alm thanks for pointing out (in personal communication) that my ﬁrst attempt to present the scheme was more confusing than helpful.

72

FOR SOMEONE’S SAKE

in some cases we acknowledge that an attitude of ours is such that we ﬁnd ourselves merely having it; we have sided neither with nor against it. The properties of the object are ‘there’ all right, in the intentional content—but they are not doing any job at all as justiﬁers. We have, as it were, been struck by the attitude.17 In fact, at times we do not merely experience the attitude as something planted in us; we also believe that there are good reasons (either moral or prudential) for us not to have the attitude in question. A similar phenomenon has been discussed at great length by theorists analysing emotions. Take a textbook example. Many people suffer from fear of ﬂying although they know very well that it is much safer to ﬂy than it is to drive a car. Their fear remains unaffected by their belief. Another example is a person who has a certain preference for, say, a particular sexual activity, but who believes that there are good reasons for him not to live out his sexuality.18 Justin D’Arms and Daniel Jacobson refer to these kinds of emotion as “recalcitrant emotions”, and this label catches very well what these affective states are all about.19 As far as I can see, emotions are not the only recalcitrant mental states we have. There are many attitudes that we somehow regard ourselves as having, but which we look upon as alien to the kind of person we wish we were; these attitudes are like uninvited guests that we try to show the door but often just have to learn to cope with. Notice that I am here describing a general phenomenon. I am not saying, for instance, that love should be regarded as an alien, uninvited attitude. It might be like that, but it need not be. I am merely pointing out that sometimes we do not intend to, and cannot, justify a discerning attitude in terms of properties of the object on which the attitude is directed. This strongly suggests that the properties of objects can play various roles in attitudes. The distinction is therefore not a pure invention. There is a sound basis for making it. In fact, a further observation—besides the purely phenomenological one that we sometimes experience ourselves as having an attitude we recognize ourselves as just having—a fortiori strengthens the idea that properties which ﬁgure in intentional content may be playing various roles. Thus sometimes we are ready to accept that it is logically legitimate for a person not to provide a justiﬁcation, cast in terms of the properties of the object, for having an attitude. Of course, it is far from obvious exactly how logical legitimacy should be determined. Much could be said here. I shall merely hint at what I have in mind. In some cases, we seem to accept that people have attitudes that they simply recognize as theirs—we allow them, if you like, and crudely put, to be struck by the attitude. Moreover, we recognize that there is no linguistic ground on which to demand justiﬁcation for the attitude. Certain attitudes are merely had. 17

This observation has been made by some authors: e.g. see Frankfurt (1999a). A case of this is the Swedish priest who declared in the newspapers in the summer of 2007 that he was homosexual. While he thought it was not a sin to be a homosexual, he did think it was a sin to perform homosexual acts, and he had therefore decided to live in celibacy. 19 See D’Arms and Jacobson, ‘The Signiﬁcance of Recalcitrant Emotions (or anti- quasi-judgementalism)’ (2003): “We will say that an emotion is recalcitrant when it exists despite the agent’s making a judgement that is in tension with it” (p. 129). 18

5.5

T WO K I N D S O F F S S AT T I T U D E

73

A typical example is having a desire for something. Consider a person who wants her house painted red. She may have no other ‘justiﬁcation’ than the fact that she desires to have her house painted red. After a period of indecision she just realizes that it is a red house she wants. By refusing to come up with an answer to the question ‘why do you want it red?’ (other than something along the lines of ‘I just do’) this person is not committing a logical mistake. Compare this case with one in which someone a admires another person. Here, a may not be able to provide a justiﬁcation for her attitude, but, for logical (i.e. conceptual) reasons, she cannot deny that a justiﬁcation is called for. To say ‘I admire her for no reason at all’ is to commit a conceptual error, to misuse the term ‘admiration’. ‘I desire a red house for no reason’ is not in the same way a logically strange thing to say. Whether the ‘desire’ case should be handled in the way I propose for love is doubtful, but I will not pursue this issue here. Another tricky issue that I will not go into requires us to say to what extent we can genuinely admire something that we acknowledge not to be admirable. However, notice that whether or not we admit this possibility, it remains the case that, for conceptual reasons, if you admire something, you can always legitimately be asked to justify your attitude in terms of the admired object’s properties. We do not so much love a person for his or her properties as we love him or her regardless of his and her properties.20 As a result, we cannot (nor do we recognize the need to) supply a justiﬁcation for our love cast in terms of the beloved’s discerned properties. This, in its turn, explains why we sometimes feel that we are not in a sense responsible for having the attitude. However, on this view of love we are stranded with a problem. For supposing that attitudes are never wholly non-discerning, we need to understand what role properties do play in the attitude. I suggest that the distinction between identiﬁers and justiﬁers accounts for this.

5.5 Two kinds of FSS attitude Let us now return to the FSS attitudes mentioned in the analysis of personal value. I repeat (3) above from Section 5.1, then: (3) a favours x for b’s sake Now it might seem tempting to ask what this kind of attitude resembles more, love or admiration. However, since ‘favours’ is but a notational alternative to ‘has a pro/ contra-attitude’, this question would appear to be too general to admit of any deﬁnite answer. Still, we might conﬁne ourselves to the kind of FSS attitudes that we believe will, as a matter of fact, be expressed when we think something has value for someone. We might then ask: Should personal values be understood in terms of love-like or 20 Recall that I have set the psychological realism of these attitudes aside. I am interested in a certain notion of love as a logical possibility.

74

FOR SOMEONE’S SAKE

admiration-like FSS attitudes? Or to put the question differently: Is favouring x for b’s sake an attitude whose persistence is conditional on the identity of b? Is it, in other words, one, like love, which accommodates identity-justiﬁcation? Or is it an attitude whose persistence is independent of the identities involved—one we might even be compelled logically to adopt, on pain of conceptual error, vis-a`-vis a person, or thing, that we agreed did not relevantly differ from b? Perhaps the question should be rephrased. We must keep in mind that love is an Identity attitude, i.e. an attitude capable of identity-justiﬁcation, while admiration is a Non-Identity attitude. The issue then, I believe, is not whether there are FSS attitudes that are dependent for their existence on the identity of a person, for clearly there are. What makes things less straightforward is that it also seems clear that FSS attitudes do not all have to be like love (in this respect, that we are not conceptually compelled to love someone who shares the same features as someone we love). That is, there can also be Non-Identity FSS attitudes—attitudes which, in common with admiration, are such that the persistence of the attitude is conditional not on the identity of a person a, but rather on universal properties of a. If these observations are correct—and in a moment I will argue that they are—the real crux of the matter appears to be whether personal values should be understood in terms of only one type, or both types, of FSS attitude. Consider the following scenario. I am at a dinner. The people at the table are all unknown to me. Suppose person a in front of me politely asks me to pass the salt to b, the person sitting to my left. This person, too, is a complete stranger. If I pass it, I will have acted on what might be described as a desire to pass the salt to b for a’s sake. Is this desire conditional for its persistence on the identity of a? This seems rather unlikely. Still, I can correctly say that I passed the salt to b for a’s sake. Consider next a different scenario. The woman sitting beside you at dinner presents herself in the following way: ‘I am John’s mother’. When you realize that you have the mother of your friend, John, beside you, you decide to take especially good care of her. You soon ﬁnd her quite boring, but you do your best to be entertaining—and you do so for your friend’s sake. When the evening is over, you realize that she was John’s mother all right, but not ‘your’ John’s mother. Here your attitude of being entertaining for your friend’s sake does seem to be conditional for its persistence on the identity of John. The ‘John’ that ﬁgures in your intentional content is not one that corresponds to your real friend. This conclusion may not seem obviously correct, but the following observation supports it. The John of your attitudinal content is obviously a person who is related to the woman sitting next to you in a mother–son relationship; it is when you realize that this ‘content-John’ does not correspond with the real John that your attitude changes. Let us now return to the question I posed earlier: Should personal values be understood in terms of Identity FSS as well as Non-Identity FSS attitudes? Here I am afraid that we have gone as far as formal analysis can take us.

5.6

SUMMING UP

75

Certainly, it might be argued that in the case where I pass b the salt for a’s sake it is rather obvious that, in some sense, I am not really doing it for a’s sake. Instead I am doing it for the sake of anyone who exempliﬁes the universal properties of a that make me pass the salt.21 In this case these properties pretty much amount to being seated in front of me, being a stranger, and asking me to pass the salt to another stranger sitting beside me. However, this does not really answer the question; it only spells out again that there is a difference between two kinds of ﬁnal attitude. Cogent argument that personal value should never be understood in terms of Non-Identity FSS attitudes would require evaluative reasoning. I think, therefore, that the best approach to this issue (at least, in a work of a formal character) is to acknowledge that personal values can be analysed in terms of both Identity and Non-Identity FSS attitudes. The more interesting personal values—those expected to have a substantial impact on a person’s life—will, in my view, be those that require for their analysis Identity attitudes; and it will be these with which I concern myself as we continue. On the other hand, in my view we should allow even Non-Identity-ﬁnal attitudes a role in our evaluative reasoning. Consider again the example of the tourists in the church, which indicated that it is possible for us to take something to have personal value even if we do not know whose value it is. The tourist believed that he should behave in certain ways for the sake of the religious believer. People (he thinks) ought to respect certain religious objects for the sake of people attending worship. In cases like this, at least, people appear to recognize that they have reasons to direct Non-Identity FSS attitudes on the relevant objects.22

5.6 Summing up I fear that this chapter covers several somewhat elusive issues. A brief summary might therefore be in place. After introducing some examples of FSS attitudes, I considered the possibility that these attitudes are, in effect, combinations of at least two or more standard attitudes. I am not closing the door on this possibility. However, since the two most salient examples that I can think of do not seem entirely satisfactory, I will continue to regard them as cases of a single attitude. Unfortunately I have had to employ a linguistically awkward formulation to persist in this regard. But, as I argued, this should not deter us. It is not necessarily incoherent to think people have such attitudes. It is possible, I argue, to make reasonable sense of these for-someone’s-sake descriptions. In Section 5.3, I moved on to explain what I have in mind by ‘identity21 The example is simpliﬁed. In point of fact two personal values could be in play here: it might be the case that you ascribe personal value to the satisfaction of b’s request or desire to make me pass the salt to a, or it might be the case that you think a’s getting the salt is what carries a’s personal value. 22 The example is complicated, though. Some of us will also think that, fundamentally, the worshipper does not have any reason to favour the Madonna, which suggests that he or she might not after all consider the Madonna to be carrying someone’s genuine personal value; our value judgement expresses rather a kind of inverted commas use of ‘value for’.

76

FOR SOMEONE’S SAKE

involving attitudes’. To this end, I discussed a typical example of such an attitude, namely love. I went on to contrast love with admiration, which I employed as an example of a Non-Identity attitude. Having prepared the ground in this way, the distinction was then, in Section 5.4, applied to FSS attitudes. They, too, which was suggested in Section 5.5, can easily be divided into Identity- and Non-Identityinvolving cases. In Chapter 6, I consider some foreseeable objections to my notion of an Identityinvolving attitude. For example, it would seem, on my account, that the dropping of such an attitude is difﬁcult to imagine. However, I will show why there is no need to worry; the objection overlooks something that plays an important role throughout this work, namely the distinction between an attitude’s cause and its intentional object. Chapter 6 also considers some other issues, the toughest of which arises when we try to analyse value in terms of these FSS attitudes. In their nature, the Identity FSS attitudes (which I think the really interesting personal values require for their analysis) appear not to be well suited to value analysis. However, as I shall argue, although this is certainly true when it comes to impersonal values, it need not be true about personal values.

6 Examining the Analysis

The preceding chapter was intended to add more ﬂesh to the bones of the rather opaque ‘for its own sake’ idiom, and to the thought that personal values accrue to objects that we value with a particular individual in mind. To accomplish this, I provided intuitive support for the idea that there is conceptual space to be ﬁlled in such constructions as ‘O ought to be favoured for a’s sake’. If the suggestion I have made about FSS attitudes (i.e. for-someone’s-sake attitudes) holds water, we now have the materials we need to analyse personal values in a novel way. Before continuing any further, I should point out that in what follows I will, unless clariﬁcation is necessary, speak of ‘favouring’ without specifying the kind of favouring involved, i.e. whether or not it is ﬁnal or instrumental. This simpliﬁcation allows me to avoid the cumbersome locution ‘favour x for its own sake/for its effect’s sake/for something else’s sake’. Now, FSS attitudes, like many other attitudes, are such that the properties in their intentional content play various roles. In fact in the philosophically interesting cases they appear to be particularly striking examples of discerning attitudes that display intentional content-properties with more than one function or role. Their peculiar directedness on something for someone’s sake strongly suggests that identiﬁers may play a role in their intentional content. Moreover, this approach implies a quite natural way of separating the kind of attitude called for by objects carrying personal value and the kind of attitude called for by objects with impersonal value. That is, the identity of the person for whose sake we have an attitude plays a particular, and crucial, role. For instance, when I care for a person b for my friend a’s sake, the properties of a will, in part at least, be identity-makers in the caring attitude. The ‘justiﬁcation’ of the attitude is intimately related to the identity of a. If I somehow come to believe that I am mistaken—that the properties identify the wrong person—my caring conduct will emerge in a different light. I will have been caring for b with an eye to a person I take to have properties identifying a. But the person in question is not a, but a*. Of course,

78

E X A M I N I N G T H E A N A LY S I S

I might care for b nonetheless—it might, for instance, be the moral thing to do. But if b is not somehow related to a, but rather to a*, and if I have no relation to a* (who is not my friend), and I initially but mistakenly cared for b for a’s sake, it will be, if not inexplicable, then at least highly unexpected that I am caring for b for a’s sake. My operative reason for caring for b will be mistaken. This is an important point, so let me dwell on it a little more. What I am trying to make clear is not that it would be impossible for someone to care for something for a perfect stranger’s sake. It would be absurd to deny that possibility. This was what the ‘pass the salt’ example was designed to show. The point is rather that, on many occasions, realizing that you have mistaken the identity of a person is sufﬁcient for you to drop your FSS attitude. I take this to be a fact about us as persons. In order to explain this phenomenon, given that properties do play a role in the intentional content of the relevant attitudes, we need therefore to come up with some sort of story. The suggestion is that in some cases the identity of the person plays no vital role, and in other cases it does. Personal values, then—at least those that interest me—should be analysed in terms of these latter attitudes. Accordingly, we can now spell out FAP in a bit more detail: FAP*: An object x’s value-for a person a (i.e. x’s personal value), consists in the existence of a normative reason for favouring/disfavouring x for a’s sake—where this favouring/disfavouring is an Identity FSS attitude.

6.1 Dropping an FSS attitude It might seem that the view of FSS attitudes outlined here faces, in effect, a reductio ad absurdum. If these attitudes (at least, those of major relevance for FAP, the so called Identity FSS attitudes) characteristically have non-fungible persons as their objects, it becomes hard to explain how people might come to stop having these attitudes, especially if the people for whose sakes we have the attitudes are qualitatively the same. If FSS attitudes are genuine attitudes, they will be expected to behave like other kinds of attitude, and this will mean, among other things, that they can be dropped and replaced by other attitudes (though not necessarily in any voluntary direct way). But supposing this is true, and given that identity plays the peculiar role it does when we have love in mind, as outlined above, we seem to face something of a mystery. In the case of regular attitudes we often explain the loss of an attitude by saying that we made a mistake about the properties of the object of our attitude: for instance, the car you once wanted turned out not to have excellent brakes as you originally thought, and so you do not want this car any more. There is nothing strange about this. But when it comes to Identity FSS attitudes, since no universal properties of the object (i.e. properties capable of being described without reference to any particular individual) justify, or explain, your direction of the FSS attitude on the object, you cannot ever, it

6.1

D RO P P I N G A N F S S AT T I T U D E

79

seems, avail yourself of a similar explanation. Therefore, it might appear that the view outlined here leads in effect to an absurd position. Does it not oblige us to ascribe to the person who has lost the attitude some sort of amnesia concerning the identity of the individual for whose sake she originally had the attitude? But that sounds odd; people do stop having attitudes, and they do so without having any doubts whatsoever about the identity of the person they once, say, loved or cared for. Someone might be quite positive that they are not suffering from amnesia, and that the person they do not love is the very same person they used to love. So there must be something wrong after all with the way love and other attitudes are described here. This objection can be dealt with quite easily. It is paramount not to blur the distinction between what ﬁgures in the intentional content of one’s attitude and what caused one to have the attitude. Causes need not appear in content. Once we accept this, it should come as no surprise at all that what causes an attitude to disappear need not be part of its intentional content. So if the explanatory reason (i.e. the explanation in causal terms of ) why I favour something for a’s sake is some universal property, then a’s loss of this property might well result in my dropping the attitude. What might have seemed mysterious then boils down to the fact that what causes one to have, and not have, attitudes is often beyond one’s epistemic reach. But this is nothing particularly troublesome for FSS attitudes, and indeed it is a feature of most of our attitudes. But then, it might be replied, we face another problem, namely that we are unwarrantedly mystifying the typical person’s psychological set-up. That is, does not the above response imply that when we lose an attitude, we must be at a loss as to why we did so—why we stopped, say, being in love, or are no longer ready to care for b for a’s sake? Are we really epistemically thwarted in this way by our own psychologies? The implication does not follow, and besides I think we are sometimes surprised by the attitudes we have or do not have. Just as we in some cases are struck by an attitude, we can be struck by its absence. Literature is full of examples of men and women who wake up one morning realizing that they no longer love the person next to them, or who ﬁnd that they have changed, overnight, into people who no longer need alcohol or drugs. Such ﬂashes of realization are, at least, reported. However veridical these reports are, we must of course grant that they are seldom followed by a genuine sense of perplexity; but this shows only that people often have beliefs about the causes involved. It is far from obvious that a certain kind of value should be analysed in terms of attitudes that take non-fungible persons as their objects. For one thing, attitudes like these are not expressible in universalizable value judgements. Value and universalizability are, in many philosophers’ eyes, intimately related. Though universalizability may in fact be an obstacle to FAP*, I will postpone discussion of it at this point. Instead I want to brieﬂy comment on an issue which, I am sure, has been lurking in the background for some time now.

80

E X A M I N I N G T H E A N A LY S I S

6.2 Non-fungible persons and identity Earlier I stated that I take it to be a psychological fact that the persistence of some of our attitudes is conditional on the identity of the person on whom they are directed.1 A rich, and in many ways intriguing, literature suggests that we should be careful when employing the notion of identity. This warning is based on the fact that it is fairly easy to imagine cases in which, for logical reasons, it seems impossible to say whether or not an individual at time t retains her identity at a later time t*. Given this background, I can imagine the following sort of response to what I have been doing so far. It might be pointed out that whether or not an attitude is, as a matter of fact, conditional on the identity of the person, some philosophers concerned with socalled diachronic identity (identity over time) have argued that attitudes ought not to be conditional in this way. So whether or not there happen to be such attitudes, we should give up on them. There are good grounds for not distinguishing between objects only in virtue of their identities, and if two objects are qualitatively identical, it is irrational not to direct the same attitude on one object you have directed on the other. The position I set out from is not inconsistent with this response—at least, so long as we clarify it by making some very plausible amendments to it. Let us therefore return to the example of love, which makes for an interesting test case. To begin with, recollect what has so far been said about this attitude. The love one has for a particular person at time t1 is not dependent, for its persistence over time, on one’s beloved’s possession of precisely those properties that he or she has at t1. This independence feature explains why love does not necessarily disappear when the properties of the person loved change. To love a particular person is to love him, or her, whatever his, or her, universal properties are—where these universal features are describable without reference to any particular individual. Love might therefore be characterized as unconditional from one perspective. However, if we turn to what causes this attitude, love may well be dependent on the fact that the beloved does not exhibit certain properties (e.g. being unfaithful or a child molester). But acknowledgement of this does not commit one to saying that one loves one’s beloved for not being, say, a child molester. All it means is that if she were one you would not love her.2 This picture constitutes a truly romantic view of love. Needless to say, it is an incomplete account. It says nothing about what it is to feel or experience love: it concerns the object of love. Furthermore, I hold it to be a clear possibility that love, as depicted here, might be an element in different kinds of love. A description of, say, parental love would not coincide completely with a description of love between adults, and the love I have for my wife is not the same as the one I have for my brother and

1

I am inclined to think that in certain attitudes—love would be an example—this conditionality is analytically built in. However, this is not a claim I shall argue for here. 2 See here Jonathan Dancy (1993, p. 77).

6.2

NON-FUNGIBLE PERSONS AND IDENTITY

81

sister. Still, love as characterized here may well be a core element in different kinds of love. In some cases it is more obviously present than it is in others. It might be objected that we have no reason to believe that anyone actually has this sort of love for anyone. We might acknowledge that there is a conception of love such as the one depicted, and agree also that this conception is pressed into service in romantic novels and the like. But in real life people do not have this attitude to each other. Perhaps this is the case. It is at least a reasonable view. I am just not sure how one would go about settling whether or not it is accurate in a convincing way. The fact that people divorce, break up or fall out of love is not conclusive evidence. As we have seen above, falling out of this (demanding) kind of love is quite possible. However, let me at least bring up some evidence that love of this kind might not after all be a merely romantic fantasy. Suppose your beloved dies and is replaced by an exact replica with the same physical appearance and mental set-up. The two persons here are so similar to each other that you would be unable to point to any difference between them in terms of appearance, behaviour, and thinking. If you think that you have a reason to feel sorrow, or at least believe that you would feel sorrow, over the loss of your beloved, the likely explanation is that (you believe) your love is of the romantically resistant kind. The person you love in virtue of whatever properties he or she has is no longer with you. So you feel sad.3 To feel sorrow in a case like this is quite consistent with the idea that you can come to love the replicant; after all, this is quite likely to happen. (As might be recalled, it is quite consistent with the view outlined here that one could replace one love attitude with another one.) We may expect that many of the properties that once caused you to (fall in) love (with) your beloved are present in the new person too. Whether or not this is the case, however, the point here is that lovers will in fact say, about the scenario above, that they would be saddened by the news that their beloved has ceased to exist: the fact that a perfectly similar (or even better) replacement would be provided does not in itself cancel this reaction. One might try to resist this by arguing as follows. Suppose, without your knowing it, we replaced your beloved with an identical copy. Would you not go on—in fact, would you not have to go on loving this person as much, and in the same way, as you did with your beloved? This kind of challenge is well known, but hardly impressive. A person who experiences the love that I have in mind will go on loving the person whom he falsely believes to be his beloved. How could it be otherwise? But from this it does not follow that on discovering that his love is directed on a copy, and that the 3 Cf. Michael J. Zimmerman’s (2001b) description of romantic love: “Suppose Kath has an identical twin, Kay, identical not only in terms of appearance but also in terms of personality, and so on. If Kath’s wit and ﬁgure give me equally good reason to love her for her own sake, then Kay’s wit and ﬁgure give me equally good reason to love her for her own sake. But this is a disturbing thought. It’s disturbing because it seems incompatible with the attachment that I have to Kath in particular. The fact is that neither Kath, nor Kay, nor I believe that reason somehow requires me to love Kay as I do Kath. Romantic love is simply not subject to duplication in this way. In this sense, it is not a rational attitude” (p. 129).

82

E X A M I N I N G T H E A N A LY S I S

original is somewhere else or no longer alive, he will not go and look for her or feel sad that he has lost the person he loved.4 This, at any rate, is my sense of the matter. Not everyone will share my intuitions. Some will say that the realization that their beloved has been replaced by a qualitatively identical copy will annul any feelings of sadness: they would not expect to experience the slightest negative feeling. To back this up, they will stress that there is something truly ﬁshy about attaching weight to particularity in the way I am doing. They will contend that, as far as love goes, nothing of importance happens when we replace the beloved individual with an exact copy. Again, how realistic these people are is a matter of intuition. I am certainly inclined to think that any person in the circumstances described would feel some sadness, and that there can be cases of the sort of ‘resistant’ love I have in mind. But suppose, as was once suggested to me, that by some metaphysical oddity my beloved disappears in thin air, and a split second later there appears on the same spot a being identical in all universal features with my beloved. Surely I am rationally required to love this new person in the same way as I loved the one who disappeared? Again, let us set aside the fact that I would probably love the ‘new’ person—I would do so for the simple reason that whatever caused me to love the original individual would cause me to love his or her replica. But this is one thing. The question I am now addressing is different. I am asking: Am I somehow irrational if I do not love this person—or, at least, do not love her in the same way? Intuitively, we distinguish between properties that are internally and externally related to an object. Suppose we consider our beloved only in terms of his or her internal properties.5 Perhaps this will place too much strain on our imaginative capacities, but suppose it can be done. In that case, I venture to say that the whole idea of directing an attitude on a that differs from the attitude you direct on b when a and b share all internal properties becomes incomprehensible. If a and b are exactly alike in their internal properties, and if you are not allowed to consider external properties at all, how could you feel differently about a and b? Things look different if we include externally related features. Which ones to contemplate is an open question. However, the most obvious are, I think, those concerning the history of the person. I am not ready to replace my beloved with a copy, however similar in internal properties, for the simple reason that the copy will not be the person with whom I have a shared past, to whom I owe obligations, have made promises, and so forth. The idea that our beliefs about people’s identity are in part based on external properties gives us occasion to further illuminate the notion of an identiﬁer. So far the identiﬁers that we have been speaking about have been assumed to be parts of the 4

Cf. Christopher Grau (2002). There is insufﬁcient space in this work to discuss the complex nature of this distinction. Its fruitfulness in application will, I hope, be displayed, though. 5

6.3

G O O D -F O R A N D W E L FA RE

83

intentional content of the attitude. This picture needs now to be slightly modiﬁed. Although some relational properties seem discernable, obviously, this cannot be the case with all features. In cases in which something was related to an object in the past, we are forced to say that the relevant property cannot be what we discern. This seems incontestable. The implication, then, is that in some cases (I would even go as far as to say, in most cases) the set of identiﬁers is not entirely made up of properties belonging to the content of the attitude.6 Now, to return to the earlier issue, one might be morally required to love the replica, and it certainly would not make any sense to deny that one would love him or her if one were unaware of the switch. However, since the replica is not the person one shared one’s past with, this suggests a way of explaining the fact that one would be saddened by the news of a switchover: the replica is not the one who, say, gave birth to my daughters (she believes so, but she is mistaken). In fairness, by referring to the ‘new’ person as a replica I might be accused of evading the issue. Why not imagine a possible world in which this person has all of the beloved’s properties, internal and external? Then what? At this point it becomes hard to follow the example. But notice that the reason for this is not that the notion of love outlined here rests on some metaphysical idea of a ‘personal substratum’ (e.g. that to which a person’s features attach). Nothing I have said so far saddles me with this commitment. (Keeping our metaphysics to a minimum, we should allow for the possibility that we love persons even if we happen to believe that they are nothing but, say, bundles of properties or, if you prefer, tropes.) Of course, I can imagine my beloved in a possible world, but I am not sure it makes sense ﬁrst to think of someone, and then to imagine this someone having his or her own properties (this we must assume) replaced by all of my beloved’s properties in some possible world. So I will not pursue questions about what is going on in this strange kind of example.

6.3 Good-for and welfare FAP* captures, I believe, something important about value-for. Does it also exhaust this notion? Most probably, it does not. I have already noted in passing that we should not expect there to be a perfect match between value and normativity. Some values in particular do not perfectly ﬁt the idea that values are nothing but reasons for attitudes. Beauty is perhaps the most arresting example. For instance, people who are beautiful (or ugly, for that matter) have been said to strike their observers with their looks.7 6

Should we allow these relational properties to be part of the attitude indirectly? Perhaps this makes sense. However, it does not affect the point being made here, i.e. that certain properties cannot be said to be discerned. 7 Recently Kevin Mulligan has voiced this sort of objection to the buck-passing account; see his ‘Emotion, Value, and Morality’ (2009b). For a similar point, see also Johan Brännmark, ‘Goodness, Values, Reasons’ (2009).

84

E X A M I N I N G T H E A N A LY S I S

A particular kind of experience is involved in such cases, and it is not obvious that we have experiences of reasons—at least, not in the sense that we are, for instance, taken aback by the beauty of a certain face. This discrepancy is worth examining, but to take such cases at face value and conclude that we actually experience sui generis value properties, and that a ﬁtting-attitude pattern of value analysis therefore cannot be the whole truth, would be unwarranted. There is ample evidence that we should not automatically trust phenomenology of this sort to be veridical. Ideally, then, a buckpasser will provide a suggestion as to why we should not rely on phenomenology in cases of ‘value experiences’. This is not something I will be engaged in here, though. Rather, I prefer to explore another possibility, and one that seems to me more important—namely, that there might be more than one notion of value-for, and that FAP accounts only for one of these. This possibility certainly looks to be worth pursuing, especially in connection with positive or negative expressions of value-for—that is to say, in connection with ‘good-/bad-for’. Let me therefore say something about this matter here. I shall focus on the relation between good-for and welfare, although the approach I shall take to this matter could equally well be applied to the many attempts to link good-for to interest or beneﬁt.8 I will follow up on this discussion in Chapter 7, where I will consider some examples of good-for analysis that appear to involve a notion of good-for differing from the one FAP* aims to clarify. One way of conceiving of the relation between good-for and welfare is quite consistent with FAP*. I shall refer to this as the thin-thin synonym relation. Suppose some x is constitutive of, or is, the welfare-making base of a person a’s welfare. For the time being we can leave out the details of x. It seems perfectly natural to say then that x is good for a (where ‘good for’ thus expresses a positive value-for). This seems inoffensive because many of us would like, at least, to say that whatever is good for a person is welfare-constitutive, and that whatever is welfare-constitutive is good for the person. Effectively, we regard the connection as holding between two value conceptions (or, if you prefer, between two evaluative notions). It is merely a verbal matter, on this view, whether we should use ‘good for’ or ‘welfare’ (or indeed ‘of beneﬁt to’) in order to express these two value conceptions. So whatever we take to be good-for making properties are also, analytically, welfare-making properties, and vice versa. But notice that in acknowledging this we have agreed neither that the notion of goodfor is analytically related to a particular non-evaluative term, nor that welfare is so connected. I doubt that welfarists (who can be treated here as anyone who ascribes value-for to welfare) would analyse the relationship between good-for and welfare in the way just described. More likely they would take welfare to be either a natural (descriptive) E.g. see Zimmerman in ‘Understanding What’s Good for Us’ (2009), who takes “a beneﬁt account of goodness-for to be correct; what is good for someone (in the relevant sense) is what beneﬁts him and, all else being equal, his life is good to the extent that it beneﬁts him”. 8

6.3

G O O D -F O R A N D W E L FA R E

85

notion, rather than a value notion, or a thick value notion. Later on I will have more to say about especially the latter possibility. Here I want to introduce the thin-thin conception, and to outline yet another possibility, which is the one I favour myself. Welfarists (whether or not they are in a position to explain what they mean by welfare in terms of natural, non-evaluative properties) need not be treating good-for/ welfare in accordance with any of the approaches mentioned. They might be thinking, say, that welfare, understood as a certain mental state, is just an example of what is good for the agent. ‘Good-for’ and ‘welfare’ are not regarded as synonyms, since for one thing the former can be applied to things other than welfare. But welfarists are generally not quite that permissive when it comes to what can be good for an agent. They might acknowledge that something other than mental states of the kind they focus upon can be good for the agent. However, such a good-for claim would then be regarded by this sort of welfarist as being about a different kind of good-for value. It would be an employment of a notion of good-for of a different kind to the one they employ when they think welfare is good for a person. Their good-for notion becomes necessarily (i.e. analytically) connected with, for example, some mental state, such as feeling pleasure, or happiness, that is regarded as being what welfare/good-for is all about. Welfarists of this kind tend, in other words, to think that there is sufﬁcient descriptive content in the notion of welfare to allow one, on purely conceptual grounds, to rule out a number of candidates as good-for/welfare constituents. So, for instance, when Connie Rosati,9 as we shall see later on in Chapter 7, makes a sharp distinction between ‘good for’ as a term referring to personal welfare and other senses of this expression (e.g. in the assertion ‘water is good for the plant’), she has in mind a narrower notion than the one I am seeking to capture in FAP. A third notion also seems to be intimately related to welfare and good-for, namely ‘better off ’. Thus, we might say that something is good for a person if and only if it makes him better off; and that something makes a person better off if and only if it increases his welfare. Again, I am strongly inclined to think that if we are not merely regarding it as a synonym of ‘good for’, ‘better off ’ will at some point or other have to refer to either an impersonal or personal value, in which case what I have said so far will also apply to ‘better off’ approaches. I am not assuming that there are things which, in virtue of the meaning of ‘value for’ or ‘good for’, cannot exemplify these values. In my view, it is an evaluative (or, if you prefer, a normative) issue what things are good-for/welfare-constitutive. If we regard the connection between ‘good for’ and ‘welfare’ as analytic in the thin-thin sense, no obstacles to FAP* arise. However, I suspect that for many observers this will be to strain the meaning of ‘welfare’ too much. Be that as it may, the view that welfare must be understood in terms of natural properties, such as the enjoyment of certain mental states, and that good-for accrues only to those properties, squares badly with a ﬁtting

9

Rosati, ‘Objectivism and Relational Good’ (2008).

86

E X A M I N I N G T H E A N A LY S I S

attitude analysis which, like FAP*, involves a normative element. I remain unconvinced that there is a notion of good-for that FAP fails to capture. This notwithstanding, the important thing is that what FAP does denote is at least as interesting a valuekind as the other notions of good-for outlined here.

6.4 Two sorts of personal value In my view FAP* has a strong intuitive appeal as an analysis of personal and not merely relative value. The object x does not only have value relative to some person. It ought to be favoured in a special way; it should be favoured for some person’s sake. To do something (in a general sense of ‘do’) for someone’s sake is to act with an eye to this particular person. A value understood in these terms is more appropriately described as a personal, rather than impersonal, value. Presently, I will point to some other positive features of FAP, but in the meantime let me continue to comment a little more on what we have just discussed. Suppose we take ‘welfare’ to have some kind of descriptive meaning. For example, we assume that a person is in a certain state of welfare if and only if she is in a certain mental state. A great many writers discussing welfare tend to agree that the presence of such a state is at least a condition of welfare. Of course, they also disagree not only over the precise nature of the mental state, but also over whether this condition is sufﬁcient. Now, I have no problem with the claim that welfare is to be understood in terms of mental states. However, if this is how we look upon welfare, it will be an entirely evaluative issue whether we should ascribe value-for to these states. This is consistent with the idea that many of the things we take to be welfare states (and, for that matter, states of well-being)10 are typical examples of something that carries personal values. If I am right about this substantial claim, we might wish, for purely classiﬁcatory reasons, to group together a range of what are typically considered bona ﬁde welfare-making properties of a person a in virtue of which we can reasonably favour x for a’s sake (whether or not the pro-response is ﬁnal or instrumental). The thought that a single, large group of ‘welfare’ states is part of the set of objects carrying personal values seems intuitively plausible. It is, at any rate, very likely that the supervenience bases to which we normally ascribe ‘welfare value’ are such that they supply (or, at least, are believed to supply) us with reason for favouring them for the person’s sake. Personal values, I shall say, consist of two fundamental categories of value, one of which contains those values-for whose realization is welfare-enhancing or welfarepreserving (i.e. such as to enhance or preserve the kinds of state that are typically regarded as constituting welfare). For the sake of illustration, I assume here that ‘welfare’ is not a synonym of ‘good for’, but merely something that might provide

10

Recall that I do not take well-being and welfare to be the same.

6.4

T WO S O RT S O F P E R S O N A L VA L U E

87

one kind of supervenience base for good-for. We need not go into the precise nature of this something, but what I have in mind are objects that contribute to the existence of certain feelings, beliefs, and attitudes of the person whose welfare is being affected. The second category of personal values contains a related, but nonetheless different, kind of value-for. Objects carrying values of this other kind are what we have reason to favour for a’s sake without it being necessarily the case that this favouring will somehow contribute to a’s welfare, even if welfare is conceived in quite a broad sense. Despite the fact that a favouring of x would not have any welfare-contributing properties, values belonging to this category would nonetheless exemplify something that is good for a. We need to dwell a little longer on this last matter, since it might appear somewhat counter-intuitive. Setting aside any substantial reservations we might have, the question we are addressing is this: Is it really possible to favour something for a’s sake that is not in some minimal way welfare-promoting (in this narrower mental-state sense of welfare)? I think it is possible, and I shall argue as much in a moment. However, it might be protested right away that favouring something for someone’s sake simply has to be understood in terms of what makes the person better off (especially if we take favouring to be equivalent to preferring) or what is in the person’s interest. Perhaps this is correct. Again, it depends how we understand these terms. If we take ‘better off’ or ‘interest’ here to be synonyms of ‘value-for’/‘personal value’ understood in terms of FAP*, then, obviously, we should go along with the protest. But otherwise I think it is best to be sceptical. It would surely strain the words ‘better off’ a little too much to say that everything which we have a reason to favour for a person’s sake will make him better off. Indeed a moment of reﬂection should show that it is possible to favour something for a’s sake that is not in the slightest bit welfare-promoting (again, in this narrower mental-state sense of welfare). First, we need to recall that not all attitudes have the same kind of intentional object. Consider the attitude respect. Showing a person respect is not necessarily something one does in order to contribute to this person’s welfare. It might even have welfare-impairing effects. Respect has rather to do with what the person is or represents, and not with what you hope to achieve for the sake of this person’s welfare. Once upon a time it was not uncommon for soldiers to respect their enemies. This did not prevent them from wishing that the welfare of their enemies should come to an abrupt and most deﬁnitive end. Of course, many attitudes like this one are oldfashioned, and some are quite ridiculous. But this fact does not controvert the notion that, conceptually speaking, we can do something for someone’s sake that does not necessarily involve enhancing her welfare. Doing something for a’s sake must, however, in some minimal sense, involve acting in favour of a—in the sense (to return to the example) that being respected is regarded as positive, while being ignored or disrespected counts against a person.

