Value in Ethics and Economics 0674931904, 9780674931909

Elizabeth Anderson offers a new theory of value and rationality that rejects cost-benefit analysis in our social lives a

802 131 41MB

English Pages 262 [129] Year 1995

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Value in Ethics and Economics
 0674931904, 9780674931909

Citation preview

VALUE IN ETHICS AND ECONOMICS Elizabeth Anderson



Harvard University Press Cambridge, Massachusetts London, England





*

»

*

.



Copyright © 1993 by the President and Fellows of Harvard College All rights reserved Printed in the United States of America Second printing, 1995

Library of Congress Cataloging-in-Publication Data Anderson, Elizabeth, 1959Value in ethics and economics / Elizabeth Anderson, p. cm. Includes bibliographical references and index. ISBN 0-674-93189-0 (cloth) ISBN 0-674-93190-4 (pbk.) 1. Values. 2. Value. 3. Reason. 4. Economics—Moral and ethical aspects. 5. Markets—Moral and ethical aspects. 6. Decision -making. I. Title. BD232.A48 1993 121'.8—dc20 93-365

For David

Contents

Preface ! 1/ A Pluralist Theory of Value

1

1.1 A Rational Attitude Theory of Value

1

1.2 Ideals and Self-Assessment

5

1.3 How Goods Differ in Kind (I): Different Modes of Valuation

2

3

1.4 How Goods Differ in Kind (II): Social Relations of Realization

11

An Expressive Theory of Rational Action

11

2.1 Value and Rational Action

11

2.2 The Framing of Decisions

22

2.3 The Extrinsic Value of States of Affairs

26

2.4 Consequentialism

30

2.5 Practical Reason and the Unity of the Self

38

Pluralism and Incommensurable Goods

44

3.1 The Advantages of Consequentialism

44

3.2 A Pragmatic Theory of Comparative Value fudgments

41

viii •

Contents • ix

Contents

3.3 Incommensurable Goods

. 55

3.4 Rational Choice among Incommensurable Goods 4

59

168

8.1 The Case of Commercial Surrogate Motherhood

168

8.2 Children as Commodities

110

8.3 Women's Labor as a Commodity

115

65

4.1 The Test of Self-Understanding

65

8.4 Contract Pregnancy and the Status of Women

182

4.2 The Hierarchy of Values

66

8.5 Contract Pregnancy, Freedom, and the Law

185

4.3 Agent-Centered Restrictions

13

Cost-Benefit Analysis, Safety, and Environmental Quality

190

9.1 Cost-Benefit Analysis as a Form of Commodification

190

9.2 Autonomy, Labor Markets, and the Value of Life

195

9.3 Citizens, Consumers, and the Value of the Environment

203

9.4 Toward Democratic Alternatives to Cost-Benefit Analysis

210

Conclusion

211

Notes

223

References

231

Index

241

Criticism, Justification, and Common Sense

86 91

5.1 A Pragmatic Account of Objectivity

91

5.2 The Thick Conceptual Structure of the Space of Reasons 5.3 How Common Sense Can Be Self-Critical 5.4 Why We Should Ignore Skeptical Challenges to Common Sense Monistic Theories of Value\ 6.1 Monism

9

19

4.5 A Self-Effacing Theory of Practical Reason?

6

Is Women's Labor a Commodity?

Self-Understanding, the Hierarchy of Values, and Moral Constraints

4.4 Hybrid Consequentialism

5

8

91 104 112

111 V

6.2 Moore's Aesthetic Monism 6.3 Hedonism 6.4 Rational Desire Theory 7 \ The Ethical Limitations of the Market 1.1 Pluralism, Freedom, and Liberal Politics 1.2 The Ideals and Social Relations of the Modern Market 1.3 Civil Society and the Market

111 119 123 129 141 141 143 141

1.4 Personal Relations and the Market

150

1.5 Political Goods and the Market

158

1.6 The Limitations of Market Ideologies

163

Preface

Why not put everything up for sale? I first began wondering about this question more than a decade ago, when political theories that advocated virtually unlimited market expansion were enjoying a resurgence that continues to this day. Since then the market itself has expanded into new domains, such as human organs and women's reproductive powers. We have seen the labor union movement, which once imposed powerful constraints on labor markets, in dramatic retreat from the private sector in the United States. Market deregulation and privatization have dominated economic policies in North America and Western Europe, while Eastern Europe is now opening up to capitalist development. People have increasingly withdrawn from civic life to malls and privately developed, sheltered "communities," while public spaces in inner cities are used to "house" the homeless and the mentally ill. Most of the debates about these developments have concentrated on questions of efficiency and income distribution. Although these are important issues, I do not--believe~itet~th^^ concerns we should have about theve^t^aHimitations of the markej. We should also care about what sorts of people and eeimrnirritiH^ of ourselves when we treat women as commercial baby factories, public spaces of social interaction as places either to shop or to avoid, and the natural environment as just another economically exploitable resource. In this book I attempt to articulate and justify these other sorts of concerns. When I first turned to philosophy and social science to help me think about the proper scope of the market, I didn't find what I was looking for. The dominant models of human motivation, rational choice, and value in

xii • Preface

these disciplines seem tailor-made to represent the norms of the market as universally appropriate for nearly all human interaction. According to the prevailing theories of value, people realize their good in having their wants satisfied. Markets are represented as generically appropriate vehicles for satisfying anyone's wants. According to the prevailing theories of rationality, people act rationally when they maximize their "utility" (welfare or want-satisfaction). Market choices provide the paradigm for this kind of rationality, which social scientists have eagerly generalized to cover the entire domain of human action. So markets are represented as the generically rational form of human organization. To count as rational, any other domain of human interaction would have to be governed by the same principles as the market. People can maximize utility only if they can find a common measure of value for all their options. Markets seem to provide such a common measure because they can put a cash value on almost anything people want. One could find room within the prevailing .theories of value and rational choice to question much of what markets do. Still, these theories share with economistic political theories several features that make this task more difficult. One is a socially impoverished conception of the individual. TQie^ejheojie^reE£ esen ^^ individual adult as freely forming and expressinj^^ apart from any particular social contexts or relations to others. TmTiE3rvT3uaIis^c picture of1 aTational person, as self-sufficient and independent, of others, supports a consumerist ideology that represents the individual as most free and rational in his market choices, where he need not concern himself with anyone else in deciding what to buy. This obscures the role of dialogue with others in making sense of ourselves and the role of social norms in shaping reasonable desires. It also leads to a psychologically impoverished conception of an individuals concerns. The prevailing theories of value and rationality suppose that when people value or care about something, they are engaging only one basic attitude or response—desire, perhaps, or pleasure—which can vary quantitatively but not qualitatively. And this view, in turn, leads to a drastically reductionistic or monistic view of value. Being valuable becomes a matter of having a single property or arousing a single response in us. Goods differ in quantity, as they arouse more or less of the same response, but not in quality or in kind. My original interest in the limits of markets led me to formulate a new theory of value and rationality that avoids the defects of the dominant theories. My theory emphasizes the richness and diversity of our concerns and finds a place for the full range of our responses to what we value. We

Preface • xiii

don't respond to what we value merely with desire or pleasure, but with love, admiration, honor, respect, affection, and awe as well. This allows us to see how goods can be plural, how they can differ in kind or quality: they differ not only in how much we should value them, but in how we should value them. In trying to make sense of the different ways we have of valuing things, we arrive at a socially integrated conception of the rational person. Being rational is a matter of intelligibly expressing our varied concerns to others. To do this, we must govern our conduct by shared norms established in dialogue with others, norms that are constitutive of different spheres and roles of social life. This socially grounded view of value and rationality, in turn, provides the key to understanding the ethical limitations of markets. If different spheres of social life, such as the market, the family, and the state, are structured by norms that express fundamentally different ways of valuing people and things, then there can be some ways we ought to value people and things that can't be expressed through market norms. We have to govern their production, circulation, and enjoyment through the norms of other social spheres to value them adequately. This book covers a lot of ground, from theories of value and rational choice, to disputes about justification and the objectivity of values, to theories of freedom, autonomy, markets, and politics. Different readers are therefore likely to be interested in different parts of this book. Those who are primarily interested in markets and politics should read Chapter 1, and then Chapters 7, 8, and 9. Those who are primarily interested in the theory of value should read Chapters 1 and 6, then §2.3, §§3.2—3.3, and §4.2. Those who are primarily interested in the theory of practical reason should read Chapters 1 through 5. Those who are interested in asking what value judgments mean, whether they express beliefs or emotions or other attitudes, and whether they refer to really existing values will not find me much engaged with these issues. Because my own inclinations are pragmatic, I prefer to set aside these semantic and metaphysical disputes and concentrate on normative questions. However, my investigations expose some features of our evaluative practices and experiences that any metaethical theory should accommodate. In Chapters 1 through 6 I discuss phenomena potentially relevant to these disputes. The last three chapters of this book contain material I have published before. Chapter 7 is a revised and expanded version of "The Ethical Limitations of the Market," Economics and Philosophy 6 (1990): 179-205. Chapter 8 is a revised version, with replies to my critics, of "Is Women's Labor a Commodity?" Philosophy and Public Affairs 19 (1990): 71-92.

Preface

Portions of Chapter 9 are drawn from "Values, Risks, and Market Norms," Philosophy and Public Affairs 17 (Winter 1988): 54-65. I thank Cambridge University Press for permission to reprint the first article, and Princeton University Press for permission to reprint the last two. I am grateful for the support of several institutions in aiding my research. The Woodrow Wilson National Fellowship Foundation awarded me a Charlotte W Newcombe Fellowship in 1986-1987. The Institute for the Humanities at the University of Michigan supported my research in 1989-1990. The University of Michigan and the Philosophy Department provided course relief in Fall 1990 and Fall 1991. Numerous departments and academic conferences have invited me to present work in progress. I thank them for the opportunity to open my views to critical scrutiny before I committed them to print. I take great pleasure in thanking the many friends and colleagues who have offered me support, advice, and criticism while I was writing this book. I owe an immense debt to Hugh Lacey, who first set me on the path of philosophy and has provided insight and inspiration to me ever since. John Rawls, Thomas Scanlon, and Burton Drdben advised me on my first attempts to formulate my views about markets\and values when I was in graduate school. My colleagues at the University^of Michigan, including Stephen Darwall, Allan Gibbard, Don Herzog, David Hills, Peter Railton, and David Velleman, read various chapters in painstaking detail and offered criticisms that have made this book immensely better than it otherwise would have been. I learned a lot from collaborating with Rick Pildes on a paper that helped me frame some central arguments of the book. I have profited from conversations with Jim Conant, Ann Cudd, John McDowell, Michael Pakaluk, Michael Sandel, Marion Smiley, Michael Stocker, Cass Sunstein, Charles Taylor, and Paul Weithman. Alan Wertheimer, Debra Satz, Charles Beitz, and Richard Zeckhauser offered invaluable criticisms of published or presented versions of the last three chapters of my book. Ruth Chang and Heidi Feldman also provided numerous helpful comments. I am deeply grateful to my parents and my husband, David Jacobi, for their unflagging support of my philosophical pursuits.

Value in Ethics and Economics

1 • A Pluralist Theory of Value

1.1 A Rational Attitude Theory of Value People experience the world as infused with many different values. Friendships can be intimate, or merely convenient, charged with sexual excitement, or mellow. A subway station can be confining, menacing, and dumpy, or spacious, welcoming, and sleek. When people attribute goodness or badness to some thing, person, relationship, act, or state of affairs, they usually do so in some respect or other: as dashing, informative, or tasty, delightful, trustworthy, or honorable, or as corrupt, cruel, odious, horrifying, dangerous, or ugly Our evaluative experiences, and the judgments based on them, are deeply pluralistic. I aim to explain and vindicate this pluralism of ordinary evaluative thought and to develop some of its practical and theoretical implications. This requires an investigation^ in J in valuing or caring about things, in expressing and ^6^^2^^^^ these phenomena will help us home in on what it is to be good and how we know things to be good. The suggestion that we have evaluative experiences has struck many philosophers as metaphysically eerie: science has discovered no "evaluative facts," or any organs of "moral sense," that enable us to discern the properties of "good" and "bad" in the world (Mackie 1977, pp. 38—42). We can dispel this mystery by recalling what ordinary experiences of value are like. We experience things not as simply good_o^jba^^buL^s_gQQcl or J3aj__in_particular Ijspgctsjjiat elicit distinct responses in us. There is nothing mysterious about finding a dessert delectable, a joke hilarious, a soccer match exhilarating, a revolution liberating. We also can find someone's compliments cloying, a task burdensome, a speech boring. To experience something as good is to be favorably aroused by it—to be

A Pluralist Theory of Value • 3 2 • A Pluralist Theory of Value

inspired, attracted, interested, pleased, awed. To experience it as bad is to be unfavorably aroused by it—to be shocked, offended, disgusted, irritated, bored, pained. Evaluative experiences are experiences of things as arousing particular positive or negative emotional responses in us. Evaluative experiences are relevant to questions concerning the good because they typically arouse or express our concerns about what we experience. Valuing or caring about things is more fundamental to understanding values than are experiences of value, for many things can be good which are not directly encountered in experience, but are known only through theory or description (Johnston 1989, p. 142). No particular qualities of experience need to accompany knowledge of the literacy rate, the justice of patterns or processes of wealth distribution, or the stability of habitats for endangered species. What makes such things candidates for goodness seems to be that we can care about them or value them. To value something is to have a complex of positive attitudes toward it, governed by distinct standards for perception, emotion, deliberation, desire, and conduct. People who care about something are emotionally involved in what concerns the object of care. Parents who love their children will normally be happy when their childlen are successful and alarmed when they are injured. They will be alert\o their needs, take their welfare seriously in their deliberations, and want ro take actions that express their care. These all express the way loving parents value their children. To experience something as valuable and to value it are not to judge that it is valuable. A person may laugh at a racist joke, but be embarrassed at her laughter. Her embarrassment reflects a judgment that her amusement was not an appropriate response to the joke. The joke was not genuinely good or funny: it did not merit laughter. A person could also judge that a joke is funny, but be so depressed that she can't bring herself to laugh at it. Such a judgment could be the occasion of further depression, because it makes her aware of her own deficient state of mind, too miserable even to appreciate a good joke. These observations support the following proposal: to judge that something is good is to judge that it is properly valued. And to judge that UlsT bad is.tp judge that it is prop,QrJy_4isyalued. Often people judge that something is good in some particular respect, as in being charming, or inventive. I suggest that the proposition "x is F," where F is a respect in which something is judged to be genuinely valuable, entails that x meets a particular standard F, and that x merits valuation in virtue of meeting F.1 One intrinsically values something when one values it in itself—that is,

apart from valuing anything else. I propose that the judgment that x is intrinsically valuable entails that (under normal conditions) x is properly intrinsically valued, independent of the propriety of valuing any other particular thing. Extrinsic values include but are not confined to instru-1 mental values. One may treasure an ugly, useless gift because it was given 1 ^~~ by a loved one. Such a gift is extrinsically valuable, in that one's valuation/ of it depends upon one's valuation of the giver. — R^ejlej^iyE-Jzaluej^ forms of self-assessment which are_embodied in second-order attitudes, or attitudes about «~^ other attitudejL_As we saw above, one may be embarrassed or depressed by one's failure to respond appropriately to what one judges to be good. One may be pleased by or proud of one's appropriate valuations. Ij3ror>ose_lhat thi^s so because the concepts of m e r i t i ^ : valued are rationality ^concepts. When we wonder whether something is appropriately valued, we wonder whether we would be making sense in valuing it. On my view, the investigation into what is worth our caring about is a quest for self-understanding, an attempt to make sense of our own valuational responses to the world. In §5.1, I will tie the project of rational self-understanding to social practices of justification. Here I will offer a provisional account of the story to come. The link between selfunderstanding and justification is provided by the fact that valuations are expressive states. They are bearers of meanings and subject to interpretation. Since meanings are public, I can understand my own attitudes only in terms that make sense to others. Attitudes are also partly constituted by norms that determine their proper objects. So the interpretation of attitudes involves their evaluation as well. I will argue that people interpret* and justify their valuations by exchanging_reasons Jor_them with the aim of reaching a common point of view from which others can achieve and reflectively endorse one another's valuations. To judge that one's valua- ^ tions make sense is to judge that they woulcTrSe endorsed from that \s!f^^^^^o^x^J2L^£SN. To Be rational is to be suitably responsivejtp reasons ojgered by those a t t e m p t i n g t o J ^ ^ L I ^ ^ J ^ ^ s ^ ^ S S ^ The terms in which we make sense of our valuations are given by our evaluative concepts. The opening of this chapter sampled some of the rich variety of concepts through which we describe evaluative experiences and express value judgments. Call a person's values whatever standards she accepts for evaluating persons, actions, and things. Evaluation is the process by which a person judges how far and in what ways different things meet her standards. An object's values consist of whatever properties it has, in virtue of which it meets various standards of value. I have proposed that

A Pluralist Theory of Value • 5

4 • A Pluralist Theory of Value

the judgment that an object meets an authentic standard of value entails that its meeting that standard makes it sensible for someone to value it. The standards of value for objects are standards of rationality for our responses to them. One of my values could be that bedrooms be cozy. If a given bedroom is cozy, then coziness is a value it has. Its coziness gives me a reason to feel comfortable in it and makes sense of my feeling snug when I retire there. Standards rationally adjust our valuations to their appropriate objects. Although all authentic values set standards for rational valuation, not every rational valuation of something depends upon its meeting some standard of value (Gaus 1990, pp. 70—71). Some ways of caring about things do depend upon their measuring up to particular standards of value—people don't admire athletes or musicians who lack dedication and skill—but other ways of valuing things do not. Parental love is like this. Parents can love infants independent of any valuable qualities they may have. Of course, loving another person will usually involve delight in some of that person s qualities, as when parents rave over the fact that little Melissa has her father's eyes. But this doesn't imply thatjthe parents think that having father's eyes merits anyone's raving, much less that their love for Melissa depends upon her having her father's eyesl Rather, parents express their love for an infant in part by adoring whatever features she has which can be adored. These features need not merit Valuation in their own right: parents can dote even on an ugly face. It follows that we have_two_conceptions of goods that_do not exactly coincide. On one view, a good is something that is appropriately valued. On the second, a good is a bearer or bundle of equalities that meet certain standards or requirements we (correctly) setJbrjt (Mackie 1977, pp. 55— ; 56). The_^e£oj^c^nce£tion defines a subset of the_ob^j^thatJa31-under the first: those things that merit valuation by meeting prior standards of /value? But jie^figt^amcepuon^ij^rnore basic, for it can be appropriate to / Yjlue.SQrne-thi»g&.-ox.persons mxeitain ways without.. their ni&etiagjndeV pendent standards oXevaluation—that is, without their meriting valuation. _]•__. twO conceptions of goods lead to two conceptions of the plurality of goods. On the first, goods are plural in that they are sensibly valued in fundamentally different ways. The opposing monistic view holds that all goods are the proper objects of a single evaluative attitude, such as desire, pleasure, or admiring contemplation. On the second conception, goods are plural in that the authentic evaluative standards they meet are fundamentally diverse. The opposing monistic view maintains that the apparently diverse standards for rational valuation can be reduced to some single

ground or explained by reference to a single good-constituting property, such as being desired or pleasant. The^rst conception of pluralism is more basic than the second because it explains why the second is true: we neecTa plurality of standards to make sense of the plurality of emotional responses and attitudes we have to things. The things that sensibly 'elicit delight are not generally the same things that merit respect or admiration. Our capacities for articulating our attitudes depend upon our understandings of our attitudes, which are informed by norms for valuation. To attempt to reduce the plurality of standards to a single standard, ground, or goodconstituting property threatens to obliterate the self-understandings in terms of which we make sense of and differentiate our emotions, attitudes, and concerns. To adopt a monistic theory of value as our self-understanding is to hopelessly impoverish our responsive capacities to a monolithic "pro" or "con" attitude or to mere desire and aversion. In identifying what is good with the proper objects of positive valuation, my theory follows Franz Brentano's. Brentano (1969, p. 18) hddjhat an object is good if and only if it is correct to love it, and bad if and only if it is correct to hatejt. My theory adds two main points to Brentano's. First, it views the concept of "correctness" as a rationality concept, H^jltP , ti^ My theory of value could be called a "rational_attitiicie_lhejor^'' according to which the attitudes engaged when we care about things involve not just feelings but judgment, conduct, sensitivities to qualities in what we value, and certain ways of structuring deliberation concerned with what we value.2 Second, thej^ej.s not justjoneJ^ay^to love or have a ^ro-attimde" toward things. There^are: different forn^^^ and there are ways of valuing things that are not love at all, such as respect and admiration. The variety of ways of caring about things is the source of^ pluralism in my theory of value. 1.2 Ideals and Self-Assessment Valuing and evaluation are distinct activities. In evaluation, people determine how far something meets the particular standards they set for it. In valuing something, people meet certain standards for caring about it, although they may be unaware of, may not endorse, and may not try to govern their actions by those standards. A person could care about something but judge himself contemptible for caring about it. For example, Max could discover to his dismay that he is absorbed by his own good looks, even though he judges his vanity contemptible. Evaluation is a

(Tie

A Pluralist Theory of Value • 7 6 • A Pluralist Theory of Value

means by which people come to rational self-understanding and selfgovernance of their own valuations. BepiiS£JtlLe.-S£aDiia^ J b j t j x e th^s^andards o£ i d valuations luations into into harmony, harmony, people people judge judge In bringing their evaluations and themselves worthy of positive valuation, or at least not worthy of negative valuation. This suggests that the grounds of a person's reflectively held values (if she has any) lie in her conceptions of what kind of person she ought to be, what kinds of character, attitudes, concerns, and commitments she should | have. I call such self-conceptions ideals. Ideals are objects not merely of desire, but of aspiration. The desires to be an exemplary mother or a U.S. Marine, to be a suave, sophisticated cosmopolitan or a self-made man, to be a champion of science over superstition or a zealous missionary devoted to spreading God's word are aspirations toward ideals with which we are familiar. Members of communities may have shared ideals, such as to be a citizen republic, culturally or racially pure, to be the artistic avantgarde, to live in holy matrimony or in harmony with nature. As these examples suggest, to call a self-conception an ideal is not necessarily to endorse it, but to imply that it is a possible object of admiration or condemnation, honor or disdain, and that the people who adapt it regard it as worthy. J Ideals set the standards of conduct and emotion people expect themselves to satisfy with regard to other people, relationships, and things. A U.S. Marine is supposed to be patriotic—to love his country, obey its leaders, and fight to the death for the causes it esteems. A connoisseur of fine art is supposed to cultivate an appreciation of subtle qualities in painting and sculpture and to be appalled at damage done to great works. A labor union activist is supposed to build solidarity with fellow members of the working class and to feel that "an injury to one is an injury to all." Such standards of conduct and emotion tell us how to care about things and people. We care about things and people in different ways, which express what I call different modes of valuation, such as love, respect, and admiration. Ideals give us perspectives from which to articulate and scrutinize the ways we value things. The core of an ideal consists in a conception of qualities of character, or characteristics of the community, which the holders regard as excellent and as central to their identities. Associated with this core is a conception of admirable conduct or worthy practices and projects that demand the cultivation, exercise, and expression of these qualities. An ideal is constitu-

tive of a person's identity if it governs her self-assessments and her responses to her achievement and failure and if she uses it to discipline her desires and frame her choices. Failure to live up to one's ideals will prompt shame, guilt, self-contempt, or other negative self-assessing emotions. Circumstances which prevent a person from realizing her ideals are likely to be experienced as humiliating and degrading, not just as frustrating. Ideals ground some crucial distinctions in the theory of value. One is between value and importance to a person. I have claimed that goods are things whose valuation is rational. An ambiguity exists here between what anyone could rationally value if she were in appropriate circumstances and what it makes sense for a particular person to value, given her circumstances and characteristics. I reserve the impersonal sense of rationality for the attribution of value to something and the personal sense for what is important to a person. There is a great diversity of worthwhile ideals, not all of which can be combined in a single life. Different ideals may require the cultivation of incompatible virtues or the pursuit of some projects that necessarily preclude the pursuit of others. Individuals with different talents, temperaments, interests, opportunities, and relations to others rationally adopt or uphold different ideals. Since ideals direct a person to specially value some worthwhile projects, persons, and things over others, they distinguish from among all goods those that are particularly important to the individual. That incompatible ideals are properly adopted by different persons explains why it doesn't make sense for everyone to take up the same attitudes toward the same things. There are far more potentially worthy objects of valuation than could occupy any one person's concern. The different relations individuals have toward persons and things help determine their proper attitudes toward them. This is obviously true for love. Radically different kinds of love are appropriate to different members of one's family, depending on one's relationship to them. That an individual stands in a particular relation to some persons or objects—say, as daughter, business partner, or inventor—partly determines the ideals rationally available to her, the importance these persons and objects have for her, and hence the appropriate attitudes she should take up toward them. So ideals distinguish among goods that play a more or less important role in a person s life. They also distinguish between goods that are important to a person just because she happens to care about them and goods that are important to her because they command her concern (Frankfurt 1988). In the former case, as long as the goods don't violate minimal impersonal standards for rational valuation, it doesn't matter for her self-

8 • A Pluralist Theory of Value

regard whether she cares about them or loses interest in them. In the latter case, whether she cares about them can reflect well or poorly on herself. A person sees her failure to live up to her core ideal aspirations in this light. Call goods of the former type weakly valued and those of the latter type strongly valued.3 People use ideals to cultivate and discipline their desires. Ideals function in this way because they are expressed in second-order desires, or desires to have or change other desires. If I uphold an ideal of integrity, I want myself to be motivated to stand up for my beliefs, and I want this desire to govern my actions even when it conflicts with my desire to maintain a favorable reputation. Not every second-order desire expresses an ideal. I could want to get rid of a desire simply because it is inconvenient. Perhaps my desire to linger on the telephone prevents me from getting on with my evening. Here I engage only my weak valuations, for I regard the desires in question as merely optional. I could choose to adopt a more leisurely attitude toward my affairs rather than to get rid of my desire to carry on with my friends over the phone. But I don't regard my desire for integrity as merely optional. No simple, unobjectionable change of perspective is available which would allow me to pander to others' opinions when my integrity is at stake. If I lack the desire for certain weakly valued ends, such [ as physical comfort, this might make me weird or quirky but not worthy I of contempt. If I lack the desire for strongly valued ends, such as integrity, ] this makes me base or deplorable in my own eyes.4 J In telling us how to value different goods, and in tying our valuations to our judgments of self-worth, ideals help structure the world of goods into different kinds. They draw boundaries between different classes of goods, setting them into circulation within distinct networks of social relations governed by distinct norms. This differentiation of ways of valuing things, socially embodied in different social spheres, provides the key to understanding how goods differ in kind. 1.3 How Goods Differ in Kind (I): Different Modes of Valuation Kant's moral philosophy provides a particularly illuminating example of how goods differ in kind: "In the kingdom of ends everything has either a price or a dignity. Whatever has a price can be replaced byjomething else as its equivalent; . . . whatever is above all price, and therefore admits of no equivalent, has a dignity" (Kant 1981, p. 40). In thjs-passage^ Kant expresses the view that there are two kinds of value, (elative worthand , intrinsic worth. £ very thing is either a mere means, wfthTTprice or

A Pluralist Theory of Value • 9

relative value, or an end in itself, with an intrinsic worth which Kant calls "dignity." Things that differ in the kind of worth they have merit different kinds of valuation. People value mere means by using them, but they value persons with dignity by respecting them. People express these different modes of valuation in part by deliberating about their objects in different ways—engaging in prudential calculation for use-values and in deliberation according to the categorical imperative for ends-in-themselves. Kant's ideal of human rationality grounded his distinction between the way we should value persons and the way we should value things. By considering other ideals that are widely recognized in U.S. culture, we can see that Kantian ethics is hampered by the fact that it recognizes only two ways of valuing things, use and respect. These two modes of valuation are not enough to account for the richness of our experiences of value and our practices. Three examples from his Lectures on Ethics, concerning the status of animals, inanimate nature, and adultery, illustrate some problems a two-valued ethic has in attempting to account for our concerns in a many-valued world (Kant 1979, pp. 239-241, 169). Although Kant recognized aesthetic value as a distinct category of non-moral worth, he failed to see that even the domain of morality is many-valued. Animals cannot be respected in a Kantian ethic, for to respect something in the Kantian sense is to act toward it in accordance with laws it would accept as a legislating member of the Kingdom of Ends. Animals are incapable of entering into the reciprocal relations based upon a conscious acceptance of common principles which membership in the Kingdom of Ends requires. But Kant's conclusion does not follow—that animals are mere means and may be used by us for any purpose that does not violate our duties to humans. We shouldn't be cruel to animals. Kant tried to account for this commonsense view by arguing that we have an indirect duty to humans to refrain from animal cruelty, because cruelty to animals makes us more likely to treat humans cruelly. This attempt to account for our duties to animals is strained. If someone is cruel to her pet, people condemn her action whether or not this behavior will increase her cruelty to people. Neither Kantian respect nor mere use captures the appropriate treatment of pets. The ideal of a pet owner includes much more than even the avoidance of cruelty and the provision of basic necessities—we criticize an owner for failing to show proper affection for her pet. Although we make fewer demands for our treatment of animals in the wild, there is a base line of care which we should show for all animals. I suggest that we call this

10 • A Pluralist Theory of Value

kind of valuation "consideration." Consideration is a way of caring which pays due regard for the interests of sentient beings, apart from whether they are rational.5 Kant also regarded inanimate nature as a mere means. The only duty we have to conserve natural habitats follows from our duty to leave future generations enough resources. Kant saw no reason to preserve natural habitats from destruction through consumption, only reason to ration this destruction over time. In the United States today, we recognize ideals expressed in environmental movements to preserve ecosystems and natural wonders which express a deeper concern for nature. Most U.S. citizens view the redwoods and the Grand Canyon as beautiful and wondrous things to be intrinsically valued. To regard these wonders only from the standpoint of their use-value to humans is base. But inanimate nature can neither be respected in a Kantian sense nor given the consideration owed to animals, since it has no interests of its own. What seems to be an appropriate mode of valuation for inanimate nature is rather what we may call "appreciation." A third problem for Kantian ethics concerns the difference between the badness of cheating on a business deal and the badness of cheating on one's husband or wife. Kant condemned both actions for one reason: theyf reflect a lack of respect for persons. He argued that adultery is a graver sin I than fraud because the marriage contract is more important than any business contract. This does not explain why the victims of these actsy typically experience different kinds of diminishment. The significance of adultery seems to lie not so much in its failure of respect—which it shares with fraud—as in its betrayal of love. Modern ideals of marriage demand of partners deeper forms of care for each other than commercial contracts do. When these forms of care are no longer forthcoming, their loss is felt more personally Use, respect, appreciation, consideration, and love are five different ways of valuing things. A little reflection suggests more modes of valuation, such as honor, admiration, reverence, and toleration. We are familiar with numerous modes and expressions of disvaluation as well: to shun, humiliate, mock, despise, ignore, desecrate, and so forth. My provisional account of how goods differ in kind is thus that they differ in kind if they are properly valued in different ways. Talk of different kinds of goods may be somewhat misleading, if we think of kinds as non-interbreeding species. I think of kinds of goods as more like literary genres: they can be hybridized, like the comedy-thriller; they can stand in different relations to different audiences, as heroic odes do to oral and written cultures; and

A Pluralist Theory of Value • 11

they can be categorized differently by different cultures, as myths are by cultures having and lacking a scientific cosmology. Ideals tell individuals how they should value different things, depending on their value and personal importance. Some goods merit a particular mode of valuation because they meet a standard of value: beautiful things are worthy of appreciation, rational beings of respect, sentient beings of consideration, virtuous ones of admiration, convenient things of use. Here the pluralism of values or standards underwrites the pluralism of kinds of goods. Other goods are appropriately valued in a particular way because of their relation to the valuing agent, which makes them important to him. People who have helped someone are owed gratitude, brothers and sisters are to be loved, one s children to be nurtured. Romantic love, patriotism, loyalty, the treasuring of heirlooms, and the cherishing of friends are modes of valuation connected to importance judgments, not just to impersonal value judgments. Here the kind of good a thing is for a person depends on her particular biography and social situation, her place in a network of relationships. To value or care about something in a particular way involves a complex of standards for perception, emotion, deliberation, desire, and conduct that express and thereby communicate one's regard for the objects importance. To love someone involves the performance of many actions which express that love, which show the beloved that he or she has a special importance to the lover. It entails particular ways of deliberating about questions concerning what is valued, questions which distinctively engage the agents perceptual dispositions and set certain considerations in priority over others. Parental love involves perceiving and attending to a child's needs and wants and giving the child's needs a certain priority in deliberation (over his wants and over other concerns). Finally, a mode of valuation includes distinctive emotional responses to the apprehension, achievement, and loss of things related to what is valued. Romantic love involves feeling grief when the beloved dies, despondency at her lack of reciprocation, exultation at her confession of a reciprocal love, jealousy when her affections are turned to another, alarm at her being harmed. These different ways of flourishing and suffering with regard to the beloved show her that she is loved, as opposed to merely liked or tolerated. 1.4 How Goods Differ in Kind (II): Social Relations of Realization I have thus far explained how goods differ in kind in terms of the different ways people properly care about them. Individuals are not self-sufficient in

12 * A Pluralist Theory of Value

their capacity to value things in different ways. I am capable of valuing something in a particular way only in a social setting that upholds norms for that mode of valuation. I cannot honor someone outside a social context in which certain actions, gestures, and manners of speaking are commonly understood to express honor. More important, I do not adequately express my honor for another unless others recognize my honor as appropriate. To care about something in a distinctive way, one must participate in a social practice of valuation governed by norms for its sensible expression. So the difference between, for example, appreciating something and using it lies in the social relations and norms within which we produce, maintain, distribute, preserve, and enjoy or otherwise realize the value of that thing. To realize a good as a particular kind of good we place it in a particular matrix of social relations. The following shall be my primary account of the heterogeneity of goods: goods differ in kind if people properly enter into different sorts of social relations governed by distinct norms in relation to these goods. It is proper for them to do so if it makes sense to value the goods in the ways expressed by these norms. For example, consider the status of music in the United States. We enjoy live "classical" music in special social settings—music halls—governed by distinctive cultural norms that express a regard for this art form as worthy of awe. Silence is to be observed as soon as the orchestra starts playing; even the pauses between movements may not be interrupted by applause. We are supposed to concentrate all our attention on the music itself. The audience may not openly criticize a performance in progress or suggest alterations. We are to be humbled by the majesty of the work and its performance, to receive it as instructive and uplifting, as somehow above us, yet as ennobling us through our polite, restrained admiration of it. To value the music in this way demands a clear separation between the audience and the music, expressed spatially in the separation of the orchestra pit or the stage from audience seating, temporally by the strictly separate times in which the musicians and the audience may express themselves, and functionally by the fact that audience members don't participate in the creation of the music itself. This ideal of classical music often involves the subordination of orchestral musicians to conductors and composers, who are regarded as superior aesthetic authorities. One extreme expression of this ideal can be found in the authentic music movement, where the virtuosity and interpretive skills of the musicians themselves are subordinated to the goal of reproducing as exactly as pos-

A Pluralist Theory of Value • 13

sible the musical sounds, and the techniques for producing those sounds, as they existed in the composer's time. The composer is considered a genius whose original intentions regarding performance must be absolutely respected, lest we fail to do justice to his music. Through these kinds of social norms, classical music in this culture is deemed a kind of sacred good. North Americans didn't always value classical music as a sacred good, worthy of awe (Levine 1988). Until the late nineteenth century, they celebrated classical music, especially Italian opera, as a highly popular form of entertainment, to be valued as audiences value athletic contests—and closer to the ways they value professional wrestling than golf or tennis! Audience members regarded the music as theirs to criticize, applaud, change, and perform at their own inclination. They made a raucous crowd, prone to rioting when performers did not heed their wishes. They often demanded that popular songs of the day be included in operas at a moment's notice, interrupted performances with critical comments, sat on stage, talked loudly and ate during performances. The social norms of music appreciation gave the audience a powerful set of claims on how the music was to be performed and enjoyed that expressed a view of its value as properly reflecting popular taste and sentiment rather than as educating or uplifting it from a higher standpoint. And they often participated en masse in performing the music itself. At the National Peace Jubilee of 1869, the Anvil Chorus from // Trovatore was performed with one hundred Boston firemen beating anvils with sledgehammers (Levine 1988, p. 105). So the kind of good classical music is—how we value it—is determined by the norms governing the relations among audience, composers, and performers. These norms in turn are governed by different aesthetic ideals. In the ideal of classical music as a sacred expression of supreme genius, performers are subordinated to composers, and audience to performers. The sharp distinction between classical and popular music also functions as a class-marker, giving the wealthier and more educated classes claim to a higher standing in the cultural hierarchy than those who prefer other genres, such as rock and country-Western. In the ideal of music as a popular expression of public taste, the hierarchy is reversed, or rather blurred, since the social roles of composer, performer, and audience are not as sharply distinguished as in the former case and the arbiters of good taste are not confined to a specific class. The aesthetic conceptions are ideals, because they give us standards for self-criticism as well as for criticism of art itself. We make ourselves different kinds of persons by creating, performing, disseminating, and

14 * A Pluralist Theory of Value

appreciating music in different ways, through different kinds of social relations. Music mediates our relations to one another and thereby creates different forms of community with different virtues and vices. Aesthetic ideals are highly contestable. By upholding the sacralized ideal of art, do we heighten our aesthetic appreciation or merely make snobs of ourselves? By upholding popular ideals that celebrate virtuosity and public participation, do we corrupt works of genius and debase ourselves by pandering to uneducated taste? Or do we rejuvenate our cultural identities by providing outlets for creative reinterpretation of our musical heritage? Our answers to questions like these help determine how it makes sense to value music. In §5.1, I will consider the prospects for justifying answers to such questions. An ideal-based pluralistic theory of goods does not concern itself exclusively with the qualities of the goods people enjoy. It also focuses on the realization of distinct ideals of the person and community, and it views goods as mediating these relations among people. Ideals require people to care about goods in particular ways, by embedding them in appropriate relations of production, protection, distribution, and enjoyment. Treating a good as a particular kind of good is as much a way of realizing and expressing appropriate relations among people as it is a way of properly valuing the good itself. So far I have just sketched the outlines of a map of the world of goods, taking ordinary practices and commonsense judgments as my guide. Most theories of value acknowledge a pluralism of goods, such as friendship, knowledge, and pleasure. My map reveals a proliferation of pluralisms beyond this. First, it recognizes a plurality of evaluative attitudes, such as love, admiration, and appreciation. Second, it recognizes a plurality of values or standards, such as beauty, convenience, and loyalty, by which we evaluate different goods and adjust our attitudes toward them. Third, it recognizes a plurality of different kinds of goods, distinguished by the complexes of attitudes it makes sense to take up. toward them and by the distinct social relations and practices that embody and express these attitudes. Finally, it recognizes a plurality of contestable ideals, by which we try to govern the development of our attitudes, character, values, and aspirations. In dividing goods into different kinds, I do not claim that for any one good there is just one mode of valuation appropriate to it. Inanimate nature is a proper object of both use and appreciation, as well as of awe and wonder; animals are proper objects of kindliness and even admiration, as well as of consideration and use. These modes of valuation are often incompatible. The pluralism of ideals and the relational character

A Pluralist Theory of Value • 15

of importance also imply that the ways one person should value a particular thing or person need not be the ways another person should value it or him. The respects in which anything is properly valued, and the ways and circumstances in which it makes sense to value it, remain problems. In introducing the notion that goods differ in kind, I suggest that these are the kinds of problems we should be posing ourselves, not that the answers are to be found in establishing a rigid classification of things into kinds. My socially grounded, ideal-based, pluralistic theory of value goes against the grain of a long philosophical tradition. Philosophy has traditionally expressed impatience with the pluralistic, contestable, historically contingent and socially informed evaluative practices in which ordinary people participate. Since Socrates, a^conmon^rnlosoghical aspiration has been to find some means of grasping the good or the rigliF directly, 13irune3i^ed"by"ffi"e"plurffistic hodgepodge of socially particular evaluative concepts and ideals (Plato 1961a). To reach sound ethical judgments, we are thought to require an entirely new mode of ethical justification, independent of the historical and social contingencies in which common* ^sense"'ev^uaidye'reasb1riingis mired. Many motivations support this aspiration: the determination to make value judgments unconditionally universal (Plato 1961a) or to represent them as subject only to purely personal intuition (Moore 1903); to overcome ethical disagreement (Bentham 1948; Plato 1961b); to find a determinate rational decision procedure in ethics (Bentham 1948; Brandt 1979; Hare 1981; Harsanyi 1982); to naturalistically reduce "values" to "facts" (Brandt 1979; Railton 1986); to enable critical reflection on our own practices (Brandt 1979; Hare 1981). The attempt to bypass the varieties of pluralism I affirm leads to a monistic or drastically reductionist theory of value. In emphasizing the intimate connections between the plurality of our evaluative attitudes and the plurality of our ideals, evaluative concepts, and social practices, I aim to highlight the problems involved in adopting such monistic and reductionist programs. If we bypass the plurality of values and ideals in attempting to get a direct grasp on what is good and right, we will lose the resources to make sense of our attitudes and even to have highly differentiated and nuanced attitudes. We could be reduced to expressing a crudely generic "pro-" or "con-" attitude. Monistic theories of value tend to overlook this problem, because they assume that value is normative for just one attitude or response, such as desire, mere liking, or being pleased. It is no accident that the moral psychologies of such monists rarely acknowledge the existence, much less the importance, of other attitudes besides their favored one (§§6.2—6.4). But if it makes sense

16 • A Pluralist Theory of Value

for us to have a variety of evaluative attitudes, we can't do without our commonsense pluralistic practices. Monism is inherently defective, because it cannot make sense of the phenomena of values and valuation that any theory of value must account for. Some of the following chapters will be devoted to elaborating this argument, considering monistic replies to it, and defending pluralism against monistic challenges. My larger ambition is to explore some of the practical implications of my socially grounded, pluralistic rational attitude theory of value. In the next three chapters, I will show how it supports an alternative to the dominant theories of rational choice. In the last three chapters, I will explore some of the political implications of pluralism. In providing an account of how economic goods differ in kind from other kinds of goods, pluralism sharpens our view of the ethical limitations of the market and helps us determine what goods should and should not be treated as commodities.

