Reasons and Recognition: Essays on the Philosophy of T.M. Scanlon 9780199753673, 0199753679

For close to forty years now T.M. Scanlon has been one of the most important contributors to moral and political philoso

604 93 2MB

English Pages 379 [394] Year 2011

Report DMCA / Copyright


Polecaj historie

Reasons and Recognition: Essays on the Philosophy of T.M. Scanlon
 9780199753673, 0199753679

Table of contents :
Cover......Page 1
CONTENTS......Page 8
Preface......Page 6
Contributors......Page 10
1. The Activity of Reason......Page 18
2. Valuing......Page 38
3. Aims as Reasons......Page 58
4. Scanlon on Desire and the Explanation of Action......Page 94
5. Of Metaethics and Motivation: The Appeal of Contractualism......Page 116
6. Contractualism on the Shoal of Aggregation......Page 144
7. Immoral, Conflicting, and Redundant Promises......Page 170
8. The Trouble with Tolerance......Page 194
9. Rescuing Conservatism: A Defense of Existing Value......Page 218
10. Global Political Justice and the “Democratic Deficit”......Page 246
11. Establishment, Exclusion, and Democracy’s Public Reason......Page 271
12. The Significance of Distribution......Page 291
13. The Trouble with Psychopaths......Page 322
14. Blame, Italian Style......Page 347
15. Dispassionate Opprobrium: On Blame and the Reactive Sentiments......Page 363
B......Page 388
E......Page 389
M......Page 390
Q......Page 391
S......Page 392
V......Page 393
W......Page 394

Citation preview

reasons and recognition

This page intentionally left blank

Reasons and Recognition essays on the philosophy of t. m. scanlon Edited by R. Jay Wallace, Rahul Kumar, and Samuel Freeman

Oxford University Press, Inc., publishes works that further Oxford University’s objective of excellence in research, scholarship, and education. Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2011 by Oxford University Press, Inc. Published by Oxford University Press, Inc. 198 Madison Avenue, New York, New York 10016 Oxford is a registered trademark of Oxford University Press All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Library of Congress Cataloging-in-Publication Data Reasons and recognition : essays on the philosophy of T. M. Scanlon / edited by R. Jay Wallace, Rahul Kumar, Samuel Freeman. p. cm. ISBN 978-0-19-975367-3 (alk. paper) 1. Scanlon, Thomas. 2. Philosophy, American--20th century. I. Wallace, R. Jay. II. Kumar, Rahul, 1967- III. Freeman, Samuel Richard. B945.S254R43 2011 191—dc22 2010039086

1 3 5 7 9 8 6 4 2 Printed in the United States of America on acid-free paper


t. m. scanlon is one of the most important philosophers working today. Through his writing and his teaching, he has decisively shaped the questions with which research in moral and political philosophy now grapples, exerting a virtually unrivaled influence on contemporary discussions. In moral philosophy, Scanlon is best known for the development of the theory he calls contractualism. This theory offers a comprehensive account of the central moral requirements that regulate our interactions with each other (the “morality of right and wrong,” in Scanlon’s words). On the contractualist approach, these requirements are determined by principles for the general regulation of behavior that nobody could reasonably reject. Thus, morality has its source in the idea of a kind of hypothetical agreement amongst individuals who are seriously concerned to discover a common basis for social life. The principles that are arrived at through this agreement in turn define a valuable way of relating to other people, which Scanlon refers to as “mutual recognition.” By complying with core moral requirements, we ensure that our behavior can be justified to anyone who might be affected by it on grounds that it would be unreasonable for those persons themselves to deny. This way of thinking about morality represents a sophisticated and attractive alternative to the consequentialist approaches that have dominated discussion in the tradition of Anglo-American philosophy. Furthermore, Scanlon’s development of the contractualist theory exhibits the comprehensive ambitions that are characteristic of the great works of moral philosophy of the modern period. Contractualism is, in the first instance, an account of the normative structure of the morality of right and wrong. But Scanlon takes its appeal to derive in part from the light it is able to shed on other issues that are of perennial concern to moral theory, such as the normative significance of moral requirements for the agents who are subject to them and the v



nature of responsibility and blame. His contributions in moral philosophy also include penetrating treatments of fundamental questions in metaethics and moral psychology, concerning (for example) the nature of value and the sources of moral motivation. Taken together, Scanlon’s writings on these topics represent one of the most powerful bodies of contemporary work in philosophical ethics. Scanlon’s contributions in political philosophy are similarly influential and important, though they have assumed a somewhat different form. Rather than develop a single, comprehensive theory of political morality or justice, Scanlon has offered detailed and thoughtful investigations of a series of discrete questions that are at the center of social and political life: What is the rationale for the constitutional protection of the right to freedom of expression? What makes different forms of inequality objectionable? Is an attitude of tolerance one that we should value in our political institutions and in ourselves as citizens? Scanlon’s treatments of these important questions combine analytical rigor, philosophical insight, and a deep appreciation for the complexities that emerge when normative ideals interact with institutional and cultural realities. The present volume brings together fifteen new essays by both younger and more established philosophers on themes from Scanlon’s work. The topics that have been selected reflect the full range of Scanlon’s interests, touching on issues in all parts of moral and political philosophy. Together, the essays contribute to a deeper appreciation of Scanlon’s views, while also advancing our understanding of the important questions that are addressed in his groundbreaking work. We offer the collection with gratitude and admiration, as a tribute to the significance of Scanlon’s philosophical accomplishments and to the value of his influence on our field. Drafts of several of the essays in the volume were discussed in the company of Scanlon and most of the contributors at a workshop that took place in Princeton in October of 2008. We are grateful to Michael Smith and to Princeton University (in particular, to the Philosophy Department, the Council of the Humanities, and the Office of the Dean of the Faculty) for the generous support that made possible this stimulating and constructive event. We would also like to thank Jordan MacKenzie, who prepared the index and provided conscientious assistance with the production of the volume. R. Jay Wallace Rahul Kumar Samuel Freeman

Contents Preface v Contributors ix part one | reason, value, and desire 1. The Activity of Reason 3 Christine M. Korsgaard 2. Valuing 23 Samuel Scheffler 3. Aims as Reasons 43 Niko Kolodny 4. Scanlon on Desire and the Explanation of Action 79 Michael Smith part two | ethical themes: contractualism, promissory obligation, and tolerance 5. Of Metaethics and Motivation: The Appeal of Contractualism 101 Pamela Hieronymi 6. Contractualism on the Shoal of Aggregation 129 Rahul Kumar 7. Immoral, Conflicting, and Redundant Promises 155 Seana Valentine Shiffrin 8. The Trouble with Tolerance 179 Angela M. Smith

part three | political themes: conservatism, justice, and public reason 9. Rescuing Conservatism: A Defense of Existing Value 203 G. A. Cohen 10. Global Political Justice and the “Democratic Deficit” 231 Charles R. Beitz 11. Establishment, Exclusion, and Democracy’s Public Reason 256 Joshua Cohen 12. The Significance of Distribution 276 Aaron James part four | responsibility 13. The Trouble with Psychopaths 307 Gary Watson 14. Blame, Italian Style 332 Susan Wolf 15. Dispassionate Opprobrium: On Blame and the Reactive Sentiments 348 R. Jay Wallace index 373


Charles R. Beitz is Edwards S. Sanford Professor of Politics at Princeton University, where he directs the University Center for Human Values. His publications include The Idea of Human Rights, Political Equality, and Political Theory and International Relations, as well as articles on a range of topics in political philosophy. He recently ended a tenyear term as editor of Philosophy & Public Affairs. G. A. Cohen was, from 1985 until his retirement in 2008, Chichele Professor of Social and Political Theory and a Fellow of All Souls College, Oxford. He occupied the Quain Chair of Jurisprudence at University College London at the time of his death in August 2009. A posthumous collection of his essays, On the Currency of Egalitarian Justice and Other Essays in Political Philosophy, was published in 2011. Joshua Cohen is Martha Sutton Weeks Professor of Ethics in Society at Stanford University, where he teaches political science, philosophy, and law; he also directs Stanford’s Program on Global Justice, and is editor of Boston Review. His many publications include Philosophy, Politics, Democracy, and Rousseau: A Free Community of Equals. Samuel Freeman is Avalon Professor in the Humanities at the University of Pennsylvania. He works in social and political philosophy, ethics, and the philosophy of law. His publications include Rawls and Justice and the Social Contract. He edited the Cambridge Companion to Rawls, as well as John Rawls’s Lectures on the History of Political Philosophy and his Collected Papers.




Pamela Hieronymi is Associate Professor of Philosophy at the University of California, Los Angeles. Her primary research concerns agency, responsibility, and the mind. She has published a series of articles developing an account of the agency we exercise with respect to those states of mind for which we are responsible. Aaron James is Associate Professor of Philosophy at the University of California, Irvine. He has published in meta-ethics, moral theory, and political philosophy in a variety of journals, including Philosophy & Public Affairs and Philosophy and Phenomenological Research. He is currently writing a book on fairness in the global economy from a social contract perspective. Niko Kolodny is Associate Professor of Philosophy at the University of California, Berkeley. He works mostly in moral and political philosophy, and has written on love, relationships, promising, rationality and reasons, and the semantics of “ought.” Christine M. Korsgaard is Arthur Kingsley Porter professor of philosophy at Harvard University. She is the author of Self-Constitution: Agency, Identity, and Integrity; The Constitution of Agency; The Sources of Normativity; and Creating the Kingdom of Ends, and one of the editors of Reclaiming the History of Ethics: Essays for John Rawls. Rahul Kumar is Associate Professor of Philosophy at Queen’s University, Canada. He works mainly in moral and political philosophy. He has written on deontic constraints, the nonidentity problem, obligations to past and future generations, and the justification of risk imposition. Samuel Scheffler is University Professor at New York University, where he teaches in the philosophy department and the law school. He works mainly in the areas of moral and political philosophy. His publications include The Rejection of Consequentialism, Human Morality, Boundaries and Allegiances, and Equality and Tradition. Seana Valentine Shriffin is Professor of Philosophy and Pete Kameron Professor of Law and Social Justice at the University of California, Los Angeles, where she directs the Law and Philosophy Program. She works on ethics, political philosophy, the philosophy of law, freedom of speech, contracts, and other legal subjects. Angela M. Smith is Associate Professor of Philosophy at Washington and Lee University. Her main research interests concern the nature of moral agency and moral responsibility, and she has published papers on these and related topics in journals such as Ethics, Philosophical Studies, Philosophical Topics, and The Journal of Ethics. Michael Smith is McCosh Professor of Philosophy at Princeton University. His interests include ethics, moral psychology, the philosophy of mind and action, political philosophy, and philosophy of law. He is the author of The Moral Problem and Ethics and the A Priori, and co-author of Mind, Morality, and Explanation.



R. Jay Wallace is Professor of Philosophy at the University of California, Berkeley. His publications include Responsibility and the Moral Sentiments, Normativity and the Will, and numerous articles on practical reason, moral psychology, and ethical theory. Gary Watson is Provost Professor of Philosophy and Law at the University of Southern California and professor emeritus at the University of California, Riverside. His research concentrates on questions of human agency and accountability. Much of this research is collected in Agency and Answerability: Selected Essays. Susan Wolf is Edna Koury Distinguished Professor of Philosophy at the University of North Carolina, Chapel Hill. She is the author of Freedom within Reason and Meaning in Life and Why It Matters. Her work ranges broadly over issues in ethics, moral psychology, and value theory.

This page intentionally left blank

reasons and recognition

This page intentionally left blank

part one Reason, Value, and Desire

This page intentionally left blank

Then you have a look around, and see that none of the uninitiated are listening to us—I mean the people who think that nothing exists but what they can grasp with both hands; people who refuse to admit that actions and processes and the invisible world in general have any place in reality. —plato, Theaetetus 155e . . . as regards mere perception and receptivity to sensations he must count himself as belonging to the world of sense, but with regard to what there may be of pure activity in him (what reaches consciousness immediately and not through affection of the senses) he must count himself as belonging to the intellectual world. . . . —kant, Groundwork of the Metaphysics of Morals 4:451

1 The Activity of Reason* Christine M. Korsgaard

i. introduction

a venerable tradition holds that the difference between human beings and the other animals is that human beings are rational animals, that is, animals with reason. In the philosophical tradition, reason is often identified as the active capacity or power of the mind. This identification is implicit in the contrasts generally made between reason and sensation or perception, in the theoretical realm; and between reason and passion or desire, in the practical realm. It is also explicit in the work of some of our major philosophers: in Kant’s association of reason with the mind’s spontaneity, and in Aristotle’s doctrine of the active intellect or nous, for example. Putting these two ideas together—that reason is what distinguishes us from the other animals, and that reason is some special way the active dimension of the mind—we get the thesis that the human mind is active in some way that the minds of the other animals are not, and that this activity is the essence of rationality. My project in this paper is to articulate and defend that idea. I will offer an account of why the human mind is different from the minds of the other animals, and suggest a way of characterizing the activity of reason. Although my aim is to present a view 3


Reason, Value, and Desire

that I find compelling rather than to argue against alternative views, I do wish to contrast my view with two other views that have currency on the contemporary scene. First, there is the view that particular substantive reasons are the primary locus and source of normativity, a view I will call “substantive realism” about reasons. According to this view, the work of reason is to recognize and respond to these reasons. So reason as conceived by this view is a receptive rather than a purely active faculty. Second, there is the resulting view that rationality is something different from reason. I will begin by discussing some worries I have about these views, some possible problems which I think can be traced to their failure to do justice to the activity of reason.

ii. views that give priority to substantive reasons

When we talk about reason, we seem to have three different things in mind. One is the general faculty or capacity of reason, often, as I have just said, identified with the active dimension of the mind. Reason has also traditionally been identified with either the employment of, or simply conformity to, certain principles, namely rational principles, which may be taken to include some of the following: the rules of logical inference, the principles which Kant identified as principles of the understanding, canons for the assessment of evidence, mathematical principles, and the principles of practical reason. A person is called “rational” when her beliefs and actions conform to the dictates of those principles, or when she consciously guides her deliberations by them. And then finally, there are “reasons,” the particular considerations that count in favor of belief or action. To avoid confusion, I will refer to these particular considerations as “substantive reasons.” There are a variety of views that one might take about the relationships among these three things. According to substantive realism about reasons, versions of which have been advocated by Tim Scanlon, Derek Parfit, and Joseph Raz, among others, the primary item here is the third thing I mentioned, the substantive reason, a consideration that counts in favor of some belief, action, or attitude, and that has normative force. Reason, according to these philosophers, is the faculty that enables us to recognize and respond correctly to these substantive reasons. There are familiar worries, which I share, about both the metaphysics and the epistemology of substantive realism about reasons. There are worries about whether it is consistent with a naturalistic view of the world, and about how exactly we are supposed to know what reasons there are. But rather than going over this well-trodden ground here, I want to begin by talking about a problem less commonly discussed. The problem is that this conception of reasons seems to make them atomistic, in a way that makes it hard to see how they can do their job. What I mean is that, on this conception of reasons, it is hard to see why there is any general reason to expect that there will be, in any given case, one course of action, or even one belief, that the reasons, all things considered, will favor. Why should we expect reason so conceived to give us determinate answers about what to believe and to do, even in ideal conditions, when we know what all the relevant reasons are? That is, why shouldn’t it just turn out to be

The Activity of Reason


the case that some considerations favor a certain belief or action, while others are against it, and there is no more to say? At this point an interesting disanalogy may emerge between theoretical and practical reasons. In the case of theoretical reasons, many philosophers would want to argue that we expect the reasons to give us an answer to the question what to believe because belief aims at truth. Substantive reasons in favor of believing something are reasons for considering it to be true; and since the truth is one thing, the reasons must ultimately count in favor of believing that one truth. So the faculty of reason, on its theoretical side, is a faculty that enables us to recognize considerations that will guide us to the truth. In the case of practical reason, on the other hand, some philosophers would give a very different kind of answer to the worry that the reasons might not point to a conclusion: they will claim that practical reasons have a measurable dimension—weights or strengths—and that we arrive at a decision about what to do by determining where the balance of these weights or strengths, that is, the “balance of reasons,” lies. Before I go on, I should perhaps make it clear why I think these two explanations are different. We do sometimes talk about the weight of the evidence, and that may mislead us into thinking that the two cases are analogous. But they are not, for ordinarily we only talk about the weight of the evidence in cases where we have no conclusive reason for believing one thing rather than another, and the weight of the evidence only shows us how probable it is that something is true. More importantly, no one thinks that anything is true in virtue of being supported by the weight of the evidence, and therefore those who suppose that belief aims at truth would deny that a belief is everything that it should be in virtue of being supported by the weight of the evidence. But practical conclusions are supposed to be everything that they should be in virtue of being supported by the balance or weight of the reasons involved. For actions, being supported by the weight or strength of the reasons is all that is wanted—it is essentially the same property as being “right” or at least “all right”—while for beliefs, being supported by the weight of the reasons is a way of getting at something else, the truth. So if we accept this combination of views, we are left with a real, and rather puzzling, disanalogy. And because of that, it seems hard to construct a consistent story about why beliefs and actions are both the kinds of things that need to be supported by reasons. In the case of practical reasons, the idea of weights does have intuitive appeal in some cases, namely those in which we are comparing plainly commensurable things. The reason for giving me the painkiller is weightier than the reason for giving it to you, perhaps, if my pain is the kind that results from invasive surgery and yours is the kind that results from stubbing your toe. But few philosophers have anything much to say about how we go about balancing reasons in other kinds of cases, or have attempted to explain why substantive reasons should have this supposed dimension, “weight,” or “strength” that makes them commensurable. So I do not see any in-principle reason, on this theory, why our recognition of reasons should not leave us with a set of incommensurable considerations for and against various actions. Of course, on any theory of practical reason, there are problems getting decisive answers about what we should do


Reason, Value, and Desire

in all cases, so perhaps it would be unfair to press this objection too hard. Or some might be willing to accept it—in practical cases, perhaps, sometimes it really is the case that all we can get is a bunch of considerations for and a bunch against. But if reasons do not, at least in principle, tend to favor a single belief or action, they will be not be very good at doing their job, which is, after all, to enable us to determine what to believe and to do. This way of putting my complaint brings me to another worry I have about this theory. Human beings, for reasons I will explicate shortly, need reasons. We cannot determine our beliefs or actions without them. And according to this theory, when we look around us, we find them. But this seems like a mere piece of serendipity. The reasons are in no way generated by the problem that, as it happens, they solve; they just happen to be there when we need them. We need to make decisions, and lo and behold, we find around us the reasons we need in order to make those decisions, equipped with weights or strengths that will enable us to balance them up and arrive at a decision. I might put it this way: if reasons did not exist, we would have to invent them. To me, this suggests that, contrary to the theory under consideration, that is what we do. Finally, let me mention one last concern. Those who favor this theory think that reasons are indefinable: they simply exist, and cannot be analyzed or defined in terms of anything else. I also find this worrying, because it seems to me that reasons and causes have something in common, something that suggests that they are species of a genus. Causes are certainly not considerations in favor of their effects, but we sometimes call them “reasons,” and we use both reasons and causes to answer “why” questions. They are both, to adopt Aristotle’s term, aitiai, becauses. But if reasons and causes are species of a genus, it should be possible, at least in principle, to define them. What is this genus? This is something that, admittedly, is so basic that it is a little hard to talk about. The Greeks called it a logos, Kant called it a ground, we might also call it a story, in the widest sense, as when philosophers say to each other, “you have to have some story to tell about that.” On my own view, as I will explain later, reasons and causes are species of this genus, whatever it is, for we might define a reason as a ground of belief or action that has been endorsed by the person who believes or acts.

iii. disconnected rational requirements

But before I turn to that, I want to say a few words about another contemporary theory, which concerns the middle category of the three things we associate with reason, namely, “rationality.” Some contemporary philosophers would call a person “rational” when she responds correctly to substantive reasons, and some would say that that is what being “rational” amounts to. But others have reserved the word “rational” for what they consider to be a somewhat different set of ideas. For instance, in the work of John Broome, “rationality” is associated with a particular set of requirements, which govern the relations among our mental attitudes, but which do not seem to be connected to substantive reasons in the way we might expect. So for

The Activity of Reason


instance, we seem to be rationally required to conform our beliefs to the principles of logic. But, according to these philosophers, that does not mean that if you believe P and you believe that P implies Q, you then have a substantive reason to believe Q. Why not? Because your belief that P or your belief that P implies Q may be false, in which case we should say that you have no substantive reason to believe Q. Or to take a practical example, we seem to be rationally required to take the means to our ends. But according to these philosophers that does not mean that we have reasons to take the means to our ends, for if your end is a bad one, then you have no substantive reason to take the means to it. Of course, these philosophers grant that if P is true and implies Q, then there is a reason to believe Q, and so they grant that it will seem to a person who believes P and that P implies Q that he has a reason to believe Q. (I assume that they think this because they think we know in advance that the world is logically unified, a point I will come back to.) So one standard response to the sorts of argument that I have just mentioned is to distinguish “subjective” and “objective” reasons or “oughts”: the agent who falsely believes that P and also believes that P implies Q has a subjective, though not an objective, reason to believe Q. But some philosophers deny that there is such a thing as a subjective reason. For one thing, the notion of a subjective reason is slippery in a particular way: it is not clear which facts about your beliefs I am supposed to hold fixed when I identify your subjective reasons. Just your beliefs about the facts, say, or your beliefs about reasons as well? In order to avoid the resulting conundrums, some philosophers prefer to avoid talk of subjective reasons, and say simply that people are sometimes mistaken about what they have reason to do, which is whatever is implied by the substantive reasons that actually pertain to their case. They would deny that there is any sense in which people should act on the considerations they take to be reasons, just because they take them to be so. If you start from the view that I canvassed before—that reasons are objective, mind-independent, normative facts—this may look like a perfectly sensible response. Then all we mean by a “subjective reason” is your best estimate of what the objective reasons are, where those exist independently of your own thought processes, and are things about which you can be straightforwardly mistaken. But that view has its own conundrum, going back to medieval debates about the erring conscience. For isn’t there a sense, after all, in which if someone thinks he ought to do something, then he really ought to do it? Are we just confused when we say that a person should act on his own best judgment, and do what his conscience says? On my own view, there is a reason why we say things like this. As a Kantian, I believe that the normative force of reasons arises from autonomy: from the laws an agent gives to himself. That implies that if an agent tells himself to do something, there really is a sense in which he ought do it, even if he should have told himself to do something else. This view gives ontological priority to the kind of item usually identified as a “subjective reason.” On this view, a subjective reason is not your best estimate about what the objective reasons, independently of your thought processes, are. Rather, an objective reason is just a subjective reason that has arisen from those thought processes correctly carried through: one that upon full reflection you could will as a universal law, say.


Reason, Value, and Desire

Some philosophers will think that there can be no difference here: the fact that there is a correct answer, or an objective reason, means that all we can be doing when we reflect on what reasons we have is tracking that objective reason. To address this view, it would be necessary to explain why not every way of arriving at an answer counts as “tracking” it. The first step is to say that the “tracking” metaphor describes a way we arrive at mind-independent truth, whereas I am claiming that truths about reasons are mind-dependent. But we then need a characterization of mind-dependence and independence that makes sense of the claim that one can be wrong about a mind-dependent truth. Philosophers who think that what is meant by mind-dependence is that “thinking makes it so” will not be able to see their way to this conclusion. What I mean by mind-dependence is rather that the body of facts in question would not exist were it not necessary for human beings to conceptualize the world in a certain way, where the aim of that mode of conceptualization is not simply one of describing the way things are. Admittedly, for a Kantian, this leaves little that is mind-independent; but truths about empirical matters, where the concepts are ones we have adopted to describe things we find in the environment, are mind-independent. We use the empirical concepts in part simply because of what we find around us. Or to put it more carefully, what makes it necessary for us to use them is not just the way our minds work, but also the way the world independently of us works. (I mean what makes it necessary for us to use the concepts, not what it makes it correct to apply them in this or that way. That always depends on the world.) Those who believe that objective reasons are mind-independent in this sense treat them just like things we find in the environment. Those are large claims and I will not try to defend them further here. Let me return to the view under discussion. If rational requirements do not actually give us reasons, then what are they? Broome has proposed that they are “wide-scope” requirements on our attitudes, that is, roughly speaking, requirements that tell us not to have certain combinations of attitudes, rather than requirements to do or to believe particular things. A wide-scoper thinks that the requirement of modus ponens, for example, tells us that we ought not to combine a belief that P, a belief that P implies Q, and a belief that ~Q. We ought not to hold those three attitudes together. But, for reasons we have already canvassed, it does not tell us that if you believe that P and you also believe that P implies Q, then you ought to believe that Q. At most, it tells us that you ought to change one of your attitudes. The “ought,” as Broome puts it, is not “detachable.” But the claim that rational requirements are requirements on our attitudes seems to me to ignore the context in which we deploy rational requirements. On my view, rational requirements do not govern combinations of our attitudes. They govern thinking, the activity of thinking; and that means that they govern someone who is actively trying to determine what she has reason to believe or to do. And thinking has a certain temporal direction. To be rational is not just to have a set of attitudes that happen to conform to a rational requirement. It is to follow to a rational requirement, to take it as an instruction. Imagine trying to follow a recipe written by a widescoping chef. Normally one might say, “after you sauté the tomatoes and the

The Activity of Reason


mushrooms, you should add a little salt to the mixture.” The wide-scoping chef would insist that this cannot be right, since you might have put in black olives rather than mushrooms by mistake, and in that case, it would be much better not to add any salt. So the most that that the recipe can tell you is that either you should add a little salt or you should previously have added black olives by mistake. And of course, since the “should” is not detachable, there is no way to take a step. If you hope ever to get your dinner made, you want to avoid recipes written by the wide-scoping chef. If the job of rational requirements is to govern the activities of thought and deliberation, and the point of those activities is to direct us to belief and action, then rational requirements cannot be wide scope, since wide scope requirements cannot do that job. Of course, some contemporary philosophers have also denied that rational requirements do govern thinking, or at least that they govern it in the way I am proposing here, that we can be consciously guided by them. These philosophers argue that, because we reason from the content of our beliefs and intentions rather than the fact that we have them, we do not normally consciously employ rational principles. Their point is that, when I reason, I do not normally say to myself, for instance, “because I believe P and I believe that P implies Q, I ought to conclude that Q.” Furthermore, for the reasons I canvassed a moment ago, they also suppose that if I did say this to myself, I would be saying something false. This is because, after all, it is not really because I believe P that I ought to believe Q. It is because P provides evidence for or some other ground for an inference to Q. My believing P is neither here nor there. The fact that I might be wrong about P just brings this out vividly. These views have the odd implication that if human beings became self-conscious about what we are doing when we engage in reasoning—that is, when we conform our beliefs and actions to rational principles—we would suffer from a kind of inability to carry on the activity of reasoning, or at least to carry it on with any confidence that we have any reason to engage in it. And this is more than an abstract possibility, because lately a number of philosophers who hold these views have concluded that rational requirements are not normative after all. For them, the standards of rationality would be normative only if there were some general reason for conforming to them, and some philosophers, such as Niko Kolodny, have concluded that there is no such general reason. To me, it seems especially surprising that philosophers should see a problem here, for philosophers are, by profession, self-conscious about what we are doing when we are reasoning. While doing philosophy, we frequently say things like “because you are committed to this, you must also accept that,” and we say them to ourselves as well as to each other. Are we just mistaken in thinking that this makes any sense? I have already complained that, on the view that gives priority to substantive reasons, there is no explanation of why reasons exist: they seem, rather magically, to be on hand to meet our needs. On the views that leave rational requirements detached from the rest of the normative realm, similarly, we can begin to be puzzled about why rational requirements exist. Why do they exist, and where do they come from, especially if we are deluded in thinking they are normative? Here again, we may come to a distinction between theory and practice. As I said before, I suppose that those who favor these views think that logical requirements exist, or seem to exist, because they