88

E X A M I N I N G T H E A N A LY S I S

It might be argued that we are not respecting a for a’s sake at all. We are rather respecting him for his dignity’s sake (e.g. see Stephen Darwall, Welfare and Rational Care, 2002). Now, it is quite acceptable to say that we respect a for his dignity’s sake as long as this is not taken to imply that the proper object of our attitude is dignity. Surely, what we respect is persons and not properties. It would be odd to argue the latter. Respect belongs to a group of attitudes presenting the starkest counter-examples to the idea that doing something for someone’s sake is necessarily doing it to preserve or promote that person’s welfare. Other examples are: to admire, to honour, to esteem, to like, and even to desire that something is the case. Even a response like desire, which has traditionally been intimately associated with welfare, need not always be understood as welfare-related. For instance, a person may live an immoral (or inauthentic, or irrational) life, and we might well desire that certain states of affairs obtain that would make her life moral (or authentic, or rational) despite any negative impact this might have on her welfare. We would still be favouring the relevant states of affairs for this person’s sake.11 It would be possible to try to counter this point by analytically building into the very notion of welfare something that blocks this eventuality, but this would just land us with a rather implausible and uninteresting notion of welfare. We still lack a full understanding of what the ‘for a’s sake’ idiom comes to. But Chapter 5 showed, I think, that we now have enough to go on to conclude that there is at least a conceptual gap that the complex construction ‘O ought to be favoured for a’s sake’ ﬁlls—even when we are talking about attitudes such as respect.

6.5 Personal intrinsic value Let me next bring out another advantage of FAP, namely that it does not necessarily make personal value conditional on some de facto favouring. (We saw in Chapter 4 that this was an implication of DEFA, i.e. the idea that some x has personal value for person a if and only if we ought to favour x in virtue of the extrinsic property that x is de facto favoured by a.) An object of personal value is one that we ought to favour for a person’s sake, and this might well be the case whether or not anyone actually favours the object for this person’s sake. This point is less counter-intuitive than it may appear to be. Suppose a did not know what her parents looked like; a knew them only by name. Furthermore, assume that there exists a photo of her parents, unknown to her. Such a photo could plausibly be said to carry a personal value for her whether or not

11 Here is another example. A person whom I once knew well lived a very self-destructive life. We eventually lost contact with each other as time passed. Later on, when I learned she had passed away from the after-effects of her destructive lifestyle, I immediately felt sadness. However, I also realized that my mourning of this person’s death had a different character from what I experienced when, for instance, my father died. In the latter case, I was sad for my own sake at having lost someone I cared about greatly. In the former, I cannot sincerely say that I actually cared much for this person. My sadness was rather of the following kind: I was sad for her sake. Her disastrous life called forth a special feeling, sadness, which I experienced for her sake.

6.5

P E R S O N A L I N T R I N S I C VA L U E

89

she knew about it. The point is not, however, that the value of the photo would be only hypothetical—say, conditional on her cherishing the photo. Here is a different argument that suggests that we take it to be possible, at least, for something to have value for a person despite the fact that he or she rejects the idea. Suppose you believe that a certain object x has personal value for your friend a. However, when you show a the item, she rejects the idea that it has this sort of value: x has no value for her, she says. Suppose, further, that a cannot give you an explanation that you are satisﬁed with as to why x is different from other objects of the same kind that do have value for her. In such a case you may draw the conclusion that you were somehow mistaken about what has value for your friend. Alternatively, you may think your friend is mistaken. Both possibilities are open. You may take the latter option because you believe your friend is being inconsistent. But this is not the only possibility. Surely you may also think that she fails to see something in the item that you think you see—something that is such that you believe there is a reason for you to care for the object for her sake. You expect that at some point a will see what you see in the object, and then come to realize that x has in fact personal value for her. Notice, too, that nothing of relevance changes if we add that the attitudes are hypothetical. For example, whether or not the photo has personal value is, on the analysis outlined here, a question of whether there is reason to favour the photo for a’s sake, and the truth or validity of this normative statement is not dependent on the truth of a hypothetical statement. However, suppose a declares, when we draw her attention to the photo, that she could not care less. Suppose further that this is actually true. There is no reason to deny this possibility: a might be absolutely indifferent to the photo. Does it follow that the photo now has no value—of a personal or some other kind? It might of course follow if we adhere to a meta-valuational theory, according to which, say, values are constituted by attitudes (e.g. classical subjectivism, or a more sophisticated non-cognitivism). But theories apart, does it follow that the photo has no value if a is indifferent to it? I do not think so!12 FAP does not assume the truth or falsity of value subjectivism or value objectivism. The deontic statement in the analysans might well be given a subjectivist or an objectivist analysis. I fear that I have repeated this point one time too many, but the fact that FAP accommodates both perspectives is crucial for understanding that, on it, it is possible for personal value to be something other than extrinsic. Of course, to what extent there are intrinsic ﬁnal personal values is, as I have mentioned earlier, in part a substantive issue. The really interesting question is whether there is a notion of intrinsic personal value. Is that notion really coherent? Given the way FAP is formulated it is not obvious that we can rule it out on purely formal grounds. To see why, consider to begin with a new version of FAP:

12 Of course, this is a substantive question (which it would be a pity to rule out on formal grounds). To answer it, we would need to engage in substantive reasoning—something there is no room for here.

90

E X A M I N I N G T H E A N A LY S I S

FAP-intrinsic: An object x’s value for a person a (i.e. x’s personal value) consists in the existence of normative reasons for favouring/disfavouring x in virtue of its internal features alone for a’s sake. As can be seen, there are no straightforward obstacles here. That is, it seems to be psychologically quite possible for a person to take into account only the internal features of an object and favour it for someone’s sake. Suppose, for example, that you have a friend who collects small bronze sculptures. You ﬁnd one of these at an auction. Your friend is unaware of this sculpture’s existence, and he cannot therefore buy it himself. It seems that I could ascribe value-for to such a sculpture in virtue of its internal features alone. That is, I could take myself to have a reason to bid for it for my friend’s sake. My point is not, however, that this is necessarily the most plausible interpretation of the case. It might be objected that we cannot avoid somehow relating the object to the friend. Perhaps this is true, but it is not enough to establish that these relation features turn the object into an extrinsically valuable object. The relational properties might simply not ﬁgure in the supervenience base of the object. Though the example is psychologically plausible, it does not, however, establish that there are intrinsic personal values.13 The bronze sculpture case is obviously about someone who takes certain features to be reason-making, but it establishes that something has been considered a normative reason, not that something is a normative reason. Hence the case does not supply a knock-down proof of the existence of a real normative value. Notwithstanding this, the fact that we seem to recognize such intrinsic personal values is not irrelevant. After all it underlines our openness to the idea of something being intrinsically valuable for some agent. That FAP can be seen as consistent with it is therefore a great plus, I think. But what if we take seriously the suggestion that all reasons are agent-relative? Surely we must then conclude that there are no intrinsic values on the ﬁtting attitude analysis? Perhaps. Certainly, I am inclined to think that this conclusion is correct. The fact that I still can see some small openings for a notion of intrinsic value-for, even if it is accepted that all reasons are agent-relative, is not something that I will go into here; currently my thoughts on this matter are not sufﬁciently worked out to be presentable. The important thing to bear in mind—and I would readily agree that it is easy to forget this—is that ﬁtting attitude analyses do not, in general, analyse value in terms of speciﬁc sets of reason-making properties. The analysis—and this goes for FAP, too—says only this much: to be valuable is to be an object that we have (the right kind of ) reason to favour. In other words, it is not an analysis of the reason-making properties of valuable objects. It is important to keep this in mind, especially in this context. Since the analysis leaves

13

Here I leave out other complications. It might be argued that the example is only about instrumental, rather than intrinsic, values. I would resist this, but to make this resistance really plausible I would need to (and here cannot) present a fuller description of the case. I think this could be done, though.

6.6

A G E N T- R E L AT I V I T Y A N D P R I VAT E VA L U E S

91

questions about the normative driving forces, and the precise nature (objective or subjective) of this normativity, open, it also leaves the question whether there can be intrinsic ﬁnal personal values unanswered. If normative reasons are facts involving the agent for whom they are reasons, it is hard, as I shall argue in Chapter 9, to understand the claim that there can be intrinsic values. Reasons qua facts all appear to be agentrelative. I am inclined to think that reasons are indeed facts of the kind mentioned above. Given this inclination, I think it is difﬁcult to make room for the notion that there can be personal intrinsic values. But I do not think that ﬁtting-attitude analyses in general analyse value in terms of a speciﬁc ‘reason notion’. Rather, they understand value in terms of a general notion of reason or normativity. I am sure many will see this as a weakness. For my own part, I think this is very much its strength.

6.6 Agent-relativity and private values FAP has inherited some serious problems from the ﬁtting attitude pattern of analysis. Most notably, the WKR problem discussed in Chapter 3 is an issue that it would have been nice to have buried once and for all. But FAP faces several problems of its own. I am bound to overlook some of the difﬁculties, but in the remaining part of this chapter and the whole of Chapters 7 and 8, I shall attempt to improve our understanding of the analysis by considering a number of more or less tough responses to it. The ﬁrst concerns a foreseeable reaction to FAP raising some difﬁcult issues. The idea is that FAP runs the risk of making personal values private. Two features of the analysis appear to trigger this kind of Moorean reaction.14 One has to do with what we have just discussed, namely the great likelihood that FAP is formulated in terms of agent-relative reasons. Agent-relativity makes personal value too private: just as there cannot be private truths—propositions that are true for me, but not for anyone else— there cannot be private values. What is more, it makes even so-called impersonal values too private, since they too are most likely to be understood in terms of agent-relative reasons. The second trigger has to do with the fact that personal value is being analysed in terms of so-called FSS attitudes some of which take non-fungible persons as intentional objects (i.e. the persistence of these attitudes is dependent on the identity of the object on which they are directed). This last feature seems to run into a problem with a generally acknowledged characteristic of value, namely universalizability. This second reaction is the more serious of the two. However, I will begin by considering the ﬁrst. Personal values would indeed be strange if value-for/good-for was similar to ‘true for’. We have difﬁculty making sense of ‘true for’, because we read it as referring to

14

See the discussion of G. E. Moore’s objection to ‘good for’ in Section 4.1.

92

E X A M I N I N G T H E A N A LY S I S

something true which, in addition, is only true for a particular person; and that just seems confused. But the value-for expression is not like this. It does not refer to value, period, and require us to understand that to be a value for a particular person. Value-for is a particular variety of value: a relational variety. The fact that this relational value is understood in terms of agent-relative reasons does not alter this picture; even if we were to understand FAP* in terms of agent-relative reasons, this would not turn personal values into items which, by their nature, are inaccessible to persons other than the one whose values they are. As we have seen, it is to be anticipated that an object may be good for person a despite the fact that a has no knowledge of that object. Personal values do not somehow become someone’s private property. The relativization of the reason (in the analysis) does not give the reason-holder ownership of it; nor, for conceptual reasons, is it impossible for anyone else to have a corresponding agentrelative reason. Naturally, this is often quite likely to be the case, as a matter of fact. Still, these comments might not be seen as going to the heart of the issue. Recall the example from earlier about the bronze sculpture that was for sale. Suppose I am at the auction, but that I do not know that my friend collects these artefacts. Here, since I am unaware of my friend’s interest in the sculptures, it seems that I cannot reasonably be said to have a reason for, say, wanting to buy one for my friend’s sake. Must I not therefore conclude that the personal value accruing to the bronze ﬁgure is ‘not there for me’? Must not the conclusion be, therefore, that FAP makes value too private? This is a strange line of reasoning. The scenario described does not show that, because I am unaware of the facts of the case, the reason does not apply to me or someone else. A fact can constitute an agent-relative reason regardless of whether you are aware of it—or so I will assume. Moreover, we should not forget that agent-relative reasons are probably universalizable (see the next section, however). In the case just described, the facts prevent me from knowing that I have a reason in this case. The conclusion to draw, then, is therefore that the facts are such that I am unable to appreciate that a value for the friend accrues to the bronze statuette. There is nothing peculiar about this, nothing that makes values into private entities.

6.7 Personal value and universalizability Let us next return to the second, and tangibly more complex, matter. This concerns the extent to which so-called FSS attitudes can ﬁgure in an analysis of value. A possible response to such a view would set out, I believe, from the following commonly made claim: if two objects differ in value, they must differ as to some universal feature—where a universal feature is one the description of which does not necessitate mention of an individual. This idea is, then, bolstered by the claim that value judgements are universalizable: if we say that an object a is valuable, we are committed to saying, of any other object with the same universal features as a has, that it too is valuable. If, then, universalizability is a requirement on all judgements of value, as Richard Hare has so painstakingly argued over

6.7

P E R S O N A L VA L U E A N D U N I V E R S A L I Z A B I L I T Y

93

the years,15 then an analysis like FAP might appear to run into problems here. However, notice to begin with that these alleged problems have nothing to do with the fact that the notion of reason employed in FAP is likely to be agent-relative. As I have just pointed out, it is generally recognized that such agent-relative reasons are as universalizable as the allegedly neutral ones. The problem has a different origin. It originates in the peculiar nature of the FSS attitude. Speciﬁcally, in the present context, it originates in the peculiar nature of Identity FSS attitudes. As may be recalled, these take non-fungible persons in their intentional content, and this does seem to prevent them from being employed in analysis of a universalizable notion of value.16 That is, suppose x carries person a’s personal value. On FAP*, this means there is a reason to favour x for a’s sake, i.e. there is a reason to direct an FSS attitude on x. Suppose, too, that the value is in fact an extrinsic personal value and is such that there is an a-relational property among the reason-making properties of x. This implies that another object (if there is one) which is exactly similar to x in its universal features might still not carry a value-for. The fact that it did not stand in this relation to a was, so to speak, a sufﬁcient reason for not ascribing value-for to it. But obviously, to reject the implication that if x has value, any exact copy as regards universal features must also have value will not sit well with an advocate of universalizability. A quite obvious reply waits around the corner at this point.17 Since value-for is a relative notion the comparison base will have to include, albeit in universalizable terms, not only the valuable object, but also the person for whom it is a value-for. Thus, suppose I ascribe value-for to an object x with an eye to a. I am now committed, it seems, to ascribing value-for to any other object y with an eye to a* if x and y, and a and a*, share exactly the same universal features. And this does seem to be something that many would agree with. If y and a* stand in the same relation, and are identical, in universal respects, with x and a, then it would not make sense, it seems, to deny that y has value for a*. That is, it would not make sense if you were a believer in universalizability. And perhaps someone with an FSS attitude need not reject this possibility. But suppose she did? She might not even deny that y has value for a*; she might conﬁne herself to claiming that she does not take a stand on this issue, and that no logic in the world can force her to say that y has value for a*. Now, universalizability is a very strong requirement, and it might therefore be argued that it would be far from devastating to discover that FAP is inconsistent with it. It is simply too strong a demand; value analysis can manage with something less demanding: supervenience.

15 For some problems with Hare’s views, see my Logic, Facts, and Representation: An examination of R. M. Hare’s Moral Philosophy (1993). 16 The problem might seem to arise only for those FSS attitudes that have identiﬁers in their intentional content. However, unfortunately this is not the case. The invocation of identity in any sort of justiﬁcation (whether the identiﬁers are internal or external to the attitude) will run into problems with universalizability. 17 This reply is not available to so-called Mooreans, like Donald H. Regan (2004) and Thomas Hurka (1987), who reject the idea of a relativized notion ‘value for’ that can be deﬁned independently of value (period).

94

E X A M I N I N G T H E A N A LY S I S

Although universalizability and supervenience are intimately related, they must not be confused with each other. The latter, in contrast with universalizability, is compatible with unicity, i.e. there being only one object exemplifying a given set of properties. Supervenience requires us to ascribe value to any object y that has the relevant properties of x, and the relevant properties in question may well pick out a single, unique individual such as the being who shared the past with me, gave birth to my children, and so on. So if I ascribe to x value-for having in mind the person who shared the past with me (and so on), value-for is uniquely determined by these external relational properties. One option, at this juncture, is therefore to throw universalizability overboard, and to stick with the idea that values are supervenient features. I do see this as a viable option. My problem is that I am not altogether sure whether universalizability is too strong a requirement. At least, it might well appear that a rather low price attaches to its acceptance once supervenience is agreed to. Surely it cannot be all that difﬁcult to acknowledge that any other object y which has the relevant subjacent features of x is also valuable, i.e. ought, also, to be an object of an FSS attitude directed on a*? True, even a ﬁrm believer in non-fungibility attitudes, if you will pardon the description, ought to agree with this. But the real crux of the matter then becomes whether the next step follows too: does it follow that the person who directed an FSS attitude on x having a in mind must also direct one ony with an eye to a*? Here I feel I have no good argument; that is, I do not think this step is motivated by the logical features of our value notions (alone). Nor do I think rationality (consistency) obliges one to take this step. I am therefore unable to provide an argument to show that the rejection of universalizability is a mistake. The above issue also has a clear bearing on something discussed in Chapter 5. I am thinking here of the distinction, made there, between two sorts of attitude-justiﬁcation: identity- and property-justiﬁcation. I suggested that identity-justiﬁcation might not, after all, really be a sort of justiﬁcation. If you have strong intuitions in favour of universalizability, quite obviously, you will not want to allow the identity of a person to be among an object’s value-making properties, and so you will eventually reject the idea that identityjustiﬁcation is a genuine variety of justiﬁcation. However, if you think invocation of the supervenience base justiﬁes the claim that an object is valuable or explains whether there is reason to favour the object, you will probably agree that identity-justiﬁcation is, after all, a bona ﬁde form of justiﬁcation. I am inclined to subscribe to the latter view. For-someone’s-sake favourings constitute a loose-knit set of very different kinds of attitude. In this chapter I have mainly focused on non-universal varieties, i.e. those that take non-fungible persons as their intentional objects. These are particularly interesting for two reasons. First, they are philosophically more problematic inasmuch as it is not obvious how to make room for such attitudes in a value analysis. However, as I argued, although it may be difﬁcult to analyse value simpliciter in terms of these attitudes, we can include them in an analysis of value-for. Second, non-universal attitudes deserve our attention because they appear to have a great impact on our lives. It is reasonable to look upon them as the kind of attitude that matters more to us: most of us would divert considerable resources to avoiding their frustration or accomplishing their satisfaction.

7 Mo(o)re Objections

So far my main concern has been to pave the way for what I believe is an interesting approach to understanding value-for. The time has now come to have a look at some more or less obvious objections to, and enquiries about, my approach. There are a number of reasons why one might be sceptical about this sort of value, some sounder than others. I will begin with a readily entertained suspicion. Given the long tradition of value theorists who have dedicated their work exclusively to the notion that something is good, period, it seems quite natural to suspect that we do not need to add any ‘personal values’ to our value typology; we can take care of alleged cases of personal value within extant views. Obviously, such a line of reasoning is bound to come in an array of forms. I will conﬁne myself to commenting on only two of these. They have in common that they do not purport to show that we have to throw ‘good for’ overboard: this expression can remain in the lexicon of value. However, they argue that it would be fundamentally confused to think that ‘good for’ is an independent evaluative notion—that is to say, a notion capable of being analysed independently of ‘good’. To believe this is to fall into serious confusion. But before we proceed, it should be stressed that the distinction between good, period, and good-for (and value, period, and value-for) raises questions that I have had to set aside. Perhaps the most important of these centres on an issue for those who claim there is only one kind of fundamental positive value notion, i.e. only one kind of goodness. The issue is then whether this value is goodness simpliciter, or goodness-for? Historically important as this discussion is, I have nonetheless chosen to disregard it. In my view it is basically a misconstrued debate between objectivists and subjectivists about value;1 since good-for is not in any really interesting sense a more subjectivist 1

The idea that there is not one but two independent notions of goodness has been a major topic in moral philosophy. Henry Sidgwick, in his Methods of Ethics (1907 [1874], pp. 120–1, 403–5), famously claimed that the mark of modern ethics, as opposed to ancient Greek philosophy, was precisely the realization that practical reasoning faces a choice; it can help us determine what we should do to secure either maximum person-relative goodness or maximum impersonal goodness. Much of this historical discussion is, in my view, closely tied to the question ‘What is the precise relation between value and people and their attitudes?’

96

MO(O)RE OBJECTIONS

notion than good period, I ﬁnd this discussion less interesting. A more interesting position rejects both these “monist” views in favour of a view embracing good simpliciter and good-for; both are perfectly meaningful value notions, and neither therefore should be dismissed. Since I do not think we should give up on the notion of good period, I will be concerned chieﬂy with arguments against good-for.

7.1 Hurka’s views on good for Various philosophers have, over the years, joined ranks with G. E. Moore in attacking the view that good-for is a normative notion on its own, one that is not eventually understandable in terms of the notion that something is good, period. In particular, Thomas Hurka (1987), and more recently Donald H. Regan (2004), have launched objections to certain ways of understanding good-for to which I suspect Moore would have consented. Their criticisms are in many ways quite understandable, and in fact I share many of their misgivings. But, like Moore, they have shown only that there are ways of understanding good-for that are confused. An analysis like the one I propose is, as far as I can see, untouched by their attacks. Obviously, one reason for this is that there is chance, as mentioned earlier, that what I am analysing has little in common with the concept that Hurka and Regan attack (and Connie Rosati defends: see below). This might be the case. I shall return to this matter in due course. However, since the examples these authors discuss all appear to involve what are personal values in my sense, it will be wise to see whether their objections have any real bearing on what I am proposing.2 Hurka identiﬁes four meanings of ‘good for’, three of which he thinks we should get rid of for the sake of clear thinking. These are rehearsed below; only (iii), which captures what might be properly referred to as the Moorean sense of ‘good for’, meets with his approval: (i) “ . . . ‘good for’ means satisﬁes the desires of, or furthers the interests of . . . [or] answers (somehow) to the subjective states of ” (p. 72); (ii) ‘Good for’ can sometimes be used to mean what a person believes good; (iii) “ ‘Good for’ can also be used to mean good from the point of view of ”; (iv) “‘Good for’ can also mean that portion of the good (from whatever point of view) that falls within a person’s own life”.3 Hurka’s objections to (i) and (ii) seem sound to me. Meaning (i) fails, in his view, because the deﬁning claims are descriptive, and we cannot account for the normativity

2 In other words, the assessment of alternative views on good-for from my perspective is not as straightforward as it might initially seem. Recall what I said at the outset of Chapter 6: that there are ways of understanding good-for which, in comparison with the idea I am analysing, involve a narrower notion. 3 Hurka, ‘“Good” and “Good for”’ (1987). Hurka’s main concern is to attack the position that “Good [ . . . ] is always good for someone” (p. 71). This is not a position that I defend.

7.1

HURKA’S VIEWS ON GOOD FOR

97

of ‘good’ with these senses. Of course, if you want to be picky, the strength of Hurka’s point is not that clear-cut. Why should an advocate of (i) feel vulnerable because the deﬁniens in (i) does not live up to the normativity of another notion that it does not try to deﬁne, namely ‘good’? It is not ‘good’ that is at issue here, but ‘good for’. It is not obvious how Hurka would get around this, since he does not want to acknowledge any independent good-for notion. However, once we accept that not only ‘good’ but ‘good for’ is normative—which I think we should—(i) fails to account for this normativity, since it reduces good-for to a mere psychological state. It should be stressed that Hurka’s own (1987) account of impersonal good places him among buck-passers: ‘good’ is reduced to what we ought to morally desire and pursue.4 So his claim that the senses given in (i) are insufﬁciently normative is quite understandable. The meaning described in (ii) is, I believe, fairly common: what is good for a is what a believes to be good. Whether or not we ﬁnd this sort of relativism appealing or plain appalling, it clearly has nothing to do with what FAP is intended to disclose, so it need not detain us. It is a sense that acknowledges only one kind of goodness, namely ‘good’, not the relational ‘good for’. That is, it understands the expression ‘good for’ in terms of what is good, period, for some person. As to (iii), if we are equating ‘good for’ with ‘good from the point of view of ’, we should probably, for the sake of clarity, be speaking about ‘good from a point of view’ to begin with. Hurka reduces this latter expression to what we ought to desire or pursue. In other words, sometimes when we are speaking about good-for we have impersonal goodness in mind. Finally, Hurka has a similar reply to (iv)—the notion that ‘good for’ means that “portion of the good (from whatever point of view) that falls within a person’s own life”. If this is what we mean, there is no need to allow that we have an independent notion of good-for. It would make for greater clarity if we were to speak instead of what is “good in her life”, where ‘good’, here, is understood as impersonal good. Hurka does not have much to say about this latter sense; but he is not alone in thinking along these lines. Regan has more recently argued that this latter sense is the only sensible thing we can have in mind when we speak about something being good for a person. Michael J. Zimmerman (2009) calls this kind of sense “an ownership account of goodness-for”. He thinks there are two versions of it—“‘x is good for P’ = def. x is good and belongs to P” and “ ‘x is good for P’ = def. x’s belonging to P is good”— neither of which he thinks corresponds to what people mean today by good-for. Regan defends the former version, so here I will only be concerned with this sense. I will return to Zimmerman’s own favoured sense shortly. However, let us ﬁrst consider Regan’s approach.

4 Hurka recently expressed during a visit to Lund (personal communication) doubts about the buckpassing account, and so I want to emphasize that what I discuss here are his views from (1987).

98

MO(O)RE OBJECTIONS

7.2 Regan’s views on good for Regan (2004) contrasts good-for with good simpliciter. That something is good simpliciter gives us an agent-neutral reason to promote it, according to Regan. Regan also assumes that good-for must be understood as a source of agent-relative reasons. (Notice by the way that this is not necessarily a ﬁtting-attitude analysis; a number of theories might accord reason-making status to moral features.) In his view good-for combines two inconsistent ideas, one about goodness and another about some inexplicable ‘for’; and the ﬁrst of these does not capture anything of interest over and above Moore’s idea of ‘good possessed by a person’. As might be recalled (see Chapter 4, note 10), Regan refers to this Moorean notion as a “good occurring in a person’s life” (cf. Hurka, 1987). He also proposes an argument against the usefulness of good-for—if it is taken to be independent of this Moorean good. By lack of ‘usefulness’ he seems to have in mind the fact that such an independent notion can make no claim on persons other than the individual whose good it is. Once this worry was my own,5 and so I am sympathetic to it. An analysis of good-for must avoid turning this kind of value into something entirely private—a goodness that is only available to (or ‘there’ for) the agent whose good it is.6 Why cannot good-for be a genuinely independent relational goodness? Regan’s idea is that at least one of the incompatible claims made by the good-for advocate has to be jettisoned: suppose, for instance, that the good-for theorist agreed with the Mooreans that ‘good for’ only denotes ‘good in the life of’. The advantage derived from accepting this, according to Regan, would be that the good-for theorist could now claim that what is good for a person gives us an agent-neutral reason. If some x is good for a in the sense that it is a “good in the life of a” anyone has a reason to care about and promote x for a. Of course, to accept this would be to acknowledge that ‘being good for’ is no longer a distinct normative property. However, on the other horn of the dilemma, the goodfor theorist could be obliged to give up the claim that she is, in effect, talking about a ‘good occurring in the life’ of the person whose goodness it is. But then the problem, according to Regan, is that good-for is no longer a normative property for everyone. Regan’s latter worry is exaggerated, though. It would indeed be problematic if ‘good for’ came out as a property which, in its nature, only carried normative force for a single individual; and certainly if good-for was somehow rigged to a particular person not in virtue of, say, the circumstances obtaining, but because the relevant goodness was in some way uniquely someone’s value property, I would join Regan in thinking that such an entity would be difﬁcult to comprehend. But FAP does not give rise to such worries. As a ﬁtting attitude analysis it understands goodness in terms of normativity. This fact also soothes the privacy worry. The fact that something that is good for 5

See Rønnow-Rasmussen (2007a; 2008a). But notice: the fact that something that is good for me requires me to be mentioned in its supervenience base does not necessarily make the value private; since FAP is formulated in terms of reasons that apply to more than the person whose good-for is in question, the privacy argument has no bearing here. 6

7.3

RO S AT I A N D G O O D O C C U R R I N G I N A L I F E

99

me probably requires me to be mentioned in its supervenience base does not necessarily make the value private; since FAP is formulated in terms of the existence of a reason which, in the circumstances, may, and probably will, apply to people other than the person whose good-for it is (most reasons being, like this one, such as to apply to many persons), the privacy argument has no bearing.7 In fairness, Regan, whose position renders problematic a number of views on welfare/well-being, does not consider the kind of view outlined in this work. He does consider Michael Smith’s theory on relational good, and he gives reasons for setting it aside that would probably also apply to my approach. I will return to these reasons in a moment. Meanwhile, I wish to draw attention to what Regan thinks ought to replace the ‘good for’ idiom (at least if good-for is understood in terms of welfare). This is relevant to my analysis regardless of how closely the good-for Regan is attacking coincides with my analysandum. From my point of view, the really interesting feature of Regan’s account is actually something different—namely, whether the job done by ‘good occurring in the life of a person’ is what we take ‘good for’ to be doing. According to Regan, the former, Moorean notion is preferable in that it suggests a way of understanding the (in his view) problematic ‘for’ in ‘good for’: “If ‘for’ means ‘occurring in the life of’ this gives us an empirical relativization to the agent, but the normativity involved is still the universal normativity of ‘good’”. But, again, since normativity is built into the very notion of good-for, Regan’s worry here need not be shared by the ﬁtting-attitude analyst.8

7.3 Rosati and good occurring in a life Rosati has recently questioned Regan’s suggestion that we should replace ‘good for’ with ‘good occurring in a life’, or GOL. I support her conclusions, albeit, I suspect, for somewhat different reasons. She is, I think, quite right to maintain that GOL is problematic in its own terms, and that none of the more reasonable ways of spelling out what ‘occurring in a life’ refers to satisfactorily replaces the notion good-for. Rosati begins by asking whether GOL should be understood as stating the following (where ‘TP’ refers to time and place): GOLTP: “Good occurs in the life of P when good occurs in the time and place in which P lives.” (2008, p. 334) 7

See also Zimmerman (2009), who suggests we can dispose of the dilemma posed by Regan by drawing a distinction between prudential reasons and moral reasons. 8 Regan seems to be saying that if the normativity of good-for is not relativized, good-for coincides with good. Rosati questions this, claiming that the normativity of good-for need not be relativized. It is still possible to argue that good-for is distinguishable from good simpliciter. Rosati’s point is that there are many things that are universally normative (e.g. rights, rationality) but which nonetheless differ from good (see 2008, p. 329). In Rønnow-Rasmussen (2007a) I maintain something to this effect. However, if normative reasons (at least, in an interesting sense) are all agent-relative, as I shall argue in Chapter 9, Regan has no point in the ﬁrst place.

100

MO(O)RE OBJECTIONS

The occurrence of goodness is, in other words, understood as a spatio-temporal person-related event. I suspect Mooreans would not be happy with this proposal. However, since I am uncertain how it might be improved, I will treat it as a candidate. According to Rosati, GOLTP may well lead to some rather counter-intuitive suggestions about what we ought to do: If good occurring in a life required only spatiotemporal coincidence, then individuals could sleep through lives in which a good deal of good occurs and this would have peculiar results for what we ought to do. We ought, other things equal, to redecorate meticulously the rooms of the permanently comatose, pipe beautiful music into their rooms, send in the clowns. To be sure, we will promote more good by expending our energies elsewhere. But the suggestion that we could have any reason to promote good occurring in the lives of the permanently comatose, at least in this sense, is dubious at best. If this were all the Moorean had in mind, then I think we could reasonably doubt that good occurring in a life makes a claim on just as much as we can doubt that the self ﬂagellator’s Sumnerian good for makes a claim on us. Good occurring in the life of P, in the bare sense sketched by TP, does not capture what the good-for theorist has in mind. (2008, pp. 334, 335)

Rosati might ﬁnd the idea of a good occurring in a comatose person’s life dubious for reasons other than my own; I am not quite sure. Her idea seems to be that no point can be found here—nothing can come from making goodness occur in the life of someone who cannot appreciate it. Perhaps the idea is rather that the alleged values mentioned, like beauty, are such that they cannot play any role in the life of a comatose person, and hence that adding such value cannot plausibly be said to be good for the person. If the comatose individual cannot use her senses, what is the point in adding value objects that require to be sensed for their appreciation. This reasoning makes sense on many plausible substantive accounts of value, but I cannot be certain Rosati subscribes to it.9 I am also inclined to think that a defender of GOLTP need not be that worried about it; but that is because I think the notion of good-for is consistent with the idea that the promotion of what is good for a person need not have any impact on the person. Needless to say, this is a contestable claim, but, as I have mentioned, fundamentally the issue is an evaluative one, and not a matter that can be settled with the ﬂick of a conceptual lever. This notwithstanding, I am not impressed by GOLTP, and I share Rosati’s conclusion that it cannot replace good-for. Let me explain. Imagine two separate, beautiful rooms ﬁlled with the same beautiful music, each containing a comatose patient. (I am afraid many readers will think this example is too strained. How can a silent room have any value for a comatose person? This is certainly a reasonable question. Perhaps the correct reply is that it cannot, the reason being that comatose people cannot have personal values—or cannot, at least, have personal values 9

Is she perhaps thinking that nothing can be good for a comatose individual? Rosati might be read as claiming, for instance, that only conscious people can have a good-for; but it is clear, I think, as we shall see in a moment, that this is not what she believes.