An Expressive Theory of Rational Action

2.1 Value and Rational Action A theory of value should help us rationally guide our actions. A rational attitude theory of value must solve two puzzles to show its relevance to rational action. First, it represen^s_^ue_a^immediately normative for our favorable attitudes, not for our choices. Ijts_associa^3^€ory of gractical reason must jtherc^ Second, my rational attitude theory locates intrinsic value in persons, animals, communities, and things, whereas action aims at the realization of states of affairs. Hence my theory must show how the intrinsic values of people and things are related to the values of states of affairs. The theory of rational action that I propose to solve these problems can be called an expressive theory. theory p y An expressive p ^^^££^ ^ our rational attitudes toward people and other to the ^ ^ H l H H H of value, something is valuable if and only if it is rational for someone to value it, to assume a favorable attitude toward it. And to adequately care about something requires that one express one's valuations in the world, to embody them in some social reality. This is a demand of self-understanding (Taylor 1979, p. 73). To fully make sense of one's rational concerns, one must be motivated to actually establish the relationship to the object of one s concern which is implicit in one's attitudes toward it. If this project leaves one unmoved and one does not suffer from weakness of will, weariness, or other motivational deficiency, one cannot sincerely ascribe to oneself the attitude it expresses. The rational requirement that attitudes seek their expression is confirmed, not undermined, by the thought that an expressive project may leave one deeply conflicted and ambivalent because one holds attitudes that require incompatible projects.

18 • An Expressive Theory of Rational Action

Practical reason demands that ones actions adequately express ones rational attitudes toward the people and things one cares about. Because expression is a meaningful activity, it requires a publicly intelligible vehicle to make its point. This is provided by the social norms that are constitutive of rational attitudes. To have an evaluative attitude toward something is in part to govern one's deliberations and actions by social norms that communicate distinctive meanings to others. By distinguishing the kind of kiss romantic lovers may exchange from that which "just friends" may exchange, social norms for kissing enable people to effectively communicate distinct attitudes toward others. Social norms typically tell us to direct our desires and actions to the realization or prevention of particular states of affairs. Norms for expressing charitable benevolence direct us to satisfy people's basic need for food, clothing, and education. Norms for expressing civility direct us to avoid embarrassing others. We acquire our rational aims partly by determining what the norms for adequately expressing our attitudes require, encourage, or make apt. Thus, a fully rational action expresses a way of valuing something in being governed by norms constitutive of that mode of valuation. In tying rational action to social norms, the expressive theory may appear to endorse a form of conventionalism. Conventionalism identifies appropriate action with action governed by whatever norms prevail in society. The expressive theory need not endorse extant social norms for expressing attitudes, however. A social order can be criticized for failing to provide adequate normative vehicles for the expression of attitudes that have come to make sense to its members. The social aspect of the expressive theory reflects not a conventionalist but an anti-individualist theory of rationality. It claims that individuals^ are not self^sufFicient bearers of practical reason: they require a context of social norms to express their, atti^"tucles^aclequately and intelligibly in action, to express them in ways others can grasp. - ff Tsociety lacks the social norms needed to adequately express its members' reflectively endorsed valuations, the rational thing to do is to invent and institute such norms. West European and North American societies lack adequate normative vehicles for expressing heterosexual affection on egalitarian terms, although many members of these societies seek to establish loving relationships on such terms. Norms for bodily contact between heterosexual lovers—for example, that the man may express his affection by wrapping his arm around his lover, or by leading her on the dance floor—also express a status hierarchy in which the man is the protector and leader, the woman the dependent follower (Tannen

An Expressive Theory of Rational Action • 19

1990, pp. 283—287). Until alternative norms for expressing heterosexual affection can be instituted, egalitarian couples will not be able to express fully and adequately the kind of love they have for each other. This gives them a reason to invent and institute norms adequate to their attitudes. The expressive theory of rationality distinguishes between two different sorts of ends for the sake of which we act. First, there are the people, animals, communities, and things toward which we direct our actions. These are the things it makes sense for a person to care immediately about, independent of its making sense for her to care about any other particular thing. Call these intrinsic.^oo^s^ Intrinsic^oods are the immediate objects of our intrinsic valuations (|1JJ7- Persons are the immediate oBjects of our respect, benevolence, and love; beautiful paintings of our admiring contemplation; pets of our affection; and so forth. These are the things we rationally value in themselves. Extrinsic goods^Jbyjcontrast, are goods which it makes^ensejfor a person to value onlyjbecause it makes sense for her to value some other particular thing. The value of an extrinsic good Sepencirupon tKe value of seTin that one's rational valuation of it is mediated by one's rational valuation of something else. Sharon may cherish an ugly bracelet because it was given to her by a dear friend. The bracelet, valued as a token of friendship, is an extrinsic good. Sharon's valuation of the bracelet is mediated by her valuation of her friend. Were they to become enemies, it would make sense for her to stop cherishing it. Two points should be noted about the definitions of intrinsic and extrinsic goods. The first is that the definitions are agent-centered: a good is extrinsically valuable if one's rational valuation of it depends on one's rational valuation of some other particular thing. It may be a condition of any given person's rationally valuing something in a particular way that other people also rationally value it. This is true for all impersonal valuations, such as respect, although not for personal valuations, such as love. The second is that the definitions refer to the valuation of particulars, not of universals. Universals—the standards for rational valuation—provide the grounds for our valuations, not their objects. A condition of any person's rationally valuing something in a particular way may be that it merit valuation by meeting certain general standards. This is also true for impersonal valuations. The only condition that makes a thing extrinsically valuable is that one's rationally valuing it depends on one's rationally valuing some other particular good. Kant's famous imperative to regard humanity as an end in itself expresses something like the first sense of "end" I have in mind when I say that

20 • An Expressive Theory of Rational Action

intrinsic goods are the ends for the sake of which we act. To take humanity as one's end is to act for the sake of or with due regard for the persons affected by one's actions. Such action involves not only promoting their welfare, but can also include such activities as participating in projects important to them or taking their opinions seriously in discussion. Kant, however, restricted the possible objects of unmediated rational concern to rational beings alone. The second kind of end for the sake of which we act is our final aims or goals, the states of affairs we seekxo bring about in our actions. These~en3s are contrasted with means, which are the actions and states of affairsTriair "are* rationally desired or chosen because they tend to bring aboufisome offier slates of affairs (our ultimate goals). It is important to distinguish between intrinsic goods and final aims (Korsgaard 1983). The distinction is often conflated by theories that contrast intrinsic with instrumental goods or that identify intrinsic goods with the states of affairs we rationally and ultimately aim to bring about. Although all instrumental goods are also extrinsic goods, there are some extrinsic goods, such as the ugly bracelet, which are not instrumentally good. According to the rational attitude theory of value, states of affairs, whether they be final aims or mere means, are for the most part only extrinsically valuable. It makes sense for a person to value most of them only because it makes sense for a person to care about the people, animals, communities, and things concerned with them.1 This follows from the fact that our basic evaluative attitudes—love, respect, consideration, affection, honor, and so forth—are non-propositional. They are attitudes we take up immediately toward persons, animals, and things, not toward facts. Because to be intrinsically valuable is to be the immediate object of such a rational attitude, states of affairs are not intrinsically valuable if they are not immediate objects of such attitudes. Evaluative attitudes take up states of affairs as their mediated objects through the desires, hopes, wishes, and other propositional attitudes that express them. Jack's love for Margaret can be expressed in the hope that he will be able to see her soon. His favorable attitude toward her is what makes sense of his favorable attitude toward the state of affairs in which he sees her soon. Margaret is the immediate object of his love. The states of affairs he desires, hopes, or aims at are the mediated object of his love. They are mediated by norms for desire, hope, and intention that express his evaluative attitudes. I do not claim that people actually value states of affairs only because they value the people, animals, or things involved in them. Many of our

An Expressive Theory of Rational Action * 21

motivational states, such as appetites, whims, habits, compulsions, and addictions, can express a concern for the realization of states of affairs without any regard for ourselves, other people, or other things. When we care about states of affairs in these ways, the fact of our caring does not depend upon our caring about anything else. However, these motivations also do not generally depend or change upon reflection on their objects or on our own attitudes and reasons for action. This is why these motivational states are neither rational nor irrational. They are not the motivational states by which we rationally or reflectively govern ourselves. Because what is intrinsically valuable is the object of a rational favorable attitude, not just the object of any favorable attitude, the fact that we have favorable attitudes such as appetites and whims toward states of affairs does not show that these states of affairs are intrinsically valuable. Although appetites and similar motivations are arational, it can be rational or at least not irrational to act on them. It is rational to do so when they fulfill the aims that would anyway be given to us by our rational valuations of people and things. It is not irrational to act on these motivations, provided that our acting on them does not violate the expressive norms constitutive of our rational attitudes. Consider a person with a gluttonous appetite, who is motivated to eat without any regard for himself. It is rational for him to indulge his appetite to the extent that this promotes aims that are rationally related to his self-concern—for example, to the extent that it promotes his health or pleasure, or to concern with others, as when eating realizes communal relations among people. It is irrational for him to indulge his appetite to the destruction of his health, pleasure, or relationships with others, supposing it makes sense for him to care about these. Between these two extremes, he has considerable scope for indulging his appetites, an activity neither rationally required nor prohibited, but simply permitted. Raw appetites and similar motivational states do not express rational attitudes toward people and things because they are not the kinds of motives governed by reflection and meaningful social norms. Their expression is not mediated by norms but is at most constrained by them. We can say then that we act with full rationality when we govern ourselves by objectively valid expressive norms constitutive of our rational valua^ "tions^—that is, when we adequately express our rational valuations. And we act in accordance with reason when we act on other motives within the constraints posed by the norms that adequately embody our rational attitudes.

22 * An Expressive Theory of Rational Action

2.2 The Framing of Decisions The fundamental task of a theory of rational choice is to select, from among the many actions a person could perform, the action which it makes most sense to perform, or at least some action which makes sense. The theory must generate and rationally ground a ranking of actions from which the top ranked option, or at least some option above a threshold of appropriateness, is to be chosen. The dominant view of rational action, which I call consequentialist (§2.4), characterizes the end of rational action as the realization of valuable states of affairs. On this view, it is difficult to locate any other basis for ranking actions except the value-rankings of the states of affairs they tend to bring about. Rational action maximizes the value of states of affairs. The expressive conception of action specifies fully rational action as characteristically having dual ends: it seeks to bring about states of affairs for the sake of the people and things we rationally care about. In acting rationally, we generally express our rational valuations of people and things by pursuing particular states of affairs. In exposing the incompleteness of the consequentialist conception of action, the expressive theory opens up an alternative basis for ranking actions besides the value of their consequences. Actions are ranked according to how well they express our rational valuations, and this is determined by judging how well our actions live up to the norms constitutive of these valuations. This conclusion may seem puzzling. While there are some types of action, such as those expressing civility, for which our deliberations are preoccupied with living up to social norms, there are others, such as those expressing benevolence, in which the consequences of actions appear to be decisive in ranking alternative courses of action, quite independent of social norms. This thought has prompted some theorists to postulate two radically distinct types of action: one rationally oriented toward bringing about consequences, the other (non-rationally) oriented toward obeying social norms (Weber 1968; Elster 1989). The expressive theory rejects this false dichotomy. Rational action is characteristically oriented in both ways. Where consideration of consequences alone appears to be relevant to justifying an action, closer examination reveals that background expressive norms implicitly set a context of response and decision which makes some consequences of action more important than others and which determines how they will be incorporated in deliberation. Where benevolence is the primary way we value the people for the sake of whom we act, it makes sense to follow norms that

An Expressive Theory of Rational Action • 23

bring welfare considerations to the foreground and ignore many nonwelfare consequences of action. Emergency aid workers do not worry about the aesthetic consequences of setting up refugee camps where they do. And no coherent form of benevolence is generically oriented toward all the welfare consequences of action. Different social norms govern the expression of distinct benevolent attitudes appropriate to persons who stand in different relations to their beneficiaries. These norms identify different welfare consequences of action as important for choice. So, philanthropists provide institutionally given food, shelter, and education to the needy. Friends offer their sympathy, companionship, personal effort, and advice. Parents involve themselves in promoting their children's welfare in ways which are none of a benevolent acquaintance s business. Thus, our preoccupation with the consequences of action in much of our deliberation reflects not the irrelevance of social norms to rationally ranking actions, but rather the fact that to express our concern for what we intrinsically value we must generally follow social norms that direct our attention to consequences concerning them. These norms are often embodied in unreflective habits and become objects of deliberation only when their expressive significance is called into question. Expressive norms typically tell us to pay attention to particular consequences of, action described in terms of particular evaluative distinctions and to incorporate these consequences into our deliberations in a particular way. They select from all the authentic candidates for rational valuation those states of affairs which are important to the agent at this time and place. A state of affairs becomes important to evaluating action when the conditions for its having extrinsic value are satisfied—that is, when pursuing it would express ones rational valuations of persons and things. Call th£rwa^s«a«^^^ relevant options and her conceptiorTof what is at stake in her choice^. £er "decision^ ). The norms for expressing a persons vaTuiaHons fundamentally shape the decision frame she uses to ground rankings of her actions. She solves the problem of deciding what frame to use in deliberation when she successfully reaches an interpretation of her predicament that enables her coherently to continue her life. This task amounts to a continuation of the project of rational self-understanding (§§1.1, 5.1). I believe that this project issues in two global norms for making sense of one's actions: one synchronic, the other diachronic. The synchronic norm tells a person at any given time to act in such a way as to adequately express the ways she rationally values all the persons and things for whose sake she should act. This norm tells a person to

24 • An Expressive Theory of Rational Action

appropriately distribute her concern among the different persons and things she properly cares about in acting. The diachronic norm tells a person to act in such a way that over time her actions can be fit into a coherent narrative (Maclntyre 1981, ch. 15; Velleman 1991). The demands of this norm have only recently begun to be explored. Thus, I will suggest only that the coherence of a narrative of one's life will depend .upon an account of how our valuations and evaluations can rationally evolve and develop in the light of new experience (Anderson 1991). Both of these global norms are regulative ideals, which can rarely, if ever, be completely satisfied. Many conflicts arise when we cannot satisfy the demands of caring about one person without violating the demands of caring about another, or when we no longer have the context or resources to coherently continue our lives. I have argued that only in the context of a decision frame do particular consequences of actions emerge as relevant for evaluating action. This is because the consequences of action generally have no intrinsic value. Their importance emerges only in a setting in which an agent's rational attitudes toward people and things are interpreted through a decision frame. But what determines the rational choice of a decision frame? Ideals that embody conceptions of how goods differ in kind play an indispensable role here. Recall that the plurality of goods arises from the fact that people care about different goods in different ways, care about the ways they care about goods, and institutionalize different ways of caring about goods by embedding them in distinct social practices of production, distribution, and enjoyment. These social practices are governed by norms that highlight some features of the goods in question as important for action concerning them and subordinate others. In classifying a good as one kind or another, by embedding it in one set of social practices over another, people select the relevant decision frames which will be applied to it. For example, to classify dogs as pets is to call for decision frames regarding dogs as proper objects of affection and domestication and to rule out decision frames which consider their edibility, or their potential life in the wild, as relevant to choices concerning them. To adopt the ideal of being an outstanding defense lawyer in an adversary system of justice is to call for decision frames that reject the justice of punishing ones guilty clients as a consideration important to preparing a defense before trial. Thus, a fundamental implication of the thesis that goods differ in kind is that people should deliberate about them in different ways, according to different frames. In determining which frame a person should use to describe the

An Expressive Theory of Rational Action • 25

optiojisjitjiand, she consults how she cares about the_2eogle_concerned ^wjjflijJie_^r2donsl_her ideals of how she ought to care about them, ancTTrTe social roles she occupies that embody these ideals. Teople tend^to take decision frames for granted because they are often e"mBodied injiabits and social roles. SociaTroTe differentiation, in enabling peope to occupy different roles at different times and places, enables them to establish different priorities in different parts of their lives. The "same" action described in terms of its consequences can have a different expressive significance, and hence a different degree of appropriateness, depending on the social context in which it is performed. When a parent sets aside his child's demands for attention in order to deal with a client's needs, it typically makes a big difference for the expressive meaning, and therefore the appropriateness of the act, whether the parent should be acting in his role as parent or in his role as businessperson at that time, and this in turn typically but not always depends upon whether he is at home or at work. In the former case, the act is more likely to express an inappropriate neglect or indifference toward his child than in the latter. What things a person cares about, as well as how and how much she cares about them, are not solely a function of the social practices, roles, and relationships she participates in. Her character, history, mood, energy, actions, and reading of her predicament play a profound part in influencing what she values, especially in influencing which practices, roles, and relationships she will make her own, how she will interpret, criticize, and change them, and so forth. In emphasizing how a person's ways of valuing things are structured through social roles, practices, and relationships, I do not want to imply that these structures are to be regarded as simply given to agents, unmediated by their own understandings, or beyond critical scrutiny. Although decision frames embedded in social roles are frequently taken for granted, it is often important to make them an object of deliberation. Do the norms constitutive of these frames adequately express the ways we should value the persons whose interests are at stake in the choices they guide? The parent/worker example presented above offers material for deliberation about appropriate decision frames, since society does not structure social roles and decision frames suitable for parents of either gender. It assigns different meanings to mothers and fathers making the same tradeoffs of work and parental responsibilities, which express such views as that children need to be with mothers more than fathers, that mothers and fathers should value their children differently, and that paid work is more important to fathers than to mothers. As these judgments, along with the gender hierarchies they

26 • An Expressive Theory of Rational Action

help sustain, come to make less sense to people, new social roles and decision frames must be devised. 2.3 The Extrinsic Value of States of Affairs I have argued that states of affairs are generally only extrinsically valuable, because our intrinsic evaluative attitudes do not generally take them as their immediate objects. It makes sense for a person to value most states of affairs only because it makes sense for him to value people, animals, and other things. This claim may seem counterintuitive. Because its implications for practical reason are dramatic, it is worth exploring in greater depth. Reflection on a few examples should convince one of its truth. All states of affairs that consist in someone's welfare are only extrinsically valuable. If it doesn't make sense to value tHe"persqn (in a particular way), then it doesn't make sense to carelibout promoting her welfare (in the way that expresses that mode of valuation). Enemies, who hate each other, have no reason to promote each other s welfare. Mary may rationally feel self-contempt for betraying her profession as a journalist. (Perhaps she published a story she knew to be false, as a favor to a government official.) Under this condition of self-disvaluation, it doesn't make sense for her to seek her own advancement in it until she has made amends, for she regards her advancement as undeserved and, hence, unworthy of pursuit. Some believers in the intrinsic values of states of affairs agree that welfare is not intrinsically valuable (Moore 1903; Regan 1989). They find intrinsic value in such states of affairs as knowledge and the existence of art. But states of affairs which consist in the existence of something are valuable only if it makes sense to care about the thing that exists. It doesn't make sense to care about the existence of a painting unless it makes sense to care about the painting itself, perhaps because it is beautiful. And beauty is a valuable attribute of the painting, not of the fact that the painting exists. One may suppose that it doesn't makes sense to care about something unless it makes sense to care about its existence. This would suggest a mutual dependence of the values of a thing and the value of its existence and, in this case, the collapse of the intrinsic/extrinsic value distinction. But the supposition is not true. It may make sense for me to love a person, but this does not imply that I must want that person to continue living. If he is gravely ill, it may be the best expression of my love for him to wish that he die quickly and mercifully. A remarried widow may still love her long-dead husband, but be appalled if he were to pop back into existence.

An Expressive Theory of Rational Action • 21

Knowledge is not intrinsically valuable, for it makes sense to care about knowing something only if the object of knowledge is interesting or important or the knowledge itself is useful. It doesn't make sense to care intrinsically about knowing a boring subject. Can bare facts be intrinsically interesting? I am inclined to say that facts are interesting only if the things they are about are interesting. But I won't press my case here. Perhaps some things are interesting because we have discovered interesting facts about them. Interest does seem to be an evaluative attitude that can take a state of affairs as its immediate and independent object. This is an exception to the general rule that states of affairs have no intrinsic value. Infamous utilitarian population paradoxes follow from the thought that welfare states have intrinsic value (Parfit 1984, ch. 17). Utilitarianism identifies the morally best state of affairs as that in which total or average welfare is maximized, and it defines the right act (rule, motive, and so forth) as that which tends to bring about the morally best state of affairs. On the total utilitarian view, we should increase the population until each individual is so burdened from overcrowding that an additional birth takes away more welfare from others than the newborn could be expected to enjoy. On the average utilitarian view, we should do away with people of below-average happiness, provided that we can do so secretly and without hurting others. There are, of course, ways utilitarians can dodge these paradoxes, which depend on contingent facts making these recommendations seem not so bad or very unlikely to maximize welfare. But they don't explain the root of the problem. Who is better off under the utilitarian population policies? No one need be better off in the average utility case. The newborn may be the only person "better off" in the total utility case. But even the rationale for bringing the newborn into the world seems to get things backward. To pile up people so that more "welfare" can exist, or to get rid of them so that a higher average level of welfare can exist, is to regard people as merely the extrinsically valuable containers for what is supposedly intrinsically valuable—states of affairs in which welfare exists (Sen and Williams 1982, p. 4). The mistake in both cases is to lose sight of the fact that what gives the pursuit of or desire for welfare its only point is that we ought to care about the people who enjoy it. Utilitarians can take either of two paths of response to this. The first is fo accept the dual-ended logic of attitudes and to concede that welfare has /only extrinsic value. It is not good in itself, but good only for persons. This path is taken by theorists who derive utilitarianism from an ideal observer theory of morality (Firth 1952). They identify the moral point of

28 * An Expressive Theory of Rational Action

view with the point of view of a perfectly impartial, benevolent, omniscient observer. The state of maximum welfare gets its moral value from the fact that such an impartially benevolent observer desires it. This rationale for utilitarianism avoids the mistake of thinking that people are just the extrinsically valuable containers for welfare. Impartial concern for people mediates between the ideal observer's desires and the value ascribed to maximizing welfare. Such concern could not endorse the average utilitarian population policy because benevolence would not approve of a policy that makes no person better off. Although it would endorse the total utilitarian population policy, it would do so because here at least the newborn is better off, and benevolence would approve of that. The cogency of this response depends on the claims that impartial benevolence toward people issues in a desire that welfare be maximized and that impartiality makes one indifferent to which people are enjoying it. Whereas concern for people mediates between the moral point of view and the conferral of extrinsic value upon the maximum welfare, impartiality could be said to make transparent its mediating role. The impartially benevolent observer sees right through people to the welfare states they enjoy, which guide desire immediately, without being modified by any regard for the particular people that mediate between them and the moral point of view. This position does not offer a plausible interpretation of impartial benevolence. Once the dual-ended logic of evaluative attitudes is accepted, then so should be its constitutive expressive norms. Benevolence is not just a desire that welfare exist: it is a concerned attitude toward people. It does not just give us states of affairs to desire or achieve. It gives us norms by which to pursue it, by which we express our concern toward the people for whose sake we act. Among these are norms for distributing our care among persons. If benevolence is to be impartial, it must be expressed toward each person. The total utilitarian population policy, in gravely burdening those other than the newborn, does not seem to express benevolence toward them. Benevolence is also especially concerned with the people who are most needy and is not satisfied when welfare is taken from them to gratify those who already enjoy much or when the number of needy people or the severity of their neediness increases. (Neediness can increase simultaneously with total welfare.) This concern for the needy is not a violation of impartiality, which demands only that one abstract from one s personal likings, interests, and biases in expressing ones valuations of persons. Neediness is an impartial basis for determining where benevolence is most called for. Benevolence is also

An Expressive Theory of Rational Action • 29

constrained by norms expressing the other ways we ought to care about people (Foot 1985). Actions that promote welfare ruthlessly, manipulatively, or unjustly are not benevolent. Actions that maximize welfare by exploiting some so that others may benefit do not express benevolence toward those exploited. If the dual-ended logic of attitudes is accepted, then impartiality toward persons cannot be reduced to indifference over which equal-sized instance of welfare occurs. Hence it cannot be expressed in a desire that welfare be maximized, without regard for the distribution of welfare across persons or for how this distribution is brought about. The mediating function of concern for people can never be made transparent: the appropriateness of desires for states of affairs must be subject to the constraint that they adequately express their correlative attitudes toward people. If the concern we ought to show for people causes such trouble for utilitarianism, perhaps people should be removed as distinct objects of concern. This is the second path of utilitarian response to the population paradoxes. Derek Parfit (1984, ch. 11) denies that persons have any distinct existence beyond the bundles of mental states they contain. If this were the case, it would not make sense to take persons as distinct objects of evaluative attitudes, which should take only states of affairs as their end. This argument supposes that our psychologies can be exhaustively analyzed in terms of phenomenal states and propositional attitudes. But most evaluative attitudes are essentially non-propositional: they essentially take persons and other things as their objects. We cannot make practical sense of ourselves without grasping ourselves as beings who take up attitudes such as love, hate, respect, and contempt toward other people. I do not believe that we are making any metaphysical mistake in doing so. But if we are, the mistake is one we have to live with, because we cannot get rid of our non-propositional attitudes by reflecting on the metaphysical mistake they supposedly embody. We cannot act as if we don't have these attitudes, nor could we make reflective sense of ourselves if we did. Practical reason, necessarily taking the logical structure of our evaluative attitudes as given, must postulate the existence of persons to make sense of them. If this contradicts the deliverances of theoretical metaphysics, this shows only that theoretical metaphysics has no practical import (Kant 1981, sec. 3). The preceding analysis of the population paradoxes suggests the following conclusions about practical reason. Firstjractical reason cannot get by without employing the logic of evaluative attitujje^^whlcK take £ e r s o n s and other things as their imrnediate^objeciL Second, because states of affairs do not generally hav^ intrimir VQ1^PI o^r attitudes toward them

30 * An Expressive Theory of Rational Action

must be conditioned by our attitudes toward the things that do have intrinsic value. This means that the rational thing to do cannot be defined simply in terms of the pursuit of states of affairs. The rationality of pursuing states of affairs is mediated by expressive norms. The ground for pursuing them is that in doing so, we express our appropriate regard for people and other things. Third, the content of expressive norms is not given simply by specifying states of affairs to be preferred. No adequate interpretation of a way of valuing something can reduce its motivational component to a desire or preference that some states of affairs occur. They must be brought about in the right ways, by the right agents, in the right context (Aristotle 1985, 1106b20—24). Whether desiring, aiming at, or achieving a given state of affairs adequately expresses the right attitudes toward people and things depends on the context that determines its expressive meaning. The argument that states of affairs generally lack intrinsic value depends upon two types of consideration. One is the survey of our intuitions about what makes sense, which has occupied this section. The other is the rational attitude theory of value^, which identifies^jjiejcandidates for tive attitudes. Many theorists hold that intrinsic values are signified by propositional attitudes such as desire. I will argue in §6.4 that the only value-signifying desires are those that express rational non-propositional evaluative attitudes. If this is true, then the only value they can signify or confer upon states of affairs is extrinsic. The claim that states of affairs do not generally have intrinsic value has enormous implications for the theory of rational choice, for the dominant theories of rational choice suppose that they do have intrinsic value. I call these theories "consequentialist." The confrontation between expressive and consequentialist theories of rational choice will occupy the remainder of this and the next two chapters. 2.4 Consequentialism I call any theory of rational or moral action consequentialist if it meets the following three conditions: First, it gives people the sole ultimate airn^of maxjinmng^toaa^yjlue. Second, it holds the fundamental object of intrinsic value to be the state of affairs. It assesses the value of a state of ajjairs independent of the values of persons, actions, motives, norms, practices, states of character, or anything else. Third, it assesses the values of these other sorts of things, or at least actions, rules, or practices, solely in terms of

An Expressive Theory of Rational Action • 31

their consequences, broadly construed—that is, in terms of how effectively they bring about or embody the best states of affairs. My use of the term "consequentialism" is more expansive than most people's. In my usage, egoism as a theory of rationality, which gives each agent the sole ultimate aim of maximizing his own personal welfare, is a consequentialist theory. In most philosophical writing, "consequentialism" refers only to a subset of the theories fitting my definition: only moral theories that give all agents the common aim of maximizing impersonal value. Such theories add to the above three conditions a fourth: that all values are "agent-neutral." Avalue is agent-neutral if it gives everyone a reason tojralue it. A value is agent-relative if it gives only some people a reason to value it, or different agents reasons to value it in different ways (compare Sen 1982; Nagel 1986). Together, these conditions state an agent-neutral principle of morality—a principle that gives all agents a common moral aim, the maximization of agent-neutral value (Parfit 1984, P- 27). I reserve the term "consequentialist" to fit theories meeting just the first three conditions because they share a common conception of the relation of action to states of affairs. They_jdlmaintain that statesof affair^ai^, immediately normative for actions or for other practical responses such as jrules, choices, preferences, or desires. The relevance of the states of affairs produced by action to the evaluation of action is not mediated by norms that express the worth of persons, motives, or anything else. Alternatively, the mediating norms are transparent, as in the case of the utilitarian interpretation of impartial benevolence: the norms themselves unconditionally specify some states of affairs, identifiable independent of expressive contexts, to be desired, chosen, preferred, or promoted. In either case, a given state of affairs is always relevant in the same way to its correlative response, regardless of its context. For example, the standard decision theoretic account of rational action begins with a preference ranking of states of affairs and generates a ranking of actions according to their expected utility—the sum of the products of their possible outcomes with the probabilities that they will bring about those outcomes. The act to be chosen is that which maximizes expected utility (Jeffery 1965). This immediate and unconditional relevance of consequences to choices in consequentialist theories contrasts with the expressive theory. In the expressive theory, expressive norms determine whether and in what way a given state, of affairs is normative for practical responses. The same state of affairs may have a different relevance, depending on the decision frame that applies to a persons responses.

i

32 * An Expressive Theory of Rational Action

My definition of consequentialism may be thought to be too narrow. It appears to fit Moore's brand of consequentialism best, since Moore explicitly formulated his theory in terms of the intrinsic value of states of affairs. But some theories I wish to include in my account don't take an explicit stand on the nature of value or what has it. The economic theory of rational choice says that the rational thing to clo isTo~actasl£one were Ft leaves to the metaphysicians to decide whether utility as definecTlnTKe behavioristic manner of decision theory is really an intrinsically good thing. And decision theory, like Moore's theory, rejects the account of intrinsic value I offer, which ties it essentially to favorable attitudes. That people ought to act as if they are maximizing expected utility does not demand that they actually value this end or desire to bring it about. My definition of consequentialism requires a suitably theory-neutral definition of intrinsic value for it to cast its net as widely as I would like. The following minimalist account of intrinsic value seems to capture the common features of the theories I call consequentialist. To ascribe value to something is to make it normative for some response in us, such as attitudes, value judgments, emotions, desires, actions, preferences, rules of action, or motives. These things ought to be favorably responsive to that to which value is ascribed. To ascribe intrinsic value to somethingis to make it immediately normative for something else; that is, its normativity is not conditioned on the normativity of some other thing. The theoriesj^ call consequentialist all make states of affairs immediately normative JCQX actions, rules of action, preferences, desires, or motives. They all identify, independent of the expressive context in which they appear, certain states of affairs as unconditionally to be chosen, desired, preferred, or brought about, such as the maximum welfare, maximum pleasure, maximum desire- or preference-satisfaction (in each of these cases, either for oneself or for everyone), or the maximum intrinsic value in some thicker sense, as in Moore's theory. (In §4.4 I discuss a hybrid form of consequentialism that makes the values of some states of affairs dependent on their expressive context.) Consequentialist and expressive theories of rationality pose sharply contrasting ways of thinking about value and action. Consequentialist theories make intrinsic value immediately normative for desire, preference, choice, or rules of action. Expressive theories make it immediately normatiyejor evaluative attitudes and derive desires, preferences, and choices from norms for expressing attitudes. Consequentialist theories conceive of rational action as directed toward one end: the production of conse-

An Expressive Theory of Rational Action • 33

quences. & 4 m ^ £ i y £ ^ action as^directed toward twojends: a final end—the production of consequences—and an end for the sakejpf which£he fmal end is sought—the people, animals, and things the agent cares about. Cojis^c|u^ntmlisttheories evaluate action in terms J theories evaluate action in terms of its . gyaluate the: relevance of consequences through * expressive norms. Consequentialist theories justify action by showing that it maximizes value. Expressive theories justify action by showing that it is normatively appropriate, that it conforms to the expressive norms constitutive of a person s rational valuations. Consequentialist theories recognize just one norm for action—that it maximize intrinsic value. Expressive theories recognize a wide variety of norms, which have several features that are puzzling from a consequentialist point of view: they are intentional, backward-looking, distributive, and non-instrumental. Each of these features of expressive norms merits a closer investigation. Expressive norms are intentional. They tell people to intend or aim at certain things. The basic form of an expressive norm is: act so as to adequately express attitude B toward Z. Although it is possible for a person to express attitudes inadvertently, or by mistake, as when one unintentionally insults someone, to adequately express one's attitudes requires that one intend their expression in some way This means not that people must always have their minds on what they do, but rather that they must be prepared to acknowledge their actions as expressing their attitudes. Consequentialist norms, by contrast, simply tell people to achieve certain consequences, whetheir they intend them' or riot. of a. cqnsequejntiajis^^^ so as tc^bring about consequericeJK, Jkis_ possible that the best way to bring about X is to aim at Qjr u thing else. The, famous "paradox of hedonism" maintains that the best way '"to achieve happiness is not to try to become happy, but to devote oneself to distinct ends, success in which will bring happiness as an unintended consequence (Sidgwick 1981, p. 136). And the best way to devote oneself to distinct ends may be to care about them independent of happiness, perhaps not to care about one's happiness at all. Consequentialism may give people an end whose achievement requires that they not identify with it. The intentional character of expressive norms supports fundamental distinctions, deeply rooted in commonsense morality, that have no evident rationale in consequentialist terms. They include the distinctions between doing something and allowing it to happen, between foreseeing and intending certain consequences, between the consequences you and I are