Reason, Value, and Desire

think the world is in fact logically ordered. So if your views are not logically ordered, you know that one of them must be false. But no such explanation can be given for the requirements of practical reason, so lately some philosophers have been trying to come up with alternative explanations of why, for example, it seems to us as if the fact that we have adopted an end gives us some reason to take the means to it. I believe that rational requirements exist because they describe the activity of reason, so now I will turn to my own view.

iv. the origin of reason

In the Kantian conception that I favor, the three aspects of reason—that is, the faculty of reason, rational principles, and substantive reasons—are closely related. The faculty of reason is not identified merely as the ability to recognize and respond to reasons. The faculty of reason is identified rather as the active dimension of the mind, and rational principles are then identified as those that describe or constitute rational activity. They are constitutive principles of rational activity. When those principles are applied by the person who is trying to work out what to believe or to do, they pick out the substantive considerations that we then regard as reasons. That description of course is schematic, and I will not be able, on this occasion, to give arguments for particular rational principles, or to show that those principles do enable us to pick out substantive reasons. What I do hope to do here is to convey why, in general, it makes sense to regard rational principles as constitutive of the activity of reason, and how the resulting view deals with some of the issues I have raised about the other views. Let me begin by explaining what I think it means that we are rational animals. I believe that the source of reason is a particular form of self-consciousness that characterizes the human mind. As human beings, we are conscious of the potential grounds of our beliefs and actions as potential grounds. A contrast will show what I mean. A nonhuman animal is guided through her environment by means of her perceptions and her desires and aversions: that is, by her instinctive responses and the other desires and aversions she may have acquired through learning and experience. Her perceptions constitute her representation of her environment, and her instincts, desires and aversions tell her what to do in response to what she finds there. In fact, I believe that for the other animals, perceptual representation and desire and aversion are not strictly separate. Either through original instinct or as a result of learning, a nonhuman animal represents the world to herself as a world that is, as we might put it, preconceptualized and already normatively or practically interpreted. The animal finds herself in a world that consists of things that are directly perceived as food or prey, as danger or predator, as potential mate, as child: that is to say, as things to-beeaten, to-be-avoided, to-be-mated-with, to-be-cared-for, and so on. To put it a bit dramatically—or anyway, philosophically—an animal’s world is teleologically organized: the objects in it are marked out as being “for” certain things or as calling for certain responses. I believe this because I think it is hard to see how perception could have

The Activity of Reason


been of any use to the relatively unintelligent animals in which it first evolved if something like this were not the case. Perception could not just provide a simple animal with information on the basis of which the animal had to figure out what to do, so it must be that it tells the animal what to do. So these normatively or practically loaded teleological perceptions serve as the grounds of the animal’s actions—where the ground of an action is a representation that causes the animal to do what she does. The exact ways in which these normatively loaded perceptions operate on an animal to produce his actions probably differ in ways that can be ranged along a scale, depending on what sort of representations the animal has, or what sort of consciousness he has of them. Primitive animals may respond more or less mechanically to these perceptions; more sophisticated animals may operate with something more like concepts or categories of “food” or “predator” or “threat” to which they respond intelligently; and yet more sophisticated animals may even be aware that they and their fellows find certain things desirable or fearful. Exactly how any given kind of animal’s representations give rise to his actions is a matter for further investigation, both philosophical and empirical. But however it may be with the other animals, there is no question that we human beings are aware, not only that we perceive things in a certain way, but also that we are inclined to believe and to act in certain ways on the basis of these perceptions. We are aware not only of our perceptions but also of the way in which they tend to operate on us. That is what I mean by saying that we are aware of the potential grounds of our beliefs and actions as potential grounds. And I believe that this awareness is the source of reason. For once we are aware that we are inclined to believe or to act in a certain way on the ground of a certain representation, we find ourselves faced with a decision, namely, whether we should do that— whether we should believe or act in the way that the representation calls for or not. Once the space of reflective awareness—reflective distance, as I like to call it—opens up between the potential ground of a belief or action and the belief or action itself, we must step across that distance, and so must be able to endorse the operation of that ground, before we can act or believe. What would have been the cause of our belief or action, had we still been operating under the control of instinctive or learned responses, now becomes something experienced as a consideration in favor of a certain belief or action instead, one we can endorse or reject. And when we can endorse the operation of a ground of belief or action on us as a ground, then we take that consideration for a reason. What this means is that the space of reflective distance presents us with both the possibility and the necessity of exerting a kind of control over our beliefs and actions that the other animals do not have. We are, or can be, active, self-directing, with respect to our beliefs and actions to a greater extent than the other animals are, for we can accept or reject the grounds of belief and action that perception and desire offer to us. We can actively participate in giving shape both to the conception of the world in light of which we act and to the motives on the basis of which we act—and ultimately, in both ways, in giving shape to ourselves. And it is the same fact that we now both can have, and absolutely require, reasons to believe and act as we do.


Reason, Value, and Desire

So here is part of the answer to one of the questions I raised earlier: why there are such things as reasons, substantive reasons. There are reasons because self-consciousness transforms the grounds of our beliefs and actions—the perceptions and impulses that would have caused them if we lacked this form of self-consciousness—into substantive reasons. This account of why reasons exist does link them to the problem that they solve: in order to believe and act, we need to endorse some of the potential grounds of our beliefs and actions, and when we do that, we get substantive reasons. And reasons and causes do have something in common, namely that the reasons for our beliefs and actions, at least the initial ones, are the very sorts of things that would have caused our beliefs and actions had self-consciousness not intervened. They are grounds of belief and action that we have endorsed.

v. identifying the activity of reason

But how do we exercise the self-directing power that this form of self-consciousness gives us? How do we pick out which grounds to count as reasons? As I said before, my view is we use rational principles to pick out the substantive reasons, and rational principles, in turn, are the constitutive principles of the activity of reason. So in order to proceed, we need to know what the activity of reason consists in. Now obviously, if all that we could say about the activity of reason is that it is “evaluating the grounds of our beliefs and actions,” or “justifying our beliefs and actions” then it will look as if the substantive reasons need to be in place before reason—the general capacity of reason—can do its job. No doubt this is part of the attraction of substantive realism about reasons. The substantive realist supposes that all we can be doing when we evaluate the grounds of our beliefs and actions is asking whether they “really are” reasons, where that is a question about whether they have some objective characteristic— intrinsic normativity, counting in favor—that cannot be specified in any other way. So I take the interesting question here to be whether there is some other way of characterizing the activity of reason, some other way of saying what we are doing when we evaluate the grounds of our beliefs and actions. Ask yourself: why do we need to evaluate the grounds of our beliefs and actions? What makes that necessary for us, and not for the other animals? Is it just because they are not smart enough to see that beliefs and actions should be supported by reasons? Or is it because they lack a receptive faculty of reason, the way we lack sonar, and therefore they just cannot see the reasons? I have already suggested that what makes it necessary for us to justify our beliefs and actions is the form of self-consciousness involved, which enables us to call the grounds of our beliefs and actions into question. But when we do that, we are, at the same time, calling two other things into question: on one side, the way of representing or conceptualizing the world that would be given by our instincts if we did not have that form of selfconsciousness, and on the other side, our own nature as the source of that way of conceptualizing and responding to the world. So when we are faced with the task of justifying our beliefs and actions, it is because we are faced with two other tasks, or

The Activity of Reason


we could just as well say two other opportunities: we both can, and need to, construct a new way of conceptualizing the world, and we both can, and need to, construct or reconstruct our own nature, as the subject of that conception and as a source of responses to the world. Those two tasks constitute the activity of reason. The other animals do not need to justify their beliefs and actions, because their way of conceptualizing and responding to the world is simply given to them by their teleological perception, by the instinctive ways in which they represent the world to themselves. There is another way of describing these tasks that I think is helpful here, because it helps us to see why these activities should be shaped and guided by rational principles. When we become aware that we are representing the world to ourselves, when we turn our attention away from what we perceive and onto the fact that what we are doing is perceiving, then there is a way in which the world loses its unity. What was once simply given to us as the environment is now given to us as a heap of perceptions, or rather experiences, and it is now up to us to put them back together into a picture of the world. And in a similar way where once upon a time we always knew what to do in response to a situation, our own possible responses are now given to us as a heap of desires and fears and impulses, and it is up to us to put ourselves back together. The principles of rationality are constitutive of the activity of reason, I suggest, because they are principles of unification. In the practical case, here is the idea: I believe that in order to regard your movements as actions that you can attribute to yourself as their author, you have to see those movements as arising from yourself as a whole, rather than from something working in you or on you. The twitch comes from your muscle, the slip from the ice below, but the walking—that comes from you, from you as a whole. Elsewhere I have written about how the principles of practical reason—Kant’s categorical and hypothetical imperatives—serve to unify our wills so that we can regard ourselves as the sources of our actions. I will not try to repeat those arguments here. Instead I want to talk about how it might work in the case of theoretical reason. To conceive yourself as a knower, in my view, is to conceive yourself as able to form a conception of the world that will enable you to find your way around in it and to act effectively in it. I include “act effectively,” because I want to emphasize that I do not just mean a conception of the world that will enable us to predict and explain events. I also mean a way of conceptualizing the world that will answer to our needs as agents. As I have argued elsewhere, to conceive ourselves as agents is to conceive ourselves as the autonomous and efficacious sources of certain events in the world: that is, as the self-determining causes of certain effects in the world. However exactly we work the details out, if something along these lines is correct, the conception of the world as causally ordered in a general way is essential to our conception of ourselves as agents, a conception that I believe is forced upon us in the first person deliberative standpoint. If that is true, then that the world is, at least in a general way, causally ordered, cannot just be an empirical discovery. For these reasons, I think we are rationally required to conceive the world as causally ordered, at least in some general way. For the world to be the sort of place in which you can find your way around and act effectively, it must be a unified place. What that means is that the relations between


Reason, Value, and Desire

the various things in the world can be traced and established. If we can say nothing about how two things or events or regions of space-time are related to each other, we cannot think of them as parts of a unified world. If we cannot trace causal relations, in particular, we cannot act effectively. So it is the business of a conception of the world to establish these various relations. Further argument is required, of course, but I suppose that we may think of the relations in question as logical, spatiotemporal, and causal. Now we may raise a question about why exactly we suppose that the world admits of a conceptualization that will unify it in these ways. Or rather, since saying “the world” makes it sound as if we already know that what we are confronted with is one unified thing, I should say instead that we may raise a question about why exactly we suppose that what we find ourselves confronted with in experience admits of a conceptualization that will unify it in this way. One familiar form of philosophical argument reminds us that the unity of the mind and the unity of its object are interdependent. Unless we conform our beliefs to logical and rational principles, our minds themselves are a mere heap of unrelated ideas or theses. And a mere heap of unrelated ideas or theses is not about anything, and therefore cannot count itself as thinking about anything or knowing anything. So our conception of ourselves as possible knowers of a world independent of our minds, a world that we can think about, depends on our idea of the world itself as something of which we might possibly form a unified conception. This explains, to take one example, why we have to take theoretical reasons to be both universal and what I call “public,” or agent-neutral, in their normative force— why that is a rational requirement. If you are to think of your experience as a perception of an object, and perception as a way of knowing that object, then you have to think that, suitably situated, another perceiver with the same sort of perceptual equipment would be having that experience too. Now you might ask, if I am constructing a conception of the world, couldn’t I just construct a world that was my world, which only existed for me and nobody else? But the answer is no, because if you are to think of your experience as perception of an object, and perception as a way of knowing that object, then you have to think that if you were to come back to the same place tomorrow, and nothing had changed in the meantime, you would have the same experience again. And that is the same thought as the thought that if another perceiver were suitably situated, he would have the same experience: both scenarios, after all, just involve a change of position. If you cannot have that thought—that if you come back to the same place later, and nothing has changed, you will have the same experience again—then you cannot think of your experience as perception of an object, and of yourself as the knower of that object, and your mind shatters into a mere heap of unrelated experiences. It follows that if you are to take “I saw it” as a reason to believe it, you must take it as a reason with universal and agent-neutral or “public” normative force. So it is not that we know in advance, somehow, that the world conforms to the principles of theoretical reason, and we should therefore expect true beliefs to do so as well. Rather, that the world conforms to the principles of theoretical reason is a presupposition of

The Activity of Reason


the world’s being the sort of place we can think about and know about at all. And I think a similar argument could be given for the normativity of the principles of the other kinds of connectedness I just mentioned—causal relatedness in space and time, the kind of relatedness that connects one event to another. Causal relatedness in space and time is a presupposition of the world’s being the sort of place we can find our way around in and also act effectively in—that is, cause things to happen in— ourselves. And since we are faced with the task of constructing a conception of the world that makes that possible, we must suppose that the world can be conceptualized in that way.

vi. an anti-realist conception of rational activity

I have been suggesting, in a very general way, that we are committed to conceptualizing the world as conforming to rational standards, because a conception of the world that does not do that cannot do its job, which is to enable us to find our way around and act in it. The theoretical activity of reason is to construct such a conception. This picture of what reason does is of course a Kantian one, and I want to emphasize one implication of that, and also to respond to a possible objection that it raises. First, the implication. I do not take it to be the only or even the primary desideratum of a way of conceptualizing the world that it should be “true.” Propositions are true when the concepts that appear in them are applied correctly; but I do not suppose that ways of conceptualizing the world are themselves simply true or false. I think of them on the analogy of maps, since they are devices that enable us to find our way around. And, as is the case with maps, they are answerable both to the world they represent, and to the conceptual capacities of their users. And in some cases they are also answerable to their suitability for specific cognitive tasks. A tourist exploring the city center on foot will prefer one of those maps on which the cartographer actually draws little pictures of the buildings with their names written across them. But this style of representation and level of detail is not wanted by someone driving across a nation on its highways. For her purposes it is better if whole towns are represented by tiny dots, so that the spatial and directional relations between them are what emerge as perspicuous. The more detailed map does not give us more truth, or less. It gives us different truths and is more suitable for a certain purpose. Of course this anti-realist way of thinking about what we are doing when we conceptualize the world is controversial. But, leaving that aside—and here is the objection— it may also make it seem as if I have not after all offered an alternative description of what we are doing when we evaluate the grounds of our beliefs. For surely, you might say, when we evaluate the grounds of our beliefs, at least in an everyday way, what we are interested in is not whether we are conceptualizing the world in the best possible way for our cognitive purposes, but simply whether the belief is true. In response, I want to make a comparison, and also to pick up the practical side of the question again. In other work I have argued that whenever you make a choice, you are also at the same kind constructing your identity. The argument goes roughly like


Reason, Value, and Desire

this. From a third-person point of view, outside of the deliberative standpoint, it may look as if what happens when someone makes a choice is that the strongest of his conflicting desires simply wins. But that is not the way it is for you, from your firstperson point of view, when you deliberate. When you deliberate, it is as if there were something over and above your desires, something that is you, and that chooses which of them to act on. This means that you take the principle or law on the basis of which you choose to be expressive of yourself: your principle speaks for you. On this basis I have argued that our practical principles are expressive of our conceptions of our practical identity. The relevant point here is that the picture I have in mind is not that there is a two-step process: step one, you first choose some way of identifying yourself, and step two, you proceed to act in accordance with its principles, like someone following a list of rules. Rather, the idea is that determining what we have reasons and obligations to do—that is, adopting maxims or practical principles—is at the same time engaging in the work of identity construction, the ongoing project of a human life. And I am not claiming that when we make everyday choices, we are normally thinking about our identity, rather than about what it is right to do, although I suppose we do think explicitly about our identity in this context sometimes. I do argue, however—again I will not try to summarize the argument here— that the fact that we are engaged in identity construction helps to explain why the process of thinking about what we have reason to do is governed by rational standards, because of the ways in which those standards secure the unity of the self and of agency. In the same way, I am proposing now that determining what we have reason to believe is at the same time engaging in the ongoing work of constructing a conception of the world, and that this helps to explain why that process must be governed by rational standards. To see this it helps to think about the nature of believing. Almost all philosophers would agree that believing P is related to the following things: being prepared to affirm P; being prepared to treat P as a premise in your reasonings about other matters; being prepared to accept the logical consequences of P; and being prepared to act as if P were true. Let us call these things the concomitants of belief. Some philosophers suppose that a belief is a particular mental state, something that simply exists or not, and that the concomitants of belief serve as evidence as to whether someone is in that mental state or not. If, say, someone sincerely affirms something but does not act as if it were true, the evidence is unclear. Other philosophers suppose that the concomitants of belief are constitutive of belief: to say that you do those things is what it means to say you believe something. If someone sincerely affirms something but does not act as if it were true, we seem to have a contradiction on our hands; perhaps we will be tempted to deny that he could have been sincere after all. I myself take the concomitants of belief to be constitutive principles of believing: normative standards that arise from the very nature of believing. In other words, what I am proposing is that, for a rational animal, believing itself is an active state, it is doing something, it is an activity: it is representing the world to yourself in a certain way. I think that failure to see this is part of where the second of the two views that I described earlier—the view that rationality is something

The Activity of Reason


separate from reason—goes astray. Those who hold these views tend to conceive of beliefs and intentions statically, as mental states or attitudes, and therefore regard rational standards merely as standards by which we evaluate combinations of attitudes. The point is somewhat difficult to articulate, but I take this tendency to exemplify a general source of philosophical problems, especially in the philosophy of mind. People tend to reify mental activities into mental states. It is symptomatic of this that philosophers with these views talk about “forming intentions” rather than “intending.” Being “formed” makes the intention an entity, something that can take up space in the mind. And the mind is then conceived as a kind of place that these states occupy. All mental phenomena then seem rather like qualia, in the sense that they are held, or can be held, in an interior gaze. This makes consciousness, and mental life more generally, seem more mysterious than it is, for it is not all as mysterious as qualia. Much of we call the mental, I believe, is actually things that we do. However that may be, on my view, belief is not simply a mental state or attitude: it is a commitment to going on in a certain way. I think it is a commitment to constructing your conception of the world in a certain way, where that involves a commitment both to certain truths and to the possibility of forming a unified and useable conception of the world that includes those truths. The fact that someone may affirm something sincerely but not act as if it were true, and other such divided responses to the concomitants of belief, simply shows that for a rational animal, believing is something that can be done well or badly—and if badly, the failures can be of various kinds. One may be inconsistent, or wavering, or fail to follow through. That is why our beliefs, like our actions, call for justification—because in a rational animal, believing can be done more or less well. And notice that on this view of what rational believing is, it makes perfectly good sense for us to say, both to ourselves and each other, that because you believe both P and that P implies Q, you ought to believe Q. It is exactly like saying that because you promised to do A and you cannot do A without doing B, you ought to do B. It is a reminder of the normative commitments that are constitutive of taking a certain kind of action, in this case mental action—believing something, that is, representing the world to yourself in a certain way.

vii. conclusion

Let me conclude by summarizing the view I am proposing. What does it mean to be a rational animal? A nonhuman animal finds herself in a teleological world, a world in which things are already marked out for her as her food, her mates, her offspring, her enemies. It is a conception of the world in her own image, as we might say, that is given to her by her instincts, and it tells her what to believe and to do. But in the human mind, the development of a certain form of self-consciousness—consciousness of the potential grounds of our beliefs and actions—breaks up this teleological conception of the world. It shatters the world into a mass of perceptions or experiences, and the self into a mass of desires and fears and impulses, and in doing so, it creates both the opportunity and the necessity for reconstruction. We are faced with


Reason, Value, and Desire

the task of unifying the mass of perception into a conception of the world that enables us to find our way around and act effectively, and of unifying the mass of desires, fears, and responses into a self that can stand behind its movements as their author and so claim them as its actions. It is by imposing rational principles upon on the self that we unify ourselves into agents, and it is by imposing rational order on our perceptions that we form a unified conception of the world. This conception of reason differs from the views I described earlier in systematic ways, ways that spring from the fact that it conceives reason as an active rather than as a receptive faculty. On this conception, rational requirements exist because they describe the activities of reason, and reasons exist because we need them in order to determine our beliefs and actions. Reasons and causes do have something in common, for at least in the first instance, reasons are the descendants of causes, the sorts of things that would have caused our beliefs and actions had self-consciousness not intervened. Our rational beliefs and intentions are not mere mental attitudes, but active states of normative commitment, and it makes perfect sense to say that they can commit us to other beliefs and intentions. The work of reason in theory and practice is parallel rather than disanalogous, except that theoretical reason aims at unifying the experienced world, and practical reason aims at unifying the self. And the reason that we human beings, unlike all the other animals, must justify our beliefs and actions, is because we alone among the animals must actively carry out the work of constructing a conception of the world and a self who is both a knower of that world and an agent within it: because we alone among all the animals have to engage in the activity of reason. notes * “The Activity of Reason” originally appeared in the Proceedings and Addresses of the American Philosophical Association 83(2) (2009). 1. Quoted from the translation by M. J. Levett, revised by Miles Burnyeat, in Plato: Complete Works, ed. John M. Cooper (Indianapolis, Ind.: Hackett, 1997). 2. Quoted from the translation by Mary Gregor (Cambridge: Cambridge University Press, 1998). 3. When her beliefs and actions conform to these principles, or, some philosophers would prefer to say, when her beliefs and intentions conform to these principles. As I will note later, some contemporary philosophers think that rationality concerns only the relations among a person’s mental attitudes, and some of these philosophers also think that an intention is something separable from an action itself, perhaps a mental state that causes it. Those philosophers would not allow that rationality concerns the relation between attitudes and actions themselves. I believe that intentions are embodied in actions and inseparable from them (see my “Acting for a Reason,” in Constitution of Agency (Oxford: Oxford University Press, 2008) especially 227–29, and Self-Constitution: Agency, Identity, and Integrity (Oxford: Oxford University Press, 2009), 6.3.2, 124–25), so I would reject these ideas. Essentially there are three reasons for believing in the separability of intention from action. One is the idea that we sometimes form intentions well in advance of the time of action. We have the intention, and yet we have not acted, and so they must be separate. I think, however, that as soon as you “form an intention”—that is, make a decision—you begin to act, because you must immediately begin

The Activity of Reason


to deliberate about all of your actions in such a way that the future action will be possible, and that is part of what it is to carry out your intention. The second is the idea that someone may be prevented from carrying out an intention. Suppose, for instance, that I form an intention but then I am immediately seized with an attack of paralysis. (Making it “immediately” blocks the force of my first argument.) Surely I had the intention, but did not act, so they must be two separate things? But it does not follow from the fact that we can identify two aspects of a thing as separate in a defective case that they are separate in a nondefective case. When you are dead, your life becomes something we identify as separable from your body, but that does not show that when you were alive, you must have had a separable soul. An intention is like that—it is the life of an action, its form, its soul, and as such, it makes it the action that it is. I mean this in the Aristotelian sense. In Aristotle’s account, we can distinguish the soul from the living body conceptually, but the soul is the cause of the body only in the sense of “formal cause” not “efficient cause.” In the same way, we can distinguish the intention from the bodily movements (not from the action), but that does not mean that the intention is a separate thing that caused those bodily movements, any more than the soul is a separate thing that causes the living body. Rather, the intention is the form of the bodily movements. (I thank Drew Schroeder for drawing my attention to this objection.) The third reason is that separating intention from action allows us to describe akrasia as the failure to carry out an intention. I do not think that is the correct way to describe akrasia, but explaining why would raise issues too large to be raised in this already bloated footnote. 4. For Scanlon, see, for instance, What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998), ch. 1; “Metaphysics and Morals” (in Proceedings and Addresses of the American Philosophical Association 77(2) [2003]); “Reasons: A Puzzling Duality?” in Reason and Value: Themes from the Moral Philosophy of Joseph Raz, ed. R. Jay Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith (Oxford: Oxford University Press, 2004); “Structural Irrationality” (in Common Minds: Themes from the Philosophy of Philip Pettit, ed. Geoffrey Brennan, Robert Goodin, Frank Jackson, and Michael Smith (Oxford: Oxford University Press, 2007); and his forthcoming Locke Lectures, Being Realistic about Reasons. For Raz, see Practical Reason and Norms (Princeton, N.J.: Princeton University Press, 1990); Engaging Reason (Oxford: Oxford University Press, 1999); “Reasons: Explanatory and Normative” in New Essays on the Explanation of Action, ed. C. Sandis (New York: Palgrave/McMillan, 2009); “Reasons: Practical and Adaptive” in Reasons for Action, ed. David Sobel and Stephen Wall (Cambridge: Cambridge University Press, 2009). For Parfit’s views, see his forthcoming book, On What Matters; and “Rationality and Reasons,” in Exploring Practical Philosophy: From Action to Values, ed. Dan Egonsson, Jonas Josefsson, Björn Petterson, and Toni Rønnow-Rasmussen (Aldershot, UK: Ashgate, 2001), 17–39. 5. Notice that the claim here is not merely that we have a faculty that enables us to represent the world to ourselves with some accuracy. We do have that, but it is not reason; every animal with perception has that. The claim is rather that we come equipped with a faculty that enables us to recognize the right way to move from that perceptual representation to what, according to this view, is supposed to be a uniquely accurate conception of the world. 6. The view might be more complicated: it might involve certain reasons that exclude or silence certain other reasons, for instance. But for my purposes here, it is all the same: I do not see what justifies the assumption that the reasons themselves will come equipped with some property that makes it clear how they bear on one another and so that enables us to “balance” them and reach a conclusion. I thank Barbara Herman for the reminder.