7.3

RO S AT I A N D G O O D O C C U R R I N G I N A L I F E

101

of the kind accruing to objects detectable through the senses. However, it is also possible that none of the objects that carry value are (in some way) capable of sensory detection. Respecting a dead or comatose person’s wish might be like this. So although it is somewhat strained, I think the example serves a purpose.) It seems to me fairly obvious that we need a notion of good-for that allows us sensibly to claim that we should leave patient a in his room, but let b move to a silent room, since she has all her life hated music. Silence has a value for patient b. It might be contended that Mooreans like Regan can acknowledge this. They can say that we should move b because she does not appreciate certain values. For sure, this is certainly a possibility. Here we must proceed carefully, though. We must not overlook something fairly obvious. We might justify moving b out of the room by appealing to the fact the b had certain desires in the past. However, whether or not we believe that this involves, at some level, taking an evaluative stand, I think we should (also) be able to evaluatively motivate our decision to move b. To do this, however, at least as far I can see, we require an agent-relative conception of goodness/badness. That is, we need to be able to appeal to something more than the non-evaluative fact I take ‘b as a matter of fact did not appreciate certain values’ to express. We must appeal to what is good for or bad for the patient; not what is good or bad according to the agent. Again, I think we should be able to say that remaining in her room is not good for b, or that the sound of music in the room has a negative value for b. I would move her or stop the music, and I would do so for her sake. Notice, too, that nothing here hinges on the substantive position we take on what is of ﬁnal value. In the example it is beauty, but we could pick more or less anything that is good, period. It still makes sense that we should be able to differentiate the two agents’ presence in these rooms, and to do so with an eye to what we should do for b’s sake (rather than a’s). That is, we need to be able to evaluatively express the relationship between an agent and what is good, period. Good-for, as handled in FAP, is in my view a serious shot at this. The invocation of merely impersonal values (i.e. of good simpliciter) would not explain why we should either ﬁnd b a new room or turn off the music in the room b is currently in. How could it? After all, non-relational values are the same for everyone, so they cannot be employed in a case like this. A possible rejoinder at this point is to suggest there is negative value in b’s room that is not present in a’s room, and that is why we should move b but not a. But suppose we imagine both patients in the same room. Would we not be obliged to say, in such a case, and given GOLTP, that this negative value is present in a’s life as well as in b’s? It seems so. But in that case nothing is really gained by introducing a negative value. We would still need, I believe, to ﬁnd a value that is particularly related to b in order to justify treating b differently from a. Ultimately I am not quite sure what to say about the coma case. It seems quite clear that Mooreans cannot reasonably have GOLTP in mind when they think of a good

102

MO(O)RE OBJECTIONS

occurring in someone’s life; it simply does not sound like a convincing account of what might be understood by that phrase. Before moving on, let me therefore voice a possibility that should not be readily dismissed, namely that we understand ‘good occurring in a life’ in terms of FAP*. Perhaps ‘good’ and ‘good occurring in the life of someone’ are two quite different notions of value (such that ‘good occurring in a life’ is a variant of Moorean good simpliciter). ‘Good’ would refer to a monadic property, while ‘good-occurring in the life of p’ would refer to a dyadic property. So far we have merely assumed that the notion of goodness picked out by the expression in ‘good occurring in a life’ is the same notion as Moorean good. But if they are different we need a better way of understanding the former notion than that provided by GOLTP. The sort of analysis I am outlining in this book might, in effect, be an interesting way of understanding ‘good occurring in a life’. An adjusted FAP* might do the necessary work here. The idea would be to look for reasons for favouring something for the person’s sake—facts that have something to do with the person’s (whole) life. It would take some time to set out such an approach in full, and I will not try my hand at this here. In fact the case would clearly be a special case, subsumable under FAP*, and I prefer to focus on the more general pattern. The idea sounds worth pursuing, though. Interestingly, Sidgwick in The Methods of Ethics (1907 [1874]) put forward a ﬁttingattitude (FA) analysis of “my own good” and good-for that Moore might well have accepted. Hurka points this out in “Moore in the Middle”:10 By “my own good” and sometimes “good for me” Sidgwick meant what I ought to desire “assuming my own existence alone to be considered”. This should not have been objectionable to Moore: it deﬁnes my own good as that portion of the good, however that is understood, that is located in me. (2003, p. 611)

Sidgwick’s proposal is interesting. The problem is that it is not obvious how one should understand “assuming my own existence alone to be considered”. But perhaps we could take it as implying that something is good for me, if and only if it is what I ought to desire for my own sake, in which case FAP might not have been objectionable to Sidgwick. Rosati considers another interpretation of GOL. We can label this GOLTPC: Perhaps we should say . . . that good occurs in the life of P when good occurs in the time and place in which P lives and P is conscious of the occurrence. Call this characterization of good occurring in the life of the time-place-consciousness characterization, or ‘TPC’. (2008, p. 336)

Rosati also dismisses this suggestion. Her reason is interesting: If all that good-for talk were about was good occurring in the life of as prescribed by TPC, goodfor theorists who do not accept a consciousness requirement would not be talking about

10

Hurka, ‘Moore in the Middle’ (2003).

7.3

RO S AT I A N D G O O D O C C U R R I N G I N A L I F E

103

good for, despite what they may think, These good-for theorists could be mistaken, but I do not think their position should be ruled out prior to our efforts to advance and assess analyses of good for P. (2008, pp. 336, 337)

Rosati has a point here. The discussion between mental-state theorists of good-for and their opponents is, I would say, a substantive discussion; it concerns what ‘good for’ judgements we ought (and ought not) to make. Before moving on to problems concerning Rosati’s own position, I want brieﬂy to comment on Regan’s reason for setting Smith’s views11 on relational good aside in his discussion of good-for theorists.12 Regan, who thinks good is the pleasurable experience of objectively appropriate activities, considers different ways of understanding good-for that purport to retain its normativity. Like Hurka, he dismisses various naturalist positions, all of which, he thinks, fail to secure the necessary normativity. However, considering the example ‘good for Brown’, he believes another approach is possible: . . . we might take it to mean just ‘what Brown ought to promote’. The most interesting development of this idea is Michael Smith’s, although Smith sensibly prefers not to refer to his concept as a version of ‘good for’ but rather as a subscripted ‘goodagent’. Interesting though this possibility is, it is clearly remote from any concept of welfare. The goodBrown in this sense may include any number of states of affairs outside Brown’s life; and furthermore there is no plausibility at all in suggesting that the goodBrown as such makes any claim on Jones. It may be, of course, that the goodBrown and the goodJones overlap, or are even identical in extension, but it will not be the goodnessBrown, the fact that Brown should promote it, that accounts for anything’s being goodJones. So we can ignore this subscripted good. ‘Good for’, as used in this essay, is a normative concept, and a synonym for the other normative concepts of ‘welfare’, ‘well-being’, and ‘good of’. (2004, p. 208)

It is a pity Regan does not explain in greater detail why Smith’s subscripted goodness should be ignored in discussion of good-for and welfare. First, it does not strike me as necessarily sensible to disassociate this sense from good-for. Surely it is possible that this subscripted sense of ‘good’ is just what welfare theorists apply to welfare, or, to be more speciﬁc, what they apply to whatever it is they take to be constitutive of personal welfare; and quite why they do not apply it to something else is in part a substantive Smith, ‘Neutral and Relative Value after Moore’ (2003). As Rosati points out, in a footnote Regan seeks to clarify his view: “It may seem that by restricting the good to pleasurable experiences, the Moorean reveals that his real concern, like the ‘good for’ theorist’s, is with value for the subjects of these experiences. Why else should an experience of a beautiful sunset, or a great mathematical theorem, or a Bach cantata, need to be enjoyable in order to be valuable? I cannot answer the point fully here. The short answer is that what is really valuable (non-relatively) is the appreciative engagement of the subject with a worthy object.” A few lines later Regan adds: “To my mind, when there is the right sort of engagement, we could as well say that the value created is value ‘for’ the sunset, or the theorem, or the cantata, as insist that the value is ‘for’ the subject” (2004, p. 221). Now, this is an interesting point, but it is not necessarily problematic for the advocate of FAP. Fundamentally there is nothing in the kind of attitude required by FAP that necessarily prevents us from having such attitudes (mutatis mutandis) to objects, such as sunsets and theorems, as opposed to people. 11

12

104

MO(O)RE OBJECTIONS

issue. The fact that goodagent may cover a “number of states of affairs outside Brown’s life” (and, I would add, concrete things) is not a reason for saying that it is a different concept from the one good-for theorists talk about. It might be—I have admitted this. However, it just might be that people are (merely) evaluatively disagreeing about what carries goodness-for. (Consider a slightly different case, one involving value, period. The fact that some people want to say that this kind of value only attaches to pleasure does not necessarily mean they cannot disagree with a preferentialist about what is valuable.)13

7.4 Personal and impersonal value Let me now say something about Rosati’s views. She has in recent years produced several perceptive papers on good-for. As she notices herself, my notion of personal value differs from her view of personal good or welfare.14 In fact I suspect our perspective on good-for is quite different. I suspect, for instance, that what I take to be the desiderata of an analysis of personal good/value are not what she has in mind, and vice versa. But I will not pursue this comparison here; instead I refer the reader to Rosati’s own illuminating work. A couple of differences should be underlined, though, since this will help to highlight certain features of FAP. Here is how Rosati describes certain key features of good-for: Good-for value is relational, not relativist value. It is objective in that it gives anyone a reason for action, though the good for might . . . be otherwise subjective; perhaps the normative relation good for itself involves a mental-state requirement, or perhaps the relational property being good for P is instantiated only by items that produce particular mental states. Finally, good for is a form of extrinsic rather than intrinsic value. (2008, pp. 344–5)

Despite certain terminological discrepancies, I ﬁnd much to agree with here. Perhaps most importantly, I think Rosati is right in keeping a mental state requirement out of the analysis. I have some queries too, though. As I suggest in Chapter 9, I am not at all sure that it really makes sense to say that good-for gives anyone a reason for action. I have also voiced my suspicion that good-for may be an intrinsic value. Now, it must be said that by ‘extrinsic’ Rosati has something in mind other than the meaning I have when I use the term. She relies on Christine Korsgaard’s distinction between two ways something can have value—a distinction we considered in Sections 1.5 and 1.9. ‘Extrinsic’, then, is to be understood in terms of where the source of the value is located. Perhaps it is with regard to this matter that our views more obviously deviate. On Rosati’s account, good-for is a derivative value. Thus something’s being good-for derives from “its relation to a being with value” (2008, p. 343), and more speciﬁcally

13

Regan’s worry, expressed at the end of the passage quoted, about goodagent not being sufﬁciently normative need not concern an advocate of FAP (not even if reasons are agent-relative). 14 See Rosati (2008, p. 318).

7.5

THIN-THIN AND THIN-THICK CONCEPTIONS

105

the being whose good-for it is. This means that if a being has no value, nothing can be good for it (or, alternatively, nothing can be considered good for it). Rosati also thinks this has a bearing on the sorts of value judgement we can make: The property of being good for a person . . . well matches the normative role of judgments about what is good for a person. Whether from the ﬁrst-person or third-person standpoint, our focus in making these judgments is on the person herself and what preserves and advances her, presupposing her value and attending to the particulars of her nature and circumstances. (2008, p. 346)

If I read this correctly, the idea is that someone who asserts, say, ‘x is good for a’ cannot at the same time deny that a has a value, period; to do so would be either to misuse ‘good for’ or to mean something by ‘good for’ other than what Rosati has in mind. But I fail to see why good-for has to derive in this sense from impersonal goodness. It is my belief that, often, when we ascribe personal values to someone, we also regard this person as valuable. But even if it were true that we always regard people as valuable, we would still need to provide an argument for the claim that good-for necessarily derives from this impersonal value. Rosati suggests the following approximation of what we mean when we say that something is good for someone. When we say: . . . X is good for P, we indicate that there is a reason to promote X with P as the beneﬁciary of the action in light of or out of regard for the value of P. (2008, p. 344)

Now it strikes me that Rosati here points out just one possibility; surely quite a number of reasons, or considerations “in light of or out of regard for”, might be invoked when we say that something is good for P, especially if we have ﬁnal goodfor in mind. For instance, why must the value be impersonal? Since I suspect Rosati thinks that we all have this impersonal value qua persons, such value cannot explain why we “should promote X with P as the beneﬁciary” rather than Q as the beneﬁciary. The real issue is a different one, however. Why must the promotion of X be out of a regard for a value? To put the question in terms of judgement, why do we have to evaluate P in the ﬁrst place? Is it not sufﬁcient, for example, that we like the person, love him or her, or some such thing? Of course, we might make an analytic connection, as certain naturalists would want, between the propositions ‘I have a pro-attitude towards P’ and ‘I evaluate P’. However, I doubt that Rosati would agree that her own analysis is, even in effect, naturalistic in this sense. Be that as it may, I think it is a mistake to build into the analysis a certain kind of metaethical position. Of course, I might be wrong about this.

7.5 Thin-thin and thin-thick conceptions Now that we have seen some examples of good-for analyses, let me address an issue that has concerned me while reading not only Hurka, Regan, and Rosati, but a

106

MO(O)RE OBJECTIONS

number of other philosophers who have expressed views on what is good for us (or how we should understand welfare or well-being). The issue can be put thus: To what extent, if at all, am I and these other writers discussing the same notion of good-for? As I pointed out when we touched upon this issue earlier, there is a chance that we are dealing with different notions. However, since I also believe there is a chance that we are concerned with one and the same notion, let me explain why I think this just might be the case. One way of regarding the relation between good-for and welfare is to say that goodfor can accrue to great many different things, including welfare. The kind of thing it will accrue to is not, in the end, something the analysis of good-for should determine (anymore than an analysis of good should determine which things are in fact good); and since the analysis does not specify the supervenience base of good-for, it may accrue to a wide variety of objects. As an upshot of this, people may disagree about whether welfare is good for a or not—and when they do so they are disagreeing about whether welfare is a supervenience base of good-for. In other words this approach does not treat ‘welfare’ as a synonym of ‘good for’. Perhaps this is too counter-intuitive to imagine. Suppose therefore that there is a tight relation of synonymy between the two notions. But such a relation is not straightforward (and not merely because of familiar problems concerning synonymy and analyticity). On one account, the relation corresponds to what I earlier referred to as the thin-thin relation. Just as ‘good for’ expresses a thin notion, so does ‘welfare’.15 However, one might agree that there is this tight relation between ‘good-for’ and ‘welfare’, but nonetheless deny that the synonymy relation holds between two thin notions. It is probably welfare, then, that will be regarded as a thick notion—I mean, a notion with a ﬁxed descriptive (non-evaluative) content. I suspect that at least some of the philosophers I have been considering here conceive of the relation in this way. But whether or not this is the case, I think there is reason to question such a relation. (I suppose one might deny that welfare is a thick evaluative notion on the grounds that it is not an evaluative notion at all. Welfare would then be a purely descriptive notion and open to a naturalistic analysis. I shall set this position aside; I doubt it would be endorsed by any of the philosophers I am engaging with.) We would then have two possibilities, each of which seems to be salient in the present context: Now, I think (1) is more plausible than (2), and that neither is as good as the approach I am taking—which is to regard the welfare as just one candidate as what is valuable for someone, and hence not to treat ‘welfare’ as a synonym of ‘good-for’. Before I explain why we should be cautious about (1) and (2), I want to enter the following proviso. If welfare were indeed a thin value notion, I would have little

15 Zimmerman (2009) is ready to accept such a synonymy relation, and he is not particularly troubled by its being trivial or circular. Whether he conceives of this relation as thin-thin or thin-thick is not clear.

7.5

THIN-THIN AND THIN-THICK CONCEPTIONS

Whatever is good for person x is constitutive of x’s welfare, and whatever is constitutive of x’s welfare is good for x

107

(1) Thin-thin relation? (2) Thin-thick relation?

Figure 7.1

objection to the thin-thin approach. For one thing, it would have the advantage that Rosati, Regan, and I might after all be discussing the same notion. The approach would not allow for much in the way of analysis, of course, but my real problem with it is that I am not convinced that welfare is such a thin notion. The notion of welfare seems, at least, to be thicker than good-for. On the other hand, (2), the thin-thick synonymy relation, strikes me as implausible, because I do not see how a thin and a thick notion could be synonymous. Rosati and Zimmerman, and many others, seem to treat ‘good for’ as a synonym of ‘welfare’. It is merely a verbal matter whether we use one or the other term. So whatever we take to be the good-for making properties in a speciﬁc case are also, analytically, welfare-making properties in that case, and vice versa. But notice that, in acknowledging this, we neither agree that the notion of good-for is analytically related to what we express by a certain non-evaluative term nor accept that welfare is so connected. We might simply have the thin-thin relation in mind. As mentioned above, Rosati makes a sharp distinction between ‘good for’ when it is used about a person’s welfare and the other senses this expression has in statements like ‘water is good for the plant’. This strongly suggests that she has in mind a narrower notion than the one I am trying to capture with FAP. If Rosati treats welfare and good-for in accordance with (1), the differences between her position and mine are substantive. We disagree over the good-for/welfare making properties. However, we are probably concerned with the same notion. However, if she treats welfare as a thick notion, or at least as a notion thicker than good-for, which she in turn regards as a thin notion, we do not have the same notion in mind. Could she not regard the relation between good-for and welfare as one obtaining between two thick notions? Perhaps. However, this strikes me as no less implausible than the idea that the synonymy relation obtains between a thin and a thick notion. Good, as well as good-for, are in my view bona ﬁde thin notions. However, I confess that I have no compelling argument for this view—other than to say that the idea that good-for has more in common with good than with various kinds of thick notion, such as admirability, strikes me as the more credible view.

108

MO(O)RE OBJECTIONS

What I have said here applies, I think, mutatis mutandis, to Zimmerman’s (2009) suggestion that good-for should be understood in terms of what beneﬁts the person. Is the relation between these notions of the thin-thin kind, though? If it is, the analysis looks acceptable; we should be able to accept it. However, the notion that something beneﬁts someone seems to me to be a thick notion, and if this is right, we have an implausible synonymy relation.

8 Problems and Possibilities

Connie Rosati’s view that good-for is a derivative value is shared by many authors, but it is not an element of FAP. In FAP the core idea is that the personal value of an object x for a person a consists in there being normative reasons to favour x for a’s sake. It is to be expected that a will often (perhaps always) carry, in the eyes of the ‘evaluator’, some impersonal value. However, in my view this arises from the fact that most people, most of the time, are ready to ascribe some value to most other people. Whether or not I am right to make this empirical claim about what people in fact evaluate, I wish to make it clear that I am not assuming that personal values are necessarily derivative values. I want to turn, in the section below, to a feature of FAP that appears to be worth expanding. Later, in Section 8.2, I will respond to some objections that have been raised to my account. FAP seems vulnerable to various kinds of counter-example. For instance, if values are understood in terms of the existence of agent-relative reasons, FAP seems to imply interpersonal value incompatibility (i.e. that persons a and b et al. will have values that are incompatible). Another kind of case concentrates on the apparent fact that FAP obliges us to care for people that we dislike. The general idea goes something like this: it is plausible to assume that people we dislike have personal values; according to FAP we therefore have reason to favour things for someone’s sake that we in effect do not care at all about, which is counter-intuitive. In this connection we shall have occasion to look at Stephen Darwall’s insightful approach to welfare in Section 8.3. His analysis appears to handle these more or less problematic cases better than my own. There are also some more detailed objections to my account that need to be addressed. I do this in Sections 8.4 and 8.5. Finally, I will suggest in Section 8.6 what I would like a second volume on personal values to be about. Pace the Tractarian Wittgenstein, there is always unﬁnished business to attend to in philosophy.

110

P RO B L E M S A N D P O S S I B I L I T I E S

8.1 Janus values So far the idea has been that objects of personal value should be understood in terms of there being a reason to take up some attitude vis-a`-vis those objects. But the question remains whether some personal values might not have more than one side—one facing the person a for whom the value is ‘personal’ and another facing the rest of us (at any rate, those of us who have a reason to favour the object for a’s sake). This issue seems to be pointed up by the fact that the attitude that it is ﬁtting to take to an object such as a drawing can vary depending on whether that drawing is, say, your daughter’s or your own. Perhaps the attitude can also vary with regard to the strength of the reason. Some reasons appear to have a stronger normative force than others, and it is perhaps conceivable that one and the same object is Janus-like, also, in this regard. This strikes me as less likely, though. The ﬁrst possibility seems more worthy of examination. I ﬁnd the bare idea that personal values might (in the ﬁrst sense) be Janus-values intellectually inspiring. Moreover, once one starts thinking along these lines, the question arises whether there might be attitudes that everyone except a (the person whose personal value we are considering) ought to take towards some object. There might be attitudes/emotions that we may direct onto, or have in relation to, others but not direct onto, or have in relation to, ourselves. Perhaps there are even attitudes/ emotions for which only the attitude holder can be the proper object of the attitude. The following may be an example of this: it might be (morally) unbecoming, but it is nonetheless the case that many people sometimes feel pride about something they have done. However, although people often say that they are proud of what others, such as their children, have achieved, or what their country or favourite soccer team has accomplished, it is not obvious in what sense one can be proud of something that someone else has done. It is like taking credit for another’s accomplishments, and although we often seem to try to do it, the notion that we can do so can be queried on conceptual grounds. As to the former possibility, promises are often presented as examples of attitudes that we cannot direct on ourselves. Certainly, one can say to oneself ‘I promise to stop smoking’, but such utterances are often said to express an avowal of some kind (say, stating your intention) rather than something that creates a promissory obligation. To promise is to promise someone other than oneself; otherwise it is not a real promise. These two examples show, perhaps, that personal values can be ‘double-sided’. On the other hand, if some personal values are Janus-like in this way, I strongly suspect that they are quite rare. The examples are hardly beyond question, and since I cannot come up with any more convincing cases, it is perhaps best to continue to regard this doublesided feature as a mere possibility. FA versions formulated in terms of, not reasons, but ﬁttingness may be better suited to account for this Janus-like character of value. Thus it might be ﬁtting to feel pride for something one has done, but quite unﬁtting to feel pride for something someone else has done (especially if you are unrelated to the person of whose actions you are proud).

8.2

SOME COUNTER-EXAMPLES

111

It might be objected that I am overlooking a problem here, and that this problem jeopardizes the whole analysis. Suppose therefore that there are attitudes (not necessarily of feeling pride in another’s accomplishments or promising oneself something) that we can have vis-a`-vis an object x for a’s sake but which a conceptually cannot have for her sake. Does this show that x cannot be a personal value? Does it even show, perhaps, that something can be a personal value for everyone except the person for whose sake we are favouring it? Both of these worries can safely be ignored. An analysis in terms of reasons would be incredible if it did not leave room for the possibility that reasons are at least agent-relative in the sense that only people who can favour something have a reason to do so. So if I cannot, on conceptual grounds, promise myself something for my own sake, then this is to be looked upon as any other case in which someone does not have a reason to favour something because she cannot favour the object. That the analysis is sensitive to this possibility is part of its very force. Accordingly, the reason FAP does not contain any clause to the effect that person a’s personal value must be understood in terms of at least a’s having an agent-relative reason has much to do with the fact—as I take it to be—that we should accommodate the possibility of personal Janus values.

8.2 Some counter-examples Here is another worry. FAP might fail to account for everything we consider to be of personal value. Think of a teenager a who wants to be ‘different’, and who values a band because it appeals only to him. He would stop valuing the band if he realized that his parents liked the group or its music.1 It seems, in other words, that the band might have personal value for a only if certain others do not favour it or even consider themselves as having a reason to favour it. Although teenagers’ attitudes do sometimes display this kind of conditionality, this does not afford a counter-example to FAP. It may nonetheless prove illuminating to discuss why it misses its target. To offer a counter-example, the teenager case would require us, ﬁrst, to establish on intuitive grounds that the object of the conditionalized attitude can carry personal value. It would then be necessary to say that FAP implies that the parents have a reason to favour the band (it has, after all, a personal value for their son). The ﬁnal move would be to insist that this is incompatible with the valuemaking base (the subjacent properties of the band): the fact that the parents see themselves as having no reason to favour the band. It is, I suppose, rather obvious why this example fails to present a problem to advocates of FAP. Surely, whether or not the parents have a reason to favour the band—or, to be more speciﬁc, that which a favours in the conditionalized manner— for a’s sake is not obviously settled by the fact that the band is favoured by a on

1

I owe this example to Michael Brady (personal communication).

112

P RO B L E M S A N D P O S S I B I L I T I E S

condition that the parents do not think they have a reason to favour the band. The latter does not exclude the former. Moreover, suppose that a’s favouring is of the following sort: a desires to listen to the band at a very high volume, on condition that the parents do not desire, or take themselves to have a reason to desire, the same thing. Again, this is not itself a reason to rule out the possibility that the parents have reason to favour the eventuality that the states of affairs a wants to obtain obtains for a’s sake. This is quite consistent with the fact that they do not have to desire to listen to the band. It is important to underline that FAP is formulated in terms of reasons that have an ‘objective’ reading in the following sense: quite what reasons there are to favour an object is not thought to be dependent on our beliefs. We may not know what features of an object are reason-providing, but the relevant features are nonetheless reasonproviding. In fact, objectivity in this sense requires not only ignorance, but also error. For instance, to take an extreme case, though we might all believe that something is a reason to favour an object, it is nonetheless possible that it is not a reason, since we all hold a false belief. There are yet other, more or less problematic, sides to FAP. For example, it implies that all that it takes for x to carry a’s personal value is that there be some reason for someone to favour x for a’s sake; but that seems too permissive. Moreover, it does not take much effort to imagine prima facie counter-examples. Here is such an example. It suggests that FAP forces us to ascribe conﬂicting personal values to a person. Suppose your father is a convinced theist, and that your mother an equally ﬁrm atheist. Your father wants you to believe that God exists, and he maintains that he wants this (among other things) for your sake. Your mother claims, on the other hand, that she wants you, for your sake, to believe that God does not exist. Surely we want to say both that your mother has an agent-relative reason to favour your adoption of atheism and that your father has an agent-relevant reason to favour your adoption of theism. But now, since what FAP requires for something to be a’s personal value is that there exists an agentrelative reason to favour this something for a’s sake, we have to conclude that you have two incompatible personal values; and this is hard to swallow. Atheism and (simultaneous) theism cannot both be states carrying your personal value. The conﬂicting-parents example has a number of facets. It opens up some interesting issues that must certainly be addressed. I will begin with the main issue: whether something can carry conﬂicting or incompatible personal values. To discuss this matter let us accept pro tem that your parents do, in the case described, have agent-relative reasons to favour your being and acting in certain ways. This seems plausible. It is less obvious, though, that they have a reason to favour your being an atheist, or a theist, for your sake. As might be recalled, it is one thing to favour x (instrumentally, or for its own sake) and another to favour x (instrumentally, or for its own sake) for someone’s sake. We need therefore to keep in mind that, on the most plausible interpretation of the conﬂicting-parents example, the kind of attitude involved is perhaps not an FFS (i.e. for-someone’s-sake) attitude at all—and that, if it is one, it may not be of the right kind. By ‘plausible’ I have in mind how easily, or to what extent, we can make sense of the

8.2

SOME COUNTER-EXAMPLES

113

example (rather than the frequency of such cases); and by ‘right kind of attitude’ I mean that we cannot exclude the possibility that each of the parents are in fact favouring x for their own sake rather than for their child’s. Where this is the case, we are dealing with the wrong kind of attitude with regard to the personal value of the child. So does the conﬂicting-parents example show, or at least render it plausible, that people can have conﬂicting personal values? I do not think so. In fact, if we are willing to make an assumption about reasons that appears quite persuasive, we can straightforwardly reject the intended conclusion of the example. Moreover, since there are, as I have just mentioned, other apparently reasonable and realistic ways of interpreting the example, our rejection of the example need not come with a costly price tag. The assumption I would be willing to make about reasons can be stated in a few words: one cannot have a reason to favour x and a reason to favour not-x. Certainly, this need not be true of every sort of reason. Thus, perhaps you can have a pro tanto reason to favour x and a pro tanto reason to favour not-x. But not all reasons are, I think, pro tanto reasons. So, setting these (and the corresponding pro tanto personal values) aside, I think there is something fundamentally counter-intuitive about the very idea that one and the same maximal set of relevant facts could give you a complete (or perhaps allthings-considered) reason to favour both x and not-x (whether that reason is agentrelative or not). In some of its features or parts, the world can, for example, be vague or indeterminate; hence we can anticipate that there will be ‘vague’ or ‘indeterminate’ facts (and reasons and personal values). It remains hard to believe, however, that the world could incorporate incompatible facts. Without doubt, I am here giving expression to a basic metaphysical view, and I am not sure that I can come up with some heavy artillery in its defence. Be that as it may, the view seems inoffensive when it comes, at least, to certain kinds of reason. Now, the scenario in the conﬂicting-parents example is different. It involves two persons; and since we are, I assume, dealing with agent-relative reasons, this opens up the possibility that these persons have different reasons—one has an agent-relative reason to favour x and the other has a similar reason to favour not-x. This makes sense. However, if we accept that facts cannot give one and the same person an agent-relative reason to favour both x and not-x, this implies that there must be some relevant difference between the father and the mother which gives rise to these two different reasons. Facts about x cannot on their own constitute an agent-relative reason to favour x and not-x. This is consistent with the claim that facts about x and the mother (and her relation to x) and facts about x and the father (and his relation to x) can generate two agent-relative reasons in this situation. Assuming this observation is correct, two connected points can be made. The ﬁrst concerns the fact that we need to turn our attention to the mother and father in order to determine whether the example involves two kinds of reason. We need, that is, to ﬁgure out whether there are relevant differences between these two persons, and then ask ourselves whether these differences are, together with other facts, such as to

114

P RO B L E M S A N D P O S S I B I L I T I E S

constitute two different reasons. My guess is that, often, once the parties agree about the facts of the case, they will also agree that they have the same kind of reason, or that only one of them will have a reason (this reason not being shared by the other person). People are most certainly different. However, at the risk of oversimplifying matters, I would add that it also seems clear that we are more alike than we are unlike. I am therefore inclined to think that most of the time facts will give people the same reasons (even if they are agent-relative such reasons); and that, depending on the circumstances, the facts will sometimes generate a reason for one but not another person. Much less commonly the facts will be such that two persons have different agentrelative reasons. As to the second point, suppose I am wrong to take this optimistic view of people’s similarities. So let us stick with the original assumption and suppose that the father does have an agent-relative reason to favour x while the mother has one to favour not-x. This does not imply that being an atheist and a theist is something that has personal value to the child. Since no one has an agent-reason for favouring both x and not-x for his or her own sake, we can safely conclude that in this very case, where x and not-x are features of the child (being an atheist and a theist) that the child could not have a reason to favour his own sake, it is clear that the issue of incompatible personal values does not arise at all. The more plausible interpretation of the conﬂicting-parent example is, I suppose, the following. Given the content of the mother’s conviction, she has a reason to see to it that she convinces the son to become an atheist (for his own sake); and given the content of the father’s belief, he has a reason to see to it that he convinces the son to become a theist (for his own sake). But from this it does not follow that the son has a reason to be a theist and an atheist for his own sake. Other examples appear to show that FAP is counter-intuitive. These set out from the very plausible idea that we do not conﬁne ourselves to speaking about the welfare of (or what is good for) people we care about, since we also express judgements about the welfare of people we are indifferent to, or even dislike. For example, although we dislike, or would have disliked, Hitler (to put it mildly), many people would nevertheless be inclined to agree, if asked, that Eva Braun was good for Hitler. FAP appears to have counter-intuitive results here. Why ought we to favour Braun for Hitler’s sake?2 As I shall explain in a moment, the advocate of FAP can deal with this kind of objection in various ways. Meanwhile, I would like to comment on Darwall’s analysis of welfare, which seems to handle cases like the Hitler example without any counterintuitive results.3

2 If you think the example has excessively strong ‘welfare’ connotations, you might prefer a different example: instead of letting Eva Braun carry Hitler’s personal value, we might think of one of his paintings or his summer residence. 3 For a review (in Swedish) of this book, see my ‘Recension av Stephen Darwalls Welfare and Rational Care’ (2004).

8.3

DA RWA L L O N W E L F A R E

115

8.3 Darwall on welfare The central tenet of Darwall’s proposal, which he calls a rational care theory of welfare, is the following: a person’s good is constituted . . . by what one (perhaps she) should want insofar as one cares about her. (2002, p. 4)

The care-proviso is, of course, what stands out here. For Darwall it is crucial; it makes it possible, according to him, to account for the normative feature of welfare: To understand the normativity of welfare . . . we must see it in relation to care . . . What is a conceptual truth is that to care for someone is to be in a relation to him such that considerations of that person’s welfare are normative for one’s desires and actions with respect to him. What is for someone’s good or welfare is what one ought to desire and promote insofar as one cares for him. (pp. 6–7) The normativity is not, however, the agent-relative kind of rational preference. It is rather an agent-neutral normativity grasped from the perspective of someone who cares for the person. (p. 45)4

Darwall’s approach seems promising. For instance, thinking that something is good for Hitler does not require that we have some actual attitude; it only requires that we believe that it is something that there is reason to want insofar as we care about Hitler. On the other hand, this approach also raises a number of questions. I discuss these problems in (2004). For instance, one problem concerns the fact that the analysis seems to give us something more than an analysis, namely a criterion with which to single out what is in fact beneﬁcial to individuals. See, for instance, page 31, where Darwall says “Something is for someone’s good if it is what that person would want for herself, as she actually is, insofar as she is fully knowledgeable and experienced and unreservedly concerned for herself ”. This is presented on the assumption that “any informed-desire standard can serve” as a “plausible criterion of welfare”. Moreover, the criterion resembles his analysis of welfare quoted above. Admittedly, it is not always easy to determine the distinction between analysing x and suggesting a criterion for what counts as x. But by adding the care-proviso Darwall appears to be trying to ensure that the analysis will not lead us to call things good-for that do not make the person better. If this is what Darwall in fact suggests (I might misunderstand him here, as mentioned), we should not, I believe, follow him. Here I will conﬁne myself mainly to just one feature that I ﬁnd problematic. According to Darwall’s analysis, a person a’s good accrues to an object x if and only if x is what one should want “insofar as one cares about” a. My own analysis resembles this claim, but there are also a couple of important differences between my account and Darwall’s. The more obvious one is, of course, that Darwall’s analysandum is much narrower than mine. He is interested in welfare. Second, this might account for the fact 4 Darwall assumes that care, or sympathetic concern, as he also refers to it, is like a psychological natural kind; see his (2002, p. 50).

116

P RO B L E M S A N D P O S S I B I L I T I E S

that he formulates his analysans in terms of only one sort of attitude, whereas FAP admits (and even encourages us to consider) a whole spectrum (perhaps even spectra) of attitudinal kinds. Objects carrying what I consider to be personal values typically provide us with reasons to be considerate. In fact, at one point I thought of the distinction between personal values and impersonal values in terms of what I called ‘(person) considerative’ and ‘non-(person) considerative’ values. To believe there is reason for you to be considerate in your dealings with something is merely to evaluate it as a personal value, as something that is good or bad for someone. However, this does not necessarily mean that you care for the person. Often you will, but caring is only one sort of attitude that objects of personal value call for. A number of attitudes can be directed at someone, or something, for someone’s sake, and it is not always the case here that we care for the person. This notwithstanding, we may still think there is reason to behave in a certain way, in the light of some fact or with regard to some object, and we might come to think this merely out of regard for the person. Again, our doing so reveals that we are treating something as a personal value. But once again, to be a considerate person is not necessarily to be someone who cares for the person involved. In this sense, being considerate is a much less demanding attitude than caring; to care for someone is to show among other things that you are considerate. The reverse is not necessarily the case. Admittedly, ‘caring’ might be given a broad sense so as to cover a variety of different attitudes. However, even granting this, it would, in my view, strain the sense of this term too much to include under it a number of the attitudes (including respect and desire) which, on FAP, we have reason to have. Third, Darwall’s analysans is not a straightforward FA analysis; it is formulated in terms of what we want insofar as we have some other attitude, namely one of caring. Why this insofar-condition, and how precisely should it be interpreted?5 As we have seen, Darwall is quite explicit about this matter. Were it not for the fact that we care about this person there would be no way of saying what we have reason to want for this person’s sake. Thus ‘caring for’ guarantees the normativity of ‘good for’; if we did not care, we would not have a reason to want anything for a’s sake. I am puzzled by this line of reasoning. It would be quite plausible to see in it an underlying and well-known view of practical reason. My problem is that I am unsure whether Darwall actually endorses this view. The core idea of the view is that what we have normative reason to do is all ultimately desire- or at least attitude-based—or (as one referee suggested), if we focus on welfare, based on certain kinds of attitude.6

5 Michael Zimmerman (2009) shares this worry: “It’s hard to know just what ‘insofar as’ should be taken to mean here. It cannot be equivalent to the ‘if ’ of the material conditional, since then the failure to care for P would sufﬁce for everything’s being good for P. Perhaps what is intended is the ‘if ’ of the strict conditional” (p. 437, n. 24). 6 In his earlier work Impartial Reason (1983), Darwall put forward the idea that “the content of the judgment that there is reason for one to do A is simply that were one rationally to consider facts relevant to doing A, then one would be moved to prefer doing A” (p. 128). Plainly, this notion makes the relation between what we mean by ‘reason for acting’ and our rational preferences analytic.

8.3

DA RWA L L O N W E L F A R E

117

Exegesis apart, suppose we did in fact endorse such a view. Would it be a problem? Not necessarily; but it would surely be preferable not to make the analysis dependent on a particular way of understanding reasons that is acceptable only to defenders of this view. Whether or not attitude-based views of either sort are plausible—I am inclined to think they are not—it would be a considerable drawback to reserve the analysis for advocates of one just theory of reasons. As far as I can see, an analysis of good-for is capable of being illuminating even if the deontic notion is regarded as a primitive. It is one thing to understand value (be it impersonal or personal) in terms of a deontic expression like ‘ought’ or ‘reason’. It is quite another to require that only insofar as certain kinds of attitude are present does something, such as a certain fact, supply a reason. Darwall’s insofar-condition is, in my view, an interesting, but nonetheless superﬂuous, element in an analysis of good-for, and it serves only to ensure that the analysis is less interesting to those unwilling to accept the relevant, special view of reasons. I share Darwall’s conviction that good-for must be understood in terms of reasons. I also share his interest in interpreting what these reasons are. However, this is not something I include in the very notion of reason I employ in FAP. Let me ﬁnally raise a minor point. The project of accounting for normativity in terms of Darwall’s care-proviso raises an interesting issue: should we regard welfare as normative when we need to formulate its analysis in terms of a condition, or should we say that it merely has conditional normativity? It might be suggested that this is a terminological quibble. I suspect it is not. Consider the distinction between ‘a belief on a condition’ (e.g. my belief that there is tiger in my living room, given that I see the animal with my own eyes) and a ‘conditional belief ’ (e.g. that if there is a tiger in my living room, I will believe I am in danger). Whereas the former is not a belief that I have unless the condition is met, the latter is a belief—it is just not a categorical one. This suggests, it might be replied, that we should rather formulate the analysans by analogy with a conditional belief. In the Eva Braun example we would thereby obtain something along the following lines: if you care for Hitler, you ought to favour Eva Braun ( for Hitler’s sake). However, as was mentioned before, I think it is a mistake if we introduce the ‘attitude (care)-clause’ in order to secure the normativity of the consequent. I would rule this out for the reasons mentioned earlier: I side, that is, with those who think that reasons must be attitude-given. If this is not what lies behind the ‘conditional form’, I suspect the disagreement between an advocate of FAP and someone offering this kind of analysis will boil down to a substantive disagreement about what reasons there are. I do not want to exclude the possibility that there are such ‘conditional values’; but in refusing to impose such an exclusion I am not obliged to deny that there are values in a stronger sense requiring a categorical analysans— which is what I try to deliver.