34 • An Expressive Theory of Rational Action

responsible for when either of us is in a position to cause or prevent them. These concerais_ajl^ mark distinctions in the. expressive significance, of ^cHonTalid, hence, b e ^ u p o n t h e i r appropriatenejss.j&di±iin theterins^f expressive tneory DeEBerately tripping someone is worse than accidentally doing so, because the former expresses contempt or hatred for someone, whereas the latter does not. Committing adultery with a stranger's husband is expressively worse than failing to prevent him from committing adultery with someone else. In the latter case, one may properly be respecting others' privacy, including the victim's, but committing adultery cannot express respect for the victim. Yet these distinctions do not track differences in the consequences of the actions or happenings they are applied to: in either case, a tripping or a betrayal has occurred. (Granted, people are more likely to take offense at a deliberate tripper or the adulterer, but it isn't evident what consequential are to make of this. For the grounds for offense presuppose an expressive logic not itself endorsed by consequential. The world may be better off if people didn't respond differently to actions with the same consequences.) Because the concepts marking these distinctions do not track differences in consequences, they cannot be fundamental to a consequentialist theory. At best, their use in guiding moral practice could be justified indirectly, as leading to better consequences overall. Ex£ressjvejiorms are backward-looking: what it makes sense to do now essentially depends on what one has done in the past. This follows from &e requirement of narrative unity (§2.2). The past sets a context that confers expressive meaning on present choices. Had the past been different, the same present act could have a different meaning and therefore a different appropriateness. Consider a couple who struggle for years through long workdays and financial difficulties to establish a distinctive family restaurant.' Now, just as the restaurant is at the threshold of steady success, a franchise operation wants to buy it from them and build dozens of similar restaurants around the country. In return for relinquishing control of the restaurant, they would get far more money than they could by continuing to operate it. The couple might think of their choice as follows: Selling the restaurant would offer them important financial security, but it would also undermine the point of their lives' personal investments and struggles, which were aimed not just at making money but at creating an alternative to the humdrum, homogenized, and predictable chain restaurants taking over the area. Dropping their life projects for this reason would leave them with life stories as "successful" sell-outs, rather than as people who had made something of their early struggles and

An Expressive Theory of Rational Action • 35

fulfilled the dream of a lifetime. They did not work all those years to make millions for some brand-x corporation. A concern for the narrative unity of their lives, for what meaning their present choices make of their past actions, could rationally motivate them to turn down the offer. Consequentialist theories reject thisJbnd of reasoning. The point of action^j^jta-jziaximize.,value, and action realizes value only through its niture consequences. Consequentialists view the couple's reasoning as irrationally weighing "sunk costs" in their calculations. If a greater amount of future good can be achieved by taking up an entirely new path than it could by sticking with one's past investments and personal commitments, one should disregard the past and take the option with the greater future payoff. This does not mean that people should weigh only crude material gains in their calculations. If having alternatives to mass-produced commercialism in restaurants is a good thing, perhaps because it gives the town a certain charm, this consideration counts in a calculation of future good. The consequentialist point is that it should count no differently for a couple who had devoted their lives to promoting it than for a couple who judge that it is as good as the devotees say, but who had just come into the restaurant by inheritance. The meanings a choice confers on one's past actions are irrelevant to the future payoffs, which are by hypothesis the same in either case. Another distinctive feature of expressive norms is their distributive structure. They tell people to express their concerns toward each person, animal, or thing for the sake of which they act. Concern is something people distribute to each of their ends. It is not an attitude that is held only toward some aggregate. Consequentialist norms, by contrast, have an aggregative structure: they tell people to maximize something. As the utilitarian population paradoxes show, the distributive structure of expressive norms typically imposes constraints on a person's maximizing behavior, since not every way of maximizing something expresses the appropriate respect, benevolence, or other attitudes we owe to each person affected by our actions. The distributive character of expressive norms explains some phenomena that are thought to be puzzling from a consequentialist point of view. It explains why it can make sense to feel ambivalence, regret, or guilt over a decision we rightly judge to be the best, all things considered (Williams 1981a; Stocker 1990, ch. 4). The expressive view calls for mixed feelings toward an action whenever the diverse ends for the sake of which we act are not each properly and adequately served by it. For example, one may feel both satisfaction and guilt in the same action, if this action both

36 * An Expressive Theory of Rational Action

adequately expresses one's love for one's family and neglects one's business clients. Became action is characteristi£ajly_j^iTrOTQe^jwith many ends at once, it can be evaluated For its expressive adequacy toward each end. Each evaluation of the act calls fora different emotional response, leading to mixed emotions toward the act overall. If no available alternative adequately expresses one's attitudes toward each person whose interests are at stake in a decision, then even the best alternative will merit regret. And if the attitudes are obligatory, even the best alternative will merit guilt. William Styron's Sophie, forced to choose one child to turn over to certain death at the Nazis' hands as a condition of being able to save any of her children, had no alternative by which she could properly express her love for the child she chose to relinquish. However justified her choice was under grave coercion, it would have been absurd for her to have viewed it with satisfaction or righteousness rather than anguish. (And the anguish is over her choice, not just over its unavoidably disastrous consequences.) There is no guarantee that circumstances will be such that we can satisfy the demands of all the expressive norms binding us and, therefore, no guarantee that we will always be able to act in ways that are above reproach. The world is not that cooperative (Stocker 1990, ch. 4). The puzzle over cases like Sophie s arises from two thoughts tempting to consequentialists. One is that all evaluations are inherently actionguiding. Regret, then, would be appropriate only if one judged that one shouldn't have performed the act. The second thought is that the appropriate consequentialist attitude to take toward a value-maximizing choice is satisfaction, righteousness, or some other unified, unmixed pro-attitude. If one really did one's best, and things turned out as best they could under the circumstances, there is no ground for regret, because the components of a state of affairs produced by an action have no distinctive significance. They are significant only as enlarging or diminishing an aggregate. Consequentialists are not committed to this reasoning, for it employs the expressivist notion of a state of affairs giving us grounds for or meriting regret. A pure consequentialist theory has no room for the idea of an emotion or attitude being merited by the intrinsic characteristics of a situation. Emotions should be evaluated in terms of their causal consequences. A person s regret for a consequentially justified action may be instrumentally valuable in deterring her from overlooking important consequences of her actions that could be rightly action-guiding in the future. If this were so, then a consequentialist could endorse a person's feelings of regret for having committed a justified act that brought harm

An Expressive Theory of Rational Action • 37

to someone. The real difficulty for consequentialism is that it cannot make sense of the reasoning that leads the agent to feel regret, since this reasoning makes use of an expressive logic not accessible to consequentialist reasoning. This logic, unlike consequentialist logic, does leave room for evaluations, such as that an act is regrettable, which are not actionguiding but rather immediately emotion-guiding (Stocker 1990, ch. 4). Consequentialists argue that if it is a good thing for the agent to feel regret, but she can feel it only by thinking in terms of expressive rationality, then she should believe the expressive theory even though it is false. A final distinctive feature of expressive norms is that they are noninstrumental. They are justified not by reference to any independent value that the consequences they recommend are thought to have, but by reference to the rational attitudes they express. In fact, expressive norms determine whether and in what way a state of affairs has any value or importance for an agent in the context at hand. Because different contexts alter the meaning of pursuing the same state of affairs, expressive norms give agents different final ends in different contexts. In consequentialist theories, by contrast, agents are given just one final end—the maximization of value. The fact that non-instrumental norms give agents different final ends in different contexts explains the "silencing" of considerations in virtuous people (McDowell 1978, p. 26; Wiggins 1980, pp. 234-235). Consider the captain of a sinking ship, who must coordinate the efforts of the crew to get people into lifeboats. It could be a fact about this captain that slipping off to his quarters to catch a nip of sherry would be a great relief and pleasure for him in these stressful hours. If the captain is virtuous, this fact will not even enter his mind as a potentially relevant consideration bearing upon what he should do now. The consideration is "silenced." This is not just a psychological fact about virtuous people; it reflects the point of view of virtue itself, which regards the consideration as strictly irrelevant. The expressive theory has an easy explanation for this. In the context of a sinking ship, the fact that the captain could get pleasure and relief from drinking really is irrelevant to what he should do. The relevant decision frame bearing upon the captain's actions accords it no value at all, for in this context there is no way his self-indulgence could mean anything other than a gross failure to respect his passengers. From a consequentialist point of view, the captain's pleasure and relief are intrinsically valuable and carry some small but positive weight in favor of his slipping off. If he should not consider this fact relevant, it is not because it is really irrelevant,

38 * An Expressive Theory of Rational Action

but because thinking about it would likely have worse consequences than not, perhaps by distracting him from the duties at hand. These differences reflect the fact that consequentialists recognize only one frame for justifying actions, whereas expressivists recognize different frames for different contexts. Consequentialists assert that tHere is'One, fixed, canonical description of states of affairs under which they are uniformly relevant for assessing all actions. They disagree over the relevant terms of description—are states of affairs relevant qua pleasures, qua satisfactions of desire or preference, or qua instances of intrinsic values such as knowledge, beauty, and friendship? But they all agree that there is some noncontextual value inherent in states of affairs which is given independent of interpretations of expressive meanings. Expressive theories deny that states of affairs generally have such intrinsic value. Apart from a deliberative context structured by expressive norms, states of affairs have no value or immediate action-guiding relevance. Their value is derived from the values of the people, animals, and things for the sake of which people act. Because states of affairs have only a context-dependent extrinsic value, it doesn't make sense to globally maximize the value of states of affairs. This is as incoherent as trying to globally maximize the instrumental value of tools, apart from the contexts which give them any usefulness. The latter incoherence is just an instance of the former, since instrumental value is one kind of extrinsic value. Thus, expressive theories of rationality contradict all three tenets of consequentialist theories. 2.5 Practical Reason and the Unity of the Self How are we to assess these disagreements between consequentialist and expressive theories of practical reason? If my rational attitude theory of value is right, then consequentialism is a nonstarter, because its criterion of rational action—that which globally maximizes the value of states of affairs—is incoherent. Extrinsic value can't be coherently maximized on a global scale. But this argument is not conclusive, since consequentialists naturally propose theories of intrinsic value incompatible with mine. These will be critically examined in Chapter 6. Another way to assess the relative merits of the two theories of practical reason is to consider what sense these theories can make of ourselves, of our motivations, desires, emotional responses, attitudes, and intuitive value judgments. Does either theory give us an adequate framework for self-understanding? In particular, does either theory enable us to under-

Expressive Theory of Rational Action • 39

stand ourselves as unified persons whose psychological states, when they are rationally justified, are coherently related to one another? This section poses some di^^Ue^xQIiSSauentiaHst theorie^have in accounting for the unity of the self. ***Consider first how expressive theories conceive of the unity of a person s psychological states. A person's psychological states are rationally justified if they are reflectively endorsable from a common point of view established in dialogue with others (§1.1, §5.1). This point of view is one from which a person's psychological states can be grasped as having a fully coherent, publicly communicable meaning. When a person's psychological states are rationally justified, or come tolerably close, they bear expressive relations to one another that give them an internal coherence and unity. A persons emotional responses and attitudes are merited by or appropriate to their objects. A person s values provide grounds for simultaneously evaluating objects and supporting the attitudes she takes up toward them. A person s rational desires, motives, and actions adequately express her attitudes in being governed by the norms constitutive of those attitudes. The entire ensemble of a person's attitudes, desires, motives, and actions is globally unified over time by her attempts to make her life meaningful in accord with the requirements of narrative unity. (^nrK^qi^e.ntLili^, jn rnnfrast^ foVe iip.an instrumental perspective on the self. They represent the self as a locus of causal forces projected into tneTuture and significant only as they bring intrinsically valuable states of ^ISaiftririfo anH out of existence^ This conception of the self is instrumental Because it views a person's psychological states as rationally justified on instrumental grounds. Motives, beliefs, emotions, attitudes, value judgments, internalized norms, and modes of deliberation are justified to the degree that they bring better states of affairs into existence. This is not to deny that the state of being constituted in a certain way—say, believing a truth—can be intrinsically good. But the overall justification of a belief is fundamentally a function of its consequences, including itself. The unity of the instrumental self is based on the unity of its preferences, that is, a person's dispositions to act so as to bring about certain states of affairs over others. Preferences are unified if they can be ranked in a single complete, consistent preference ordering. They are complete if for any two items, one is either more, less, or equally preferred to the other. They are consistent if transitive: for any three items, if A is preferred to B and B to C, then A is preferred to C. On some theories the unity of a person's preferences is secured by the supposition that it tracks independently identifiable intrinsic values of states of affairs. On other theories the

40 • An Expressive Theory of Rational Action

fact that a state of affairs is an object of a person's already unified preferences is what makes it intrinsically valuable (immediately normative for choice). A choice or action is rational if and only if it tends to bring about the person's rationally preferred states of affairs. The instrumental theory can explain the rational coherence of prefer-' ences with choices and true value judgments, provided that its conception of intrinsic value can support the strong ordering principles of completeness and transitivity. I will argue in §3.3 that these principles cannot be sustained. The instrumental theory runs into deeper trouble in attempting to account for the rational unity of our emotions, attitudes, internalized norms, intentions, and ways of deliberating. In unifying a person's preferences and choices around the achievement of particular consequences, the instrumental view creates discord among other aspects of the self. Consider the cases that arose in contrasting consequentialist and expressive views in §2.4. The paradox of hedonism shows that consequentialism may require a person to aim at ends that aren't really valuable, so that the really intrinsically valuable states of affairs can be produced as an unintended by-product. It may require people to focus on and intrinsically care about what, from its own perspective, is not really or ultimately important. The tripping and adultery cases show that consequentialism has difficulty^ making sense of the different h h same consequences brought bh about through different meansr have to the The rationally justified emotion is that wTiich^constitm states of affairs. But, as the above cases show, the emotions people actually have are responsive not to consequentialist but rather to expressive considerations. People can't reflectively endorse an emotional state which is induced by mechanisms or thoughts that don't internally warrant it on expressive grounds, even if consequentialism says it would be instrumentally good to have that state. In other words, the instrumental account of what makes an emotional state rational cannot provide a coherent basis for self-understanding. This fact is abundantly illustrated in popular cultural representations of androids and robots, who exhibit the purely instrumental, calculative, unemotional rationality extolled by consequentialism and thereby reveal themselves to be social clods and emotional dolts. Lieutenant Commander Data, an android officer of the Starship Enterprise on the television series "Star Trek: The Next Generation," provides a case in point. Attempting to forge his first romantic relationship with a woman, he feigns anger and picks a fight with her for no apparent reason. She objects to his behavior, pointing out that she has done nothing to warrant such anger. Data replies

An Expressive Theory of Rational Action • 41

that he knows this perfectly well. The point of his instigating a fight was not to express any real objections he had to her but to establish an occasion for reconciliation after a falling out: his anthropological studies had taught him that such events tend to strengthen and deepen romantic friendships. She politely tells him that their relationship simply can't make sense if conducted on these terms. Data failed to grasp that the instrumental value of emotions and of the acts that express or elicit them is parasitic upon people understanding them in terms of the non-consequentialist, expressive logic of appropriateness. If his anger isn't proper and sincere, and if their reconciliation is not based on a mutual agreement about what behaviors warrant anger, on sincere apologies by whoever misbehaved or misjudged the other, and on resolutions to act and feel appropriately as judged in terms of expressive logic, it won't be an authentic reconciliation on the basis of which the relationship can coherently continue. The same failures of reflective coherence mark other proposals to reform ordinary practices along consequentialist lines. Blame should be meted out to wrongdoers only if it would deter future wrongdoing, and to anyone else unlucky enough to be in a position where being made an example of (however unjustly) would deter further wrongdoing (Smart 1973, pp. 69—71). Honors and prizes should not be meted out to those who merit them, if distributing them on some other basis would produce better consequences, say, by raising the self-esteem of those who don't merit them. Again, th£Sjejtnaiiir^ practices can have their intended effects only if people think they are being governed in accord with expressive logic. Blame will rationally inspire outrage and resentment rather than guilt if people know it is being assigned on grounds of expedience rather than justice. Prizes will hardly inspire selfesteem if people know that they are awarded just to make them feel better, because the prize-giving would then express a misguided, patronizing "benevolence" rather than genuine honor or admiration. Consje^uentiaHst reasoning does not provide a coherent basis for us to understand and reflectively endorse our own emotions arid attitudes. Exrrre?siwYeasoriing does provide such a ground for rational self-understanding. If the practices and relationships that require an expressive self-understanding are intrinsically valuable by consequentialist lights, then consequentialism recommends that we adopt the expressive theory for our own self-understanding, even if it is false. This move creates a dramatic division between the consequentialist standpoint of authentic justification and the purportedly mistaken standpoint we are supposed to take up for purposes of deliberation, self-evaluation, and self-government. Consequentialism

42 * An Expressive Theory of Rational Action

tells people to cultivate states of character, emotions, and modes of deliberation that accept as genuine reasons for action considerations which do not really justify the actions. The consequentialist explanation of "silencing" in the virtuous sea captain's case reflects a similar division of justification from deliberation and motivation. It represents the standpoint of virtue as one that mistakenly regards certain consequences of action as unimportant. Agent-neutral versions of consequentialism, which give all agents a common final end, produce other divisions in the self. The motives it is best to cultivate may not be the rational ones to act on in a particular case. Cultivating love toward those close to oneself may generally lead to the best consequences, but sometimes, by forgoing the rescue of loved ones, one may be able to save a greater number of strangers' lives. Such an act would be contrary to loving motivation, but would lead to impersonally better consequences and would thus constitute the right action (Slote 1985, p. 93). Consequentialism rejects conceptions of personal responsibility that recognize an evaluatively significant boundary between one's own and others' actions. If another person, however unreasonably, makes his action contingent on some choice of mine, then he establishes a causal link between my choice and the consequences of his action. My aims thus become hostage to his unreasonable ones. So, famously, consequentialist moral theories tell people to cooperate in evil projects if others can credibly threaten to perform worse deeds if they don't cooperate (Williams 1973, pp. 95-99). \ Thus, consequentialism demands that the instrumental self aim at and \care about what isn't really important, deliberate in terms that don't mthenticaHy justify its actions, accept false beliefs and value judgments, md have motivations that it would be wrong to act on. These are all udgments that consequentialism makes from its own point of view on the ;orts of instrumental selves it recommends we become. In addition, consequentialism tells people to have emotions and attitudes that aren't really j warranted by their objects, to abandon any robust conception of personal | responsibility, and to repudiate (via adoption of the rule to ignore sunk Icosts) any attempt to construct meaningful connections between its past land future actions. These are judgments that expressivism makes on the ^sorts of instrumental selves consequentialism recommends we become. From an expressive point of view, all these divisions in the instrumental self make a life conceived in its terms a gigantic fraud. Consequentialists reply by pointing out that by their own lights, there is nothing bad about leading a fraudulent life. In fact, it is the intrinsically best way of life.

An Expressive Theory of Rational Action • 43

This response, however internally consistent, is repugnant to common sense. The instrumental conception of the self fails to provide us with a coherent basis for self-understanding and requires disturbing divisions among different aspects of the self. Common sense agrees with the expressive view that to fail to integrate the perspectives ofjustification, deliberation, motivation, and intention is to engage in expressively incoherent and hence irrational action. If an alternative theory is available that provides a coherent basis for self-understanding and that does not require these repugnant divisions in the self, it has these two advantages over consequentialism. The expressive conception finds a basis for self^understanding and for the unity of the self where the instrumental conception often demands disunity. On the expressive view, a person's conception of justification, of'who, what, where, when, and how to sensibly value, is directly expressed in her attitudes, motives, deliberations, intentions, and actions. She need not adopt a false or delusional standpoint of deliberation and motivation to do the rational thing. Consequentialists have three responses to this argument. First, they point out some advantages they purport to have over expressive theories and argue that expressive theories suffer from their own incoherencies. These advantages are mainly tied to the appeal of a calculating, maximizing logic of rationality and will be discussed in Chapter 3. Second, they attempt to incorporate some of the advantages of expressive logic in a hybrid theory that still retains the fundamental features of consequentialism. This response will be examined in §§4.4—4.5. Third, they reject the criteria by which expressive theories claim an advantage in this preliminary comparison of the two theories. Because these criteria are based on commonsense intuition, consequentialists have a stake in attacking the authority of common sense. This response will be considered in Chapter 5.

Pluralism and Incommensurable Goods •

3 • Pluralism and Incommensurable Goods

3.1 The Advantages of Consequentialism Expressive theories have an advantage over consequentialist theories in providing a coherent basis for self-understanding, in accounting for the unity of the self, and in making sense of ordinary intuitions about intrinsic value and norms of appropriate behavior and feeling. But consequentialism would not have such a firm grip on philosophical and social scientific accounts of rational choice if it could not claim important advantages of its own. Often rational choice and deliberation are explicidy oriented toward achieving optimum results. Consequentialism, unlike the pluralist-expressive theory, has a straightforward account of this. Call this consequentialism s explanatory advantage. Consequentialism promises to provide a single, simple, precise, and determinate procedure of justification that employs objective calculation to overcome disputes about what to do. The pluralist-expressive theory calls for action to be guided by norms described in terms of ideals and evaluative concepts such as "respect," "friendship," and "charity." These norms require interpretation to be applied. Because their constitutive concepts are essentially contestable (§5.2), the expressive theory cannot provide a determinate procedure for resolving conflicting interpretations. Call this consequentialism s pragmatic advantage. This will be discussed in §§3.2-3.3 and in Chapter 4. Consequentialism follows a long tradition in Western philosophy that contrasts reason with emotions and social norms and that seeks an independent perspective from which our emotions and social practices can be criticized. Consequentialism provides a dispassionate, ostensibly asocial method for criticizing emotions and social norms: see whether they produce the best consequences. The pluralist-expressive theory rejects the

45

dichotomies between reason and emotion and reason and social norms. It represents rational action as expressing emotions and attitudes through social norms. It needs to show how it can do this and still be able to criticize our emotions and social practices. Call this consequentialism's critical advantage. This will be discussed in Chapter 5. In this section I critically examine consequentialism's explanatory advantage. The phenomenon to be explained is that people often seek to maximize value: to get the best buy for their money, to make the best of a bad day, to send their children to the best schools, to perform their best in a musical competition. Consequentialism has a natural explanation for this, since maximizing value is just what practical reason demands. The expressive theory denies the coherence of the project of globally maximizing the value of states of affairs. States of affairs mostly have only extrinsic value relative to the context of decision. They have little actionguiding significance independent of a person's attitudes, relations to others, ideals, past, and present predicament. It makes no more sense to globally maximize the value of states of affairs, independent of a person's context of decision, than it does to globally maximize the value of tools, independent of the purposes to which they will be put. This difference between consequentialism and expressivism is not conclusive. For the phenomena to be explained include only cases of local maximization. Consequentialism tries to explain maximizing value within a decision context as instrumental for globally maximizing acontextual value. The expressive theory claims that global rationality need not be conceived as local rationality writ large. It needs to show how maximizing value can play a local role in a theory of practical reasoning globally governed by expressive norms. To see whether it can do this, we need to ask: what requirements must the theory of value meet if local value maximization is to be possible? This question may be answered by considering what the theory of value would have to do to meet stronger demands of rational choice. As the conditions are relaxed, we can see what flexibility becomes available to the theory of value. A venerable tradition in philosophy upholds three claims about the demands of practical reason: that reason can settle all questions about what to choose; that it requires the global maximization of value; and that the grounds for rational choice must be fully and decisively articulable, leaving no room for judgment and hence none for dispute (compare Rawls 1971, pp. 548—560). These claims provide a powerful argument for monism, the view that there is only one ultimate standard of value. The first two claims imply that the values of all options are com-

46 * Pluralism and Incommensurable Goods

mensurable, that their values can comprehensively be measured on a single scale.1 More technically, two goods, A and B, are commensurable if and only if there is a scale of overall value by which they can be at least ordinally ranked. An ordinal ranking says that either A is better than B, B is better than A, or they are equal in value. Unlike a cardinal ranking, it doesn't say how much more valuable one good is than the other. If goods are not commensurable, then it does not make sense to maximize their values. The third claim supports the inference from commensurability to the existence of just one value. For the only plausible scheme of fully articulate and decisive reasoning applicable to all goods seems to be the monistic one of dominant-end reasoning. If there is only one value, reason demands that people make its maximization their sole ultimate or dominant end. Choices among seemingly diverse goods are justified by their instrumental value in promoting this end. Monism offers a simple and compelling way to satisfy the three claims about rational choice. This argument for monism has often appeared in philosophy, usually in defense of hedonism (Kant 1956, pp. 20-22; Plato 1961b, 356a-357b; Sidgwick 1981, p. 406). Monism triumphs for lack of an alternative that promises to settle decisively all questions of choice. Many contemporary consequentialists find monism unattractive. They have dropped the third requirement of exhaustive articulability and accordingly relaxed the demands placed on the theory of value. If only the first two claims are accepted, the theory of value must be reductionist. While it can be pluralist in accepting many distinct standards of value, it must hold that there is a single measure of value which commensurates all goods. This theory is reductionist because even if every standard provides its own measure of value, it supposes that these measures can be reduced to a common overall measure. James Griffin (1986, pp. 26-31) offers a reductionist theory of value in arguing that there are many objective standards of value, but that informed preference provides the common measure of goods meeting any of these standards. Reductionist theories lack a fully articulate justification for action, for at some point they must appeal to brute preferences or brute facts about the relative weights of different values. Welfare economists often appeal to the first two demands of rational choice to argue for a single measure of value. The necessity of choice between two goods, such as money and life, requires their commensurability, preferably in terms of money (Arrow 1967, p. 5; Maler and Wyzga 1976, p. 6). A theory that meets the first two claims of practical reason by showing

Pluralism and Incommensurable Goods * 41

how to globally commensurate options promises important pragmatic advantages. It would be extremely simple and applicable everywhere. It would offer precision in practical matters and, if it were monistic, offer complete articulateness as well. It would yield a determinate result better than or equal to any alternative. Most important, it would eliminate the problem of apparently conflicting values by reducing deliberation to a matter of calculation. These are the pragmatic advantages that can be claimed on behalf of consequentialism. The pluralist-expressive theory cannot offer these pragmatic advantages, for it rejects the notion that the values of states of affairs can be globally compared. Suppose the second claim about practical reason were relaxed and demanded only the possibility of local value-maximization, within a decision context. Then, instead of providing one measure of value that commensurates goods for all contexts, the theory of value would have to provide only, for each context, a measure of value that commensurates goods within it. It isn't obvious that the pluralist rational attitude theory can provide this. According to it, intrinsic value judgments guide emotional and attitudinal responses. These are expressed through actionguiding norms, which tell us to realize a plurality of values in states of affairs. Sometimes these norms tell us to maximize the value of some good that possesses many different valuable features. Maxirnization requires commensurability. How are the plurality of values to be combined" in one measure of overall evaluation? There cannot be one answer here as in rea^trdmsttheories, for different values are relevant in different ways in different decision contexts. 3.2 A Pragmatic Theory of Comparative Value Judgments The pluralist-expressive theory yields a pragmatic theory of comparative value judgments. The theory is called pragmatic because it represents comparative value judgments as constructed for particular purposes, rather than as discovered in the intrinsic natures of the items being compared or in their immediate relation to some response in us, such as preference or pleasure. The pragmatic character of comparative value judgments follows from the fact that the goods it compares are extrinsically valuable. They derive their importance or action-guiding relevance from the decision frame in which they appear. And the decision frame, in turn, is conditioned by the purposes and attitudes of the agent. There is no single thing people are doing when they seek the "best" and hence, contrary to reductionist theories, no single formula for calculating the "best." Dif-

48 • Pluralism and Incommensurable Goods

ferent comparative strategies are appropriate for different purposes and functions and for different kinds and objects of overall evaluation. The pragmatic theory explains comparative value judgments by considering their functions, why people care about them, and what practices of evaluation serve these functions. Authenticity and importance are two values of standards of value particularly significant for this investigation. If a standard lacks authenticity, it is arbitrary, unfounded, or illusory. A standard is authentic if and only if it is a candidate for some rational response-guiding function—if it could make sense for a person to guide her responses by it. Authentic standards must meet the following conditions: they must identify differences among the objects being evaluated which people are capable of caring about, and their claims to authority must not rest on delusion, error, or other cognitive defects. The second condition rules out evaluative standards people care about, such as racist ones, whose authority depends upon pseudoscientific claims or selfdeception. Not every authentic standard can claim to guide a person's responses, since it may not make sense for the person to care about it. I don't ski, so it doesn't make sense for me to care about authentic differences in the value of snow for skiing. A standard is important to a person if it makes sense for her to care about it. Importance measures the degrees to and the ways in which an authentic standard properly informs a person's concerns and actions. Suppose we are considering whether to adopt a new standard or method of evaluation to govern a practice. We must judge how important the standard is. It could be important if it serves functions valued by the participants in the practice or if the practice would be improved by adopting it. Such a judgment was recently made in competitive figure skating. This sport has been strongly shaped by a contest between two standards of value—"athleticism," expressed in the execution of difficult jumps and spins, and "artistry," expressed in precisely tracing geometrical patterns on the ice. (The latter is distinct from the standard of "artistic impression," which a skater meets by expressively coordinating her movements to accompanying music.) Figure skating began as a purely artistic endeavor, but gradually the athletic component gained prominence. People recently challenged the inclusion in international competitions of the "school figures," in which skaters must trace perfect circles and other shapes in the ice. They successfully argued that, although figure skating originated in such endeavors, today they are just training exercises. Including them in a skating competition now makes as much sense as including scales in a piano competition. Skating competitions would be improved if the school

Pluralism and Incommensurable Goods • 49

figures were dropped, because then the choice of winners would more truly reflect the athletic values that practitioners of the sport really care about. No one claimed that artistry was an unauthentic standard, that past awards informed by this standard were fraudulent. They claimed only that it was no longer important. The power of the pragmatic approach to explain comparative value judgments can best be illustrated by considering examples. I will focus on two types of comparative evaluations: impersonal goodness-of-a-kind judgments, typically applied to products and performances of athletic, artistic, academic, and similar practices; and personal judgments of which good would be better to have or choose. Each of these kinds of multicriterion evaluation serves different functions, has different action-guiding implications, and is arrived at by different methods, depending upon its objects and functions. Goodness-of-a-kind judgments applied to achievements within practices compare different goods as exemplars of a particular kind of excellence. These judgments are impersonal: in evaluating a gymnastics routine, judges consult not their personal preferences but shared standards constitutive of excellent performance, such as balance, difficulty, and confidence of execution. Goodness-of-a-kind judgments consider only values internal to and constitutive of the practice, not values the object might realize externally to its contribution to the excellences definitive of the practice. The overall excellence of a portrait as a portrait could depend upon its subtle revelation of a persons character, but not upon how good it is as a commercial investment, or how good it is for covering a stain on a wall.2 Goodness-of-a-kind judgments involve evaluation against several standards of value. The component-value strategy attempts to commensurate goods by representing the overall value of a good as an objective function of its component values. If each standard can be assigned a weight, then the overall value of an object is the weighted sum of its component values. The component-value strategy possesses many of the pragmatic advantages of reductionism, without assuming that it provides a global solution to commensuration. It represents evaluation as essentially a matter of calculation, with the aim of making the process precise and decisive. Sometimes the component-value strategy offers a successful solution to the multi-criterion evaluation problem. An excellent example is found in decathlon scoring, in which times and distances in different events, such as the hundred-meter dash, the shot put, and the long jump, are converted to a common point scale. This is achieved by assigning equal weight to each

50 * Pluralism and Incommensurable Goods

event (twelve hundred potential points)—zero points to just below the worst recorded performance, twelve hundred points to a performance just beyond the present world record—and dividing the resulting interval by twelve hundred to get the incremental speed or distance per point. Adjustments are made for the peculiarities of different events, as well as to ensure that increments close to the world record earn more points than increments further away (MacKay 1980, pp. 61-71). The pragmatic theory explains the adequacy of this scheme of evaluation better than reductionist or monistic theories of value. The decathlon scoring system is reasonable because it serves the functions assigned to it: it fairly, decisively, and precisely ranks different athlete s performances across different events according to a process that experienced devotees of the practice find reasonable and relevant to what they care about (Luban 1990). It offers one authentic scheme for answering the question "who is the best all-around athlete?" But other scoring systems, or even other kinds of competition, could do so as well. It makes sense for judges, athletes, and sports fans to care about scoring differences on this system because it is the only authentic scheme that has been socially instituted. Consider how rival reductionist theories would try to explain the decathlon scoring scheme.3 Three basic measures of value are available to reductionists: (1) the degree to which we take pleasure in it; (2) the degree to which it satisfies our preferences; (3) the degree to which it possesses a distinct property, "intrinsic value," that supervenes upon its component values. The decathlon point scale determines which performances in different events represent equivalent athletic achievements. For instance, 10.1 seconds in the hundred-meter dash earns the same number of points (1043) as 812 centimeters in the long jump. Intrinsic value theory must explain the equivalence of the two performances in terms of the values of their intrinsic features. But it is absurd to suggest that the decathlon scoring system tracks independently existing relations among values, that there is something inherent in the value of an 812 centimeter jump that makes it as good as a 10.1 second run of 100 meters. This would not explain why equivalent performances change as greater advances are made in some events than in others. The pleasure and preference theories claim to set either the standard of value or a measure of value that tracks the authentic standard. On the standard view, they must be able to explain the equivalent performances as reflecting the judges' independently equivalent preferences or pleasurable feelings in witnessing the two performances. This gets the explanation backward, both logically and causally. We construct a point-scale in the

Pluralism and Incommensurable Goods • 51

first place so that we can guide and train our preferences and pleasures over the performances being judged. Without the point-scale, our preferences and pleasures would not be discriminating enough to ground precise, calculable, and articulate comparative value judgments. On the measure view, we must suppose that preferences and pleasures faithfully track the authoritative standard. Bias, fan loyalty, and considerations of meaningfulness disrupt this tracking. One performance can be more pleasing than another because it represents a greater triumph for the athlete who performed it than the highest-scoring performance represents for the winning athlete. Often we would like to see the most meaningful performance be the winning performance, for example, the one that vindicates an athlete who had until then choked in crucial contests. This preference does not track the best performances. The component-value strategy cannot be extended to cover all best-ofa-kind value judgments for at least two reasons. First, Arrow and Raynaud have proven the following theorem (1986, pp. 18—21). If goods can only be ordinally ranked by a standard of value, if their rankings along one standard vary independent of rankings along others, and if the relative ranking of any two goods is not affected by the entry or elimination of other items, then the component-value strategy will either yield inconsistent evaluations or collapse into the ranking generated by one of the standards. An inconsistent evaluation is intransitive: it says that A is better than B, B better than C, and C better than A. Since many reductionists concede that their favored measures of value yield only ordinal rankings, this theorem undermines pluralism and pushes them toward unattractive monistic views. Pragmatists have more flexibility in using the component-value strategy than reductionists, for they are free to give up the entry/elimination constraint.4 This constraint is plausible only if we assume that comparative value is derived from a prior measure of the intrinsic values the items possess independent of the context of comparison. Pragmatists deny that we have any reason to postulate such acontextual values. The reasonableness of scoring in skating competitions confirms the pragmatic view. The winner of a contest is determined by combining skaters' scores in two events, an original and a long program. Although cardinal rankings are reported for each event, only the ordinal rankings they generate are used to compute the winner.5 This permits a skater to win in a three-person competition who would have lost had one of the other two dropped out. (Skater A could win the contest if the rankings in the original program are (A, C, B) and in the long program (B, A, C). B would beat A if C hadn't

52 • Pluralism and Incommensurable Goods

competed because the long program is weighted more heavily than the short. But B loses to A when C enters because B skates third-best in the first event and A never skates worse than second.) Rules for scoring skating competitions can't be rationalized in terms of intrinsic values, but can be rationalized on pragmatic terms. The second reason the component-value strategy is not generalizable is that some practices don't support the calculations it requires, nor would they be improved if they did. Consider the evaluation of fiction. Limitations in literary standards prevent the precise calculations available in athletic scoring. It makes sense to say (in decathlon scoring) that one athlete s jump was twice as good as another's but not that one author's story was twice as suspenseful as another's. This is not due to a lack of precise measuring instruments. We could invent an operational definition of "suspensefulness" susceptible to precise measurement, but the differences this would capture would be unlikely to be authentic. We wouldn't even know how to begin to measure other aesthetic values, such as originality, in operationalizable terms. This problem cannot be solved by inventing more precise evaluative concepts. The different dimensions of aesthetic value would still have to be ranked against one another. How should we weigh suspensefulness against the ingenuity of the plot in evaluating thrillers? Any formula for weighting these standards would be groundless. The functions of aesthetic evaluation are not well served by precise, quantifiable standards placed in fixed relations to one another. The precision, articulability, and decisiveness of the component-value strategy comes at the cost of more important functions of aesthetic evaluation, such as promoting interpretive richness and creative growth. The point of much Western artistic production has been to challenge received standards and invent new ones, so as to develop new ways of perceiving, appreciating, and producing art. These purposes are not served by substituting new sets of fixed standards for the old. Progress is achieved in part by reinterpreting existing standards in surprising, interesting, and contestably new ways. This kind of invention, growth, and dynamism is facilitated by evaluative standards that are openended, fluid, susceptible to competing interpretations, but not so empty of content that anything could satisfy them. These are standards described in terms of essentially contestable evaluative concepts, which lack the features (full operationalization, fixed boundaries, quantifiability) required to be amenable to calculative strategies of reasoning. Goodness-of-a-kind evaluations serve other functions besides guiding choices. Adherence to them defines membership and authority in a prac-

Pluralism and Incommensurable Goods • 53

tice devoted to an excellence or in a camp contending for ownership of the practice. They set standards for admiration and aspiration, for training members' sensibilities, and for disciplining their participation in the practice. They provide a framework through which the practice can progress. These goods would be undermined by a system of aesthetic evaluation that generated precise and fixed rankings. The choice-guiding functions of impersonal aesthetic evaluations—the ones which are supposed to be served by precision and decisiveness—are, by contrast, relatively unimportant. They are helpful in awarding prizes and honors. A person's other aesthetic choices are not bound by goodness-of-a-kind evaluations. I might judge that one book is a much better novel than another, yet choose to read the inferior work for many good reasons connected to my varying moods, interests, tastes, and cares (perhaps I like trashy novels, or I'm in the mood for something light). Because options may belong to more than one kind, best-of-a-kind judgments can conflict. A thriller with welldeveloped characters but a less suspenseful plot may be better as a novel but worse as a thriller than one with a gripping plot. The pragmatic theory explains better than intrinsic value theory which standard of comparative value is important to my choice: it depends on the purposes I have for choosing the book. The aims of entertainment and insight make relevant different standards. Personal judgments of what option is best for one to choose constitute a different type of multi-criterion evaluation from impersonal goodness-ofa-kind judgments. Unlike the latter, they take into account the individual's personal circumstances, tastes, moods, interests, and responsibilities. They also permit comparisons of options that differ in kind or share few valuable features that could afford a basis for intrinsic, impersonal comparison. The higher-order good strategy provides a way to compare different choices in a personal evaluation by relating them to a larger whole. In choosing between getting a job and going to school, a person may consider how either choice would contribute to her having a good life, or a good family life, or a good career. Conceptions of a good life, a good career, and so forth are conceptions of second-order goods, which consist of particular arrangements and connections of first-order goods.6 A good day in the life of a person is a second-order good consisting of a sequence of activities and events. The overall worth of a higher-order good is not a function of the individual values of its component goods, for it may be judged by standards, such as balance, variety, meaningfulness, and harmony, that may not be applicable to their component first-order goods. It is an organic unity