Reason, Value, and Desire

7. Actually, I think that justifying this judgment is more complicated than it looks. See the discussion of aggregation in my “Interacting with Animals” in The Oxford Handbook on Ethics and Animals, ed. Tom L. Beauchamp and R. G. Frey (Oxford: Oxford University Press, 2011.) 8. Broome introduced the idea in “Normative Requirements,” in Normativity, ed. Jonathan Dancy. (Oxford: Blackwell, 2000), and has written a number of papers on it, including “Normative Practical Reasoning” (Proceedings of the Aristotelian Society 75(suppl.) (2001): 175–93; “Practical Reasoning” (in Reason and Nature: Essays in the Theory of Rationality, ed. José Bermùdez and Alan Millar, [Oxford University Press, 2002], 85–111), “Reasons” (in Reason and Value: Themes from the Moral Philosophy of Joseph Raz, ed. by R. Jay Wallace, Michael Smith, Samuel Scheffler, and Philip Pettit [Oxford: Oxford University Press, 2004]) and “Does Rationality Consist in Responding to Reasons?” (Journal of Moral Philosophy 4(2007): 349–74). 9. Another worry that motivates this line of thought is the worry that we can “bootstrap” reasons into existence. By adopting an end we create a reason to take the means; at the limit, if every action is a means to itself, we can create a reason for doing something just by deciding to do it. Since I think there is a clear sense in which we do create reasons, this does not worry me in general. If you can create a reason by making a promise, why can’t you create one by making a decision? But of course not everyone thinks you can create a reason by making a promise; some philosophers think making a promise only gives you a reason by activating other, standing, reasons, like the reason not to disappoint someone’s expectations. These issues are all connected. In any case, I do not think that we can “just decide” to do something any more than we can “just decide” to believe something: in both cases, we have to at least persuade ourselves that we are determined by a consideration that has the form of a law. 10. See The Sources of Normativity (Cambridge: Cambridge University Press, 1996), 254–58. 11. Later I will explain why human beings have to conceptualize the world in terms of reasons and causes. See also my “Realism and Constructivism in Twentieth-Century Moral Philosophy,” in Constitution of Agency. 12. See Sources of Normativity, 1.4.8, 44. 13. Niko Kolodny comes close to saying this when he identifies rational requirements as applying to processes rather than states. See his “Why be Rational?” (Mind 114[455] [2005]: 509–63) and “State or Process Requirements” (Mind 116[462] [2007]: 371–85). 14. Niko Kolodny, in “Why be Rational?” Of course, if you think that rational requirements are wide scope, then rationality does not require that we have any particular beliefs or attitudes: the most that rationality could require of us is that we do a little housekeeping on our attitudes, making sure that the contradictions and other incompatibilities somehow get weeded out. It is rather like cleaning out the attic. So it is no wonder that those who conceive of being rational this way think we might have no reason to do it. 15. As I will argue below, it is a particular feature of human life that we human beings have control of, and therefore take responsibility for, our beliefs and actions. The way we do that is by reasoning—thinking in accord with rational principles. And that is a feature of human life that we try to realize in a special way when we do philosophy. So I find it rather staggering that a philosopher should suggest that rationality is not normative. To me, this conclusion seems like a reductio ad absurdum of the views that lead to it. 16. See Jay Wallace, “Normativity, Commitment, and Instrumental Reason” (in Normativity and the Will. Oxford: Clarendon Press, 2006) and Joseph Raz, “The Myth of Instrumental R ationality,” Journal of Ethics and Social Philosophy 1(1) (April 2005): 1–28). 17. It is worth noting that if there are narrow-scope requirements, there are also wide ones. If having an end requires you to take the means, then you certainly should not both have an

The Activity of Reason


end and not take the means. You can derive a wide-scope requirement from a narrow-scope one, but you cannot derive a narrow-scope requirement from a wide-scope one. So if we can explain the narrow-scope requirements in terms of the activity of reason, as I suggest below, then we can explain the wide-scope ones as well. If only wide-scope requirements exist, I believe, their existence is inexplicable. 18. For more on the notion of constitutive standards and principles see Self-Constitution, 2.1, 27–34. 19. I give arguments for the instrumental principle in “The Normativity of Instrumental Reason,” (in Constitution of Agency), and in Self-Constitution, 4.3, 68–72. I give arguments for the categorical imperative in “Self-Constitution in the Ethics of Plato and Kant,” (in Constitution of Agency) and in Self-Constitution, 4.4, 72–80. I give arguments concerning the ability of the categorical imperative to pick out substantive reasons (i.e., to give us determinate moral obligations) in “Kant’s Formula of Universal Law,” in Creating the Kingdom of Ends (New York: Cambridge University Press, 1996), and in chapter 9 of Self-Constitution. 20. If you are inclined to say, “no, it’s instinct that tells the animal what to do,” I will reply that I am describing what I think is the form that instinct takes: the animal comes equipped to respond in certain ways to certain perceptual cues, and then expands this set of responses through learning. See Self-Constitution, 6.1, 109–32. 21. Obviously, this is an empirical claim, and I cannot prove it. Were we to find another animal with this kind of self-consciousness, it would be a rational animal. 22. This account of the nature of reason is taken with some modifications from Sources of Normativity, especially 3.2.1, 92–94 and “Fellow Creatures: Kantian Ethics and Our Duties to Animals,” (in The Tanner Lectures on Human Values, ed. Grethe B. Peterson. Vol. 25/26 [2005]: 85–87).). Salt Lake City: University of Utah Press; and on the Tanner Lecture Web site at: 23. See Self-Constitution, 6.2.1–6.2.5, 121; 6.4.1–6.4.2, 125–26. 24. See “The Normativity of Instrumental Reason,” and “Self-Constitution in the Ethics of Plato and Kant” (both in The Constitution of Agency), and Self-Constitution, especially 4.4–4.5 and ch. 6–9. 25. See Self-Constitution, ch. 5. 26. I do not think it commits us to the view that every event has a cause. 27. The ancestor of these arguments is Aristotle’s argument, at Metaphysics 4.41006a15, that you can get someone to agree to the principle of non-contradiction if you can just get someone to say something and mean something by it. 28. I discuss the publicity of practical reasons in Sources of Normativity, 4.2.1–4.2.12, 132–45, and in Self-Constitution, 9.4.5–9.7.6, 191–206. 29. In Sources of Normativity, ch. 3, and in Self-Constitution. See especially 1.4, 18–26 and 2.4, 41–44. 30. In Sources of Normativity, 3.1.1, 100–2. 31. See Self-Constitution, 1.4.4–1.4.6, 20–22; 2.4.1–2.4.2, 441–44. 32. See Self-Constitution, 4.3–4.4, 68–80, and ch. 7–9. 33. These arguments parallel the ones I made about the nature of volition in “The Normativity of Instrumental Reason.” those arguments treat “willing the means” as what I am here calling a “concomitant” of “willing the end.” 34. It does not follow from the idea that believing is an activity in the sense described in the text that one simply can “decide to believe,” but then as I mentioned earlier, I do not think that


Reason, Value, and Desire

one can simply “decide to act” in the sense that would be parallel to the worrisome sense of “deciding to believe,” either. 35. Kolodny, with his emphasis on “process” requirements, is an exception to this last point. 36. This comparison will work for those who think that normative commitments are constitutive of promising, but not for those who think that the obligation of promising arises from the need to avoid certain harms or disappointed expectations. This is what I meant in note 9 when I said that these issues are all connected.

2 Valuing Samuel Scheffler

human beings are valuing creatures. One may value one’s privacy, or one’s relationship with one’s brother, or a friend’s sense of humor, or the opinion of a trusted advisor. But what is valuing? David Lewis says, “It is some sort of mental state, directed toward that which is valued. It might be a feeling, or a belief, or a desire.” He adds, parenthetically, that valuing might instead be “a combination of these; or something that is two or three of them at once; or some fourth thing.” But he proposes to set these more complicated possibilities aside, and to look for a simpler account. After quickly dismissing the idea that valuing is a feeling or a belief, Lewis defends a version of the view that it is a form of desiring. The simplest version of this view, though not the one that Lewis himself endorses, is that to value something just is to desire it. This view has been surprisingly influential, and it may seem plausible in some cases. For example, it may seem plausible that to value one’s privacy just is to desire one’s privacy. But does it seem plausible that to value one’s friend’s sense of humor is to desire one’s friend’s sense of humor? What does it even mean to say that one desires one’s friend’s sense of humor? On the most natural interpretation, it means that one would like to possess one’s friend’s sense of humor oneself. But this is clearly not what is meant by saying one values one’s friend’s sense of humor. If the suggestion that valuing is simply desiring is to have any plausibility, therefore, desire must be understood in a very special—and very broad—sense. Perhaps, for example, to desire something is to have a favorable attitude toward it. To value one’s friend’s sense of humor would then be to have a favorable attitude toward it. This may seem plausible. But is it plausible to equate desiring something with having a favorable attitude toward it? Offhand, it seems that I may have a favorable attitude toward something without desiring it. I have a favorable attitude toward Vaclav 23


Reason, Value, and Desire

Havel, but I do not desire him, whatever that might mean, nor do I desire to see him, meet him, or talk to him. Philosophers have offered other reasons for rejecting the reduction of valuing to desiring. Many have argued that valuing and desiring can come apart in cases of internal conflict or psychological disturbance. In some cases of this kind, it seems that a person may desire something without valuing it. As Harry Frankfurt has emphasized, for example, an addict may desire the drug to which he is addicted, yet despise that very desire and wish to be rid of it. Similarly, a person may desire to engage in certain types of sexual activity yet disapprove of those activities and regard them as sinful or degrading. Or, as Gary Watson has suggested, a woman may have a momentary urge to drown her shrieking child in the bathtub, yet place no value at all on the drowning of her child. There are also cases of emotional disturbance in which it seems that a person may value something without desiring it. For example, Michael Smith has argued, on the basis of an example of Michael Stocker’s, that a person may value something, but as a result of depression or despair, feel no desire or motivation to pursue it. This focus on cases of compulsion and conflict has led some philosophers to suppose that although valuing and desiring can indeed diverge, so that valuing cannot be reduced to desiring, such divergence is always a manifestation of deviance or irrationality. There are, however, quite mundane cases, involving no compulsion or inner conflict or emotional disturbance, in which valuing and desiring also come apart, in the sense that one of these concepts seems straightforwardly to apply but the other doesn’t. In cases of trivial desires, for instance, the language of valuing seems overblown and out of place. I may want to thumb through a magazine in the doctor’s office, or to eat some jellybeans, or to retie my loose shoelace, but it would be peculiar to describe these as things that I value. On the other hand, as we have already seen, I may value my friend’s sense of humor, but it is not clear what it would even mean, in this context, to say that I desire my friend’s sense of humor. Similarly, I may value the lessons I learned from my grandparents, but it is not clear what it would mean to say that I desire the lessons I learned from my grandparents. Off hand, it seems that one significant difference between valuing and desiring is that the things we value are things that matter to us, or that we prize or cherish, or that have a certain importance for us. Not everything that we desire matters to us. And not everything that matters to us is a suitable object of desire. So although there are many things that we both value and desire, valuing cannot be reduced to simple desiring. But if valuing cannot be reduced to simple desiring, then how is it to be understood? Gilbert Harman, after offering objections to a variety of proposals, says, “I conclude that valuing is a particular kind of desiring, but I am unable to say more about what kind.” Some philosophers, although not Harman, have suggested that the answer lies in the idea of a second-order desire, whose importance for our understanding of personhood has long been urged by Harry Frankfurt. This suggestion is in some ways surprising. Frankfurt himself does not attempt to explain valuing in this way. Although he places great emphasis on the related yet distinct notion of caring, he uses the term valuing and its cognates only sparingly, and it is not clear how, if at all, he sees his



model as bearing on the understanding of valuing. Furthermore, one of the earliest and most influential objections to Frankfurt’s position was Gary Watson’s argument to the effect that the distinction between different orders of desire is an unsatisfactory substitute for the distinction between wanting and valuing. Nevertheless, others have sought to explain what valuing consists in by reference to Frankfurt’s “hierarchical” model. David Lewis, for example, says that to value X is to desire to desire it. This is meant to help with the case of the unhappy addict mentioned earlier. The addict craves his drug, but since he wishes he didn’t, this view implies that he does not value it. Although what the addict values diverges from what he desires, it does not diverge from what he desires to desire. But does valuing always coincide with desiring to desire? In some cases, it seems, I can desire to desire something that I neither desire nor value. For example, I may want to want to become a regular operagoer—my opera-loving friends may have convinced me that I am missing out on a rich and rewarding dimension of aesthetic appreciation—but so long as I do not actually want to become a regular operagoer, it does not seem true that I value operagoing. It seems closer to the truth to say that I wish I valued operagoing. Perhaps, then, we need to modify Lewis’s proposal. Perhaps to value X is (1) to desire it and (2) to desire to desire it. In the case just described, I do not value operagoing because I fail to satisfy the first condition. The unhappy addict, by contrast, does not value the drug he craves because he fails to satisfy the second condition. Yet there are, it seems, other cases in which I may value something without desiring to desire it at all. I may value the lessons that my grandparents taught me, but as earlier observed, it is not clear what would be meant by saying that I desire those lessons, and it is no clearer what would be meant by saying that I desire to desire them. Perhaps it would become clearer if desire were taken to mean “have a pro-attitude toward.” Then what it would be for me to value the lessons that my grandparents taught me would be for me to have a pro-attitude toward my having a pro-attitude toward those lessons. But while the idea of having a pro-attitude toward my having a pro-attitude toward the lessons may be easier to understand than the idea of desiring to desire the lessons, it does not seem to capture what is meant by saying that I value the lessons. Off hand, it seems that I might value the lessons my grandparents taught me without having any attitude at all toward my own attitude toward those lessons. At any rate, it may be clearer that I value the lessons than it is that I have a pro-attitude toward my having a pro-attitude toward the lessons. And even if I did have a pro-attitude toward my having a pro-attitude toward the lessons, this second-order attitude would be an attitude toward something about myself and my attitudes, whereas what I value are my grandparents’ lessons and not some facts about myself. So how can my valuing of those lessons be equated with any attitude toward my own attitudes? There are also cases in which one’s second-order desires conflict with one’s firstorder desires, and yet it is the first-order desires, rather than the second-order desires, that are naturally taken to express what one values. In the movie Casablanca, Ilsa, played by Ingrid Bergman, says to Rick, played by Humphrey Bogart: “Richard, I tried to stay away. I thought I would never see you again, that you were out of my life.  .  .  . Now? I don’t know. I know that I’ll never have the strength to leave you


Reason, Value, and Desire

again. . . . I can’t fight it anymore. I ran away from you once. I can’t do it again. . . . I wish I didn’t love you so much.” Here, by contrast with the structurally parallel case of the unhappy addict, the viewer is presumably meant to understand that it is Ilsa’s first-order attitude toward Rick, rather than her second-order desire not to love him so much, that indicates what she really values. Charles Taylor also relies on Frankfurt’s notion of a second-order desire, but he uses the idea in a different way than Lewis does. He takes the general lesson of Frankfurt’s discussion to be that “what is distinctively human is the power to evaluate our desires, to regard some as desirable and others are [sic] undesirable.” And he appears to equate second-order desires with evaluations of first-order desires. For example, he says that “a crucial feature of human agency is the capacity for second-order desires or evaluation of desires.” Taylor is less concerned to defend these claims than he is to argue for an additional distinction between what he calls strong and weak evaluation of our desires. For our purposes, however, the pertinent point is that, like Lewis, Taylor sees a close connection between second-order desires and value. But whereas Lewis’s position is that to value X is to desire to desire X, Taylor’s focus is on the role of second-order desires in the evaluation not of X but of the desire for X. There are a number of questions that might be raised about Taylor’s account. One question is whether what he means by evaluation is the same as what I mean by valuing, or, if not, how the two notions differ. A second question, which Taylor anticipates, is whether his focus on the evaluation of our desires neglects the extent to which we are concerned with the evaluation of the objects of those desires. A third question is whether the close connection he draws between evaluation and second-order desires is as faithful to Frankfurt’s view as he takes it to be. But for our purposes, the most important question concerns the way that connection is to be understood. Do second-order desires in and of themselves constitute evaluations of first-order desires, as Taylor appears at times to suggest? Or does he think of second-order desires as issuing from independent evaluations of the first-order desires that are their objects? Either way, there are difficulties. On the one hand, as Gary Watson argued in response to Frankfurt, it is unclear why a simple second-order desire should count as an evaluation at all, any more than a first-order desire does. A simple second-order desire is, after all, just another desire, which happens to have a first-order desire as its object. But if, on the other hand, second-order desires issue from independent evaluations, then it seems that the concept of evaluation is presupposed rather than illuminated by Taylor’s appeal to second-order desires. Perhaps, then, it is a mistake to construe valuing as a kind of desiring. Perhaps it is better thought of as a kind of believing. But what kind of believing? The most obvious suggestion is that to value X is simply to believe that X is valuable. This leaves open the question of what it is to believe that something is valuable, and there is an obvious threat of circularity if we are tempted to think, as some philosophers do, that to believe that something is valuable is to believe that it is properly valued. But the proposal that to value X is simply to believe that X is valuable is unsatisfactory in any case, for it is not only possible but commonplace to believe that something is valuable without valuing it oneself. There are, for example, many activities that I regard as valuable but which I myself do not value, including, say, folk dancing, bird-watching, and studying



Bulgarian history. Indeed, I value only a tiny fraction of the activities that I take to be valuable. Of course, most of the valuable activities that I do not myself value are activities in which I also do not participate. So perhaps to value an activity is simply to regard as valuable an activity in which one participates. But this will not do. I can regard an activity in which I participate as valuable without valuing it myself. I may, for example, go to the opera from time to time, and I may regard operagoing as a valuable activity, and yet I may still not value it myself. Even though I participate in the activity and believe that it is a valuable activity, operagoing may leave me cold. Besides, this proposal only applies to activities. Yet there are many things besides activities that I regard as valuable, and many of these are also things that I myself do not value. There are, for example, many paintings, historical artifacts, and literary genres that I believe to be valuable but do not value myself. With respect to these things, the suggestion that to value them is to believe that they are valuable and to participate in them oneself would obviously make no sense. It goes without saying, I hope, that I do not take these remarks to be merely autobiographical. I do not think that it is an idiosyncrasy of mine—or a sign that I am an especially crass or shallow person—that I fail to value so many of the things that I believe to be valuable. I think, instead, that the same is true of most people. The world contains many things that we regard as valuable; yet each of us values—indeed, I will offer reasons for thinking that each of us can value—only a small fraction of these valuable things. That is because valuing requires more of us than just belief. There is no limit in principle to the number of things whose value we can recognize or acknowledge. By contrast, there are severe limits to how much one can value within the confines of a single human life. The points I have been making have implications for the view of valuing defended by Michael Smith. Smith argues that “valuing is believing valuable and believing valuable is believing that we have a normative reason.” Or as he puts the point in another passage, “A has a normative reason to Φ iff A’s Φing is valuable,” and “A accepts that he has a normative reason to Φ iff A values Φing.” In light of our discussion, two related difficulties with these formulations are apparent. First, they neglect the distinction between believing that something is valuable and valuing it. Second, they overlook the fact that we can value many things other than our own activities. As we have seen, for example, one may value one’s privacy, or one’s friend’s sense of humor, or the opinion of a trusted advisor, or the lessons one learned from one’s grandparents. None of these is an action one performs or an activity in which one engages; none of them is a type of Φing. So Smith’s formulations wrongly equate valuing and believing valuable, and they wrongly suppose that what people are capable of valuing is limited to their own activities. The scope of what we value is in one way narrower and in one way broader than Smith suggests. On the one hand, we do not value all the things that we regard as valuable, but, on the other hand, we value many things in addition to our own actions or activities. But if valuing is neither a kind of desiring nor a kind of believing, then what is it? The third of the three possibilities mentioned by David Lewis was that valuing is a feeling. Lewis himself quickly rejects this possibility, on the ground that “the feelings


Reason, Value, and Desire

we have when we value things are too diverse.” This does seem like a reason to reject the simple view that valuing consists in a single, distinctive type of feeling, but there may be more plausible ways of analyzing valuing in terms of feelings and emotions. One way to explore this possibility is by returning to the thought that the things we value are things that matter to us, or that we prize or cherish, or that have a certain importance for us. How are the relevant notions of mattering and importance to be understood? Part of the answer, at least, has to do with the role in our emotional lives of the things that matter to us. If my relationship with my brother matters to me, then I may feel pleased at the prospect of spending time with him, saddened if we rarely have occasion to see one another, eager to help him if he is in need, distressed if a serious conflict develops between us or if we become estranged, and shocked and betrayed if he harms me or abuses my trust. Of course, these are only examples, and variations are possible depending on the details of our relationship. Perhaps my relationship with him matters to me, but in light of features of his character that are well known to me, I would not actually be shocked—only disappointed and resigned—if he abused my trust. Still, if the relationship does matter to me, then there will be a range of emotions that I am disposed to experience in various contexts, depending on how my brother fares and how our relationship fares. So if what it means to say that I value my relationship with my brother is that the relationship matters to me, then my valuing the relationship may consist, at least in part, in my being emotionally vulnerable in certain ways. We might, therefore, propose, not that valuing is a particular kind of feeling, but rather that valuing something makes one vulnerable to feelings and emotions of many different kinds concerning that thing. Following Niko Kolodny, we may say that to be “emotionally vulnerable to X” is to be vulnerable to a wide range of emotions regarding X depending on context and circumstance. So the proposal might be that valuing X makes one emotionally vulnerable to X. Although this seems like a promising suggestion, care is required in developing it, and a number of significant qualifications are necessary if some version of the suggestion is ultimately to be defensible. First, any plausible thesis connecting valuing with emotional vulnerability must acknowledge that the emotions to which valuing makes us vulnerable—call this our value-dependent vulnerability—will depend to some extent on the nature of the things that we value. For example, if I value our friendship, then I may feel a sense of betrayal if you fail to support me when I am in need. But if what I value is my privacy or some old family photographs, then I do not make myself vulnerable in the same way to feelings of betrayal. Neither my privacy nor the old photographs can betray me; only a person can do that. So the contours of one’s emotional vulnerability depend on the nature of what one values. Second, there are certain kinds of cases in which the degree of emotional vulnerability that one incurs by virtue of valuing something is likely to be highly attenuated. In general, one’s value-dependent vulnerability is apt to be greatest when the things that one values are themselves things that can fare well or poorly, can be in good or bad condition, or can flourish or be damaged or destroyed. In short, one’s emotional vulnerability is apt to be greatest when the things that one values are themselves vulnerable



to changes in their condition. People, personal projects and relationships, and physical objects all fall in this category. But consider, by way of contrast, the lessons I learned from my grandparents. If we assume that these lessons are to be identified with the propositional content of what my grandparents taught me, then what I value, in valuing the lessons, is a set of propositions. Because those propositions cannot in any straightforward sense fare well or poorly, or flourish or be destroyed, I am not vulnerable, by virtue of valuing them, to the wide range of emotions that arise in response to variations in the fortunes or in the condition of the things that one values. This does not mean that I incur no value-dependent vulnerability at all by virtue of valuing the lessons that my grandparents taught me. Perhaps, for example, I am liable to feel sad or frustrated if I do not succeed in teaching those lessons to my own children and grandchildren. Perhaps I am liable to feel angry or indignant if other people mock or belittle the importance of those lessons. Still, the degree of emotional vulnerability one incurs in cases of this kind seems considerably reduced. This suggests the third and most important qualification, namely, that valuing cannot be understood solely in terms of emotional vulnerability. There are other considerations that also support this conclusion. One of these has to do with the relation between valuing and caring. There is a great deal of overlap between these two notions. We can speak as easily, for example, of caring about one’s relationship with one’s brother as we can of valuing that relationship, and it is clear that caring, no less than valuing, renders one emotionally vulnerable. It is not surprising, then, that some writers treat the two concepts as interchangeable, as, for example, Elizabeth Anderson appears to do in the first chapter of Value in Ethics and Economics. But, although caring and valuing are clearly related, there appear to be some significant differences between them. As noted earlier, Frankfurt treats caring as a central notion, but seldom uses the language of valuing. And after arguing, in an early paper, that many patients with Alzheimer’s Disease remain capable of valuing, Agnieszka Jaworska distinguishes, in her later work, between valuing and caring, and argues instead that some Alzheimer’s patients are capable of caring even if they are not capable of valuing. What might the differences between the two notions be? English usage provides some clues. Although, as already noted, there are many cases in which it seems equally natural to speak either of valuing or of caring, there are other cases in which it seems more natural to use one of these terms rather than the other. To my ear, for example, it seems more natural to say that one values one’s friend’s sense of humor than that one cares about one’s friend’s sense of humor. Similarly, it seems more natural to say that one values a colleague’s contributions to departmental discussions than that one cares about those contributions. On the other hand, it seems more natural to say that one cares about one’s appearance or about how the Red Sox do against the Yankees than that one values those things. Indeed, to say that one values how the Red Sox do against the Yankees would be quite odd. Although we should not assume that each nuance of English usage affords deep conceptual insight, some of the differences between the way valuing is used and the way caring is used do seem to me significant. Perhaps the most striking difference is that the objects of caring, unlike the objects of valuing, need not be things toward


Reason, Value, and Desire

which one is favorably disposed. So, for example, it would be perfectly appropriate to describe someone as caring deeply about poverty and injustice. We have no difficulty in interpreting this description, and we do not take it to mean that the person is favorably disposed toward poverty and injustice. To describe the same person as valuing poverty and injustice, by contrast, would imply that the person was favorably disposed toward poverty and injustice. The second description, in other words, is naturally interpreted as meaning just the opposite of the first. This should make us wary of using valuing and caring interchangeably. Perhaps, however, to value something is simply to care about and be favorably disposed toward it. But this will not do. What is meant by saying that I value my friend’s sense of humor is not captured by saying that I care about and am favorably disposed toward her sense of humor. As noted earlier, it is not even clear that it makes sense to speak of caring about one’s friend’s sense of humor. And if it does not, then adding that one is favorably disposed toward her sense of humor won’t help. In any case, to say that one is favorably disposed toward the things that one values is an underdescription. One may be favorably disposed toward something without valuing it. The pertinent point instead seems to be that valuing, unlike caring, involves a view of the object of one’s attitude as being good or worthy or valuable. To view something as valuable goes beyond having a generically favorable disposition toward it. It is, one might say, a distinctive way of being favorably disposed toward it. And it is the fact that valuing involves this distinctive view of its objects that explains why there is such a difference between saying that someone cares about poverty and injustice and saying that the person values poverty and injustice. If this is correct, then even though, as we have seen, valuing cannot be reduced to or equated with believing valuable, nevertheless, valuing something does involve, in addition to emotional vulnerability, a view of the thing that one values as being good or worthy or valuable. These two dimensions of valuing are related to each other in at least two different ways. First, the emotions to which valuing makes us vulnerable normally present themselves to us as appropriate or merited, and this fact is closely connected to our belief that the things we value are good or valuable. The sadness I feel when a valued relationship ends is experienced as an appropriate response to what has happened, as is the annoyance I feel when the privacy I value is repeatedly disrupted, or the disappointment I feel when the colleague whose contributions to department meetings I value declines to offer an opinion on a matter of special importance. In each case, my experience of my reaction as merited or appropriate is closely connected to my view of the valued item as valuable. How is this “close connection” to be understood? On one interpretation, the fact that I view the item as valuable helps to explain why I think my reactions are merited: why sadness is an appropriate reaction to the end of the relationship, why annoyance is an appropriate response to the violation of my privacy, and why disappointment is an appropriate response to my colleague’s reticence. On another interpretation, my belief that the valued items are valuable just is, in part, a belief that these reactions, and other comparable reactions in other contexts, are merited or appropriate. To put it another way, it is a belief that there are good reasons for these reactions. On either interpretation, the fact that the



items we value are seen as valuable has implications for the character and content of our emotional vulnerability. Second, the fact that valuing involves emotional vulnerability helps to explain why there is a gap between believing valuable and valuing, that is, why believing that something is valuable does not automatically translate into valuing it. There are limits to our capacity for emotional vulnerability. Some of these limits are practical or psychological; they are limits of time or emotional resources. Some of them may instead be conceptual. To be emotionally vulnerable to X is for X to have a certain kind of importance in one’s life, and there may be conceptual limits to the number of things that can be important to a person. If one has too many best friends, then one has no best friends, and if one has too many hobbies, then one has no hobbies. Similarly, perhaps, if too many things are important to us, then nothing is important to us. But whether the relevant limits are conceptual or “merely” practical and psychological, the fact is that most people have a capacity for the recognition of value that far outstrips their capacity for emotional vulnerability. That is one reason why we normally value only a small subset of the things that we regard as valuable. We have identified two mutually related dimensions of valuing. To value X is normally both to believe that X is valuable and to be emotionally vulnerable to X. However, there is more to valuing than this. In particular, valuing also has deliberative significance: to value X involves seeing X as a source of reasons for action. By this I mean that one who values X is disposed to treat X-related considerations as constituting reasons for action in relevant contexts. For example, if I value my privacy, then I will be disposed in relevant contexts to treat considerations about the impact of proposed courses of action on my privacy as having deliberative relevance. Of course, the relevant contexts may not arise very often, and even when they do arise the privacyrelated considerations may not be decisive. But if I never see those considerations as having deliberative relevance, then it is not true that I value my privacy. Similarly, if I value my relationship with my brother, then I will be disposed in suitable contexts to treat considerations about the impact of proposed courses of action on the state of our relationship as having deliberative relevance. If I had no disposition at all to do this, then it would not be correct to say that I valued the relationship. And so on for other cases. If I value my colleague’s contributions to departmental meetings, then in suitable contexts I will treat considerations having to do with my ability to learn his opinion as relevant to my deliberations. In general, to value X is, inter alia, to be disposed in relevant contexts to treat X-related considerations as having deliberative relevance. Valuing, we may say, is deliberatively consequential. It is clear from the examples I have mentioned that, just as the emotions to which valuing makes us vulnerable may vary depending on what it is that we value, so, too, the considerations that valuing disposes us to treat as reasons for action may vary depending on the nature of the things that we value. There is no precise formula that will cover every case. That is why I have spoken simply of a disposition to treat “X-related considerations” as reasons for action. This should not strike us as surprising. We should not expect the details either of our emotional engagement or of our practical orientation to be invariant across the wide range of things that we value. Instead, the emotions to