118

P RO B L E M S A N D P O S S I B I L I T I E S

8.4 Different strategies Let us return to the defensive strategies that are open to an FAP-theorist confronted by problems like the one set by the Eva Braun case. Recall the key objection: most of us could not care less about what was done for Hitler’s sake, but it is nevertheless plausible to say that Braun had personal value for Hitler. This clearly is imaginable whether or not we adhere to FAP. However, if an analysis of personal value would imply that we should favour things for Hitler’s sake just because we recognized that Hitler had personal values, this would, the objection goes, talk against the analysis. Now, in contrast to Darwall’s account, FAP implies that if Eva Braun has value for Hitler, then we have a reason to favour her for his sake. This suggests then that FAP has counter-intuitive implications. There are in fact two quite familiar strategies open to advocates of FAP at this point: one explains why these examples appear counter-intuitive in such a way that the analysis does not require that we hold actual pro-responses with an eye to Hitler. On the second approach we are forced to swallow the bitter pill; if Braun has personal value for Hitler then we should have pro-responses vis-a`-vis Braun for Hitler’s sake. Considering the ﬁrst strategy, a combination of some well-known replies will be helpful. First, we must not forget that it is quite possible that Braun had no personal value for Hitler. But notice that to say this is not to deny that Hitler did in fact favour Braun. Apparently he did. But, again, from the fact that he loved and cherished Braun it does not follow that she was a carrier of personal value. It means only that she was the object of some of Hitler’s pro-responses. It might be argued that this fact does not explain our intuition that Braun was not merely liked by Hitler but had value for him. Perhaps we seem ready to hold that Braun had value for Hitler because we are so accustomed to ignoring the meta-value issues that arise when we speak about personal values—the reason for this inattention being that we take the following as obvious: if someone cherishes, or desires, an object, it has value for that person. But the fact that people often reason in this way is no reason to say that the reasoning expresses a correct position. For my own part, I fail to see why personal value should be exempted from the standard issues raised when we talk about (impersonal) value. There is one possibility that is closely related to the above point. Suppose we admit that “welfare” has more than one sense. Besides the one FAP is trying to capture, there is a purely descriptive non-normative notion. What if welfare constitutive in this sense might also be considered as what makes something good for someone? Perhaps “good for” likewise is ambiguous. This ambiguity would help to explain why we believe it is ﬁne to think that Eva Braun is good for Hitler despite our detesting Hitler to such an extent that we are not willing to favour anything for his sake. When this happens, the explanation is that we have a sense of ‘good for’ in mind which, plainly, is not normative. Of course, the reason we want to say that Braun carries value-for need not be that we are metaethically naive or that we have a notion of good for in mind that FAP is not trying to analyse. Perhaps the judgement ‘Braun is good for Hitler’ expresses what

8.4

D I F F E R E N T S T R AT E G I E S

119

R. M. Hare once called an ‘inverted-commas judgement’. What we say is that Braun has value for Hitler according to Hitler. In other words, the ‘good for’ in the judgement is parasitic on the genuine evaluation of Hitler. These replies seem plausible. It is their generality that worries me. Is it really always the case that we are either metaethically naive or expressing some kind of inverted-commas judgement when we judge that something is good for a person we dislike? The second strategy acknowledges that the judgement ‘Braun is good for Hitler’ is a genuine evaluative judgement; it handles the counter-intuitiveness of FAP in this sort of case by underlining that the reasons involved here are pro tanto reasons, i.e. reasons that persist even in the presence of weightier counter-reasons. To think that Braun is good for Hitler is to think that we ought, pro tanto, to have some pro-response regarding Braun for Hitler’s sake. In admitting this we are not, of course, admitting that Hitler is worthy, all-things-considered, of our pro-responses. Nor does it mean that he should be favourably treated until there is a stronger counter-reason to not do so. All it means is the following: whatever it is that we think are the good-for making properties of Braun, it is only in regard to these properties that we should favour Braun for Hitler’s sake. The approaches can in fact be combined. The fact that they can strengthens our defence of the idea that FAP accounts for people’s personal values—whether or not they involve malicious people. Still, patchwork solutions are rarely ideal—in particular, if they have to be backed up by a complex, primitive notion such as the notion of a ‘genuine evaluation’. It is therefore wise to look for more plausible ways of dealing with these alleged counterexamples. One appealing approach, suggested by a referee, deals with the Braun–Hitler case by distinguishing between moral and non-moral reasons to favour Braun for Hitler’s sake. The atrocities committed by Hitler are perhaps such that we have no moral reason to favour anything for his sake. However, this is consistent, or so it might seem, with there being non-moral reasons to favour Braun for Hitler’s sake. The suggestion that the reasons to favour Braun for Hitler’s sake would be non-moral has the obvious merit that it kills two troublesome birds with one stone: the counterintuitiveness of the patchwork explanation is gone; Braun has value for Hitler. However, recognition of this does not necessarily commit us to favouring Braun for Hitler’s sake, although such a commitment would follow if we were non-responsive to moral reasons, or if moral reasons did not override other sorts of reason. Consider again the pro tanto solution. It was argued that Braun’s being good for Hitler supplies us with a pro tanto reason to have a pro-response to Braun for Hitler’s sake, but that it does not follow from this that Hitler is in toto worthy of our proresponses. The referee’s suggestion might in fact strengthen this claim given the following premise: in cases where a moral reason is a pro tanto reason and is in conﬂict with another non-moral pro tanto reason, the non-moral reason is overridden. Just how passable a road this is remains to be shown. A serious obstacle certainly remains, since we must explain what, exactly, makes a reason a moral one—a task I will not undertake here.

120

P RO B L E M S A N D P O S S I B I L I T I E S

Let us move on to yet another worry about FAP. Like the Hitler example, this worry encourages us to regard FAP as too demanding. However, this time the issue looks more psychological than moral in nature. Consider the following example, then. Suppose winning the New York marathon has personal value for Charlie. Furthermore, assume that you want to win this race, too. Is it really the case that you should then favour his victory? Is that not plain counter-intuitive?7 Christian Piller (personal communication) has suggested another example: if your headache is good for your opponent in a game of chess because it increases his chances of winning, you surely do not thereby have reason to favour your headache for his sake as FAP requires. But since it does seem intuitively plausible to suppose that your headache is good for your opponent, there seems to be something wrong with FAP. Of course, in a case like this we might not be in agreement over what is good for the person. However, suppose you do believe that your headache is genuinely good for your opponent, and that you do not take this to be what is sometimes called a ‘natural normative fact’. The interesting thing to ask then is whether FAP provides a way of explaining why the headache is good for the opponent. When we take into consideration, ﬁrst, that favouring your headache for his sake might involve a number of very different kinds of attitude (say, liking or preferring the headache for his sake), and second, that the reason you have to favour your own headache for another’s sake may well be outweighed by other reasons (reasons having to do, perhaps, with the fact that you dislike the headache for your own sake), cases like that of the headache do not seem to present a serious problem for FAP. More generally, before surrendering to this sort of objection—the sort of objection generated by the chess and marathon examples, I mean—we should investigate to what extent ‘O has personal value for a’ expresses a genuine evaluation. I think that much of the resistance to the normativity of personal value derives from insufﬁcient recognition that the examples concern inverted-commas uses of ‘value for’, or from the assumption that ‘O has personal value for a’ is equivalent to ‘O is de facto favoured by a’. Still, suppose someone claims to recognize that x has personal value for a, but denies that this gives him any reason to favour x for a’s sake. Pointing out that reasons here might have pro tanto status will hopefully clear away much of the misunderstanding, but what should we say to someone who continues to deny that there is any reason at all to favour x for a’s sake? We could try reminding him that ‘favouring’ is a technical term covering a wide range of attitudes. In other words, a variety of favourings might be relevant here, such as to be glad for a’s sake, or to respect the winner a’s sake (and not, say, spit on it) at the end of the marathon. Many more examples could be given. Of course, some might insist that nothing whatever follows, regarding our attitudinal stance to x, from the fact that x has personal value for C. But now we have arrived at a point where I, at least, ﬁnd it hard to understand what ‘value’ in ‘value-for’ stands for in this person’s view.

7

The examples are inspired by objections raised by Jonas Olson and Michael Brady.

8.5

H E AT H WO O D ’ S O B J E C T I O N

121

An alternative approach to this issue would be to ask the persistent sceptic the following question: Would you not accept, at least, that if you want to do everything that can be done for a’s sake, you ought to favour O? I owe this informal suggestion to Margaret Gilbert (personal communication). These suggestions are worth considering. However, they regard (personal) value as normatively conditional on a person’s attitudes (wants and carings-for), whereas my suggestion sets out from the idea that what we have reasons to do is not necessarily a matter of what attitudes we have.

8.5 Heathwood’s objection There are further grounds for hesitating before seeking to interpret what is good for you in terms of FAP. For instance, Chris Heathwood (2008) has recently argued in ‘Fitting attitudes and Welfare’ that welfare, or what is good for a person, is not susceptible to a ﬁtting-attitude analysis. However, I think his conclusion can be avoided. His argument sets out from the idea that welfare is an intrinsic kind of value. But it is also, he thinks: . . . a relational kind of value. We express the idea when we say that something would be good for someone. Welfare value is intrinsic because we are saying the thing is good in itself for the person, and not merely good for what it leads to for the person. It is relational because it is a relation between the thing that’s good and the person for whom it’s good. (2008, p. 52)

Since welfare is an intrinsic value, its value will, he thinks, be unchanging; so if something is good for me yesterday, it will be good for me today and in the future in the sense that it will make “my life go better”. He then argues that “while an event’s value for a person is unchanging, the attitudes he has reason to have towards such an event can change over time” (p. 52). Suppose a person experienced a great pain yesterday, and that today, as the result of taking a drug, he does not recall the pain he felt yesterday. Does he now have “reason to be intrinsically averse to that ordeal for his own sake?” (p. 56). Heathwood thinks the answer is negative, and that therefore welfare, and what is good-for a person, should not be understood in terms of a ﬁtting attitude analysis. This reasoning is based on some assumptions that I do not share. However, even if we focus on the case he describes, I am not convinced that the example has quite the implications he seems to think it has for an FA analysis.8 He is correct that welfare might have relational value. However, it is unclear why he simply assumes that welfare is an intrinsic value. He overlooks the possibility that welfare is relational in more ways than one. Notice that Heathwood understands the “intrinsic value” of welfare as 8 Heathwood’s argument against an FA analysis of welfare is part of his general argument against ﬁttingattitude analyses. He argues that since it “would be theoretically unsatisfying to be forced to say that these two kinds of value [good simpliciter and good-for] have radically different natures”, we should give up the ﬁttingattitude analysis for good simpliciter too. As an illustration he uses the case ‘My Past and Future Operations’ presented by Derek Parﬁt in (1984, p. 165).

122

P RO B L E M S A N D P O S S I B I L I T I E S

welfare being good, in itself, for the person. However, as we have seen, ﬁnal value as well as ﬁnal value for might be extrinsic as well as intrinsic. Heathwood never really considers the idea that the pain I had yesterday might have an extrinsic ﬁnal value— extrinsic in the sense that the time of the experience might well be one of the bad-for making features of the event. If this were the case, the good-for accruing to the event would perhaps need to be time-indexed. But whether or not we want to accept this idea, if the ﬁnal value of the pain is an extrinsic ﬁnal value for, there is nothing peculiar about the idea that our reasons for favouring this event alter over time. In fact reasons are not the only thing that might vary. The sort of favouring may also change. For instance, I might have reason to dislike now the pain I have now, but to wish tomorrow that the pain I had yesterday is not to be repeated. Admittedly, this reply will not convince anyone who still thinks that the badness accruing to the pain yesterday is unchangingly bad, so Heathwood might still have a point. However, I am inclined to think that the reason we regard the pain as something with unchanging intrinsic value is that we take it to have ﬁnal intrinsic value, period; it is at least reasonable to regard pain as having ﬁnal value simpliciter. But if that is the case, we have, on an FA analysis, a timeless reason to disfavour the pain. This also seems to be true if what is bad about the event in the past is rather that is makes our life worse. In such a case we might regard the value of the pain event as having a contributory value; what has the non-contributory value for me is, say, my life. In this case, given that my life has ﬁnal intrinsic value for me, it might be reasonable to say that the pain event is a contributory bad-for that gives me an unchanging reason to disfavour it.

8.6 The good life and the argument from fetishism Like most people I am interested in the truth. Some philosophers are more directly concerned with it than others; they are really very keen to tell us what the facts are. Not all philosophers are quite this straightforward, though. Some are concerned chieﬂy to probe and explore possibilities and plausibilities rather than to disclose truths. I believe I belong, by temperament, in the exploratory camp. At the risk of sounding pompous, truth is something to hope for; it is interesting potentials that one should keep watch for. In this work I have tried to explore the possibility of expanding our taxonomy with a new set of values.9 In doing so, as was made clear at an early stage, I have had to leave a number of interesting issues untouched. It was tempting to pursue at least some of these issues en route. Doing so would have considerably lengthened and delayed the 9 I have a twofold reason for being so bold as to speak of expanding classical value taxonomy. First, goodfor has generally been understood as being dependent on the notion of good, period (or good simpliciter). However, I am treating good-for as a value notion independent of good, period. Second, the sorts of value that I am trying to understand in terms of FAP have been regarded by certain authors, including Rosati, as values that do not really correspond to what she (and others) consider to be good for people. My notion of a personal value seems, therefore, to encompass much more than what is often considered to be good for people.

8.6

T H E G O O D L I F E A N D T H E A R G U M E N T F RO M F E T I S H I S M

123

preparation of the book, however, and I would prefer to receive reactions to it at this stage, before taking on some of these other issues. One such matter concerns the relationship between personal and impersonal values. Given that conﬂicts between these kinds of value abound, this is a matter of serious theoretical concern that needs to be addressed. Often we can only promote an impersonal value at a cost to somebody’s personal value. To take a fairly simple example, suppose justice demands that a person should return an object with personal value to her to its rightful owner. Perhaps she acquired the object from someone who had stolen it and therefore did not have title to it, and therefore could not give her title to it. Can a situation be imagined in which it would be the personal value rather than justice that we should secure? I think so (and I sincerely hope I am not alone in thinking so). I also suspect that most so-called NIMBY (not in my back yard) cases concern conﬂicts between impersonal and personal values. If I am right about this, the issue is of not only theoretical but also great practical concern. Or suppose by securing a certain personal value, I do something that has (instrumentally, or is in itself a) negative impersonal value. Should I refrain from acting? In comparison with impersonal values, many—perhaps, most—personal values appear to carry little weight when we have to choose between them and impersonal values. There are no ready-made answers or guidelines on how we should resolve these conﬂicts. There are no clear-cut values about value, i.e. we need at least to be able to compare these different values to each other in evaluative terms. Of course, it would be marvellous if there were a rich set of values that accrued speciﬁcally to value comparisons (and whose values were not classiﬁable as personal or impersonal values), but this is still an underexplored area.10 Another issue to which I would like to return in the future concerns the place of personal values in a good life. Obviously such values play a number of different roles in a person’s life. Some function as incitements, others as deterrents, to action; certain personal values require promotion, and others, as we have seen, call for quite different kinds of response. But quite what roles, precisely, personal values play in the good life is as yet unclear to me. I shall not at this point try to address the difﬁcult issue of what makes a life good. However, I would like to forestall a couple of potential misunderstandings about the contributory value of personal values in the good life. To facilitate discussion, let us conﬁne attention to positive personal values. Should we then conclude that a life full of personal values is a good life, and that a life less well endowed with personal values, all other things being equal, is a worse life? Is the relation here really that straightforward? It seems hard to believe. When we consider the nature of some of the examples discussed in this work, such as the bits and pieces of a bookshelf that I mentioned at the outset of Chapter 4, it becomes absurd to say that my life would be better just because it contained more rather than less parts of these bookshelves. Living a life rich in personal values is no guarantee that you are living

10

An exception to this is Wlodek Rabinowicz’s ‘Value Relations’ (2008).

124

P RO B L E M S A N D P O S S I B I L I T I E S

a good life (or, in one sense of these terms, that your level of welfare or well-being is high). Indeed there is an obvious problem with a life with too many personal values, particularly if they are of a certain kind. According to the argument from fetishism a life full of at least certain personal values is, in effect, a bad life. The reason for this varies, depending on the way the argument is advocated. The life might be bad because it is superﬁcial—a shallow sort of existence, one in which the person has the wrong priorities. Alternatively, its being bad may have to do with some notion of authenticity or false consciousness: if you surround yourself with material things that have value for you, you are not really living a good life. You may think you are, but you are deluded. In fact, to give weight to concrete things which you consider to have value for you is to look at life with a blindfold on. ‘Real’ values—those capable of being elements in the good life—will forever escape you. Of course, the fetishist accusation is a sweeping attack, and perhaps therefore difﬁcult to take wholly seriously. For one thing, it overlooks the distinction between my treating something as a personal value and something’s actually being my personal value. I might surround myself with things that I claim have personal value, but whether these objects really carry value is, in the end, not a matter to be settled by personal ﬁat. Since I have already addressed this matter enough, I will spare the reader yet another repetition. However, the fetishist argument does highlight an interesting issue concerning a certain group of personal values—namely, those that accrue to material things (e.g. the sheets of paper with poems inscribed upon them and bits and pieces of furniture that I have considered in the course of discussion). Since one desideratum of the good life is surely that such a life will display good-making/contributive properties that are worth striving for,11 we seem obliged to argue that we should desire to have such objects if we want a good life. But it would be ridiculous to maintain this, so things carrying personal values are not part of a good life. Moreover, if these personal values play no part in a good life, we might even question whether they are real values in the ﬁrst place. Given the close, intuitive connection between the notion of personal value and the notion of a good life, we should at least expect there to be some interesting relationship between them. If FAP cannot guarantee this, it becomes questionable whether it is about personal values in the ﬁrst place. This line of reasoning does display certain insights. What is more, I suspect that the assumption underlying it is that concrete things lack any value at all—or, that if they have value it is instrumental. If the latter is the case, then the thought is that we are probably mistaking instrumental values for ﬁnal ones. This is a classic mistake. Hume warned us about it. Cases of such conﬂation are probably real. (It is at least imaginable 11

Joseph Butler’s lesson to the hedonist, that it might be self-defeating merely to seek pleasure, must be borne in mind here. That is, it is probably a valuable piece of advice to anyone who thinks that there are things that make a life good, and so should be pursued. Butler’s further point, that in order to maximize pleasure the hedonist should pursue things that are not valuable, may not be as easily followed by those espousing other views of value. See Butler, Fifteen Sermons Preached at the Rolls Chapel (1900 [1729], Preface, Section 31, and Sermon xi, ß1.9); see also Henry Sidgwick, The Methods of Ethics (1907 [1874], pp. 136, 403).

8.6

T H E G O O D L I F E A N D T H E A R G U M E N T F RO M F E T I S H I S M

125

that, in the eyes of a miser, money has ﬁnal value, whereas the reasonable position is that it has at most instrumental value.) Still, the challenge is nonetheless ﬂawed. It sets out from an even more fundamental idea, namely that values are things to be promoted; and since promoting certain personal values (such as the value accruing to an artefact from one’s past) makes no difference to whether one has a good life or not, it is readily concluded that there is something ﬁshy about these personal values. But this reasoning is based on the false assumption that all valuable objects ought to be promoted. Some should, but only some! In particular, concrete objects (e.g. people and things) of personal value are such that they should be, for example, respected, admired, or loved for some person’s sake. They are not such that we should seek to multiply them. In most cases that would be absurd. From a buck-passing perspective, the above is suggestive of at least two lines of further enquiry. If we are interested in what makes a life good, we might start by contemplating what we suspect are the good-making properties of a good life. These may be natural properties; equally they might be different kinds of personal as well as impersonal value. If we are interested, in addition, in what makes a person considerate, I suggest we start looking at how the considerate individual deals with people’s personal values (including her own). To be considerate is, in effect, to be someone who favours something for someone’s sake. Being moral is not the same as being considerate, but it is a good start.

9 One Reason Dichotomy Less?

Since its introduction by Thomas Nagel (1970) and Derek Parﬁt (1984), the dichotomy of the agent-relative and the agent-neutral has become widely popular.1 For instance, there is at present a lively discussion going on between philosophers interested in taxonomizing moral theories who detect a potential line of demarcation in agentrelative restrictions.2 Certain theories (particularly utilitarianism) have been criticized for being unable to incorporate such restrictions. On the other hand, so-called common-sense morality is said to be full of agent-relative restrictions. So why not make use of this much-employed distinction? Why not simply draw the line between impersonal and personal value with the help of these two kinds of reason? It would be a tidy solution, certainly, but I think there are strong arguments for being cautious about this approach. My arguments have recently undergone a change, though. When, many years ago, I started thinking about the notion of value-for it was very much because I was not satisﬁed with the idea of values being agent-relative. To look for an analysans in another dimension of relativity did not, therefore, seem a good idea. This was not the way to go. Actually, I had yet another motive, besides that of avoiding making personal values private. I found the very notion of an agent-relative reason unclear. I could get a grip on a fairly unambitious way of understanding agentrelativity relatively easily, particularly when it was associated with aims, principles, statements, or judgements (propositions). But things get much more complicated when we turn to reasons—or, at least, the normative kinds of reason FAP requires. My feeling was, and still is, that people try to squeeze out much more than a modest, unambitious sense of this notion. And to analyse a notion like personal value with the help of an imprecise concept did not seem like a good idea. Recently my views on these issues have changed radically. I am still sceptical about the dichotomy, but my scepticism now concerns to a greater extent its other 1

Nagel, The Possibility of Altruism (1970), and Parﬁt, Reasons and Persons (1984). See e.g. Samuel Schefﬂer (ed.), Consequentialism and Its Critics (1986) and David McNaughton and Piers Rawling, ‘Deontology and Agency’ (1993). 2

9.1

I N T RO D U C T O RY N O T E S

127

component: it is the notion of an agent-neutral reason that I now ﬁnd obscure. This would be worrying were it not for the fact that I now also believe that there is a way to avoid ending up with private values even if we employ an agent-relative reason. In the remaining sections of this chapter I will explain my change of mind; and I will argue that, given certain plausible assumptions, we should be sceptical about the distinction between agent-relative and agent-neutral normative reasons. Just what conclusion we should draw from this scepticism is still a somewhat open matter, to which I will return at the end of the chapter. It might just be that there is a way of meeting the kind of challenge I shall pose to the coherence of the notion of an agent-neutral reason. Jonathan Dancy’s recent idea that reason considerations play different roles—they may, in short, be favourers or enablers—might help us clarify and retain a notion of agent-neutral reasons. This would be most welcome. I have some qualms, though. Dancy has argued that all reasons are contributory (pro tanto) reasons. I am not quite convinced by Dancy’s arguments, and although I do not take the matter to be settled, I ﬁnd the notion of a complete reason to be quite plausible. If I am right about this—and I am not alone in thinking it to be so—then Dancy’s favourer/enabler distinction will still be of use to us, although this use will be slightly more limited. It will help us save the dichotomy when it comes to contributory reasons. However, if there is something genuinely ﬁshy about this distinction, it would certainly be bad advice to locate the boundary between personal and impersonal values in the fact that the former, but not the latter, are agent-relative reasons. We should need to ﬁnd another boundary. It was this insight, then, which prompted me to understand personal values in terms of for someone’s sake attitudes.

9.1 Introductory notes The importance of the distinction extends to most areas in practical philosophy. Despite its popularity, it remains difﬁcult to grasp. In part this has to do with the fact that there is no consensus concerning the sort of object to which it should be applied. Nagel distinguishes between agent-neutral and agent-relative values, reasons, and principles. Parﬁt focuses on normative theories (and the aims they provide to agents). In an early paper, McNaughton and Rawling (1991) are interested in rules, and in a later work (2002) they go on to discuss reasons. John Skorupski formulates the distinction in The Domain of Reasons (2010) in terms of predicates, and there are other suggestions too. Some writers suspect that we fundamentally talk about one and the same distinction (e.g. Dancy, 1993; Ridge, 2005).3 This latter issue will not be addressed here, though. What is important to keep in mind, is that my focus continues to be on (practical) normative reasons for action rather than theoretical reasons for belief. Moreover, I will understand ‘action’ in a broad sense of the word which includes states

3

See Dancy, Moral Reasons (1993) and Ridge, ‘Agent-Neutral vs. Agent-Relative Reason’ (2005).

128

ONE REASON DICHOTOMY LESS?

such as desiring, wanting, intending, and (certain kinds of ) feeling. Skorupski (2010) makes a more ﬁne-grained division, distinguishing between epistemic (what I call theoretical), practical (reasons for desires), and evaluative reasons (reasons for sentiments, emotions, and affective responses). However, for my purposes it sufﬁces to operate with the more general distinction between reasons for beliefs and reasons for desires and emotions. What is often overlooked is that there are at least two related, but nonetheless different, ways of understanding the agent-relative/neutral reason distinction. One, outlined in Section 9.2, centres on whether reasons do or do not essentially involve, or refer to, particular agents; this is the philosophically more interesting notion. The other, brieﬂy discussed in Section 9.3, takes into consideration the number of people who as a matter of fact have a reason to do something. The remainder of the chapter asks in what way the notion of a normative agent-neutral reason is problematic. Needless to say, the fact that there are problems associated with the distinction need not mean that it is suspect. The philosophical literature is full of (some would even say thrives on there being) intuitive distinctions that have not as yet received proper analysis. However, because the prospect of getting the distinction right looks somewhat gloomy, as I shall try to show below, I think things are more serious in this case. Whether there are sufﬁciently strong reasons for giving up on the more philosophically relevant distinction as a conceptual confusion is a matter I leave to the reader’s own judgement. I conﬁne myself to setting out a diagnosis. The cure will depend on one’s attitude to a number of issues the discussion of which falls outside the scope of this work.

9.2 The essentialist sense The philosophically more interesting sense of the distinction was formulated by Nagel (1986) in The View from Nowhere: If a reason can be given a general form which does not include an essential reference to the person who has it, it is an agent-neutral reason . . . If on the other hand the general form of a reason does include an essential reference to the person who has it, it is an agent-relative reason.4

Nagel draws the distinction in terms of what does and what does not essentially refer to a particular person, the owner of the reason. Various other writers have employed similar accounts. They have also tried to be more speciﬁc about what is required if a reason is to qualify as agent-relative or neutral. Philip Pettit, for instance, describes an agent-relative reason as “one that cannot be fully speciﬁed without pronominal backreference to the person for whom it is a reason”. He describes an agent-neutral one

4

Nagel, The View from Nowhere (1986, pp. 152–3).

9.2

THE ESSENTIALIST SENSE

129

“as one that can be fully speciﬁed without such an indexical device” (1987, p. 75).5 Several other writers have suggested similar accounts, stressing that the distinction concerns reasons that do or do not contain ineliminable back-reference to the agent for whom it is a reason (cf. McNaughton and Rawling, 1995; Portmore, 2001; Ridge, 2005).6 We cannot, of course, be sure that Pettit’s and similar approaches to the dichotomy fully coincide with Nagel’s own approach. It is, for instance, a matter of interpretation when a statement is “fully speciﬁed”. But this notwithstanding, the concern about indexicals and ineliminable back-references surely reﬂects the authors’ efforts to express what Nagel was driving at—namely, that the distinction is about reasons that do and do not essentially involve particular agents.7 How we present this in language is an interesting but nonetheless secondary matter; the real issue is still what makes reasonstatements true. Another inﬂuential view denies that reasons are facts on the grounds that normative reason-statements express non-cognitive states (e.g. desires). Discussion of this alternative here would complicate the presentation of what I take to be the problem with the dichotomy considerably, and I have therefore not entered into it. The ‘essentialist’ approach to the distinction is the interesting one, in that it tries to unravel a conceptual difference (one which, among other things, is thought to help us distinguish between different sorts of moral theory). It focuses on one particular feature: whether or not a reason-statement necessarily refers to the agent for whom it is a reason. The important thing to realize, then, is that the distinction is silent about the range of agents for whom something is a reason. It leaves this matter quite open. This is not true about the second approach, where the issue of quantiﬁcation is positioned centrally. But before commenting on this second, as I will call it, ‘number approach’ to the dichotomy, it will be convenient to make the ﬁrst interpretation at least intuitively accessible. Despite the fact that there is no general agreement8 as to how to formulate the ‘essentialist dichotomy’, it is at least possible to express in a few words what seems to be its core idea. Let P refer to a ‘reason-statement’ (by which I mean no more than a Pettit, ‘Universality without Utilitarianism’ (1987). In ‘The Paradox of Loyalty’ (1988), Pettit suggests that “A reason for action [ . . . ] is the sort of proposition which may appear in the major premises of a practical syllogism”. 6 D. McNaughton and P. Rawling, ‘Value and Agent-Relative Reasons (1995), Douglas Portmore, ‘McNaughton and Rawling on the Agent-Relative/Agent-Neutral Distinction’ (2001), and Michael Ridge, ‘Agent-Neutral vs. Agent-Relative Reason’ (2005). 7 Nagel, originally in The Possibility of Altruism (1970), referred to the distinction as one between subjective and objective reasons. The terminology ‘agent-neutral and agent-relative’ was introduced by Parﬁt in ‘Prudence, Morality and the Prisoner’s Dilemma’ (1979). 8 For some attempts to clarify the distinction, see McNaughton and Rawling, ‘Agent-Relativity and the Doing/Happening Distinction’ (1991) and ‘Conditional and Conditioned Reasons’ (2002), and Skorupski, ‘Agent-neutrality, Consequentialism, Utilitarianism . . . : a Terminological Note’ (1995). For a review of this last, see John Broome, ‘Skorupski on Agent-neutrality’ (1995). Objections to McNaughton and Rawling (1991) are given in Portmore (2001). For a reply, see McNaughton and Rawling (2002). 5

130

ONE REASON DICHOTOMY LESS?

statement expressing a reason). Nagel, Pettit, and others make this move from reasons to statements, or what they call forms, expressing reasons. It will simplify the presentation if we sometimes discuss reason-statements rather than reasons. The expression ‘reason-statement’ is ambiguous. Typically it will be a statement expressing some fact—say, ‘that person a needs help’ or ‘that Anna is my daughter’. However, it may also refer to statements, which consist of two parts: one stating that (i) -ing is normatively called for, and the other listing (ii) the reason-making grounds for -ing. Here are some different examples: ‘Charlie ought to help Mary because she is drowning’, and ‘There is a reason for Tom to help Mary, because she is his daughter’. The left-hand side of the ‘because’ expresses (i), whereas the right-hand side expresses (ii). Moreover, statements of the ﬁrst kind are elliptical in the following sense: fully spelled out, they will appear on the right-hand side of a statement containing parts (i) and (ii). Since nothing important hinges on this matter, I will, for simplicity’s sake, mainly use examples of the ﬁrst kind. Returning to the core idea behind the essentialist dichotomy, it can be expressed in the following way—where a (and later b) refers to an agent, and to some act: If P states a reason for a to , then: P states an agent-relative reason for a if and only if P contains an essential reference to a, Otherwise, P states an agent-neutral reason for a to . Moreover: P contains an essential reference to a if and only if P is not logically equivalent to any other statement Q that does not refer to x. The following statement expresses, on the essentialist approach, an agent-neutral reason for a (in case a 6¼ b): (Pn) ‘Person b is drowning’. The fact that b is drowning constitutes a reason to save b that is not agent-relative but agent-neutral. The statement (Pr) ‘My daughter b is drowning’, expresses in its turn a typical example of an agentrelative reason; it gives b’s parents a reason to save b that is different from the reason that I, not being her father, have to save her. The above picture is oversimpliﬁed, however. The ‘relative/neutral’ status of these reason-statements, Pr and Pn, seems to depend on the person to whom they are addressed. For instance, if Pn is directed to person b, it should be seen as stating an agent-relative reason for b. And while Pr may state an agent-relative reason for b’s parents, it would, in many people’s minds, also state a reason for people other than b’s parents, since it seems to be an agent-neutral reason that we should save people from drowning. These dependencies can be quite confusing. They are one of the sources of the slipperiness of the distinction. That is, given that we have in mind a certain kind of

9.3

T H E N U M B E R A P P RO A C H

131

reason—which I will come back to later on—it is clear that whether a reasonstatement expresses something agent-relative or agent-neutral depends, in part, on the person to whom it is addressed. On the other hand, if the truth of a reasonstatement is not relativized to the way the person to whom it is addressed understands the statement, the above confusion concerns something else, namely what people take to be reasons. This distinction between, as I will refer to them, motivating and real reasons, will occupy us shortly. In the meantime the second important ﬂeshing out of the dichotomy must be considered.

9.3 The number approach The essentialist notion is not the only one in circulation. To say that P states an agentneutral reason is, on the number approach, just another way of saying that P states the same reason for all agents; and where this is the case P is perhaps best understood as stating the same reason for each agent. A neutral reason is therefore understood in terms of being a reason that everyone has. An agent-neutral reason is a reason that calls for the same action from every agent (to which it applies), whereas an agent-relative one calls for different actions from each agent (to which it applies). The number approach should perhaps be labelled ‘Parﬁt’s approach’, because it seems to be the one Parﬁt has in mind. In Reasons and Persons (1984), he explains it in the following passage, where ‘C’ refers to a moral theory: Since C gives to all agents common moral aims, I shall call C agent-neutral. Many moral theories do not take this form. These theories are agent-relative, giving to different agents, different aims.

Further on, he adds: . . . When I call some reason agent-relative, I am not claiming that this reason cannot be a reason for other agents. All that I am claiming is that it may not be. (1984, p. 143)

Two things particularly merit comment here. First, notice that there is no mention of the notion that some reasons are essentially an agent’s reason, so we can reasonably suspect that we are here not dealing with an essentialist way of understanding the dichotomy. Second, it is not obvious how Parﬁt’s insistence (in the latter quote) that what is an agent-relative reason may also be an agent-neutral reason should be understood. The natural conclusion here would, I think, be that whether something can or cannot be a reason for other agents has do with the sort of people who happen to exist, or be present, at the time the act is normatively called for. If persons with a certain property S have a reason to do something, then it is an agent-relative reason if as a matter of fact not everyone has the property S. On the other hand, if people were to come to have property S, the reason would be agent-neutral. In other words, if a reason applies as a matter of fact only to a particular agent (or group of agents), it is has agent-relative status. Otherwise it is agent-neutral. I confess I am not entirely certain

132

ONE REASON DICHOTOMY LESS?

how much of a conjecture this interpretative point is, so in what follows I will use the expression ‘number approach’ rather than ‘Parﬁt’s approach’.9 The relationship between these two approaches—essentialist and number-based— needs clarifying. While the essentialist picture explains (to a degree) why certain reasons prescribe different actions for different agents and others prescribe the same action to everyone, the opposite does not hold. That is, a reason’s being neutral or relative on the essentialist deﬁnition tells us why it is neutral or relative on the number deﬁnition—but not vice versa. On the number approach the dichotomy is dependent, not on a conceptual matter, but on how the world happens to look. In contrast with the essentialist notion, this approach does not purport to say (at least, not necessarily) anything about what a reason is. The essentialist notion is, therefore, the more interesting one; and accordingly my focus will be on this view of the dichotomy.

9.4 Reason-for I want next to draw attention to a particular feature of reasons—what I refer to as the personalizability feature (PEF for short). PEF consists in the fact that all reasons are always reasons for an agent to . It was realizing the implication that this property has for the distinction between agent-relative/neutral reasons that recently made me change my mind about this dichotomy. That reasons are always reasons for someone seems to have clear consequences for the distinction between reasons that are agentrelative and those that are agent-neutral. In fact, setting out from this feature, we can challenge the advocate of agent-neutral reasons to give us an account of it (the feature) and, in particular, ask why the following line of reasoning is erroneous (where PEFI refers to ‘personalizability feature implication’): (PEFI) Since all reasons are apparently reasons for someone to F, and a reason to F is only a reason for someone if it somehow involves or refers to this someone, it follows that all reasons to F are in their very form reasons that refer to the person who has the reason to F. This, in its turn, is just another way of saying that all reasons to F are, on entirely formal grounds, agentrelative reasons. Whether PEFI poses a serious challenge or not depends on various things, however. First, we must be clear about the sense in which it is true that a reason is only a reason for someone, if it somehow involves or refers to this someone. We must also specify what kind of reason we have in mind. As I have already said, I will conﬁne myself to the discussion of normative reasons rather than theoretical ones. Moreover, I will be concerned with a certain popular view of normative reasons. What then becomes problematic is the

9 Cf. Mark Schroeder in ‘Reasons and Agent-neutrality’ (2007). He discusses what he refers to as the “Quantiﬁcation strategy”, according to which “agent-neutral reasons arise when something is an agentrelational reason for everyone” (p. 280). Schroeder’s article is excellent, and I am just sorry that I read it too late to include a more substantial comment.