Pluralism and Incommensurable Goods • 55

54 • Pluralism and Incommensurable Goods

judged against a scheme for recognizing when its components are well organized, distributed, and connected. Consider how the higher-order good strategy can inform the choice between doing errands on Sunday and going on a daytrip (Stocker 1990, pp. 172—173). These activities share few advantages and disadvantages. Yet one might compare them by considering how they would contribute to the second-order goods of a good day or a good week. These are not ends distinct from the activities and events which constitute them. But they do include some conception of how these activities and events should be structured: a basic framework is provided by routines, responsibilities, and other necessary tasks, as well as by a view of how labor and leisure may be favorably scheduled. For example, one might prefer to relax after daily chores are finished, so that this time may be enjoyed without worrying about duties which lie shortly ahead, or so that the relaxation can be deserved. The choice between errands and going on a daytrip can then be made by seeing which option would make for a better overall arrangement of goods constituting a good day (perhaps a good Sunday) or a good week. The pragmatic theory of comparative evaluation has a better account than reductionist theories of how the higher-order good strategy works. Reductionism claims that there is a single measure of overall value applicable to all contexts. If overall value is organic, then any local use of a second-order good in deliberation, such as of a good day or a good week, must be rationalized as promoting the value of some more comprehensive higher-order good, such as of a good life. This good does not provide a unique standard: are we concerned with a happy, meaningful, or virtuous life? Nor does it provide a global standard: only those lower-order choices that have some impact on the worth of the higher-order good can be rationalized by moving to the more comprehensive level. Decisions that may affect the quality of one's day but have no prospect of influencing the overall meaning, happiness, or virtue of one s life cannot be justified by referring to the more comprehensive standards. It is not even true that reason demands the use of the higher-order standard when a decision could influence both the higher- and the lower-order good in different directions. There is no one aspect, dimension, or unit of life whose value we must be maximizing if our choices are to be rational. Which higherorder good it makes sense to use in justifying a person s choices depends on the context of decision, especially on the agent s reflective understanding of her own predicament. The kind of reflection called for in choosing among goods by referring to a higher-order good is more a matter of interpretation than of calcula-

tion. A conception of a higher-order good is not like a rigid template which measures first-order goods according to the precise degree to which they match its shape. Or rather, it may be like that for people of extremely inflexible habits. But it is better viewed as an open-ended and flexible schema, which can be filled out and reshaped in an infinite variety of ways, as circumstances, opportunities, purposes, principles, mood, taste, and imagination recommend. There is no fixed pattern to a good day or a good career. The schema provided by a higher-order good is too open-ended to support a fixed set of tradeoff schedules for different values; people turn to it to guide their imagination and interpretive powers, not to engage in mathematical calculation. Although deliberation using higher-order goods thus sacrifices the pragmatic advantages of consequentialist models of rational choice, it has its own advantages. The open-endedness of higher-order goods encourages creativity and growth, independence, individuality, flexibility, exploration, and play 3.3 Incommensurable Goods My investigation has proceeded thus far on the assumption that practical reason demands at least the local maximization of value. But the expressive theory of practical reason permits us to relax this assumption. If the norms properly governing choice in a context do not prescribe the maximization of extrinsic value, there may be no content to the thought that there is some extrinsic value there to be maximized. If the norms properly governing attitudes do not prescribe that one intrinsic good be valued more intensely or in a higher way than another, or that they be valued in exactly the same way, there may be no content to the thought that either one good is intrinsically more valuable than another or they are equal in value. A theory of value adequate to the demands of rational choice could thus leave open the possibility that some goods are incommensurable. Two goods are incommensurable with respect to some scale if one is neither better, worse, nor equal in value to the other in the respects measured by the scale (Raz 1986, p. 322). Incommensurability generally arises when the following three conditions are met: (1) the goods in question meet the standards measured by the scale in very different ways; (2) there are no gross differences in the degree to which each good exemplifies its own way of meeting the standards; and (3) meeting the standard in one way is not categorically superior to meeting it the other way. The standard of brilliance can be met in very different ways, for example, musically and scientifically. Neither

56 • Pluralism and Incommensurable Goods

way of being brilliant is categorically superior to the other. Bach and Darwin were each highly successful in their own ways of being brilliant. Neither was superior in brilliance to the other, nor were they roughly equal in brilliance. If this were so, then a small but significant improvement in the brilliance of one would suffice to tilt the judgment in his favor. But it is silly to claim that, say, had Darwin achieved some brilliant insights into genetic theory as well as evolution, he would thereby have exceeded Bach in brilliance. There is an intransitivity in value judgments here: our imagined Darwin is not more brilliant than Bach, Bach is not more brilliant than Darwin, yet our imagined Darwin is more brilliant than Darwin himself. Joseph Raz calls this intransitivity (in the relation "is not superior to") the mark of incommensurable value (1986, pp. 325326). The more a given scale of value encompasses very different, categorically unranked ways of meeting it, the more scope there is for incommensurability. So incommensurability is more likely to arise for more global judgments of overall value, in which a goods ranking on many standards is relevant to its overall standing. This poses a serious problem for reductionist consequentialists—those who accept the pluralist claim (that there are many distinct standards of worth) but also hold that rational or moral action must maximize the global value of states of affairs. If pluralism is right, and different ways of being a valuable state of affairs are not categorically ranked, then we should expect widespread incommensurabilities among states of affairs. If options cannot be compared in overall worth, then consequentialism will offer a very incomplete and fragmented perspective for agents to justify their actions. The pragmatic theory of comparative value judgments explains why this is so. It makes sense to construct a comparative scale of evaluation only if there is some point to doing so. Comparative value judgments make sense only if they serve some function. The more global scales of comparison serve no authentic choice- or attitude-guiding function, since the expressive norms constitutive of decision and response frames do the work that consequentialists think must be assigned to global evaluations. The pluralist-expressive theory implies that there is no fact of the matter about what is the best state of affairs from a consequential point of view and hence, no consequential fact of the matter about what is the best action. Phenomena surrounding incommensurability provide a powerful test case for the relative merits of pluralist-expressive and reductionist-consequentialist theories. Suppose we encounter deliberative situations with the following features: (a) we seem to be at a loss to make any credible relative

Pluralism and Incommensurable Goods • 57

judgments about the values of two options; (b) Razian intransitivity, the mark of incommensurability, is present; (c) the expressive theory explains why comparative value judgments have run out here; (d) the expressive theory provides other bases besides their relative values for choosing between the options. The ability of the pluralist-expressive theory to make both theoretical and practical sense of such situations would be powerful evidence in its favor. Situations meeting the above conditions are familiar. Consider Sarah, who must choose between two ways of life (compare Raz 1986, pp. 341— 344). One is secure, content, and parochial, focused on the cultivation of fulfilling and loving relationships among a small circle of friends and family. The other sacrifices lasting relationships with intimates for the sake of an outstanding career in ballet, which offers excitement, glamour, significant contributions to world culture, and broader knowledge of the world. Each way of life is good in its own way and defective in others. Suppose Sarah has equal prospects for leading either life successfully. I claim that consequentialists cannot plausibly argue that the value of Sarah s leading one way of life is commensurable with the value of her leading the other. Consequentialists can evaluate Sarah's options by either a welfare or an impersonal intrinsic value scale. (Because other reductionist scales incorporate at least one of these two, incommensurabilities in these will infect all other measures.) We have no grounds for claiming that one way of life is impersonally better overall than the other. Nor are they judged equal in value. Improvements in the degree to which one way of life exemplifies its characteristic merits need not make it superior to the other. A consequentialist could argue that this indecision only reveals our ignorance. If we reflected more deeply on the issue, we could make a determinate value judgment. The problem is that our society no longer reflects on these issues (Regan 1989, p. 1060). The expressive theory explains why reflection runs out here: in our liberal, pluralist, egalitarian society, there is no longer any point in impersonally ranking all legitimate ways of life on some hierarchy of intrinsic value. Plural and conflicting yet legitimate ideals will tell different people to value different lives, and there is no point in insisting that a single ranking is impersonally valid for everyone. The project of ranking all legitimate ways of life on an impersonal scale of intrinsic worth is action-guiding only for upper-class members of status-based societies, who have the freedom to choose the most "honorable" ways of life and to force others to perform the less "intrinsically valuable" but socially necessary functions.

58 * Pluralism and Incommensurable Goods

Welfare considerations also fail to rank Sarah's two ways of life. Let's leave aside questions of comparing welfare between persons, where consequentialists concede grave difficulties, and stick to the intrapersonal case. By hypothesis, Sarah is capable of leading each way of life happily and successfully; that is, either way of life is equally likely to turn out satisfactory to whichever of her future selves she will become. Each could be regarded by her as an attractive, but very different, prospect. Neither is superior to the other. But nor are they equal in value, for the situation meets the Razian in transitivity test. Consequentialists may object that if the values of the two ways of life are really incommensurable, then it wouldn't make sense to agonize over the choice between them. Why not just flip a coin? Agonizing must reflect the judgment that the values of options are commensurable (Regan 1989, p. 1059; Chang 1992, p. 28). I will argue in §3.4 that expressive theorists have bases for choice other than the values of options. They can also choose on the basis of what it means for them to make the choice for the reasons that motivate them. It does make sense to agonize about this. But the objection can also be turned against consequentialists. If two options are equal in intrinsic value, then consequentialism says that, leaving aside the value of valuing itself, it doesn't make sense to value one more than the other. But it always make sense to care about the radically different ways in which options meet a standard, if none is grossly less successful in its own terms than the others. It always makes sense to care about the composition of partial values making up a successful and satisfactory option (Sen 1981; Herzog 1985, pp. 123-132; Stocker 1990, pp. 291302). There just is no scale of overall value for all options about which we cannot sensibly judge that it is worth sacrificing some total quantity for a different composition of the whole.7 Some consequentialists reject a Moorean intrinsic value scale and take the measure of intrinsic value to be given simply by one's informed preference rankings (Griffin 1986, pp. 11—15). Whatever one most prefers when calm and informed is the option that maximizes value. This view avoids the above objection because it requires that a person be indifferent only among options about which she would be indifferent anyway when suitably calm and informed. If preferences are complete, then any two options are commensurable. And preferences are complete, because it is always possible to force someone to make a choice and hence express a preference between any two options. So all options are commensurable. This argument equivocates on the meaning of "preference." When preferences are identified with choices, they are complete. But when a person is

Pluralism and Incommensurable Goods • 59

forced to make a choice, she may lack any confidence that the choice is best, even if it is calm and informed. The kinds of preferences which we know are complete therefore cannot be identified with the kinds of preferences capable of being interpreted as measuring, tracking, or constituting value judgments. The completeness of a persons choices does not imply the completeness of her value rankings. These arguments put pressure on consequentialists to reject pluralism. Although monism is deeply implausible, given the obvious fact that we do find things valuable for more than one reason, it at least promises that all options are commensurable. Michael Stocker argues that even monistic views such as hedonism cannot guarantee this, if they permit, as seems inevitable, different pleasures to have different qualities. In that case, it can always make sense to prefer a different mixture of qualities of pleasure to a maximum quantity. One may find a peculiar, piquant pleasure more interesting than a greater quantity of a familiar, languorous pleasure (Stocker 1990, pp. 184—188). And certain mixtures may not be commensurable, as demonstrated by Raz's intransitivity test. Consequentialism, whether pluralistic or monistic, cannot explain how rational choice in many situations is possible. It cannot unify the self around a unified structure of aims. 3.4 Rational Choice among Incommensurable Goods The failure of consequentialism to solve the problem of rational choice between incommensurable goods would not count against it if the expressive theory could do no better. This would show only that the structure of values does not always fix a rationally determinate choice. But the expressive theory can guide choices among incommensurable goods, for it provides other bases for choice than just the values of their consequences. Recall that the expressive theory conceives of rational choice as oriented toward dual ends and, therefore, as having both an expressive and a causal dimension. People try to bring about consequences for the sake of the people and things they care about. Even if the values of the consequences of choice are incommensurable, the act of choosing one option for the particular reasons motivating it may have a different expressive significance from the act of choosing the other. These expressive differences provide grounds for rational choice unavailable to pure consequentialists. Consider again Sarah's choice between two ways of life. By hypothesis, neither is impersonally better than the other, and neither offers a superior prospect of future welfare to the person Sarah will become. But Sarah has more concerns toward herself and others than are expressed in self-love or

60 • Pluralism and Incommensurable Goods

benevolence, understood consequentially. First, she may be concerned about the sort of person she will become. Her commitment to an ideal may incline her toward one choice. This is not a welfare consideration. She may even rationally choose to sacrifice some of her welfare for the sake of achieving an ideal. This is also not a judgment of impersonal value. She need not claim that the ideals expressed in the rejected alternative are impersonally inferior to hers. It is a judgment of personal importance. Such judgments are partly independent of and normative for her preferences and pleasures and hence are not merely reflections of the other two available reductionist scales of value. Second, she may evaluate her choice by evaluating the reasons motivating it. Although there are good reasons for choosing one option, her motivations may also be informed by bad reasons. She may not want to yield to the fear of a challenge which impels her toward the secure life, or to the fear of intimacy which impels her toward the artistic life. Third, her sense of personal responsibility may demand that she make a narrative unity of her life. In choosing the secure life, would she be turning her back on everything to which she has devoted her life so far? Would this make her life up to that point a meaningless episode? Even if she would find neither future more satisfactory to herself at the times she would be living it, one future might make better sense of her whole life than the other. Fourth, she may consider her choice in the light of the other ends she cares intrinsically about. She may turn down the opportunity to live the artistic life, because choosing it might express callousness toward those she loves. Although this choice may deprive audiences of welfare and edification, it doesn't make sense for her to care about their interests at this stage in her career. She has no obligations to them yet. So, states of affairs may be incommensurable by any of the scales of global evaluation available to consequentialists. But choices among such states of affairs may be commensurable by expressive considerations. The expressive theory can explain why comparative value judgments run out, and it can guide choices even when they do. It does seem, then, that there are familiar situations for which the expressive theory has a better practical and theoretical explanation than consequentialism. A consequentialist may reply to this argument by trying to incorporate the expressive considerations just mentioned into a sophisticated accounting of the values of states of affairs. This response will be considered in §4.4. The expressive theory has a further advantage over consequentialism, in that even when it does need to commensurate goods, it does not call for the scope of commensurability to be as wide as consequentialism requires.

Pluralism and Incommensurable Goods * 61

Compare the demands made on the concept of welfare by consequentialist and expressive conceptions of benevolence. Expressive conceptions of benevolence, tied to social roles and special relationships, do not require the notion of overall welfare that consequentialist benevolence does (§2.2). The relationship between a given person or community and a benefactor circumscribes the aspects of welfare which properly concern the latter—perhaps to health, or literacy, or housing in typical charitable or welfare state projects; or to the sharing of personal intimacies, advice, and time in certain friendships, where financial well-being is not a proper concern. Comparative value judgments using scales limited to specific dimensions of welfare are considerably more plausible, justifiable, and easier to make than judgments of global welfare. A theory that doesn't require us to construct global welfare scales has these advantages over a theory that does. Consequentialists may dig in their heels at this point and argue that there is no way to avoid the need for global commensuration. They argue that the norms constitutive of decision frames, which are supposed to do the work in expressive theories that global commensuration does in consequentialist theories, inherently conflict. These conflicts can be resolved only by finding a more fundamental principle that justifies or explains the point of each in a way that allows us to determine what to do in every case (Mill 1979, pp. 53-57; Hare 1981, pp. 39-40; Sidgwick 1981, pp. 360, 406). The fundamental principle prescribes an ultimate end, such as welfare maximization, and intuitive principles are commensurated as more or less efficient instruments for achieving that end. Or, if obeying the intuitive principles is partially constitutive of the end, then weights must be attached to them to determine how they balance out in a particular choice situation. This argument misrepresents the function of expressive norms. Consider, for example, the demands of kindness and of respect. Kindness is partly expressed in obeying the norm not to hurt others' feelings. Respect is partly expressed in obeying the norm of telling the truth. Consequentialists suppose these principles function in commonsense moral thought as unreflective rules (Hare 1981, ch. 2; Sidgwick 1981, bk. 3, ch. 11). If this were true they would often conflict and reason would have to appeal to more fundamental norms to resolve the conflict. Consequentialists propose either that these two principles be evaluated as instrumental to some dominant end independent of each, such as happiness, or that their distinct ends be somehow compared in value. In either case, we need some comprehensive answer to the question: how much is this true belief

62 • Pluralism and Incommensurable Goods

worth compared to that quantum of hurt feelings? But pluralism implies that there are widespread incommensurabilities among these states of affairs, that any weights assigned to the different principles will be arbitrary, and that there is no dominant end that can reconcile conflicts between principles. We can resolve apparent conflicts within the terms of the expressive theory by reflecting more deeply on the respective demands of kindness and respect and by considering what the actions of telling the truth or avoiding hurt feelings would mean in the present context. They mean different things, depending on the agents relationships to the people for whose sake he acts. An adviser to a legitimate official is duty-bound to tell the truth about the failure of the official's policies, even at great cost to his self-esteem. In contrast, it is absurd to charge a person with dishonesty for not expressing her frank opinion, when solicited, of her elder aunt's taste in dress. A mere acquaintance may rightly be judged intrusive for telling a dying person that he is dying, even if she correctly judges that the dying person ought to know and that his relatives do him a disservice in misleading him. Yet a close friend may subvert her friendship in refusing to share this information with her friend. In each of these choices between truth-telling and sparing someone's feelings, one does not simply ask how much this piece of truth is worth in comparison with that instance of self-esteem. Rather, choice is made against a background of appropriate relations among people, as defined by institutional responsibilities, rules of etiquette, ideals of friendship, and so forth. Within a given relationship, telling the truth or sparing someone's feelings has different expressive meanings. Telling the truth in one case is simply tactless or presumptuous; in another, an act of honesty and courage; in yet another, an expression of a kind of loyalty and respect proper to friends. Sparing someone's feelings may in one case constitute an act of sycophancy; in another, of considerateness; in yet another, of love. These examples illustrate how expressive norms, such as those favoring truth-telling and sparing people s feelings, have authority to the extent that they are adequate expressions of the ways people ought to value one another. These norms, expressed at such a high level of generality, are only crude approximations of the actual demands of respect and kindness in different situations. The application of intuitive principles requires an interpretation of their underlying expressive point in light of the relations the agent has to the people for whose sake she acts.8 If different principles appear to generate conflicting recommendations, the task for a rational

Pluralism and Incommensurable Goods • 63

agent is not to weight and aggregate their consequences, but to seek more refined interpretations of their demands so that all of them can be satisfied. This is the traditional task of casuistry, reasoning by analogy, and the other commonsense modes of practical reasoning familiar to ordinary life (Jonsen and Toulmin 1988). Nothing guarantees that an agent's rational evaluative attitudes will all be reconcilable in a particular situation. Sometimes the demands of love or of piety fundamentally conflict with the demands of particular ideals, such as patriotism. It may be impossible to adequately express one's love for two people if they become enemies. In such cases, there is no guarantee that either choice will make better sense of one's valuations than the other or that either will make adequate sense of them at all. Such a choice becomes tragic, as opposed to merely unfortunate or painful, only when it threatens the very coherence of the chooser's life. No matter what he does, he will deeply violate or betray something he values highly. In such cases, an agent's choices may well be incommensurable. We need not look to tragic choices to find incommensurable actions. There may be very different and incommensurable ways of adequately expressing one's valuations of one's ends. Such incommensurabilities do not disable rational choice. If either option makes adequate, but very different sense, of one's valuations, then reason permits the pursuit of either one. Joseph Raz has argued that reason demands that one choose not on the basis of reasons that beat all others, but only on the basis of reasons that are undefeated. Either can be rationally chosen for the distinctive reasons it offers. If reason does not dictate which option to choose, people can nevertheless form a preference for one over the other and avoid perpetual indecision by giving play to motivational states, such as habit and temperament, that do not track or rely on comparative value judgments (Raz 1986, p. 339). The explanatory and pragmatic advantages claimed on behalf of consequentialism are not what they purport to be. No compelling theoretical or practical reasons demand the global maximization of value. Evidence from our actual practices and failures to construct plausible global measures of value suggests that there is no single measure of value valid for all contexts. There are many measures of value valid for different contexts and purposes. The pragmatic theory of comparative value judgments that follows from the expressive view explains these measures better than reductionist theories do. It explains why the pragmatic advantages claimed on behalf of reductionist-consequentialist theories are often not important, because many comparative value judgments are not primarily action-guiding or

64 • Pluralism and Incommensurable Goods

because the ways they guide action, by stimulating imagination, innovation, play, and so forth, are not well served by the calculative model of rational choice. The pragmatic theory also explains why attempts to construct more global measures of value have failed, and hence it explains the phenomena of incommensurability better than consequentialism can. The expressive theory makes only modest demands on our capacity to commensurate states of affairs. It offers more resources for commensurating actions than consequentialism does, thereby enabling rational deliberation to proceed even when confronted with incommensurable goods. The expressive theory can lay a better claim to explanatory and pragmatic advantages, even in the domains where consequentialism claims its greatest strengths.

Self-Understanding, the Hierarchy of Values, and Moral Constraints

4.1 The Test of Self-Understanding In the last two chapters I have shown that the expressive theory of rationality provides a more coherent basis for self-understanding than consequentialism does. It is better at explaining what we care about and what it makes sense to care about, in ways we can reflectively endorse and successfully adopt in practice. I believe this is the fundamental test of a theory of practical reason (compare Taylor 1985d, 1989). To be practically rational just is to make sense of ourselves and our actions on the basis of reasons we can reflectively endorse. Consequentialists have a complex response to this test. They argue that commonsense self-understandings are deeply irrational, inconsistent, or self-defeating. This leads to a familiar dialectic between consequentialism and common sense. Consequentialists bxginjby^ attacjdng the rationality of certain distinctions common sense takes to be fundamental to moral and practical reasoning: distinctions between higher and lower goods, between jloing and allowing, and between intending and foreseeing a consequence _of ones action. These distinctions are applied in norms that prohibit people from trading higher for lower goods in some contexts but not others and that prohibit them from intentionally sacrificing certain higher goods to preserve more of the same kind from sacrifice by others or by natural causes. Consequentialists regard such norms as irrational because they prescribe different final ends, realize different consequences, and accept different tradeoffs of the same goods in different contexts. Norms embodying these distinctions appear not to be maximizing value, but rather to be inconsistent and self-defeating. Advocates of common sense respond by noting that these distinctions are deeply entrenched in our lives and that we feel repugnance for conse-

66 •

Self-Understanding

quentialist systems that refuse to acknowledge them. This provokes two consequential responses. One is to dismiss commonsense intuitions as having no normative authority without some further grounding. Alternatively, consequentialists attempt to show how, on a sophisticated analysis of remote effects, some simulacrum of the original intuitive distinctions can be preserved in practice after all, so the gap between consequentialist prescriptions and ordinary intuitions is narrowed. In this chapter, I argue that the apparent irrationality of commonsense norms and distinctions is due to the fact that_consequoitialismJacks the conceptual resources to represent the evaluative point of thesejdistinctions. But there is a pluralist-expressive rationale for them. This adds a spin to the usual dialectic by undermining the dismissive consequentialist response. It forces consequentialists on the path of accommodation, where they stand on weaker ground, given their need to take their lead from intuitions that have no natural roots in consequentialist thought. The kinds of rationales they can offer for such intuitions must rely on obscure calculations and distant, highly contingent improbabilities. The expressive theory turns out to be superior to consequentialism because it makes better sense of evaluative distinctions fundamental to a meaningful life. At this point, consequentialists take issue with the test of self-understanding itself. They argue that the truth of a theory of practical reason must be distinguished from its adequacy as a basis for self-understanding. The true theory may tell us that the practically rational thing to do is to believe a false theory of practical reason. I will argue that this is a desperate move for a theory of practical reason to take. A theory that can avoid making it has a clear advantage. 4.2 The Hierarchy of Values In commonsense ethical debate, people often speak of some goods as incomparably higher in worth than others. Money, commodities, conveniences, luxuries, and sensual pleasures represent paradigmatic lower goods. They are seen not simply as less valuable, but as not even comparable on the same high scales as those on which paradigmatic higher goods such as human life, friendship, freedom, and human rights are measured. Pluralist-expressive and consequentialist theories can be compared according to how well they make sense of these commonsense intuitions about such "hierarchical incommensurability." Consequentialists attempt to represent such claims about relative worth in terms of preference rankings over states of affairs. Because consequen-

Self-Understanding



61

tialism gives people the end of maximizing value, any difference in relative values should be reflected in a difference in preferences. Yet attempts to interpret hierarchical incommensurability in terms of preferences lead to absurd or catastrophic recommendations or fail to capture anything like the same class of judgments intuitively made about it. Lexical preference orderings offer the most popular way of representing a hierarchy of goods. A person lexically prefers C to B if she is unwilling to give up any amount of C for any amount of B. Lexical orderings prohibit tradeoffs of one good against another. Rawls s claim that rights to basic liberties must not be traded off against greater economic benefits, Dworkin's claim that rights "trump" claims to advance aggregate social welfare, and Nozicks claim that individual rights operate as "side-constraints" on others' actions can all be represented by lexical rankings of rights over other goods (Rawls 1971, p. 61; Nozick 1974, pp. 29-33; Dworkin 1977, p. xi). Lexical orderings sometimes capture liberal practices concerning rights. In First Amendment adjudication, no amount of prospective psychic distress suffered by witnesses to a political rally can justify censoring its message, however vicious. The right to free speech is lexically ordered over the avoidance of psychic distress from exposure to it. Lexical orderings, however, fail to capture many aspects of liberal practice. Liberal states do not spend unlimited amounts of money on the law enforcement measures needed to protect rights. Other interests, such as economic security, can justify some marginal sacrifices of effective power to exercise one's rights. Poor countries reasonably spend less money than wealthy countries on protecting civil liberties. Although human life is said to be of incomparably higher worth than money, saving money on small cars and cheaper roads, which are less safe than alternatives, is reasonable. It would be absurd, even disastrous, to prohibit such tradeoffs. Modifications of the lexical ordering scheme fail to escape these problems. Rawls proposes that lexical preferences of rights over economic development should take effect only if a country has reached a threshold of prosperity (1971, p. 152). Nozick entertains a loophole permitting the violation of individual rights to avoid catastrophes (1974, p. 30n). Neither proposal accounts for the reasonable tradeoffs of effective rights and safety against other policy goals routinely made by prosperous liberal regimes. Discontinuous preferences offer another way to represent the difference between higher and lower values. C might be thought to be higher than B if no amount of B could make one indifferent to less than some quantity of C. John Stuart Mill used discontinuity as a criterion for higher pleasures: one pleasure is higher than another only if people who have

68 •

Self-Understanding

experienced both "would not resign it [give it up entirely] for any quantity of the other pleasure" (1979, p. 8). Mill's test marks a difference in quality but not a difference in the rank of goods. Many goods are such that a person would not prefer to give up either for any amount of the other. I feel this way about the pleasures of reading fiction and of eating chocolate. Even if we constrained the test to pairs of goods ordered asymmetrically by it, it would not capture anything like our ordinary judgments concerning higher and lower goods. Some people are happy to give up the pleasures of listening to LP records for the pleasures of listening to compact disks and are unwilling to accept the converse. But this does not make the latter listening pleasure a higher good than the former, in the sense this has when goods like convenience are compared with human rights, justice, and friendship. Furthermore, convenience is sometimes rationally preferred to such higher goods as justice. Correcting small injustices, such as being shortchanged a quarter at a hot dog stand, isn't worth the court's time. So courts reasonably impose a threshold of damages a case must claim before they will hear it. Other representations of a hierarchical difference in the values of goods in terms of preference have been explored and rightly rejected by James Griffin (1986, ch. 5). For example, preferences for higher goods might be non-optional, while preferences for lower goods are optional. We are not permitted to ignore human life, but it doesn't matter if we don't care about sushi. This is true, but it doesn't prevent us from comparing the relative worth of marginal amounts of the two goods. There seems to be no way to capture any rationally supportable jdaims about a hierarchy of values in terms of preferences. Consequentialists often conclude that the claim that one value is incomparably higher than another means only that it is much more valuable than the other, perhaps also that the speaker is squeamish about specifying precisely how much more. Perhaps a human life is worth several million dollars, but its monetary worth is still finite. Since we could not reasonably spend the entire GNP to save a life, it is merely a pious myth to claim that life is priceless (Melmck 1990, pp. 24, 35-36; Russell 1990, pp. 17, 21). If a claim to hierarchical incommensurability is supposed to support a norm prohibiting all tradeoffs between a higher and a lower good, the results are absurd or catastrophic. If the claim is supposed to support a norm sometimes prohibiting and sometimes permitting such tradeoffs, it is inconsistent. This consequentialist conclusion is repugnant to common sense. Common sense does appear to express judgments that one good is incomparably higher in worth than another, through norms that prohibit trade-

Self-Understanding



69

offs of higher for lower goods in certain contexts, but not others. For example, commonsense morality categorically prohibits gravely humiliating a person just for fun. A consequentialist would have to permit such a humiliation, if enough people got enough fun from it to outweigh the humiliation of the victim. Appeals to discontinuity (claims that no amount of fun can be equal in value to a given avoidance of humiliation) cannot prevent this conclusion. Considered just as an event in his life, the humiliation might be judged by the victim to be less bad than the pain he suffered from a broken leg. Many people reasonably judge their own fun to be worth some pain to themselves—certainly enough, over hundreds of hours of enjoyment, to exceed the pain of a broken leg (consider skiing). If the fun enjoyed in the humiliation is worth more than avoiding the pain of a broken leg, a consequentialist must accept, by transitivity, that it exceeds the value of avoiding the humiliation and hence that the humiliation is justified. The consequentialist focus on the values of states of affairs rather than on the ways we should value people permits this conclusion. Consequentialists resort to many devices to avoid such embarrassing conclusions—rule consequentialism, slippery slopes, remote effects, and so forth. I find these devices to be lame, but I don't think counterarguments at this level of dispute are productive. More telling is the fact that many consequentialists are so confident, without ever exhibiting the calculations, that the numbers accord with commonsense intuitions in these kinds of cases. Notably, consequentialists sometimes willingly exercise their ingenuity in thinking up remote effects to tilt the balance in favor of common sense, when in other cases they deploy their ingenuity (and credulity) in the other direction. In the absence of real calculations, these arguments show only that the costs of the questionable action may be higher than originally thought, not that the costs outweigh the benefits. The confidence and ingenuity that consequentialists display in these cases seem to be motivated not by genuine consequentialist considerations, but by an expressive intuition that deliberately harming an individual in certain ways for certain reasons is wrong, because it expresses an improper contempt for another human being. The expressive intuition masquerades in consequentialist discourse as an intuition into the outcomes of arcane and unexhibited calculations. This_re,CUrxingMneed to appeal to occult quantities to avoid conclusions they are uncomfortable with suggests that consequentialists lack the conceptual resources to represent their own concerns and, therefore, lack the resources to achieve real self-under"stahding (Williams 1973, pp. 100—101). If another theory can transparently make sense of these intuitions, this is strong evidence in its favor.

10 •

Self-Understanding

The pluralist-expressive theory can make sense of intuitions that some goods are incomparably higher in worth than others. There is an air of paradox surrounding such claims: in the same breath, one denies that two goods can be compared and goes on to compare them. The paradox is dispelled by the following expressive analysis. Two goods are incomparable in intrinsic worth if they are not candidates for the same mode of valuation. One good is of incomparably higher worth than another if it is worthy of a higher mode of valuation than the other—worthy of a love, awe, or honor beyond the respect owed to persons in general, worthy of respect beyond the consideration owed to animals, worthy of consideration beyond mere use, worthy of appreciation and care as a superb usevalue of its kind, like finely crafted tools, as opposed to being an acceptable object of abuse, as is a piece of junk that can be destroyed in use and discarded without a second thought. One way of_yaluing something is higher than another if the things^ concerning^ it make^deeper^u^litatively more significant demands on the attitudes, deliberations, and actions of the valuer. And one way to express this difference in demands, Js_ to prohibit tradeoff between.,.states of ; affairs concerning thejgoods^diat express a lower kind of valuation for the higher good than it merits. Such expressive considerations support norms that distinguish among intending, allowing, or unintentionally causing tradeoffs of different goods. Whereas it is not rational to prohibit all preventable tradeoffs between a higher and a lower good, it can be rational to prohibit certain tradeoffs of higher for lower goods chosen for particular reasons or in such a way as to express an inappropriate regard for the higher good. To the extent that causing and allowing, intending and foreseeing the same consequences express different, more or less appropriate, ways of valuing a good, these actions may be evaluated differently. For example, the same tradeoff of friendship against money has a different expressive significance, depending on the reasons people have for acting in ways that result in the tradeoff. It is usually no offense against a friend for one to move to a distant city to obtain a higher paying job, even if the friendship is attenuated because of reduced contact. But it is a betrayal for someone to accept another's offer of money on condition she not see her friend so often (Raz 1986, p. 349). The expressive analysis makes sense of liberal theorists' appeals to lexical orderings. In practice, liberal theorists categorically prohibit only those tradeoffs of rights against other goods that express disrespect for persons (Pildes and Anderson 1990, pp. 2150-2158). The idea that a given tradeoff would express disrespect or other inappropriate regard for persons is often

Self-Understanding

• 71

expressed in the thought that a higher good is at stake in a choice. Freedom of movement is not really at stake in a town's decision to enforce traffic laws for safety; that is, enforcing this tradeoff between effective freedom of movement and safety does not throw into question one's commitment to valuing persons as free. Prohibiting all overseas travel in the name of citizens' safety typically does. The categorical prohibitions required by liberal theory need not lead to disaster or absurdity. If the reasons motivating the tradeoff are serious enough, they need not fall under the prohibition, for they may express in the present context an obligatory valuation of higher goods at stake in the choice. Liberal states recognize the need to impose curfews during a riot. But this emergency restriction of liberty is justified in the name of respecting persons whose lives would be in jeopardy if the usual freedom of movement were permitted. The expressive analysis of hierarchical incommensurability explains other intuitive ideas, such as those of degradation, fetishism, and idolatry, which are difficult to grasp in consequentialist terms. A practice is degrading when it expresses a lower valuation of something than it merits. We speak of slavery as degrading to persons because in slavery, individuals are valued as mere commodities or use-values, rather than as persons, worthy of the higher valuation of respect. Idolatry and fetishism make the opposite error: worshipping or otherwise valuing in a higher way things unworthy of all but lower modes of valuation. Commodity fetishism, a characteristic vice of capitalist societies, consists in a blind devotion to consumer goods, attributing powers and values to them that are properly to be found in relations among people. A consequentialist might argue that different modes of valuation can be reduced to preferences over states of affairs. Valuing something highly merely amounts to valuing it intensely. Valuing something intensely amounts either to wanting more things of that sort around, or to weighing its interests more heavily than things valued less intensely. Neither interpretation is plausible. The first does not account for romantic love, which is directed toward a unique other. It also reflects a confusion, inherent in the instrumental theory, between the ends for the sake of which one acts and the aims of one's action. Respecting humanity as an end becomes, absurdly, a matter of increasing the number of human beings, as the population paradoxes show (§2.3). The second interpretation attempts to reduce qualitative differences in valuing things to quantitative differences in the weights one assigns to their interests. The difference between loving and respecting someone is captured in a preference structure that weights the beloved's interests as

Self-Understanding

12 • Self-Understanding

some multiple of the interests of respected persons. But the relative demands of love and respect cannot be captured by any such acontextual weightings. These demands are embodied in different networks of social relations regulated by discrete norms for different social roles. In private life, one may of course give the interests of the beloved priority over those of strangers. But in public life, as in assigning jobs, one may not weight the interests of the beloved more heavily than the interests of a random applicant. Nepotism is prohibited by the demands of respect, and it is not required by the forms of love compatible with life in modern liberal societies. No single weightedpreference ranking can explain a person's choices across all of her social roles. A consequential might argue that, in principle, she can represent different ways of valuing things in terms of preferences. She would make the rankings of states of affairs relative to the specific factual circumstances that make their being chosen expressive of one mode of valuation rather than another. For example, whenever factual circumstances make a tradeoff of life against money expressive of contempt for life, the consequentialist would represent an individual as preferring life to money. But such a preference ranking would merely record an individual's choices, not ground or justify them. It would be parasitic upon the expressive theory that provides these preferences with a rationale (Pildes and Anderson 1990, pp. 2153-2154). There would be no way to sensibly extend these preference rankings to novel situations apart from interpreting the expressive demands of different modes of valuation. There is no reason to think that occult quantities lie behind and justify the qualitative evaluative distinctions embodied in different forms of life. I can now frame the following challenge to consequentialism: I claim \ that it makes sense to value different goods in different ways and that we have, little idea what human Hfe could be if it did not engage in social practices that supported different ways of valuing things. The pluralistexpressive theory offers a perspicuous understanding of and rational support for differentiated social practices of valuation. It can effectively articulate and justify the concerns we express in common, deeply entrenched intuitive value judgments. Consequentialism cannot do so, at least not with the same clarity and ease. So it faces a dilemma. Either it must find a rationale for these concerns, and for the differentiated practices of valuation that embody them, which neither appeals to occult quantities nor is parasitic upon the pluralistic-expressive theory. Or it must dismiss these concerns as irrational and propose a more appealing way of life than that embodied in anything like our current self-understandings.