Reason, Value, and Desire

which we are liable and the ways in which we have reason to act will, to some degree, vary depending on the nature of what we value. Consider, for example, that certain emotions, such as indignation, presuppose that the object of the emotion has the capacity to recognize and respond to reasons. Valuing one’s relationship with another person involves a susceptibility to experiencing, toward that person, emotions that carry this presupposition. By contrast, valuing an inanimate object involves no comparable susceptibility. This illustrates the point that what it is to value something is conditioned by the nature of the object that is valued. There is one obvious lesson that we should learn from this—a lesson which, despite its obviousness, has not been often enough noticed. The lesson is that any account of valuing in general will inevitably be both very abstract and, in important respects, incomplete. Beyond a certain point, we can make progress in understanding what valuing involves only by proceeding in a more piecemeal way, that is, by reflecting on the specific kinds of things that people value. Nevertheless, my aim in this essay is to see how far we can get in understanding valuing as a general phenomenon. Our discussion to this point suggests that valuing involves a distinctive fusion of reason and emotion. It comprises a complex syndrome of interrelated dispositions and attitudes, including, at least, certain characteristic types of belief, dispositions to treat certain kinds of considerations as reasons for action, and susceptibility to a wide range of emotions. Somewhat more precisely, it seems that valuing any X involves at least the following elements: 1. A belief that X is good or valuable or worthy, 2. A susceptibility to experience a range of context-dependent emotions regarding X, 3. A disposition to experience these emotions as being merited or appropriate, 4. A disposition to treat certain kinds of X-related considerations as reasons for action in relevant deliberative contexts. Three points should be noted by way of qualification and elaboration. First, I take this understanding of valuing to apply in the first instance to noninstrumental valuing. We can and often do value things instrumentally as well. Whether instrumental valuing can be understood as involving the same four elements, or whether some modification is required to cover the instrumental case, is a question I will not address here. Second, we have seen that the extent (and not merely the content) of one’s emotional vulnerability can vary depending on the nature of what one values. It is likely to be greater when the things that one values are themselves vulnerable, and it is likely to be attenuated when they are not. But there are also other factors that can affect the extent of one’s emotional vulnerability. Not everything that one values plays an equally central role in one’s life. I may value both my relationship with my son and my colleague’s contributions to department meetings, but they obviously do not play the same role in my life. Other things being equal, the degree of emotional vulnerability that one incurs in valuing a given thing will depend on the role that thing plays in



one’s life. These differences in role, in turn, are largely constituted by differences in the range and character of the reasons for action with which different valued items are seen as providing us. I see my relationship with my son as providing me with compelling reasons for action in a very wide range of contexts, many of which are highly consequential. I also value my colleague’s contributions to department meetings, but these provide me with reasons for action only in a very limited range of contexts. We speak of valuing things to different degrees, and of valuing some things more than others. To some extent, these ordinal and comparative judgments reflect judgments about how valuable different things are. But to a great extent they are best understood by reference to differences of role, reason, and emotional vulnerability. Third, I said earlier that, according to one view, the belief that X is valuable just is, in part, the belief that one’s context-dependent emotional reactions regarding X are merited. If one accepts this analysis, then it will seem a redundant feature of my account that it includes, as separate items, both the belief that X is good or valuable and the disposition to experience one’s context-dependent emotions regarding X as being merited or appropriate. I will later provide reasons for doubt about this analysis of the belief that X is valuable, but for those who accept it, the redundancy in my account is easily remedied. There is a potential difficulty with the kind of account I have been developing. One way of bringing out the difficulty is by comparing my account with Thomas Scanlon’s, which is in some respects quite similar to mine, but which also differs from it in significant ways. Scanlon summarizes his account as follows: We value many different kinds of things, including at least the following: objects and their properties (such as beauty), persons, skills and talents, states of character, actions, accomplishments, activities and pursuits, relationships, and ideals. To value something is to take oneself to have reasons for holding certain positive attitudes toward it and for acting in certain ways in regard to it. Exactly what these reasons are, and what actions and attitudes they support, will be different in different cases. They generally include, as a common core, reasons for admiring the thing and for respecting it, although “respecting” can involve quite different things in different cases. Often, valuing something involves seeing reasons to preserve and protect it (as, for example, when I value a historic building); in other cases it involves reasons to be guided by the goals and standards that the value involves (as when I value loyalty); in some cases both may be involved (as when I value the U.S. Constitution). To claim that something is valuable (or that it is “of value”) is to claim that others also have reason to value it, as you do. Scanlon’s account is similar to the one I have been developing in its emphasis on the wide range of things that we value, in its assertion that valuing involves seeing oneself as having reasons for action of various kinds, and in its insistence on the diversity both of those reasons and of the actions they support. There are also some differences between the two accounts. For one thing, Scanlon does not identify


Reason, Value, and Desire

emotional vulnerability as an aspect of valuing. His reference to taking oneself “to have reasons for holding certain positive attitudes” toward the things that one values covers some of this territory, but only some of it. As we have seen, the emotions to which valuing makes one vulnerable are not all positive in character; valuing a friendship makes one as vulnerable to feelings of grief at the friend’s death as it does to feelings of delight in the friend’s company. In addition, valuing involves a disposition actually to experience these attitudes, and not merely to see oneself as having reason for them. A second difference has to do with the way Scanlon draws the distinction between regarding something as valuable and valuing it oneself. His formulation (in the final sentence of the quoted passage) suggests that the only things that one can see as valuable are things that one values oneself. In my view, this is a mistake. Valuing is not necessary for believing valuable, and believing valuable is not sufficient for valuing. I regard bird-watching and the study of Bulgarian history as valuable, but I do not value them myself. As I said earlier, it is partly the fact that valuing involves emotional vulnerability, while believing valuable does not, that explains why there is a gap between the two. So the second difference between Scanlon’s account and mine is related to the first. There is a third difference between the two accounts, and it is this difference that brings out a potential difficulty for the analysis of valuing that I have been developing. In my view, valuing something involves a belief that the valued item is good or valuable. In Scanlon’s view it does not. As we have seen, Scanlon takes valuing to be a notion that is prior to the notion of believing valuable, and he analyzes claims about what is valuable with reference to this prior understanding of valuing. Indeed, Scanlon’s primary motivation for investigating the concept of valuing is because he sees it as a “helpful stepping-stone” in the development of his buck-passing account of goodness and value. The whole point of the buck-passing account is to deny that “being valuable” is itself a substantive, reason-generating property. Instead, “to call something valuable is to say that it has other properties that provide reasons for behaving in certain ways with regard to it.” So it suits Scanlon’s wider purposes to explain valuing in terms of the perception of considerations as reasons and to explain believing valuable in terms of valuing. I have already criticized Scanlon’s statement that “To claim that something is valuable is to claim that others also have reason to value it, as you do.” As I have indicated, the suggestion that one must value something oneself in order to claim that it is valuable seems to me mistaken. A further worry is that Scanlon’s statement, taken literally, appears to imply that there are two sets of reasons: the reasons that we take ourselves to have insofar as we value something, and the reasons that we have for valuing something. In addition to taking ourselves to have the first set of reasons, in other words, we also have a second set of reasons for taking ourselves to have the first set. But the content of this second set of reasons is not explained. These difficulties can, however, be deflected by formulating Scanlon’s view in a slightly different way. He himself says in another place that the claim that “friendship is valuable is best understood as the claim that it is properly valued, that is to say, that the reasons recognized by someone who values friendship are in fact good reasons.”



This alternative formulation avoids the second of the two problematic features I have mentioned. It does not imply that we have reasons for seeing ourselves as having the reasons involved in valuing. Nor does it imply that one must value something oneself in order to claim that it is valuable. Yet this formulation retains the essentials of Scanlon’s position. It preserves the buck-passing idea, as well as the idea that believing valuable should be defined in terms of valuing. To value something does not involve seeing it as valuable, on this account; instead, to claim that something is valuable is to claim that it is properly valued. Because it avoids the difficulties of the first formulation, I will treat this second formulation as the canonical statement of Scanlon’s view of the relation between valuing and believing valuable. Scanlon’s view presents a challenge for the account of valuing that I have been developing. That account states that valuing X normally involves at least four elements, of which the first is a belief that X is good or valuable or worthy. But if the claim that something is valuable is in turn taken to mean that it is properly valued, then this element of my account looks to be circular. I noted earlier that the threat of circularity is a problem for views according to which valuing is simply equated with believing valuable. My own view, however, seems to suffer from a version of the same problem. One response to this problem is to question the force of Scanlon’s account of the relation between believing valuable and valuing. Although the second, “canonical” account of that relation does not imply that one must value something oneself in order to claim that it is valuable, it still takes the reasons of the people who do value a given thing as fixing the content of the claim that that thing is valuable. For example, if I say that bird-watching or studying Bulgarian history is valuable, my statement is to be understood as claiming that the reasons recognized by those who value these things are good reasons. But if claims to the effect that X is valuable are to be interpreted as claims about reasons, it is implausible to construe them solely as claims about the reasons recognized by people who value X. It is more plausible to suppose that they are, at least in part, claims about the reasons that even nonvaluers have. For example, if I say that bird-watching or studying Bulgarian history is valuable, despite the fact that I myself do not value those things, then my statement is most plausibly interpreted as being, at least in part, about the reasons that I—and other people generally—have with regard to those activities. We should not, for instance, cast aspersions on those who engage in the activities or disrupt what they are doing without good cause. Those who value bird-watching or the study of Bulgarian history have other reasons that the rest of us lack—such as reasons actually to engage in these activities. In general, those who value a given thing have distinctive reasons that are not shared by people who recognize the value of the thing but do not value it themselves. If claims to the effect that something is valuable are to be construed as claims about reasons, then they are best understood as being, at least in part, claims about the reasons that even nonvaluers have. These considerations may be taken to support a formulation of the buck-passing account that is not tied to the reasons of valuers in particular. And in fact, Scanlon formulates the buck-passing account itself in just this way. As we have seen, he says it asserts that “to call something valuable is to say that it has other properties that


Reason, Value, and Desire

provide reasons for behaving in certain ways with regard to it.” Here there is no reference to the reasons of valuers specifically. Instead, the relevant reasons, and the persons for whom they are reasons, are left unspecified. This brings out a point that has so far been left implicit in my discussion, namely, that there are two distinct aspects of Scanlon’s view that are, in principle, separable: the buck-passing account proper, and the claim that valuing has priority over believing valuable. The buck-passing account is the claim that to call something valuable is to say that it has other properties that provide reasons for behaving in certain ways with regard to it. The priority claim is the claim that valuing is a notion that is prior to believing valuable, and that believing valuable is to be explained in terms of valuing. The considerations we have been rehearsing provide reason to doubt the priority claim, but they do not count against the buck-passing account itself. Suppose, however, that, while rejecting the priority claim, we do accept the buck-passing account. And suppose we also accept Scanlon’s view that to value something is just “to take oneself to have reasons for holding certain positive attitudes toward it and acting in certain ways with regard to it.” Then, because the buck-passing account, unsupplemented by the priority claim, interprets the statement that something is valuable without any reference to valuing or to the reasons of valuers in particular, and because, on Scanlon’s account, valuing something involves no claim that the valued thing is valuable, the relation between valuing and believing valuable, on these assumptions, becomes elusive. On the one hand, it is no longer clear how the two phenomena differ from one another, inasmuch as both are said to consist solely in having certain very general beliefs about reasons. But if we assume, on the basis of the arguments given earlier, that they do differ from one another, then, because neither is defined with reference to the other, it is no longer clear that there is any relation between them at all. But this is implausible. Surely there is some connection between these cognate notions. One way of securing the connection it is to accept the priority claim and to define believing valuable in terms of valuing. But if we reject the priority claim, as I have argued we should, then we need to secure the connection in another way. My proposal does this by asserting that to value something is in part to believe that it is valuable. This explains the difference noted earlier between valuing and caring and, as we have seen, it is compatible with the buck-passing account of value. On one interpretation, for example, the claim that X is valuable might here be understood, compatibly with the buck-passing account, as asserting that X has properties in virtue of which all people have reasons for behaving in certain (minimum) ways with regard to it. On a second, perhaps more plausible interpretation, it might be understood as the claim that X has properties in virtue of which (1) all people have reasons for behaving in certain (minimum) ways with regard to X, and (2) some people have reasons for additional actions with regard to X and for being emotionally vulnerable to it. On either interpretation, to value X would involve believing that it is valuable in the specified sense, but it would also involve being emotionally vulnerable to it oneself and treating it as a source of additional reasons for action in one’s own deliberations. Because—on either interpretation—my proposal is compatible with the buck-passing



account, that account provides no objection to it. It is the priority claim, rather than the buck-passing account, that fuels the circularity objection to my proposal. However, I have argued that we should reject the priority claim. It overlooks the fact that to value something is in part to view it as good or worthy, and it treats the reasons of valuers as fixing the content of the claim that something is valuable. In so doing, it neglects the fact that valuable things give everyone, and not merely those who value them, certain minimal reasons for action, such as reasons not to destroy or denigrate those things. Let me conclude with two additional points about the distinction between valuing and believing valuable. First, I have said that nobody can value all valuable things. However, I believe that something even stronger is true, namely, that there are some valuable things that can only be valued by certain people, even though everyone can recognize their value. Imagine, for example, that I have just heard a glowing account of the friendship between two people whom I have never met and with whom I have no connection. I might think, on the basis of this account, that their friendship sounds like a valuable one. It would be bizarre, however, for me to say that I value their friendship. Or, at any rate, it would be bizarre in the absence of some special explanation. Perhaps, for instance, it would make sense for me to say that I value their friendship as an example of a type of relationship that I took no longer to exist. Even in this case, however, I cannot value the friendship in the same way that the participants can; it cannot play the same role in my emotional life and practical deliberations. If it seemed to play such a role, we might have doubts about my mental well-being; we might fear that I had entered the psychological territory inhabited by groupies and stalkers. Yet even though I cannot value the friendship in the way that the participants can, I am just as capable as they are of recognizing its value. So not only is it the case that nobody can value all valuable things, but, in addition, some things whose value everyone can recognize can, nevertheless, only be valued by certain people. In this sense, at least some valuing is positional in a way that believing valuable is not. That is, there are some things of which it is true (1) that only those who occupy the right position in relation to the thing are capable of valuing it, or of valuing it in a certain way, and (2) that not everyone is capable of occupying the right position in relation to that thing. This provides an additional reason, beyond those already mentioned, why nobody can value all valuable things. Nobody can occupy all the relevant positions. Second, in emphasizing the importance of the conceptual distinction between valuing and believing valuable, I have appealed at various points to ordinary usage. I believe that the conceptual distinction is firmly rooted in ordinary thought, and that this is clearly reflected in our linguistic practices. Nevertheless, ordinary usage is not so strictly regimented that our linguistic practices perfectly track the underlying conceptual distinction with which I am concerned. Despite what I have said, there are doubtless times when people use the verb value simply to mean something like “believe valuable.” Someone may speak, for example, of valuing a colleague’s contributions to department meetings when all she means is that she regards those contributions as valuable. This is especially likely to happen when, as


Reason, Value, and Desire

in this example, the item that is the object of the attitudes in question is not one that gives much scope for emotional vulnerability and whose practical relevance is limited. In such cases, the differences between valuing in my sense and believing valuable are already attenuated, so it is perhaps not surprising that the boundaries between the two linguistic expressions should become blurrier. Still, the underlying conceptual distinction is clear, and as I hope to have demonstrated in this essay, it represents a stable and significant division in our attitudes.


attaching value It is not uncommon to describe a person as “attaching value” to something, and, in context, this seems more or less equivalent to saying that he values it. For example, someone may be said greatly to value his relationship with his brother, or an old watch that his father gave him. But it may just as naturally be said that he “attaches great value” to the relationship or to the watch. We also speak—albeit, perhaps, in a somewhat narrower range of contexts—of people “assigning value” to one thing or another. These locutions may suggest that the phenomenon of valuing has—or is ordinarily understood as having—a voluntaristic dimensions that my account neglects. The idea that, when we value something, we are “attaching” or “assigning” value to it may suggest that valuing consists in making a certain kind of decision—a decision to confer value on something that would otherwise lack it. I think that this is misleading, for at least two reasons. First, it is true that the expressions in question are often used in contexts of decision, but these are decisions about what to do and not decisions about what to value. For example, I may say, in explaining why I decided to pass up an evening at the theatre in order to spend time with an unhappy friend, that I attach greater value to the friendship than I do to going to the theatre. There is indeed a decision here, but it is a decision about how to spend the evening, not a decision about the items upon which I wish to confer value. Second, our use of expressions like attach value and even assign value, in contexts where it would make equally good sense to speak simply of what we value, is partly a way of registering the fact that valuing something goes beyond merely recognizing that it is valuable, and that nobody can value every valuable thing. When we “attach” value to something, we are active not in the sense that we choose to make the thing valuable when previously it wasn’t, for it may well be valuable already, and if it is not, then our valuing it does not make it so. Instead, we are active in the sense that we take up a certain orientation toward it; it becomes a focus of our emotions and a source of what we take to be our reasons. In this sense, it becomes important or valuable to us. But we need not decide to make this so, and usually we do not. Sometimes, to be sure, things become important to us—or we come to value them—as a consequence of decisions we have made, but that is not the same as our deciding to make them important to us or to value them. Often, the process of coming to value something happens quite gradually, without any conscious awareness on our part, and we simply find at some point that we do in fact value the thing: that we treat it as a source of reasons, and that we are emotionally vulnerable to it. Sometimes, far from deciding to value things, we find it difficult even to recognize or acknowledge that we value them. One familiar psychotherapeutic aim is to help people to attain greater reflectiveness about their emotions and motivations, and thereby to come to a



better understanding of what they do and do not value, or of what actually matters to them. This reinforces the point that valuing or “attaching value” is not normally something that we decide to do. Instead, as I have argued, expressions like attach value register our implicit understanding of two points. First, valuing is a selective phenomenon, in the sense that nobody can value everything that is valuable, and that different people value different things. Second, in valuing something, we are taking up a practical and emotional orientation toward that thing, even though we do not normally decide to do this and may in some instances have a difficult time recognizing that we have done so. acknowledgments This essay is also published in Samuel Scheffler, Equality and Tradition: Questions of Value in Moral and Political Theory (Oxford University Press, 2010). Printed here by agreement with the publisher. Earlier versions of the essay were presented to the Bay Area Forum for Law and Ethics (BAFFLE), to departmental colloquia at the University of Iowa and the CUNY Graduate Center, and to my Fall 2008 graduate seminar at NYU. I am grateful to all of these audiences for valuable discussion. I am also indebted to Macalester Bell, Josh Glasgow, Niko Kolodny, Stephen Neale, Derek Parfit, and Jay Wallace for helpful comments on previous drafts. notes 1. See David Lewis, “Dispositional Theories of Value,” as reprinted in Lewis, Papers in Ethics and Social Philosophy (Cambridge: Cambridge University Press, 2000): 68–94, p. 69. 2. Ibid. 3. See Harry Frankfurt, “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68 (1971): 5–20. 4. See Gary Watson, “Free Agency,” as reprinted in Watson (ed.), Free Will (New York: Oxford University Press, 1982), pp. 96–110, at pp. 100–101. 5. Michael Smith, “Valuing: Desiring or Believing?” in D. Charles and K. Lennon (eds)., Reduction, Explanation, and Realism (Oxford: Oxford University Press, 1992), pp. 323–59, at pp. 326–27. For Stocker’s example, see Michael Stocker, “Desiring the Bad: An Essay in Moral Psychology,” Journal of Philosophy 76 (1979), at p. 744. 6. Gilbert Harman, “Desired Desires,” as reprinted in Harman, Explaining Value and Other Essays in Moral Philosophy (Oxford: Clarendon Press, 2000), pp. 117–36, at p. 135. 7. See, for example, Harry Frankfurt, “The Importance of What We Care About,” as reprinted in The Importance of What We Care About (Cambridge: Cambridge University Press, 1988), pp. 80–94; “On Caring,” in Necessity, Volition, and Love (Cambridge: Cambridge University Press, 1999), pp. 155–80; The Reasons of Love (Princeton, N.J: Princeton University Press, 2004). 8. See Watson, op. cit., esp. pp. 107–109. 9. See Lewis, op. cit. 10. Casablanca (Warner Brothers, 1942), screenplay by Julius J. Epstein, Philip G. Epstein, and Howard Koch (based on the play “Everybody Goes to Rick’s” by Murray Burnett and Joan Allison), available online at, pp. 102–104 (last accessed on March 12, 2009). 11. It may seem that Rick himself draws a different conclusion, because, despite his love for Ilsa, and despite her declaration of love for him, he insists that she leave him behind and escape from Casablanca with her husband, Victor Laszlo: “Inside of us we both know you belong with Victor. You’re part of his work, the thing that keeps him going. If that plane leaves


Reason, Value, and Desire

the ground and you’re not with him, you’ll regret it. . . . Maybe not today, maybe not tomorrow, but soon, and for the rest of your life” (Ibid, p. 121). Yet what makes him willing to help Ilsa and Laszlo escape together, which he had previously refused to do, is precisely his conviction that her declaration of love for him was sincere. Moreover, although it would have made for a worse movie, there would have been nothing jarring conceptually about having things end differently, with Rick simply accepting Ilsa’s declaration of love and their living happily ever after. 12. Charles Taylor, “What Is Human Agency?” as reprinted in Taylor, Human Agency and Language: Philosophical Papers 1 (Cambridge: Cambridge University Press, 1985), pp. 15–44, at p. 16. 13. Ibid., pp. 42–43. 14. Ibid., p. 18n. 15. Elizabeth Anderson says in Value in Ethics and Economics (Cambridge, Mass.: Harvard University Press, 1993, p. 2) that “to judge that something is good is to judge that it is properly valued.” Thomas Scanlon says in What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998, p. 89) that “the claim that friendship is valuable is best understood as the claim that it is properly valued.” 16. Smith, op. cit., p. 344. 17. Ibid., p. 329. 18. Lewis, op. cit., p. 70. 19. The connection between valuing and emotional vulnerability is emphasized by Elizabeth Anderson in Chapter One of Value in Ethics and Economics, and by Niko Kolodny in “Love as Valuing a Relationship,” Philosophical Review 112 (2003), pp. 135–89. The notion of something being important to a person figures centrally in the writings of Harry Frankfurt cited in footnote 7. Frankfurt sees importance to a person as being closely connected with caring, but rather than analyzing these ideas in terms of emotional vulnerability, he interprets both of them as picking out essentially volitional phenomena. As will soon become clear, I distinguish between valuing and caring, and my account of valuing is not volitional in Frankfurt’s sense. I regard the ideas of mattering and importance to as suggestive, but in the end I offer no analysis of them and nothing in my account of valuing turns on these notions. 20. Kolodny, op. cit., pp. 150–52. Kolodny offers an account of valuing that has a good deal in common with the one I develop here. 21. Anderson, op. cit. 22. Agnieszka Jaworska, “Respecting the Margins of Agency: Alzheimer’s Patients and the Capacity to Value,” Philosophy & Public Affairs 28(1999): 105–38. 23. See, for example, Jaworska, “Caring and Internality,” Philosophy and Phenomenological Research 74(2007): 529–68, and “Caring and Full Moral Standing,” Ethics 117(2007): 460–97. 24. One class of (partial) exceptions should be noted. These are cases in which a person values something but describes it as having merely “sentimental” or “personal” value, or as being valuable only to him or herself. Because these very descriptions imply that in normal cases the value ascribed to valued items is not similarly qualified, I regard cases of “merely personal value” as being exceptions that prove the rule. (For a possible example of such a case, see Jerry Cohen’s discussion of his eraser in Section 4 of his “Rescuing Conservatism: A Defence of Existing Value.”) It is also worth noting that even the ascription of “sentimental” or “personal” value to an item goes beyond having a generically favorable disposition toward it. 25. In this spirit, I have elsewhere engaged in a number of such piecemeal inquiries. I have asked: What is involved in valuing a personal project? What is involved in valuing a personal relationship? What is involved in valuing one’s membership in a group, community, or association?