9.4

REASON-FOR

133

combination of this widespread view and the essentialist way of understanding agent-neutral reasons. Before continuing, a caveat is in place: some reasons are not normative but explanatory. It might be that the latter, but not the former, kind of reason is consistent with the distinction. Although the neutral/relative dichotomy is usually applied only to normative reasons, in principle it can be applied to explanatory reasons. So-called explanatory reasons are psychological states that cause us to act (albeit not necessarily in their own right). Could such reasons be propositional entities? This is a somewhat awkward idea, as we shall see in a moment. But for the sake of illustration, suppose for a moment that we regard reasons as propositions (of a certain kind). Would that not pave the way for the dichotomy? This line or reasoning might then amount to the following: many different propositions may, as a matter of fact, motivate a person. Moreover, there are no compelling grounds to be found in the literature for supposing that it is only propositions that refer to the agent which can motivate her to act. For instance, believing that there is a stranger drowning may cause you to act.10 It might therefore seem that the challenge can be brushed aside if we employ a notion of reason understood in terms of the propositional content of the mental states that cause the agent to act. What causes an agent to act may be the entertaining of a proposition that refers to the agent, but it is an open issue whether it has to be this sort of proposition that motivates the agent. Beliefs expressible with agent-neutral as well as agent-relative statements may be motivating the agent. Is the above, then, the way to understand explanatory reasons? I think not! In fact, we should disregard the above interpretation. It is not really propositional content that constitutes an agent’s explanatory reasons but rather her mental states. The description presented above is therefore, I suggest, a more accurate depiction of what we might refer to as motivating reason. That there is a point in keeping motivating (Dancy; Parﬁt) or as Thomas Scanlon calls them, operative, reasons apart from explanatory reasons is, I think, obvious.11 Thus, it is the agent’s belief that p, rather than the fact that p alone, that explains why she acted in the relevant way. Consider the following example. An agent a is asked why she jumped into the water and says something like “The reason is that God told me to save the person from drowning”. In this case it would seem the explanation of why a acted as she did is hardly that God told her something; rather the explanation must be that 10 What precisely causes an agent to act is a complicated matter, and one that I do not want to go into. My point here is merely that the belief must not necessarily involve, say, the idea that the stranger is in front of me, or the idea that I can help the stranger. It is (at least, in part) an empirical question which beliefs move us to act (alone or in company with other mental states). 11 See Scanlon (1998). Cf. also the notion of an apparent reason in Ingmar Persson’s The Retreat of Reason: A Dilemma in the Philosophy of Life (2005). Cf. Philip Pettit and Michael Smith (1990), and their distinction between a ﬁrst-person and a third-person perspective on reasons. That what motivates us is what we believe and not that we believe it is defended forcefully by Dancy in his Practical Reality (2000b, Ch. 6). Cf. Parﬁt (2001). Jonas Olson and Frans Svensson in ‘Regimenting Reasons’ (2005) also suggest a tripartite distinction between what they refer to as motivating, deliberating, and normative reasons.

134

ONE REASON DICHOTOMY LESS?

the agent believed God spoke to her. Nor does it seem correct to conclude that what motivated her to act was that she believed that God spoke to her; what motivated her was something that was the case—namely, God’s telling her to do something— and not her merely believing it to be the case. As I argued earlier, we should, in other words, avoid confusing the content of the agent’s mental state with what caused her to act. It is beliefs and desires that make the agent act, not some proposition-like entity. However, if we were to speak not about the content of the agent’s mental states, but rather about the agent’s mental states, we would in fact be referring to the agent, and then explanatory reasons would all become agent-relative in character.12 It would be the agent’s belief that God spoke that explained the acting. Moreover, if these observations are correct, the notion of a reason appears to decompose into at least three varieties: (i) normative reasons, (ii) motivating reasons, and (iii) explanatory reasons. What motivates the agent is hardly ever her belief that p but rather p. However, given the falsehood of her belief we need to invoke her belief to explain why she did what she did. The division between motivating and explanatory reasons is, of course, only sketched here. Again, the general idea is that we sometimes have that which motivates the agent in mind when we talk about her reasons. Thus, what motivated Don Quixote to take up arms was that monsters stood in his way, but what explains why he spent time chasing windmills was his confused belief that these buildings were actually living monsters (and his desire to kill them). There is another reason why we should not mix up the notion of a normative reason with the related notion of a motivating reason. The expression “motivating reason” refers to that state of affairs which as a matter of fact is believed to be a (normative) reason by the agent; it is that which the agent takes to count in favour of some attitude or act. Few would be ready to say, though, that just because we believe something to be a reason, it is a reason. Only on a highly controversial metaethical view (a crude sort of relativism) are motivating reasons constitutive of normative reasons.13 The distinction between (normative) reasons and motivating reasons is important. The motivating kind appear to be quite consistent with the neutral/relative dichotomy—that is, there seems to be no basis for doubting that the beliefs which the agent has about what are her reasons need not always take a form that essentially involves the agent. Nothing necessarily prevents an agent from regarding the allegedly agentneutral fact that b is drowning as a reason for her to try to save b. But again, this does not establish that there are agent-neutral normative reasons—it indicates at most that the agent believes there to be (in some sense other than the essentialist one) agent-neutral reasons.

12

Michael Zimmerman reminded me of the latter interpretation (personal communication). The notion that a statement expresses an agent-relative or neutral reason does not show that the relevant statement expresses a truth about a reason; it shows merely that it expresses a motivating reason. The issue, discussed at the end of Section 9.2, of whether a statement expresses an agent-relative or agent-neutral reason depending on the person to whom it is addressed is an issue concerning operative, not real, reasons. 13

9.4

REASON-FOR

135

Now, since the ﬁtting-attitude analysis understands impersonal as well as personal value in terms of normative (or justiﬁcatory, or good) reasons, my real interest concerns these reasons, so I will not take the opportunity to enter into detailed discussion of explanatory and motivating reasons. I will leave it open, then, whether in particular socalled motivating reasons can be understood in terms of the dichotomy. Explanatory reasons will run into most of the problems that normative reasons do with regard to the challenge. However, this is not something that I propose to pursue here. Consider, next, normative reasons. There is much less agreement over the way these should be analysed, although most observers agree that they are reasons that justify, rather than simply explain, an agent’s performance of an act. There have been all sorts of attempt to close the gap between normative and explanatory reasons. Here I will make one assumption that rules out some of these attempts. Thus, in keeping with the way many inﬂuential writers regard reasons these days, I have supposed in this work that normative reasons are facts.14 The nature of these facts can be left unspeciﬁed— they may be natural or, say, evaluative/normative. Some observations now need to be made about the above assumption. First, it is important that facts should be understood as entities independent of what people believe to be the case (facts about people’s beliefs are no exception; if such facts are reasons, they are so independently of whether they are believed to be the case). This is also in line with much recent work on reasons qua facts. Second, ‘fact’ and (i) ‘obtaining of a state of affairs’ are considered synonyms.15 There are other options, though. For instance, we might think of a fact as (ii) an aspect of an obtaining state of affairs; or we could follow the common practice of using ‘fact’ to refer to (iii) the content of a true possible judgement, i.e. a ‘true proposition’. I remain unsure whether anything of genuine importance, for my purposes, hinges on my preference for (i), and in what follows I shall have the more robust sense of ‘fact’ indicated in (i) in mind—a sense allowing us to recognize that facts contain properties and objects. This does not square well with the idea that facts are true propositions. Propositions are generally thought to contain only concepts, whereas obtainings are generally taken to involve all sorts of objects. Again, how crucial this preference for ‘robust facts’ is for the discussion in this chapter is not entirely apparent to me. As I made clear earlier, I have no problem with the idea that we can formulate sentences that are agent-relative in the sense of being speaker-relative. Equally, I do not have any problem with propositions being either

14 See, for instance, Broome (2004) and Parﬁt’s ‘On What Matters’ (forthcoming). Skorupski, in his turn, has suggested (in correspondence) that reasons might be understood as “facts plus modes of presentation thereof ”. 15 The idea that it makes sense to establish a division like this between the non-obtainings and the obtainings of states of affairs raises difﬁcult metaphysical questions that I will largely set aside. The notion that we need somehow to draw a line between these ‘entities’ seems obviously correct, as does the claim that the obtainings differ metaphysically from chairs, persons, and other concrete objects. Quite how we should account for the obtainings is another matter.

136

ONE REASON DICHOTOMY LESS?

agent-relative or agent-neutral (i.e. not agent-relative); it is at least imaginable that we should formulate propositions in two ways; those containing cross-references to the person who expresses them, and those that do not. However, I do not think that the fact that we can draw these distinctions settles the question of whether there are any reasons at all; nor do I think it settles the issue of whether there are any agent-neutral/ relative reasons. One (but not the only) thing we need to do in order to settle these issues is consider what makes such propositions true. My assumption is that this will have to be a thicker entity, like an obtaining state of affairs, rather than a thin aspect (singled out by the proposition) of such an obtaining state. One reason I have for preferring (i) to (ii) and (iii) is that it opens up a number of interesting questions about reasons. There is not space here to go into details. However, let me say this much. If we are suspicious about the idea that a statement expressing an alleged reason will in itself settle the question whether there are any agent-neutral/relative reasons, then consideration of the proposition expressed by this statement will hardly offer further assistance. And if the true proposition does not settle the question—“does the statement express a true normative reason?”—I do not see how the aspect, i.e. the truth-maker of the proposition, can do so. Something more is needed. I have a further reason for choosing (i) over (iii). The notion that what makes a true proposition agent-relative or agent-neutral is whether it contains a cross-reference to the person for whom the true proposition is a reason seems to trivialize the distinction it is designed to illuminate. For instance, an agent-relative proposition that contains an indexical expression such as “my daughter” (e.g. “my daughter is drowning”) can easily be transformed into what appears to be an agent-neutral proposition with some rigid designators. To save the distinction from this kind of trivialization, I suppose we could say that even if the agent-relative proposition were transformed into an agentneutral one, or vice versa, by employing or replacing rigid designators, it would still be possible that only one of these propositions is a reason for the agent, so the distinction is not trivialized. However, I must confess that I am not convinced by this reply. My preference for a thick way of understanding facts has, then, to do with the fact that I follow Joseph Raz in thinking that reasons have “vague and incomplete criteria of identity” (2006, p. 109). As an effect of this, reason-statements are more often than not very incomplete descriptions (expressing incomplete propositions) that capture only some features of the fact that constitutes the reason. I shall therefore assume that the examples of reasons I give (or, more accurately, the reason-statements) are very probably incomplete descriptions. In a given situation there are almost certainly a number of facts that together call for a certain action. However, in discussing reasons we tend to single out only the salient features of a certain situation as the reason. A more detailed account would disclose a much more complex picture. Of course, it need not be like this, but in what follows I will set out from the idea that the full ‘identity of the reason’16 need not be disclosed by the

16

I use here ‘identity’ in Raz’s (2006) sense.

9.5

MEETING THE CHALLENGE

137

examples offered here. At the end of the chapter, when we discuss Dancy’s objections to the notion of a complete reason, we shall have occasion to return to this matter. I now turn to the question more directly relevant to the view I am canvassing in this work, namely: how readily normative reasons, qua facts, can be understood in terms of the agent-relative/neutral dichotomy. The idea that all reasons are reasons for an agent (see PEF above) does appear to create a problem for a view according to which reasons are facts, i.e. obtainings of states of affairs. To examine this issue, recall ﬁrst what it would be for a fact to be an agent-relative reason in the essentialist view; it would be for some obtaining of a state of affairs to involve the agent for whom the fact is a reason. Suppose next that a reason-statement Pr expresses this state of affairs, and that Pr is true, so that what Pr says obtains does obtain. Moreover, the state of affairs that Pr expresses is in some sense about the agent for whom it operates as a reason—say, that the person who is drowning is my daughter. Here it is quite evident which feature makes this fact into a reason for me: that it is my daughter who is in need of help. An analogous reply does not work when it comes to agent-neutral reasons qua facts. On the essentialist approach, given the truth of an agent-neutral reason-statement, Pn, there is an obtaining of a state of affairs such that it is a reason for x. Moreover, it is a reason for x despite the fact that the state of affairs is not in any way about x.17 The question ‘Why is this fact a reason for x?’ cannot therefore be answered by pointing to some feature of the fact that even in a minimal sense concerns x. This is quite remarkable. The conclusion we appear to be obliged to draw is that there is nothing about the fact that makes it a reason for x, but that the fact is nonetheless a reason for x.

9.5 Meeting the challenge There is something troubling about this picture of normative agent-neutral reasons qua facts. But it might be insisted that this worry should not be exaggerated. Can we not simply say that, in certain cases, the fact is that everyone has a reason to do the act, and that since I fall under the thin descriptor ‘everyone’, I have a reason to do the act? This is an important question. In discussions of PEF and the challenge (PEFI) it poses to normative agent-neutral reasons, it is often the concern people most readily press. If it is a fact that everyone has a reason to do some act, then I, like everyone else, have a reason to do this act. But this reply is nonetheless question-begging; it will not meet the challenge—at least, not if we are interested in the essentialist deﬁnition of reasons qua facts. If the fact that a person y is drowning gives me and everyone else a reason to save y, the reason I should save y is not that everyone has a reason to save y but rather that it is a fact that y is drowning. The crucial thing to realize here is that the number of people who have this reason is not a reason-making feature. The fact that a person is in 17

Again, the reason it cannot contain a reference to the agent is that a reason-statement P expressing a fact that involves an agent is a true statement containing an essential reference to the agent; the statement is not logically equivalent to any other statement that does not contain a reference to the very same agent.

138

ONE REASON DICHOTOMY LESS?

need of help may well be such that it gives everyone a reason to perform the act. This is precisely what is under discussion here. But if it is a reason for me, there must be some feature of the fact that in a minimal sense concerns me—for its being a reason for me to save y. Otherwise it would not be a reason for me. (I might believe that it is a reason for me, but we are not discussing motivating reasons or explanatory ones.) I should add, though, that there might be other ways in which an agent could relate to a fact (or, say, a true proposition) that would explain why it (the fact or the proposition) is a reason for this agent. Suppose we agree that all reasons of the kind I am interested in here are reasons-for. It might be denied that a normative reason is a reason for someone only if it involves, or refers to, someone. Michael Zimmerman (personal communication) has suggested that it seems sufﬁcient that the reason should apply to the person for whom the fact is a reason. The agent does not need to be a part of the fact or in some way involved in the fact. It is enough that the fact applies to the person for whom it is a reason. That the reason (i.e. the fact) applies to the person means, then, that the fact bears some kind of appropriate relation to the person. It does not have to be the case that the fact is about, or involves, the agent. On this suggestion we would not ﬁnd a reference to the person in the content of a reason-statement expressing this fact. But it would still be the case that the reason was a reason for the person, in virtue of the alleged relation between the person and the fact. Zimmerman’s suggestion is appealing. In fact, I am ready to embrace it; there are cases in which a fact is a reason for us without our being involved in the fact. Recall the case we discussed earlier in which it was a fact that everyone has a reason to perform some act. Certainly this fact can give you a reason, and so there is a sense in which the reason applies to you without it being the case that you are involved in the fact (i.e. a true description of the fact would not mention you). However, this is a peculiar kind of case. Such an assertion would naturally lead to the question: ‘What is this reason that everyone has?’ And this suggests that we are dealing with what is best described as a derivative reason. There is some other reason on which the derivative reason depends. If the fact that everyone has a reason is a reason, it is so because of some other fact that constitutes a reason for everyone. But although Zimmerman has a point when it comes to derivative reasons I am not sure we can extrapolate and claim that this holds for nonderivative reasons as well. I, at least, am not sure what it means to be related to this more basic sort of fact without it being the case that I am somehow related to the constituents of the fact. Notice, too, that the suggestion cannot accommodate the notion that the agent is somehow related to these constituents, since if there were such a relation, a reason-statement expressing this fact would be agent-relative (i.e. a reference to the person would be a part of the content of the reason-statement). I could be wrong about all this. The matter is complicated. Perhaps Zimmerman’s suggestion works for these non-derivative reasons, too. The explanation of why such a reason would be a reason for me would then be that I am somehow appropriately related to the fact. Suppose we set aside the question of what, exactly, is the nature of this relation; we charitably accept that there is such a relation. Will we then have met the challenge?

9.5

MEETING THE CHALLENGE

139

Perhaps. We can at least give some explanation as to why the reason is a reason for the agent; he or she is appropriately related to the reason. Whether this is a fully satisfactory answer is debatable, though. It depends, I suppose, on how one answers the further question: What is the underlying fact in virtue of which there is such an appropriate relation between me and the fact that constitutes the reason? If this fact does not involve me as an agent, I suspect that there is something unsatisfactory about this sort of reply. Of course, one might regard things in a slightly different way. Reasons could be taken to have two components. On such a view, my reason to would consist of (P1) the fact that I am related in a certain way to fact (P2)—where (P2) is some other fact that does not necessarily involve me. It seems reasonable to regard reasons in this way. Such a view has the advantage that the reason now includes, or refers to, the agent. A possible drawback is that this way of regarding reasons might easily be conceived of as making all reasons agent-relative. If so, we would not have met the challenge. These initial attempts to defend agent-neutral normative reasons did not in any obvious way meet the challenge. However, there is a series of responses that we need to examine; and it is too early to make up our minds about the status of this challenge. Here, to begin with, are two replies that question the line of reasoning: (i) Normative reasons need not be reasons for agents. (ii) An obtaining of a state of affairs may be about x and still be an agent-neutral reason. Now, (i) is a somewhat far-fetched response to the problem. It simply refuses to take up the challenge presented by PEF. This is not an unexpected response. After all, I have not offered any argument in support of the observation captured in PEF on which the challenge sets out—at least, none other than that there is an obvious ring of plausibility to PEF. (As is well known, one person’s modus ponens is another’s modus tollens.) But perhaps there is, in effect, a suggestion that would help break this stalemate—and one, moreover, which would leave the distinction intact. It can be suggested that the question ‘Why is this fact a reason for x?’ is not always legitimate (as is it would be if all reasons are reasons-for). Agent-neutral reasons are not reasons-for, but ‘merely’ reasons. Accordingly, it would be to misunderstand these reasons to ask: ‘What is it about this fact that makes it into a reason for x?’ By analogy with the core distinction in this work between good, period and good-for, and value, period and value-for, reasons might just be reasons, period, but not reasons-for. As I have mentioned, there is something tempting about this response. However, it would be a mistake to pursue it. For further thought reveals, I think, that this ‘analogyreply’ is quite unconvincing. Reasons are not like values in the suggested sense. Qua facts, they do not just ‘call for’ an action, or make it eligible, but call for an action by those who are able to perform the action. The idea that reasons can be altogether agent-unrelated is mystical. If a fact is a reason, it is a reason for at least one agent. That something might merely be a reason, period, but not for any present, past, or future agent, is, to say the least, an awkward idea.

140

ONE REASON DICHOTOMY LESS?

Although the analogy objection is unconvincing, there might be another way of saving the ﬁrst kind of response to the challenge. I will come to this saviour in a moment, but ﬁrst, a word on this contrast between values and reasons. It might be argued that if there is this dissimilarity between values and reasons, a ﬁtting attitude analysis of value cannot be correct. In reducing values to reasons something is lost—something, that is to say, that would account for this unlikeness. This is very probably true. But in my view this value-reason gap constitutes no real obstacle to the analysis. It would be naive, and quite futile, to expect there to be a perfect match between the associations we obtain from our value notions and those we obtain from our normative notions. The important thing, from the point of the analysis, is that there is match between the signiﬁcant features. That they are in some sense mystical entities is not, as far as I can see, an important feature of values.

9.6 An overcrowded boat There may be another way of saving the ﬁrst kind of response to the challenge. It might be that reasons after all need not be reasons for particular agents. The question is whether there is not a rather obvious objection to my reply above (that there cannot be reasons that are not somehow reasons for certain persons). Wlodek Rabinowicz, for instance, has suggested (in personal communication) that there might be examples that underwrite the claim that an action ought to be done by someone without underwriting the claim that it ought to be done by anyone in particular. For instance, if there are too many people in a boat, someone ought to leave it. This seems to be an example in which there is a reason for someone to act that is not a reason for any particular person to act. This does not mean that everyone has a reason to leave the boat; it means precisely that someone but not everyone has a reason to leave the boat. The example is indeed interesting. But what it shows is quite an open matter. For instance, even if we were to accept that it highlights a reason for someone that is not a reason for any particular person, this would not automatically qualify the reason in question as an agent-neutral reason—at any rate, on the essentialist approach. Here is why: the statement expressing the fact that in a boat with persons x, y, and z, someone has to leave, will not be logically equivalent with any other statement that does not mention x, y, and z. What we seem to have here, then, is a peculiar case of the agentrelative reason: one which is a reason for someone, but not for any particular person. Advocates of normative agent-neutral reasons should not draw encouragement from the existence of peculiar agent-relative reasons for action. In fact, even these reasons might be questioned. For it appears to be quite counter-intuitive, as Rabinowicz has suggested (personal communication), to regard the boat example as being, in the ﬁrst place, a scenario in which there is reason to act. Consider again the details of the case. That the boat is overcrowded is a reason for there being someone who leaves the boat. However, this is hardly describable as a reason for an act. A reason for there being someone is not about action, even if leaving the boat is. It is therefore unclear just what this

9.7

MORE REPLIES TO THE CHALLENGE

141

example shows. If the fact about the boat is a reason, it seems most plausible to regard it as a reason that is different in kind from reasons for action. Accordingly, if there are such reasons, we cannot exclude the possibility that there are non-agent-relative reasons, even if we conclude, eventually, that all reasons for action are agent-relative.

9.7 More replies to the challenge It is always logically legitimate to ask what it is about some fact that makes it into someone’s reason for action. And in the case of a so-called agent-neutral reason, the test is how does statement P, which allegedly contains no reference of any sort to x, express a reason for a to ? What is it about the expressed fact that explains why it is a reason for x? What is it that the ‘for’ in ‘This fact is a reason for x to ’ signals, with regard to the agent, when it comes to agent-neutral reasons? Response (i) should be rejected. Acceptance of it is tantamount to an admission that reasons for action can, quite literally, be nobody’s reason, which is incomprehensible. The next response, (ii), is quite different from the ﬁrst. It acknowledges the PEF feature and does not deny that reasons are always reasons for an agent. However, it suggests that the step from ‘this fact is (among other things) about x’ to ‘this fact cannot be an agent-neutral reason’ is invalid. Facts can be about a certain agent, and nonetheless be an agent-neutral reason for this very agent. There might be something to this reply. However, as far as the essentialist approach goes, (ii) appears to be a non-starter. As might be recalled, it follows from our interpretation of ‘statement that essentially refers to an agent’ that if a fact is a reason, and if it is also, among other things, about person x (whether or not it is also about every other agent y, z, . . . ), then the statement expressing the state of affairs P will not be logically equivalent to some other statement Q that does not refer to x. Hence P will contain an essential reference to x. So P will express an agent-relative reason for x. In order words, if reply (ii) is to be successful, it must provide a different interpretation of the dichotomy than the essentialist one. I would like to emphasize that I am certainly open to the idea that within what I have referred to as agent-relative reasons, we should be able to distinguish between different sorts of reason, some of which might be referred to, in a sense, as agent-neutral reasons. For instance, Portmore has recently suggested (personal communication) a way of understanding the dichotomy that preserves, he thinks, the philosophical importance. I am inclined to agree. Consider the following two examples: AR: Person x doing in c would ensure that Jones saves x’s child. AN: Person x doing in c would ensure that Jones saves Jones’s child. These counter-factuals involve the agent for whom they are a reason, and so, as Portmore pointed out, the personalizability feature has been accounted for; AR is still in a sense agent-relative in a way that AN is not. I think Portmore is right about this.

142

ONE REASON DICHOTOMY LESS?

However, I also think his distinction certainly falls within what I have referred to as agent-relative reasons. These subjunctive facts involve the agent for whom it is a reason. However, let us not rush things. It may be pointed out that not all features of a fact are equally relevant. So although a fact may to some degree be about the person for whom it is a reason, this agent-relative feature is not a reason-making property; and accordingly the statement expressing the fact need not mention this agent-relative feature.18 There is something in this reply. In a while I will also consider a version of it that perhaps constitutes the most promising way of meeting the challenge set by PEF. But speaking more generally this reply is far from compelling; it tries to solve a problem of how to clarify a conceptual issue by placing a substantive requirement on, in this case, what features of facts may or may not be relevant reason-providers. At best it will only convince people who share the same normative outlook, i.e. those who agree to the same normative statements. As far as possible, answers to formal questions should not depend on substantive views. If all reasons are reasons for someone, the PEF challenge is, in other words, to ﬁnd what it is about the fact that explains why it is a reason for the agent if this fact does not involve the agent. If there is no such feature, it will be impossible to explain why the fact is a reason for the agent. Of course, we might not know, or we might be unsure, what the feature is. We might also question the notion that the fact is, in effect, a reason in the ﬁrst place. But this is not what is at issue here. The question is: Why is this fact a reason for x? To maintain that a fact which does not involve person x might still be reason for x is to admit that there is no explanation of why the reason is a reason for x. That is, if x is not involved in the fact, it becomes inexplicable that the fact is a reason for x. Perhaps we should accept that certain facts are, entirely inexplicably, reasons for certain agents. Even if there is nothing about the fact that in the slightest way involves the agent, it might still be a reason for her to do something. That would be a case of an inexplicable reason. It could then be said that the set of inexplicable reasons coincides with the set of all non-agent-relative reasons. This is a tricky manoeuvre to deal with. If you are ready to accept the existence of inexplicable reasons, the PEF challenge can perhaps be met. But the price of this is high. Another possibility—as I will argue in Section 9.8—is to maintain that all reasons are agent-relative reasons. But there are other replies to the challenge. Could it, perhaps, rest on an implausible interpretation of the expression ‘reason for x to ’? Surely something could be a reason for x to , even if the reason failed to ‘get a grip on’ x? This reply implies that something could be an agent-neutral reason for x to despite the fact that x would not 18

As I understand Portmore’s suggestion, this is not something he would subscribe to. That the agent ensures something is essential to his view of what makes something into a reason. In my view, Portmore’s interpretation involves the wrong sort of reason-makers. For instance, in his account, the fact that my daughter is drowning is not a reason for me to jump into the water. Nor is the fact that she is drowning and that I can save her a reason for me to jump into the water. Rather, the reason I should jump into the water is the subjunctive fact that my jumping into the water would ensure that she is or will be saved.

9.7

MORE REPLIES TO THE CHALLENGE

143

even if she were aware of that reason. On such occasions the reason would still be a reason for x. This reply leaves the challenge intact, however, since it is based on a misunderstanding. That something is a reason for x does not, as I pointed out above, require x to be somehow aware of x. In fact, that some fact F is a normative reason for x to is not conditional on its being the case that x would if she believed that F. This might be true of some facts, but it would be implausible to maintain that it is always the case. Sometimes two persons will have the same beliefs about what the facts are and still not take the same normative views. More than that: even if I know a certain fact, and know this fact is a reason for me to , I might still not . Moreover, it is generally accepted that many facts are unknown to all of us. In the light of this it is reasonable to expect there to be many normative reasons that are unknown to all of us. That reasons are always reasons for x should not, in other words, be taken to imply that x must somehow be aware of, or even respond to, the reason. The expression ‘reason for’ establishes that there must be something that relates the fact to the person; but the relation in question need not be epistemic. Once there is no relation at all between the agent and the fact, the suggestion that the fact is a reason for x becomes incomprehensible. If this contention is correct, there is one sense of ‘reason for x’ that appears to be difﬁcult to understand in terms of a normative agent-neutral reasons. But why not go straight to the point and say that some facts essentially involve certain individuals, while others only contingently involve people? This looks like a straightforward solution. But it remains to be shown what sense can be given to the locution ‘this obtaining of state of affairs contingently involves x’. Metaphysicians tend to agree that propositional (abstract) entities like states of affairs have all their elements necessarily. The same holds, I would say, about obtainings of states of affairs; they too have their constituents necessarily. There is the possibility that the challenge sets out from a narrow view of acts. Suppose (a) ‘helping x’ is one kind of act, and (b) ‘helping x if you can’ is a different kind of act. The reason for performing (a) must somehow be a fact that among other things involves one’s being able to help x. Thus, the fact that x is in need of help is an incomplete reason for performing the act in (a). The fact that x is in need, together with the fact that you can help x, is a reason for you to perform the act mentioned in (a). However, the former fact is not obviously incomplete when it comes to the act in (b). That is, it might be argued that the fact that x is in need is sufﬁcient to provide a reason for performing the act in (b). Hence, there seem to be facts that are reasons which do not necessarily involve the person who has a reason to act. This is a quite strained reply. First, it sets out from a highly implausible view of what acts are. In what sense does (b) refer to an act? There is no obvious answer to this question. Certainly, it is one thing to act on a condition, but so acting is not necessarily doing something different from what one does when one acts, but not on a condition. So what this reply in effect shows is that reasons might perhaps apply to things other than acts and persons, and that there might in fact be agent-neutral reasons as long as they are

144

ONE REASON DICHOTOMY LESS?

not reasons for acts, but for other things (such as acting on a condition). The discussion above of the overcrowded boat has already suggested something along these lines.

9.8 A positive argument For all its simplicity, the challenge turns out to be a hard nut to crack. It sets off from two ideas: ﬁrst, that claims such as ‘This is a reason’ and ‘The reason is . . . ’ are, at least in the case of normative reasons, elliptical—they are all shorthand for ‘This is a reason for a to ’ or ‘The reason for a to F is . . . ’. Second, if normative reasons are facts, those facts that are reasons must somehow reﬂect or explain this ‘for-a’ feature of reasons, since otherwise it would be inexplicable why those facts constitute reasons for the agent. The challenge for adherents of agent-neutral reasons is, then, to ﬁnd a fact that does this without involving the person for whom it is a reason. Our consideration of the challenge to this point has not established that it cannot be met. In fact, a further reply quite correctly points out that since all that has been shown so far is that certain replies do not work, caution is called for. This is, of course, a sensible warning. However, there is also a strong positive argument (already invoked more than once in this work) for the claim that all normative reasons are agentrelative. It centres on a generally accepted feature of practical reasons: a feature that can be captured, in a rather rough and ready way, in the ‘Reason-Implies-Ability’ principle: (RIA) There is reason for a to ! it is possible for a to . Suppose, then, it is true that to have a reason to you must be able (physically and perhaps psychologically) to . Furthermore, suppose that this ‘ability constraint’ on reasons applies to all practical reasons. This restricts in its turn the kinds of fact that constitute reasons: fact F is a reason for a to only if F involves in part a’s ability to . A full description of the fact (that is a reason) would, in other words, mention that the agent is able to act. So if RIA does indeed identify a formal feature of reasons (or more accurately of reason-statements), it seems a fortiori to strengthen the claim that there are no agentneutral normative reasons. (The cautious ‘it seems’ claim is warranted, as we shall see in a moment.) All normative reasons are agent-relative. It might be argued, perhaps, that the ‘ability constraint’ is an uninteresting, trivial, or in some sense disregardable feature of reasons; and that we may, or should, divide reasons into different groups without paying any attention to this shared feature. This may be right, but it does not change the fact that these groups would consist exclusively of agent-relative reasons; you cannot uphold the reason dichotomy discussed here if you believe in the doctrine that practical reasons imply the ability to act. Or so it seems. Dancy’s intriguing idea, set out in Ethics without Principles (2004), that we should distinguish between reason considerations that are “favourers/disfavourers” and those that are “enablers/disablers”19 suggests how the positive argument might be refuted. 19 See Dancy (2004). In this work, Dancy also discusses a third role, namely intensiﬁers/attenuators. However, to outline the general idea, I need only focus on the favourer/enabler distinction.

9.8

A POSITIVE ARGUMENT

145

Favourers, we are told, are features (facts) that count in favour of acting in one way or another. This cannot be said about enablers, though; they do not favour acting. Instead, what they do is to clear the ground for the favourer—they make it possible for a certain feature to count in favour of acting. Thus, the idea is that the enabler allows the favourer to do its job. Here is one of Dancy’s own examples: “(1) I promised to do x, (2) my promise was not given under duress, (3) I am able to do it, (4) There is no greater reason not to do it”. Of these, (1) is a favourer and (2–4) are different kinds of enabler. The distinction between favourers and enablers might well come in handy. It suggests a way of meeting the PEF challenge: even if reasons imply ability, it does not follow that all reasons are agent-relative. We could draw a line between relative and neutral as follows: agent-neutral reasons consist of two kinds of fact: (i) the favourer, which is a fact that does not involve the agent (for whom the fact is a reason); and (ii) the enabler, which is the fact that the agent is able to perform the act in question. This is certainly a plausible response to the positive argument. Agent-neutral reasons are favourers of a certain kind (they do not involve the agent). They have as enablers the ability of the agent, but the fact that the agent can do something is not part of the favourer. Just how Dancy’s distinction should be understood in more detail is still very much discussed. Raz, for instance, ﬁnds “the category of being an enabler . . . so diverse as to be of little use, and likely to mislead” (2006, p. 106), More generally, Raz detects a number of tensions in the distinction between favourers and enablers.20 Admittedly, Dancy’s distinction is fundamentally phenomenological, and so it should come as no surprise that it gives rise to interpretative tensions. This notwithstanding, as far as I can see the idea that not all ‘reasons features’ play the same role paves the way for a defence of the neutral/relative dichotomy which is by far the most promising I am familiar with. However, I want to voice two worries right away. The ﬁrst has to do with Dancy’s view that all reasons are contributory reasons. I am inclined to believe that there is at least another kind of reason, namely something we can refer to as a complete reason; and I suspect that such (all facts inclusive) reasons are still open to the challenge. Dancy comments on an idea closely related to the notion of a complete reason. This is W. D. Ross’s view that “duty is toti-resultant”. He says: “Ross lumps all relevant features together, as part of the right-making base (the ground, as one might call it), and this metaphysical picture of the situation is far too indiscriminate” (2004, p. 45; see also pp. 91–2).21 Here, Dancy’s choice of the term ‘indiscriminate’ is

20 Dancy’s account of favourers and enablers raises interpretative issues that I will not go into here. Raz considers four interpretations. His ﬁnal one makes room for a notion of a favourer that does not actually favour anything (since there is no enabler present); see Raz (2006, p. 105). He might be right that there is textual evidence for this in Dancy’s book. However, as I read Dancy, the important part of his idea is that some features favour on condition, and that it is these features that are reasons, again on condition. 21 See also Dancy (2004, p. 39), where a related idea is discussed—namely, that the real favourer is a complex consisting of features (1) and (2) (see his example). Dancy rejects this idea on intuitive grounds. For criticism of this defence, see Raz (2006) and Strandberg (2008, especially Section 7).