• 13

The latter horn of the dilemma is difficult to grasp. Somehow, we would have to imagine a human life without love, without respect, without any of the social practices that embody our highly differentiated attitudes, emotions, and modes of valuation. And somehow, we would have to find it a worthier kind of life to live than our own. Some ancient philosophers, notably the Stoics, argued that we would lead better lives if we learned not to care about goods that were not in our complete control. They recognized that the relentlessly reductionistic value-system they proposed called for the suppression of the emotions (Nussbaum 1992). Modern-day consequentialists are not prepared to defend this austere and chilling view of human life. The former horn appears more tractable to modern sensibilities. But there consequentialism faces the challenge of a competing theory that accounts with ease for profound features of our self-understandings that it is likely to explain only through arcane considerations. 4.3 Agent-Centered Restrictions We have seen that the expressive theory supports practical principles that prohibit certain kinds of tradeoffs between higher and lower goods. Another set of intuitions deeply entrenched in commonsense practices concerns practical principles known as agent-centered restrictions. Xhes£ same principles prohibitxertain.^ kindrjh^y teU_each person that there are^some.OT may not intentionally yiolate a principle or duty^x^ejiJfjioirj^jo could i piewntmoxe violations.of exacdy...the:..same.Mnd..feom,being committed byjDthers. Agent-centered restrictions structure relations of special obligation among friends, family members, professionals and their clients, fellow members of communities, and so forth. A person may not betray his friend to prevent two others from betraying their friends. A doctor may not neglect the health of her patient, a corporate executive whose demise will cause his firm to cease neglecting its workers' health. Commonsense morality views such actions as outrageous, because it gives each person responsibility to care particularly for the others specially related to them. Agent-centered restrictions also apply to relations among strangers. One may not, in general, murder innocents to prevent others from murdering more innocents, although commonsense morality permits certain exceptions in a just war. Advocates of consequentialist morality argue that agent-centered

14 •

Self-Understanding

restrictions are instrumentally irrational, because they identify a bad to be avoided—the destruction of human life, the betrayal of a friend—and then restrict the means people may take to prevent that bad, even if these means are less bad than the bads they prevent. It seems irrational to forbid the minimization of objectionable conduct (Scheffler 1985, p. 409). Consequentialist theories of morality admit as fundamentally justifying only agent-neutral principles, which give every person the common aim of minimizing the bad, no matter who commits it. Derek Parfit goes further, arguing that agent-centered restrictions make commonsense morality directly self-defeating: in following it, we each achieve our moral aims less well than if none of us had followed it (Parfit 1984, pp. 55, 95-98). Particularly troublesome are the special relations of obligation, which give each person a different moral aim, to advance the interests or welfare of his own friends, kin, compatriots, or clients. If a storm causes a boat to capsize, pitching children into a lake, commonsense morality tells parents to try to save their own children's lives, even if it would be easier for them to save strangers' children from drowning. In obeying this principle, parents make lifesaving as a whole more difficult, and fewer children overall will be saved. Their morally given aims of protecting their children are achieved less well than if none of the parents had shown special favor toward their own children. Consequentialists conclude, from their failure to make sense of agentcentered restrictions, that these norms are irrational. The expressive theory attributes inadequacy not to the norms, but to the powers of consequentialism to represent the concerns people express in them. It provides a rationale for agent-centered restrictions by working, as Stephen Darwall (1986) has suggested, "from the inside out." Expressive theories begin not with the external aims or states of affairs a person is to bring about, but with her internal attitudes toward the ends for the sake of which she acts. Expressive theories tell people to adequately express the ways they appropriately value their ends by following norms of attention and response and by governing themselves by norms of action that express these different modes of valuation. These norms, interpreted in the context of their concrete predicaments, tell people to try to bring about certain states of affairs for certain reasons. Expressive theories distinguish between the meaning of an event brought about by an action and its value as a happening. In bringing about states of affairs for certain reasons, people endow these events with a meaning that they would not have if they were just happenings or if they were brought about for other reasons. A slap in the face has a different

Self-Understanding



15

meaning, depending on whether it is administered out of self-protection, anger, or contempt; or by mistake, reflex, or inadvertence. People care about the meanings of actions beyond the values they have when considered just as happenings. An expressive theory of morality may call for people to deplore a state of affairs as an unfortunate happening, but to refrain from certain acts that would prevent it, because those actsf would express an inappropriate meaning or way of valuing things. People want not only that one another's lives be fortunate, but that they be meaningful. Nothing guarantees that these two aspects of life will coincide. People wanL.their welfare to be achieved in the context of meaningful relationships with others. They care about living in loving, respectful, honorable, tolerant, civil, and solidary relations with one another. To live in such relations is to govern ourselves by principles that express respect, love, tolerance, and so forth. Some of these principles are agentcentered restrictions. Consider the rules of etiquette, which express civility. Etiquette tells people not to be rude to others, even if insults will deter others from delivering more insults themselves. Consequentialists complain that more rude actions are perpetrated under this rule than if rudeness were permitted as an instrument to minimize rudeness. But one does not strike a blow for civility by descending to the level of those who hold it in contempt. To suppose otherwise is to reverse the relation between principles and their consequences. It is to suppose that the point of these principles is to minimize conduct which is bad independent of principle. But rudeness is bad because, in violating rules of etiquette, it expresses incivility toward others. The point of the principles of etiquette is to establish and express relations of civility among people, relations which are constituted by conduct governed by agent-centered restrictions. This expressive rationale for agent-centered restrictions requires a noninstrumental understanding of the function of practical principles. It denies that the value of obeying a principle can be reduced to the independently conceived value of its consequences. If this were the case, then it would always be possible in principle to propose a different, agentneutral rule that would bring about more of these consequences more efficiently. This is what prompts consequentialists to accuse commonsense moralists of irrational "rule worship" (Smart 1973, p. 10). From an expressive point of view, however, the value of the consequences of a principle of conduct is extrinsic and derived from its expressive meaning. A consequentialist might object that the etiquette example peculiarly favors the expressive view, because the values realized in etiquette are purely expressive—-they are just a function of recognized obedience to

16 •

Self-Understanding

expressive principles. But friendships and family relations typically have consequences that can be evaluated partly independent of the meanings of the principles that bring them about. Parents have obligations to promote the welfare of their children above and beyond promoting the flourishing of the parent-child relationship. Friends act not just for the sake of preserving their friendship, but to promote one another's welfare. And it is true that obedience to agent-centered restrictions carries some welfare costs that would not be borne by obedience to agent-neutral principles. Parfit designed his example of drowning children to highlight the welfare costs of agent-centered restrictions. But Parfit s claim that these costs show commonsense morality to be self-defeating is mistaken. People care about living meaningful lives, even at some cost to their welfare. Such a thought informs the revulsion provoked in most people by the suggestion that it should be all right to betray one's friends, if doing so could prevent others from betraying more of their friends. The meaningfulness of friendship is pardy constituted by a set of loyalties and feelings that are invulnerable to these kinds of contingencies affecting other friendships. Impersonally fair-weather friends, who stick by their friends only provided that such loyalty does not allow other people to betray their own friends, just do not fulfill the expressive requirements most people demand of true friends. Consequentialists do not do a better job of preserving specially valued relations in requiring people to violate agent-centered restrictions to minimize total violations. They fail to preserve part of what is valued in these relations, which is a kind of integrity constituted by life when it is governed by agent-centered restrictions. That people, in their insistence on preserving such restrictions, show themselves willing to accept the cost of more violations does not indicate that they are irrational. It shows only that they care about realizing certain meanings in their lives which they cannot achieve without making themselves vulnerable to these costs. These costs are not catastrophic. Parfit overlooks the means available to commonsense moralists to reallocate responsibilities in ways that better promote the welfare of their loved ones (Kuflik 1986). Parents already partially delegate responsibility for their children's health care and formal education to professionals. A parent who, as a teacher, spends more time formally educating other people s children than he does his own, who are taught by other professional teachers, violates no special duty to his own children. What people may not do, consistent with commonsense morality, is so homogenize their responsibilities to everyone that they compromise their capacity to realize meaningful relations with others,

Self-Understanding

• 77

expressed through norms that embody special ways of valuing people. This restriction does not prevent people from calling expressive norms into question, when previously unrecognized costs of obeying them are brought to light. The revelation of such costs is evidence that obeying the norms in question may not express appropriate regard for the people who suffer as a result. Such challenges are met by refining our understandings of the demands of love, respect, and other modes of valuation and by more carefully specifying the norms expressing them in ways that suit our greater comprehension of our circumstances. Agent-centered restrictions are rational because in governing our conduct by them, we constitute ourselves as loving, friendly, civil, dignified persons living in specially valued relations. Such lives are neither irrational nor self-defeating. Still, a consequentialist might wonder whether they are the best lives available. Meanings are constituted by conventions, which are subject to change. Wouldn't we all be better off if we could so reform the meanings of respect, benevolence, love, and so forth that agent-neutral principles could adequately express them? Such a reformation would minimize the occasions in which there is a difference between the expressively appropriate action and that which brings about the most fortunate state of affairs. It would minimize the welfare costs of a meaningful life. Peter Railton suggests that respect would be better interpreted as always demanding the deliberate sacrifice of some for the greater welfare of others rather than as prohibiting such tradeoffs (Railton 1984, p. 163n32). In refusing to exploit some for the greater benefit of others, are we not failing to respect these others? An equivalent consequentialist sensibility informs the thought that it is too bad morality prevents us from using the fertilizer manufactured by Nazis from the bodies of concentration camp victims. After all, bygones are bygones, there is good fertilizer here, and can't we reform our meanings to suggest that the use of this fertilizer would be making the best of a bad thing, ensuring that the victims had not died wholly in vain? The proposed reformation of meaning is akin to asking us to speak FORTRAN instead of a natural language. The whole system of meanings in terms of which we understand our lives becomes unhinged. Proposals such as the fertilizer case that make a mockery of someone's sacrifice now are supposed to constitute some kind of honor. The expressive theory can explain why meanings cannot be coherently reformed in the convenient consequentialist direction. We need to value people in higher ways than those in which we value lower goods, such as mere use-values. The principles of appropriate treatment for mere use-

I 78 •

Self-Understanding

Self-Understanding

values are the consequentialist principles of instrumentality, aggregation, indifference to substitution, and so on. Toj^cpress.higher modes of valuation, we must govern ourselves by norms that embody a logic which contrasts with consequentialist logic. Norms expressing higher modes of valuation require two features. First, they must be distributive. A higher valuation must be expressed toward each individual object of the kind that is specially valued, not just toward the aggregate of things of that kind. Higher modes of valuation are respecters of the "separateness" of persons. For example, kindness demands that one be kind to each person taken in herself. Kindness opposes utilitarian "benevolence," which demands that one be cruel to some for the greater pleasure of others. Second, norms expressing a higher mode of valuation must tell us to be willing to sacrifice something of ourselves for the sake of upholding the relation between us and what we value. If meanings were reformed to wholly coincide with what maximizes our convenience and good fortune, there would be nothing left to contrast with lower modes of valuation, in which things are valued only for what they can do for our independently defined interests. The second requirement can be illuminated by recent debates about the prerogatives associated with special relationships. Bernard Williams (1973) has popularized the complaint that utilitarianism attacks one's integrity by demanding that one violate special commitments to others, or to life projects, whenever doing so would preserve more of others' like commitments and projects. Utilitarians, characteristically confusing idealism with egoism, object that Williams calls simply for a kind of self-indulgence in the context of impersonally more important demands. Samuel Scheffler (1982) supports this purported self-indulgence by defending "agent-centered prerogatives" that permit but do not require individuals to forgo the pursuit of the best consequences for the sake of those they specially value. But the utilitarian charge of self-indulgence contradicts their usual charge, which is that people living by agent-centered norms lose more than they gain in doing so. The logic of agent-centered restrictions requires that the occasions of sacrifice be greater than those of supposed self-indulgence. Out of respect or reverence for each special relation, each person is prepared to accept certain misfortunes rather than demand that others violate their own special duties. To value something highly is, in part, to despise the gains in good fortune that can be brought about by its betrayal, to be willing to sacrifice this good fortune for the sake of that which is valued highly.1 To treat everything people value in accordance with consequentialist logic is to abolish higher modes of valuation altogether. It is to call for the

• 79

homogenization of our highly differentiated emotions and attitudes, of love, respect, honor, consideration, and hence for the destruction of most of what makes our lives meaningful. But these forms of life are so important, and so deeply entrenched in our self^-understandings, that it is hard to see how any theory that fails to affirm them can be taken seriously. If consequentialism is to be taken seriously, it must concede their importance and offer its own account of them. 4.4 Hybrid Consequentialism The preceding arguments (§§2.4-2.5, §§4.2-4.3) have established that expres&iygL reasoning must be granted a significant place in a rationally endorsable account of human life. Both expressive and consequentialist ^optimizing) logic seem to be indispensable to practical reason. An adeqmteiHeorymust explain how they are related to each other. The expressive^theory claims that expressive reasoning has global autKority in justifying actions, emotions, attitudes, desires, and norms, while consequentialist calculations play various local roles within it. Consequentialist reasoning is everywhere constrained and explained by the expressive theory, which gives consequentialist reasoning its point (§2.2, §3.2). The burden is now placed on consequentialism to show how it could have global validity while providing a local role for expressive considerations, which can no longer be dismissed as irrational. Some pluralist consequentialists recognize the importance of expressive considerations to human life and take the path of accommodation, attempting to reconstruct meaningfulness, respect, friendship, and so forth within a consequentialist framework (Conly 1985; Gibbard 1986; Railton 1988). Expressive theories of rationality recognize three kinds of consideration relevant to the justification of emotions, attitudes, norms, and actions that are not part of a pure consequentialist theory: that an emotion or attitude is appropriate to its object; that a norm adequately expresses an attitude or emotion; and that an action adequately conforms to a valid expressive norm. On a pure consequentialist view, the value of an emotion, say, is equal to its value as an occurrence, taken by itself (perhaps as pleasurable or painful), plus the expected value of its causal consequences. In §2.5 I argued that human beings cannot coherently adopt this view of emotions because the causal consequences of an emotion depend upon people caring about its expressive relations of appropriateness to its object. But accommodationists could propose a hybrid consequentialist theory, which would allow the three types of expressive consideration to count

#0 •

Self-Understanding

toward the intrinsic value of an emotion, attitude, norm, or action. Such "expressive values" would contribute to the total expected value of an action, norm, or emotion (which would still include pure consequential values such as welfare and pleasure), just like icing adds value to a cake. Hybrid consequentialism takes expressive considerations to contribute to the intrinsic values of emotions, attitudes, norms, desires, and actions, whereas the expressive theory takes them to constitute their rationality. I will argue that it is incoherent to represent expressive considerations as considerations of intrinsic value. This criticism should be distinguished from the common criticism that a consequentialist life would be bad, because it would not be able to support people leading meaningful lives in special relations with others (Williams 1973; Stocker 1976; Kapur 1991). Hybrid consequentialists can justify special relations with others on both intrinsic and instrumental grounds (Railton 1984; Conly 1985; Jackson 1991). My complaint is not that consequentialism gives us bad advice, but that it gives us either incoherent or superfluous advice. The fundamental problem with hybrid consequentialism is that it misrepresents the nature of expressive meaning in attempting to fold it under the concept of intrinsic value. The meaning of an action is constituted by the attitudes it expresses and by its place in the narratives of the lives of the agent and those with whom she interacts. The act could express love or contempt. It could represent a triumph over adversity or the last straw. These are relational features of the act, not intrinsic properties of it. What confers expressive significance on an action is its relations to the other alternatives open to and conceived by the agent, to the ends for the sake of which she acts, and to her past and future, as these are mediated by expressive norms. A given act, such as a mother s feeding her hungry child just a cup of milk, could constitute an act of love if that milk is the last food available, neglect if there is enough food around, callous disfavor if she feeds her other children plenty, or an obsession with the past if it reflects the anxieties of a once-deprived childhood. An act expresses an appropriate meaning to the extent that the agent takes account of its consequences in the right way, according to expressive norms for adequately valuing the people, animals, and things for the sake of which she acts. Consider an agent who attempts to use the hybrid theory to determine which action is justified. The facts about meaningfulness reviewed above pose three problems for a hybrid theory that construes appropriate meaning as adding to but not exhausting the value ("to-be-doneness") of an act. First, it requires the agent to undergo an incoherent double-

Self-Understanding

• 81

counting of the consequences: once, in determining which act would express an appropriate meaning, and again, in determining which act has the greatest overall expected value, counting the 'Value" of meaningfulness along with the consequential values of the act. Because the agent has already accounted for all the relevant consequences of an act by interpreting the norms for expressing the right attitudes toward the people and things she cares about, it doesn't make sense for her to count them again in a consequentialist mode. For example, if her predicament demands kindness to strangers, expressing kindness requires her already to consider their welfare in deciding what to do. This is not a consideration in addition to her interest in expressing kindness. But the hybrid theory suggests that such considerations contribute independent quanta of value to her act. Second, an agent who regarded the meanings of her actions as contributing quanta of value to them, and who then chose the act with maximum expected value, would by this process destroy any meanings she thought she had incorporated into the optimum act. To express love or respect, to act out of friendship or solidarity, is to govern one's deliberations and actions from the internal perspective these attitudes provide, to take account of the consequences of one's actions in accordance with the norms of these modes of valuation. To confer meaning on one's actions is to globally govern oneself from the expressive point of view. To take up the hybrid consequentialist view instead is to take account of the consequences of one s actions in ways external to the norms for expressing one's valuations and, hence, to fail to adequately express one's valuations at all. A rneaningfuljife, cannot be one that views the world from a consequen^jdalist^gerspective, even a hybrid one. Foj^exjumgle, to count respect for persons as just another value of an act that may be traded off against others, ""such as maximizing welfare, is to fail to respect persons at all. One expresses respect for others by governing one's actions and deliberations by certain distributive agent-centered principles, not by maximizing agent-neutral principles (§4.3). Third, the hybrid consequentialist theory provides the agent with "one thought too many" (Williams 1981b, p. 214). It suggests that she should consider respecting or loving a person a good thing; and for that reason, she should respect or love her. But to respect or love someone because it waukLbe a good thing to do is not really to respect or love a person at all. It is to act like a teenager who dates a sweet but dull boy not out of interest in him but out of a pity mistaken for genuine love. Respect and love must be_jmrnediate responses to the person, not the results of cohsFdefatiBiTg d^jjLwQJild be a good thing for them to happen (Stocker 1981).

82 •

Self-Understanding

A consequentialist might argue that the expressive view does not take account of the consequences of responding to people in various ways. Suppose, for example, that expressing one's love for another would lead to disaster, say, by breaking up a marriage and disrupting children's lives. Not just the intrinsic worth of expressing one's attitudes, but the consequences of doing so must be accounted for in determining the rational act. Only hybrid consequentialism, it is claimed, takes account of both. This argument supposes that the expressive theory counsels people to express their attitudes blindly, regardless of the consequences. But the expressive theory tells an individual that adequately expressing one's attitudes is a matter of taking account of the consequences in particular ways and adjusting ones responses with sensitivity to what they mean for everyone whose interests are at stake in one's actions. Respect for those adversely affected by one's love for another may well demand muting or even abandoning one's expressions of love. Such considerations can be comprehended from within the expressive perspective, which needs no help from consequentialism of any global kind to know what it should be doing. Once we understand what it is to endow one's acts with certain meanings, the consequentialist proposal to treat these meanings as intrinsic values to be added to the consequential values of acts appears incoherent. Meaningfulness is not a property actions have in addition to their consequential values, it is a property they have in virtue of the way the agent has taken their consequential values into account. Consequentialists don't see this, because they suppose that the meaning of an act is contained in its motive. An act is kind, loving, or respectful by virtue of its being caused by the motives of kindness, love, or respect. Consequentialists conceive of intrinsically valuable motives as immediate desires for good consequences. If this were true, then the value of a kind act could be analyzed into two separate components: the good it causes and the desire for good that causes it (its "intrinsic expressive value"). Then it might make sense to wonder whether the former consequences might be worth achieving by other motives, even at the cost of losing out on the intrinsic value of the "nice" motives. Thus, classical utilitarians argue that the general welfare is better achieved by harnessing people's self-interest in the marketplace than by relying on their benevolence. Motives that produce welfare indirectly, as an unintended consequence, have more total value, although less intrinsic expressive value, than motives that aim at directly producing welfare. This view depends on the supposition that desires for states of affairs rather than attitudes toward people and things are the vehicles of expres-

Self-Understanding

• 83

sive meaning. But there is no state of affairs the desire for which always adequately expresses kindness, or admiration, or friendship, or any other meaning, regardless of the context (§2.3). For a desire to express a meaning_adequa.tely, it must be governed by expressive norms. These give people different final aims in different contexts and direct their attention immediately to the people and things they care about, and only mediately to the states of affairs they are to bring about. desire to be kind and yet fail to adequately express kindness if she does not jgovern her kind impulses by expressive norms that attune her to the goods at jitake in her action. Such blind activity, however well intended, lacks intrinsic value. Expressive theories demand not that particular desires always motivate us, but that whatever rational desires we have should adequately express or conform to the attitudes toward people and things appropriate to them. If attitudes toward people and not desires for states of affairs are the vehicles of expressive meaning, and if our lives are to be meaningful, then we must adopt a perspective informed by the expressive theory as our global mode of deliberating about and justifying our actions, emotions, and attitudes. We cannot regard meaningfulness as adding some quantum of value to otherwise consequentially evaluable lives. To attempt to regard meaningfulness in this way is to engage in incoherent deliberation and to derail the rational attitudes that give life its meaning and point. A consequentialist could grant this point without giving up her theory. She would interpret the above argument as follows: All of the intrinsically good human lives accessible to us must possess this global intrinsically valuable property of meaningfulness. To endow our lives with this property, we must adopt an expressive point of view for purposes of deliberation and interpersonal justification. Consequentialism is thus a self-effacing theory: it tells us to believe a different theory. But a self-effacing theory could still be true (Parfit 1984, p. 43). It tells us to believe an incompatible theory not because the latter theory is true, but because believing it is instrumentally valuable. Hybrid consequentialists are forced to accept self-effacement because once they grant a foothold for expressive considerations, there is no rational way to prevent them from assuming an increasingly global role in human life. They must govern not only emotions, desires, motives, and norms of action, but beliefs as well. One cannot love someone merely because it would be a good thing to do. The tendency of the expressive point of view to assume a global stature reflects the fact that hybrid consequentialists cannot identify a unit of intrinsic value smaller than a

84 •

Self-Understanding

whole life (Griffin 1986, pp. 34—36), a whole form of life, or even perhaps the whole universe (Slote 1985, p. 103). This fact appears in consequentialist thought as the doctrine of organic values: the claim that the value of a whole may be more or less than the sum of the values of its parts (Moore 1903, p. 28). The doctrine of organic values represents a characteristic consequentialist response to the fact that direct applications of maximizing logic to states of affairs identified as intrinsically valuable lead to unacceptable conclusions. The consequentialist test for intrinsic value is that its bearer has the same normative relevance for choice, regardless of the context in which it appears. When a context can be identified in which bringing about the state of affairs identified as intrinsically valuable would be unacceptable in itself, the consequentialist must assume that the state of affairs must not have been intrinsically valuable after all. It did not carry its value with it, but must have gained its value only in relation to its context. A new intrinsically valuable state of affairs is identified, which consists of the old state of affairs, plus the context that made it appropriate to bring it about. When this more global state of affairs is again seen to be not intrinsically valuable, an even more comprehensive state must be identified as the basic unit of intrinsic value (Korsgaard 1983, pp. 192—193). Consider how this tendency toward organic values works within hybrid consequentialism. Begin with the thought that emotions could be units of intrinsic value. We have seen, however, that we have no way of making sense of emotions except as appropriate or inappropriate responses to their objects (§2.5). The emotion, as well as the context that makes it appropriate, must be the minimum unit of intrinsic expressive value. This represents the first foothold of expressive logic into consequentialist terrain. But it can't be stopped there. For if emotions with their appropriate contexts were intrinsically valuable, then they would be immediately normative for choice or preference. This suggests that it would be instrumentally good to kill people, so that appropriate grief could occur. A consequentialist could respond: surely the actions that bring about appropriate emotions must also be expressively appropriate before we are willing to commend them for bringing about their appropriate emotional responses. But we cannot identify the individual expressive action—an action expressing gratitude or a betrayal—as a basic unit of expressive value or disvalue either. If it were a basic unit of value, then it would make sense to maximize or minimize its occurrence in agent-neutral fashion. But we have seen that acting on agent-neutral principles for the minimization of expressively bad actions does not adequately express meanings

Self-Understanding



85

such as respect and benevolence (§4.3). We also don't determine howmany times I should thank someone for a gift according to some etiquette-independent notion of the declining marginal intrinsic values of successive acts of gratitude. Rather, the norms of etiquette tell us how to express our gratitude. It seems then that the basic unit of expressive intrinsic value cannot be isolated acts, principles of conduct, or their emotional consequences. These are all made appropriate in the context of whole types of relationships constituted by distinctive modes of valuation, such as civility and respect. Are these types of relationships the units of intrinsic value, to be maximized in number, intensity, and duration? It seems not, for how many of each kind of relationship one should foster is constrained by expressive norms internal to these relationships. The meaningfulness of most marriages and certain other kinds of romantic relationships is secured in part through singularity and exclusivity. To divide one's romantic attentions between two or more people is often to betray them all; to commit oneself to more than one marriage at once is to commit oneself to none. Consequentialist, maximizing norms cannot directly regulate the realization of expressive values without disrupting the non-consequentialist norms through which they are realized. It seems as if the basic unit of intrinsic expressive value from a consequentialist point of view can thus be nothing less than a whole form of life, which includes within it a comprehensive non-consequentialist way of evaluating things. "Maximizing" this value simply amounts to living this form of life. There is no fundamental place for the principle of maximizing any intrinsic (context-independent) values of states of affairs within this form of life, because every attempt to supplant expressive norms with consequentialist counterparts ends up garbling the expressive "values" pluralist consequentialists aim to accommodate. The expressive theory explains why consequentialists are driven to this global doctrine of organic values: states of affairs generally lack intrinsic value. Their value is dependent on the expressive context in which they appear. Incorporating the context into a more comprehensive state of affairs is hopeless, for the conditions that make states of affairs Valuable are not other states of affairs, but the people, animals, and things^it makes sense to care directly about. Consequentialists are mistaken from the start in conceiving of expressive meanings as coming in quanta of value, which can then be aggregated or traded off against other values in a calculative, maximizing fashion. Meaningfulness is a global property of social life, realized in living according to expressive norms that have their own, non-

86 •

Self-Understanding

consequentialist logic. Expressive logic assumes an increasingly global role in any rationally endorsable human life because it is only in terms of this logic that we can coherently make any sense of what to do or feel. Consequentialists mistake a form of practical reasoning that sometimes makes local sense within a context set by expressive norms for a form of reasoning that can make global sense of practical life itself. 4.5 A Self-Effacing Theory of Practical Reason? We have seen that a rationally endorsable conception of human life must use expressive considerations to guide attitudes, emotions, motives, preferences, and actions. Any acceptable form of consequentialism must therefore be a hybrid variety. But we have seen that a hybrid consequentialism can use expressive considerations to guide our conative responses only if they guide our beliefs as well. We can endow our lives with meaning only by comprehensively taking up the expressive point of view, reasoning about and evaluating our responses in its terms. The only acceptable form of consequentialism must therefore be self-effacing. Consequentialists attempt to explain the self-effacing tendencies of their theory in terms of the organic nature of the intrinsic values of states of affairs. But whereas consequentialists must take the principle of organic values as a brute fact, the expressive theory explains it in a way that is damning to consequentialism—as an artifact of the merely extrinsic value of most states of affairs. Nevertheless, there is still room for a consequentialist to insist on the truth of her self-effacing theory. She could complain that I have not shown that a holistic account of the values of states of affairs is truly hopeless. Nevertheless, the resort to a self-effacing theory seems desperate. One would have to have strong reasons for accepting a self-effacing theory rather than the theory it tells us to believe, since the latter theory has the advantages of clarity, simplicity, and straightforward application. In §3.1, 1 identified three types of advantages a theory of practical reason could claim for itself: pragmatic, critical, and explanatory. The theory could guide action more successfully, provide us with better methods of self-criticism, or account for the phenomenology of our practical lives better than its rivals. I propose that a self-effacing theory would have to show at least one of these advantages over the theory it tells us to believe, if it is to lay a claim to superiority over it. Consider the relative pragmatic advantages of consequentialist and expressive theories. We have already seen that the purported pragmatic

Self-Understanding

• 87

advantages of consequentialism cannot be sustained when we confront choices among incommensurable goods or goods commensurated by the higher-order good strategy and that these pragmatic advantages sometimes come at the cost of more important goods, such as promoting creativity and growth (§§3.2—3.3). The expressive theory enjoys the latter pragmatic advantages, enables us to make choices among incommensurable goods, and places lighter demands on our capacities to commensurate goods than consequentialism does (§3.4). Once consequentialism moves toward a more holistic conception of value, its purported calculative advantages must be forfeited altogether. So consequentialists face a dilemma. The smaller the unit of intrinsic value they can define, the more they can sustain claims to resolve disputes and justify actions by means of objective calculations. But such atomistic accounts of value lead to intuitively objectionable recommendations. Once commonsense intuitions are given an internally coherent expressive rationale, consequentialists must move to accommodate them. Accommodation pushes consequentialists toward more organic accounts of value and more indirect methods of achieving it. The upshot of this trend is to reduce consequentialism to an inarticulate endorsement of a whole expressive way of life, which is not guided by ultimately consequentialist norms. Perhaps consequentialism can claim an advantage in being able to criticize different expressive ways of life by evaluating the consequences of acting on them. However, once consequentialists resort to a strongly organic theory of value, they lose the articulateness of explicit, objective calculation and can confront commonsense intuitions only with purported intuitions into occult quantitative measures of the values of whole ways of life. And once they resort to a self-effacing theory, we have no reason to be confident that consequentialist intuitions report independent evaluations of the values of alternatives, rather than evaluations that parasitically track the expressive judgments consequentialists themselves admit they ought to believe. How do consequentialists distinguish between intuitions they have because of an accurate consequentialist accounting of intrinsic values and intuitions they have because they believe the expressive theory their own theory tells them to believe? Thus, hybrid consequentialists can endorse the claim that some expressive forms of life are valuable, but they are at a loss to exhibit the calculations or justify the intuitions that declare some to be better than others. Why is it best that the different features of such a life—specific modes of valuation, distinctions between higher and lower values, norms of obligation, motives and emotions—are connected in the way they are? How

88 •

Self-Understanding

might one evaluate the adequacy of a given form of expressive relationship or compare one with the other? The expressive theory provides a rich, articulate account of the relations of the different aspects of expressive modes of life and of what makes them more or less adequate and meaningful. These considerations undermine consequentialist claims to have a critical advantage over expressive theories. Consider, finally, the powers of consequentialism and the expressive theory to explain pervasive and entrenched features of our practical lives—our emotions, attitudes, preferences, internalized and institutionalized norms, values, and so forth. Pure consequentialism appears to have an explanatory edge in accounting for explicitly value-maximizing or optimizing behavior. However, the phenomena to be explained include only cases of local, not global, maximization, and here the pragmatic theory of comparative evaluation can claim advantages of its own. It explains the great variety of commensurating measures and strategies we actually employ in different domains of evaluation. It is not hobbled by the demand inherent in consequentialism to reduce these various measures to a single ultimate measure, a demand difficult to discharge in numerous cases, such as athletic scoring (§3.2). The pluralist-expressive theory explains and justifies a wide variety of other valuational phenomena. It explains the deeply entrenched distinct tions we make between causing and allowing, intending and foreseeing consequences, between higher and lower goods, and appropriate and inappropriate emotions and attitudes (§§2.4—2.5, §4.2). It accounts for judgments of hierarchical incommensurability and judgments involving the concepts of degradation, fetishism, and idolatry. It explains our willingness to trade higher for lower goods in some contexts but not others. It elucidates our practices of social role differentiation, embodying distinct modes of valuation governed by different norms, and liberal practices concerning human rights (§4.2). It accounts for the "silencing" of morally irrelevant considerations in virtuous people, mixed or ambivalent feelings about actions we rightly judge to be justified, and non-action-guiding evaluations of regret. It explains the ways we take the past into account in evaluating future actions and also the persistence and appeal of norms for action and feeling that are intentional, retrospective, distributive, noninstrumental, and agent-centered (§2.4, §4.3). Finally, it explains why and when our capacities to commensurate the values of states of affairs run out and when the phenomena of incommensurability, such as Razian intransitivity, are likely to appear (§3.3). An ingenious and sophisticated consequentialist could explain and

Self-Understanding

• 89

endorse most of the phenomena mentioned above. The difficulties arise when we consider how consequentialists manage to do this. Often they do so by appealing to remote effects, slippery slopes, and strategies of indirection, such as rule consequentialism. These considerations may generate the intuitively endorsed results, but they fail to explain our confidence in and insistence upon them. Why should we be so confident that remote effects and indirect strategies pan out the way consequentialists need in these cases, when we are so uncertain about consequences and strategies in other cases? In other situations, consequentialists derive the "right" results by appealing to occult calculations or mysterious intuitions of brute facts about organic intrinsic values. These look suspiciously parasitic upon value judgments the expressive theory can articulately justify. Frequently, consequentialists manage to endorse our practices, choices, and feelings only in terms difficult to reflectively endorse, because they force repugnant divisions in the self between justification and motivation, truth and what is to be believed, the best motivations to have and the best ones to act on, the ends we should care about and the consequences we should achieve, the reasons we should appeal to and the reasons that actually justify our actions, attitudes, and feelings. Williams points out that these divisions require "some kind of willed forgetting . . . to keep the committed dispositions from being unnerved by instrumental reflection when they are under pressure" (1985, p. 109). It is difficult to maintain our confidence in our responses while reflecting on the consequentialist considerations that justify them, because they represent these responses as mistaken, delusionary, or dependent on distorted representations of values. Expressive theories, in contrast, explain the phenomenology of valuation in terms that exhibit internal, meaningful unities among our emotions, attitudes, motives, internalized norms, intentions, and actions and thereby reinforce our confidence in them (§2.5). Hybrid consequentialism borrows some of the explanatory power of the expressive theory for emotions and attitudes by adopting some of its premises, but only at the cost of making deliberation in its terms incoherent. This problem can be solved by making the theory self-effacing. Once a theory becomes self-effacing, however, it forfeits any claim to offer a superior account of the phenomenology of valuation. The whole point of a self-effacing theory is to concede phenomenology to the theory it tells us to believe, while retaining the claim to truth for itself. Consequentialism, especially in its self-effacing varieties, thus appears to be pragmatically useless and incapable of providing a sound basis for selfunderstanding—that is, an account that makes sense of what we care

90 * Self-Understanding

about in terms we can reflectively endorse. It is a theory of practical reason that cannot be put into practice. But it is unclear what it would be for a theory of practical reason to be true, yet pragmatically and cognitively useless. We do have a relatively clear idea of how to pragmatically justify a theory of practical reason: we test it in action, see if we find living by its lights satisfactory and if it solves puzzles unresolved by other theories, clears up confusions, gives us deeper insight into what we really care about. Assume etiquette is a worthwhile practice. Could a theory of etiquette be true if it offended everyone? The challenge to explain how it could be true puts a burden on consequentialism which need not be shared by expressive theories. The former must give an account of what the truth of a theory of practical reason could consist in that is independent of pragmatic considerations. This would require a hard-core realism about intrinsic values. Even if value realism is true, we have no reason to believe that the world of goods has an essentially consequentialist structure. I have steered clear of ontological disputes in this book. Yet I do think the pluralist-expressive theory has an advantage in not having to rely on such extravagant metaphysical assumptions, for it is consistent with a purely pragmatic account of the validity of theories of value and practical reason. The arguments presented in the last three chapters do not demonstrate that consequentialism is false. With sufficient ingenuity, consequentialists may be able to refine their theory, add a few epicycles, and unleash awesome technical apparatus to repair the disadvantages of their theory, like riveters and welders patching together a rusting, leaky, but still formidable battleship stolidly defending its home waters. But why bother? A lighter, swifter, altogether more seaworthy vessel awaits those bold enough to board it en route to thus far poorly charted territories of the self, which the old battleship can reach only by awkwardly following in its wake.

Criticism, Justification, and Common Sense

5.1 A Pragmatic Account of Objectivity I have characterized the project of determining what is good as a quest for self-understanding. It is an attempt to make sense of our own attitudes and evaluative experiences. If my theory of value is right,thenthe things that are good are the things it makes sense for us to value. The standards of value for things are the standards of rational valuation for us. Ideals are the self-conceptions through which we try to understand ourselves, to make sense of our emotions, attitudes, and concerns. Making sense of ourselves is not a matter of theorizing about an object whose properties we cannot affect. We make ourselves intelligible to ourselves by cultivating attitudes that make sense to us, by determining to act in accord with ideals we accept that have survived critical scrutiny. My account of value thus hinges on its connections to our subjective states and contestable ideals. This raises the concern that there are no objective constraints on what attitudes make sense. A rational attitude theory of value would then have to represent values as merely subjective. Subjectivism is the view that the mere existence of a favorable subjective state taking x as its object (thinking that x is valuable, wanting x, identifying with an ideal that endorses x) makes x valuable to the person in that state.1 If this were so, there would be no room for error or genuine disagreement in value judgments when people know their subjective states. In this chapter I propose a pragmatic account of how we can objectively justify our value judgments, and I defend it against several levels of subjectivist and skeptical criticism. The subjectivist concern could be put like this. I identify the good with the object of a rational "pro-attitude." To have a pro-attitude toward something is to like it. But there are no constraints on what we might

Criticism, Justification, and Common Sense • 93 92 * Criticism, Justification, and Common Sense

rationally like. So, what is valuable is whatever we happen to like. A more challenging criticism would claim that there are no objective constraints on the ideals we could adopt that could rationalize taking up any attitude toward any thing. The choice of ideals is all a matter of taste, about which there can be no rational dispute. The first objection rests on a false analogy between liking and other favorable attitudes. There are almost no constraints on what may be sensibly liked. But there are significant constraints on what can be a sensible object of other modes of valuation, such as love, respect, or admiration. It doesn't make sense to admire musical performances for being sloppy, humdrum, or out of tune. It doesn't make sense to respect people for being servile, immature, petty, or sleazy. It doesn't make sense to romantically love heartless people. Asserting an ideal that endorses these valuings is not enough to convince us otherwise. One must be able to tell a story that makes sense of the ideal, that gives it some compelling point, that shows how the evaluative perspective it defines reveals defects, limitations, or insensitivities in the perspectives that reject these valuings. One might reply that since a person could like almost anything, as long as liking something can make it valuable, my theory still places virtually no constraints on what is good. But my theory does not quite allow that liking something makes it valuable. It says that what is valuable is the object of a rational favorable attitude, not the object of just any favorable attitude. If mere likings are not subject to rational criticism, they are not "rational, but arational. Their objects, therefore, lie only at the margins of the good. Because it can make sense to disapprove of or otherwise disvalue what one likes, objects of mere liking don't count as unqualified goods. People may like what they judge to be junky food, insipid music, or kitschy art. This is consistent with the thought that what is liked usually has some value, however marginal. Mere likings or tastes are distinguished from other attitudes in that they are largely exempt from processes of justification. No one demands that another justify her picking blue, or even chartreuse, as her favorite color. "There is no disputing about tastes" applies only to tastes, and it applies there because it makes sense to have a social practice in which people are allotted emotional space for the cultivation and free play of idiosyncratic valuations exempt from demands for public justification. This social practice is governed by the norm against disputation. Tastes are valuings that are constituted by this norm. What identifies a liking as a mere liking is its relatively complete exemption from justificatory demands, its nearly complete subjectivity.