See “Relationships and Responsibilities,” Philosophy & Public Affairs 26(1997): 189–209, reprinted in Boundaries and Allegiances (Oxford: Oxford University Press, 2001), pp. 97–110; “Projects, Relationships, and Reasons,” in R. Jay Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith (eds.), Reason and Value: Themes from the Moral Philosophy of Joseph Raz (Oxford: Clarendon Press, 2004), pp. 247–69; and “Morality and Reasonable Partiality,” in Equality and Tradition (Oxford University Press, 2010), pp. 41–75. 26. Nor shall I consider how the account might need to be modified to accommodate cases of “sentimental” or “personal” value like those discussed in footnote 24. 27. Scanlon, op. cit., p. 95. 28. Actually, Scanlon’s formulation is slightly ambiguous. The phrase as you do might mean either “as you value it” or “as you have reason to value it.” In the text, for simplicity, I assume the first interpretation, and I criticize his formulation on that basis. But my arguments also apply, mutatis mutandis, to the version of his formulation that assumes the second interpretation. 29. Scanlon, op. cit., p. 95. The buck-passing account is in turn meant to provide an alternative to the “teleological conception of value,” which Scanlon takes to be associated with consequentialism but also to have wider appeal. In the course of his discussion of the teleological conception, Scanlon criticizes my article “Agent-Centered Restrictions, Rationality, and the Virtues” (Mind 94[1985]: 409–19; reprinted in S. Scheffler [ed.], Consequentialism and Its Critics [Oxford, 1988], pp. 243–60), but the extent to which our views differ is not clear to me. In the article that Scanlon discusses, I was attempting to diagnose the appeal of consequentialism and the correlative impression that there is something paradoxical about deontological prohibitions. My diagnosis was that consequentialism seems appealing because it embodies a form of rationality that I called “maximizing rationality.” I maintained (1) that maximizing rationality is a familiar form of rationality, which plays a role outside consequentialist thought as well as within it, (2) that so long as one remains within the confines of maximizing rationality, deontological constraints do look puzzling, (3) that it might nevertheless be possible to reconcile the constraints with maximizing rationality, (4) that it’s in any case “not obvious that maximizing rationality constitutes the whole of rationality” (Consequentialism and Its Critics, 258), and (5) that an alternative defense of deontological constraints would involve showing that they “embody a limitation on the scope of that form of rationality, and give expression to a different form of rationality which we also recognize and which also has its place in our lives” (Consequentialism and Its Critics, 252). Despite the critical tone of Scanlon’s discussion, it’s not clear to me which of these claims, if any, he actually rejects. 30. Scanlon, op. cit., p. 96. 31. This is also a difficulty for the view I defended in “Relationships and Responsibilities.” 32. Scanlon, op. cit., p. 89. 33. Or, at any rate, it avoids this implication provided the claim that the reasons recognized by others are good reasons is not taken to commit one to treating them as reasons for oneself. 34. What is the source of these additional reasons? Since, by hypothesis, they are reasons that nonvaluers lack, it seems that their source cannot lie simply in the fact that the valued items are valuable. But this invites the following worry. If a person who values X has additional reasons that others lack, then perhaps the source of these additional reasons lies in the very fact that the person values X. But this would seem to imply that valuing is, so to speak, a reason for itself, which may seem circular or otherwise objectionable. Although I cannot address this worry adequately here, I believe that the additional reasons that valuers possess are to be accounted for, instead, by appealing to special features of the position that valuers occupy in relation to the valued item. For example, if I value my relationship with my son, the additional reasons that I have—but which others who recognize the value of that relationship lack—are to


Reason, Value, and Desire

be explained by reference to the fact that I am a participant in the valuable relationship while they are not. Indeed, as I argue in the text that follows (at p. 37), there are some things (such as particular personal relationships) that only certain people can value, because only certain people can occupy the right position in relation to those things. Of course, what counts as being in the right position in relation to a given item will depend on the type of item it is. In the case of personal relationships, participation in the relationship may define the “right position,” but for valuable objects or activities, other accounts may be required. For relevant discussion, see T. M. Scanlon, “Reasons: A Puzzling Duality?” in R. J. Wallace, P. Pettit, S. Scheffler, and M. Smith (eds.), Reason and Value: Themes from the Moral Philosophy of Joseph Raz (Oxford: Clarendon Press, 2004), pp. 231–46, and Niko Kolodny, “Aims as Reasons” (chapter 3 in this volume). 35. It might be objected that people who value X will surely recognize, inter alia, those reasons that even nonvaluers have in regard to X. They will recognize, for example, reasons not to destroy or denigrate X. So if the claim that X is valuable is interpreted as a claim that the reasons recognized by those who value X are good reasons, then it is being interpreted, at least in part, as a claim about the reasons that even nonvaluers have. However, this interpretation fails to distinguish between the reasons that valuers alone have, such as reasons to engage in the activity (if X is an activity), and the reasons that nonvaluers also have, such as reasons not to denigrate or disrupt X. Even if it implies that reasons of both kinds are good reasons, in other words, it fails to distinguish between those good reasons that apply to valuers alone and those that apply to everyone. So even if, by virtue of its endorsement of the reasons recognized by valuers, it implies that any reasons that also apply to nonvaluers are good reasons, it makes no claim to the effect that certain reasons do in fact apply to nonvaluers. And that is the kind of implication that, in my view, a claim that X is valuable does carry. 36. Scanlon, op. cit., p. 96. 37. This interpretation is a variant of one suggested to me by Niko Kolodny. 38. If, despite my arguments, one continues to find the priority claim compelling, then one may simply eliminate the first element in my account of valuing, while retaining the other three. This would narrow the gap between Scanlon’s account and mine, though the differences noted earlier concerning the role of the emotions would remain. 39. It might be suggested, however, that the difference between valuing and believing valuable is best thought of as a matter of a degree rather than as a sharp dichotomy. On the one hand, I have conceded that the level of emotional vulnerability involved in valuing may be attenuated in some cases, and on the other hand, it may be said that even believing something valuable brings with it some degree of emotional vulnerability. For example, I may be vulnerable to feelings of mild regret if I learn that something I regard as valuable (but do not value myself) has been destroyed. It is important to distinguish between two different ways in which this suggestion might be developed. On one interpretation, being liable to experience certain (mild) emotions is part of what it is to believe that something is valuable. This seems to me implausible. On the other interpretation, which seems to me more plausible, the point is rather that a belief that X is valuable, when accompanied by a disposition to experience (say) mild X-related regret if X is destroyed, may not itself constitute an instance of valuing, but it may differ only in degree from a case of valuing in which the level of emotional vulnerability is attenuated. The two cases may still be distinguished from one another by reference to the degrees of emotional vulnerability involved and the kinds of reasons for action that the people in question recognize, but the difference may not be as stark as in some other pairs of cases. I am grateful to Kinch Hoekstra and Richard Fumerton for raising this issue.

3 Aims as Reasons* Niko Kolodny

in recent discussions, one finds the claim that, despite first appearances, attitudes do not give us reasons for action. More precisely, the “Thesis”—to give it a name—is that the fact that a person currently has or lacks an attitude does not give her normative reasons for action, where by “attitudes” we mean what Scanlon would call “judgment-sensitive” attitudes. These are attitudes that are responsive to their bearers’ judgments about the reasons for them: including beliefs, desires, and intentions, but excluding sensations, such as those of pain and pleasure. Because the motivation and contours of the Thesis are somewhat obscure, I begin by exploring these. Next I consider a challenge to the Thesis, which Scanlon raises: that in many cases, particularly cases of underdetermination by reason, having an aim seems to affect one’s reasons for action. In the bulk of the chapter, I suggest how the challenge might be met, by building on suggestions that Scanlon offers. 1. clarifying the thesis

Let us, provisionally, take the Thesis to say: No basic normative principles take the: attitude-based form: S has at t reason to do X, where X is a function of just S’s attitudes at t (or disposition at t to have attitudes) and nonnormative facts. A normative principle is a principle of the form One has reason to phi in C1. 43


Reason, Value, and Desire

Such a principle is basic iff it is not explained some other normative principle One has reason to psi in C2 and either rules of inference familiar from outside the normative domain (and possibly other nonnormative facts) or the nonnormative fact that It is likely, conditional on one’s phi-ing in C1, that one psi’s in C2 and one’s phi-ing in C1 helps (nonsuperfluously) to bring this about and General Transmission (GT): If there is pro tanto reason for one to E, and if it is likely, conditional on one’s M-ing, that one E’s and that one’s M-ing (nonsuperfluously) helps to bring this about, then there is, because of this, pro tanto reason for one to M. There is more reason the more likely it is—the more “effective” a means one’s M-ing is to one’s E-ing. This understanding of the Thesis would explain why its friends are willing to allow that, in countless cases, one’s having an attitude at a given time can affect one’s reasons at that time for action. For example: 1. The fact that I believe that everyone is out to get me usually means that my anti-psychotic medication is wearing off, and so makes it the case that I have reason to take another dose. 2. The fact that I desire a slice of pie means that I will enjoy a slice of pie, and so makes it the case that I have reason to have one. 3. The fact that I desire to be a doctor means that I am more likely to succeed in attempts to become a doctor, and so makes it the case that I have more reason to try. 4. The fact that I do not want to spend time with my aunt means that if I were tell her that I want to, I would be deceiving her, and so makes it the case that I have reason not to tell her that I want to. 5. My well-being consists in the satisfaction of my desires. So the fact that I desire to X means that X-ing would promote my well-being, and so makes it the case that I have reason to X. 6. There is a virtue of “resoluteness” which consists in doing what one intends. So the fact that I intend to X means that by X-ing, I would display this virtue, and so makes it the case that I have reason to X. 7. Friendship is partly constituted by its participants’ intentions, including their present intentions. The fact that I have certain intentions toward Kevin makes him my friend, and so makes it the case that I have reason to help him. In each of these cases, the effect of my attitude on my reason is explained by the combination of (i) GT, (ii) a normative principle (not necessarily basic) that is not of the attitude-based form, such as:

Aims as Reasons

1. 2. 3. 4. 5. 6. 7.


I have reason to avoid psychotic episodes. I have reason to enjoy my food. I have reason to become a doctor. I have reason not to deceive my aunt. I have reason to promote my well-being. I have reason to display the virtue of resoluteness. I have reason to help my friends.

and (iii) the fact that that attitude means that I am likely (for example) to avoid psychosis by taking medicine. Such explanations need not appeal to basic normative principles of the attitude-based form. With some sense of the Thesis’s content, let us ask about its point. Why would anyone deny that any basic normative principles take the attitude-based form? A common objection to attitude-based theories, which claim that all basic normative principles are of this form, is that they are extensionally incorrect. Consider the Desire-Based Theory, which claims that there is just one basic normative principle: S has reason at t to do just what would serve S’s intrinsic desires at t. The Desire-Based Theory implies, on the one hand, that given the right set of desires, one would have reason to do something that, intuitively, one would lack reason to do. If one had a desire served by committing an atrocity, and no desires served by alternatives to it, then one would have conclusive reason to commit that atrocity. And the Desire-Based Theory implies, on the other hand, that given the right set of attitudes, one would lack reason to do something that, intuitively, one would have reason to do. If one were to have no desire served by avoiding suffering, and desires served by alternatives, then one would lack sufficient reason to avoid suffering. However, the fact that some attitude-based theories are extensionally incorrect does not seem a compelling ground for denying that there are any basic normative principles with the attitude-based form. For one thing, other attitude-based theories are less clearly extensionally incorrect. There are Kantian views, for example, according to which all reasons for action derive from exercises of the will that respect the will’s structural, self-imposed standards. These standards require that some things be willed (at least if anything is willed) and that other things not be willed. For another, even if all attitude-based theories are extensionally incorrect, it is less clear that a “hybrid” must be. A hybrid would say that some, but not all, basic normative principles are of the attitude-based form. Take, for example, the: Intention-Based Hybrid: There is a basic normative principle that S has reason at t to do what S intends at t to do. But there are other basic normative principles, which are not of the attitude-based form. This hybrid does not imply that if one did not intend to avoid suffering, then one would lack sufficient reason to avoid suffering. It does imply that if one intended an


Reason, Value, and Desire

atrocity, then one would have pro tanto reason to commit it. However, it need not imply that one would have conclusive reason to commit it, because the intention-based reason might be overridden by other reasons, such as those provided by human suffering. To my mind, the mere fact that one intends an atrocity does not make it the case that one has even overridden pro tanto reason to commit it. So, by my lights, this hybrid is still extensionally incorrect. But I also find it hard to muster independent evidence for this claim, or to feel confident that, on its own, it has much importance, either for practice or theory. The Intention-Based Hybrid would also seem to imply that, when two options are otherwise tied—when there is merely sufficient reason for option X, merely sufficient reason for option Y, conclusive reason for (X or Y), and not sufficient reason for (X and Y)—forming an intention for one of these options can tip the scales, making it the case that one has conclusive reason to pursue it. But this implication actually seems intuitive, as we will see. Scanlon, one of few friends of the Thesis to explicitly consider the possibility of a hybrid, rejects it not on extensional grounds, but instead because it would involve a “troubling” or “puzzling” “duality in the sources of reasons.” It is elusive, however, what is supposed to be puzzling. Scanlon already accepts a multiplicity of sources of reasons. The value of human life is one source, the grandeur of the Grand Canyon another. Why not yet another source, consisting of desire or rational willing? If we already accept a multiplicity of basic normative principles, with a multiplicity of forms, why would one more basic normative principle, of the attitude-based form, introduce a puzzling duality? What puzzles Scanlon seems not so much a duality of sources as a duality in modes of sourcing: a difference between the way in which pleasure, friendship, etc. are “sources” of reasons, according to friends of the Thesis, and the way in which desire or the will is a “source” of reasons, according to attitude-based theorists. At several points, Scanlon suggests that he seeks to avoid the idea that attitudes (specifically intentions) “create” or “generate” reasons, or “confer” on certain considerations the status of a reason (231– 32, 239–40). But these terms leave the contrast rather elusive. Scanlon himself agrees that intentions “make it the case” that one has reasons, at least in the sense that in coming to have an intention, one comes to have an attitude that one would not otherwise have had (240). It is not clear what the difference is supposed to be between creating a reason and “merely” making it the case that there is a reason when there was not before. One might try saying that when an attitude merely “makes it the case” that there is a reason, it does so only in virtue of fulfilling the condition of a normative principle that is “just” there, to be “discovered,” independent of one’s attitudes (231). But attitude-based theories will say the same thing. When a desire “creates” a reason, the desire-based theorist will say, it does so only by fulfilling the principle: If S has a desire at t, then S has reason at t to do what will serve that desire. This principle is already there to be discovered, independent of one’s attitudes. Perhaps the difference at issue is this. For friends of the Thesis, sources provide reasons in the way in which things that we think of as values or things of value provide

Aims as Reasons


reasons. They are sources in the sense that they are the things that we most fundamentally have reason to promote, or to act for the sake of. Accordingly, the sources of reasons that friends of the Thesis recognize (e.g., pleasure, friendship) are naturally described as values or things of value. I will not try to give a positive description of the different, “attitude-based” way in which the attitude-based theorist thinks of desires or the will as “sources” of reasons. For our purposes, what matters is the negative point that the attitude-based theorist does not think of desires or the will as “sources” in the sense of: Value-Provision: A type T (or, alternatively, instances of a type T) is a source of, or provides, reasons for action only if there is a basic normative principle that is of the form either: (i) if one stands in a relationship of a certain kind to an instance of T, then one has reason to bring it about or: (ii) if one stands in a relationship of a certain kind to an instance of T, then one has reason to respect or engage with  or honor or act in a certain way for the sake of it. No desire-based theorist, for example, will say that one is supposed to bring about one’s present desires (contrast the objects of these desires), or that one’s present desires are things that one somehow respects by satisfying. Accordingly, no desire-based theorist will say that one’s desires are themselves things of value. The desire-based theorist may say that other things are of value: in particular, that the objects of one’s present desires are valuable (“for one”?) because one desires them. The point is that what the desirebased theorist takes as the source of reasons is not itself of value—or, to be less committal, is not a source of reasons in the way that values characteristically are. This may seem, at first glance, to set Kantians apart from attitude-based theorists, because what Kantians see as the source of reasons—the rational will—is something that they also see as being of value and calling for respect. In the relevant regard, however, Kantianism is like the Desire-Based Theory: the rational will is not a source of reasons in the way in which values are sources of reasons. Kantians do not advance a basic normative principle that says that one has reason to do what respects the rational will, because the rational will is something of value that calls for respect. Instead, their basic principle is something like: If the rational will takes something as its object, then, because of this, it is valuable and one has reason to treat it in certain ways. They then claim that the rational will cannot but take itself as its own object. From this they derive, as a kind of theorem, that the rational will is valuable and something that we have reason to respect.


Reason, Value, and Desire

If this is right, then it suggests that the Thesis is really: Thesis: All sources of reasons for action are sources in the way in which values characteristically are. That is, all reasons for action derive from such sources, or are value-based. Accordingly, attitude-based theories should be understood as claiming: Attitude-Based Theory: No sources of reasons are sources in the way in which values characteristically are. All are sources in the alternative, attitude-based way. and hybrids should be understood as claiming: Hybrid: Some sources of reasons are sources in the way in which values characteristically are, whereas other sources of reasons are sources in the alternative, attitude-based way. This makes it clearer why a hybrid would involve a puzzling duality. The Thesis represents practical deliberation as seeking, ultimately, to trace reasons for action back to the kind of source that values or things of value represent. Perhaps we can make sense of the suggestion of attitude-based theories that practical deliberation takes an altogether different form: that it is a kind of existential “self-legislation,” or calculation of means to ends set by groundless desire. What is harder to make sense of is the idea that practical deliberation should take both forms: that we should at once seek to ground our reasons in value and accept as reasons what lacks any such grounding. What puzzles me, and may also puzzle Scanlon, is the difficulty of conceiving of a single, unified deliberative viewpoint that could integrate these two very different stances toward our reasons. This understanding of the Thesis is compatible with the cases that we considered earlier, in which it is granted that one’s attitudes do affect one’s reasons. In (5)–(7), my present attitude helps to constitute the value that is the source of the relevant reasons: my well-being, my resoluteness, my friendship. We might put this by saying that my attitude is part of an attitude-constituted value. In (1)–(4), the relevant value is constituted by something else: my mental health, my pleasure, my aunt’s interest in knowing the truth. My attitude simply affects whether a given action facilitates my doing what the relevant value provides me with reason to do. It plays a purely facilitative role. The crucial point is that in all of these cases, my attitudes affect my reasons only by affecting my value-based reasons. It is also worth noting that, in many of these cases, my present attitudes affect my reasons in an attitude-independent way: that is, in a way in which factors that are not my present attitudes can affect my reasons. For example, a blood test can tell me just as soon as my paranoid belief that I need my meds, I can deceive my aunt about an investment just as soon as about my desires, someone can cease to be my friend because of his feelings just as soon as because of mine, and my future

Aims as Reasons


desires determine what constitutes my well-being just as soon as do my present desires. However, in theory—that is, in abstraction from further claims about what values there are and what reasons they provide—the Thesis does not require that my present attitudes affect my reasons (if at all) only in an attitude-independent way. Indeed, in theory, the Thesis is even compatible with some or all basic normative principles being of the attitude-based form. For example, if one’s present desires were things of value to be respected—like human life, or the natural environment, or cultural achievements—and if respecting them consisted in doing what served them, then there might be a value-based basic normative principle that said that one had reason to do what served one’s present desires. However, we can set aside this abstract possibility, because it is uncontroversial that there are no such values. That is something that even attitude-based theorists accept.

2. explaining the appearance that our attitudes affect our reasons

To explain why one might wish to defend the Thesis, of course, is not yet to defend it. In what follows, I want to focus on a particular challenge to the Thesis: that our attitudes seem to affect our reasons, even though they seem not to affect our value-based reasons. To defend the Thesis, we need to explain away either the first appearance— that our attitudes affect our reasons—or the second—that our attitudes do not affect our value-based reasons.

2.1 The explanatory and the normative

In some cases, the appearance that our attitudes affect our reasons arises from confusing explanatory and normative reasons. If we ask why someone did or believed something, it may be appropriate to answer that it was because he had some attitude: because he desired, intended, or believed something else. In this case, we ask for, and receive, an explanatory reason for his belief or action. But this is a different question from whether he had normative reason for, whether there was anything to be said in favor of, believing or doing what he did.

2.2 Patterns of reasons

In other cases, the appearance that our attitudes affect reasons arises from a different sort of confusion. When we have an attitude, we often have reason for it. And when we have reason for an attitude, we typically have reason for other attitudes or responses. Reasons, we might say, come in patterns. One such pattern is reflected in GT. GT implies that if I have an aim, and I have reason to have it, then I have reason to take means to it. So it may appear that my having the aim gives me reason to take the means. But what in fact gives me reason to take the means is my having reason to have that aim. I would have the same reason to take the means even if I did not have that aim.


Reason, Value, and Desire 2.3. The evaluative and the deliberative

In other cases, the appearance that our attitudes affect our reasons results from overlooking a distinction that Scanlon has done much to clarify: the distinction between an attitude’s being deliberatively significant and its being critically or evaluatively significant. Reasons, at least in the sense with which we are concerned, are deliberatively significant. They have their home in the first-person standpoint of a deliberator trying to decide what to believe or do and, by extension, in the second-person standpoint of an advisor trying to help the deliberator reach that decision. Evaluations, by contrast, characteristically arise instead from the third-person standpoint of a spectator, who aims not to guide himself, or an advisee, but instead to appraise, after the fact, what someone else came to believe or do. One can be misled to think that one’s having an attitude is a reason for one to do something—a consideration that carries weight within the first-person standpoint of deliberation—when in fact it only makes it the case that one’s doing it would be subject to a certain kind of evaluation, from a third-person standpoint of assessment. Although this evaluation can take other forms, the form most relevant for our purposes is what Scanlon calls “structural irrationality.”  One’s attitudes can make it irrational for one to, or to refuse to, form or revise other attitudes. Of particular interest to us is the putative rational requirement: Means-End (ME): It is irrational of me (to intend at t to E, to believe at t that intending at t to M is a necessary means to my E-ing, but not to intend at t to M). As Scanlon observes, we have no reason, in general, to avoid irrationality. Nevertheless, when we are irrational, we can be criticized for malfunctioning, or manifesting a kind of vice. 3. the tie-breaking effect of intention

In other cases, however, it is harder to deny that our attitudes affect our reasons. I will consider one such case, or cluster of cases, to which Scanlon draws our attention. These are cases of underdetermination, in which, to start, I have merely sufficient reason to X, merely sufficient reason to Y, conclusive reason (to X or to Y), and lack sufficient reason (to X and to Y). Then I form an intention to X, and as a result, I come to have conclusive reason to take means to X and to lack sufficient reason to take means to Y. An example: I have merely sufficient reason to attend a conference in New York, merely sufficient reason to attend a conference in Boston, conclusive reason to attend one of these, and lack sufficient reason to attend both. As things stand, I have not made up my mind. The New York organizers call me to ask whether I would like to preregister, which, as they explain, is cheaper than registering on site. It would be reasonable for me to respond: “Sorry, I don’t know whether I’m going, so I don’t have reason to register yet.” Then I make up my mind to go to New York, and the organizers

Aims as Reasons


call back. Now, it seems, I do have sufficient reason to preregister for New York, whereas I still lack sufficient reason to preregister for Boston. Given that I have decided to attend the New York conference, the fact that preregistering for New York would serve, in an efficient way, the aim of attending the New York conference is a stronger reason than it was before. What is the alternative? Would it be correct for me to think that I continue to have just as much reason to take steps to attend the conference that I have decided against attending? Should I just as soon preregister for Boston, or book a flight there? The resources of the last section seem not to explain this phenomenon: Intention Effect, first pass (IE1): If I have merely sufficient reason to E1, merely sufficient reason to E2, conclusive reason (to E1 or to E2), and lack sufficient reason (to E1 and E2), then if I intend to E1, I have, because of this intention, stronger reason than I otherwise would to take means to E1-ing. GT does not explain IE1 unless intending the end somehow gives one stronger reason for the end itself. But that would itself require explanation. Nor it is plausible that we accept IE1 only because we confuse it with some principle of rationality. To begin with, this principle could not be ME. ME applies only to what I believe are necessary means, whereas in this case the means are not necessary. Preregistration is not required; it is just cheaper than the alternative. Next, ME requires me only to intend the means. However, what seems out of place in the case described is not simply to refuse to intend the means, but also to refuse to believe that I have greater reason to take the means. Suppose that I have decided to attend the New York conference. Whether or not I intend to preregister for New York, it would be odd of me to maintain that, so long as I am decided on New York and against Boston, it is a matter of indifference whether I preregister for New York or for Boston instead. If we were confusing IE1 with a principle of rationality, it seems, that principle would have to be: Believed Intention Effect (BIE): If I believe that (I have merely sufficient reason to E1, merely sufficient reason to E2, conclusive reason (to E1 or to E2), and lack sufficient reason (to E1 and E2)), intend to E1, and believe that (M is a means to E1-ing), then it is irrational of me to deny that, because of this intention, I have stronger reason than I would otherwise have to M. The problem is not that there is no such principle of rationality. There may well be. The problem is instead that BIE seems to presuppose IE1, read literally, as a distinct fact about reasons. Suppose that IE1 were false. Now consider a reflective agent who knew that IE1 was false: that after deciding to E1, one does not come to have stronger reason. According to BIE, rationality would require her to ignore what she knows: require her to believe, falsely, that in deciding to E1, she comes to have reason that she lacked before. How can rationality, as a rule, require us to take a false view of our reasons?


Reason, Value, and Desire

There is a more basic problem with the suggestion that ME, or BIE, or any other principle of rationality might explain the phenomena that lead us to accept IE1. These principles require something of the agent when (provided the other psychological conditions are met) and only when he believes that M (or intending to M) is a means to E. Because an agent’s rationality consists merely in the coherence of his attitudes, what matters for his rationality is merely what he believes, not what is the case. By contrast, IE1 does not apply when the agent falsely believes that M-ing is a means to E-ing. If in fact there is no preregistration for the conference, and I have simply misunderstood the situation, then spectators, realizing this, will not think that my decision makes it the case that I have reason to pay. They might advise me of this. And IE1 does apply even when the agent fails to realize that M-ing is a means to E-ing. If I do not realize that I must register for the conference, then observers will still think that my decision makes it the case that I have reason to preregister. Again, they might advise me of this. In sum, IE1 appears to be a claim about reasons, not about structural irrationality. If we cannot explain away the first appearance—that my intention affects my reason—then we can defend the Thesis only by explaining away the second appearance— that my intention does not affect my value-based reason. But it is not clear how my intention could affect my value-based reason. It would be a strange kind of action at a distance for my mere decision to go to New York to affect the value of attending that conference, say, by making it more important or illuminating. What Scanlon suggests, more plausibly, is that the decision to attend the New York conference affects my reasons by affecting my relation to the value of attending the New York conference. This is entirely in keeping with the value-based view. The reasons that something of value provides us with depend, in general, on our relation to that thing of value. As Scanlon puts it, adopting a goal changes one’s reasons by putting one in a new relation to the goal, just as, say, developing specialized skills would put one in a new relation to it. It is clear how acquiring skills tailored to a particular goal might change one’s reason. Those skills make it less costly for one to pursue that goal going forward, or make it more likely that one will accomplish it. But what new relation to the goal does merely intending it bring about, and how does this new relation affect one’s reason? Scanlon writes: What a person has reason to do in such a situation is to pursue one of these eligible goals. To do this, she must employ some procedure for selecting one of the goals and then pursue that goal, rather than any of the others. Once she has selected a goal, she is no longer in this situation. This might be put by saying that adopting a goal puts one in a different relation to that goal, and this makes a difference to what a given consideration gives one reason to do. But, again, what is it about this change in one’s “situation” that changes one’s reason? Unless more is said, one worries that this just abstractly describes the kind of explanation that the Thesis needs, rather than provides such an explanation.

Aims as Reasons


Indeed, one might worry that IE1 is straightforwardly explained by the intentionbased hybrid, which says, recall, that in addition to whatever value-based reason to E one may have, one’s intending to E provides nonvalue-based reason to E. In most cases, the explanation would say, this nonvalue-based reason is negligible. Either it is superfluous, because one has conclusive value-based reason to pursue E, or it is outweighed, because one otherwise has conclusive value-based reason not to pursue E. But when one’s value-based reasons otherwise underdetermine the choice, one’s nonvalue-based reason breaks the tie. So my intention to attend the New York conference makes it the case that I have conclusive reason to attend the New York conference, and in turn, by GT, makes it the case that I have conclusive reason to take the means to it. Doesn’t that—the hybrid theorist will say—fit the facts just about right?

4. the cost explanation

It is often noted that if one takes some means to E, or makes some “investment” in achieving E, this can make it the case that one has greater reason, on balance, to E. Having already taken some means toward the end, one can realize it, going forward, with less cost. So, on balance—that is, taking into account the costs and benefits— there is greater reason to E. Because this mechanism is predicated on actually taking means to E, it is usually thought to be irrelevant to IE1. But this may be an oversight, as Michael Bratman once pointed out to me in conversation. Strictly speaking, intending to E is taking some means to E-ing. Having intended to attend the New York conference, I can realize the end of attending the New York conference with one fewer step than it would take to realize the end of attending the Boston conference. Boston requires, whereas New York does not, that I take the step of changing my mind. This step requires deliberation, which typically carries some cost. It taxes my attention and mental energy, and it involves the irritation of an unsettled future. Recall, moreover, that there is no offsetting benefit. Because, as I already know, neither choice is better than the other, thinking about it more will not help me arrive at a better choice. Of course, I and others are rarely so articulate about the fact that changing our minds is a pointless burden. But we naturally think and say things that suggest that we are in fact responding to such factors: “You could spend all day going back and forth on this. But you’ve already made a decision, and you have other things to do.” Scanlon suggests something very close to this explanation. He rightly observes that the cost of deliberation provides one with “second-order purely pragmatic” reasons not to “reconsider” the decision (unless some evidence appears to cast doubt on the choice that one made). Reconsideration is a waste; to reopen the decision is to bear needless cost. However, Scanlon seems not to take the further step of saying that the cost of deliberation provides one with reason to pursue what one now intends. But whether or not he says it, it seems to follow from what he says. If pursuing another option requires one to reconsider, and if to reconsider is to bear needless cost, then to pursue the other option requires one to bear needless cost.