146

ONE REASON DICHOTOMY LESS?

noteworthy. As indiscriminacy is not a feature of the world but more accurately of how it is described, Dancy is probably concerned with descriptions or explanations of reasons. His objection to Ross’s idea does not, therefore, bear immediately on the notion of reason I have been examining—the notion of reasons qua facts. We should certainly agree that it is not very helpful to start enumerating all the relevant facts if we want to explain, cite, or, more generally, describe a reason to someone. But this is quite consistent with the idea that there is one and only one thing I have a reason to do at t1, and that this is determined by how the world precisely is at t1. Not that we ever aspire, or should aspire, to take into consideration all there is to know about the world in the moment we think there is a reason to do something. Surely that would be preposterous. The world is epistemically muddy. Our knowledge of, say, what our actions will result in is fallible. But if, at any given moment at which we are in the presence of a reason, there is a set (very probably limitless) of relevant facts, and if it is true that what we have reason to do is, at least sometimes, determined not by a single fact but a number of facts, then it is hard to see how we can rule out the notion of a complete reason on purely conceptual grounds. That we cannot aspire to know what facts constitute a given complete reason is something we have to live with; it does not render the notion of a complete reason unusable. But, it might be objected, even if there are complete reasons in this sense, could we still not differentiate between complete reasons whose favourer(s) do not involve the reason holder and those that do? And so could we not just say, as before, that the ﬁrst kind is an agent-neutral reason and the second an agent-relative one? This response is inadequately thought through. A complete reason is composed by all the relevant facts, and on their own these do not constitute a favourer. At any rate, they do not do so in what I believe to be Dancy’s sense, namely a favourer on condition. But, more importantly, even supposing that the complex fact constituted some sort of favourer, such a favourer (what might be referred to as a complete or unconditional favourer) might well be constituted by facts which, taken on their own, constitute contributory reasons that favour things other than, or even disfavour, the act that the complete favourer (reason) calls for. The correspondence between the two kinds of favourer cannot, at least, be assumed to be necessary without argument. It is therefore not obvious how the notion of a favourer on condition applies to complete reasons: Dancy’s favourer/enabler distinction is not very helpful, then, when it comes to one kind of reason. We might tell apart those features that are favourers and those that are enablers, which together constitute a complete reason. However, it is not obvious that we can use this to differentiate between agent-neutral and agent-relative complete reasons. It remains unclear, at least, what the precise relationship is between these favourers/ enablers that constitute different contributory reasons and the complete reason. So even if a given complete reason will, in a sense, encompass, say, a number of different favourers (in Dancy’s sense), some of which will involve the agent and some which will not, it is far from clear what the relationship is between these favourers and the

9.8

A POSITIVE ARGUMENT

147

complete reason. Hence, it remains obscure why the nature of a complete reason should eventually be determined by one favourer rather than another. This point needs to be ampliﬁed a little, however. We should ﬁrst distinguish between two kinds of complete reason, both of which can be characterized as reasons that are individuated (and thus constituted) by the facts at a certain time. Thus, it might be that the complete reason at t1 is (i) constituted by all the obtainings of states of affairs at t1. This would be one way of conceiving what is involved in the exempliﬁcation of a complete reason. However, it is hard to think that there would ever be such a complete reason.22 Among other things, it would mean that, at a given time, all facts were relevant, which is hardly imaginable. A more realistic idea is that a complete reason is (ii) constituted only by all the relevant facts. However, since we are dealing with the notion of a complete reason here, one of these facts would have to be somewhat peculiar, namely the fact that all other facts are irrelevant, i.e. have nothing to do with the reason. The justiﬁcation for this proviso is then the following observation: a reason at t1 would hardly exemplify the notion of a complete reason qua the facts if there were further facts that might alter, or inﬂuence, what we have a reason to do at t1. A complete reason of the ‘all the relevant facts’ type, or of the former, more implausible, (i)-kind, should not be understood as the reason that results from weighing different (contributory) reasons against each other. Rather, the idea is that the obtaining of the relevant states of affairs at a given time together equally constitutes the reason. Here is an analogy that might be helpful. A jigsaw puzzle J1 (depicting, say, some prescriptive instruction) consists of equally shaped pieces that together constitute an image. Another jigsaw puzzle J2 (also depicting a prescriptive instruction) is not really ﬁnished—there remain some pieces that we do not have access to. Still, in its present state J2 contains enough pieces to give us a good idea of what it would look like (or of what is prescribed). However, J2 is made up of pieces all of which are unique in shape. Since the pieces are different, we might want to suggest that some pieces play roles that other pieces do not. For example, we might suggest that some are better to begin with than others: corners, for instance, can easily be spotted in the box at the outset. It would not make sense to say this about J1, whose pieces are uniformly shaped. Now although we never as a matter of fact seem to be in a position to say with certainty what a complete reason is, the notion of a complete reason is not pointless. For instance, it may well be what best explains our motivation to look for more facts even when we think we have a contributory reason for doing a certain thing. We are not—at least, not always—just looking for yet another contributory reason. We are looking for (what best approximates to) the reason, i.e. the reason constituted by the complete picture. Now, my ﬁrst worry is that such a complete reason squares badly with the idea that reasons are favourers on condition. Each fact does its equal share, as it were, of the reason-constitutive work (witness puzzle J1). The explanation of why it does not seem 22 Notice that nothing hinges on the idea that reasons are thick entities; there is an analogous problem with complete propositions.

148

ONE REASON DICHOTOMY LESS?

to be a favourer on condition is that a complete reason does not in any obvious sense have (or more accurately, admit of having) an enabler. The notion of a complete reason outlined here entails that any fact will be either (1) such that, together with all other facts, it equally constitutes the reason, or (2) such that it is either relevant or irrelevant. We are dealing with (1) if ‘complete reason’ is understood in the unrealistic sense implying that all facts are relevant, and we are dealing with (2) if we read ‘complete reason’ in the sense of (ii). Since it is not obvious how one and the same feature can be constitutive of a reason and an enabler for this very same reason, there are grounds to be concerned that this sort of reason-notion is still open to the challenge set by PEF. Dancy seems to recognize something to this effect:23 “Walter Sinnott-Armstrong suggested to me that, in a case where the mere ability to act is a reason, it is also an enabler for itself. I see no reason to deny this amusing possibility.” I am not quite sure what to make of Dancy’s admission (see footnote 24, though). If we take facts to be ontological entities in their own right, it is hard to see how one and the same ‘thing’ can be its own (condition or) enabler. Metaphysics apart, this footnote of Dancy’s is interesting for another reason; if ability can be a favourer (on condition) as well as an enabler, we cannot meet the challenge we are considering by claiming that abilities appear only as enablers. Now, Dancy discusses at length a number of reason notions that differ from his own ‘contributory reasons’. For instance, Dancy objects to Roger Crisp’s notion of an ‘ultimate reason’. Brieﬂy, he objects to what he thinks is the idea underlying Crisp’s notion of a full explanation. The notion of a complete reason qua fact that I have outlined is not, in contrast with ‘explanation’, a success notion; so Dancy’s argument against Crisp does not in any obvious way affect the notion I am deploying. Dancy’s discussion of Raz’s notion of a complete reason is of particular interest: [o]n Raz’s account of a complete reason, whose parts (which he thinks of as premises, and which I would think of as reasons, some of them) are not reasons, it will be true that whatever is a reason is always a reason, and always on the same side. (2004, p. 97)

Dancy takes Raz to endorse what he calls “cluster atomism”, and it is really this idea that is bothering Dancy. A cluster atomist not only regards (at least) some reasons as clusters of features, but also maintains that “a cluster that plays a certain role in one case, must play that role wherever it appears” (p. 97). This means, then, that if something is a reason for x in a certain situation, it will be a reason for x in any situation. This Dancy takes to be inconsistent with holism (his own view of reasons). Holism is the idea that a feature that is a reason to act in one case may be no reason, or even a reason against acting, in another case. Now, the atomism–holism issue is important. However, the belief that there are complete reasons in the sense I have outlined here commits one to neither atomism nor holism. In fact it can quite plausibly be argued that the issue does not arise in the ﬁrst 23

See Dancy (2004, p. 40, n. 1).

9.8

A POSITIVE ARGUMENT

149

place. The argument proceeds on the basis that what is a complete reason at one point in time will not be a complete reason at any other time (given the assumption, hard to deny, that the set of facts changes over time). In my view we should follow Dancy and reject atomism. That is, I share Dancy’s view that one should reject atomism when it comes to contributory reasons. But not all reasons seem to be contributory. As far as I can determine, Dancy’s objections to complete reasons do not affect an idea such as the one outlined above. The distinction between favourers and enablers is important. But even if we accept it, which I think we should, we need not accept that all reasons are favourers on a condition. We must make room for the possibility of a non-contributory kind of reason, namely a complete reason. Since such a reason cannot be depicted as a favourer on condition, we cannot avail ourselves of Dancy’s distinction to meet the PEF challenge. The categories ‘atomism’ and ‘holism’ apply more naturally to operative reasons, i.e. to what we believe to be reasons. This brings me to my second worry. The anxiety I have here concerns an idea that several writers have had, namely that Dancy’s distinction is fundamentally a pragmatic, rather than metaphysical, one. Favourers and enablers are notions we employ when we cite, or explain, what we believe to be our reasons. Something that is a favourer in one context might therefore be an enabler in another, depending on what we are interested in emphasizing.24 Raz is ready to read Dancy along these lines. Consider, for instance, the following: [Dancy] protests against the claim made by Crisp (2000b, p. 44), that citing a favourer may be good enough an explanation of one’s reason for acting as one did (Dancy, 2004, pp. 47–8, 95–7). That seems to me true since explanations can have different objectives as well as different objects. Not infrequently citing the so-called favourer is, given one’s interlocutor’s interest, the best explanation of one’s reason. But this only illustrates the importance of distinguishing between the explanation of a reason and the reason itself. The notion of a complete explanation probably does not make sense. There can always be additional puzzles calling for additional explanations. It does not follow that the notion of a complete reason does not make sense. One should not conclude from the fact that human questions have no end that how things are in the world is equally open ended and in the same way. (Raz, 2006, p. 110; my italics)25

24

Recall Dancy’s claim that one and the same feature might be a favourer on condition (reason) and its own enabler. Perhaps this should be understood as follows: there are different ways of bringing out what is salient about the ability feature. Given the context, you might want to stress to someone that ability is what favours doing the act, but you might also point out that ability is a condition of doing the act. We are, in other words, talking about the feature in different ways. However, the fact that a person is able to do something remains the same fact whether or not we invoke it as a favourer or an enabler. 25 See also Caj Strandberg (2008), who gives such a pragmatic account of the distinction between “what makes objects have moral properties and enablers” (p. 150). Moreover, Strandberg thinks his distinction is generalizable to Dancy’s related distinction between favourers and enablers. He cites the following works as putting forward related suggestions: Raz (2000, p. 59), Broome (2004, pp. 32–5), and McKeever and Ridge, (2006, pp. 72–5). That a feature may be a favourer in one case and an enabler in another is something that Dancy himself points out (see his discussion of the ability to raise one’s arm, p. 40; see also n. 1). He also claims that “it is easy to ﬁnd examples of cases where it is not clear which side of the favouring/enabling distinction a given feature is to fall” (p. 51).

150

ONE REASON DICHOTOMY LESS?

To the extent that there is something in these suggestions, we should take the distinction between favourers and enablers to concern what reason considerations can do rather than what facts (features) can do.26 The distinction has more to do with how we explain (our) reasons than it has with what (our) normative reasons are. What an explanation will look like, in terms of favourers and enablers, might vary depending, say, on the person to whom the explanation is directed (e.g. a child or an adult; see Raz, 2006, p.101). Since the challenge concerns what reasons are, rather than how we explain them, it might be argued that Dancy’s distinction does not help us defend the dichotomy of agent-neutral and agent-relative reasons. So if these suggestions are correct, the favourer/enabler distinction lacks any metaphysical basis; hence, “enability” and “favourability” are probably no more than features we ascribe, or decline to ascribe, to facts, depending on what appears to us as salient given our desires and interests at the time. These ascribed features are, therefore, better characterized as belonging to operative (normative) reasons, rather than to normative reasons qua facts that are independent of our beliefs. I confess I remain uncertain as to whether Dancy’s distinction is, or more importantly must be, contextual in this sense. Perhaps it need not be. However, since it is difﬁcult to see what would settle this issue, it is surely best to adopt a cautious attitude to the question whether Dancy’s distinction can be used to defend the dichotomy. This notwithstanding, Dancy has supplied us with the tools needed to illuminate that question. The distinction between agent-relative and agent-neutral reasons attracts as much philosophical attention as it does in part, of course, because it involves several more or less complicated issues. Thus, there are: 1. Deﬁnitional Question (I): Is it possible to draw a distinction between agentrelative and agent-neutral reason-statements? 2. Deﬁnitional Question (II): Is it possible to draw a distinction between agentrelative and agent-neutral reasons? 3. Metaphysical Question: Are there metaphysical grounds for rejecting the distinction? 4. Normative Question: Are there any agent-neutral or agent-relative reasons at all?27 The ﬁrst two questions need to be held apart. Albeit that they are intimately related, they are not the same. From the fact that we have linguistic means to formulate statements that are agent-relativized, it does not follow that we have to answer the second deﬁnitional question in the afﬁrmative. And it is equally important to see that, even if we answer ‘yes’ to the second question, it might still be the case that the metaphysical underpinnings of this distinction are too obscure for us to identify ground 26

It should be mentioned that Dancy (2004) discusses the distinction between favouring/enabling/ intensifying in terms of the different things that “relevant considerations can do” (p. 41). 27 I am grateful to Skorupski (personal communication) for reminding me of some of these questions.

9.8

A POSITIVE ARGUMENT

151

on which we can draw the distinction in an adequately clear way. Finally, even if we have no metaphysical objections to the distinction, we might insist that there are no agent-relative reasons—or, for that matter, agent-neutral reasons. We might well reject one kind of reason on purely evaluative grounds. This chapter has drawn attention to Deﬁnitional Question (II), and in particular to a feature of reasons that is an obstacle to the widely held view that normative agentneutral reasons are facts.28 There is a general objection to my approach, however. Does not my insistence that all reasons are reasons for agents in effect make the question whether all reasons are agent-relative analytically true? And if that is indeed the case, does it not follow that I am not really in the business of arguing for, or clarifying, something? Am I not merely stipulating a deﬁnition? Surely it can be quite illuminating to draw attention to a feature of a distinction, analytic or not. Hence I do not necessarily see myself as doing something as philosophically unexciting as solving an issue by deﬁnition. Rather, I see myself as underlining what follows from a much neglected feature of all reasons. I have not established that there cannot be normative agent-neutral reasons, but it ought to be clear that the notion of a normative agent-neutral reason is undermined by certain views of reasons qua thick facts. Whether, in the end, this is a problem for the essentialist approach to the dichotomy, the deﬁnition of normative reasons, or for the notion of agent-neutral normative reasons for action, is a further issue. It depends in part on one’s attitude to the idea that reasons presuppose an ability to act. However, if you do not take Dancy’s distinction between favourers and enablers to solve the issue, it is bound to seem that reasons, qua facts, are all agent-relative—on the essentialist approach, at any rate.

28 See, for instance, Christine Korsgaard, ‘The Reasons We Can Share: An Attack on the Distinction between Agent-relative and Agent-neutral Values’ (1993). She expresses doubts about the distinction connected principally with the Metaphysical Question and the Normative Question.

10 Value Bearers and Value Pluralism

The preceding chapter argued that we should be cautious about the dichotomy of agent-relative and agent-neutral normative reasons. There are good grounds to believe that reasons are, by their nature, always agent-relative. Accordingly, there are good grounds to be cautious about understanding the distinction between personal and impersonal values in terms of the dichotomy between agent-relative and agent-neutral normative reasons. As we saw, if every reason is a reason for someone, PEFI (i.e. the personalizability feature implication) deﬁes us to say what it is about a certain fact (or set of facts) that gives the agent a reason to act—if, that is, the fact does not involve the agent at all. If there is no such feature, the fact’s being a reason for the agent appears to be inexplicable. Of course, we might not know, or we might be unsure, what the relevant feature is. We might also question whether the fact is in effect a reason in the ﬁrst place. We might be speaking about operative reasons rather than normative reasons, or we may have theoretical rather than practical reasons in mind. Again, we might not have distinguished between reason statements and true reason statements, or between reason statements and reasons, or we might have confused motivating reasons with normative reasons. But if we are clear about these matters and have in mind normative reasons, qua facts, then for someone taking an essentialist approach to the dichotomy the question is the following: Why is an allegedly agent-neutral reason a reason for person a? To maintain that a fact that does not involve a might still be a reason for a is to admit that there is no explanation of why the reason is a reason for a. That is, if the reason-constitutive fact in no way involves (or even relates to) a, it becomes deeply puzzling that the fact is a reason for a. Jonathan Dancy’s distinction between enablers and favourers provides some hope when it comes to contributory reasons, but FA (i.e. ﬁtting-attitude) analyses are not conﬁned to such reasons. In fact, if it makes sense to distinguish between complete and contributory reasons, the FA analysis is naturally understood as being primarily about complete reasons. The notion of a complete reason is not that straightforward, and so Dancy might be right in his criticism. Nor is the distinction between enablers and favourers clear-cut. However, even if we believe there is a metaphysical basis for this distinction, it remains

10.1

E X T R I N S I C F I N A L VA L U E S

153

to be shown just how it provides us with a satisfying account of the relative/neutral dichotomy. Suppose it cannot do this, and that I am right that the notion of a reason is, in the way I have outlined, such as to necessarily introduce a relation to agents. In that case, on the FA pattern of value analysis we obviously need to differentiate between impersonal and personal values in some way other than by employing the distinction in question. One important premise in this book is the idea that a formal analysis of value should be accessible to as many plausible substantive value theories as possible. Two such theories are hedonism and preferentialism. Since these positions do not ascribe, or so I will argue in this chapter, non-derivative value to the same kind of value bearer, we have reason to endorse a formal view of value that is consistent with value-bearer dualism. That is, given certain plausible assumptions about the nature of pleasure and preference, the value accruing to pleasure (from a hedonist point of view) cannot be accounted for by preferentialists. The two positions are most plausibly regarded as ascribing non-derived ﬁnal value to different kinds of value bearer (even if they ascribe derived ﬁnal value to the same sort of bearer). Second, the question of how many kinds of bearers of value there are is a challenging one, so in the remaining sections I discuss in more detail both monist and pluralist arguments, concluding that a pluralist position is preferable.1 Moreover, since my discussion in this chapter will be about ﬁnal values in general, it should be recalled that personal as well as impersonal values may be ﬁnal values.

10.1 Extrinsic ﬁnal values In Chapter 1 I brieﬂy considered some examples that are inconsistent with G. E. Moore’s views on intrinsic value. An object that is valuable because it is rare or unique is a good example. Rarity and uniqueness are externally relational properties and would not be admitted in an Moorean supervenience base which only admits the value bearer’s internal features. Many other objects seem to have a value in virtue of some external relational property. Objects related to persons of historical importance appear to carry impersonal ﬁnal value; the house where I have lived for more than twenty years is an example of a personal extrinsic ﬁnal value. It seems easy to come up with an object, event, or person that is valued for its own sake because it has certain relational features. There is a somewhat different kind of extrinsic value, i.e. the kind of value that is somehow conditional on something else, which I need to set aside here, although it does seem to me to raise important issues. For instance, Dancy (2004, p. 172) has offered the pregnant example of a joke that is funny only when the butt of the joke is present; and Dan Egonsson considers the beauty of a motorcycle he owns that he thinks is beautiful

1 In doing so I will unfortunately have to cover quite a bit of ground rapidly. Since there are a lot of tigers in the area, I realize I cannot reasonably expect not to be accused of excessive speed and inadequate care.

154

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

“on condition of its having an instrumental value” (i.e. that it works).2 Shelly Kagan’s example3 of a fast racing car seems also to belong to this group of examples. One reason for resisting the notion of extrinsic ﬁnal value is that it does not combine well with certain inﬂuential substantive views on what is valuable for its own sake. Consider, for example, a theory like preferentialism. This position relates value to preferences, desires, and wants. Given certain plausible assumptions about these attitudes, preferentialism (at least, in a plausible form) can be depicted as containing the following core claims: (1) Final value accrues only to the objects of (ﬁnal) preferences (desires, wants). (2) A preference has as its object the realization, or obtaining, of a state of affairs. From these two theses there follows a third claim about what the value bearers are according to preferentialism. (3) Final value accrues only to obtainings of states of affairs. If there are non-obtaining states of affairs, such as ‘that everyone is happy’ or ‘that I know all there is to know about values’, preferentialism is most plausibly regarded as a view that ascribes value to the realization of such states—i.e. to what I have referred to as facts, or the obtaining of states of affairs—rather than to the states of affairs themselves. It is evident that there are different ways of understanding what a preference is, and hence that the claim that all preferentialists subscribe to (1)–(3) is debatable. Still, we need a name for the kind of view that embraces (1)–(3), and preferentialism is, I think, quite suitable here. Assuming therefore that (1)–(3) are what preferentialists endorse, they cannot straightforwardly endorse the following claim: H1: Final value accrues to concrete objects (not states of affairs or the obtainings thereof ), e.g. stamps, wild nature, dresses, and experiences. Hedonists actually support the stronger claim that ﬁnal, non-derivative value accrues only to experiences. However, the point I wish to establish requires only the weaker H1. It is not only hedonists, then, that might defend this thesis. An objective list theory, for instance, may well regard different experiences as the only value bearers there are, and another version might share with hedonism the idea that experiences are value bearers, but nevertheless maintain that there are in addition other things (e.g. certain obtaining states of affairs) that have ﬁnal value. If preferentialists are faithful to (3), they will have to say that objects such as a unique stamp or the Brazilian rainforest or (a personal value such as) one’s own pleasure experience have a kind of value, if any, other than ﬁnal value. It is not the stamp that is valuable for its own sake; it is rather some realization of a state of affairs involving the stamp that is the genuine bearer of ﬁnal value. For instance, it might be that value accrues to the fact that See his ‘Can Intrinsic and Final Preferences be Irrational?’ (2003). In ‘Rethinking Intrinsic Value’ (1998, p. 283). For a somewhat different and interesting approach, see Rae Langton, ‘Objective and Unconditioned Value’ (2007). 2 3

10.1

E X T R I N S I C F I N A L VA L U E S

155

a certain stamp is in my possession or that having a certain dress ensures that I am related to someone famous. The stamp and the dress have at most derivative value. The value of these concrete objects is thus taken from some other entity that carries ﬁnal value.4 I presume a reply like this one will work in some cases. However, it will not take care of every case. In fact, a hedonist who regards certain experiences (notably pleasure) as having ﬁnal value has an argument against the sort of preferentialism presented here. What the hedonist submits is that preferentialists, in saying that value accrues only to the obtainings of states of affairs,5 miss what hedonists value for their own sake. So from H1, the hedonist draws the conclusion: C: Preferentialists are debarred from valuing what hedonists value for its own sake. Notice by the way that the preferentialism which I consider here, what Wlodek ¨ sterberg (1996) have called object-preferentialism, should be Rabinowicz and Jan O distinguished from what they refer to as the satisfaction version of preferentialism. The latter makes the following claim (rather than (1)): intrinsic value is assigned to the circumstance that our intrinsic preferences are satisﬁed.6 But this version is even less likely to convince a hedonist about the falsity of H1 (i.e. the idea that ﬁnal value accrues to experiences and not to the obtaining of states of affairs or circumstances). The simplest way of satisfying a preference is to make something the case—to see to it that some states of affairs is realized. I will not defend the hedonist claim that ﬁnal non-derived value accrues only to experiences of a certain sort, namely those with hedonic properties. The hedonist perspective on value is restricted. However, whether or not one is a hedonist, and I am not one, I think hedonism does capture something correct in insisting that value cannot simply be understood as something accruing to the realization of a preferred, or desired, abstract state of affairs. But if this is true, there is reason to resist the attempt to view all ﬁnal, non-derivative value as supervening on states of affairs. This issue is therefore of great importance. A formal analysis of ﬁnal value should, after all, be embraceable by the proponent of any major substantive view on value, and hence by the hedonist and preferentialist. In the next section I will brieﬂy explain why I think the hedonist Preferentialists might therefore accept the amendment H1*: ﬁnal derivative value accrues to concrete objects. This, however, would not settle the issue they have with hedonists who challenge their view on ﬁnal, non-derivative value. 5 Very often hedonism is described as a theory about the value of abstract entities (e.g. states of affairs) rather than concrete objects (e.g. sensations). Consider, for instance, Richard Brandt’s comment “philosophers, from Epicurus to Bentham to Sidgwick to J. J. C. Smart and other contemporaries, have thought that there is one and only one state of affairs that is intrinsically good: pleasant (liked) experiences (and bad states: disliked experiences)” (1996, p. 36, my emphasis). In what follows I will not pay any attention to different versions of hedonism that differ in their views on whether value is related to the duration of the pleasant experience or its intensity or a combination of these. There are all sorts of issues involved here; these have been discussed at length by Michael Zimmerman (2001a). In particular, see his intriguing discussion of evaluative inadequacy (pp. 142–8). 6 With regard to the point I wish to make here, nothing hinges on the fact that the two different versions are formulated in terms of intrinsic, rather than ﬁnal, preferences. 4

156

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

conclusion C ought to force preferentialists to reconsider their position, and in effect, why a pluralist view on value bearers is to be preferred to a monist one.7 If I manage to do this, the idea that personal values accrue to many different sorts of objects, which plays an important role in this book, will have been bolstered.

10.2 Hedonism and preferentialism Preferentialists, I imagine, can pursue various strategies to deal with the hedonist attack. For instance, a preferentialist could agree without difﬁculty that she is debarred from valuing what hedonists regard as pleasure as long as it is established that the hedonist has a mistaken view of pleasure. Once the mistaken view is replaced by a more plausible one, C will lose its appeal; and so the hedonist argument does not in the end pose any real threat to preferentialism. This particular strategy tries to take the sting out of the argument by replacing the hedonist’s view of pleasure with one that would render her substantive claim—that ﬁnal value accrues only to pleasure, i.e. to experiences with hedonic qualities, or tones, being had at a certain time, in a certain place—less convincing. We may question the scope of experiences with hedonic tones, but to completely deny that there are these kinds of experience is phenomenologically counter-intuitive.8 Of course, we need not claim that all pleasant experiences have hedonic tones. A mixed view between the ‘hedonic tone’ version and other versions— say, Richard Brandt’s well-known ‘desire-version’, or Fred Feldman’s recent attitudinal version of hedonism—may well be what, in the end, best captures our usage of ‘pleasure’.9 But that is a merely a verbal matter. Brandt’s position is, perhaps, particularly interesting in this context, since it explicitly leaves room for hedonic tones. However, he thinks that an account of pleasure in terms of hedonic quality is elusive. Actually he makes it clear at an early stage that his way of understanding ‘pleasure’ does not, and is not intended to, correspond to the common usage of this term. His deﬁnition is rather intended to be “suited for a scientiﬁc psychological explanatory conceptual framework” (1979, p. 25), and he seems to think that, given this, a hedonic tone account is disqualiﬁed. I shall not here take a stand for or against such a view. However, it is important to realize that the mere fact that hedonic tones may be difﬁcult to ﬁt into a scientiﬁc conceptual framework does not mean that we can leave them out of an axiological framework. If there are experiences with hedonic tones, they should surely be strong candidates on anyone’s 7 I will not attempt to explain why, in combination with their own favoured view, hedonists should embrace the preferentialist idea that value accrues to the objects of at least some preferences. 8 Gilbert Ryle’s view of pleasure is another in which it is denied that pleasure is an experience with a distinctive positive tone. In his account, an activity that we ﬁnd pleasant becomes an activity that we are prone to engage in (or to continue being engaged in). As T. L. S. Sprigge has put it, “This gives a strikingly joyless picture of pleasure or happiness” (The Rational Foundations of Ethics, 1988, p. 132). 9 One possibility is that one analysis ﬁts sensory pleasure and another ﬁts non-sensory pleasure.

10.2

HEDONISM AND PREFERENTIALISM

157

list of possible valuable objects. Thus, for the hedonist, the important thing to notice is that value accrues to those experiences that have hedonic tones. How they are connected to a person’s desires is another issue. Brandt may have put his ﬁnger on a conceptual link between desires and pleasant experiences. Even so, his view fails to bring out what is valuable about pleasure. Incidentally, this may not apply accurately to David Sobel’s more recent idea that a pleasure is nothing more than an experience “which is intrinsically wanted for its own sake and what is wanted is the way it feels when it is occurring”. Here the idea is that it is our desiring (or, more accurately, our liking) some present conscious state that makes the state pleasant. Of course, a quite natural response to this is that we desire or like something not because doing so is pleasant, but rather because what we like, the object of our liking, is something pleasant. Sobel rejects this phenomenological approach, though. Some of his reasons are familiar objections to such approaches. However, his main idea seems to be that it is only if we understand pleasures as experiences that are desired when they are occurring that we can account for their normativity. It is, in his view, only the desire element in pleasure that makes pleasure into a reason-providing state. This is no place to go into detail. I must simply state that I fail to see that Sobel has really established this conclusion.10 Brandt is right to say that the notion of a hedonic tone is elusive. For instance, it might well be that if we try to direct our consciousness so as to single out this hedonic quality we actually lose sight of it. At least, this has been my own experience; trying to grasp and keep the pleasure, or pain, that I am focusing on is a bit like scooping up a handful of sand and then trying to make a ﬁst—the grains simply sift through the ﬁngers as we tighten them. Some hedonic tones vanish as we focus on them. But to take this as evidence that there are no experiences that have hedonic properties is a mistake too commonly made. Anyone who has tried focusing on a minor pain will probably be able to conﬁrm that they ended up being conscious of certain sensations— e.g. that something is burning or pulsating. The earlier experience—the one with, say, a dishedonic property—is replaced. Sometimes, this act of replacement is taken as evidence that there were no ‘hedonic tones’ involved in the ﬁrst place. But I have not yet come across any argument that secures this point.11 The observation that we cannot necessarily hang on to the hedonic quality of an experience by consciously focusing on it helps throw some light on another issue which has for some time been on the hedonist’s agenda: given the heterogeneous nature of our pleasant experiences, is it not wrong to talk about a hedonic tone? It might be claimed that even if we focused only on sensory pleasures, the pleasure we experience when having, say, an orgasm has nothing (or very little) in common with

See Sobel, ‘Pain for Objectivists: The Case of Matters of Mere Taste’ (2005). That certain phenomenological states are not available to introspection has been conﬁrmed by recent scientiﬁc research. For discussion of the relevant literature see, for instance, Leonard D. Katz, ‘Pleasure’ (2006). 10 11

158

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

the pleasure of feeling a cool hand when you have fever. Pace Epicurus, Bentham, and Sidgwick,12 this Mill-inspired13 observation might seem plausible. Obviously a lot of things happen in the ﬁrst case that do not happen in the latter. They are experiences of different things. But it is not clear what we should conclude from this. If there is something to the observation I made earlier—namely, that we tend to lose sight of the hedonic qualities when we try to focus on them—it is not so obvious that the two experiences need, after all, to be entirely different kinds of pleasures. The often-voiced claim that there is no special quality of pleasure, but only various sensory experiences, needs to be backed up with something more than the observation that all we experience are different bodily sensations. Different experiences need not have the same hedonic property to be cases of pleasure. Let us not press this phenomenological point, though. Even if we grant (as seems not implausible) that the two cases do not differ only in respect of how intense, or durable, the experiences are, the point has no really vital consequences vis-a`-vis the present hedonist argument. That different experiences may have different hedonic tones does not alter the fact that what has value, according to the hedonist, are certain experiences (whether or not there is a multitude of them) rather than different states of affairs. Another issue that it is not crucial to take a stand on is whether or not the pleasant experiences should include moods. I see no reason to leave these out of the hedonist account and good reason to include them, but this need not be argued here.

10.3 Preferentialism The disagreement between preferentialists and hedonists goes deeper than the question whether or not there are experiences with hedonic tones. Recall that what at least one common kind of preferentialist actually has in mind when she speaks of the value of experiences is the value accruing to the obtaining of some abstract proposition-like entity involving, or being about, experiences. Hedonists, on their part (whether or not they maintain that there are hedonic tones), would be ill-advised to give up the idea that pleasure is a concrete entity. The idea that pleasure is something proposition-like squares badly with the way we ordinarily regard pleasant experiences, namely as particulars that exist in space and time. Still the distinction between what is abstract and what is concrete is, as we shall see in the next sections, far from clear, and we should be careful not to put too much weight on it. Fortunately, the issue between preferentialism and hedonism about what are the fundamental bearers of non-derived ﬁnal value can be illuminated further. For 12

Roughly speaking, the three writers mentioned here maintained that we should differentiate between so-called higher and lower pleasures by considering how long or intense the pleasure is. If it is very intense or lasts for a long time, the pleasure belongs to the higher category. 13 Mill would have pointed out that my examples were all examples of pleasures with an inferior quality. In Mill’s eyes pleasures of the intellect are superior quality pleasures. For an interesting discussion which questions Mill’s choice of superior pleasures, see Egonsson, Dimensions of Dignity (1998).

10.4

VA L U E B E A R E R S

159

phenomenological reasons, I committed myself above to the existence of experiences with hedonic tones. This gives me another reason to dispose of the idea that the value of pleasure should be understood as value associated with the obtaining of a particular state of affairs, namely one involving the pleasure. The argument sets out from the plausible assumption that hedonic tones are obviously not tones of proposition-like entities such as states of affairs. A state of affairs may, of course, involve a pleasure (or a hedonic tone), but it would not make sense to say that a phenomenological feature accrued to something like a propositional entity. (Indeed it might be argued, more generally, that states of affairs do not have features in the ﬁrst place, hedonic or otherwise.) But if this is the case, we should ask the person who thinks obtainings might have such hedonic properties the following question: How can the realization of such a proposition-like entity as a state of affairs give rise to an entity—an obtaining— that has a hedonic property if the state of affairs that it is a realization of does not have such a feature in the ﬁrst place? Surely the only realistic difference there could be between an obtaining and the state of affairs it realizes is that the former, but not the latter, is being realized. I see no reason why the obtaining of a state of affairs would somehow acquire hedonic tones. The only entity I can reasonably imagine having these hedonic features is a concrete experience (or, more generally, the brain). As mentioned at the outset, I believe hedonists are wrong to think pleasure is the only thing of value. However, you do not need to be a hedonist to accept that certain experiences are valuable, namely those that have hedonic properties. In my view this is hard to deny, especially if we make it clear, as I have tried to do here, that this is not the same as to say that they are experiences of hedonic tones. Moreover, I would add that experiences with hedonic tones are concrete objects—if, and to the extent that, an object is something that can be described in different ways. Only our own limitations seem to limit how the concrete can be depicted. This is not true of abstracta like the state of affairs that Stockholm is the capital of Sweden in 2006. The same, I suggest, holds for obtainings of states of affairs. These obtainings may be slightly less abstract than states of affairs, but they are nonetheless too abstract to carry hedonic tones. Whether we should refer to the concrete object as a physical thing, or perhaps conﬁne ourselves to saying that it is a condition of the brain (leaving it open how to understand ‘condition’) is a less acute issue for the value theorist—at least, as long as we keep in mind that it is the condition that is valuable, not its obtaining.

10.4 Value bearers We have seen, then, that the claim that bearers of value may be quite different in nature, though it makes it hard to endorse any one particular view of value, remains viable. However, other, more compelling arguments might show right away why we should not recognize stamps, dresses, and other concrete entities as bearers of value. So far the discussion has concerned the obtaining of states of affairs, on the one hand, and concrete objects, on the other. But these are not the only cases we need to examine.

160

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

Recent work in the philosophy of value has brought questions about the bearers of (ﬁnal) value to the fore.14 Among the many accounts, here it is possible to detect two general positions: value-bearer monists claim that (non-derivative) ﬁnal value accrues only to one kind of object; value pluralists deny this, claiming there is more than one kind of bearer of ﬁnal value. Both preferentialism (at least, on the attitudinal reading of this theory) and hedonism are bone ﬁde examples of the monist15 position. The following value bearers have recently been discussed: facts (the obtaining of states of affairs), properties (universalia), tropes and individual physical and mental objects (things, persons, experiences). As I have stated, by ‘fact’ I mean the obtaining of a state of affairs. Regarding facts in this way allows me to distinguish between states of affairs as one kind of entity (such as that mermaids live in the sea) and obtainings of states of affairs, which are a second kind of entity. Since there are no mermaids, there is no obtaining of the state of affairs that they live in the sea.16

10.5 Value-bearer monism Monists have made a number of reductionist objections to pluralism. To make discussion of these arguments more manageable, I shall begin by making some general observations about the issue without considering particular proposals (we shall turn to those proposals in due course). Monists wish to reduce one thing, the reductandum, to another, their preferred reductans. An example is: Reductandum: a rare stamp is valuable for its own sake Reductans: that a rare stamp exists is valuable for its own sake. This monist will argue that the statement ‘a rare stamp is valuable . . . ’ is somehow misleading in that it is reducible to the statement ‘that a rare stamp exists is valuable . . . ’ In Section 10.7 I will consider some proposed reductions in detail. Meanwhile, I need to differentiate between some different kinds of monistic approach, since there is one in particular that I want to set aside. A fully-ﬂedged version of monism—what one might call ‘strong monism’—claims that only its favoured kind of reductans is strictly speaking true, and that any reductandum that does not coincide with the reductans is false. Simplifying matters somewhat, one reason for adopting this sort of monism would be that there is no bearer of value such as that referred to in the reductandum. If there are no such entities (pace the implicit 14

For some of the latest discussions of this topic, see Rønnow-Rasmussen and Zimmerman (eds.), Recent Work on Intrinsic Value (2005). 15 Hedonist monism should be understood as monism regarding ‘positive’ ﬁnal value. An extreme and absurd form of value monism is imaginable in which the view that ﬁnal value accrues to the concrete object ‘experience of pleasure’ is combined with rejection of a corresponding view about what has ﬁnal negative value, i.e. experiences of pain. 16 Zimmerman (2001a, 49 ff.) goes a step further, suggesting that the obtaining of a state of affairs is best understood as consisting in some individual(s) exemplifying a certain property at a certain time. For a discussion of this suggestion see Rabinowicz and Rønnow-Rasmussen (2003a, especially the appendix).