Although rational evaluative attitudes are partly constituted by social norms that determine their appropriate objects, we still disagree about what attitudes to take toward particular things and persons. Should Pete Rose be admired as a great baseball hero, even though he betrayed the game by betting on it? Here, demands for justification come to the fore, where people offer reasons for and against rival proposals. People s attitudes are rational to the degree that they respond properly to these reasons. The norms of appropriate response are objective to the degree that they are determined by objective practices of justification. What would it be for a claim arising from a process ofjustification to be objectively valid? An objective claim requires two things: the possibility of error or deficiency (the mere fact that one accepts a claim does not make it true or valid) and a basis for free agreement by different people on the same claims. I propose that a process of justification is objective if its participants can reach significant agreement or progress on the matters under discussion when they adhere to norms like the following:2 All participants acknowledge the permanent possibility of a gap between their actual attitudes and judgments and what would be the most rational attitudes and judgments for them to hold. They acknowledge the equal authority of others to offer criticisms and proposals, giving them weight in discussion. For example, they may not dismiss others' criticisms out of hand or bully or belittle the people making them; they must instead offer reasons for rejecting others' proposals and accepting their own. (A group counts a consideration as a reason if its members commonly acknowledge it as counting for or against proposals.) No one capable of participating in justification is excluded from it. Participants must be consistent: they must be willing to apply reasons in the same way to their own and others' proposals. They are committed to making themselves mutually intelligible. This means that they aim for agreement or a common point of view f. and agree to work from common ground (mutually accepted reasons) toward resolution of their disagreements. Finally, the practice contains methods for introducing novel considerations as reasons and for criticizing what participants currently take to be reasons. The practice of justification has been described as a process by which different travelers arrive at a common point of view in the "space of reasons" (Sellars 1963, p. 169). Justification is called for when people endorse different attitudes and judgments and when they have some interest or need to come to agreement. It is possible when people share some common territory in the space of reasons (the considerations each party accepts as counting for or against attitudes and judgments overlap) or

94 • Criticism, Justification, and Common Sense

when they have the capacity to find such common territory. Reasons function like traffic signs in this space, directing people away from defective paths of reasoning toward a point of view all can endorse. The parties to a dispute try to reach a common destination by pointing to those signs that indicate defects in one another's points of view. They try to show that the perspective of those who disagree embodies defects such as inconsistency, ignorance, partiality, confusion, double standards, insensitivity, or pragmatic self-defeat. These are features that the others cannot rationally endorse and thus have reason to eliminate from their perspective. People provisionally call their attitudes and judgments rationally justified and objective when they are reflectively endorsable from a common point of view achieved in such normative discussion. Objectivity can vary by degrees, with respect to both the scope of the justifying community and how well that community lives up to the norms of objectivity. Existing practices ofjustification fall short of this ideal. The norms against bullying and ad hominem attacks and in favor of universal, equal participation have arrived very late in human history and even now are weakly enforced. It is an open question to what degree human communities are capable of achieving objectivity in judgments or whether an objective agreement reached by a local community can be extended to the entire human community, or to all rational beings. The importance of achieving objectivity in judgments about a given subject is also an open question. Liberal theory tends to sharply divide the morally right from the good, reserving objectivity in the strictly universal sense to the former and relegating the good to individual, subjective desires or tastes. Although morality demands objectivity of a wider scope and with more urgency than other values because disagreements here more often lead to violent conflict, it is a mistake to assimilate the good to pure subjectivity. The difference between the right and the good in respect of objectivity is a difference in degrees, not in kind. The good is grounded in communities of valuing, not just in individualistic liking (Walzer 1983). These communities usually should not cover all of humanity. Pluralism implies that different individuals and communities properly aspire to different ideals which need not be ranked in relative worth. The space of reasons is wide enough to accommodate diverse ideals. Why does it make sense to engage in practices ofjustification that have a potential for objectivity? The project of figuring out what is valuable is a project of self-understanding, of making sense of one s own valuings. This cannot be a purely individual project, for the attitudes one has that tran-

Criticism, Justification, and Common Sense • 95

scend mere liking are partly constituted by social norms of appropriateness that inhabit the public space of reasons. One can make sense of one s own attitudes only by taking up a point of view from which others can also make sense of them. To refuse to criticize and justify ones attitudes is to \ withdraw from this space and, consequently, to deprive oneself of the capacity to have and express coherently any attitudes beyond mere liking. Because we do have attitudes that transcend mere liking, we can make sense of ourselves only by participating in practices ofjustification. These practices have the potential for objectivity, since emotional communication, like all meaningful expression, requires a commitment to mutual intelligibility and must make room for the possibility of error and of common agreement. But why do people have reason to engage in normative discussion with others on terms of equality? For most of human history, mutual intelligibility has been achieved mainly by some people's forcing others to accept their attitudes as their own. I am skeptical of proposals that trace the norm of equal and universal participation to the internal logic of communication (Habermas 1975, pt. 3) or to its advantages for reproductive fitness (Gibbard 1990, pp. 76-80). Because this norm is of extremely recent origin (as Habermas 1989 showed), I suggest that its rational appeal be traced to historically contingent practices that make egalitarian social relations for the first time both conceivable and attractive. This account of objectivity raises the possibility of a more sophisticated subjectivism than the one about mere liking. On this view, the subjectivity of value follows from the fact that ideals are essentially contestahle: they inherently invite disagreement (Gallie 1955—1956). They invite disagreement, not just a parting of ways, in that the parties dispute with one another, seeking and so presupposing the possibility of agreement based on the exchange of reasons. But they inherently invite disagreement, suggesting that no rational discussion can settle the issue. If all ideals are essentially contestable, then one might think that no objectivity in value judgments is possible. Consider the prospects for objectively criticizing and justifying the "reverential" and "populist" aesthetic ideals for classical music that were discussed in §1.4. I shall argue that (1) criticizing an ideal requires interpreting its associated attitudes; (2) interpretations can be supported by empirical evidence; and (3) interpretations of attitudes undermine or support their endorsement. A populist might interpret the reverential attitude as an example of the emperors new clothes phenomenon—in truth, those who attend high-class symphonies and operas are bored and stifled, but, fearing

f;i|

96 • Criticism, Justification, and Common Sense

ridicule from their peers and social superiors, they feign appreciation. This can be empirically tested: would one feel relief and seek other musical enjoyments' if those whom one regarded as one's social peers or superiors confessed their own boredom at concerts? This interpretation, if true, is clearly damning. But the highbrow music lover replies that he feels uplifted at concerts, not stifled. The populist attributes this feeling not to intrinsic appreciation of musical merits, but to snobbery—the feeling of superiority one gleans from appropriating aesthetic objects to create social distance between oneself and purported inferiors (Bourdieu 1984). This can also be empirically tested: does the highbrow aesthete abandon his reverence for particular pieces as soon as they become popular among the masses (consider Beethoven's "Moonlight Sonata")? This interpretation is also clearly damning, if true. The highbrow aesthete may reply that his abandonment is due not to popularity but to excessive repetition and exposure, which makes any piece tedious. But then why do he and his peers so insist that the New York Metropolitan Opera stick to the same old repertory that it hasn't staged a new opera for decades? He replies: the musical geniuses of the past are peerless, and we are simply upholding the grand tradition of classical music in the same way those artists whom we revere would have upheld it. The populist interprets this tradition as betrayed by the very reverence thought to uphold it. Mozart, Beethoven, and Chopin were revolutionaries, eagerly embracing new techniques, exploiting the richness of folk music and other popular genres through shameless quotation and outright plagiarism, breaking old conventions and making new ones, composing for instruments of the future. In revering the tradition by fixing it in an imaginary exalted past, highbrow aesthetes merely embalm a tradition whose vitality they have destroyed. Populists hold out the prospect that the dynamism and creativity of the classical tradition could be restored by breaking down the sharp line between highbrow and lowbrow music so carefully cultivated by their rivals. This dialectic illustrates the ineluctable intertwining of interpretation and evaluation inherent in the quest for self-understanding. It can lay a legitimate claim to progress or improvement. For the challenges posed by adherents of the rival ideal, if supported by empirical evidence, cannot be honestly ignored by adherents of the ideal being criticized. They demand a response, which may consist in a change of attitude on the part of those challenged: they recognize features of their attitudes that they cannot reflectively endorse, and they alter these attitudes so they make sense in the context of an enlarged self-understanding. This explanation of valuational change supports the claim that the change constitutes a genuine

Criticism, Justification, and Common Sense •

91

improvement, because it represents the change as clearing up confusions in self-understanding, as achieving greater consistency, as enabling one to carry on one's commitments more fruitfully than before (Taylor 1985b, 1985d). But even if the response involves sticking to one's guns, improvement can be registered in the form of extending one's understanding of what one is committed to in doing so and acting accordingly. Such extensions will affect one's attitudes and practices, just as judicial interpretations of legitimate statutes, in clarifying their implications, tend to make behavior more consistent with the law. Where the best explanation of change of attitudes represents it as progressive, there is a legitimate claim to objectivity. This claim does not require a proof that there exists some end point of agreement, not yet discovered through rational discussion, that will settle all disputes. There is no way to know this independent of participating in the discussion of disputants. Their assumption of potential agreement is supported by every progressive change of attitudes in response to reasons offered by the other side. The continual eruption of yet new disagreements inherent in essentially contestable ideals need not undermine this assumption, as long as each side can register improvements in its own progressively shifting terms along the way. Actual disagreements rarely exhibit the idealized rationality illustrated in the example above. Many vices and psychological obstacles, such as stubbornness, glibness, smugness, stultification, defense mechanisms, and repression, stand in the way of rational self-understanding and change. Debates are often suppressed or distorted through the exploitation of power relations between disputants, which rationalize the trivialization, ignorance, dismissal, or misrepresentation of challenges from the less powerful. Finally, disputants often speak at cross-purposes, appealing to considerations in a part of the space of reasons not within the horizon of those challenged, while at a loss to find ways to move them to a point where they can be recognized as reasons. Eventually, disputants may simply part ways and cease to discuss their differences or to care about them. One could then say that their practices embodied values that were simply different—as opposed to common values whose interpretation is contested—for which there can be better and worse answers objectively valid for both parties. 5.2 The Thick Conceptual Structure of the Space of Reasons To justify an evaluative claim is to appeal to reasons that make sense of particular attitudes toward the evaluated object. To interpret an attitude is

98 • Criticism, Justification, and Common Sense

to represent it as endorsable or not from an adequate evaluative perspective informed by such reasons. Therefore, a great deal hangs on the conceptual structure of the space of reasons. Pluralism maintains that the evaluative concepts by which we express our reasons for valuation and action are fundamentally diverse. They are mostly what Bernard Williams (1985, p. 140) has called "thick" evaluative concepts. The concept "snob" is one such thick concept, employed in a reason for rejecting the reverential ideal of music. Thick concepts apply to particular domains of action and guide particular feelings. They contrast with the "thin" evaluative concepts— "good," "bad," "right," "wrong," and "ought." The thin evaluative concepts are the most general evaluative concepts we have and are applicable to all domains of action and feeling. According to pluralism, the application of the thin concepts is largely determined by evaluative standards expressed in terms of thick concepts. Monists hold that we should bypass the thick concepts and try to directly access what is good, bad, right, and wrong. They think we should do so largely because they are skeptical about the fundamental reason-giving authority of claims expressed in terms of thick concepts. Before examining these skeptical challenges, consider how pluralism accounts for the reason-giving authority of such claims. The distinctive feature of authentic thick evaluative concepts seems to be that they are simultaneously "world-guided" (particular facts must obtain for them to be applied) and "attitude-guiding" (they offer reasons for valuing and acting). In contrast, thin concepts are said to be only action- and attitude-guiding. This contrast may not hold up. For example, "good" is conceptually tied to such concepts as benefit or advantage, which are world-guided (Foot 1978a, 1978b). A significant distinction between thick and thin evaluative concepts can still be drawn if, as I contend, the thin concepts derive their world-guidedness only through their conceptual ties to thick concepts. Some thick concepts describe objects as meeting standards defined in terms of the attitudes they merit: as humiliating, ridiculous, wonderful, deplorable, titillating, fascinating, and so forth. Others describe objects as meeting standards defined without direct reference to attitudes. These include concepts of virtues and vices, such as sincerity, integrity, brutality, and stinginess, and non-moral evaluative concepts of qualities of character, such as being cool, cheeky, macho, independent, witty, and vivacious. That people meet standards defined by such concepts gives us reason to adopt various attitudes toward them, depending on their relation to us: pride, shame, respect, contempt, admiration, ridicule, approval, or disapproval. Concepts of human flourishing

Criticism, Justification, and Common Sense •

99

and diminishment, such as autonomy, dignity, loneliness, and neurosis, guide the expression of various kinds of self- and other-concern, such as pity and respect. Aesthetic evaluative concepts, such as of the beautiful, the goofy, and the quaint, guide such responses as admiring contemplation, disparaging amusement, and nostalgia. Other evaluative concepts refer to institutionalized norms or rules that guide particular attitudes: the rude is that which warrants feelings of offense and indignation; the unjust warrants resentment; the immoral warrants guilt, shame, and outrage. We need a plurality of thick concepts to make sense of the variety of evaluative attitudes we have toward persons and things. Most commentators have focused on the logic of thick concepts to support claims about the cognitive status of value judgments or about the ontology of values (Hare 1952; Foot 1978a; McDowell 1979; Williams 1985; Wiggins 1987a). I set aside these issues and draw attention instead to their reason-giving functions in practices ofjustification. Three features of thick concepts (besides their essential contestability) are important for understanding their role in making sense of our valuations. First, their applications are determined by interpretive processes that employ evaluative reasoning (McDowell 1979; Williams 1985, pp. 141-142). Second, their coherence depends on the social practices and contexts that make their proper attitudes intelligible (Maclntyre 1981, ch. 1). Third, they tend to evolve in reciprocal interaction with their proper attitudes (Wiggins 1987a). The first feature of thick concepts defeats attempts to fix their applications in neutral factual terms that could be determined without employing value judgments. To apply them to new factual circumstances, we must be able to interpret their evaluative point. This demands that we engage in distinctively evaluative reasoning—reasoning which engages judgments about what standards rationally govern our attitudes. Consider how the rules of etiquette apply in the context of changing gender roles (Martin 1989, pp. 304—305). Traditional rules of etiquette encourage men to give women personal compliments. They discourage women from calling attention to their achievements or openly objecting to others' opinions. They also tell businesspeople to avoid personal compliments in a business context, while permitting them to seek credit for their achievements and to frankly express their disagreements with others about business projects. So, if a man compliments a woman on her clothing at the office, is he being gallant or rude? Can a woman succeed in business without being rude? Does women's liberation demand that we disregard etiquette? If etiquette were just a matter of descriptively fixed rules, it

100 • Criticism, Justification, and Common Sense

would be in trouble here. In fact, the application of such rules in novel or apparently conflicting cases is open, pending an interpretation of their underlying evaluative point. Judith Martin argues that the apparent conflict between rules of etiquette is generated by the fact that different rules apply to business and private sociable relations, combined with the sexist and obsolete assumption that women exist only in the latter sphere. In the sphere of private sociability, etiquette permits persons of either gender to compliment others on their clothing and generally to recognize their gender. This suits a function of sociable conversation—to facilitate the development of personal relationships. Social manners also discourage both men and women from aggressively asserting their opinions or calling attention to their achievements in polite conversation, for social occasions are supposed to add to the charm of life, not facilitate competition. In a business context, etiquette teUs people to identify others according to their job, not their gender, and permits them to act competitively. These rules suit the functions of business, where performance is what matters, where gender doesn't matter to the performance of most jobs, and where competitiveness improves performance. On that account, when a man compliments a woman on her clothing in a business context or takes offense at her competitive office behavior, he is mistakenly applying private social manners to her. To do so is to imply that "when ladies are around, serious business is suspended" (Martin 1989, p. 305). This is an insult to women and disrupts their careers and the effective performance of the business as a whole. Etiquette therefore demands that men and women alike be treated by the rules of business etiquette in the business world and by the rules of social etiquette in social life. (This leaves open the possibility that many gendered rules of social etiquette can be internally criticized as demeaning to women.) Martin's interpretation of the demands of etiquette (the extension of the thick concepts "polite" and "rude") constitutes a justification of the norms she recommends because it makes sense of our attitudes. We make sense of our attitudes when we meet various pragmatic demands: when, by means of our self-understandings, we articulate, cultivate, and refine our attitudes in satisfactory ways, when we overcome confusion and contradiction, orient ourselves consistently and successfully to our world, and do so in a way that withstands a reflective understanding of how we manage it. The activity of making sense of our attitudes by articulating the reasons for them or the underlying evaluative points toward which they are oriented guides our attitudes in ways more coherent and focused than before. Having accepted Martin's arguments, we are in a position to

Criticism, Justification, and Common Sense •

101

engage in sociable and business interactions with persons of both genders with greater confidence and success and less ambivalence and confusion about what we are doing and how we should feel about it. The refinement of attitudes that comes from reflection on thick concepts in justification sets the stage for further extensions of these concepts, often in radically new directions. Thick concepts and attitudes evolve in reciprocal interaction through history, in such a way that an extension of a thick concept that did not make sense at one point in time may make sense at another. This interaction is mediated by social practices that provide the background conditions for the coherence of the attitudes expressed in them. Thus, the application of the thick concepts "rude" and "polite" was once regarded as inescapably gendered in all social contexts. Even today, women who assert their opinions or claims on men—that is, women who act as equals with men—are judged more harshly by conventions of etiquette than men who behave the same way. But modern sensibilities about etiquette are evolving under the pressure not only of the moral norm of equal respect for all, but of concrete social practices that enable women to participate as juridical equals in marriage, politics, and the workplace. The gendered thick concept of social order that demands female submissiveness to male authority is being supplanted by more egalitarian conceptions of social order which make sense only against the background of the social practices that embody them. Such practices enable people to experience their social world as successfully ordered through more egalitarian norms and to cultivate sensibilities which help justify the norms that make sense of them. In the absence of such practices, people who rejected gendered norms of etiquette would be left with a sense of vertigo; they would be at a loss as to how to conduct themselves. Rejection in such a context would not make sense of people's attitudes or successfully guide their conduct. The properties singled out by an egalitarian etiquette as calling for offense or indignation depend on the cultivation of a certain egalitarian sensibility. There is no way to identify the trajectory of an egalitarian sensibility for etiquette without assuming its own perspective, for this sensibility is responsive to culturally specific meanings internal to the particular social practices embodying it. Attitudes and their associated thick concepts evolve in reciprocal interaction and, when they inform a practice with vitality, tend toward ever-greater articulation, differentiation, and refinement. Thick concepts, because of their open-endedness and essential contestability, have a dynamic and generative character which enables us to envision new possibilities for living.

102 • Criticism, Justification, and Common Sense

These features of thick concepts reflect the need to preserve their attitude-guiding functions. In interpreting the underlying evaluative point of such concepts, people seek to make sense of the valuations at stake in their disputes by characterizing their appropriate objects. The descriptive content of thick concepts cannot, therefore, be determined independent of a process of justification that engages our understandings o( our attitudes. Judgments expressed in terms of thick concepts give us reasons for valuing and action to the degree that their scope is guided by a reflectively endorsable understanding of our concerns. Because our concerns and attitudes often make sense only against a background of socially contingent and historically evolving social practices and conditions, our evaluative concepts evolve in concert with changing social circumstances and offer opportunities for divergent interpretations in the face of social conflict. Some theorists claim that thick concepts are inherently incapable of providing the terms in which authentic reasons and intrinsic value judgments can be framed. Authentic reasons and value judgments must motivate anyone who sincerely avows them. But people can accept claims expressed in terms of thick evaluative concepts without being motivated to follow them. This reasoning motivates the demand to bypass thick concepts and directly access the thin evaluative concepts (such as "good" and "right"). Only claims expressed in terms of the thin concepts are thought to have the inherent link to motivation that qualifies them as authentic. R. M. Hare (1981, pp. 72-75, 21-22) uses this argument to justify a monistic theory of value, according to which non-moral intrinsic value judgments are simply expressions of personal preference. Hare's motivational requirement is unreasonable. For something to count as an authentic value judgment or reason, it must be reflectively endorsable. But actual motivational states are not always reflectively endorsable. One of the functions of value judgments is to note when one's motivational states are deficient because they fail to track what one judges to be good. Boredom, weariness, apathy, self-contempt, despair, and other motivational states can make a person fail to desire what she judges to be good or desire what she judges to be bad (Stocker 1979). This prevents the identification of value judgments with expressions of actual desires and preferences, as Hare insists. The rational attitude theory of value says that to judge that something is good is to judge that it makes sense for someone to value it. This makes intrinsic value judgments at least six times removed from actual first-order desires for the apparent good. First, they are immediately

Criticism, Justification, and Common Sense •

103

normative for attitudes, not desires. Second, they usually take persons and things, not states of affairs, as their objects. A person's judgment that a historic coal mine rescue was an excellent deed may inspire her awe for the rescuers, but what must it make her want to do? Third, they say only what a person's attitudes ought to be, not what they actually are. A depressed person could judge that her accomplishments merit pride, but be incapable of rousing herself to feeling it. Fourth, even when they inspire the right attitude, that attitude's motivational effects may depart from its rational demands. Appropriate guilt may induce paralysis rather than desires to make amends for any wrongdoing. Fifth, they may express impersonal judgments of value, not personal judgments of importance. A person could think it would be a very good thing if the couple next door kissed and made up, but she could also think it is none of her business to do anything about it. Finally, even intrinsic value judgments of high personal importance can still rationally leave a person wide latitude to indulge in caprice, impulsiveness, and sheer bad taste. A poet may judge that her dedication to writing fine poems is good and important. But this needn't prevent her from whimsically trying her hand at Hallmark greeting card doggerel. Nor does her choice commit her to the judgment that these mawkish ditties are any good or that it is good to write them. Thus, no evaluative considerations necessarily motivate choice, for there can always be a gap between what one judges to be valuable and what one finds oneself actually caring about at a given time. Engaging in objective discussion with others is one of the ways we try to get our attitudes in line with what makes sense. Value and importance judgments framed in terms of thick concepts give people reasons for valuation because they provide the sensibility conditions for different ways of caring about things. Take away norms expressed in terms of thick concepts, permit only reasoning in terms of a homogeneous "good," and one wouldn't know whether it made sense to admire, honor, love, or merely like the object in question. Emotions, feelings, and cares would be reduced to a uniform, inarticulate blur if we were deprived of the thick concepts by which we delineate different kinds of goods. The application of thin evaluative concepts to the world therefore depends upon the outcome of discussions by which people try to make sense of their attitudes through the exchange of reasons. There is no hope of identifying what is good (right, wrong, and so on) or of defining a comprehensive, empirically determinate standard of goodness in terms that completely avoid thick concepts (Hurley 1989).

I 104 * Criticism, Justification, and Common Sense

5.3 How Common Sense Can Be Self-Critical My pragmatic account of justification takes as its starting point the commonsense evaluative intuitions that constitute the space of reasons for a community at any given time. An intuition is an opinion endorsed under conditions favorable to sound judgment—when one is reflective, calm, coherent, informed, and responsive to others' perspectives. Most commonsense intuitions are expressed in terms of thick concepts such as "kindness" "friendship," and "dignity." But philosophers have often attacked the use of intuitions and thick concepts to justify evaluative claims. Many worry that reliance on intuitions traps people into following judgments informed by superstition, prejudice, cultural bias, and obsolete practices. The intuitions of a racist are untrustworthy but supposedly impervious to criticism from an intuitive point of view. Intuitive thinking, incapable of critical self-reflection and tied to the status quo of received opinion, secures the smug in their prejudices, the hidebound in their habits, and the oppressed in their lowly positions (Singer 1974, p. 516; Brandt 1979, p. 21; Hare 1981, p. 76). Only a non-intuitive form of critical reasoning can rescue peopte from these failures, by giving them a standpoint independent of their social practices. Such critical thinking would bypass the culturally contingent thick concepts embodied in intuitions and would reason directly about the good and the rightjusing-onfe logic and value-neutral scientific facts (Hare 1981, ch. 1; Brandt 1979, pp. 22, 1990). This argument is the basis for consequential claims of having a critical advantage over pluralist theories such as mine that are grounded in social practices (§3.1). I contend that all the genuine critical practices that make sense can be included, in intuitive reasoning. Commonsense critical practices can objectively endorse the intuitions they employ because they already contain methods for criticizing what people take to be reasons and for introducing novel reasons in normative discussions. These practices, or ordinary extensions of them, provide all the reasons we have to reject or refine old intuitions and create new ones. They can meet all the demands for objective justification that it makes sense to care about. Lets recall some of the conditions for objective justification (§5.1). Justification is a responseJo criticism, complaint^an^ ariseTlrrine context of conversation among people who aim to reach some common point of view, and it is addressed to those who disagree. It is pointless to engage in justification when the parties have no interest in reaching agreement, when there is no concrete complaint, or when there

Criticism, Justification, and Common Sense * 105

is no common ground from which to begin a dialogue. Common ground could consist in shared intuitions or in curiosity, trust, and a willingness to try alien practices. Common ground determines the starting point of justification. We try to reason or explore from it to a new position that Resolves the disagreement (Kawls 1971, pp. 580-582). Consider three_j^ay^ in which one person might criticize another's evaluative claim. She_cauld challenge the importance of the__other\ reasonsi injayorofhisjudgment; she co apply to. the case; or she could challenge the authenticity of these reasons. To support a disagreement of the first type, a person must offer an overriding reason for judging differently. For example, in opposing the construction of a new intensive care unit, a hospital administrator may argue that it is more important to devote the resources the new unit would require to the prenatal care clinic. Reasons like this appeal to comparative value judgments of the kind discussed in §3.2 and §3.4. To support a disagreement of the second type, a person must offer inter^retimr^som for thinking that the facts don't support the first party's claims. Against Sharon's complaint that Mark owes her payment for a loan, Mark could offer evidence that both of them understood her transfer to be a gift when it took place. People invoke interpretive reasons when they try to extend the application of a thick concept by interpreting its evaluative point (§5.2). To support a disagreement of the third type, a person must offer undermining reasons^^ against the authenticity of the first person's, values. That is, she must show that they don't make sense^that they don't reflect or support anything worth caring about.3 The point of view from which they seem to make sense is shown, from a more objective point of view, to be confused, limited, or founded on error. This section will vindicate the use of intuitions and thick concepts in critical thinking by showing how they can generate undermining reasons. When we inquire into the authenticity of values, what we wonder, generally, is whether it makes sense to value them for the reasons they purport to offer. No plausible account of making sense comes close to offering its sufficient conditions. But no such account is needed. We should ask no more of ethics than of science. Science provides no test that guarantees the veracity of its starting points. It is enough that it provides means for detecting and correcting errors and for introducing superior theories, concepts, and methods. Commonsense evaluative practices provide similar means. They offer a catalogue of critical strategies that generate undermining reasons and enable expansions of the space of reasons. More critical strategies exist than are listed here, and more could be

\ \\

106 * Criticism, Justification, and Common Sense

invented. It is reasonable for a person to think that a value is.authentic when it seems to make sense to her and when it surviyes^he^gamuLof critical strategies that can be launched against it in discussions-gavexnediy the norms of objectivity (§5.1). Critical strategies can be roughly divided into three classes. Internal or "ethnocentric" strategies rely only on commonsense shared intuitions, armchair reflection, and ordinary observation. Scientific strategies draw upon empirical knowledge obtained through scientific investigations. Experiential and persuasive strategies enable people to grasp novel intuitions. Most worries about the conservatism of theories that appeal to intuitions result from the mistaken view that such theories can accept only strategies of the first type. Some intuitive theorists such as Walzer (1987) and Rorty (1989) accept the ethnocentric constraint. Their position, though needlessly narrow, has the merit of demonstrating how rich are the internal intuitive resources for criticism. Consider three such resources: internal coherence testing, narrow reflective equilibrium, and idealistic self-criticism. Thick concepts can be tested for internal coherence and found to be irresolvably unclear. Or analysis could reveal that a purported thick concept cannot simultaneously perform its reason-giving and descriptive functions. John Stuart Mill (1977) used this strategy to undermine the use of nature as an evaluative standard. He showed that the concept of nature was deeply equivocal. Any interpretation of "nature" that had descriptive content had no normative force (for example, the natural as the usual). And any interpretation of "nature" that seemed to have normative force reduced to some other value (for example, the natural as the functional). In the method of narrow reflective equilibrium people attempt to organize their intuitions into a coherent, consistent, systematic whole (Daniels 1979, p. 258). Critical development of their views works through exploiting the tensions and contradictions they find between the general evaluative principles they accept and their intuitions about particular cases. This strategy is driven by a desire for consistency and a sharper, more effectively action-guiding articulation of principles. Narrow reflective equilibrium can provide reasons that undermine intuitions about principles. One such reason could be that we can't find any particular cases in which they offer more sensible guidance than rival, simpler principles. Narrow reflective equilibrium can also provide reasons that undermine intuitions about particular cases. An example could be that we can't discover any evaluative point expressible in a principle that endorses the particular intuition. Reflective equilibrium does not merely offer a

Criticism, Justification, and Common Sense •

107

strategy for constructing a consistent input-output decision-making mechanism: as the etiquette case illustrated in §5.2, reflective equilibrium accepts and rejects intuitions according to their capacity to express or promote some intelligible evaluative point. Walzer's (1983, 1987) strategy of idealistic self-criticism moves beyond armchair reflection to a study of social practices. Its materials for criticism are the gaps between the actual practices of a community and the ideals by which a community justifies them. Ideals always stand at some distance from their supposed embodiments. This allows us to criticize the practices, institutions, or persons attempting to realize them by articulating their demands more adequately. Although the ideal of democracy in the United States is partly constituted by a conception of such social institutions as elections and representative assemblies, it is not exhausted by their present forms. It provides reasons for thinking that democracy would be better realized through reforms. For example, public financing of elections would reinforce the democratic principle that popular support, as opposed to the support of well-financed special interests, be the effective determinant of who gets elected. An interest in integrity motivates this critical method. It is not simply a matter of adjusting practices to fixed principles. Meanings can be implicit in practices which people haven't articulated at the level of principles. If they can articulate new ideals or principles which better account for practices they find fulfilling, then the practices offer grounds for accepting the principles. By accepting them, people can engage in their practices more self-consciously and effectively than before. In other cases, accepting some principles that purport to account for our practices might make them go less well than before. This would provide powerful evidence that there is something wrong with our practices, our principles, or both (Taylor 1985d). Criticism does not stop at interpreting the demands of intuitions and practices. It can also undermine the factual beliefs underlying them by drawing upon scientific knowledge. This is how people discover that their intuitions are founded upon prejudice, superstition, cultural bias, and other cognitive distortions. There are at least six ways science can be used to undermine intuitions. First, science can show that a factual concept used for normative purposes is radically at odds with causal knowledge. This was shown for the teleological conception of nature needed to sustain Aristotle's theory of the good and for the concept of race as a biological category which is needed to justify certain racist practices. This kind of criticism is especially important for undermining claims about instrumental goods and bads, for example, the notion of witchcraft as an

i

Criticism, Justification, and Common Sense • W8

109

* Criticism, Justification, and Common Sense

instrumental evil. Second, science can provide evidence that an ideal cannot come close to being realized, and so is merely Utopian. This need not undermine the ideal's authenticity, since the options open to human beings just might be miserable. But it does undermine attempts to generate action-guiding norms from the value, because such norms would be futile. Defenders of capitalism employ Utopian criticism against democratic socialism when they argue that there is no "third way" between capitalist democracy and totalitarian communism (Hayek 1944). Third, science can show that there are viable alternatives to practices that are justified on grounds of necessity or are thought to reflect an inevitable framework of thought. Anthropologists have exposed alternatives to social practices based on a bi-modal concept of gender; and radically different conceptions of masculinity and femininity than those structuring Western gender roles (Ortner and Whitehead 1981). Fourth, historical reflection can undermine an intuitively accepted norm by showing that it has lost its function or point. Early-rising was a virtue when nearly everyone was a farmer and the productive use of daylight hours was a condition of responsible farm management. It is obsolete for people in urban settings, who can fulfill their responsibilities at other times. Fifth, social theory can show that the background social conditions needed to make sense of a thick concept do not exist. Social practices may not support the application of normative distinctions that once made sense. Surviving legal documents from early medieval Europe invoke the rich vocabulary of Roman law. But historians discovered that early medieval institutions of property and contract were too primitive to support full-blooded applications of Roman thick legal concepts. The words survived their meaningful uses and functioned as little more than magical phrases conferring legal authority to contracts actually enforcing simpler obligations than the words in their Roman context would suggest (Cheyette 1978). Finally, genealogical criticism can expose the incoherence of a value by showing that vicious or self-deceptive motivations are required for its genesis, evolution, and endorsement. The purported reasons for supporting the suspect value are masks for attitudes their own adherents cannot reflectively endorse. Nietzsche (1969) used genealogical criticism to attack Christian morality. While Christians and moralists claim to support morality as an expression of universal love, it in fact expresses ressentiment against noble, powerful, vital people. Morality involves a pragmatic contradiction, for the only motive that can move people to embrace it is one that morality must condemn. Genealogical criticism of different

forms also underlies the critical methods of psychoanalysis and consciousness-raismg (Ricoeur 1970; Fay 1987; MacKinnon 1989). A final class of critical strategies appeals to experiences whose most illuminating or compelling descriptions invoke alien intuitions. Some philosophers have suggested that a person's evaluative experiences are mere creatures of the intuitions she already accepts (Harman 1977, ch. 1). If an individual has the intuition that persons of color are inferior and thus need not be respected, then she will experience disrespect toward them without outrage, horror, remorse, or other emotions that embody a contrary intuition. This empirical claim is false. People often experience events in evaluative terms that are at odds with their intuitions.4 Such experiences, if not accounted for in terms that enable a person to discount their putative claim on herself, pose a challenge to her evaluative perspective. They can cause crises whose rational resolutions require the creation of a new evaluative perspective that does justice to the experience. Crises can be brought about by factors other than anomalous experiences. Practices may fall into crisis, as new circumstances and experiences render them incapable of performing their functions or make their participants lose confidence in their evaluative point. They may cease to provide a useful map of the practical landscape. Moral and political theories that relied on the idea of a hierarchical order of beings, with God and the King at the top, nobles and clerics next, and different ranks of commoners at the bottom, each inferior subject to the face-to-face authority of some superior, had to break down once classes of people arose who were "masterless"—such as vagabonds, who, wandering the roads, had no immediate superior. These people couldn't be fit into the old political map, which could find no norms to govern them. Liberalism provided a new account of legitimate political order, appealing to the thick concept "consent," which was designed to accommodate "masterless men" (Herzog 1989, ch. 2). In testing political regimes against the consent of the governed rather than against conceptions of the cosmic order, liberalism introduced new intuitions and thick concepts into political debates and undermined old ones. This innovation can be justified by the fact that it provided a perspective which explained why the older system fell into crisis, and that it enabled people to resolve or dissolve the crisis, while successfully performing the practical tasks demanded of it.5 A justified change to a new normative perspective need not be motivated only by crisis. Sometimes persuasion is sufficient. One person or culture can present a new perspective in an especially appealing way, opening up possibilities never before imagined. People commonly

M

1

110 • Criticism, Justification, and Common Sense

change their aesthetic intuitions in this way. They are invited by others to see if they can experience and appreciate a work of art in the terms proposed by them. Rational persuasion need not operate by direct appeal to the thick ethical concepts endorsed by an agent. Typically, the persuading agent must establish her authority with those she wishes to persuade. She may do so by joining the community she wishes to persuade. Having gained the members' trust, she may convince them to try the norms she endorses, but they don't. Or a stranger may manifest extraordinary qualities of courage, mercy, or charisma recognized by others who do not share her intuitions. Her possession of these exemplary qualities may give them reason to credit her perspective with authority, for they may view her admirable qualities as signs of the worthiness of her perspective. John Stuart Mill's defense of equality in marriage uses a persuasive strategy (1975). While he defended marital equality on grounds of justice and the welfare of women, Mill also sketched an appealing picture of an ideal of marriage as a friendship between equals. This ideal was alien to much of his audience. Yet Mill rightly believed that exposure to such marriages could exert a powerful attractive influence. Unequal marriages in which wives are treated merely as servants are notorious for their emotional sterility. Wouldn't men who reflected upon the contrast between their own lives and the richer, more fulfilling lives of self-confident men living in more equal marriages feel a tinge of envy toward them and perhaps even of self-contempt in recognizing that their own sense of self-esteem is staked upon bullying domination? If persuasion did not work for men set in their ways, it held promise for their sons, who knew their fathers' failures all too well and had the flexibility and ambition to seek something better. This catalogue of critical strategies is not exhaustive. But it is sufficient to put to rest the objection that intuitive thinking is incapable of critical self-reflection. None of these critical strategies requires wholesale rejection of appeals to intuitions. They all work through people's common sense, intuitions, and experiences. Even the scientific strategies rely on second-order intuitions about the reasonableness of intuitions. This is true of all the critical strategies thus far put to use outside arcane philosophical contexts. The indeterminacies and tensions in intuitive thinking, combined with changing social circumstances, personal experience, and scientific knowledge, provide people with ample reasons and materials for selfcriticism. They have reason to find a new perspective superior to their old one if it articulates the concerns they were inarticulate about, resolves

Criticism, Justification, and Common Sense •

111

their contradictions, clears up their confusions, enables them to determinately apply them to their old predicaments (or to new ones they confront), allows them to settle practical conflicts that were unresolvable under the old perspective, or simply permits them to lead their lives in terms they find more compelling (Taylor 1985b). The critical strategies outlined above enable people to find such superior perspectives, without resort to heroic attempts to transcend ordinary modes of reasoning. If an intuition is defective because of cultural bias or other factors, people can discover this fact through ordinary means of investigation. My view of criticism might suggest that I accept the coherentist, antifoundationalist account ofjustification known as "wide^eflective equilibrium." According to this view, a judgment is justified to a person if it is part of a coherent, reflectively stable system of beliefs she holds, including (a)~ intuitions about.particular cases; (b) intuitions about general principles; and (c) various background scientific and ideal theories, including theories of the person, of moral development and human motivation, of social order, rational choice, and so forth (Daniels 1979). I resist identifying my position with theories of wide reflective equilibrium. Wide reflective equilibrium demands that justified evaluative judgments form part of a theoretical system, but it isn't evident that our evaluative intuitions can or ought to be systematized into theories in the way supposed (Baier 1989; Noble 1989). More important, theories of wide reflective equilibrium usually fail to think through the implications of the social character of justification. Justification is concerned with making sense of our concerns and attitudes. But rational attitudes are essentially constituted by social norms the authority of which can be established only in dialogue with others. A person may be in personal wide reflective equilibrium but know that his attitudes are poorly developed as a result of inexperience, defective character, neuroses, or other problems. These facts give him reason to distrust the deliverances of his own attitudes and judgment and to trust the intuitions of more experienced, wise, reflective, and virtuous people. Every person has reason to take seriously the judgments of others just from the fact that any individual's own point of view, no matter how reflective and informed, is still limited by his personal biography and particularity. In emphasizing the availability of methods for learning from and persuading others, I avoid the charges leveled against coherence accounts that they are merely subjective or give us no way to adjudicate disputes between incompatible but internally coherent systems (Singer 1974, p. 494; Hare 1976, p. 82; Brandt 1979, p. 22; Copp 1984, p. 161). Discussion, persua-