Reason, Value, and Desire

This “cost explanation” supports something stronger than IE1, in two respects. Intention Effect, second pass (IE2): When there is reason for me to E, if I intend to E, then I have stronger reason, on balance, than I would otherwise have to E and thus, by GT, to take means to E (which means presumably includes sustaining the intention to E). First, intending E makes it the case that one has stronger reason, on balance, not only to take means to E, but also to E itself. The cost of E-ing as a whole is lower than it was, and so, to that extent, it has become the better bargain. Second, the effect is not limited to tie-breaking situations. Suppose the New York conference was the only option, so that I had conclusive reason to attend it. It would still be the case that, having decided to attend it, I no longer need to take that step. To that extent, then, the prospective costs of attending are lower than they were, and so, again, it is an even better bargain than it was. Even if I lacked sufficient reason to attend New York, the lower prospective cost of attending would give me greater, although perhaps still not sufficient, reason, on balance, to attend New York. Of course, it is clear why this effect may be overlooked in non tie-breaking situations. It is most often, if not always, in tie-breaking situations that the modest change in cost makes it the case that one has conclusive reason to pursue the option that one has decided on. On this explanation, the effect of intending to E is purely facilitative. It makes it the case that an end can be achieved at lower cost. The general principle is: Cost: If some condition lowers the cost, going forward, of achieving some end, E, then, because of this, there is less reason against E-ing. As a result there is more reason, on balance, to E and hence, by GT, to take means to E. The cost explanation is compatible with the Thesis, therefore, so long as the reasons to pursue the end and to avoid the cost are value based. The effect of intention here is also attitude independent. Factors other than my having some attitude can lower the cost of achieving some end. The thing that I intend to buy, for example, can be put on sale. One worry about the cost explanation is that revising one’s intention is not always costly. There are situations, such as being stuck in an elevator or lying awake in bed, in which one has “nothing better to do” than to revisit the decision. The cost explanation would imply that, once one enters into such a situation, one no longer has greater reason to pursue the option one intends. But this seems like the right implication. If I have nothing better to do, then presumably I do not have the opportunity to take any means to my ends other than sustaining or revising my intentions themselves. So the test is whether, if I find myself in such a situation, reopen the decision to go to New York, and change my mind, I have done something that I lack sufficient reason to do, or failed to do something that I have conclusive reason to do. It seems, as the cost explanation would predict, that I have not. A different worry is that even when revising one’s intention is costly, this cost seems, in most cases, too modest to explain the more pronounced change that one’s intention

Aims as Reasons


seems to work on one’s reasons. So we turn to another effect of intention, which, together with the effect that we have been discussing, may help to explain this more pronounced change.

5. the effectiveness explanation

In general, one’s reason to intend to X depends on the epistemic probability that other conditions, that affect one’s reason to X, or whether one will succeed in X-ing, will or will not obtain. I have more reason to intend to picnic tomorrow, for example, if it is less likely to rain. These conditions are sometimes the actions of a person. I have more reason to intend to attend this conference if you are more likely to attend. And sometimes the person is the agent himself. I have more reason to intend to find accommodation where the conference is being held, if I am more likely to attend the conference. And I am more likely to attend if I intend to. When I intend something, this changes what the future is likely to hold. And, in general, changes in what the future is likely to hold, whether or not those changes are brought about by my intentions, change what I have reason to intend and do. This is the phenomenon that Scanlon calls the “predictive significance” of intention. Why, Scanlon asks, is it wrong for me to buy rat poison with the intention of using it to kill my wife, but not wrong to buy it with the intention of using it to kill only rats? His initial answer is that when I intend to kill my wife, but not when I intend to kill only rats, I am likely to perform the other acts (such as putting it in her soup) that together with my buying rat poison will be jointly sufficient for killing her. Thus, when I intend to kill my wife, but not when I intend to kill only rats, buying rat poison puts her at risk. In buying rat poison for myself, I am, so to speak, an accomplice to murder, just as I would be if I gave my neighbor rat poison, knowing that he was likely to use it to kill his wife. Likewise, when I intend to attend the New York conference but not the Boston conference, my registering for the New York conference makes it more likely that I will attend the New York conference—is a more effective means to attending the New York conference—than registering for the Boston conference would make it likely that I will attend the Boston conference. That is, if I intend to attend the New York conference, but do not intend to attend the Boston conference, then: the probability, conditional on registering for the New York conference, that registering for the New York conference helps to bring it about that I attend the New York conference is greater than: the probability, conditional on registering for the Boston conference, that registering for the Boston conference helps to bring it about that I attend the Boston conference.


Reason, Value, and Desire

The reason is that if I do not intend to attend the Boston conference, it is all but certain that I will not take other means that, along with registration, would be jointly sufficient for attending the Boston conference. I will not book a flight to Boston, I will not board a Boston-bound plane, etc. So preregistering for Boston will do next to nothing to increase the chances that I attend the Boston conference. It will be a waste. By contrast, if I do intend to attend the New York conference, it is likely that I will take the other means that, along with registration, will be jointly sufficient to attending the New York conference. Such is the “predictive significance” of that intention. More generally put: (i) When I intend to E, I am more likely to take some insufficient means M* to E-ing. (ii) When I am more likely to take some insufficient means M* to E-ing, there is a higher probability, conditional on my taking some other insufficient means M, that my M-ing helps to bring it about that I E: that is, M-ing is a more effective means to E-ing. (iii) Effectiveness (implied by GT): If some condition makes my taking some step, M (which different from this condition) a more effective means to some goal, E, that there is reason for me to pursue, then, because of this, there is more reason for me to M. In our case, the relevant condition is my intending to attend the New York conference, which makes preregistration for New York a more effective means. My intention plays a purely facilitative role. So this explanation is consistent with the Thesis, so long as the reasons to pursue the goal are value based. The effect of my intention is also attitude independent. Other factors—my legs’ being rested, my gas tank’s being full, my bank account’s being flush, and so on—can also play this role, by making some means more effective. The effectiveness explanation supports: Intention Effect, third pass (IE3): When there is reason for me to E, if I intend to E, then I have stronger reason than I would otherwise have to take insufficient means to E-ing. This differs from IE2 in that intending to E does not affect my reason to E, or to take sufficient means to E (which entails E-ing). This is because, even if I intend to attend the New York conference, the probability, conditional on attending the New York conference, that I will attend the New York conference and: the probability, conditional on attending the Boston conference, that I will attend the Boston conference

Aims as Reasons


are equally one. Instead, it affects only my reason to take insufficient means to E: for example, to preregister for the conference. IE3 also differs from IE1 in that the effect is not limited to underdetermined cases. Suppose that I have conclusive reason to attend the New York conference, but I have not decided to. Do I have sufficient reason to preregister? Probably not. If I decide not to attend the New York conference, then I will not take other means. So I will not attend the New York conference even if I preregister. So whatever reasons I might have for attending it do not translate into reasons for preregistering. And there may well be reasons against preregistering, such as that the money could be spent elsewhere. Thus, although the best option may be to preregister and to take other means, the second-best option, assuming that I will not take other means, is not to preregister at all. Although I have most reason (to intend to attend New York and to preregister), I have more reason (not to intend to attend New York and not to preregister) than (not to intend to attend New York and to preregister). Nevertheless, it is easy to overlook this effect in determined cases. The IE3 effect is more pronounced in underdetermined cases. In most cases, one is more likely to do something that one hasn’t yet decided to do if one has conclusive reason to do it than if one has merely sufficient reason to it, other things equal. Therefore, deciding to do something that one has conclusive reason to do has a smaller effect on the probability of one’s taking other means than has deciding to do something that one has merely sufficient reason to do. I turn now to some objections to the effectiveness explanation. First, it is not always true that intending the end makes insufficient means more effective. On the one hand, one’s intending the end might not make it significantly more likely that one takes other means, because, say, one will refuse to take those other means. On the other hand, intending the end might not make it significantly more likely that one takes other means, because it was already so likely that one would take those means. In these cases, according to the effectiveness explanation, one’s intending the end does not affect one’s reason to take the means in the way it otherwise would. Viewed from the outside, this seems true and unproblematic. If observers believe that I will never board the plane, say, then it will seem to them a waste for me to preregister. And if they are convinced that I will opt in the end to go to the New York conference, however sincere my professions of indecision, then it will seem to them that I ought to preregister now, before the fees increase. It will probably seem to them that I ought to decide to attend now too; I’m wasting my time turning it over in my mind, simply forestalling the inevitable. When we consider my point of view, however, things may be more puzzling. It would be strange for me to view my own intentions for the end as making no difference to whether I should take the means. On the contrary, we would expect me, upon making my decision, to see myself as having reasons that I did not have before. The explanation, I think, is that it would be difficult or impossible for me to believe what observers know, so long as my intentions are as described. If, in the first case, I knew what observers know—namely, that I will not take other means—then I could not sustain my intention for the end. One does not intend what one believes one will not do. If, in the second case, I knew what observers know—namely, that


Reason, Value, and Desire

I will decide on the end—then I would likely feel pressed to decide on it myself. Once one knows how one will decide, it is bound to strike one as idle to continue wrestling with it as though it were a live question. The second objection asks why premise (i) in the effectiveness explanation is true when it is. Why am I more likely to take other insufficient means to attend New York when I intend to attend New York? Given that I am a rational agent, the explanation why I will do something is, in general, that I believe that I have reason to do it. So it is natural to answer: Because I believe that, having intended to attend New York, I have stronger reason to take those other means to attend New York. But why do I believe this? If I believe it because it is true, then we seem to be appealing to the very fact that we sought to explain: that intending an end gives me stronger reason to take insufficient means to it. How can we break out of this circle? One reply would be to deny that I am more likely to take other insufficient means to attending New York because I believe that, having intended to attend New York, I have stronger reason to take those other means. Instead, it might be suggested, it is because I believe that, having intended to attend New York, it would be irrational of me to refuse to intend those insufficient means to attend New York, and when I believe that I am rationally required to do something, I am likely to comply. One problem with this reply is that, if ME is the only relevant requirement of rationality, then I am rationally required to take insufficient means to what I intend only insofar as I believe that intending those insufficient means is itself a necessary means. However, I need not always believe this. The more serious problem with this reply is that, as I suggested in Section 2.3, we do not consider what rationality requires of us in deliberating what to do. Considerations of our own rationality are deliberatively inert; they have only evaluative significance. So, if I am more likely to take other insufficient means to attending the New York conference, it must be because I believe that, having decided to attend the New York conference, I have stronger reason to take those other means. The reason why I believe this lies, I think, in the cost explanation. Having decided to attend the New York conference, I correctly believe that I have stronger reason to take insufficient means to attend the New York conference, because it is now the cheaper option. And because I expect that I will continue to believe this, I expect that I will take insufficient means to attend the New York conference. This appeal to the cost explanation does not make the effectiveness explanation superfluous. The increase in the effectiveness of insufficient means gives one further reason, in addition to the cost of reconsideration, to take insufficient means. Recall that the main objection to the cost explanation was that it accounted for only a modest effect on one’s reason to take means to one’s end. The addition of the effectiveness explanation accounts for the more pronounced effect that decision seems to have. 6. defending predictive significance: too “theoretical”?

Some may protest that the idea of predictive significance, which underlies the effectiveness explanation, attributes to agents an overly “theoretical” stance toward their own present intentions and future actions. To caricature the idea of predictive

Aims as Reasons


significance: Agents are supposed to observe that they intend certain things, to base predictions of their own future behavior on those observations, and then to “work around” that future behavior and its effects, as though they were obstacles heaved up by brute nature or some alien intelligence. That is fine for predictions of inanimate events, or the behavior of others, but it makes no sense in one’s own case, when the intentions and actions are under one’s control. The first distortion of this caricature is that the idea of predictive significance requires that our knowledge of what we intend be somehow observational. The idea of predictive significance is compatible with the fact that our knowledge of what we ourselves intend is radically different from our knowledge of what others intend. The second distortion is that there needs to be, or usually will be, explicit reflection on what one intends. I suspect that our expectations about what we will do in the further future, and perhaps even our awareness of what we are currently intentionally doing, are, insofar as they are reasonable, based on inferences from the fact that we intend what we do. But these inferences, like many other simple inferences, do not require us to reflect explicitly on their premises: the facts that we intend this or that. The last distortion is that the idea of predictive significance asks the agent to treat his present intentions or future actions as not under his control, or not his responsibility, or not things for which he can have or lack reasons. Predictive significance asks the agent to treat his present intentions or future actions like natural occurrences or the actions of others in one respect: namely, as factors that affect the likelihood of, or determine, what will happen in the future, which in turn affects what he has or lacks reason to intend or do now. But nothing in the idea of predictive significance requires the agent to treat his present intentions or future actions like natural occurrences or the actions of others in another respect: namely, that they are not under the agent’s control, or not his responsibility, or not things for which he has or lacks reasons.  Presumably, the agent needs to view his own agency in both lights. On the one hand, he needs to see his present intentions and future actions as things under his control, to consider the reasons that he has for them, and to respond accordingly by forming, revising, or abandoning his present intentions. On the other hand, he also needs to be sensitive to what those formations, revisions, and abandonments mean for the future, and to what this future then means, by a kind of feedback, for his reasons. What, after all, is the alternative? Suppose that we could not base any expectations of our future behavior on what we currently intend. Practical reasoning, which relies pervasively on such expectations, would be crippled. It is not clear how we could ever take means to our ends, or coordinate our actions. Theoretical reasoning would also be crippled, because many of the future events about which we form expectations depend on what we will do. Fissures would appear elsewhere. Our generalizations about peoples’ behavior would break down, without explanation, in our own case. And others’ predictions of our own behavior would seem to us groundless.


Reason, Value, and Desire 7. defending predictive significance: scanlon on self-facilitation

In motivating the idea of predictive significance, I invoked an explanation of Scanlon’s, which we might call the “appeal to self-facilitation.” This is the explanation that, when I intend to poison my wife, my buying rat poison is wrong, for the same reason that, when my neighbor intends to poison his wife, my giving him rat poison is wrong: because it facilitates a murder. The difficulty is that Scanlon later distances himself from the appeal to self-facilitation, for reasons that might seem to threaten the idea of predictive significance itself. Scanlon observes that, whereas in the neighbor case it seems natural to say that what I am doing is wrong because it facilitates a murder, it seems odd about to say this when I myself am the prospective murderer, at least if I expect that I will control whether or not I actually poison my wife. If instead I expect that I will not be able to control my murderous urges, then buying the rat poison for my future self is much like giving it to my neighbor. But if I do expect that I will control whether I poison my wife, then “the idea of facilitation involves an odd relation between [me] and [my] own future conduct” (Scanlon, Moral Dimensions, ch. 2). I agree that there is something odd about the appeal to self-facilitation. But I do not think that the explanation of what is odd undermines the idea of predictive significance, or the suggested uses of it. What’s odd about the appeal to self-facilitation, at least in the first instance, seems to lie in how the appeal expects the agent to reason. The initial parts of his reasoning, however, seem unproblematic: (1A) I intend to poison my wife. (2A) If I intend to poison my wife, I have reason to believe that if I get rat poison, I will poison her (even though I know that this will be under my control). (3A) If I have reason to believe that if I get rat poison, I will poison my wife (even though I know it will be under my control), and if I have no other compelling reason to get rat poison, then because of this, it is wrong for me to get rat poison. (4A) So, it is wrong for me to get rat poison. One might worry that (1A) and (2A) are “too theoretical,” but, if the argument of the last section is correct, this worry has little merit. At worst, (1A) and (2A) are unnaturally explicit. (3A) follows from applying to the facts of the case the following plausible moral principle: Control Principle: If I have no compelling reason to X, and I have reason to believe that X-ing will significantly increase the risk of harm to someone, then it is wrong for me to X. Of course, (3A) would not follow if we rejected the Control Principle in favor of: Noncontrol Principle: If I have no compelling reason to X, and I have reason to believe that X-ing will significantly increase the risk of harm to someone, but not through events that I control, then it is wrong for me to X.

Aims as Reasons


But the Control Principle seems more compelling morally, especially within a contractualist framework. On the one hand, the objections to permission are weaker against the Control Principle. From the victim’s point of view, it seems irrelevant whether the route from X-ing to the harm depends on events that the agent controls. What matters is simply that X-ing increases the risk of harm. On the other hand, there seems no valid objection to prohibition against either principle, because the agent has no compelling reason to X. So what makes the reasoning in the appeal to self-facilitation so odd? Amplifying this question is the fact that many structurally similar explanations, which Scanlon himself describes, do not seem odd in the same way. Withdrawing from the appeal to self-facilitation, Scanlon suggests the fallback position that, although buying the poison is not wrong, refusing to abandon my intention to kill my wife is wrong. But why is it wrong? Scanlon suggests that it would not be wrong to stick pins in a voodoo doll with the intention of killing my wife, if I had no reason to believe that it would actually kill my wife. This suggests that it would not be wrong to have the intention to kill my wife unless I had reason to believe that the intention would actually lead to my killing her. But if so, then my reasoning about keeping the intention is similar to my reasoning about buying the poison: (2D) I have reason to believe that if I intend to poison my wife, I will poison my wife (even though I know that it will be under my control to revise this intention). (3D) If I have reason to believe that if I intend to poison my wife, I will poison my wife (even though I know that it will be under my control to revise this intention), then, because of this, it is wrong for me to intend to poison my wife. (4D) So, it is wrong to intend to poison my wife. Moreover, when explaining how intentions can make certain actions permissible, rather than, as in the rat poison case, impermissible, Scanlon suggests that one might correctly reason: (1C) I intend to use embryos for therapy. (2C) If I intend to use embryos for therapy, I have reason to believe that, if I create an embryo, I will use it for therapy. (3C) If I have reason to believe that, if I create an embryo, I will use it for therapy, it is permissible, whereas it would otherwise be impermissible, to create an embryo. (4C) So, it is permissible to create an embryo. (1D) I intend to fight the fire. (2D) If I intend to fight the fire, then I have reason to believe that, if I cross your land, I will use this crossing to fight the fire. (3D) If I have reason to believe that, if I cross your land, I will use this crossing to fight the fire, then it is permissible, whereas it would otherwise be impermissible, to cross your land. (4D) So, it is permissible to cross your land


Reason, Value, and Desire

These analyses are even closer to the appeal to self-facilitation. They involve an agent’s judging something about the permissibility of a present action on the grounds that it would facilitate a future action of his that he has reason to believe that he will perform because of what he currently intends. And yet they do not seem odd in the same way. Why? The answer, I think, lies in the steps the reasoning takes next: (4A) So, it is wrong for me to get rat poison, because putting my wife at risk is wrong. (5A) So, I will not get rat poison. (6A) But, of course, I still intend to kill my wife, even though it puts my wife at risk. This further stretch of reasoning is incoherent, because it involves simultaneously taking and not taking the risk to my wife as a reason. That is, I judge in (4A) that the risk to my wife means that I lack sufficient reason to buy the poison, and I respond accordingly in (5A) by not buying the poison. But in (6A) I fail to drop the intention to poison her, either because I fail to judge that the risk to my wife means that I lack sufficient reason to intend to poison her, or because I fail to comply with that judgment. Why do we assume (6A): that I do not drop the intention to kill my wife? Because if I did drop the intention, then I could not sustain the judgment in (4A) that my getting rat poison puts her at risk. It would be obvious to me that (1A) and (2A), on which (4A) depends, were no longer true. To put it paradoxically, my intention is in tension with the very reasoning that it supports. The general structure of the incoherence is: (i) One intends to X. (ii) One believes that, because one intends to X, it is likely, conditional on one’s taking some insufficient means, M, to X-ing, that one’s M-ing will help to bring it about that one X’s. (iii) One concludes that, because it is likely to help to bring it about that one X’s, one lacks sufficient reason to M, and so refuses to M. (iv) One does not similarly conclude (or does not respond to the similar conclusion) that, because it is likely, conditional on one’s intending to X, that intending to X will help to bring it about that one X’s, one lacks sufficient reason to intend to X—as is evidenced by one’s continuing to intend to X. Notice that the same incoherence does not arise in the other, contrasting cases that we have considered. The reasoning in (2B)–(4B), in which I abandon the intention to kill my wife, because I believe that that intention puts her at risk, leaves no opening for such incoherence, because there is only one judgment about my reasons, and I comply with it. Consider next the neighbor case. I can coherently (a) judge that I lack sufficient reason not to give my neighbor rat poison because, given that he intends to

Aims as Reasons


poison his wife, my giving it to him puts her at risk, (b) judge that he lacks sufficient reason to intend to poison her, because his having that intention puts her at risk, (c) comply with (a), and (d) judge that he fails to comply with (b). Or suppose that I expect that I will not be able to control my future actions. No matter what I intend now, if I get rat poison, I will poison my wife. In this case, I can coherently judge that I lack sufficient reason to intend to poison my wife, because this puts her at risk, and that I lack sufficient reason to buy the poison, because that puts her at risk. And responding to the first judgment now does not falsify the second judgment. Even if I do not intend to kill her now, buying rat poison still puts her at risk, because I will intend it later. Cases with the structure of the conference decision also escape the incoherence. My intending to attend the New York conference fits the following coherent pattern: (i) One intends to X (e.g., to attend New York). (ii) One believes that, because one intends to X, it is likely, conditional on taking some insufficient means, M, to X-ing, that M-ing will help to bring it about that one X’s. (iii) One concludes that, because it is likely to help to bring it about that one X’s, one has sufficient reason to M and so M’s. (iv) One does not reject the conclusion that, because it is likely, conditional on intending to X, that intending X will help to bring it about that one X’s, one has sufficient reason to intend to X—as evidenced by one’s continuing to intend X. The embryo creation and land-crossing cases, (1C)–(4C) and (1D)–(4D), share this coherent structure. I can coherently judge that fighting the fire is sufficient reason to cross your land given that I intend to fight the fire and that fighting the fire is sufficient reason to intend to fight the fire in the first place. And my responding to the second judgment—that I have sufficient reason to intend to fight the fire—verifies, rather than falsifies, the first judgment—that I have sufficient reason to cross your land. Thus, the oddity of the appeal to self-facilitation does not threaten the idea of predictive significance or the applications of it that we have considered.

8. what matters to us

So far, I have been asking how having an intention, for whatever object, might affect one’s reason for action. However, there is something else that often comes under the label “having an aim.” To “have an aim” in this sense is to “identify” with, to be “committed” to, to “care” about, to “value” some life-structuring pursuit. It is for some cause, career, or calling to be “important” or “matter” to one. “Having an aim” in this sense differs from having an intention at least in taking a narrower range of objects. Although one can intend to pick up dry cleaning or to order a sandwich, these are not activities that can be “important to” one or that one can “care about” in the relevant


Reason, Value, and Desire

sense. “Having an aim” in this second sense also differs from having an intention in the psychological orientation it involves. Merely deciding to pursue a career is not the same as “caring about” or “valuing” it. For one thing, when one has to decide whether (say) to stick with a career, one’s decision often is preceded by an attempt to figure out what one really cares about or values. For another, the decision is something that one can get right or wrong, by sticking with, or giving up, what really “matters” to one. When the judge’s son opts for law school over a promising future as a jazz musician, we may think that he has given in to expectations, or shrunk from the risk of failure, instead of pursuing what is truly “important to” him. It is not clear how either of these things could be so if “caring” consisted in the decision itself. What is it to “care” about something? The most plausible account, I think, is that it is to be emotional vulnerable to it: to be disposed to experience positive emotions when it fares well or is properly regarded by others, and to experience negative emotions when it fares poorly or is not properly regarded by others. If I care about my work, I will feel elated when it advances and dejected when it is set back. I will feel affirmed when others admire my work, crestfallen when they call its flaws to my attention, and resentful when they dismiss it too hastily. It is natural to think that having an aim, in this sense, gives me reason that I would not otherwise have. If I care about my life’s work, if it matters to me, then I have reason to pursue it that someone else, who does not have that relationship with it, lacks. And if I care about my life’s work, I have reason to pursue it that I do not have to pursue other aims, to which I am not attached in the same way. To be sure, the work itself must be independently worthwhile for my being invested in it in this way to provide me with reasons. The work must be something that everyone, whether or not he or she cares about it, has reason at very least to respect, if not also to support. I have no reason to pursue the project of counting blades of grass on the asylum grounds, for example, no matter how much, in my madness, I may care about it. Nevertheless, when a pursuit is independently valuable, my caring about it seems to give me further reason beyond those that its independent value gives everyone. The challenge for the Thesis is to explain how caring about a pursuit can affect one’s value-based reasons in this way. It might first be suggested that the fact that an aim matters to me can make it more likely that I will succeed in it. The more likely I am to succeed in an aim, other things equal, the stronger my reason to pursue it. As Scanlon writes: [B]eing drawn to a pursuit is (at least under favorable conditions) a condition for having a good reason to undertake it as a career. For one thing, if one takes up a career which “leaves one cold,” then one is unlikely to succeed in it. One problem with this suggestion is that it is not always true that I am more likely to succeed in what matters to me more. First, it is a familiar, if tragic, phenomenon that one can be less likely to succeed at something precisely because it matters more to one. Because one cares about an aim so much, one may be overly anxious about it, or perversely inclined to self-sabotage. Second, people sometimes reliably pursue aims that

Aims as Reasons


they do not care much about. They need to earn a living, or are creatures of habit. Third, one may simply have greater talent for an aim that matters less to one. Even so—even though I am less likely to succeed in what matters to me more—this does not seem to extinguish the force of my reason to do what matters to me more. Of course, the fact that I am more likely to succeed in something else may outweigh this reason. But that is compatible with the present point: namely, that the reason that arises from its mattering to me does not itself depend on whether I am more likely to succeed in it. A second suggestion appeals to an implication of our account of caring: that when I care more about a pursuit, I will experience more intense positive emotions when it goes well. The prospect of these welcome experiences, the suggestion runs, is what gives me reason to pursue aims about which I care. One problem is that caring about a pursuit also makes me vulnerable to more intense negative emotions when it goes badly. It is enough to call to mind the cliché of the tormented artist, who “suffers for his work,” to see that pursuing what matters to us need not make our lives more pleasant. And yet the prospect of such suffering seems not to extinguish the force of the reason to do what one cares about, although, of course, it may outweigh it. Moreover, this suggestion seems too extrinsic. Even if pursuing art would be pleasant for a talented, well-adjusted artist, it would be oddly alienated for him to pursue his artwork rather than some no less valuable project, on the grounds that, given his psychological make-up, it promised more pleasant experiences. A similar problem recurs for a third suggestion: that I have reason to pursue an aim that I care about because pursuing it enhances my well-being. It is plausible that one’s well-being consists, inter alia, in the successful pursuit of independently valuable aims that one cares about. What is less plausible is that I have reason to enhance my own well-being, or at least that the reasons that I have to pursue what I care about are reasons to enhance my well-being. Again, this seems too extrinsic. As both Scanlon and Raz have argued, agents typically view their aims themselves, and not those aims’ contribution to their well-being, as the source of their reasons to pursue them. This is thrown into relief by an example of Raz’s, in which, as he puts it, “a person’s life loses nothing by a forced change of career, and yet where this person is acting reasonably in trying to stave off the change, being willing to pay a considerable price in the process.” Let us imagine a ballet dancer enjoying a reasonably successful career with a small provincial ballet company, the only one within hundreds of miles. Then dwindling audiences threaten the future of the company. If it is forced to disband our dancer will have to abandon ballet, to change career and look for something else to do. It does not surprise us that he regards the prospect as a great personal disaster. I think that it would be agreed that it is reasonable for him to try to prevent the collapse. . . . Let us assume that the company has to close. Our dancer looks for other possibilities and starts a new career as a theatre director with the local theatre company where he remains until his retirement. He quickly comes to like his new work, enjoying a success comparable to his success in his first career.