10.6

SIMPLICITY

161

suggestion of pluralists), then the reductandum cannot be taken at face value; it must be understood in some other way. It would seem that if this claim can be paired with a convincing refutation of alternative pluralist suggestions as to the bearers of ﬁnal value, monists will have a strong case for their own view. In its most extreme form, strong monism entails that there is, in fact, only one kind of thing: the things that are the bearers referred to in the reductans. Supposed bearers that do not coincide with what is depicted in the reductans are then considered metaphysically queer. A somewhat less breathtaking variant conﬁnes itself to saying that the valuable objects of the reductandum are not among the things that exist, i.e. do not obtain, or are not instantiated. I am not sure what the metaphysical rationale for the stronger version is, and I shall not discuss it here. Sufﬁce it to say that that version strikes me as much too radical a view. The milder view might, in certain cases, be true, i.e. there might be cases when a certain reductandum is capable of being reduced to an alleged reductans. But as a general rule this does not seem to be the case. But not every form of value-bearer monism is based on metaphysics. Thus in versions of ‘moderate monism’ one ﬁnds the idea that there is something ﬁshy, but not necessarily false, about the reductandum. To say that ﬁnal value accrues to x might, according to the moderate monist, express some truth, but it is not the whole truth; only the reductans conveys the whole picture. If we were to admit that both the reductandum and the reductans expressed the whole truth of the matter, it would be hard to see how this view could be described, strictly speaking, as monist. The point must rather be this: the truth there is to the reductandum (e.g. ‘a rare stamp is valuable for its own sake’) consists in the fact that there is some such thing as x and x is a value bearer. However, the value that accrues to x is a derivative value that stems from the ﬁnal value accruing to the bearer suggested in the reductans, y (e.g. ‘that a rare stamp exists is valuable for its own sake’). Saying that a value is derivative need not be, but often has been, I suspect, a way of somehow downgrading ‘reductandum-value’. Just what this means is another question that will engage us later on.

10.6 Simplicity I suppose that one important driving force behind monism is its simplicity. A parsimonious ontology is preferred to one containing many different kinds of entity of the kind that cannot be reduced to a fundamental single kind of entity. Still, it is difﬁcult to see why simplicity in itself is a mark of truth. Pluralists will eventually counter by saying that the cost to monism is that their ontology is oversimpliﬁed. There is a version of the simplicity argument which, in standard textbooks on the matter, is set aside as being ﬂawed. I am thinking of the idea that there might be various external reasons for wanting a simple theory. For one thing, such a theory might be looked upon as easier to deal with. Thus, giving priority to a theory that is simple may on occasions be the rational thing to do—it will, for instance, be time-saving in that,

162

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

most likely, it can be tested and eventually dismissed more rapidly than more complex theories. However, although simplicity may be an incentive to test one theory rather than another, it does seem like a wrong kind of reason for settling the truth-value of the theory. That the life of the investigator, or the evaluator, will go more easily is hardly a compelling reason to become a monist. David Alm has suggested (personal communication) that it might only be a bad reason when it comes to empirical science, and that similar reasoning does not obviously apply to theories of value. Perhaps he is right when it comes to substantive theories of value. However, I am quite sure that the fact that things are made easier for the formal axiologist should not have an impact on his or her views on the monism/ pluralism issue. A more common argument nowadays has it that pluralism should be rejected because it has a number of non-desirable consequences. Monists with regard to value bearers have, for instance, claimed that pluralism makes value comparisons impossible. How can we compare, much less weigh one value against another, if these values are not of the same kind? That would be just like weighing one metre against one litre. A version of this external (to truth) argument has also been voiced among value-bearer monists. The idea is that if there were only one kind of value bearer, we might be able to provide an informative account of the computation of value. If pluralism is correct, and there are different kinds of value bearer, then such computation seems impossible. This objection is perhaps not as easily brushed away. Nonetheless, the pluralist’s basic comment, I take it, will be that there is not much to computation if values are not computable. Here is another explanation of why monism has been preferred. During the last ﬁve decades or so the development of a formal logic for operators such as “It is good that ___” or “that___is better than that___” has continued steadily. The opportunity to lean on a well-established framework is probably one reason why some value theorists have been inclined to reduce all value to proposition-like objects. (Cf. Rabinowicz and Rønnow-Rasmussen, 2000, p. 44). Again, this is a quite understandable motive that has nothing to do with the monism/pluralism issue. There is also an evaluative side to this issue. It might be argued that the pluralist position rests somehow on the idea that there is a subject matter here to be discovered. But if what has value is in some sense up to us, it is surely better to have a value theory that allows values to be compared and computed. This is an intriguing argument. However, as I argued in Chapter 1, it is a mistake to think that pluralists are somehow committed to objectivism, or realism, about values. Still, the gist of the argument is, I suppose, that a world that contains no moral conﬂicts is a better world than one with such conﬂicts, so if we want to systematize our thoughts on value, we might just as well opt for one that allows us to solve conﬂicts. But is that really that obvious? I, at least, am not sure what the substantive argument is for the notion that a world in which values are computable is better, in some sense, than a

10.7

SOME REDUCTIONS

163

world in which value is not computable and where value conﬂicts may emerge. I might be wrong, but this is clearly a substantive issue.17

10.7 Some reductions Suppose x refers to a pleasant experience. It may now be argued that the claim ‘x has ﬁnal value in virtue of its property F ’ (F referring to a hedonic quality) should actually be reduced to a claim about some existential fact like ‘what has ﬁnal value is that x, which has F, exists’. An alternative to this might be ‘what has ﬁnal value is the fact that there is a state of affairs of the form there exists some object that has F ’. Given an eliminativist reduction, the alleged value of the experience is now localized in the existential fact. This kind of reductionist manoeuvre is open to a very natural objection, namely that it starts from the wrong end of the story—or, as Rabinowicz and I (2000) have remarked: it puts the cart before the horse. The reason we think that the fact ‘the experience x which has property F exists’ is valuable is surely that the experience is valuable. The value of the experience grounds the value of the existential fact. To suggest otherwise by contending that it is the fact that grounds the experience is to misplace what is of ﬁnal value here. If anything, the fact derives its value from the value of the experience. The reverse of this relation is unconvincing. The advocate of this particular reductandum—say, a hedonist—could, in other words, accept that value might also accrue to facts. In fact it would be strange if she denied them value altogether. However, these facts would have merely derivative value. In the case under consideration it seems natural to say that the existence of a pleasant experience is valuable not because some existential state is non-derivatively valuable, but in virtue of the value that accrues to the object itself (i.e. the experience). There may be other examples in which it is conceivable that the value of the state need not be drawn from the value of the concrete object. The latter might be of no value at all. A further possibility should be mentioned. Even if what we value does not accrue to the fact that the experience that has property F exists, this alternative may still be regarded as being on the right track. Perhaps what is valuable for its own sake has something to do with the experience having the property P.18 On the other hand, it does not seem very plausible to say that it is properties, in the sense of universalia, that are value bearers.19 A different, but related, proposal does come with intuitive appeal, however. This is the proposal that it is rather the instantiation of properties that is valuable. So when person a experiences pleasure, what is valuable is the instantiation of pleasure in a, and not some

17

Value-bearer monism is consistent with value pluralism. For instance, one might argue that happiness is one sort of value and freedom another, but that these values accrue to only one sort of bearer—say, to facts or to states of affairs. Amartya Sen, for instance, is often interpreted as being a value pluralist but a value-bearer monist (Development as Freedom, 1999). 18 Parayot Butchvarov (1989), for instance, argues that the bearers of ﬁnal value are properties. 19 Noah Lemos (1994) and Zimmerman (2001a) criticize this view.

164

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

universal which, perhaps, nobody exempliﬁes.20 Of course, that the instantiation of F occurs in a rather than in another object (person) b is of no evaluative importance (at least, as long as we have moral values in mind). It is not the fact that this object a has F that is valuable. What is valued is rather the instantiation of F in a. Value accrues in the same way to each instantiation of F, in whatever object it occurs. On this proposal, then, we value various instantiations of P rather than the states that a has F, that b has F, and so on. Some metaphysicians have suggested that the instantiation should be regarded as an irreducible ontological entity sui generis—what, nowadays, is referred to as a ‘trope’ (following Donald Williams, 1953). In contrast with properties (i.e. universalia) tropes are particulars. Whereas a single universale can be instantiated in many different places, this is not the case with a trope. Two or more instantiations of a universale are identical with each other. Tropes, on the other hand, are never identical with each other. Thus a’s experiencing pleasure at t1 and b’s experiencing pleasure at t1 are two tropes (not the exempliﬁcation of one and the same pleasure property). There are two particular entities involved in the two cases. Moreover, unlike concrete individual things such as people, these particular entities are regarded as abstract particulars.21 This argument has a certain ring of plausibility. For one thing, if it is the instantiation of a property that is valuable, the cart-before-the-horse argument is inapplicable. It would be counter-intuitive to suggest that the value that this instantiation of property F at t1 in a has derives from the value of a. On balance, however, I think a hedonist should eschew this manoeuvre. The precise metaphysical nature of these ‘instantiation-entities’ is still somewhat opaque. Are they abstract or concrete entities? I am inclined to see them as abstract. The notorious distinction between abstract and concrete entities is a vexed issue. However, recently Zimmerman has lent us a hand with it: It’s been suggested, for instance, that abstract entities exist necessarily, whereas concrete entities exist only contingently. This is problematic, since God is often taken to be a necessarily existing concrete entity, while sets of contingently existing objects are often taken to be contingently existing abstract entities. It’s also been suggested that concrete entities have spatiotemporal location, whereas abstract entities do not. But this is again problematic, since individual souls are often taken to be concrete entities that do not occupy space, while times and places are often taken to be concrete and to have no spatiotemporal location themselves; and again, God is often taken to be a concrete entity that lacks spatiotemporal location. Another suggestion is that abstract entities can have instances, whereas concrete entities cannot. But this too is problematic, since sets are often taken to be abstract and yet

20 Another possibility, suggested by Ingmar Persson (private communication) and Zimmerman (2001a; 2001b), would be that it is rather that x has P, which is valuable. For criticism of this idea, see Rabinowicz and Rønnow-Rasmussen (2000). 21 There is no consensus among trope theorists over what else there is besides tropes. On what we might refer to as a nominalist position, no other entities exist apart from tropes. This is not classical nominalism; but in virtue of its denial of universalia and its insistence that there is only one fundamental ontological category (namely, of particular tropes) it is a kind of nominalism (contrast Keith Campbell, Abstract Particulars (1990, p. 27)).

10.7

SOME REDUCTIONS

165

are incapable of being instantiated, something which is also plausibly thought to be true of ‘impossible’ properties (such as the property of being a square circle).22

Zimmerman next suggests that: . . . [a]ll abstract entities (except sets, if they exist) exist necessarily, whereas no concrete entity (except God, if he exists, and states of God) does, that only abstract entities are instantiable, and that only concrete entities can have spatiotemporal location.23

So where do tropes ﬁt in? Take the claim that abstract entities exist necessarily. To understand this claim we have to be aware of an idea that I think underlies it, namely the idea that there is a conceptual distinction between the existence of an (abstract) entity and its instantiation or obtaining. Consider, for instance, properties—say, the property of being a mermaid. On this (platonic) account, this property qua universale exists necessarily, but in this world it happens to be the case that no such property is instantiated. Should we consider tropes as entities that exist necessarily? Modern trope theory developed as an attempt to throw light on our notion of a property. It might therefore be thought that if the latter were assumed to have necessary existence, so too should tropes. But it is in fact hard to detect anything that speaks in favour of this suggestion. The distinction between something’s existing and something’s being instantiated is, I believe, inapplicable to tropes. While properties qua universalia are capable of instantiation, properties qua tropes (if there be such) are not. Given this, we cannot make use of this distinction in order to explain why tropes would exist necessarily. That tropes are not necessary is pretty obvious: my beard is greyish, but who would deny that I (in at least some sense) could have had a red beard. However, tropes can be abstract in some other sense than that given by the phrase ‘entity existing but not obtaining’. I will return to such a possibility in a moment. Meanwhile, a word must be said about the possibility that tropes are concrete entities. Recall that, in Zimmerman’s sense of the term ‘concrete’, it was (with a few possible exceptions) only concrete particulars that have spatio-temporal location.24 Are, then, all tropes spatio-localizable? This is of course speculation, but it seems to me that at least some tropes, if they exist, are more easily located in space and time than others. My hair colour would be an example, whereas my pain would not be just as easily positioned in space. (Is it in my thumb? My brain?) And what should we say about dyadic tropes such as my relation to the truth that ‘2 + 2 = 4’?25

22

Zimmerman (2001a, p. 34). Ibid. 24 Zimmerman’s own view must be mentioned here. In his (2001a) he suggests that it is neither properties nor abstract facts, but rather concrete states of individuals that are value bearers; see also Zimmerman (2001b). For a detailed examination of the idea that value can be reduced to such states of individuals, see Rabinowicz and Rønnow-Rasmussen (2003a). 25 In Rønnow-Rasmussen (1998, pp. 293, 294) I expressed doubts about using indexical terms such as ‘here’ and ‘there’ about phenomenal qualities. Today my doubts are less insistent, at least when it comes to some kinds of quality. 23

166

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

10.8 Separating the concrete from the abstract Actually there is an alternative approach to the abstract/concrete division, and on this approach—which, I think, comes closer to the meaning of ‘abstract/concrete’—there is no doubt that tropes are abstract particulars. Comparing different lollipops with each other, Williams says: To borrow now an old but pretty appropriate term, a gross part, like the stick, is ‘concrete’, as the whole lollipop is, while a ﬁne or diffuse part, like the color component or shape component, is ‘abstract’. The color-cum-shape is less abstract or more concrete or more nearly concrete than the color alone but it is more abstract or less concrete than color-plus-shape-plus-ﬂavor, and so on till we get to the total complex which is wholly concrete. (1953)26

In my view the main idea expressed here catches what we have in mind when we perform an act of abstraction: the colour of this lollipop, for instance, occurs together with other qualities of the lollipop, and what I do, when I bring this quality before my mind, is set aside other qualities of the lollipop. An act of abstraction is, as Keith Campbell puts it, “an act of selective ignoring” (1990, p. 3). What trope theorists then do is to claim that the object of such an act of abstraction has existence. Williams and Campbell also want to take a further step and say that tropes are located in the same place; for instance, the concrete lollipop is the totality of being where the colour, shape, (etc.) of the lollipop are. In other words, the doctrine that two different things cannot be at the same place at the same time is either false or inapplicable to abstracta. But as far as I can see there is no need to accept this latter claim, even if we do regard abstract entities in the above way (i.e. as thin entities that constitute the building material of thick entities). Nor, as far as I can see, need we endorse another claim that is implicit in the above passage, namely that things can be more or less concrete/abstract. An object might be thicker or thinner, more or less diffuse, and in that sense we may speak of an abstract entity getting closer to, or further away from, being concrete. But, intuitively, to claim this is in no way to remove the gap between the concrete and the abstract. What should we say, then, about the obtaining of a state of affairs? Are these abstract or concrete entities? I am not sure. Following Williams, we might want to say that they are more concrete than the features Williams mentions as abstract; but being more concrete is not necessarily the same as being a concrete object. I am inclined to deny that obtainings are concrete entities, but I cannot presently offer any good argument for this.

26 The passage comes from D. H. Mellor, and A. Oliver (eds.), Properties (1997, p. 113). Cf. Campbell (1990, pp. 2, 3). See also D. M. Armstrong, Nominalism & Realism (1978). Armstrong discusses the case of a coloured cube (pp. 120–1). This, as he calls it, ‘concrete’ cube he then compares with the tactual and visual cubes, which he refers to as abstract particulars. Notice that Armstrong does not endorse trope theory. See also Zimmerman (2001a, p. 66, n. 6), where Campbell’s use of ‘abstract’ and ‘concrete’ is criticized.

10.8

S E PA R AT I N G T H E C O N C R E T E F RO M T H E A B S T R A C T

167

Let us return to the question of why we should accept that it is the instantiation of pleasure that is valuable. Surely, in the end, the hedonist will say that what we value is something concrete and not some thin, abstract entity such as a property-instantiation. Obviously it cannot merely be an instantiation that prompts the hedonist to ascribe value to a sensation; after all, most property instantiations do not carry any value according to hedonists. It is true, of course, that the sensation is valuable because it is a certain kind of instantiation, namely one involving a certain kind of experience. But again, the very fact that hedonists say that there is something about this kind of sensation that makes it valuable explains why it still makes sense to ask “But what is so valuable about these experiences?”, and this suggests that “instantiating . . . ” only gives us what has derivative value. The natural thing to say here is that the hedonic property P is a good-making property that renders the complex experience valuable. To this it might be objected that if an object is valuable because it has certain properties, these properties (or, at least, the instantiations of them) must themselves be valuable. The idea underlying this objection would be that whatever is a condition of something of value must also be valuable. However, as has been said before, this is a non sequitur: it is not logically necessary for the condition of the valuable to be valuable itself, either intrinsically or extrinsically. The good-making property need not itself be valuable to qualify as a good-making property. We now need to consider a further aspect of the hedonist position. So far I have been speaking about its being a certain kind of experience that has value, according to (at least one kind of) hedonist. However, it might be pointed out, quite rightly, that I have not said very much about what this experience is. Recently Ben Bradley has suggested the following: If there is a feeling of pleasure itself, akin to the feeling of warmth or sweetness, and if the possession of such a feeling is intrinsically good, the bearers of intrinsic value need not be ﬁnegrained. They might be coarse-grained events consisting of a person having that feeling of pleasure at some time. Nevertheless, the bearers of value will be things that happen or obtain, rather than, say physical objects. And that is really the crucial distinction here: between, on the one hand, those who attribute intrinsic value to things and people, and on the other hand, those who attribute intrinsic value to events, tropes, states of affairs, property-instantiations, facts, or propositions. (2006, pp. 115–16)

Bradley is right, I think, in suggesting that the value bearer in this sort of hedonism is coarse-grained. And he is, of course, also right in thinking that such bearers are not physical objects. It is in his further characterization of the bearer that I think he goes wrong. That is, what makes one experience valuable, rather than some other one, according to the hedonist, is that it is an experience with a hedonic property. That is what makes the experience into a feeling. An event, or happening, or obtaining, contains the hedonic property only indirectly, and these kinds of object will therefore only carry indirect or derivative value, according to the hedonist.

168

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

Bradley gives us no reason to give up the idea that what has value, in classical hedonism, are experiences of a certain kind, namely those with hedonic properties (which, incidentally, does not necessarily mean that they are experiences of hedonic properties). But he is on the right track when he maintains that there is a crucial distinction between two groups of value theorists, and that members of these groups regard the nature of value bearers quite differently. Nevertheless, it seems to me that the real difference is not so much between those who think that physical things may carry value and those who disagree (and who think that it is rather proposition-like entities that are the value bearers). This simpliﬁes the issue too much. The demarcation line runs rather between those who think that value bearers must be possible objects for attitudes that take propositional entities as their object, and those who think that other kinds of attitude may well ﬁgure in the analysans of value. So I would not draw the line where Bradley does. In particular, I would reject the idea that pleasures are the direct object of attitudes such as desires and preferences. Pleasures, in my view, are concrete entities; they are concrete, not in the sense of being material objects, but rather in the sense that we can go on describing them in ways that seem, at least, inexhaustible. This characteristic is not shared by abstract entities. It is shared by obtaining states of affairs, though. But they are not what a hedonist non-derivatively values. Although some of them certainly involve hedonic properties we would not say that a hedonic property accrues to these entities in the way a hedonic property accrues to pleasant experiences. Pleasures are nonetheless too concrete to be objects of attitudes that have abstract entities as their intentional objects. A hedonist may prefer that a certain person, or a certain consciousness, or brain, is in this sort of pleasure-state. But any value that accrues to the fact that this consciousness is in this particular hedonic state, derives, on the standard hedonist position, from the value of this sort of state.27

10.9 Recapitulating This chapter has addressed some quite complicated issues, and it might therefore be a good idea to go over the main points that have been argued for. It might even be worth recapitulating the key ideas expounded in this book. This will help situate Chapter 10 in its context. In this work I have defended the idea that we have reason to expand our value taxonomy with a set of values that it is appropriate to refer to as personal values (rather than impersonal values like justice or equality). We refer to personal values with relative terms such as ‘good-for’ and ‘value-for’. The notion of ‘good-for’ plays an

27 David Brax has recently developed an account of pleasure in Hedonism as the Explanation of Value (2009). His central idea is that pleasures are experiences liked by the person having the experience. However, in contrast to, say, Brandt, he suggests that the liking is actually part of the experience liked. This, of course, makes pleasure into a self-reﬂecting attitude. A somewhat similar approach can be found in Murat Aydede, ‘An Analysis of Pleasure vis-a`-vis Pain’ (2000). See also Katz (2006) and Sobel (2005).

10.9

R E C A P I T U L AT I N G

169

important role in many ethical discussions. However, there is little agreement among philosophers over its correct analysis, and so there is an obvious need for a new analysis. This work purports to meet this need by offering a novel way of understanding the notion of personal value. For various reasons this kind of value has been assumed, in one very inﬂuential tradition in value theory, to be either incoherent or explicable in terms of (and therefore less important than) what is good simpliciter. But the analysis defended in this work shows that you can analyse personal value without appealing to such non-relative goodness. Moreover, the analysis supplies us with a notion of goodfor which steers clear of standard objections. It does so by ﬁne-tuning a certain pattern of value analysis—one that has recently attracted much attention and which has roots in the writings of the Austrian philosopher Franz Brentano. Personal ﬁnal values come in two versions, depending on the nature of their valuemaking properties. If all of these are internal to the object of value, we are dealing with ﬁnal intrinsic value. If some of the value-making properties are external, i.e. externally relational, we are dealing with an extrinsic value for its own sake. As to the kind of analysis I have employed, one attraction is that it remains neutral when it comes to the issue of subjectivism versus objectivism. It is sometimes said that subjectivists cannot conceive of objects having intrinsic or ﬁnal value. If this is true, there is not, it seems, much point in an analysis that remains neutral on the matter. However, since I believe subjectivists need not give up on ﬁnal value, I have described in some detail (in Chapter 1) just how I conceive of this distinction. It will be recalled, I hope, that what the subjectivist might ﬁnd hard to accept is the following claim: The Invariance thesis: The ﬁnal value of an object is invariant over possible worlds. But this thesis is not sacred, and indeed if it is the non-essential properties of an object that make it valuable, we should reject it. Not all value will then be invariant over possible worlds. For this reason we cannot employ the thesis as a demarcation line between subjectivists and objectivists. Subjectivists can in fact choose between several more or less interesting ways of understanding value. To bring this out I distinguished between value supervenience and value constitution. Understanding values in terms of reasons has a number of advantages. However, it also faces some weighty objections. Chapter 3 was devoted to the most serious of these—the so-called ‘wrong kind of reason’ problem. Although, for obvious reasons, I would welcome a solution to this problem, I am not convinced by the solutions I have seen proposed. In Chapter 4 I gave my reasons for not accepting the (naturalist) attempt to understand value-for in terms of the de facto attitudes of its beneﬁciary (i.e. the person whose value it is). My suggestion is instead the following: an object x’s value for a person a (i.e. x’s personal value) consists in the existence of normative reasons for favouring/disfavouring x for a’s sake. The idea that something could be good for a person has often been criticized. I considered Moore’s famous attack on the very notion of a personal good, arguing that it is not very convincing.

170

VA L U E B E A R E R S A N D VA L U E P L U R A L I S M

The for a person’s sake-attitude, which plays a leading role in my analysis, is difﬁcult to characterize, certainly. In Chapter 5, in order to describe one of the more intriguing features of this attitude, I introduced a distinction between two kinds of property. These differ from each other in the role they play in the intentional content of the attitude. I argued that this distinction between what I called justiﬁers and identiﬁers helps us to separate two kinds of for someone’s sake-attitude, both of which are pertinent to the analysans of personal value. It is important to acknowledge that good-for has more than one sense. In particular, I admit that the idea that whatever is welfare-constitutive is good for the agent might express a notion of good-for that differs from the one I try to illuminate. Another complex question, which I addressed in Chapter 6, is whether universalizability is a problem for an analysis like mine. It would seem to be so. For someone’s sake-attitudes that take non-replaceable persons as their intentional objects are not obviously universalizable. This may not be an insurmountable difﬁculty, though. It is important to keep in mind that universalizability is a very strong requirement, and that value analysis can manage with something less demanding: supervenience. I have also distanced my views on good-for from some alternative suggestions, made by other philosophers, and I considered the possibility that values in general, and personal values in particular, are so-called Janus values. If there are any such values I suspect they are rare. The idea that personal values should be understood in terms of agent-relative reasons, and that impersonal ones should be interpreted as agent-neutral reasons, seem plausible. However, in Chapter 9 I argued that we should be cautious about this blend of ideas. Finally, in Chapter 10 I present an argument whose aim was to pave the way for value-bearer pluralism, i.e. the idea that entities belonging to quite different general metaphysical categories can be bearers of ﬁnal value. The structure of the argument was fairly simple. It set out from the idea that a common form of preferentialism and the standard version of hedonism both make plausible claims about what is valuable. Brieﬂy, preferentialists say that it is the satisfaction of our preferences and the obtainings of certain states of affairs that have value. Hedonists, by contrast, claim that pleasure is valuable. Of course, the argument will only gain traction with those who take these two views to be credible, but since I feel conﬁdent that most people would be prepared to agree that pleasure and preference satisfaction (in the above sense) are valuable, I take the argument to be of general interest. The core idea of the argument is as follows. It is likely that pleasure will be among the many kinds of thing that preferentialists take to have value. However, the value preferentialists ascribe to pleasure will, the argument goes, always be regarded by hedonists as derivative in nature; and since hedonists think that pleasure has nonderivative value, they will conclude that preferentialists are debarred from valuing precisely what they, as hedonists, value—at least, in the way hedonists value pleasure. As we saw, the reason for this is that although the obtaining of a state of affairs may be about hedonic properties, its obtaining will not itself be an experiencing entity, and

10.9

R E C A P I T U L AT I N G

171

hence it cannot have hedonic properties in the way that a pleasurable experience has hedonic properties. The value accruing to pleasure, from a hedonist point of view, cannot, therefore, be accounted for by the preferentialist. To think it can is—again from the perspective of the hedonist—to put the cart before the horse.28 So if we want to retain substantial value dualism we have to give up the idea of value-bearer monism. The effect of this reasoning is to encourage abandonment of value-bearer monism. Moreover, if we recognize that hedonists and preferentialists regard different kinds of metaphysical entity as the fundamental bearers of non-derivative value, and if we agree that a formal analysis of value should be accessible to as many plausible substantive theories of value as possible, the case for preferring pluralism to monism becomes even stronger.

28 Notice the qualiﬁcation here. If you are not a hedonist, the cart-before-the-horse argument might not apply to obtainings of states of affairs.

Bibliography Aiken, Henry D. (1950). ‘Evaluation and Obligation: Two Functions of Judgement in the Language of Conduct’. Journal of Philosophy, 47 (1): 5–22. Alexander, Samuel (1920). Space, Time and Deity. London: Macmillan. Alfano, Mark (2009). ‘A Danger of Deﬁnition: Polar Predicates in Moral Theory’. Journal of Ethics and Social Practice, 3 (3): 1–13. Allen, R. T. (1993). The Structure of Value. Aldershot: Avebury. Anderson, Elizabeth (1993). Value in Ethics and Economics. Cambridge, MA: Harvard University Press. Armstrong, D. M. (1978). Nominalism & Realism. Cambridge: Cambridge University Press. Aydede, Murat (2000). ‘An Analysis of Pleasure vis-a`-vis Pain’. Philosophy and Phenomenological Research, 61 (3): 537–70. Baron, Marcia W. (1997). ‘Kantian Ethics’, in Marcia W. Baron, Philip Pettit, and Michael Slote (eds.), pp. 3–91. Baron, Marcia W., Pettit, Philip, and Slote, Michael (eds.) (1997). Three Methods of Ethics. Oxford: Blackwell. Bradley, Ben (2006). ‘Two Concepts of Intrinsic Value’. Ethical Theory and Moral Practice, 9: 111–30. Brady, Michael (circulating ms.). ‘The Value of Valuing’. Brandt, Richard B. (1946). ‘Moral Valuation’. Ethics, 56: 106–21. ——(1979). A Theory of the Good and the Right. Oxford: Clarendon Press. ——(1996). Fact, Values, and Morality. Cambridge: Cambridge University Press. Brännmark, Johan (2001) ‘Good Lives: Parts and Wholes’. American Philosophical Quarterly, 38: 221–31. ——(2002). Morality and the Pursuit of Happiness: A Study in Kantian Ethics. Lund: Media-tryck, Lund University. ——(2008). ‘Excellence and Means: On the Limits of Buck-Passing’. The Journal of Value Inquiry, 42: 310–15. ——(2009). ‘Goodness, Values, and Reasons’. Ethical Theory and Moral Practice, 12: 329–43. Brax, David (2009). Hedonism as the Explanation of Value. Lund: Media-tryck Sociologen, Lund University. Brentano, Franz (1969 [1889]). The Origin of Our Knowledge of Right and Wrong. London and Henley: Routledge & Kegan Paul. Broad, C. D. (1925). The Mind and Its Place in Nature. London: Routledge & Kegan Paul. ——(1930). Five Types of Ethical Theory. London: Routledge & Kegan Paul. ——(1942). ‘Certain Features in Moore’s Ethical Doctrines’, in P. A. Schilpp (ed.), The Philosophy of G. E. Moore (The Library of Living Philosophers). New York: Tudor Publishing Company, pp. 43–67. Broome, John (1995). ‘Skorupski on Agent-Neutrality’. Utilitas, 7: 315–17. ——(2004). ‘Reasons’, in R. Jay Wallace, Philip Pettit, Samuel Schefﬂer, and Michael Smith, (eds), pp. 28–55.

BIBLIOGRAPHY

173

Brülde, Bengt (1998). The Human Good. Go¨teborg: Acta Universitatis Gothoburgensis (PhD dissertation). Burgess, J. A. (2009). ‘When is Circularity in Deﬁnitions Benign?’ The Philosophical Quarterly, 58 (231): 214–33. Butchvarov, Panayot (1989). Skepticism in Ethics. Indianapolis: Indiana University Press. Butler, Joseph (1900 [1729]). ‘Fifteen Sermons Preached at the Rolls Chapel’, in J. H. Bernard (ed.), The Works of Bishop Butler, vol. 1 (London: Macmillan). Bykvist, Krister (2009). ‘No Good Fit: Why the Fitting Attitude Analysis of Value Fails’. Mind, 118 (469): 1–30. Campbell, Keith (1990). Abstract Particulars. Oxford: Basil Blackwell. Carlson, Erik and Sliwinski, Rysiek (eds.) (2001). Omnium-gatherum. Philosophical Essays Dedicated ¨ sterberg on the Occasion of his Sixtieth Birthday (Uppsala Philosophical Studies 50). to Jan O Uppsala: Department of Philosophy, Uppsala University. Chisholm, Roderick M. (1981). ‘Deﬁning Intrinsic Value’. Analysis, 41: 99–100. Crisp, Roger (2000a). ‘Review of “Value . . . and What Follows”, by Joel Kupperman’. Philosophy, 75: 458–62. ——(2000b). ‘Particularising Particularism’, in Brad Hooker and M. Little (eds.), Moral Particularism. Oxford: Oxford University Press, pp. 23–47. ——(2005). ‘Value, Reasons and the Structure of Justiﬁcation: How to Avoid Passing the Buck’. Analysis, 65 (285): 80–85. ——(2008). ‘Goodness and Reasons: Accentuating the Negative’. Mind, 117: 257–65. Dancy, Jonathan (1993). Moral Reasons. Oxford: Blackwell. ——(2000a). ‘Should We Pass the Buck?” in A. O’Hare (ed.), Philosophy, the Good, the True, and the Beautiful. Cambridge: Cambridge University Press, pp. 159–75. Also in Rønnow-Rasmussen and Zimmerman (2005), pp. 33–45. ——(2000b). Practical Reality. Oxford: Clarendon Press. ——(2004). Ethics without Principles. Oxford: Oxford University Press. Danielsson, Sven (2001). ‘The Supervenience of Intrinsic Value’, in Erik Carlson and Rysiek Sliwinski (eds.), pp. 93–103. ——(2007). ‘On Geach on Good’, in T. Rønnow-Rasmussen, B. Petersson, J. Josefsson, and D. Egonsson (eds.), pp. 1–11. Danielsson, Sven and Olson, Jonas (2007). ‘Brentano and the Buck-Passers’. Mind, 116: 511–22. D’Arms, Justin and Jacobson, Daniel (2000a). ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions’. Philosophy and Phenomenological Research, 61: 65–90. ——(2000b). ‘Sentiment and Value’. Ethics, 110: 722–48. ——(2003). ‘The Signiﬁcance of Recalcitrant Emotion (or, Anti-Quasi Judgmentalism)’, in A. Hatzimoysis (ed.), Philosophy and the Emotions. Cambridge: Cambridge University Press, pp. 127–47. Darwall, Stephen (1983). Impartial Reason. Ithaca and London: Cornell University Press. ——(2002). Welfare and Rational Care. Princeton: Princeton University Press. ——(2003). ‘Moore, Normativity, and Intrinsic Value’. Ethics, 113: 468–89. Dickie, George (1979). Aesthetics: An Introduction. Indianapolis: Pegasus. Egonsson, Dan (1998). Dimensions of Dignity. Dordrecht: Kluwer. ——(2003). ‘Can Intrinsic and Final Preferences be Irrational?’ in W. Rabinowicz and Toni Rønnow-Rasmussen (eds.) (2003b), pp. 114–30.

174

BIBLIOGRAPHY

Egonsson, Dan, Josefsson, Jonas, Petterson, Bjo¨rn, and Rønnow-Rasmussen, Toni (eds.) (2001). Exploring Practical Philosophy: From Action to Values. Aldershot: Ashgate Publishing. Ewing, A. C. (1947). The Deﬁnition of Good. London: Macmillan. ——(1959). Second Thoughts in Moral Philosophy. London: Routledge & Kegan Paul. ——(1973). Value and Reality. London: George Allen and Unwin. Feinberg, Joel (ed.) (1969). Moral Concepts. Oxford: Oxford University Press. Feldman, Fred (2002). ‘The Good Life: A Defense of Attitudinal Hedonism’. Philosophy and Phenomenological Research, 65: 604–28. ——(2004). Pleasure and the Good Life: Concerning the Nature, Varieties, and Plausibility of Hedonism. Oxford: Oxford University Press. Findlay, J. N. (1970). Axiological Ethics. London: Macmillan and St Martin’s Press. Frankfurt, Harry G. (1999a). ‘Autonomy, Necessity, and Love’, in Harry G. Frankfurt, Necessity, Volition, and Love. Cambridge: Cambridge University Press, pp. 129–42. ——(1999b). ‘On Caring’, in Harry G. Frankfurt, Necessity, Volition, and Love. Cambridge: Cambridge University Press, pp. 155–66. Gauthier, David (1986). Morals by Agreement. Oxford: Oxford University Press. Geach, Peter (1956). ‘Good and Evil’. Analysis, 17: 33–42. Gert, Joshua (2004). ‘Value and Parity’. Ethics, 114: 492–520. Gibbard, Alan (1990). Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press. Grau, Christopher (2002). ‘The Irreplaceability of Persons’. Unpublished PhD dissertation, Johns Hopkins University. Hare, R. M. (1952). The Language of Morals. Oxford: Oxford University Press. Hatzimoysis, Anthony (ed.) (2003). Philosophy and the Emotions. Cambridge: Cambridge University Press. Hartmann, Eduard von (1908). Grundriss der Axiologie oder Wertwägungslehre. Bad Sachsa: Haacke. Heyde, Erich (1926). Wert: Eine philosophische Grundlegung. Erfurt: Kurt Stegner. Heathwood, Chris (2008). ‘Fitting Attitudes and Welfare’. Oxford Studies in Metaethics, 3: 47–73. Hieronymi, Pamela (2005). ‘The Wrong Kind of Reason’. The Journal of Philosophy, 102 (9): 437–57. Humberstone, I. L. (1997). ‘Two Types of Circles’. Philosophy and Phenomenological Research, LVIII: 249–80. Hurka, Thomas (1987). ‘“Good” and “Good for”’. Mind, 96: 71–73. ——(2003). ‘Moore in the Middle’. Ethics, 113 (3): 599–628. Kagan, Shelly (1992). ‘The Limits of Well-Being”, in Ellen Frankel Paul, Fred D. Miller, Jr., and Jeffrey Paul (eds.), The Good Life and the Human Good. Cambridge: Cambridge University Press, pp. 169–89. ——(1998). ‘Rethinking Intrinsic Value’. Journal of Ethics, 2: 277–97. Katz, Leonard D. (2006). ‘Pleasure’, in Edward N. Zalta (ed.), Stanford Encyclopedia of Philosophy (Summer 2006 edition) . Kim, Jaegwon (1993). Supervenience and Mind. Cambridge: Cambridge University Press. Korsgaard, Christine M. (1983). ‘Two Distinctions in Goodness’. The Philosophical Review, 92: 169–95. Also published in Korsgaard (1996), pp. 249–74. ——(1993). ‘The Reasons We Can Share: An Attack on the Distinction between AgentRelative and Agent-Neutral Values”. Social Philosophy & Policy, 10: 24–51. ——(1996). Creating the Kingdom of Ends. Cambridge: Cambridge University Press.