112 * Criticism, Justification, and Common Sense

sion, and interaction can provide new conceptual resources for mutual understanding and adjudication of disputes. 5.4 Why We Should Ignore Skeptical Challenges to Common Sense The conception of justification in which thick concepts and commonsense intuitions function does not satisfy certain skeptical doubts. On the view I defend, justification involves meeting specific intelligible complaints and criticisms by means of arguments that begin from some common starting point. There is no need to justify the entire framework of justification or to justify the starting point in the absence of evidence that the point in question involves some specific ethical or cognitive error. Although any particular intuition or thick concept can be intelligibly criticized, it makes no sense to criticize the whole lot at once, for the only way we can frame an intelligible criticism is in terms of some intuitions and thick concepts whose authenticity must provisionally be presupposed. Skeptical critics of intuitions reject intuitive claims in the absence of independent reasons to accept their authenticity. Some moral theorists used to believe that intuitions report observations about an independently existing, nonnatural realm of values. But few can accept the extravagant platonic metaphysics needed to sustain this foundationalist account of justification. Alternatively, one could take a coherentist approach to justification and argue that intuitions are among the beliefs with which any satisfactory evaluative system must cohere. But for this to be the case, we must have a reason for granting them some initial credibility. Lacking an account of the authority or credibility of intuitions, they have no probative value at all (Brandt 1979, pp. 20-22; Hare 1981, p. 76). This criticism of intuitions is sometimes expressed in the claim that the point of justification is to answer the skeptic. To justify intuitions to a skeptic, an account of justification must explain how intuitions could be justified in themselves. But the best a coherentist account ofjustification can do is explain how intuitions can be justified relative to a person's beliefs and concerns, which themselves may be faulty. No appeal to intuitions can justify evaluative claims in themselves without begging the question against the skeptic (Copp 1984, pp. 142-143, 147-149). Critics of thick concepts draw two normative inferences from these arguments. First, lacking some answer to the skeptic, we must prefer skepticism to the use of intuitions in evaluative argument (Brandt 1979, p. 3). Second, we should search for some way to reason about the good and the right which

Criticism, Justification, and Common Sense •

113

bypasses intuitions and thick concepts. Only by finding a direct route to the right and the good that is independent of evaluative intuitions can we justify value judgments in themselves (Brandt 1979; Hare 1981). The need to respond to the skeptic motivates a reductive, thin account of values. I will argue that it doesn't make sense to care about meeting skeptical standards ofjustification. This response is unlikely to disarm philosophers driven by skeptical doubt. So I also offer a diagnosis of the motivations that lead philosophers to entertain skeptical doubts. I argue that such doubts are incoherently expressed in skepticism toward evaluative intuitions. These motivations are tied to misleading disanalogies between science and ethics, as well as to intuitions about the normative authority of science which are inconsistent with its official skepticism toward all intuitions. We have no reason to take skeptical challenges seriously, because their practical implications are absurd. Brandt claims that if no independent grounding can be given to our intuitions, then we should prefer skepticism. Like most critics of intuitions, Brandt confines his skepticism to moral intuitions—intuitions about moral right and wrong and perhaps also about a persons good. But none of the skeptical arguments against intuitions hangs on any supposed peculiarity of moral intuitions. If they work against moral intuitions, then they work against all evaluative intuitions, including intuitions about rationality, good grammar, good arguments, and good scientific experiments.6 Should we cease to speak grammatically, or to correct one another's grammar, because we have no account, independent of our intuitions about grammar, of what it is to be grammatically correct? This is absurd. We have no way of making sense to one another apart from following the rules of grammar. It might be suggested that the concept of making sense allows us to draw a. distinction between intuitions about grammatical and epistemic values and intuitions about other kinds of value, for the latter are not needed to make sense of anything. This is a grave error. j\s^rgue.d.abov^ W£ji£ecWalue judgments to make sense of many of our basic emotions, cares^con£erns, and practices. The moral skeptic is in a position to argue that we would be better off without guilt. Perhaps morality is bad for us. But moral skepticism makes sense only against a background of intuitions about other goods, such as human flourishing, health, or perfection. The skeptical arguments employed by the critics of intuitions cannot stop at morality. If they work, they work against all evaluative intuitions. They amount to the recommendation that, in the absence of some transcendent justification, we should cease to care about things in any of the ways that

114 * Criticism, Justification, and Common Sense

involve intuitive value judgments. We must cease to admire or aspire to be anything and cease all the practices that embody value judgments. This kind of life is conceivable. Fish and birds lead such lives. But it is absurd to claim that humans must live like birds and fish if they can't "justify" living like humans. In any event, the skeptic has nowhere to stand in deriving normative implications from his skeptical claims. To assert that skepticism is to be preferred to any value judgments is itself to rely on a normative intuition that somehow escapes the skeptical demand for justification. To this, the skeptic might reply that he is merely making a theoretical point, that no account is forthcoming of how intuitions can justify evaluative claims in themselves, apart from any relation they may have to any person's beliefs and concerns. This retreat from practice to theory in discussions of justification is incoherent. Justification is an inherently normative concept directed to what claims we ought to accept. If it doesn't make sense to adjust one's beliefs according to their relation to some standard, the standard doesn't count as a criterion of justification at all. The skeptic might respond: but surely it makes sense to care about whether one's values are right or wrong in themselves! After all, being justified relative to some set of beliefs and concerns isn't satisfactory if the beliefs and concerns are mistaken. This claim is not strictly true. Some mistakes are harmless or inconsequential. More important, the skeptic must come up with a notion of what it would be to be mistaken, or to fail to make sense, which is not discoverable by means of any of the critical strategies outlined above or by any analogous strategy that makes use of evaluative intuitions. On the pragmatic view of justification defended here, all our evidence for the soundness of value judgments expressed in intuitions is contained in the following kinds of facts: that such judgments express what we actually find to be valuable; that they successfully orient our lives, actions, and feelings, providing them with points we can reflectively endorse; and that they survive the kinds of criticisms sketched above. If our intuitions enable us to overcome frustration, confusion, irresolvable conflict, irresolution, and similar pragmatic defects, why should we purge them from our lives? The skeptical demand seems to be for some ontological underwriting of our intuitions, some demonstration of how they can track a realm of objective normative facts or "values in themselves" that can be characterized independent of our concerns. It is difficult to see how the success or failure of intuitions to track some realm of facts characterized independent of our interests and concerns could or should matter to us. Being valuable

Criticism, Justification, and Common Sense • 115

is just a matter of meeting standards that it makes sense to care about; making sense, in turn, is just a matter of achieving a perspicuous selfunderstanding justified in pragmatic terms. The skeptic, then, has no argument that we should care about meeting his standards and, hence, no argument that his standards constitute authentic demands of justification. But skeptical arguments against intuitions are so popular that some diagnosis of their appeal should be made. All the above-mentioned skeptics of evaluative intuitions believe that science provides a model of how to justify claims in themselves. Their skepticism about the justificatory power of evaluative intuitions is derived from a supposed contrast with the justificatory power and authority of science. Specifically, they are struck by a presumed disanalogy between observation statements in science and particular intuitions in ethics. We appear to have an account of how observation statements can provide evidence for theoretical claims about a world that exists and operates independent of our concerns. But particular ethical intuitions cannot provide an analogous kind of evidence for evaluative claims without presupposing an extravagant platonic metaphysics. So the kind of justification available to science is not available to ethics (Harman 1977, ch. 1). Skepticism about evaluative intuitions is a way of expressing reverence for science in conjunction with a normative intuition that any practices that command comparable reverence must exhibit the same structure of transcendent justification supposedly available to science. This line of thought undermines itself in at least two ways. First, the relevant analogy to evaluative intuitions in ethics is not observation statements in science, but its evaluative standards of evidence, method, and argument. What evidence do we have that our norms of evidentiary relations and scientific method are authentic? We know only that they successfully guide the construction of theories that realize various epistemic values which make sense to us, such as predictive power, simplicity, fruitfulness, and coherence. We have no account either of how these epistemic values track values-in-themselves or of how the realization of these values enables us to track truth-in-itself. The interpretation of these epistemic values is also as essentially contestable as those in any other domain (Kuhn 1977). Justification in science depends upon evaluative intuitions that are on a par with the evaluative intuitions we follow in any other practice. Second, the skeptical attitude that reflects a reverence for science depends upon an intuition about norms for reverence that stands in need of the same justifications as any other. In fact, the reasons for our rever-

I 116 * Criticism, Justification, and Common Sense

ence for science—its satisfaction of aspirations toward mastery over nature, autonomy, and mature, objective understanding, unsullied by childish superstition, wishful thinking, and slavish obedience to authority—are themselves dependent upon acceptance of the normative authority of claims expressed in thick concepts (Taylor 1985a, pp. 235, 244). The skeptical demand is incoherent, because it makes the warrant for holding science in awe dependent upon the very norms it calls into question.

6 • Monistic Theories of Value

6.1 Monism Mojriists contend that the good is fundamentally unitary. Because goods can be understood in two ways—as meeting various evaluative standards or as the proper objects of various favorable attitudes—monism can be of two types. Monists could hold that the apparently diverse evaluative standards we use can be explained in terms of their relation to a single, unitary, good-constituting property. Or they could hold that all goods are the proper objects of a single favorable attitude (§1.1). I have argued that the fundanaental plurality-of authentic evaluative standards is grounded in our need to differentiate the plurality of evaluative attitudes we have toward things. This suggests that attempts to reduce the normative authority of plural standards to a single good-constituting property ultimately depend on the assumption that just one attitude or response lies behind all evaluative claims. This suggestion is confirmed by hedonism and rational desire theory, the two most important naturalistic versions of monism. Hedonists claim that the good is what is pleasant. An evaluative standard is authentic if and only if it tracks our pleasurable responses to things. Rational desire theorists claim that the good is what is rationally desired. An evaluative standard is authentic if and only if it tracks what we would rationally desire. These theories make pleasure, or desire, the sole value-signifying response. (I say that a response signifies value if its appropriate object is valuable.) Monism also comes in nonnaturalistic versions, such as Moore's theory. Moore (1903) held that the good is whatever bears a simple, nonnatural property, "good." The apparent diversity of intrinsic values—beauty, friendship, knowledge—is unified by the fact that they all share this common good-constituting property. I will argue in §6.2 that nonnatural-

118 * Monistic Theories of Value

istic monism ultimately relies on the assumption that admiring contemplation is the only value-signifying attitude. • Monism represents an attempt to bypass the complex, socially grounded, essentially contestable evaluative standards defined by thick concepts so as to reason directly in terms of the thin, general concepts "good" and "right." The single good-constituting property is supposed to be characterized and knowable independent of thickly described, plural evaluative standards. The single value-signifying attitude is supposed to be something individuals can take up independent of relating to others in particular social contexts. Thus, the "rationality" of a person's desires in rational desire theory may depend only on conditions, such as being calm and fully informed, that are supposedly specifiable independent of thickly described social contexts. (It does not depend on their being endorsable in dialogue with others who relate to one another as equals.) Pluralists deny that it makes sense to reason about the good and the right independent of thick evaluative concepts. When monists try to do so, they either abolish discursive reasoning about values altogether, or confine it to an arbitrarily narrow set of considerations. This has disastrous implications. In adopting a theory of value, we adopt a way of understanding and appreciating what is worthwhile in life and of exploring new possibilities for living. Monism drastically impoverishes these possibilities. It disables us from appreciating many authentic values. It suppresses the parallel evolution of evaluative distinctions and sensibilities that make us capable of caring about a rich variety of things in different ways (§5.2). It cuts off fruitful avenues of exploration and criticism available on a pluralistic self-understanding. Adopting a theory of value can have these effects, because our capacities to experience and realize values partly depend on our understandings of why and how things are valuable. If we do not distinguish between liking and loving someone, then we will be incapable of a fully adequate love, which requires some self-reflective awareness of its distinctive demands. Evaluative experience also requires an appropriate social context. If we lack shared beliefs in the distinctive values of certain goods, we will not be able to sustain the social practices needed to make sense of different ways of valuing them. Certain ways of appreciating music would be unavailable if our only exposure to music were in stores and elevators. Music could then be valued as a stimulant to spending or as a vaguely pleasant "atmosphere," but not as something intrinsically worthy of appreciation and study. We have two ways of testing monistic theories against pluralist theories

Monistic Theories of Value •

119

in experience. First, we can see whether all and only the things the theory identifies as good by its criteria are things we can value on reflection. A theory of value must describe a point of view, partly characterized by a set of circumstances one could occupy, from which one is favorably disposed toward all and only the things the theory identifies as good. We must be able to reflectively endorse this point of view as authoritative, as one we rationally adopt in determining what is fundamentally good. Call this the reflective self-endorsement test. It would be difficult to accept a theory of value that refused to submit to this test. If one were favorably disposed toward something only under circumstances in which one distrusted one's attitudes, this would surely ground doubts about whether it would make sense to value it. And if one could find no reason to distrust the perspective from which one values something, one would have no reason to think it not authentically valuable. The second way we can test theories in experience is to see whether they can account for the full range and variety of our evaluative experiences and practices. For example, a theory should make sense of our practices of discursive reasoning about values. Alternatively, certain practices may no longer make sense to us on reflection. A theory of value, therefore, must either make intelligible and endorsable the full range of our evaluative experiences and practices or persuade us that it would make sense to do without those it cannot make sense of. Call this the test of practical

understanding.

I will argue that monism fails both tests of experience. There are three main versions of monism: Moore's nonnaturalism; hedonism; and rational desire theory. Whatever version it assumes, monism excludes things we reflectively find good, includes things we reflectively find bad, or fails to account for distinctive features of our evaluative experiences and of the practices that make sense of them. 6.2 Moore's Aesthetic Monism According to G. E. Moore, "good" refers to a simple, undefmable, nonnatural property. Everything that is good has this property. Moore's view has been attacked for relying on a mysterious and extravagant metaphysics (Mackie 1977, pp. 38—42). What could a nonnatural property be, and how can anyone detect it? I prefer to leave aside these metaphysical concerns and instead consider whether Moore's theory can explain deeply entrenched features of our actual practices and experiences. Consider the reflective self-endorsement test. Moore believed that only

120 • Monistic Theories of Value

personal relations, beauty, and knowledge are great intrinsic goods. He excluded meaningful work, athletic achievement, justice, and freedom from the list of intrinsic goods. These exclusions do not agree with many peoples reflective judgments. Moore's neglect of the goods of activity and of impersonal social relations reflects the limitations of his test of intrinsic value. According to Moore, something is intrinsically valuable if and only if one judges that a world in which it exists in isolation is good (1903, p. 187). To make such judgments, Moore and his followers removed themselves from active engagements in the larger world, withdrew to private spaces in the company of intimate friends, and introspectively contemplated the isolated objects of their imaginations. It is not surprising that many goods were not salient to people in such a privileged, exclusive aristocratic setting, insulated from experiences of work and practical activity with strangers. The demand that intrinsic goods be valued in isolation from their social context, through undisturbed contemplation, mirrors the norms of appreciation for objects in a museum. Thus, Moore's test eflfectively restricted intrinsic goods to the states and proper objects of an aesthetic attitude of admiring, passionate contemplation and private communion (Keynes 1949, p. 83, 96). John Maynard Keynes, the famous economist and onetime member of Moore's circle, articulated a devastating pluralistic critique of Moore's monistic aestheticism: The ways in which states of mind can be valuable, and the objects of them, are more various, and also much richer, than we allowed for. I fancy we used in old days to get round the rich variety of experience by expanding illegitimately the field of aesthetic appreciation . . . classifying as aesthetic experience what is really human experience and somehow sterilizing it by this mis-classification. (1949, p. 103) The same sterility accompanies any theory of value that identifies the good with just erne state of mind or its objects. Moore's monism is additionally handicapped by the fact that his valuesignifying aesthetic attitude was insulated from rational criticism. Moore resisted the demands of justification because he supported a radical individualism, maintaining that an individuals value judgments are accountable to no one else (Regan 1986, pp. xii, 129). This view fails the test of practical understanding. It fails to explain our practices of resolving evaluative disagreements by offering reasons for our judgments. People do not merely assert that something is good; they point to features it has and to standards it meets that support the claim that it is good. If "good" were a

Monistic Theories of Value •

121

simple, immediately accessible property, discerned through an ineffable aesthetic attitude, there would be no room for argument and appeal to evidence and facts to resolve disagreements. Either one would "see" that something is good, or one wouldn't. Value-blindness, like color-blindness, cannot be cured through discursive reasoning. Moore's circle of friends, the Bloomsbury group, adopted his theory of value. Notoriously, they did not offer reasons for their value judgments. Keynes brilliantly described evaluative discourse in Moore's company: Victory was with those who could speak with the greatest appearance of clear, undoubting conviction and could best use the accents of infallibility. Moore . . . was a great master of this method—greeting one's remarks with a gasp of incredulity—Do you really think that, an expression of face as if to hear such a thing said reduced him to a state of wonder verging on imbecility, with his mouth wide open and wagging his head in the negative so violently that his hair shook. "Oh!" he would say, goggling at you as if either you or he must be mad; and no reply was possible. Strachey s methods were different; grim silence as if such a dreadful observation was beyond comment and the less said about it the better . . . [Woolf] was better at producing the effect that it was useless to argue with him than at crushing you . . . In practice it was a kind of combat in which strength of character was really much more valuable than subtlety of mind. (Keynes 1949, pp. 85, 88) Alasdair Maclntyre (1981, ch. 3) has argued that the social practice of Moorean value realism is indistinguishable from the social practice of crude emotivism, in which value judgments are identified with raw assertions of personal preference and influence others through browbeating and emotional manipulation. Keynes s account shows that Moore sustained consensus on his value judgments through the sheer force of his personality. The social practices upheld by Moorean monism, incapable of grounding a publicly accessible distinction between impersonal value judgments and idiosyncratic preferences, cannot sustain significant practices of reason-giving. They therefore cannot sustain any attitudes beyond mere liking (§5.1). Moore's theory effectively identifies the good with the objects of a peculiar aesthetic liking. Moore's theory would not fail the test of practical understanding if he could show that practices of reason-giving don't make sense and should be abandoned. Moore's aestheticism might have supported his distrust of discursive reasoning about the good. Call a good aesthetic if its worth depends primarily on what it is like to experience it. Then no third party description of it can conclusively support a claim to its value. One must actually experience it for oneself to tell whether it is really good. This

122 • Monistic Theories of Value

shows the insufficiency of reasoning uninformed by direct experience of what a good is like. It does not show that ineffable experiences are the only basis for aesthetic judgment. If this were the case, we could not write art, music, and movie reviews. Moore might also have thought that his principle of organic value undermined our ability to reach correct value judgments through discursive reasoning. The principle states that the value of a whole is not equal to the sum of values of its parts. A whole consisting of two bads (crime and punishment) may be better than a whole consisting of one of these bads and a good (crime and enjoyment of its fruits). It might be thought to follow that the discursive citation of the good and bad elements of a whole could never provide conclusive grounds for an evaluation of the whole. There is no such thing as a deductive proof of an organic value judgment. But this is just to say that there is no monological method for demonstrating value judgments that can replace or track the dialectic of public justification. Discursive reasoning retains an indispensable role in organic value judgments. One must articulate a reason for thinking that two bads are better than one of the bads and a good: perhaps the latter whole is worse because the good is undeserved. A modification of Moore's view might support discursive reasoning about values. Perhaps Moore goes wrong not in advocating monism, but in thinking that "good" is a property immediately accessible to discerning people. Perhaps it is better to construe "good" as a theoretical property that supervenes upon empirical properties of things. Every difference in value is grounded in a difference in natural properties, and our only access to value differences is mediated by knowledge of natural differences (Falk 1986). Practices of reason-giving could then be viewed as akin to scientific practices of offering observational evidence for unobservable, theoretical entities such as electrons. This analogy is invalid. We are entitled to infer that electrons are flowing through a circuit when we observe the lamp light up because electrons flowing through lamp filaments cause them to glow. But the nonnatural property "good" does not cause the gymnast to execute an exceptionally graceful and daring back-flip on the balance beam. We are not entitled to infer from the grace and daring of the back-flip that there is some single nonnatural property "good" which it has. Offering a reason for judging something good is not the same as inferring its cause. In any event, it seems that no other properties need exist to support the claim that the back-flip is good, other than its grace and daring, and the relation these properties have to our evaluative sensibilities and norms for making

Monistic Theories of Value •

123

sense. The nonnatural property "good" is dispensable in accounting for how we know things are valuable. Certain features of aesthetic discourse might offer another way to salvage Moorean monism. Perhaps "good" is a property directly accessible to observation, as beauty is thought to be, but one's evaluative attention must be turned in the right direction to notice where "good" occurs. Reasongiving could be understood as a way of pointing out good-bearing features of a thing that passed unnoticed by those who disagree. Just as people must be trained to listen to music, to notice complex rhythms and echoes of earlier themes, before they can fully appreciate a musical composition, so in evaluative matters people s attention must generally be drawn to the natural properties of things that bear the nonnatural property of goodness. This story might account for appeals to overriding reasons in evaluative discourse, but not to undermining reasons. On Moore's account, if one judges something good one supposedly thinks that one detects a simple, nonnatural property it has. An undermining reason would have to be seen as a consideration which shows that this apparent detection of the property "good" was either an illusion or a detection of a different property that resembled "good" but was not identical with it. It is difficult to fit the catalogue of undermining reasons discussed in §5.3 under either account. It is difficult to make sense of the thought that there are properties (simple? nonnatural?) that resemble "good" but are not identical with it, and it is difficult to make sense of value illusion without an account of what veridical evaluative perception comes to. Aristotle argued that human practices do not make use of Plato's transcendent Form of the Good and that we have no idea how they could be improved by knowledge of it (Aristotle 1985, 1097a4-14). The case is no different today with respect to Moore's simple, nonnatural "good." Worse, people who adopted a Moorean self-understanding would not be able to sustain intelligible, non-manipulative practices of discursive reasoning about the good. The deterioration of these practices would stunt the processes of articulation, interpretation, and discrimination among thick concepts essential to the development of our capacities of discernment and valuation, which in turn enrich our lives. Given the failure of nonnaturalistic monism, monists had better pin their hopes on naturalism. 6.3 Hedonism Consider hedonism, the doctrine that pleasure is the sole intrinsic good.1 Hedonism is appealing because it appears to satisfy several seemingly

124 * Monistic Theories of Value

compelling philosophical demands. One is that the theory of value support a determinate rational decision procedure in ethics. Hedonism is the almost inevitable product of this quest once it is interpreted as the search for a dominant end (Rawls 1971, pp. 552-557). A second reason for the appeal of hedonism is its seemingly easy reconciliation of ethics to a scientific worldview. Values appear to be mysterious, but if they can be reduced to pleasure, and pleasure reduced to some physiological or behavioral phenomena, such as neuron firings, endorphin concentrations, or persistent motivations, then ethics can be scientifically grounded. Finally, hedonism represents one interpretation of the compelling thought that intrinsic values are intimately connected with favorable consciousness. Let s consider this last thought in more detail. I have argued that something is valuable only if it makes sense for someone to value it, to take_up a favorable attitude toward it. A favorable attitude toward something includes the disposition to be favorably aroused by it. So my theory also acknowledges an intimate connection between value and favorable consciousness. Hedonism departs from pluralism in three crucial respects: (1) in ascribing intrinsic value only to favorable consciousness and not to its object; (2) in supposing that all pleasures exist independent of the realization of non-hedonic intrinsic goods; and (3) in allowing only one kind of favorable response toward objects—being pleased—to signify the presence of intrinsic value. Classical hedonists identify the intrinsically good with favorable consciousness. Pluralists identify it with the object of favorable consciousness. On the classical hedonist view, external objects and objective relations are good only instrumentally, to the degree that they produce pleasurable states of consciousness in us. Three accounts of pleasure are consistent with the classical view. First, pleasure could refer to intrinsically liked physical sensations that do not refer to anything beyond themselves, such as the pleasures of an orgasm or of a full stomach. It is universally agreed that we find many things to be valuable which do not cause, except incidentally, such physical sensations—such as life in relations of equality with others, meaningful work, and outstanding athletic accomplishment (which is often physically painful). Second, pleasure could refer to psychic feelings with no particular object, such as being in a good mood. A person could be in high spirits without feeling good about anything in particular. This view of pleasure is as impoverished as the first. It suggests that the ideal life would be that of a drug addict on a perpetual high, permanently absorbed in his own states of consciousness. To any self-respecting person,

Monistic Theories of Value •

125

and to anyone who cares about other people, this solipsistic view of the good is too degraded, passive, and meaningless to take seriously. Most hedonists take a third view, broadening the intrinsically good states of consciousness to include pleasurable emotions, feelings that refer to something beyond themselves, such as being pleased with boating, friends, and conversation. What is really valuable here, the objects of favorable response—boating, friends, conversation—or just the favorable response itself? Classical hedonists support the latter view. They do not deny that people rationally desire objective states other than pleasurable consciousness. People rationally desire such things as real knowledge of truths, real accomplishments, and genuinely faithful friends, as opposed to the corresponding false beliefs and illusory experiences of having these things. But such desires are rational only because they are instrumentally valuable for maximizing pleasure. One-cannat -obtain the-pleasuTes of real accomglishment without aiming at accomplishment.....for its own sake. Sidgwick, a sophisticated hedonist, appealed to the reflective self-endorsement test in his support: To me at least it seems clear after reflection that these objective relations of the conscious subject, when distinguished from the consciousness accompanying and resulting from them, are not ultimately and intrinsically desirable . . . [when] we "sit down in a cool hour," we can only justify to ourselves the importance that we attach to any of these objects by considering its conduciveness, in one way or another, to the happiness of sentient beings. (1981, pp. 400-401)2 Today the philosophical consensus has turned against Sidgwick. Some have appealed to exotic thought experiments to refute him: Would you willingly step into an "experience machine" that would induce in you fictional but incorrigible experiences of everything you like (Nozick 1974, pp. 42-45)? Would you be indifferent between the existence of a beautiful and an ugly world, even if no one would ever experience either (Moore 1903, pp. 83—84)? I think the error of hedonism can be discovered in ordinary life, without resort to such esoteric reflections. People reflectively endorse many ideals, even when commitment to them causes little favorable consciousness, or even grief. Environmentalists endure long hours of often boring, poorly compensated work to save remote ecosystems, such as Antarctica, that few people will ever see and enjoy. The Enlightenment ideal of rational autonomy, of facing up to hard facts no matter how disenchanting or disappointing, reflectively values truth over favorable but illusory feelings. If feelings were all that intrinsically mat-

126 • Monistic Theories of Value

tered, then people should indulge in fantasies, denial, and drugs to suppress the displeasing ones and arouse the pleasing ones. But such a life is either contemptible, if authentic joys are available to the person who leads it, or pathetic, if they are not. And the reason it is "better to have loved and lost than never to have loved" is not that a lost love on balance delivers more kicks than a loveless life. It often does not. People reflectively value life in loving relations with others even when this results in grave misfortune and grief, in part because love invests life with a meaningfulness and depth that a "happy" but loveless life lacks. So it seems that the object of value is the object of a rational, favorable response and not the favorable response itself (Gaus 1990, p. 108). (Of course, favorable responses may also be rationally valued, if they are the object of a favorable rational attitude.) A fourth conception of pleasure accommodates this thought.3 It understands "pleasure" in the sense in which we say that swimming or music or drinking are pleasures. These are the activities and experiences that we enjoy or that please us, which we pursue just because we like them. A hedonist of this fourth sense of pleasure claims that the sole ground of value is that it please someone. This sense of pleasure does not come close to comprehending all the activities and experiences people reflectively value. It includes only activities that are pursued as pleasures or for the sake of pleasure, the leisure-time activities of the vacationer, the weekender, and the idle rich (Bond 1983, pp. 113-114). But many activities, while enjoyable, aim at excellence rather than pleasure. Professional practitioners of the arts, sciences, and athletics value their work for the level of excellence it achieves, and they subordinate their efforts to standards of excellence defined independent of pleasure. Although they often take pleasure in their pursuits, it is absurd to suppose that the standards of excellence they strive to meet are merely instrumental means to pleasure. It is much easier to enjoy oneself without so much hard work, frustration, and discipline. "It would be absurd if . . . our lifelong efforts and sufferings aimed at amusing ourselves . . . Serious work and toil aimed [only] at amusement appears stupid and excessively childish" (Aristotle 1985, 1176b28-33). A closer examination of the kind of pleasure people take in excellent achievements shows that pleasure is not the sole ground of value. For people could not be pleased by these achievements if they did not recognize them as meeting standards of value defined independent of pleasure. The baseball pitcher who perfects his curveball takes pleasure in his superior athletic achievement, in a good he recognizes to be distinct from pleasure. The phenomenon of taking pleasure in excellence, in the

Monistic Theories of Value •

121

meeting of non-hedonic standards of merit, would thus seem to show that other grounds of value besides pleasure are authentic. The phenomenon of taking pleasure in excellence also puts to rest a common hedonist argument. Some believe that we desire an end only if it pleases us. If this is so, then pleasure is the only thing that fundamentally drives our valuations and hence constitutes the sole standard of intrinsic value (Railton 1990, pp. 167-170). The inference is mistaken. People might take pleasure in an end because it meets independent standards of excellence. The experience of pleasure would then be driven by other evaluative considerations. This is consistent with the claim that awareness of the realization of every desired non-hedonic value always pleases. It seems clear from experience that pleasures often depend upon nonhedonic evaluations. A mathematician may be thrilled by what she thinks to be an ingenious discovery, but will be disappointed if she finds out that it has been proved much more cleverly. A hedonist could claim that the only fundamental favorable response one can have to something is to be pleased by it. All other positive feelings are supposedly reducible to pleasure conjoined with various beliefs.4 But many feelings which cannot plausibly be reduced to the state of being pleased signify awareness of value. We can be awed by something, such as a volcanic eruption, which is too overwhelming at close quarters to be pleasant. Such experiences elicit a certain reverence for nature that informs much of the spirit of environmentalism. If nature were ectSj we would not care about preserving its harsher, awesome jasj>ectsJdangerous predators, desolate climes) or constrain our recreational and economic uses of it as we currently do for many nature preserves. Thus, awe can ground values independent of pleasure. Similar arguments cafTBelnade for other non-hedonic feelings associated with higher modes of valuation, such as respect and admiration, both of which are compatible and sometimes mingled with the unpleasant feeling of fear. Hedonists also are mistaken in reducing the bad to whatever is painful. Many unfavorable responses to things—annoyance, boredom, embarrassment, contempt, dismay, shock—need not be painful, yet they are clearly responses to what we find bad. It would be more plausible to reduce the bad to whatever arouses displeasure. But this does not explain contempt of others, which is often inextricably tied to a pleasing judgment of one's own superiority. Contempt signifies that its object is bad, but as a state of consciousness considered in itself, it may be quite satisfying. I conclude that an object s power to please someone does not constitute a necessary condition for its being considered good. Hedonism also errs in

128 * Monistic Theories of Value

claiming that arousing pleasure is a sufficient condition of goodness. We are appropriately ashamed, shocked, dismayed, or offended at many pleasures. One needn't have a puritanical or even a moralistic sensibility to be ashamed of childish or dishonorable pleasures. Children acquire a sense of shame in juvenile pleasures as part of growing up and learning to appreciate adult pursuits. This is a matter not of a pleasure's being intrinsically good, but instrumentally bad. In striving after independence, a maturing child recognizes her pleasure in childish dependency as something to be despised. It is not just an accidental obstacle to development, but unworthy in itself, since it reflects an attachment to an inferior way of behaving, appropriate only for "babies." Hedonists have a familiar response to this line of criticism. Confronted with the manifest pragmatic failure of their theory to make sense of the full variety of our attitudes and concerns, and with the absurdity of advocating a life without them, hedonists retreat to a self-effacing theoretical posture. What is good (pleasing) is that we lead lives filled with such richly differentiated ways of valuing things. Hedonism thus recommends that we pretend that there are authentic values besides pleasure, indeed, that we fool ourselves into really believing such falsehoods. If this is how we get our kicks, relieve our boredom, soothe our consciences, then let us live an illusion whose falsehood can be grasped only from a hedonistic perspective. This response is inadequate. Hedonists have not discharged the burden of exhibiting the defect in the perspectives from which we intrinsically value other things besides pleasure or its objects. And on what does the hedonist base his confidence that lives filled with respect and contempt, appreciation and derision, dignity and dishonor, pride and shame enjoy more net pleasure than lives insensitive to these distinctions? No one has offered empirical evidence for this claim or even figured out a plausible way to test it. It seems that what really grounds one's confidence in the greater worth of a life which embodies non-hedonic evaluative distinctions is not the empirical evidence on pleasures, but one's sense of dignity (Mill 1979, p. 9). It is an undignified, lesser way of life to fail to respond to distinctions between the honorable and the dishonorable, the admirable and the despicable, the moral and the immoral. But the sense of dignity cannot be reduced to a certain way of being pleased without begging the empirical questions at stake. If the sense of dignity is just another kind of susceptibility to certain species of pleasure, why cultivate it at the expense of our capacities to wallow in more abundant and easily obtainable base pleasures?

Monistic Theories of Value • 129

One way to answer this question is to point out that people rationally prefer a life with dignified pleasures to a life without them. This suggests a move from pleasure to rational preference or desire as the standard of value. Perhaps rational desire theory is the best option for monistic value theory. 6.4 Rational Desire Theory The most prominent naturalistic theories of the good identify it with the object of rational desire. What is good (for a person) is what she would desire if she were fully rational. One thing is better (for a person) than another if and only if she would rationally prefer it. The theory is naturalistic if it describes in nonevaluative terms the conditions of rational desire.5 Welfare economics identifies rational preferences with actual preferences. This is a crude approximation of an acceptable standard of value. Ignorance and violent emotions make people want things they find bad once they get them. Philosophical versions of rational desire theory try to refine the conditions of ideally rational desire to satisfy the reflective selfendorsement test. Widely, proposed conditions ..include that one be calm, reflective, and competent in calculating probabilities and consequences, as well as fully and repeatedly informed of all relevant facts (Rawls 1971, p. 408; Brandt 1979, p. 10; Sidgwick 1981, pp. 111-112; Darwall 1983, pp. 103-105; Griffin 1986, pp. 12-15; Railton 1986, pp. 173-174). A desire js_ saidjta_he rational if and only if a person would have it in these conditions.

hi'*

i11

My pluralist theory and rational desire theory agree in identifying the good with the object of a favorable attitude, not with the favorable attitude itself They also agree that a value-signifying attitude can fail to meet various rationality conditions: one can therefore objectively criticize a person s valuations. Pluralist and rational desire theories part ways on two issues. First, they disagree about which favorable attitudes signify values. Pluralist theories deny that desires or preferences directly signify value, insisting instead that other rational attitudes, mostly non-propositional, perform this role. Second, they disagree about how to specify the conditions that make a value-signifying attitude rational. Naturalistic rational desire theories claim that these conditions can be specified in nonevaluative terms. Pluralism claims that these conditions must refer to thick evaluative concepts. Consider the first point of disagreement. I have three objections to identifying the good with the object of rational desire. First, we find many

I

130 • Monistic Theories of Value

things to be valuable besides the possible objects of rational desire. Second, desires don't have the right structure to signify values in the way rational desire theory demands. Third, not all desires can signify values, because certain sources of motivation, such as appetites, whims, habits, and compulsions, often incline us to seek things we find to be bad. The desires which do seem to be reliable guides to value are all merely derivative of the rational attitudes that the pluralist-expressive theory identifies as value-signifying. To the extent that rational desire theory tracks reliable judgments of value, it is parasitic on the pluralist-expressive theory. Rational desire theory has no hope of offering a comprehensive theory of value. If desire is understood to be an actual motivational state that disposes a person to try to bring about its object, then its only coherent objects are states of affairs that a person can reasonably think she can bring about. People find many things to be valuable which are not coherent objects of desire. They value many things besides states of affairs. Rational desire theory cannot account for the intrinsic values of animals and persons. It accounts at best for the values of aims, not for the intrinsic values of the ends (persons, communities, animals) for the sake of which people act. Some states of affairs that people find valuable cannot be coherent objects of desire, because they can't be changed or because one cannot bring them about. One may judge certain laws of nature to be inconvenient or dangerous, such as those that make one sick from too much radiation, and others to be fortunate, such as those that make life on earth possible. These value judgments are not tied to any desires humans could coherently have to change or preserve the laws of nature. People may judge events in the past and events of a cosmic order to be fortunate or beautiful, although they can do nothing to help or hinder their occurrence, and hence cannot frame rational desires with respect to them. A romantic lover may be thrilled at the thought that his beloved might spontaneously fall in love with him without any prompting on his part. Such an event can be valuable, it can be coherently hoped or wished for, but it cannot be coherently desired in any motivational sense of desire. To try to bring this event about would be self-defeating, because the event is valued as one that the valuer has no active part in bringing about. Many goods are not objects of actual desire, since they come by surprise, and are valuable in part because they came that way (Bond 1983, p. 46). My theory, which ties value to valuings rather than desires, suffers from none of these limitations. Some attitudes, such as aesthetic appreciation, take things other than states of affairs as their object and need not imply that the agent has desires for its object.