Reason, Value, and Desire

Even if we idealize the case, so that the dancer knows that this is what will happen, that his life will be no worse when he abandons ballet, Raz observes: it is still reasonable for him to regard the threat to his dancing career as a disaster for him. He wants to be a dancer not a theatre director. He knows that if he is forced into theatre he will come to like it and be reasonably successful. But that is not what he wants and he is willing to try hard to stay with dancing. These admissions of his entail, on the assumption that his feelings and actions are rational, that he is not concerned with his well-being. His concern is to have the life he wants to have, meaning the life he has become committed to. The dancer’s reasons to stay with dancing neither derive from the prospects for his well-being, nor from his greater likelihood of success at dancing, nor from any supposed greater value of dancing. Instead, his reasons seem to derive from the simple fact that dancing, not directing, is the life he “wants to have.” This may seem a telling admission: a concession that, after all, we need desire-based reasons, or attitude-based reasons of some kind, to account for the reasons that the dancer takes himself to have. But this is not, I take it, what Raz has in mind. Consider these two passages from his earlier book, The Morality of Freedom: Previously, I have argued that wanting something is not a reason for doing it. We can see now that, while fundamentally right, in one respect that claim was exaggerated. Saying “I want to . . .” can be a way of indicating that one is committed to a project, that one has embraced a certain pursuit, cares about a relationship (389). In embracing goals and commitments, in coming to care about one thing and another . . . [o]ne creates values, generates . . . reasons which transcend the reasons one had for undertaking one’s commitments and pursuits (387). One might put it this way: The fact that the dancer cares about dance means that his pursuing dance constitutes something of further value—a value that would be not constituted if he did not care about it: if he pursued it coldly, say, from a sense of obligation or inertia. This further value is what provides him with reason to stave off a change, even a change for something in which he would be no less happy or successful. The suggestion, put generally, is that a certain kind of attitude-constituted value is realized by one’s pursuing an independently worthwhile aim that one cares about. It is this attitude-constituted value that, compatibly with the Thesis, provides one with further reasons to pursue such an aim. There are, however, two straightforward worries about this proposal. The first worry, to which I will return, is that it may seem not to explain Raz’s dancer after all. Shouldn’t the proposal lead us to say that by bringing about the change in career and coming to care about his new life as a theater director, the dancer will realize another instance of the same kind of attitude-constituted value, no worse than the first? If so, doesn’t it follow that, as far as the value is concerned, he has no more reason to stave off the change than to accept it?

Aims as Reasons


The second worry, more immediately pressing given the concerns of this chapter, is that the proposal does nothing more than to affix the honorific title “value” to a more straightforward hybrid explanation: that some of our reasons to pursue an aim derive from the independent value of that aim, whereas other reasons derive, independently of value, from our present caring about it. Why think that caring ushers a new value onto the scene? One reason is that the hybrid explanation, on closer inspection, is unconvincing. First, as we noted earlier, caring about a pursuit appears to provide one with reason only if the pursuit is independently valuable. The hybrid theorist might try to explain away this appearance. Even when the pursuit is not independently valuable, caring about it provides one with reason. It is just that this reason is usually outweighed by the opportunity cost: the value-based reason to devote one’s resources to a pursuit that is independently valuable. This is similar to the earlier suggestion of the intention-based hybrid that the nonvalue-based reasons that arise from our intentions are so weak that we notice them only in tie-breaking contexts. But this explanation seems less plausible in the present case, for the simple reason that caring about a pursuit, at least when the pursuit is independently valuable, seems to provide one with stronger reasons than would a mere intention to pursue it. One might think, for example, that one has reason to stick with a career that one cares deeply about rather than to abandon it for a career that would be, in the grand scheme of things, more valuable. But it is not clear that a bare decision to pursue the former career would have this effect. Second, the fact that one cares about an independently valuable pursuit typically also gives others reasons: for example, not to undermine one’s pursuit of it. The hybrid theorist might suggest that this is because undermining one’s pursuit of what one cares about also undermines one’s well-being. But this need not always be so, as Raz’s example suggests. Or the hybrid theorist might suggest that respect for one’s autonomy calls for us not to undermine one’s pursuit of what one cares about. But what we respect, in such cases, is not, or not simply, one’s self-regarding choices. We would not have the same reason to respect a choice to pursue something, such as counting blades of grass, that is not independently valuable. Indeed, our reason may not be to respect a decision at all, because the pursuit about which the agent cares, and whose claims we recognize, may be one that he himself has chosen, for bad reasons, to give up. On closer inspection, the relevant kind of “autonomy” seems little different from the attitude-constituted value that we have proposed. Perhaps it just is that value, viewed, as it were, from the outside. Finally, even when the pursuit is independently valuable, currently caring about a pursuit provides one with reason only if one has also had a history of pursuing it, with the right sort of concern, up until now. Someone who just found himself, out of the blue, caring about pursuing a life of dance—even if he had the same athleticism and feel for music—would not have the same reasons as Raz’s dancer, who has, as Raz puts it, “become committed to” that life over time. Indeed, it would seem out of place— “premature,” as we might say—for the former to invest dancing with the same importance in the first place. (Of course, people have reason to embark on pursuits to which


Reason, Value, and Desire

they are attracted. But being attracted, in this sense, is something different from caring about the pursuit, in the way that Raz’s dancer does. Indeed, the normative significance of being attracted to a pursuit may be just that it indicates that one might, in due course, come to care about that pursuit, and so constitute the relevant value.) The attitude-constituted value in question, when fully spelled out, is that of pursuing an independently valuable aim, that one cares about, where one has a history with that aim of a kind that makes sense of that caring. Whatever the merits of the alternative hybrid explanation, it should not be a foreign idea that there is a value with this structure. We are familiar with other such values, such as friendship. One might “go through the motions” (or at least some of the motions) of friendship, to save the feelings of a pathologically sensitive acquaintance, without actually caring about that relationship. One would not have a friendship, or the reasons that a friendship provides. One has a friendship only if one currently cares about one’s friend and one’s relationship with him, and has the sort of history with him that would make sense of this. Perhaps, if one has a friendship, then one has additional reason for interacting with one’s friend, because his well-being and one’s own are now hostage to it, or because one is more effective in ministering to his needs, or because one takes greater pleasure in doing so. But surely the more fundamental and important reasons flow from the friendship itself. Friendship just is a distinctive kind of value, which is constituted by a present emotional vulnerability to someone and to relations with him and by the kind of history that makes sense of this vulnerability. The present suggestion is just that a distinctive kind of value is similarly constituted by pursuing something of independent value, when one is currently emotionally vulnerable to its pursuit, and when one has a history of pursuing it that makes sense of this vulnerability. The analogy to friendship in turn helps answer the first of our two worries: that the proposal might not explain why Raz’s dancer has more reason to stave off the change in career than to accept it. After all, the alternative is to come to pursue and to care about pursuing being a theater director, which would instantiate the same kind of attitude-constituted value. But just as the value of friendship need not give me as much reason to cultivate a new friendship as it gives me to cherish an existing one, likewise this attitude-constituted value need not give me as much reason to cultivate a new pursuit as it gives me to sustain my existing one. The mistake, put broadly, is to think that things of value are sources of reasons only in the sense that, when we are able to bring about something of value, we have reason to do so. But as Value-Provision reflects, and as Scanlon has cautioned, this is overly narrow. Things of value can provide us with reasons when we stand in relations to them other than being able to bring them about—such as being currently engaged with them—and the reasons that they provide us with may be to do things other than to bring them about—such as to honor or respect them in suitable ways. This brings to the foreground, however, my assumption that the attitude-constituted value can play a role in the agent’s own deliberation, which might seem open to question. People who are fully engaged in pursuits that they care about are usually focused on the features that make it independently valuable. Thus, we would expect

Aims as Reasons


the dancer to be concerned with the aesthetic and expressive qualities of choreography and execution, and not to be given to retrospection on his history with dance, or introspection on his current feelings about it. It seems likely that this special value rises to deliberative salience only from what Jay Wallace calls the standpoint of “eudaimonistic reflection,” where we step back from our relationships and pursuits, and survey our lives and their sources of meaning as a whole. When we are contemplating some life change, for example, it is entirely natural to reflect on our past with some project and to engage in “soul-searching”: an attempt to get clear on whether it is what we really care about. Much the same is true of friendship. When we spend time with friends, we focus on their charms and the pleasure of their company, not on our histories with them or our own present feelings. But, in other contexts, such as when our loyalties are tested, considerations about what we have been through together and how our friend matters to us regain prominence.

9. conclusion

I have been addressing only one version of the challenge that our attitudes seem to affect our reasons without affecting our value-based reasons. There are other versions of this challenge, of course, arising from the perceived effects, not of aims, but instead of beliefs and desires. And there are other challenges, which stem from the perceived “metaethical,” or “second-order,” advantages of attitude-based theories. Only such theories, it is thought, can explain the metaphysics, epistemology, or motivating force of reasons. If I have not dealt with such challenges, it is partly from a conviction that the resilience of attitude-constituted theories owes largely to their “first-order” or “substantive” plausibility. They would not enjoy their perennial appeal were it not obvious—an utterly familiar part of human experience—that our beliefs, wants, and decisions make a difference to what we have reason to do. Of course, it is also obvious—an unavoidable consequence, I think, of taking up the deliberative point of view—that we see our reasons as flowing from what is valuable, or good, or worthwhile. As I hope to have suggested, insistence on reconciling these commonplaces, as well as grounds for hope that reconciliation is possible, are among Scanlon’s many contributions to practical philosophy. notes * I am very grateful to private comments from Tim Scanlon, Sam Scheffler, Jay Wallace, and Wai-Hung Wong, as well as to public comments from participants in the Seminar on Ethics and Normative Theory (SENT) at Stanford; at the University of Valencia, where Eduardo Ortiz delivered prepared comments; and at the University of Stirling, where Kent Hurtig delivered prepared comments. 1. For instances of this claim, focusing variously, and often selectively, on belief, intention, or desire, see Michael Bratman, Intention, Plans, and Practical Reason (Cambridge, Mass.: Harvard University Press, 1987) and “Intention, Practical Rationality, and Self-Governance,” Ethics 119 (2009): 411–443; John Broome, “Normative Requirements,” Ratio 12 (1999): 398–419; “Are


Reason, Value, and Desire

Intentions Reasons? And How Should We Cope with Incommensurable Values?” in Practical Rationality and Preference: Essays for David Gauthier, ed. Christopher Morris and Arthur Ripstein (Cambridge: Cambridge University Press, 2001), 98–120; and “Practical Reasoning,” in Reason and Nature: Essays in the Theory of Rationality, ed. José Bermùdez and Alan Millar (Oxford: Oxford University Press, 2002), 85–111; John Brunero, “Are Intentions Reasons?” Pacific Philosophical Quarterly 88 (2007): 424–44; Garrett Cullity, “Decisions, Reasons, and Rationality,” Ethics 119 (2008): 57–95; Jonathan Dancy, Practical Reality (Oxford: Oxford University Press, 2000); Richard Moran, Authority and Estrangement (Princeton, N.J.: Princeton University Press, 2001), ch. 3; Derek Parfit, “Reasons and Motivation,” Proceedings of the Aristotelian Society 71 (suppl.) (1997): 99–130; and On What Matters (Oxford: Oxford University Press, forthcoming); Joseph Raz, The Morality of Freedom (Oxford: Oxford University Press, 1986), ch. 6; Engaging Reason (Oxford: Oxford University Press, 1999), ch. 3; “The Myth of Instrumental Rationality,” Journal of Ethics and Social Philosophy 1(1)(2005): 1–28; and “Instrumental Rationality: A Reprise,” Journal of Ethics and Social Philosophy, Symposium 1 (2005): 1–19; T. M. Scanlon, What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1997), esp. ch. 1, sec. 9; “Reasons: A Puzzling Duality?” in Reasons and Value: Themes from the Moral Philosophy of Joseph Raz, ed. R. Jay Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith (Oxford: Clarendon Press, 2004), 231–46; and Moral Dimensions: Permissibility, Meaning, and Blame (Cambridge, Mass.: Harvard University Press, 2008); Dennis Stampe, “The Authority of Desire,” Philosophical Review 96 (1987): 335–81; Barry Stroud, Engagement and Dissatisfaction: Modality and Value (New York: Oxford University Press, 2010), ch. 4; Judith Jarvis Thomson, Goodness and Advice (Princeton, N.J.: Princeton University Press, 2001), 24–29; and R. Jay Wallace, “Normativity, Commitment, and Instrumental Reason,” Philosophers’ Imprint 1(3)(2001): 1–26. I should note that Bratman allows that intentions can provide “framework” reasons. This may be compatible with the Thesis, however, if framework reasons are identified with requirements of structural rationality of the kind discussed in section 2. For discussion of a similar thesis about belief, see Hannah Ginsborg, “Reasons for Belief,” Philosophy and Phenomenological Research 72 (2006): 286–318. 2. See What We Owe to Each Other, 18–22. 3. I discuss GT (as well as its relation to the Facilitative Principle of Raz, “The Myth of Instrumental Rationality,” 6, to which it is indebted) in my “Toward a Principle of Instrumental Transmission,” unpublished. In the rest of this chapter, I will often drop “(nonsuperfluously)” and “that one E’s and . . .” for less cumbersome formulations. Judgments about what there is reason to do and judgments about what is likely are relative to a body of information. The body of information need not be the agent’s. After all, when we have more information than the agent, our advice about what he has reason to do is typically in light of our information, not his. In general, I believe, the relevant body of information is selected by the context of the person assessing the reasons- or likely-claim for truth or falsity. See my and John MacFarlane’s “Ought: Between Objective and Subjective,” unpublished ms. 4. See, for example, Dancy, Practical Reality, 40–41, 55–56; Ginsborg, “Reasons for Belief”; Parfit, On What Matters; Raz, Engaging Reason, 59–60; Scanlon, What We Owe to Each Other, ch. 1, sec. 9, and Moral Dimensions, ch. 2; Thomson, Goodness and Advice. 5. Parfit, On What Matters. 6. Broome, “Are Intentions Reasons?” 7. Scanlon, What We Owe to Each Other, ch. 1, sect. 9, and Moral Dimensions, ch. 4. 8. I am not understanding attitude-based theories as metaphysical accounts of relation of being reason for, which claim either (i) that the relation that we naively call “being reason for”

Aims as Reasons


is reducible to some nonnormative relation involving attitudes, or (ii) that what we naively call “beliefs that something is reason for something else” are reducible to attitudes, other than belief, toward nonnormative contents. Instead, I am taking attitude-based theories to be substantive, although broad, theories of what reasons we have. This is not to deny, of course, that some of the appeal of attitude-based theories, understood as substantive normative theories, may derive from their being confused with such metaphysical theories. See Stroud, Engagement and Dissatisfaction, ch. 4; and Parfit, On What Matters. Nor is it to deny that some deliberately present attitude-based theories as metaphysical theories. See Mark Schroeder, Slaves of the Passions (Oxford: Oxford University Press, 2007), esp. ch. 4, 11. Nor am I addressing views, such as that of Michael Smith, The Moral Problem (Oxford: Oxford University Press, 1994), that understand the relation of being a reason for not in terms of nonnormative relations involving attitudes, but instead in terms of rational relations among attitudes. 9. Broome, “Normative Requirements,” and, following Broome, Brunero, “Are Intentions Reasons?” stress that, on some attitude-based theories, attitudes would be reasons for themselves. For example, if any belief is a reason to believe its logical consequences, then, because p is a logical consequence of p, any belief is a reason for itself. But Broome and Brunero never explain what the sting of this observation is supposed to be. Is it simply a way of dramatizing the problem of extensional incorrectness, the point being that the content of the belief, p, could be anything? Or is it supposed to have some independent force? Is the worry perhaps that if an attitude were a reason for itself, then it would be a reason that one could not go against, making deliberating about it, or offering it as advice, superfluous? But how troubling would this be? After all, some have explicitly argued that attitudes are self-justifying. For example, Gilbert Harman, Change in View (Cambridge, Mass.: MIT Press, 1986) argues, not implausibly, that any belief has some provisional justification simply in virtue of being held. 10. See, for example, Christine Korsgaard, The Sources of Normativity (Cambridge: Cambridge University Press, 1996). 11. Moreover, Mark Schroeder has suggested a pragmatic explanation of our reluctance to accept that the intention makes it the case that one has pro tanto reason. We expect claims about reasons to be made only when the reason is strong, and this reason is weak. See his “Instrumental Mythology,” Journal of Ethics and Social Philosophy, Symposium 1 (2005): 1–12; “The Scope of Instrumental Reason,” Philosophical Perspectives 18 (2004): 337–64; and Slaves of the Passions, ch. 5. 12. Parfit, On What Matters considers a hybrid desire-based theory, but of a different form. According to this theory, desires give one additional reason for action, but only when one has reason to have those desires. Although Parfit finds such a theory implausible, he is not particularly troubled by it, finding it “fundamentally value-based.” This hybrid theory is structurally similar to the “further attitude-constituted value” that I describe in Section 9 below. 13. Scanlon, “A Puzzling Duality?” 246, see also 233, 237, 239. 14. One might suggest: “The attitude-based form is strictly: If S has an attitude at t, then S has reason at t to do certain things. My attitude “creates” the reason in the sense that I did not have the reason prior to coming to have the attitude. Contrast this with a case in which, having had no friends, I make one, by coming to have the relevant attitude. Here the relevant principle is of the form: S has reason to help S’s friends, whoever they may be. In this case, my attitude does not “create” the reason. Even prior to having the attitude, I already had reason to help every member of the set of my friends. It’s just that the set was empty.” (As Raz writes, it is not problematic that adopting goals should change our reasons by “activat[ing] conditional reasons that we have anyway” (“The Myth of Instrumental


Reason, Value, and Desire

Rationality,” 23).) One problem with this suggestion is that it seems at least as natural to interpret the relevant principle as: If S has a friend at t, then S has reason at t to help him. And to say that, prior to making a friend, I did not already have reason to help my friends. Instead, it was true of me that, if I made a friend, I would then have reason to help him. 15. Raz, “Myth of Instrumental Rationality,” 3–4, and elsewhere, is explicit about the grounding of reasons in value. This is not necessarily to say that the concept of value is somehow explanatorily prior to the concept of providing reasons in this way. It is just to say that there is a way of providing reasons that is characteristically associated with what we are prepared to call values or things of value. So it is compatible with (although not committed to) Scanlon’s “buck-passing” proposal that “to call something valuable is to say that it has other properties that provide reasons for behaving in certain ways with regard to it” (What We Owe to Each Other, 96). 16. These activities of “respecting” and “engaging with” things of value are elucidated by Joseph Raz, Value, Respect, and Attachment (Cambridge: Cambridge University Press, 2001). For our purposes, the intuitive connotations of the terms are enough. 17. Contrast the way in which Scanlon, What We Owe to Each Other, ch. 3, says that we have reason to respect the value of human life, or Raz, Value, Respect, and Attachment, ch. 4, says that we have reason to respect people. 18. Christine Korsgaard, “Kant’s Formula of Humanity,” in Creating the Kingdom of Ends (Cambridge: Cambridge University Press, 1996), 106–32 at times suggests a shorter argument: (i) When the rational will takes something as its object, then we have reason to treat it in a certain way. So it is a source of reasons. (ii) Because the rational will is a source of reasons, it is itself something that we have reason to respect. This argument just equivocates on the two senses of “source of reasons” that we have been distinguishing. In order to get the conclusion that the rational will is itself something to be respected, the Kantian needs the claim that the rational will must take itself as one of its objects. 19. This might explain why it is so natural to label the opposition to the Desire-Based Theory the “value-based” view. See Dancy, Practical Reality, 29–31, who asks why the opposition to the desire-based view must involve any reference to value. 20. Note that even if we understand the Thesis in the earlier way, as the claim that no basic normative principles are of the attitude-based form, it still may be compatible with a theory extensionally equivalent to a hybrid or attitude-based theory. In theory, a normative principle of the attitude-based form might be derived from a normative principle not of that form. 21. See Dancy, Practical Reality; Ginsborg, “Reasons for Belief”; Parfit, On What Matters; Stroud, Engagement and Dissatisfaction, ch. 4; Thomson, Goodness and Advice. 22. Raz, Morality of Freedom, ch. 6, (and, following Raz, Dancy, Practical Reality, ch. 2; and Parfit, On What Matters) makes a similar point about desires. Usually, when one desires something, one has reason to desire it. And usually when one has reason to desire something, one also has reason to act to achieve it. So it may seem that desiring something makes it the case that one has reason to achieve it, when in fact it is having reason to desire it that does so. For further discussion of patterns of reasons, see Niko Kolodny, “How Does Coherence Matter?” Proceedings of the Aristotelian Society 107 (2007): 229–63.

Aims as Reasons


23. See Scanlon, Moral Dimensions, ch. 1. 24. Speaking of “the first-person standpoint of deliberation” and “the third-person standpoint of evaluation” helps, I think, to indicate the relevant distinction. But it may be a ladder that one kicks away after ascending. The more one thinks about the distinction, the less confident one is of any neat correspondence with these standpoints. After all, we can make claims about the reasons that third parties have, purely hypothetically or after the fact. And we can consider before the fact, indeed in deliberation, how we ourselves would properly be evaluated for a particular performance. Compare Scanlon, Moral Dimensions, ch. 1. I am indebted to Sam Scheffler for pressing this point. 25. See “Structural Irrationality,” in Common Minds: Themes from the Philosophy of Philip Pettit, ed. Geoffrey Brennan, Robert Goodin, Frank Jackson, and Michael Smith (Oxford: Oxford University Press, 2007), 84–103. Other modes of evaluation are especially relevant to the narrower thesis, which Scanlon explores in Moral Dimensions, that an agent’s present intentions, or reasons for which he acts, affect the moral permissibility of his so acting much less often than is commonly thought. Opponents of this narrower thesis, who insist that intentions do affect permissibility, need not embrace attitude-based, or hybrid, theories of the kind that the Thesis rejects. Opponents of the narrower thesis may well accept that acting with a good (or bad) intention realizes a kind of value. Scanlon’s reply, I take it, is that although acting with a good intention can be said to realize a value, this value need not have deliberative significance. It need not provide one with reason to realize it, or with any other reasons to act as one does. It may have, instead, only critical significance. It may provide only others with reasons to respond in certain ways to what one does. Less personally, it may give others reason, say, to admire one for displaying a virtue. More personally, it may give them reason to reaffirm, or alter, the attitudes and interactions that constitute their “relations” with one. This kind of value (or disvalue) constitutes what Scanlon calls “meaning.” Even though malice, recklessness, and culpable ignorance are all impermissible ways of harming another, they give different meanings to what is otherwise the same action. In particular, they make the action blameworthy in different ways. 26. For further discussion of ME, see Bratman, Intentions, Plans, and Practical Reason; Broome, “Practical Reasoning”; and Wallace, “Normativity, Commitment and Instrumental Reason.” 27. See also Niko Kolodny, “Why Be Rational?” Mind 114 (2005): 509–63; and Raz, “Myth of Instrumental Rationality.” In “Why Be Rational?” I try to explain why considerations about one’s irrationality can seem to play a role in deliberation and advice. 28. See also Brunero, “Are Intentions Reasons?” and Cullity, “Decisions, Reasons and Rationality,” 64–66. 29. Scanlon, “Reasons: A Puzzling Duality?” 236 suggests such a principle of rationality. But he does not suggest that IE1 is plausible only because it is confused with this principle. He agrees that this principle of rationality must be backed up by a truth about reasons, for much the same reasons as given in the text. 30. As the wording of Value-Provision suggests, the relevant relation to the thing of value need not be the purely facilitative relation of being more likely to do what it gives one reason to do. Rather, what it gives one reason to do in the first place may depend on one’s relations to it. For example, my friendship gives me reasons to do certain things that it does not give strangers. This is part of what is meant by saying that friendship gives rise to “agent-relative” reasons. (In this case, however, the thing of value cannot be neatly distinguished from my relation to it, because what constitutes my relation to my friendship also partly constitutes that friendship itself.)


Reason, Value, and Desire

31. Scanlon, “Reasons: A Puzzling Duality?” 239. See also Moral Dimensions, 93–94. Likewise, Raz, “The Myth of Instrumental Rationality,” 23 suggests that intentions to pursue certain goals can affect our reasons to take certain subsidiary actions, but not by changing the value of the goals themselves. “Crucially, the way goals acquire their normative relevance is by being conditions on the applicability or stringency of reasons. Therefore, they can have that effect only if the goals are worth pursuing in the first place.” 32. See, for example, Broome, “Are Intentions Reasons?”; Cullity, “Decisions, Reasons, and Rationality,” 64; Raz, “The Myth of Instrumental Rationality,” 22. 33. The point is sometimes put by saying that, if someone is disposed to reopen decisions, then he will achieve less of what he has reason to achieve over the long run. This may be true, but there is no need to appeal to dispositions, or to what will happen over the long run. If the agent reopens this very decision, he thereby incurs unnecessary cost in this very instance. 34. See “Reasons: A Puzzling Duality?” 240–46. The following line of thought may make it seem as though the costs of deliberation provide reasons only not to reconsider, not also to pursue what one intends. “Suppose that one reopens the decision and, having done so, revises one’s intention. Granted, one did something wrong in reopening the decision. But, having reopened it, one did nothing wrong in then revising one’s intention. This shows that although one had conclusive reason not to reopen the decision, one did not have conclusive reason to pursue the option one had decided on.” But what is actually shown is that, once one reopens the decision, one no longer has conclusive reason to pursue the option one had decided on. But this is precisely what the cost explanation predicts. By reopening the decision, one erases the cost advantage. Now, going forward, pursuing either option will require some deliberation and choice. This does not show that, before reopening the decision, one did not have conclusive reason to pursue the option one had decided on. 35. See “Toward a Principle of Instrumental Transmission” for discussion of why and when (because there are exceptions) the fact that an action requires some cost provides one with reason against that action. 36. Scanlon, “Reasons: A Puzzling Duality?” 241; Parfit, On What Matters. 37. However, as suggested in note 34, even in a case in which one does have something better to do, once one has reopened the decision, changing one’s mind may well not be against reason. 38. See Moral Dimensions, ch. 2. In “The Myth of Practical Consistency,” European Journal of Philosophy 16 (2008): 366–402, I suggest that predictive significance explains many of the phenomena that are taken to be evidence for a rational requirement of consistency in intention. If we take predictive significance seriously, in other words, we may have no need to posit such a requirement. 39. Brunero, “Are Intentions Reasons?” arrives independently at a similar explanation (although he suggests, arbitrarily in my view, that the effect is restricted only to necessary insufficient means). I have learned a great deal from his illuminating paper. 40. These are structurally similar to cases discussed by Frank Jackson, “On the Semantics and Logic of Obligation,” Mind 94 (1985): 177–95; and Frank Jackson and Robert Pargetter, “Oughts, Options, and Actualism,” Philosophical Review 95 (1986): 233–55. The claims in the text do not deny (taking a helpful example from Skip Schmall) that failing to intend (say) to save a drowning child is against reason if one has conclusive reason to intend it, and irrational if one believes that one has conclusive reason. It would also be dishonest to try to justify oneself by treating one’s lack of intention as an impediment beyond one’s control, or by saying, “Alas my diving in would be of no use. But I would help to save the child, if only there was some

Aims as Reasons


way that I could help.” For there is a way one could help: namely, by intending to save the child. But this does not affect the point: that I have more reason (not to intend to save the child and not to dive in) than (not to intend to save the child and to dive in). 41. See Bratman, Intention, Plans, and Practical Reason, 37–41. 42. I am particularly grateful to Jay Wallace for pressing me on this. 43. Brunero, “Are Intentions Reasons?” suggests this. 44. How exactly to understand and explain the difference is, of course, a longstanding problem. In this connection, Moran, Authority and Estrangement, suggestively emphasizes the fact that we adopt a deliberative stance toward our own intentions, assessing the reasons for and against them with an eye to forming, retaining, or revising them in light of these assessments. One wonders, though, whether this explains knowledge of our intentions in underdetermination cases. In such cases, it seems, one might have intended the other thing, without having assessed the reasons any differently. 45. Sarah Paul, “How We Know What We’re Doing” Philosophers’ Imprint 9 (2009): 1–23 gives several persuasive arguments for this conclusion and against the claims of G. E. M Anscombe, Intention (Oxford: Basil Blackwell, 1957); Kieran Setiya, Reasons without Rationalism (Princeton, N.J.: Princeton University Press, 2007); and J. David Velleman, Practical Reflection (Princeton, N.J.: Princeton University Press, 1989). 46. There is a kind of Sartrean “bad faith,” illuminated by Moran, Authority and Estrangement, ch. 3, that consists in a similar slide from treating one’s present intention, like a natural event or the mind of another, as an object of belief—a “facticity”—to treating it, like a natural event or the mind of another, as something not under one’s control—a mere facticity. Moran has no objection, at least not here, to treating our intentions as objects of belief, or evidence of what we will do (see 79). The problem lies in not treating them as things that we sustain or abandon by our own deliberation. I would add that although there is bad faith in treating one’s present intention as something not under one’s control, there is not the same bad faith in treating the influence of one’s present intention on one’s future behavior as something not under one’s present control—although it is less clear to me whether Moran would agree. What one now intends is under one’s present control, but what one will do in a week, say, is not under one’s present control, except via one’s present intentions (or self-manipulation). If one is convinced that one’s present intention will fail to influence one’s future behavior, then I think there is nothing amiss in judging that present intention futile and abandoning it. Of course, much may be amiss in the conviction itself, which may be the product of self-deception. And much may be amiss in the back-sliding that it anticipates, which will be akratic unless one changes one’s mind about whatever reasons underlie one’s present intention. 47. It might be replied that we accomplish this by directly conforming to rational requirements on our intentions, such as ME, in a way that bypasses expectations about what we will do. For example, when we intend the end, we come to intend what we believe are necessary means, without any reflection on how we are likely to behave. Among other problems with this suggestion, our deliberations often display finer-grained sensitivities, to our chances of success and to nonnecessary means, which the suggested mechanism cannot explain. 48. I take it that the point is not simply: “An intention to murder one’s wife would have to be akratic, or impulsive. So, if it remains under the agent’s control, there is little reason to expect that he will carry it out. Sooner or later he will come to his senses. Therefore, manipulation is unnecessary, unless it will not be under his control.” For one thing, this would not provide any contrast with the third-person case.