BIBLIOGRAPHY

175

——(1998). ‘Motivation, Metaphysics, and the Value of the Self: A Reply to Ginsborg, Guyer and Schneewind’. Ethics, 109: 49–66. Lang, Gerald (2008). ‘The Right Kind of Solution to the Wrong Kind of Reason Problem’. Utilitas, 20 (4): 472–89. Langton, Rae (2007). ‘Objective and Unconditioned Value’. Philosophical Review, 116 (2): 157–85. Lapie, Paul (1902). Logique de la Volonte´. Paris: Fe´lix Alcan. Lemos, Noah (1994). Intrinsic Value: Concept and Warrant. Cambridge: Cambridge University Press. ——(1998). ‘Organic Unities’. Journal of Ethics, 2: 321–37. Mackie, J. L. (1977). Ethics: Inventing Right and Wrong. New York: Viking Press. McKeever, Sean and Ridge, Michael (2006). Principled Ethics. Oxford: Clarendon Press. McNaughton, D. and Rawling, P. (1991). ‘Agent-Relativity and the Doing/Happening Distinction’. Philosophical Studies, 63: 167–85. ——(1993). ‘Deontology and Agency’. Monist, 76: 81–100. ——(1995). ‘Value and Agent-Relative Reasons’. Utilitas, 7: 31–74. ——(2002). ‘Conditional and Conditioned Reasons’. Utilitas, 14 (2): 240–48. Mellor, D. H. and Oliver, A. (eds.) (1997). Properties. Oxford: Oxford University Press. Morgan, C. Lloyd (1923). Emergent Evolution. London: Williams and Norgate. Moore, G. E. (1993 [1903a]). Principia Ethica. Cambridge: Cambridge University Press. ——(1903b). ‘The Refutation of Idealism’. Mind, 12: 433–53. ——(1922). ‘The Conception of Intrinsic Value’. Philosophical Studies, 253–74. Mulligan, Kevin (1998). ‘From Appropriate Emotions to Values’. The Monist, 84 (1): 161–89. ——(2006). ‘Geist (and Gemüt) vs Life—Max Scheler and Robert Musil’, in R. Calcaterra (ed.), Le Ragioni del Conoscere e dell’Agire. Scritti in onore di Rosaria Egidi. Milan: Franco Angeli, pp. 366–78. ——(2008). ‘Scheler: Die Anatomie des Herzens oder was man alles fühlen kann’, in H. Landweer and U. Renz (eds.), Klassische Emotions-theorien von Platon bis Wittgenstein. Berlin: de Gruyter, pp. 589–612. ——(2009a). ‘Values’, in R. Poidevin, P. Simons, A. McGonigal, and R. Cameron (eds.), The Routledge Companion to Metaphysics. London: Routledge, pp. 401–11. ——(2009b). ‘Emotions and Values’, in P. Goldie (ed.), Oxford Companion to the Philosophy of Emotions. Oxford: Oxford University Press, pp. 475–500. Nagel, Thomas (1970). The Possibility of Altruism. Oxford: Oxford University Press. ——(1986). The View from Nowhere. New York: Oxford University Press. Newton-Smith, W. (1973). ‘A Conceptual Investigation of Love’, in Alan Monteﬁori (ed.), Philosophy and Personal Relations. London: Routledge & Kegan Paul, pp. 113–36. Nozick, Robert (1981). Philosophical Explanations. Oxford: Oxford University Press. Olson, Jonas (2003). ‘Revisiting the Tropic of Value: Reply to Rabinowicz and RønnowRasmussen’. Philosophy and Phenomenological Research, 67: 412–22. ——(2004). ‘How to Pass the Buck’. (circulating manuscript). ——(2005). Axiological Investigations. Uppsala: Acta Universitatis Upsaliensis. ——(2009). ‘The “Wrong Kind of Solution to the Wrong Kind of Reason Problem” ’. Utilitas, 21 (02): 225–32. Olson, Jonas and Svensson, Frans (2005). ‘Regimenting Reasons’. Theoria, LXXI, (3): 203–14.

176

BIBLIOGRAPHY

Olson, Jonas and Timmons, Mark (forthcoming). ‘A. C. Ewing’s First and Second Thoughts on Metaethics’, in Thomas Hurka (ed.), Underivative Duty: British Moral Philosophers from Sidgwick to Ewing. Oxford: Oxford University Press. O’Neill, John (1992). ‘The Varieties of Intrinsic Value’. Monist, 75: 119–37. Parﬁt, Derek (1979). ‘Prudence, Morality and the Prisoner’s Dilemma’. Proceedings of the British Academy, 65: 539–64. ——(1984). Reasons and Persons. Oxford: Clarendon Press. ——(2001). ‘Rationality and Reasons’, in Dan Egonsson, Jonas Josefsson, Bjo¨rn Petterson, and Toni Rønnow-Rasmussen (eds.), (2001), pp. 17–41. ——(circulating ms). ‘Climbing the Mountain’. ——(circulating ms). ‘On What Matters’. Pascal, Blaise (1995 [1867]). Pense´es and Other Writings, translated by Honor Levi (Oxford World’s Classics). Oxford: Oxford University Press. Paul, E. F., Miller, Jr., Fred D., and Paul, Jeffrey (eds.) (1992). The Good Life and the Human Good. Cambridge: Cambridge University Press. Persson, Ingmar (2005). The Retreat of Reason: A Dilemma in the Philosophy of Life. Oxford: Oxford University Press. ——(2007). ‘Primary and Secondary Reasons’, in T. Rønnow-Rasmussen, B. Petersson, J. Josefsson, and D. Egonsson (eds.), pp. 1–16. Pettit, Philip (1987). ‘Universality Without Utilitarianism’. Mind, 72: 74–82. ——(1988). ‘The Paradox of Loyalty’. American Philosophical Quarterly, 25: 163–77. Pettit, Philip and Smith, Michael (1990). ‘Backgrounding Desire’. The Philosophical Review, 99: 565–92. Portmore, Douglas W. (2001). ‘McNaughton and Rawling on the Agent-Relative/AgentNeutral Distinction’. Utilitas, 13: 350–56. Price, Richard (1897). A Review of the Principal Questions in Morals, ed. L. A. Selby-Bigge. Oxford: Clarendon Press. Pucecetti, Roland (1993). ‘Does the Universe Exist Because it Ought To? A Critique of Extreme Axiarchism’. Dialogue, 32: 651–57. Rabinowicz, Wlodek (2008). ‘Value Relations’. Theoria, 74: 18–49. ——(2009). ‘Values Compared’. Polish Journal of Philosophy, III (1): 73–96. ¨ sterberg, Jan (1996). ‘Value Based on Preferences’. Economics and Rabinowicz, Wlodek and O Philosophy, 12: 1–27. Rabinowicz, Wlodek and Rønnow-Rasmussen, Toni (2000). ‘A Distinction in Value: Intrinsic and for its Own Sake’. Proceedings of the Aristotelian Society, 100 (1): 33–51. ——(2003a). ‘Tropic of Value’. Philosophy and Phenomenological Research, 66: 389–403. ——(eds.) (2003b). Patterns of Value: Essays on Formal Axiology and Value Analysis, vol. 1. Lund: Lund Philosophy Reports. ——(2004a). ‘The Strike of the Demon: On Fitting Pro-Attitudes and Value’. Ethics, 114: 391–423. ——(eds.) (2004b). Patterns of Value: Essays on Formal Axiology and Value Analysis, vol. 2. Lund: Lund Philosophy Reports. ——(2006). ‘Buck-Passing and the Right kind of Reasons’. Philosophical Quarterly, 56: 114–20. Raz, Joseph (2000). ‘The Truth in Particularism’, in Brad Hooker and Margaret Little (eds.), Moral Particularism. Oxford: Clarendon Press, pp. 48–78.

BIBLIOGRAPHY

177

——(2006). ‘The Trouble with Particularism’. Mind, 115: 99–119. Regan, Donald H. (2004). ‘Why Am I My Brother’s Keeper?” in R. Jay Wallace, Philip Pettit, Samuel Schefﬂer, and Michael Smith (eds.) (2004), pp. 202–30. Reisner, Andrew (2009). ‘Abandoning the Buck Passing Analysis of Final Value’. Ethical Theory and Moral Practice, 12: 379–95. Ridge, Michael (2005). ‘Agent-Neutral vs. Agent-Relative Reason’, in Edward N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Winter 2005 edition) . Roberts, Robert (2003). Emotions: An Essay in Aid of Moral Psychology. Cambridge: Cambridge University Press. Rønnow-Rasmussen, Toni (1993). Logic, Facts and Representation: An Examination of R. M. Hare’s Moral Philosophy. Lund: Lund University Press. ——(1998). ‘Moral Realists and Moral Experts’, in Carsten Bengt Pedersen and Niels Thamassen (eds.), Nature and Lifeworld: Theoretical and Practical Metaphysics. Odense: Odense University Press, pp. 281–99. ——(2001). ‘L. W. Sumner’s account of Welfare’, in Juan Jose´ Acero, Francesc Camo´s Abril, and Neftalı´ Villanueva Ferna´ndez (eds.), Actas del III Congreso de la Sociedad Espan˜ola de Filosofı´a Analı´tica. Granada: Universidad de Granada, pp. 281–85. ——(2002a). ‘Hedonism, Preferentialism, and Value Bearers’. Journal of Value Inquiry, 36: 463–72. ——(2002b). ‘Instrumental Values–Strong and Weak’. Ethical Theory and Moral Practice, 5: 23–43. ——(2003). ‘Subjectivism and Objectivism; An Outline’, in W. Rabinowicz and Toni Rønnow-Rasmussen (eds.), pp. 246–63. ——(2004). ‘Recension av Stephen Darwalls Welfare and Rational Care’. Sats—Nordic Journal of Philosophy, 4 (2): 171–80. ——(2007a). ‘Analysing Personal Values’. Journal of Ethics, 11 (4): 405–35. ——(2007b). ‘Dislodging Butterﬂies from the Supervenient’, in Stephen Voss (ed.), Philosophical Anthropology, vol. IX, Proceedings of the 2003 Istanbul World Congress. Ankara: Philosophical Society of Turkey, pp. 236–37. ——(2008a). ‘Buck-Passing Personal Values’, in David Chan (ed.), Values, Rational Choice, and the Will: New Essays in Moral Psychology. Stevens Point: Springer, pp. 37–51. ——(2008b). ‘Love, Value and Supervenience’. International Journal of Philosophical Studies, 16 (4): 495–508. ——(2009a). ‘Normative Reasons and the Agent-Neutral/Relative Dichotomy’. Philosophia, 37: 227–43. ——(2009b). ‘On For Someone’s Sake Attitudes’. Ethical Theory and Moral Practice, 12: 397–411. Rønnow-Rasmussen, T., Petersson, B, Josefsson, J., and Egonsson, D. (eds.) (2007). Hommage a` Wlodek. Philosophical papers dedicated to Wlodek Rabinowicz. <www.ﬁl.lu.se/hommageawlodek>. Rønnow-Rasmussen, Toni and Zimmerman, Michael J. (eds.) (2005). Recent Work on Intrinsic Value. Dordrecht: Springer. Rosati, Connie (2008). ‘Objectivism and Relational Good’. Social Philosophy & Politics, 25 (1): 314–49. Ross, W. D. (1930). The Right and the Good. Oxford: Clarendon Press. ——(1939). Foundations of Ethics. Oxford: Clarendon Press.

178

BIBLIOGRAPHY

Royce, Josiah (1891). ‘Discussions: The Outlook of Ethics’. International Journal of Ethics, I: 106–11. Scanlon, Thomas Michael (1998). What We Owe to Each Other. Cambridge, MA: Harvard University Press. ——(2002). ‘Reasons, Responsibility, and Reliance: Replies to Wallace, Dworkin, and Deigh’. Ethics, 112: 507–28. Schefﬂer, Samuel (ed.) (1986). Consequentialism and Its Critics. New York: Oxford University Press. Schilpp, P. A. (ed.) (1942). The Philosophy of G. E. Moore (The Library of Living Philosophers). New York: Tudor Publishing Company. Schroeder, Mark (2007). ‘Reasons and Agent-Neutrality’. Philosophical Studies, 135: 279–306. ——(2009). ‘Buck-Passers’ Negative Thesis’. Philosophical Explorations, 12: 341–47. Selby-Bigge, L. A. (ed.) (1897). British Moralists, vol. 2. Oxford: Clarendon Press. Sen, Amartya (1999). Development as Freedom. Oxford: Oxford University Press. Sidgwick, Henry (1907 [1874]). The Methods of Ethics, 7th edn. London: Macmillan. Skorupski, John (1995). ‘Agent-Neutrality, Consequentialism, Utilitarianism . . . : A Terminological Note’. Utilitas, 7: 49–54. ——(2007a). ‘Buck-Passing about Goodness’, in T. Rønnow-Rasmussen, B. Petersson, J. Josefsson, and D. Egonsson (eds.), pp. 1–15. ——(2007b). ‘What is Normativity?’ Diputatio, II (23): 247–69. ——(2010). The Domain of Reasons. Oxford: Oxford University Press. Smith, Adam (1759/1790). The Theory of Moral Sentiments. London: A. Millar. Smith, Michael (2003). ‘Neutral and Relative Value after Moore’. Ethics, 113: 576–98. Sobel, David (2002). ‘Varieties of Hedonism’. Journal of Social Philosophy, 33: 240–56. ——(2005). ‘Pain for Objectivists: The Case of Matters of Mere Taste’. Ethical Theory and Moral Practice, 8: 437–57. Sprigge, Timothy L. S. (1988). The Rational Foundations of Ethics. London: Routledge. Strandberg, Caj (2004). Moral Reality: A Defense of Moral Realism. Lund: Media-tryck, Lund University. ——(2008). ‘Particularism and Supervenience’, in Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, vol. 3. Oxford: Oxford University Press, pp. 249–60. Stratton-Lake, Philip (2005). ‘How to Deal With Evil Demons: Comment on Rabinowicz and Rønnow-Rasmussen’. Ethics, 115: 788–98. Sumner, Leonard. W. (1996). Welfare, Happiness, and Ethics. Oxford: Clarendon Press. Swanton, Christine (1995). ‘Proﬁles of the Virtues’. Paciﬁc Philosophical Quarterly, 27: 335–44. Urban, Wilbur M. (1909). Valuation, Its Nature and Laws, Being an Introduction to the General Theory of Value. London: Sonnenschein/New York: Macmillan. Vlastos, Gregory (1969). ‘Human Worth, Merit and Equality’, in Joel Feinberg (ed.), pp. 141–52. Wallace, R. Jay (2002). ‘Scanlon’s Contractualism’. Ethics, 112: 429–70. Wallace, R. Jay, Pettit, Philip, Schefﬂer, Samuel, and Smith, Michael (eds.) (2004). Reason and Value: Themes from the Moral Philosophy of Joseph Raz. Oxford: Clarendon Press. Wiggins, David (1987). ‘A Sensible Subjectivism?’ in D. Wiggins, Needs, Values, Truth: Essays in the Philosophy of Value. Oxford: Blackwell, pp. 185–214. Williams, Bernard (1972). Morality. New York: Harper and Row.

BIBLIOGRAPHY

179

——(1985). Ethics and the Limits of Philosophy. Cambridge: Cambridge University Press. Williams, Donald Carry (1953). ‘On the Elements of Being’. Review of Metaphysics, 7: 3–18. Zimmerman, Michael J. (2001a). The Nature of Intrinsic Value. Lanham: Rowman and Littleﬁeld. ——(2001b). ‘Intrinsic Value and Individual Worth’, in Dan Egonsson, Jonas Josefsson, Bjo¨rn Petterson, and Toni Rønnow-Rasmussen (eds.) (2001), pp. 123–38. ——(2009). ‘Understanding What’s Good for Us’. Ethical Theory and Moral Practice, 12: 429–39.

This page intentionally left blank

Index abstract 4, 12, 20, 25, 26, 59, 158, 159, 166, 167 entities 143, 155, 164, 165, 168 admirability-makers 69 agent-neutral/relative reason distinction: essentialist sense 128–9 number approach 129, 131–2 Alexander, Samuel 4 n. 5 Alfano, Mark 29 Allen, Richard T. 52 n. 12 Alm, David xiii, 71 n. 16, 162 amnesia 79 analysis: ﬁtting attitude analysis of value 20, 22, 24, 26, 33, 90–1, 98–9, 121, 135, 140 naturalistic analysis 28, 105–6 Anderson, Elisabeth 25 n. 21 Anna v, 1, 130 Armstrong, David M. 166, 184 n. 26 atheism 112 atomism 148–9 attitude: and non-fungible persons 78–9, 80–1, 91 constitutive ground of value 15 dual role 37 experienced as correct 21 identity-involving attitudes 56 in the background 15 in the foreground 15 non-discerning 37, 68, 73 object-given reasons for 35, 38, 41, 43 state-given reasons for 35, 41, 43 Aydede, Murat 168 n. 27 Baron, Marcia W. 25 n. 21 Baudelaire, Charles 1 Baumann, Peter xiv being valued and being valuable 16 n. 29 beneﬁt(s) 2, 65, 84, 108 Bentham, Jeremy 155 n. 5, 158 better off 85, 88 biconditional account 42–43 Bjo¨rnsson, Gunnar xiv BPA; see also buck-passing account 27 Bradley, Ben 8, 167–8 Brady, Michael xiv, 111 n. 1, 120 n. 7 Brandt, Richard 21, 155n. 5, 156–7, 168 n. 27, 173–4 Brännmark, Johan xiii, 4 n. 5, 6 n. 11, 7 n. 14, 52 n. 13, 53 n. 14, 61, 83 n. 7 Brax, David xix, xiii, 168 n. 27

Braun, Eva 114, 117–19 Brentano, Franz viii, x, 34 n. 1 and 2, 21–4, 39–40, 42 British emergentists 4 n. 5 Broad, Charlie D. 4 n. 5, 21, 49 n. 4 Broome, John xiii, 44–5, 129 n. 8, 135 n. 14, 149 n. 25 Brülde, Bengt 15 n. 29 buck-passer 29, 43, 45, 97 buck-passing account 27, 29 n. 28, 42–5 Burges, J. A. 32 n. 44 Butchvarov, Panayot 163 n. 18 Butler, Joseph 124 n. 11 butterﬂy 11 Bykvist, Krister xiii, 33 n. 1 Campbell, Keith 164 n. 21, 166 Carlson, Erik xiii cart before the horse 163–4, 171 ‘Catch 22’ 34 n. 4 Chan, David xiii child molester 80 Chisholm, Roderick M. 5 n. 9 and n. 10, 21 n. 6 circularity 30 n. 36 and 37, 31–32, 48 n. 2 cognitivism 23 n. 11 Collin, Finn xiv Colonel Cathcart 34 n. 4 Colonel Korn 34 n. 4 concept of obligation 22 concrete: the distinction between abstract and 164–6 entities 12, 25–6, 31, 59, 104, 124–5, 135 n. 15, 154–5, 158 Conﬂation Problem 34 n. 2 consciousness 157, 168 false 124 constituted: values 12, 16–17, 89, 115 reasons 146–7 constitutive: attitude 14 grounds of (ﬁnal) value 9, 13, 19 of reason 134, 147, 148, 152 of welfare 84–5, 103, 107, 118, 170 powers 54 properties of the car 7 content-reason 40–1 correct to prefer 22 Cruft, Rowan xiv Crisp, Roger 27–9, 34 n. 4, 42–4, 148–9

182

INDEX

Dancy, Jonathan xiv, 10 n. 19, 20 n. 1, 24 n. 19, 80 n. 2, 127, 133, 137, 144–6, 148 n. 152, 153 Danielsson, Sven xiii, 7 n. 14, 11, 21, 24 n. 18, 34 n. 3, 40–2 D’Arms, Justin 31, 34 n. 2, 72 Darwall, Stephen 47, 88, 109, 114–18 DEFA 53–4, 88 De Facto Attitude analysis of Personal Value 53 Deigh, John 25 n. 20 delicacy and gracefulness 28 Demon 34–42, 45 deontic notion 24 n. 18 derivative value 104, 109, 155, 161, 163, 167 diachronic identity 57, 80 Dickie, George 52 dignity 88, 158 distinctive feeling view 21 n. 5 duplication 81 n. 3 Dworkin, Gerald 25 n. 20 Egonsson, Dan xiii, 153, 158 n. 13 Ehrenfels, Christian von 21 Elly v, 70 enabler 10 n. 19, 127, 144–6, 148–52 end 2–3, 9, 15–16, 20–2, 26–7, 33, 55, 60–4, see also sake; objective Epicurus 155 n. 5, 158 Ewing, A. C. xi, 21, 22–4, 25 n. 19, 30–1, 39, 42 Ewing’s metaethical views 23 n. 11 and n. 13. Ewing’s notion of pro attitude 23 Explanatory reasons 79, 133–5, 138 extrinsic ﬁnal value 5–7, 15–16, 26, 53–4, 122, 153–4 FA 20, 32, 33, 102 fact: as obtaining of state of affairs 143, 154–5, 159–60 fact-independently 11 FAP 109, 111–12, 114, 116–22, 124 FAP* 78–9, 83–7, 92–3 FAP-intrinsic 90 Favour x for its own sake for b’s sake 58–62 favourer 127, 144–51 favourer on condition 146, 148–9 Feldman, Fred xiv, 21 n. 5, 156 Fetishism 122–4 FFSS attitude 58–61 ﬁnal value: extrinsic 5–7, 15–16, 26, 53–4, 122, 153–4 the FA analysis of 27, 33 is invariant over possible worlds 12, 169 intrinsic 89, 91 isolating extrinsic 5–7 thick and thin varieties of 27

Findlay, John N. 21 n. 6 ﬁne-grained conceptual tools 19 Fitting-attitude analysis 20, see also FA founding father 21 history of 22 Fittingness: as an ethical concept 23 the ‘ought’ of 23 as primitive notion 39 Foot, Philippa 7 n. 14 Frankfurt. Harry J. 64, 72 n. 17 Fritzson, Fritz-Anton xiv, 14 n. 27 FSS attitude: 55 dropping an 78–9 faces a reductio ad absurdum 78 identity-involving attitudes 56 Non-identity attitude 74–6 two kinds of 57–8 Geach, Peter 6–7 Gert, Joshua 24 Gilbert, Margaret xiii, 121 God 10, 66, 112, 133–4, 164–5 GOL 99, 102 Goldman, Alan xiv GOLTP 99–102 GOLTPC 102 Good/goodness: absolutely 50 intrinsic 3, 50, 52 life 122–5 like yellow 4 plain 6–7 relational 98–9, 103 two distinctions in 9, 15 good-for: and welfare 46, 83–8, 103–4, 106–7, 114, 121–2 Grau, Christopher xiv, 65 n. 9, 82 n. 4 H1 154–5 H1* 155 Haddock, Adrian xiv Hadji, Ish xiv Hare, Richard M. 4 n. 5, 92, 119 Hartmann, Eduard von 3 n. 3 Heath W. White xiv, 36 n. 11 Heathwood, Chris xiv, 121–2 hedonic tone 156–9, see also properties: hedonic Hedonism attitudinal version of 156 Heeger, Robert Heyde, Erich 17 n. 32 Hieronymi, Pamela 44–5 Hitler, Adolf 114–15, 117–20 holism 148–9

INDEX

Humbertone, I. L. 32 n. 44 Hurka, Thomas xiv, xxvi, 50–1, 93 n. 17, 96–8, 102–3, 105 Identiﬁers 56, 77, 82–3, 93 n. 16 Identity attitudes 56, 75 identity-involving attitudes 56 I-justiﬁes 71 individual souls 164 Instrumental value In the Strong evaluative sense 51 In the Weak evaluative sense 52 interest 87 intrinsic personal values 89, 90 Invariance thesis 12–13, 169 Isolation test: intentional version 5 ontological version 5 Jacobson, Daniel 31, 34 n. 2, 72 Janus values 110–11 Juliet 65–6 justice 1, 123, 168 Justiﬁcation: two sorts of 71 identity-justiﬁcation 71, 74, 93 n. 16, 94 is called for 73 property 71, 94 Kagan, Shelly 8, 16 n. 30, 53 n. 15, 154 Kantians 8 Kath 81 n. 3 Katz, Leonard D. 157 n. 11, 168 n. 27 Kay 81 n. 3 Kim, Jaegwon 11 Kirshin, Simon xiv Korsgaard, Christine M. 9, 15, 16 n. 31, 61, 104, 151 n. 28 Kupperman, Joel 34 n. 4 La Giaconda 13 Lang, Gerald 41 n. 21 Langton, Rae 26, 154 n. 3 Lapie, Paul 3 n. 3 Lemos, Noah M. xiv, 4–5, 12 n. 23, 25 n. 21, 163 n. 19 Love: a volitional state 64 as apprehension 64 characterized as unconditional 80 higher mode of taking pleasure 21 occurrent 67 Pascal's view of 65 romantic view of 64, 80–1 Mackie, John L. 18 n. 34 Maurin, Ia xiv

183

Massin, Olivier xiv McKeever, Sean 149 n. 25 McNaughton, David 126 n. 2, 127, 129 Meinong, Alexius 21, 30 n. 36 Mellor, D. Hugh 166 n. 26 metaethical 6–7, 23 n. 13, 52, 105, 134 metaphysically queer 161 Miller, Adrian xiv Moore, G. E. 3–6, 8, 12, 17, 25 n. 19, 34 n. 3, 91 n. 14, 96, 98, 102, 103 n. 12, 47–50, 153, 169 Moorean, 8–9, 22, 91, 93 n. 17, 96, 98–102, 103 n. 12, 153 Moral: necessity 11 obligation 23 ought 24 reason 23, 119 thing to do 99 n. 7 wrongness 32 Morgan, C. Lloyd 4 n. 5 Mulligan, Kevin xiv, 27 n. 24, 30 n. 36, 31 n. 42, 55 n. 1, 64 n. 8, 83 n. 7 Musschenga, Albert xiv my own good 49–50, 102 Nagel, Thomas 126–30 Napoleon’s hat 53 Newton-Smith, W. 67 n. 10 nominalism 164 n. 21 non-derivative value 62, 153–5, 160, 163, 168, 171 non-fungible persons 67–8, 78–9, 91, 93–4 non-natural property 6, 29 normative concepts 24, 103 normative reason: see also reason Broome’s view of 44 Hieronymi’s view of 44–5 statements of 129 ultimately desire- or attitude-based 116 Nozick, Robert 52 n. 12 objective list theory 154 objective 26, 61–2, 64 Objectivism 14–15, 17–18, 54, 89, 162, 169 obtaining of states of affairs 12, 135, 143, 147, 154–5, 158–60, 166 Oddie, Graham xiv Olson, Jonas xiii, xiv, 7 n. 14, 34 n. 3, 40–2, 21, 23 n. 11, 24 n. 18, 25 n. 21, 28, 29 n. 27, 120 n. 7, 133 n. 11 Oliver, A. 166 n. 26 organic unity 61 ¨ sterberg, Jan xiii, 14, 155 O Ought: of moral obligation 23 of reasonableness 23–4

184

INDEX

parental love 80 Parﬁt, Derek xiv, 35, 38–41, 121 n. 8, 126–7, 129 n. 7, 133, 135 n. 14 Parﬁt’s approach 131, 132 particularity 82 Pascal, Blaise 65 patchwork solutions 119 PEF 132, 137, 139, 141–2, 145, 149 PEFI 132, 137, 152 personalizability feature 132, 141, 152 Personal value: Distinguished from impersonal value 55 being intrinsic 90 Persson, Ingmar xiv, 38 n. 15, 133 n. 11, 164 n. 20 Petersson, Bjo¨rn xiii Pettit, Philip 128–30, 133 n. 11 Piller, Christian xiv, 56 n. 2, 120 pleasing in itself 21 pleasure: distinctive feeling view 21 n. 5 feeling of pleasure 21, 167 higher mode of taking 21 instantiation of 163, 167 of the intellect 158 n. 13 Ryle’s view of 156 n. 8 sensory 156 n. 9, 157 pluralistic position 7, 22 Portmore, Douglas W. xiv, 129, 141–2 positive tone 156 n. 8, see also hedonic tone practical reason 95 n. 1, 116, 144, 152 Preferentialism, 153 cannot endorse H1 154 core claims 154 object version 155 satisfaction version 155 Price Richard 4 n. 7 private value 47, 91, 127 primitive notion: ‘correctness’ as 39, 42 ‘ﬁttingness’ as 39, 40, 42 Princess Diana’s dress 9 Properties: attitude-justifying roles 70 essential 35 dependent 10 hedonic 20, 155–9, 163, 167–8, 170–1 instantiations of 67, 163–5, 167 lower-order 27–8, 42–3 non-essential 9, 30–1, 34, 169 non-natural 6, 29 reason-providing 25 n. 20, 29, 43–5 relational 5, 6, 9, 36, 53, 83, 90, 93–4, 153 subjacent 10, 24, 25 n. 20, 52, 69, 71, 94, 111 supervenient 4, 10–11, 14–15, 94 value-making 14, 25, 36, 69–70, 94, 169

Quante, Michael xiv quasi-judgmentalism 31 R1 44–5 R2 44–5 Rabinowicz, Wlodek xviii, xix, 8–9, 14, 16 n. 30, 20 n. 1, 21 n. 6, 25 n. 21, 29, 30 n. 32, 33, 34 n. 5, 35–7, 39, 41 n. 20, 61, 140, 155, 160 n. 162–3, 164 n. 20, 165 n. 24 Rawling, Piers 126 n. 2, 127, 129 Raz, Joseph 136, 145, 148–50 realism 73 n. 20, 162 reason: see also moral reason content defeasible holding-reason 40–2 holding 40–2 qua facts 91, 135, 137, 139, 146, 150–2 right kind of 35, 39, 41, 70, 90 object-given 35, 38, 41, 43 state given 35, 41, 43 three varieties 134 reason for, see personalizability; recalcitrant emotions 72 reductandum 160–1, 163 reductans 160–1 reductio ad absurdum 78 Regan, Donald H. 48, 51 n. 10, 96–9, 101, 103, 104 n. 13, 105, 107 Reisner, Andrew xiv, 27 n. 24, 39 n. 17, 42 replica 81–3 RIA 144 Ridge, Michael 127, 129, 145 n. 25 Roberts, Robert C. 70 n. 15 Robinson, Paul xiv Romeo 65 Rnnow-Rasmussen, Toni 9, 12, 14, 16 n. 30, 20 n. 1, 21 n. 6, 25 n. 21, 32 n. 43, 34 n. 5, 36 n. 10, 37, 39 n. 18, 52 n. 11, 53 n. 15, 56, 61, 67, 98 n. 5, 99 n. 8, 160 n. 14 and 15, 162, 164 n. 20, 165 n. 24 and 25 Rosati, Connie 48–9, 51 n. 19, 85, 96, 99–105, 107. 109, 122 n. 9 Ross, W. D. 4, 26, 30–1, 145–6 Ryle, Gilbert 156 n. 8 Sake 61–3 saucer of mud 34–35, 36 n. 10, 37–8, 40, 45 Scanlon, Thomas xiv, 21, 25 n. 20, 27 n. 23, 48 n. 2, 133 Schefﬂer, Samuel 126 n. 2. Schroeder, Mark xiv, 45 n. 27, 132 n. 9 self-destructive life 88 n. 11 Sen, Amartya 163 n. 17 sexuality 72 Shah, Nishi xiv Shakespeare 1

INDEX

Sidgwick, Henry 21 n. 3, 47, 50 n. 9, 95 n. 1, 102, 124 n. 11, 155 n. 5, 158 Simplicity 7, 161–2 Sinnott-Armstrong, Walter 148 Skorupski, John xiv, 21, 32, 38–41, 127–8, 129 n. 8, 135 n. 14, 150 n. 27 Smart, J. J. C. 155 n. 5 Smith, Adam 21 n. 3 Smith, Michael 99, 103, 133 n. 11 Smith’s being happy 4 Sobel, David 157, 168 n. 27 Sobel, Howard xiii Socrates 3 n. 1, 11 Soﬁa v soldiers 87 source 9, 15, 16, 104 spatiotemporal location 100, 164–5 Sprigge, T. L. S. 156 n. 8 Stocker, Michael xiii Strandberg, Caj xiv, 10 n. 19, 145 n. 21, 149 n. 25 Stratton-Lake, Philip 36 n. 10 subjacent properties, see properties, Subjectivism 12, 13 n. 25, 14–18 subjectivists 12–13, 95 Supervenience: base of 9, 13, 24 n. 16, 25, 36, 68–9, 90, 106 consistency intuition 10–11 go from intra-world to inter-world 11 Strong supervenience 11 Svensson, Daniel xiii Svensson, Frans 133 n. 11 Swanton, Christine 25 n. 21 Swedish priest 72 n. 18 Tersman, Folke xiv, 34 n. 5 Thin-thin synonymity 84–4, 105–8 Thin-thick synonymity 105–7 Thomson, Judith Jarvis 7 n. 14 Timaeus 3 n. 1 Tropes 83, 160, 164–7 dyadic 165 uniqueness 68, 153 universale 164–5 universalizability 79, 91–4 Urban,Wilbur Marshall 3 n. 3 Value: see also derivative, extrinsic; ﬁnal, intrinsic, instrumental; private

185

bearer monism 13, 160–1, 163 n. 17, 171 bearer dualism 153 bearer pluralism 170–1 computation 162 constitution 9, 14–15, 17, 169 end 2–3, 9, 15–16, 25–6, 61–4 realism 162 source 9, 15–16, 104 Value judgements: being universalizable 79, 92 Value-for: Abel 48 a teenager 111 Cain 48 a comatose person 100–1 Hitler 114–15, 117–20 more than one notion of 84 Vieth, Andreas xiv Vlastos, Gregory 3 n. 4 Voorhoeve, Alex xiv Wallace, Ray J. 25 n. 20 Welfare 19, 84–8, 103–4, 106, 114–15, 117–18, 122, 124 dependent on the attitudes of the welfare subject 17 good-for 46, 83–5, 107 intrinsic value of 121 respect 88 Welfarists 20, 84–85, 99 Well-being 19, 46, 86, 99, 106, 124 White, Heath W. xiv, 36 Wiggins, David 30 n. 37 Wikforss, Åsa xiv Williams, Bernard 7 n. 14, 27 n. 22 Williams, Donald. C. 164, 166 WKR, see wrong kind of reason problem wrong kind of reason problem 33, 35, 37–8, 40, 41 n. 21, 44 n. 26 Yeats, W. B. 64 Yossarian 34 n. 4 Zimmerman, Michael J. xi, xiii, xiv, 3 n. 1, 4–5, 7, 12 n. 22, 21, 24 n. 16, 25 n. 21, 81 n. 3, 84 n. 8, 97, 99 n. 7, 106 n. 15, 107–8, 116 n. 5, 134 n. 12, 138, 155 n. 5, 160 n. 14, 160 n. 16, 164 n. 20, 165, 166 n. 26

Greek Personal Names: Their Value as Evidence

Value Investing For Dummies (For Dummies (Business & Personal Finance))

Value Investing For Dummies (For Dummies (Business & Personal Finance))

Sentimental Value

Natural Value

Shock Value

Personal Finance

Personal Demon

Personal Demon

Personal Demon

Personal Demon

Personal Recognizance

Personal Demon

Personal Demon

Personal Trainers

Personal Jesus

Personal Demons

Personal Demon

Personal Demon

Nothing Personal

Personal Training

Personal Tutor

Personal Nutrition

Added Value

Schlock Value

Personal History

Personal Demon

Personal Leave

Personal Tutor

Personal Demon

Personal Value

Greek Personal Names: Their Value as Evidence

Value Investing For Dummies (For Dummies (Business & Personal Finance))

Value Investing For Dummies (For Dummies (Business & Personal Finance))

Sentimental Value

Natural Value

Shock Value

Personal Finance

Personal Demon

Personal Demon

Personal Demon

Personal Demon

Personal Recognizance

Personal Demon

Personal Demon

Personal Trainers

Personal Jesus

Personal Demons

Personal Demon

Personal Demon

Nothing Personal

Personal Training

Personal Tutor

Personal Nutrition

Added Value

Schlock Value

Personal History

Personal Demon

Personal Leave

Personal Tutor

Personal Demon

Recommend Documents