Monistic Theories of Value •

131

Most rational desire theorists concede the incompleteness of their theory of value. They may take it narrowly to account only for human welfare or the good for a person (Railton 1986). The identification of a person's good with what she rationally desires seems both too wide and too narrow. People rationally care about many things besides their own welfare. And they sometimes fail to care about themselves, or suffer from impoverished desires, because of a lack of self-respect or self-esteem, depression, self-abnegating ideals of altruism, and adaptation to oppressive circumstances. There is no reason to think that full information, calmness, calculative competence, or any other naturalistically described conditions need make people care about themselves or arouse their desires. The first problem can perhaps be addressed by identifying a person's welfare only with the rational desires she has for herself. This risks confusing well-being with self-interest in the selfish, atomistic sense. Desires one has for communities of which one is a member can be for objects that contribute to one's well-being. Alternatively, the good for a person could consist in the satisfaction of those rational preferences she has that logically entail her existence (Overvold 1982). But the guilty or self-loathing desire that I be the one to be punished for a crime does not signify that I would be better off if I were punished. The second problem is also hard to resolve. One cannot simply add selfvaluation to the conditions of rational desire. Valuing oneself too much, or inappropriately, can also lead one to desire things that are not good for oneself. Vanity and egotism tend to pose obstacles to the appreciation of friendship and other relations that are constitutive of personal well-being. It is hard to imagine how to specify the "right" kinds of self-concern without using thick evaluative concepts. Uncertainty also remains about how to account for serendipitous and other goods for a person that cannot be brought about or prevented through her own efforts. One way to deal with this problem is to define a person's welfare in terms of the desires she would have if she could choose whole life histories that include events beyond her control in real life (Harsanyi 1982). This retreat from the desires people actually have to the desires they would have if their causal powers were wildly inflated, and if they could somehow step outside their lives and choose at once how their whole life would go, suggests trouble for rational desire theory. Why have any confidence in our ability to frame rational desires or make good decisions in such inhuman circumstances? We will return to these concerns. The second objection to desire accounts is that desires don't have the

.11

lit

132 * Monistic Theories of Value

structure required to generate coherent value judgments. Rational desire theories assume that value can be maximized. If rational desires are to serve as the sole ground of value, then they must be representable in a single, complete, transitive preference ordering. If rational motivations could be represented only in multiple, conflicting preference orderings, then some grounds other than rational desire would be needed to determine which preferences, if any, signified value. If preferences were incomplete, then some goods would be incommensurable and value could not be maximized. If preferences were intransitive, they could not ground coherent judgments of superior value. That is, because the relation "is better than" must be transitive, the relation "is preferred to" must also be transitive if it is to serve as the sole ground for asserting the first relation. Rational desire theories of value simply assume that a person's motivations can be represented in a structure that mirrors the logical structure of quantitative value judgments. But preferences are not always complete (§3.3). And appetites and whims can without contradiction express themselves in intransitive preference orderings. One may hunger for corn when faced with the choice between corn and peas, for peas when peas and asparagus are available, and for asparagus when it and corn are present. Such intransitive motivations are not incoherent but cannot signify relative values. Human motivations come from such diverse sources, and conflict so pervasively, that they cannot fit into a single preference ordering. They conflict because they arise from multiple sources influenced in different ways by different factors. Some desires spring from sources (rational attitudes) that are influenced by judgments concerning what we value, whereas others come from sources that are unaffected by such judgments. Given the pervasive conflicts among desires even when we are rational, we must appeal to standards external to desire to judge the authority of conflicting desires. Motivations vary according to the degree and ways they are affected by judgments concerning what a person values. Some desires lack the twofold structure of ends characteristic of the expressive self. Although aiming at the realization of some state of affairs, they are not done for the sake of anyone or anything else (§2.1). Such motives resist being changed in direction or intensity by judgments about their consequences or expressive significance for who we care about. These include desires prompted by instincts, drives, and appetites. People often submit to the cravings of hunger and thirst without regard for their health. Motives prompted by impulses and whims often incline people to do things harmful to people

Monistic Theories of Value •

133

they care about, as is notoriously the case with impulsive gambling and sexual behavior. Actions prompted by habits tend to persist after the original rationale for acquiring them has disappeared or been forgotten. Addictions, compulsions, and obsessions often overcome a persons best efforts to govern himself by reflective judgments. Finally, some emotions can be "blind": even though people acting on them act for the sake of some end, the action prompted by the emotion may not be responsive to reflections about what concerns the end. Anger can prompt one to actions, such as lashing out at those one loves, which one knows are contrary to one's interests and values. These motivations contrast with desires that express rational emotions or valuings. These desires display the twofold structure of ends and are influenced by reflection on what one rationally values. Call these desires evaluative. If a person's desire is evaluative, it will tend to diminish upon j h e discovery that it does not adequately express the way he values ie. Ben's enthusiasm for throwing his friend a surprise birthday party will flag if he finds out that his friend will only be embarrassed by it. An evaluative desire to do something for someone will also diminish upon the discovery that the person for whom one wants to act does not deserve the kind of valuation the act expresses. Ben will no longer want to throw his friend a party when he finds out that his "friend" has just defrauded him in an insurance scheme. Most motivations fall between perfectly rational evaluative desires and "blind" passions, appetites, and habits. Different attitudes are subject t a different degrees of influence by rational reflection on the merits of their objects. All attitudes can be ingrained in habit, and if inculcated at an early age, they can be especially resistant to rational reflection. So the disgust people feel for certain foods is hard to overcome even when prudence demands it. Once disgust becomes real nausea, it is hardly subject to normative control. Some motives such as fear can be "blind," but are usually sublimated or redirected by cultural norms and brought under the influence of evaluative reflections. Men in societies governed by an honor-revenge ethic sublimate their "instinctive" self-concern so that they fear dishonor more than death or physical injury. The difference between these two classes of motivations is reflected in the different means people use to control, suppress, arouse, and direct them. People use cognitive means to control rational evaluative desires. They judge how well the objects of their underlying attitudes meet relevant standards and how well their desires express their underlying attitudes. They undergo psychotherapy to get a better interpretive grasp of

;

!,Vi"

Ji

134 • Monistic Theories of Value

what their attitudes are really directed toward and why Faced with motivational mechanisms insensitive to cognitive changes of this sort, people turn to noncognitive means of control, such as behavioral conditioning, drugs, habituation, sheer exertion of willpower, and manipulation of the external conditions under which the desire mechanism is triggered. Sometimes people rig the conditions so they cannot act or so some stronger countervailing desire is triggered at the same time. The famous case of Ulysses, who had himself tied to the mast of his ship so he could hear the Sirens sing without being driven to steer his ship into the rocks, illustrates this mode of self-control (Elster 1979). Other times people rig the conditions so that the acts prompted by arational desires and passions end up having consequences desired for other reasons. Football coaches rouse the aggressive passions of their teams so they can win games. Here instrumental reason makes use of arational action-mechanisms for its own purposes. That people try to control, suppress, and redirect their desires suggests that they do not exist in the pre-established harmony required by the rational desire account. The different sources of motivation push and pull people in conflicting directions. Impulses conflict with habits, appetites with evaluative desires, blind emotions with rational attitudes. These conflicts are not merely due to the fact that people can't get everything they want, so that in satisfying one desire, a person frustrates others. A single preference ranking can represent the conflicts resulting from such contingencies, where a person wants A, wants B, can't have both, and articulates a preference for A over B. But the conflicts arising from different sources of motivation are often noncontingent: they occur when a person wants both A and not-A, when part of her prefers A to B and another part prefers B to A. A smoker may have a craving that makes her prefer smoking to abstinence—she may also have a concern for her health, or for an ideal of cleanliness, that makes her prefer abstinence to smoking. She may simultaneously take pleasure in smoking and despise it. Here conflicting attitudes give rise to essentially conflicting preferences. Preferences that directly or noncontingently conflict cannot be represented in a single preference ordering. If human beings are conflicted, their desires must be represented by multiple, conflicting preference rankings. But then which preference ranking constitutes the authoritative standard of value, according to rational desire theorists? It might be replied that motivational conflict will disappear once a person assumes the ideally rational standpoint of calm, calculatively competent, fully informed repose.

Monistic Theories of Value •

135

I have two reasons for doubting that this is a realistic expectation for human beings. One is rooted in human biology. Some of our motivational dispositions, notably appetites, drives, and instincts, are impervious to rational reflection. Even after people are fully informed of the dangers of drinking saltwater, their thirst for it does not go away when they are adrift in the ocean with no other water available. A little speculation about human evolution suggests why we are inherently conflicted beings. Because humans evolved from non-rational animals, we inherited many of their motivational capacities, which of course were not sensitive to rational reflection on the implications of acting on them. As human ancestors developed new motivational capacities subject to more sophisticated controls—for example, foresight and planning, commitment, and social attitudes such as shame and honor, governed by social norms—these were simply overlain on more primitive capacities without completely supplanting them or subsuming them under the new controls. The noncognitive control mechanisms for these primitive motivational capacities are also not completely effective in reigning them in. A second reason for thinking that motivational conflict is inevitable is rooted in society. Many of the most important interests, desires, and attitudes people have are attributable to their social roles and responsibilities. But there is no pre-established harmony among the different social roles and responsibilities a given individual might assume, or among the social norms governing the attitudes appropriate to each role. Even if, in some idyllic primitive society, there was once some such functional harmony of roles and attitudes, historical change would have guaranteed its disruption. Social roles often lose their point, whereas the attitudes informing them persist and get reinterpreted and redirected. Individuals assume new social roles, or they combine roles in unanticipated ways. The demands of wage labor and professional life, originally tailored to suit the life cycles of men with minimal dependent care responsibilities, conflict with the demands of family life once women enter the workplace and the gendered division of domestic labor is called into question. Because social norms are informed by essentially contestable concepts, there is no way to put a close to the conflicts generated by reinterpreting them, or any way to finally reconcile the attitudes informed by them. Because individuals cannot tailor their social role-given aims and attitudes to suit themselves, they will always find themselves conflicted. They can of course join with others to produce social changes to eliminate contradictions that have become intolerable. But history gives us every reason to believe that such change will produce new contradictions elsewhere. (This is not a conser-

,r ;

136 • Monistic Theories of Value

vative argument against social change. Some contradictions are manifestly better to live with than others, particularly in respect of the opportunities they afford for people to develop new capacities in meeting the challenges they pose.) Given that human beings' motivations must be represented by conflicting, multiple preference rankings even when they are ideally rational, which ranking is supposed to represent the order of value according to rational desire theory? One might propose that the strongest preference one has under these circumstances is the authoritative one for us. Preference strength can be measured either by felt intensity or by motivational effectiveness. Neither interpretation satisfies the reflective self-endorsement test. Desires differ in their phenomenological feel—some are visceral cravings, like hunger and thirst; others are emotional, like the desires expressing passionate love and hate; others are urgent, like the need for information about a disaster involving loved ones; still others are virtually without affect, as are most habits, such as the desire to get dressed in the morning. But the authority of desire has little to do with its felt intensity—if it did, I should let my cravings always override my habits. Yet if I get up late for work, I may judge it more important to get dressed fast than to eat breakfast, although my hunger is intense. The need to work may be urgendy felt without its being an object of craving or felt attraction. And how are we to compare the intensity of a craving with the intensity of felt urgency? The authority of desire does not correspond with its motivational effectiveness. If it did, people would never suffer from weakness of will, because they would always endorse whatever desire actually moved them to action. Perhaps second-order desires indicate which first-order desires are authoritative for a person. The preference that sets the standard of value is the preference that one wants to have. So, in the smoker's case, it is not only that part of her wants to smoke and part of her wants to abstain but she also wants to not want to smoke. And this would seem to tilt the judgment in favor of the desire for abstinence. But this can't be right. There is no guarantee that second-order desires are characterized by any less conflict than first-order desires. And why should one accord more authority to a desire, just because its object is a desire rather than some other state of affairs? One can be afflicted by second-order desires which persist from early childhood indoctrination, but which one no longer endorses. Can the type of desire indicate its authority for a person? Clearly habits, instincts, appetites, impulses, whims, and other nonevaluative desires have

Monistic Theories of Value •

137

no automatic authority for us. We all know occasions when we would not endorse our acting on these motivations, even when they are felt most strongly. Perhaps, then, the authoritative preferences are those that represent our evaluative desires, the desires that express our rational attitudes or valuings, such as respect, admiration, love, and contempt. But the mere fact that a desire expresses admiration as opposed to some non-rational response such as nausea or hunger does not, as such, give it special authority. One may well have better reasons for acting on a desire prompted by nausea than on one prompted by admiration. A person may feel charmed by someone, such as a charismatic religious leader, who she thinks is a charlatan. In such a case, her feeling of charm is not wholly responsive to her judgments of merit, and she withholds endorsement from her feeling and from desires that express it. I conclude that neither the strength, nor the order, nor the type of a desire can determine its authority for us. When rational preferences conflict, we cannot turn to any of these intrinsic attributes of desire to determine which sets the standard of value. Any proposal that constructs the ground of value just by the kind, strength, or order of a desire one has in naturalistically described "rational" circumstances fails the reflective self-endorsement test. One might say that the authoritative, value-signifying preference is the preference one endorses, all things considered. But if this formula is to transcend the limitations of the second-order preference criterion, it must include among the things considered standards external to rational desire itself, to determine which preferences signify action-guiding values. My pluralistic-expressive theory provides such standards. It claims that if a person has an appropriate attitude toward an object, and if she has a desire that adequately expresses that attitude, then that desire is valuesignifying—its object is valuable. The value-signifying character of rational evaluative desires is parasitic upon the value-signifying character of rational emotions or valuations. Only because some desires express underlying rational attitudes do they indicate what is valuable to us. Being the object of a rational desire is neither necessary nor sufficient for being good. It is not necessary because some things can be the object of favorable rational attitudes that do not generate desires. It is not sufficient because some desires that persist when one is fully informed, calm, and so forth fail the reflective self-endorsement test. The fundamental value-signifying attitudes are emotion-laden, mostly non-propositional valuings, not desires or preferences. This resolves the first conflict between pluralism and rational desire theory. The two theo-

$8 * Monistic Theories of Value

ries agree that the good is the object of a favorable rational attitude, while disagreeing on which attitudes set the standards for goodness. The second disagreement concerns the standards for determining when an attitude is rational. Rational desire theory specifies these standards in naturalistic or nonevaluative terms, pluralism in thick evaluative terms. I do not think that any set of standards specified in strictly nonevaluative terms captures all the ways in which attitudes can be rationally criticized. There is no set of naturalistically described circumstances such that one is bound to endorse whatever attitude one has in those circumstances. Consider the naturalistic conditions for rationality widely proposed in the literature: that one be in the "cool hour" of reflection, fully and repeatedly informed of all relevant information, and competent in processing information and calculating probabilities. There are many ways one could criticize the value-signifying purport of even the attitudes one has under these conditions. First, consider the concept of "relevant" information and "competent" processing. Is it relevant information that a certain object meets a particular thick standard of evaluation—that a deed was courageous or mean, that a sentiment is smarmy or sincere, that an excuse is lame or compelling? If this information is excluded, then we are at a loss to formulate any attitudes toward the objects in question beyond such primitive ones of like and dislike. Some attitudes like respect, appreciation, admiration, and moral approval are essentially constituted by norms for response that discriminate between thickly described objects. It is absurd to confine value-signifying attitudes to those that are not governed by thickly described norms. But if thick evaluative information is included, then it would seem that the conditions for rationality are no longer naturalistic. One response to this is to refuse to discriminate between authentic and unauthentic thick descriptions in the definition of "full information." Let "full information" include all characterizations of the objects in question from all evaluative perspectives, and let people respond however they may to this barrage. This refusal to evaluate the evaluations presented could restore the naturalistic character of the definition of full information. I suspect that a characteristic human response to such presentation would be confusion, ambivalence, and conflict. Surely one need not endorse whatever attitudes one has under information overload. The naturalist may respond by identifying the good not in terms of the actual responses one would have under full information, but with the responses one would have if one s cognitive capacities were enormously inflated to accommodate such information (Railton 1986). Must I

Monistic Theories of Value •

139

endorse the attitudes my cognitively enhanced self would have? The answer partly depends on an evaluation of my own character. Suppose I have a tendency toward arrogance and superciliousness which would expand proportionately with an increase in cognitive capacity. My cognitively enhanced self might regard my actual self with patronizing contempt. But surely this does not give me a reason to so regard myself or for me to adopt the groveling attitudes my cognitively enhanced self may think are appropriate for a being of my limited capacities. Enhanced information and cognitive capacities can corrupt character and judgment as surely as they can improve it. Even if we could postulate away the confusions created by full information, we could still raise questions about how a computationally competent person will process it. Naturalists hope that this information can be processed in "value-free reflection" (Brandt 1979, p. 113). It just wafts over ones consciousness, and desires emerge without one having consciously reflected on the merits of one's responses to that information. If sophisticated emotions such as love and respect are involved in the generation of desire, this hope is absurd. One can try to suppress reflection on the merits of one's emotional responses to facts. But one's emotions would still be unconsciously reacting more or less according to norms for feeling that distinguish between objects described in thick evaluative terms. Evaluation would still be going on, but without conscious monitoring. This kind of unreflectiveness obviously cannot be reflectively endorsed. It amounts to the recommendation that we act on the immediate promptings of emotion, without considering whether our emotions are appropriate or whether the desires prompted by them adequately express such emotions. Alternatively, the demand for "value-free reflection" could amount to the demand for information processing from which one's sophisticated (norm-governed) emotions have been disengaged. This is not unimaginable. Lobotomy patients are known to have extremely dulled and detached emotions. Perhaps, then, naturalistic rational desire theorists should persuade us that our good consists in what we would want if we were lobotomized but could still do math. Naturalists try to substitute for the question: do these facts merit this attitude? the question: do these facts cause this attitude? I believe that no matter how the facts are presented to a person however naturalistically constituted, she always has room to ask whether her resulting attitudes are rational or merited or endorsable. She has no reason to give up evaluative reasoning. Naturalistic rational desire theorists have tried to persuade people to give it up by attacking the intuitions on which it is based. We

140 • Monistic Theories of Value

have seen in the last chapter how unconvincing these attacks are. It is true, as naturalists insist, that our intuitions may always be prone to errors such as prejudice and superstition. One could guarantee that one would never make such errors by giving up these intuitions. But to do so would be to give up the entire panoply of emotions that make our lives meaningful, to reduce our lives to a wormlike existence. Worms, indeed, do not make evaluative errors. But this hardly means that their "way of life" is appropriate for us. Some rational desire theorists, such as James GrifFm, agree that the conditions of rationality cannot be fixed in nonevaluative terms. Full information must include an understanding of thoughts expressed in terms of adequately evolved thick concepts (1986, pp. 12—13). Welfaretracking desires must be cultivated under conditions of justice and appropriate self-regard, so that they do not merely reflect adaptation to oppressive circumstances. In rejecting naturalism, such a theory approaches the evaluations of pluralist-expressive theory. Because the good for a person includes states of affairs that she cannot influence, rational desire theory must interpret "desire" to include more than motivational states—for example, wishes, hopes, and likings. Once the theory has gone this far, why resist the idea that all rational favorable attitudes are value-signifying? At this point, the distinctive function of desire disappears and is replaced by other favorable attitudes. This would be the superior solution, since it avoids the trap of endorsing all nonevaluative desires that resist rational reflection.

7 • The Ethical Limitations of the Market

7.1 Pluralism, Freedom, and Liberal Politics Political theorists have often justified liberal practices by appealing to pluralism. My theory of value is pluralistic in two ways that are relevant to political theory: it acknowledges a plurality of authentic but conflicting ideals and conceptions of the good, andjt_xlajjns jh^,.di]£exexilLJdnxis_of goojds^are^ajioj^ally^^ Liberals appeal to the pluralism of ideals to justify individual rights to liberty against state interference in their personal choices. Thj^grounds^ the liberal jiivisian public andprivate spheres. The_second kind ojjjlumHsmj^ robiistjystern of social spherejiifferentiation that jrequ^sjhar^^liir^te jDih th£J>£or2e_jQi^ theory has not yet come to grips with the full implications for human freedom and flourishing of this most expansionary institution of the modern world. In the next three chapters of this book I will focus on the limits of the market. I will show that an adequate grasp of liberal commitments to freedom, autonomy, and welfare supports more stringent limits on markets than most liberal theories have supposed. The need to limit markets is based on a pluralistic theory of the social conditions for freedom and autonomy. Call a person free if she has access tg a wide range of significant options through which she can express her diverse valuations. Individuals require social settings, governed by distinct | social norms recognized and endorsed by others, to develop and express; their different valuations (§1.4). Because people y^h^e^diShj^nX^QQ^in difiercn-t.ways^rthjeit freedom requires the availability of a variety of social sgheres that embody these different modes of valuation. Freedom thus requiresIjnultigle[sphere differentiation—•boundaries^ not just between the staJ^jmdjjbLe^mar^ self-expression, such as family, friendship,.,xlub^ art, science,

142 * The Ethical Limitations of the Market

*

*

"

religion, and charitable and ideal-based associations (compare 1989, pp. 173-174). Call a person autonomous if she confidently governs herself by principles and valuations she reflectively endorses. Autonomy can be undermined internally by addictions, compulsions, phobias, and other neuroses, which motivate a person in ways she cannot reflectively endorse. It can be undermined externally by social stigmatization and relations of domination. Stigmajization under mines the self-jesp^ct^j^^on^ngeds to take ho]iyaluatiojris.senQusly. Relations of domination give others^h tell one what~xa.dQ or to force., one tp_dp_sojm^hing^wjthaujLliayirig_to J c^nisaltxu:xespQnd..to.Qnes own judgmen&^Autonomy can be realized on a collective scale through democratic institutions. Collective autonomy consists in collective self-governance by principles and valuations that everyone, or the majority, reflectively endorses. It can be undermined by any non-democratic institution that controls political outcomes. Autonomy, like freedom, requires social conditions for its realization that demand significant constraints on the scope of the market and private property rights. The discriminatory use of business property stigmatizes members of excluded groups and diminishes the range of significant options open to them. Liberals propose a differentiation within the private sphere between the personal sphere and civil society to solve this problem. While individuals are free not to befriend or marry members of groups they hate, they may not close their businesses to customers or job-seekers who belong to such groups. Protection of autonomy may sometimes regujre.^rohibiting the commodification of some things. Prohibiting the sale of addictive drugs can help preserve autonomy for individuals suscer>tible to drugjbuse. Prohibiting the sale of votes helps preserve collective^ autonomy_by blocking one way the wealthy may try to control political outcomes. ^ M o s t important to the preservation of autonomy^are goods embodied in the person, such as frejejbm^ojfj£tion and the powers of productive d reproductive labor. To selljihese goods to another ith retaining rights to^consultation^ self-judgment, and control over^the conditions in which one acts is to redu£e^one^s autonomy by_subjecting ojiesejfjto another s domination. TLihej^Ddlf o™1** recognize this as a reason \ for prohibiting thesale of persons into slavery, but they often fail to think through its imrjEcations for employment contracts,,and for contract involving a person's sexuaj_and regroductive powers.1 Autonomy requires that many fights ip oursejve^r^^ali^ab^) ^TKe seco'nH kind ot pluralism affirms the liberal commitment to •

• > * . ? / • /

The Ethical Limitations of the Market •

143

freedom and autonomy buj^_dej^pen^ -^ under which they are realized. These conditions provide two grounds for constraining the scope of markets. Constraints may be needed to secure the robust sphere differentiation required to create a significant range of options through which people can express a wide range of valuations. And they may be needed to protect individual or collective autonomy. Lfbx^ to secure .freedom ion" can-be practiced to promote other Hbej^aims, such as, equality,-justice (Walzer 4983, 19R4)r,iricliyiduality, sndLB.ejjtr^lity (Herzpg 1989). These aims may require further constraints on the market. My theory of sphere differentiation resembles Michael Walzer s (1983). But Walz^r_^ontends that the properJbojandajies,,bej^ejen jodaX.sgheres can be dexiyj^^^ isg^Qf-^^-Sd^which are taken...as,&Q^ This view encounters familiar difficulties: shared understandings, if they exist at all, are often riddled with contradictions and confusions, are^stel^hedjn relations of domination that sikna^ t h e ^ e x ^ society, and fail to meet the pragmatic demancls,. sja^h as_ the £^ oLsoeial order, that people ask of them (Dworkin 1984; Daniels 1985; Cohen 1986; Herzog 1989, pp. 162-171). Walzer is right in maintaining that shared understandings are the proper starting point of political argument. But justification need not be confined to such understandings. It allows for conceptual innovation in the space of reasons (§5.3). Justification also requires equality of the participants, so as to avoid a false consensus achieved by force or domination. With these qualifications in mind, let us consider in detail the ethical limitations of the market. 7.2 The Ideals and Social Relations of the Modern Market Pluralism says that goods differ in kind if they are properly valued in different ways that are expressed by norms governing different social relations. Economic goods are goods that are properly valued as commodities and properly produced and exchanged in accordance with market norms. The proper Hmits xdC the market are partly defined by answering the following questionsCjFirst/do market norms do a better job of embodying thejwaysj^ve properly value^ a particuIaFgood than nonSFof other spKeres? If not, then^we^n^SdnT*'treat: th^nTas commocUties but rather locate them in non-majr^e££pheres. Second, do market norms, when they govern the circulation of a particular good, undermine importantTcfeals

The Ethical Limitations of the Market •

145

144 • The Ethical Limitations of the Market

such as freedom^^autonomy, and equality, or important interests lggitimately protected by the state? If so, the state may act to remove the good from control by market norms. We can understand the nature of economic goods by investigating the ways we value commodities; the social relations within which we produce, distribute, and enjoy them, and the ideals these relations are supposed to embody. I call the mode of valuation appropriate to pure commodities "I-KP"^Hsp^L^J^gggrC^ exclusive mode of valuation. It is contrastedjwjtfa higher rriode£ of valuation, such as respect. / To merely use something i§ to subordinate if tn onf's.._QwnT_^mi^jWTfhnnt regard for its.,intrinsic value. When owners of David Smith sculptures stripped the sculptures' paint to enhance their market value, they treated UrV\ •A* them as mere use-values, disregarding their intrinsic aesthetic worth in favor of their usefulness for independently defined ends.v^The^imriexsQnality of use is contrasted with valuing something for its, p£XSQjnaLaJ£a.chments to oneself, as when O ne cherishes an heirloom. Mere usg-Yalues^re d w iit h commodity at some f u n g i b l e a n d are tracLed n i , equanimity e i q u a m i m . i ^ j for v i . i Sany M f ,other v «***•.=•_=. _. . i d i p l a c e a b l e It price. But a cherished item is valued as unique and irreplaceable. is often sold only under duress, and its loss is felt personally (Simmel 1978, pp. 123, 404, 407; Radin 1982, 1987).2 The exclusivity of use-values is contrasted with shared goods. Commodity values can be enjoyed in use by oneself or by private groups, excluding those with whom one exchanges the good. But the value to oneself of shared goods is dependent upon other people in civil society, or the people with whom one exchanges the good, also enjoying the same items according to shared understandings of what it means. For example, the site of a historical event may be valued as part of the national heritage, or the layout of a neighborhood valued as the locus of a community. The most important ideal the modernjnarketai&empts to embody is^an iije^onoi^ Economic freedom)consists in having ^both a large menu of choices in the mai^etplzc^Snaexclusive power to use what one buys there at will. ItjeavelTone free fromTthe constraints on use required to realize goods as hjgherv personal, or shared: it permits one to disregard or destroy the intrinsic value of what one owns; it gives one access to goods independent of one's personal characteristics or relations n^ s to others: and it leaves one free from uncontracted obligations to others, | free to disregard their desires and value judgments, and free to exclude ern from xcren a 1 qtan H prHS 4 > ] ^ th£ worker u worker in in ^esjjnej^ojcjek^ the sjmej^oicele^ position positiQi> a§fa a^ the customer. I shall say that a thing is a pure economic good if its production, distribution, and enjoyment are properly governed by the five norms above and its value can be fully realized through use. This defines what may be called the ideal type of an economic good, tied to an ideal typical account of market norms. Ideal type analysis has limitations. Any particular social institution or practice may diverge from its ideal type in many ways and include mixtures of norms from other social spheres.^rhejxorms that govern our actual practices at present often inadequately express the ideals these practices are supposed to embody. My arguments focus on concerns that find little place in the standard models of welfare economics and justice. The standard models highlight other important concerns, such as efficiency. Any move from an evaluation of ideal types to an evaluation of actual practices must be informed by a detailed empirical investigation of the actual norms they embody, how well these norms express the ideals in terms of which they are justified, and how well the practices fare by other criteria such as justice and efficiency. Because I cannot provide such details here, the cases I discuss in this chapter should

The Ethical Limitations of the Market • 147

be taken as illustrations of the kinds of arguments I wish to endorse, and not as comprehensive evaluations of the practices in question. In the next three sections I will consider limitations of the market with respect to partially commodified goods of civil society, goods of personal life, and goods of the political sphere. 7.3 Civil Society and the Market Civil society defines a sphere of interaction that ideally is open to all its members on the same terms, independent of their personal relations to others, their social status, or their occupation of government office. Civil society includes markets, profit-making firms, nonprofit institutions such as hospitals and schools, professional associations and labor unions, political parties and action groups, and philanthropic and ideal-based organizations. Although the state may regulate and even fund these institutions, individuals pursue their own purposes in them, which are defined by internal institutional ideals and functions rather than by state fiat. And though individuals may engage in market transactions in their non-market institutional- or role-given capacities, their activities are not and should not be comprehensively governed by market norms. The scope of the market is limited by other roles and institutions in civil society. The proper relationship of these other institutions to the market raises delicate questions of boundary setting. Consider the status of professionals such as doctors, lawyers, academics, athletes, and artists who sell their services. Excellent performance in professional roles is judged by the standards of goods internal to the practice rather than by external instrumental criteria such as profitability. Academics pursue understanding, athletes win games, artists produce aesthetic value, and so forth. Adherence to their defining ideals and goals often involves forgoing opportunities for making money. When professionals sell their services, they enter into market relations that impose norms on their activities which potentially conflict with the norms of excellence internal to their professional roles. The goods internal to these professions become partially commodified. Pluralism does not repudiate such mixed practices. Sphere differentiation should not be confused with complete sphere segregation. The freedom of professionals to sell their services promotes equality of opportunity and autonomy. Achieving excellence in the professions is a fulltime activity. If professionals could not be paid for their work, only independently wealthy people would be able to pursue it. Professionals can make a living at what they do by selling their services on the market,

148 * The Ethical Limitations of the Market

seeking private patrons, or obtaining state funding. The market offers some advantages over the other two sources of support. When professional artists and doctors depended on private patrons, their fates were closely tied to the whims and fortunes of one or a few individuals, and they often lost voice over their activities. Reliance on the state can also be dangerous, since one must be careful not to offend the political interests of those in power. The option of marketing one s products or services to the general public can enhance autonomy by restoring voice to self-employed professionals who set their own terms of sale before making contact with their customers. But the market is no panacea, either. If professionals are employees of profit-making firms, their autonomy can be compromised by the firm's demand to make a profit. Greed can also corrupt self-employed professionals. Artists may pander to public taste rather than challenging it. Lawyers may act merely as hired guns for their clients, harassing those against whom their clients have no genuine legal case. Doctors may perform profitable but medically unwarranted services on ignorant or demanding patients. Some of these dangers can be alleviated through employment by nonprofit organizations, such as museums, hospitals, and universities, whose defining aims are the promotion of goods internal to professional practice. Professional autonomy and integrity are enhanced by differentiation within civil society between profit and nonprofit institutions. This differentiation can be sustained only if market norms do not wholly govern exchanges of money for professional products or services. Yet some regard market norms as providing the only normatively valid framework for such transactions. This thought lies behind the case for government censorship of the art that it funds. This argument accepts as ethically axiomatic that whoever pays for a good may refuse to pay for any goods that fail to meet her specifications.4 Hence the National Endowment of the Arts acts within its rights when it refuses to fund art it regards as obscene or politically offensive. The argument views the state as a customer with absolute rights to exit, while it regards state-sponsored art as a mere commodity, the production of which should be purely wantregarding. The same argument would uphold the former policy of the exSoviet Union of funding only Lysenkoist genetic theory and of firing scientists who opposed this fraudulent and politically corrupt research program. In general, the state would have the right to fund only public universities and academics whose research and teaching slavishly parroted the party line of state officials, or perhaps the political opinions of the majority of citizens.

The Ethical Limitations of the Market •

149

If the state should promote citizens' freedom and autonomy, it may not regard itself as a customer with respect to all the projects it funds. Its proper aim in funding projects is not to serve the political interests of the state, the self-interest of its officials, or even the tastes of the majority, but to expand the range of significant opportunities open to its citizens by supporting institutions that enable them to govern themselves by the norms internal to the modes of valuation appropriate to different kinds of goods. Art and science constitute significant domains of human pursuit, each requiring institutions enabling people to regulate their activities in accordance with standards of excellence internal to them. Firms that produce art or science for profit do not fully meet these enabling conditions because they subordinate truth-seeking and aesthetic production to external commercial standards of profitability. Profit and nonprofit art and science institutions that need to charge high prices for access to their products in order to survive exclude all but the wealthy from the opportunities they provide and also mistakenly treat nonrival goods as if they were rivals in consumption. Private philanthropy often undersupplies opportunities to the general public and distorts supply in the interests of snobbery and elitism. State funding of artistic and scientific projects, in which the assignment of grants is determined by peer review rather than by political criteria, can therefore play an irreplaceable role in enhancing the freedom and autonomy of citizens (compare Dworkin 1985, pp. 227—233). Such funding will often result in artistic production that is offensive to public tastes, and scientific theories, such as evolution, that are offensive to popular beliefs. But autonomy is enhanced by providing opportunities for people to learn and grow, even when these opportunities offend, by challenging, their present beliefs and desires. Neither freedom nor autonomy is correctly defined in terms of the satisfaction of given desires or conformity to given beliefs. Both scientific valuation and modern aesthetic appreciation engage a potentiality for self-transcendence, which can be actualized only by permitting experiments that may also fail. The state therefore should recognize the boundary between itself and civil society by not regarding itself as a customer for all the projects it funds. One might object that state control always accompanies state funding; consequently, the hope that the state can enhance autonomy is illusory. Whatever its source, external funding tends to influence the content of what is funded. Because professionals must rely on some outside funding, this is an argument that their autonomy is best preserved by a diversity of funding sources, public and private.5 The autonomy-compromising effects of reliance on outside funding can often be more severe in the market or

150 • The Ethical Limitations of the Market

the private domain than in state funding. Many fields of research in the humanities and the basic sciences would receive very little if they had to be commercially profitable or interest a private philanthropist. Liberal institutions have managed to secure high degrees of autonomy for many professionals whose activities are funded by the state. State universities have not done a notably worse job than private universities in protecting academic freedom, and in both cases, despite serious lapses, their performances have been remarkably good. The choice of institutional sponsorship of professional acjdvity^sJLess impoxtaut than the norms governing its sponsorship. The state compromises autonomy and hinders the pursuit of the non-economic goods internal to professional practices when it adopts market norms for allocating funds. Profit-making firms can promote non-economic goods internal to professions if they regulate their activities by non-market norms. The commissioner of baseball and the draft system introduce nonmarket norms into professional baseball which help preserve goods internal to the sport that are undermined by pure competitive markets in players and complete private property rights in teams. tDiiaxkel norms pose a constant threat to the autonomy of professions jind the integrity of goods internal to them, they may be to some degree Jndispensable, because the professions require external sources of funding. Since funds are limited, efficiency considerations should influence allocative decisions. Physicians would have no difficulty spending the entire GNP on health care if they were guided solely by the aims internal to their practice. Incentive systems structured by market norms might play a useful role in preventing waste. Unfortunately, no incentive system simple enough to save more than the costs of monitoring its implementation reliably tracks the standards of good care internal to medical practice. Compromise is required here. The goods of professional practice are, perhaps inevitably, partially commodified, and they require hybrid institutiojas- in civil society that combine market with non-market norms for their proper provision. 7.4 Personal Relations and the Market The modern Western opposition of personal and with market relationships is the product of historical processes that separated economic production from the household (Nelson 1969; Lasch 1977; Zaretsky 1986). In this section I will focus on two of the many ideals distinctive to the sphere of personal relations: intimacy and commitment. Living on inti-

The Ethical Limitations of the Market * 151

mate terms with another person involves sharing private concerns and cherished emotions attuned to the others personal characteristics. This is the romantic side of personal relations, involving passion, affection, and trust, but not necessarily devotion, as the romantic relationship may end as soon as the passions that animate it subside. The deepest ideal of commitment involves dedicating oneself to permanently living a shared life with another person. The goodness of such a life for each partner is shared. It partially consists in the fact that the other partner also enjoys this life, that each partner realizes this, and that she knows that the other knows. One committed and loving partner cannot unequivocally rejoice in his life with his partner if he knows that the other finds the relationship oppressive in some way. Commitment to a shared life, such as a marriage, requires redefining one's interests as part of a couple. A person's committed interest in the aims of the marriage can be neither defined nor satisfied independent of her being joined with, her partner in marriage. These ideals inform the ways we value the people with whom we have personal relationships and the goods we exchange with them. The goods exchanged and jointly realized in friendship are not merely used but cherished and appreciated, for they are expressions of shared understandings, affections, and commitments. The jjoods proper to the personal sphere can be fully realized only through!gifl^excnang?] They cannot be procured by paying others to produce them, because the worth of these goods depends upon the motives people have in providing them. Among these goods are trust, loyalty, sympathy, affection, and companionship. The norms of gift exchange differ from the norms of market exchange in several respects (Mauss 1967; Titmuss 1971; Sahlins 1972; Hyde 1983). Gift exchange affirms and perpetuates the ties that bind the donor and the recipient. To refuse an appropriate gift is to insult a friend by failing to acknowledge or sustain a friendship. Gift exchange aims to realize a shared good in the relationship itself, whereas market exchange aims to realize distinct goods for each party. Although both forms of exchange involve reciprocity, the form and timing of the return of goods differ in the two cases. In market exchange, an uncontracted delay in reciprocation is cause for legal action. But the exchange of gifts among friends usually incorporates an informal understanding of reciprocity only in the long term. To be anxious to "settle accounts" of small sums, as when one person insists upon splitting a restaurant tab exactly in half, calculating sums to the penny, is to reject the logic of friendship. The delay in reciprocation expresses an intrinsic valuation of the recipient: gifts are given for the

152 • The Ethical Limitations of the Market

friend's sake, not merely for the sake of obtaining some good for oneself in return. The accounting mentality reflects an unwillingness to be in the debt of another and, hence, an unwillingness to enter into the longer term commitments such debts entail. The debts friends owe to one another are not of a kind that they can be repaid so as to leave nothing between them. (Debts involving large sums are another matter, since they threaten the relative financial self-sufficiency that is presupposed by modern friendship in market-based economies between equals who are not kin.) Friendly gift: exchange is responsive to the personal characteristics of friends and to the particular qualities of their relationship. We seek to give gifts to our friends that have more than a merely generic meaning. For gifts express friends' mutual understanding of how their relationship stands (or how the giver wishes it to be) and not merely a good of impersonal use-value to the receiver. This is evident not only in cases of such material gifts as engagement rings, but also in the exchange of compliments, affections, and jokes. This is why cash is usually an inappropriate gift between friends: because it can be used by anyone to acquire any commodity, it expresses nothing of the giver s personality, of any particular thought the giver had for the receiver, or of the receiver's interests.6 These differences between personal and market norms can help us explore how personal goods are undermined when market norms_govern their circulation. The thought that authentic personal goodsj*re jmdernunedjwhen this happens has been challenged by femimsjjhe^ 1 n,f 3fp