Reason, Value, and Desire

49. It may not even undermine the truth of the appeal to self-facilitation itself: the claim that buying the poison is wrong, precisely because the agent thereby facilitates his own murderous plans. See note 54. 50. See Scanlon, What We Owe to Each Other, ch. 5 for further explanation of “objections to permission” and “objections to prohibition.” 51. Even if we assume the agent has compelling reason to X, the objections to prohibition against the Control Principle do not seem significantly stronger. If X-ing leads to harm only through events that the agent will control, then the agent has a way to X without violating the Control Principle. He only needs to intend now to control future events in such a way that Xing will not lead to harm. Then X-ing no longer increases the risk of harm, and so does not violate the principle. 52. Scanlon actually writes that I “should abandon” this intention, but it is hard to see what else “should” could mean in this context. 53. See Scanlon, Moral Dimensions, ch. 2. 54. The objections to predictive significance in this and the previous section have been claims that an agent cannot take certain considerations as reasons. It is a further question whether, if these claims are true, this means that the considerations are not reasons. On the one hand, I have myself suggested that reasons are to be understood by reference to their role in deliberation. So there must be some connection between reasons and deliberation. On the other hand, it seems a mistake to construe this connection so tightly that any psychological impediment to taking them into account deprives them of standing as reasons. My inability to face a certain terrible fact, for example, should not, it seems, disqualify it as a reason. Likewise, the fact that it would be incoherent to think about the fact that I lack sufficient reason to continuing thinking about my reasons (“Stop dithering and act already!”) should not mean that there is no such fact. Perhaps the sort of inability alleged in the previous section—the alleged inability to form predictions about or on the basis of what one believes one controls—would have been sufficiently basic and structural to constrain what reasons we have. But the kind of inability that we have acknowledged in this section seems more derivative and local. So I am not persuaded that we should conclude that, because responding to the judgment that buying the rat poison is wrong would involve the sort of incoherence that we have discussed, buying the rat poison is not in fact wrong. 55. These expressions no doubt have different shades of meaning. For example, as Samuel Scheffler, “Valuing,” in Equality and Tradition (New York: Oxford University Press, 2010), 15–40 points out, valuing involves, whereas caring need not, believing that the thing valued is valuable. My topic is the area where they overlap. 56. See Elizabeth Anderson, Value in Ethics and Economics (Cambridge, Mass.: Harvard University Press, 1993), ch. 1; Niko Kolodny, “Love as Valuing a Relationship,” Philosophical Review 112 (2003): 135–89; and Scheffler, “Valuing.” Scheffler argues persuasively against other proposals: that caring, or rather the relevant component of valuing, is desiring, desiring to desire, believing valuable, or having a particular feeling. 57. Scanlon, What We Owe to Each Other, 49. See also Parfit, On What Matters. 58. Derek Parfit, Reasons and Persons (Oxford: Oxford University Press, 1984), 502, for example, suggests an account of well-being along these lines, in which it consists in the combination of “objectively” valuable activities and the right psychological orientation to them. “Pleasure with many other kinds of object has no value. And, if they are entirely devoid of pleasure, there is no value in knowledge, rational activity, love, or the awareness of beauty. What is of value, or is good for someone is to have both; to be engaged in these activities, and to be strongly wanting to be so engaged.” It is worth asking, though, why well-being should

Aims as Reasons


consist in caring about (or being otherwise psychologically attuned to) independently valuable activities. One explanation is, first, that well-being consists in engaging with things of value and, second, that caring about an independently valuable pursuit constitutes a further value. If this explanation is correct, then the present suggestion, that caring matters because of its effects on well-being, actually depends on the idea, discussed below, that caring constitutes a further value. 59. Scanlon, What We Owe to Each Other, ch. 3; and Moral Dimensions, 94–95; Raz, Engaging Reason (Oxford: Oxford University Press, 1999), ch. 13. The following quotations are from 316–17. 60. Ruth Chang, “Can Desires Provide Reasons for Action?” in Reasons and Value, ed. Wallace, Pettit, Scheffler, and Smith, 56–90, suggests something like this at 86. 61. See also Engaging Reason, 63–64. I here suppress, or downplay, language in Raz’s surrounding discussion that might suggest two ideas that we have questioned. The first is that a mere choice—the formation of a bare intention—creates a value of the relevant kind. (Recall the judge’s son.) The second is that the reasons really flow from the value of living a successful life and that the effect of caring is to determine what counts for one as a successful life. (Recall the dancer.) 62. Compare Susan Wolf, “Meaning in Life and Why it Matters,” Tanner Lectures on Human Values at Princeton University, November 2007, who suggests that such a value, fusing “objective” and “subjective” elements, is what makes for meaning in life. 63. I am implicitly assuming here that the caring-based reasons that the hybrid theory recognizes would simply be “added” to the value-based reasons that it recognizes. Caring and value would not interact in any more significant way to determine what reasons an agent has. This rules out, for example, the possibility that caring provides reason, but only when the thing cared about is independently valuable. Given the fundamental difference in kind between caring-based and value-based reasons, it is hard to see how they could interact in any more significant way. 64. See Raz, Morality of Freedom, 382. 65. See Niko Kolodny, “Which Relationships Justify Partiality? The Case of Parents and Children,” Philosophy & Public Affairs 38 (2010): 37–75; and “Which Relationships Justify Partiality? General Considerations and Problem Cases,” in Partiality and Impartiality: Morality, Special Relationships and the Wider World, ed. Brian Feltham and John Cottingham (Oxford: Oxford University Press, 2010), 169–193 for further discussion. If the history provides reasons for one’s present caring, then one might wonder whether one’s present caring actually contributes anything to the value. Perhaps the history suffices for the value, and then this value provides one with reasons both for present caring and for pursuing the aim. This may be the case with interpersonal relationships, such as one’s relationship to one’s child. The history gives one reason to care about the child, but the history would provide one with reasons for action even if one did not care about the child. But the kind of aims—causes, careers, callings—that we have been discussing seem different in this respect. If the dancer finds that he no longer cares about dancing, then that settles the matter. I am grateful to Sam Scheffler for raising this issue. 66. See R. Jay Wallace, “Caring, Reflexivity, and the Structure of Volition,” in Autonomes Handeln, ed. Monika Betzler and Barbara Guckes (Berlin: Akademie Verlag, 2000), 213–34; and “The Rightness of Acts and the Goodness of Lives,” in Reason and Value, ed. Wallace, Pettit, Scheffler, and Smith, 385–411. 67. Do the resources that we have canvassed—cost, effectiveness, the further attitudeconstituted value—explain Raz’s observation, noted in 31 above, that “goals acquire their


Reason, Value, and Desire

normative relevance by being conditions on the applicability or stringency of reasons”? I think so, although, of course, other explanations may also be possible. Raz’s main example is the following: “For someone intent on running a marathon every day during August, running a marathon today, the 20th of August, is crucial to the realization of his ambition. For me it is just an opportunity to know what running a marathon feels like—a matter of much less moment.” So described, the example may be complicated by factors independent of our different goals. If I have not run a marathon on each of the first nineteen days of August, then, whatever my goals are, my running a marathon today does not bring me any closer to achieving the end of running a marathon every day in August. If the August marathoner has run a marathon on each of the first nineteen days of August and remembers what running them felt like, then, whatever his goals, his running a marathon today does not bring him any closer to achieving the end, long since achieved, of acquiring knowledge of what running a marathon feels like. So, to split hairs, suppose that the date is August 1, neither of us knows what running a marathon feels like, and yet each of us is capable, and equally capable, of running a marathon every day during August. Why does the August marathoner have stronger reason to run a marathon today than I have? To begin with, one imagines that the August marathoner cares about this goal, and has been preparing it for some time. If so, then the attitude-constituted value that we have been discussing will be in play. But even if we set this aside, the effectiveness explanation seems to account for the difference. Given that I have no plans to run a marathon every day in August, there’s no chance that I will run a marathon every other day in August, even if I run one today. And if I will not run a marathon every other day in August, then running one today does not make it more likely that I run one every day in August. By contrast, given that the August marathoner does have plans to run a marathon every day in August, there is a chance that he will run a marathon every other day in August, if he runs one today. And if he will, then running a marathon today does make it more likely that he will run one every day in August. Because my running one today is not a means to my running one every day in August, whereas his running one today is a means to it, I lack the reasons that flow from realizing that value, whereas he has them. (My running one today is still a means to something of value, namely knowing what running a marathon feels like. But then so too is his running one today.) 68. For a recent revival of this challenge, see Chang, “Are Desires Reasons for Action?”

4 Scanlon on Desire and the Explanation of Action Michael Smith

in the first chapter of What We Owe To Each Other, T. M. Scanlon tells us that he has “become convinced that insofar as ‘having a desire’ is understood as a state that is distinct from ‘seeing something as a reason,’ it plays almost no role in the justification and explanation of action.” Because desire, understood more traditionally as a behavioral disposition, is such a state—an agent can see something as a reason without having any behavioral disposition, and he can have a behavioral disposition without seeing something as a reason—Scanlon commits himself to the conclusion that desire, understood in this more traditional way, “plays almost no role” in the explanation of action. My aim in what follows is to assess his reasons for thinking this is so. Since the standard story of action, defended by Hume and developed by the likes of Hempel and Davidson, tells us that desire, understood as a behavioral disposition, is part of the explanation of every action, I begin by outlining this standard story (§1). I then consider and evaluate what I take to be Scanlon’s official reasons for thinking that at least this part of the standard story is wrong (§2, §3). In the process of doing so, I say a little about what he has to say concerning the justificatory role of desire. I close by discussing some more positive claims Scanlon makes about a subclass of desires that he calls desires in the “directed-attention sense” (§4). If what he has to say about desires in the directed-attention sense is correct, then that may seem to give some additional support to the conclusion that desires, understood as behavioral dispositions, have little role to play in the explanation of action. However I argue that what Scanlon has to say about desires in the directed-attention sense gives no support to that conclusion. 79


Reason, Value, and Desire 1. desires, behavioral dispositions, and the explanation of action 

Suppose someone makes something happen. What makes it the case that, in making that thing happen, he acts, as opposed to merely being involved in something’s happening? Equivalently, when someone makes something happen, what makes it the case that he is an agent, as opposed to being a mere patient? The standard story of action tells us that the answer lies in the causal etiology of what happens. If the agent acts then, according to the standard story, we can trace what happens back to some bodily movement that the putative agent knows how to make happen, where his knowledge how to make that bodily movement happen is not itself explained by his knowledge how to do something else. In Arthur Danto’s terms, his moving his body in the relevant way has to be a “basic action” for him, something that he can just do. If these conditions are met, then we can establish whether the agent acts by seeing whether his bodily movement is caused and rationalized in the right kind of way by some desire he has that things be a certain way, and some meansend belief he has that some basic action of his, namely, his moving his body in the way we have identified, has a suitable chance of making things the way he desires them to be. If these further conditions are met then, when the agent makes the relevant thing happen, he acts; if not, he does not. Suppose, for example, that John flicks a switch. Is his flicking the switch an action? According to the standard story, we answer this question by first of all tracing back from the imagined action’s effects, the switch’s flicking, to one of John’s bodily movements. If John flicked the switch then he has to have done something that caused the switch to flick, and what he has to have done is to move his body in some relevant way. Let’s suppose that we trace back to a movement of his finger. If John’s flicking the switch is an action then this bodily movement has to be one that John knows how to make happen, and his knowledge how to make it happen must not be explained by his knowledge how to do something else. Moving his finger in the relevant way must be something that he could just do: a basic action. If it is not—if (say) the bodily movement is a ghastly contortion caused by a bomb blast that catapults John onto the switch—then we can conclude straightaway that his flicking the switch is not an action, but is rather a mere happening in which he is involved, albeit not as an agent. Now suppose that John’s making his finger move in the relevant way is something that he can just do. He moves his finger against the switch in the ordinary way in which we all move our fingers. In that case, the standard story tells us that whether or not John acts, when he flicks the switch, depends on the causal antecedents of that movement of his finger. Is that movement caused and rationalized by a desire John has that things be a certain way and a belief he has that his moving his finger in that way has some suitable chance of making things the way he desires them to be? Does he (say) desire to relieve an itch and believe that he can do so by moving his finger against the switch? Or does he desire to illuminate the room and believe that he can do this by moving his finger against the switch? If so, do his desire and belief cause his finger movement in the right way? If some such questions as these get an affirmative answer,

Scanlon on Desire and the Explanation of Action


then John’s flicking the switch was an act; if not, then we have to conclude that John was involved in a mere happening in which he was not an agent. Note that if we accept this standard story of action, then it follows immediately that not just desire, but means-end belief as well, is part of the explanation of every action. Whenever an agent acts he moves his body in some way, where that bodily movement has its causal origins in a desire-and-means-end-belief pair. But of course this prompts an obvious question. What is it about desire and belief that enable them to play this explanatory role? The answer to this question, according to the standard story, lies in the dispositional natures of these states. As Robert Stalnaker puts it: Belief and desire . . . are correlative dispositional states of a potentially rational agent. To desire that P is to be disposed to act in ways that would tend to bring it about that P in a world in which one’s beliefs, whatever they are, were true. To believe that P is to be disposed to act in ways that would tend to satisfy one’s desires, whatever they are, in a world in which P (together with one’s other beliefs) were true. An agent who acts is thus one in whom these two behavioral dispositions interact in a manner that makes it appropriate to describe him as having exercised his capacity to be instrumentally rational. In the standard story, this is what’s captured by the idea that desire and belief cause action in the right way. Given that belief and desire have these dispositional natures, the fact that there is a distinctive kind of behavior whose causal source lies in these very dispositions, a kind of behavior that constitutes a basic manifestation of agential control, becomes completely unmysterious. Agents who have these dispositions, and who act as a result, exercise their capacity to be instrumentally rational. Their behavior is differentially sensitive to what they desire and believe: they would have behaved ever so slightly differently if they had had ever so slightly different desires and beliefs from those they in fact have. Though much more sophisticated kinds of control are of course possible—for all that we have said, the means-end beliefs that are manifested in such behavior could be the result of brainwashing, and the desires could be utterly compulsive—it seems that these more sophisticated kinds of control would have to be built upon a foundation of agential control of this more basic kind. Every action must be the product of an exercise the agent’s capacity to be instrumentally rational. The standard story of action, when belief and desire are understood in these dispositional terms, thus looks to be all but analytic. Having said this, it is important not to suppose that what has been said is more ambitious than it really is. Means-end beliefs are more complex psychological states than mere dispositions to behave in certain ways, given our desires. Beliefs are acquired and sustained by our appreciation of evidence that bears on their truth, and they in turn provide an evidential basis for other beliefs that we either already have or go on to acquire. The inferential role of beliefs in sustaining and generating other beliefs is thus also a part of that state’s dispositional nature. When we said that beliefs are dispositions to behave in certain ways, given our desires, what we said was thus


Reason, Value, and Desire

only a part of the story, albeit the only part of the story that is absolutely crucial for belief to play the role it plays in controlling behavior in the basic way we have just described. If there were some psychological state distinct from belief that shared this aspect of belief’s dispositional nature, but which lacked its inferential role— imagining or fantasizing, perhaps—then it follows that imagining or fantasizing could also manifest itself in behavior, and there would therefore be a sense in which, when that happens, the person who engages in such imagining or fantasizing would be in control of what he thereby does. It is, I think, a great virtue of the standard story of action that it can so easily explain what the deep metaphysical similarities are between this kind of behavior, and cases of what it says constitute actions. The question “Should we call this sort of behavior ‘action’ too?” would seem to require something more akin to semantic decision than metaphysical reflection. Likewise, desires are much more complex psychological states than mere dispositions to behave in certain ways, given our means-end beliefs. For one thing, notwithstanding Stalnaker’s apparent suggestion to the contrary, certain desires do not have any obvious connection with behavioral dispositions. Desires that things be a certain way, where their being that way could not be the result of any intervention on our behalf, are like this. We simply have an affective orientation towards things being that way: we would experience a positive affect, were things thus—in short, we would like it.  For another, even when we restrict ourselves to desires that are plausibly constituted by behavioral dispositions, there seems to be more to such desires than just having those behavioral dispositions. People who are disposed to make things a certain way also generally like it when things are that way. Desires that manifest themselves in action thus ordinarily have an indirect connection to affect as well. Since there is a conceptual distinction between our being such that we would experience a positive affect if things were a certain way and our being such as to make them that way, we might well therefore ask whether desire, as we ordinarily understand it, is a complex state having these two dispositions as components, or whether there are two kinds of desire corresponding to motive and affect (“wanting” and “liking,” perhaps?), kinds that bear a certain sort of relation to each other. But however we answer this question, what’s crucial is that it is the fact that desire either is, or is inter alia, a behavioral disposition that is absolutely crucial for us to understand how desire can play the role it plays in the explanation of action. If we decide that desire is the name for the complex state, then whether or not we should give the name “action” to behavior that manifests dispositions to behave unaccompanied by a disposition to have some positive affect seems, once again, to be more a matter of semantic decision than metaphysical reflection. As before, it is a great virtue of the standard story of action that it can so easily explain the metaphysical similarities between this kind of behavior and cases of what we would more ordinarily call “action.” The fact that to act at all, an agent’s behavior has to be the causal upshot of desire and belief, where these are behavioral dispositions that may themselves be susceptible of further explanation, helps us understand how agents may come to exercise more sophisticated kinds of rational control over their behavior. For example, since rationality is on the side of an agent’s believing what is supported by his evidence, when an

Scanlon on Desire and the Explanation of Action


agent does what he does because of some belief he has, where this is understood as a behavioral disposition, but that disposition is (or is not) in turn the product of an exercise of his capacity to be so disposed in the light of the evidence available to him, we can cite his rationality (or his lack there-of) in explaining his action, and we can see him as manifesting (or failing to manifest) a corresponding kind of control. His actions may reveal his differential sensitivity (or his lack there-of) to the evidence available to him. Likewise, because rationality would seem to be on the side of an agent’s being disposed to make things ways he would like them to be, when an agent does what he does because of some desire he has, where this is understood as a behavioral disposition, but that disposition is (or is not) in turn the product of his capacity to be so disposed in the light of how he would like things to be, we can similarly cite his rationality (or his lack there-of) in explaining his action, and we can see him as manifesting (or failing to manifest) a corresponding kind of control. His actions may reveal his differential sensitivity (or his lack there-of) to his likes and dislikes. These are just a few examples; more could easily be given. To sum up, we have seen that the special role played by desire in the explanation of action, and means-end belief too, is underwritten by these states’ natures as complementary behavioral dispositions together with a conception of action as a basic manifestation of agential control: the differential sensitivity of what an agent does to what he desires and believes. The upshot is that if Scanlon is going to succeed in calling desire’s claim to play this special explanatory role into question, then he will need to provide us with some reason to question either the conception of desire as a behavioral disposition, or the conception of action as a basic manifestation of agential control. Let us therefore turn to consider and assess the reasons he provides. 2. scanlon’s objection

The first, and most important, move Scanlon makes in arguing against the claim that desire, understood as a behavioral disposition, has a crucial role to play in the explanation of action lies in his classification of desire as a “judgment-sensitive attitude.” Let’s therefore begin by examining what judgment-sensitive attitudes are and the role that the idea of such attitudes plays in his argument. According to Scanlon, judgment-sensitive attitudes are those “ .  .  . that an ideally rational person would come to have whenever that person judged there to be sufficient reasons for them, and that would, in an ideally rational person, ‘extinguish’ when that person judged them not to be supported by reasons of the appropriate kind.” The reasons to which judgment-sensitive attitudes are supposed to be sensitive are what he calls reasons in the “standard normative sense,” the paradigmatic examples of which are considerations that support the truth of our beliefs. In Scanlon’s view, desire is thus an attitude for which reasons, much like reasons for belief, can be given, and it is also an attitude that is responsive to beliefs about the reasons that there are to have it. Moreover, as he sees things, intention, hope, fear, admiration, respect, contempt, and indignation are also such attitudes.


Reason, Value, and Desire

The identification of the members of the class of judgment-sensitive attitudes is in turn important, according to Scanlon, because these attitudes “constitute the class of things for which reasons in the standard normative sense can be asked or offered.” As he notes, this commits him to the view that reasons for action, which are also reasons in the standard normative sense, reduce to reasons for some judgment-sensitive attitude or other. His preferred candidate is intention: “[R]eason for action” is not to be contrasted with “reason for intending.” The connection to action, which is essential to intentions, determines the kinds of reasons that are appropriate for them, but it is the connection with judgmentsensitive attitudes that makes events actions, and hence the kind of things for which reasons can sensibly be asked for and offered at all. Here, then, Scanlon provides us with his own preferred account of what makes an event an action: an event is an action in virtue of its connection with judgmentsensitive attitudes, specifically, with intentions. It might be thought that we are already in a position to see why Scanlon rejects the view that desire, understood as a behavioral disposition, has a crucial role to play in the explanation of action. For, it might be suggested, Scanlon evidently thinks that that role is reserved for intentions, where intentions differ in some important way from desires. But although there are theorists who think that intentions differ in an important way from desires—Michael Bratman and Richard Holton are notable examples—even they think that intentions share a crucial feature with desires, at least as desires are understood in the standard story, as in their view intentions are also behavioral dispositions. To the extent that Scanlon thinks that the state plays a crucial role in the explanation of most action is taking something to be a reason, where this is a different state from a behavioral disposition, he presumably thinks that this is so even when the behavioral disposition in question is intention: what really does the explaining, even when an agent acts on an intention, is the agent’s taking something to be a reason. In any event, Scanlon himself seems to side with those who think that intentions just are desires, in a suitably broad sense of the term. “Desire” is sometimes used in a broad sense in which the class of desires is taken to include any “pro-attitude” that an agent may have toward any action or outcome, whatever the content or basis of this attitude may be. Desires in this sense include such things as a sense of duty, loyalty, or pride, as well as an interest in pleasure or enjoyment. It is uncontroversial that desires in this broad sense are capable of moving us to act, and it is plausible to claim that they are the only things capable of this, since anything that moves us (at least to intentional action) is likely to count as such a desire. Putting these two passages together, Scanlon seems to be saying that events are actions in virtue of their connection with judgment-sensitive attitudes, specifically with

Scanlon on Desire and the Explanation of Action


intentions, and he also seems to be saying that the only states capable of moving us to act at all are desires, in a suitably broad sense of the term “desire,” from which I take it to follow that Scanlon would be happy to admit that intentions are a certain kind of desire. In what follows I will therefore just assume that his skepticism about the claim that desires, understood as behavioral dispositions, have a special role to play in the explanation of action is not grounded in his conviction that that role is reserved for intentions. So why does Scanlon reject the claim that desires, understood as behavioral dispositions, have a special role to play in the explanation of action? What is the relevance of the fact that desires are judgment-sensitive attitudes? As I understand it, the answer comes in the very next passage. [M]any elements of this class [that is, the class of desires in a suitably broad sense, whose members are judgment-sensitive attitudes] are what Nagel calls “motivated desires”; that is to say, they do not seem to be sources of motivation but rather the motivational consequences of something else, such as an agent’s recognition of something as a duty, or as supported by a reason of some other kind. A substantial thesis claiming a special role for desires in moving us to act would have to be based on some narrower class of desires, which can be claimed to serve as independent sources of motivation and perhaps also of reasons. Natural candidates for this role are what Nagel calls “unmotivated desires” (that is to say, desires that are not dependent on some other state for their motivating and reason-giving force). Scanlon’s objection to the claim that desires, understood as behavioral dispositions, play a crucial role in the explanation of action thus seems to be that, since desires are judgment-sensitive attitudes, it follows that when the desires on which we act are themselves produced by our beliefs about the reasons that there are to have them— that is, when they are motivated desires—then they are not themselves the explanation of what we do, but are rather part of what gets explained in our doing what we do by the thing that really does the explaining: namely, our beliefs about our reasons. A good question to ask at this point is whether the argument we gave earlier for thinking that desires, understood as behavioral dispositions, play a crucial role in the explanation of action, required us to take a stand on whether the desires that play that role are motivated or unmotivated. Somewhat surprisingly, the answer is much more difficult to provide than might initially be thought. In order to get these difficulties in focus, let’s ask whether Hume should agree that desire is a judgment-sensitive attitude. Or, to put the question somewhat differently, should Hume think that the desires that move us to act are motivated or unmotivated? The evidence is equivocal. Consider the following passage from the Treatise: A passion is an original existence, or, if you will, modification of existence, and contains not any representative quality, which renders it a copy of any other


Reason, Value, and Desire

existence or modification. When I am angry, I am actually possest with the passion, and in that emotion have no more a reference to any other object, than when I am thirsty, or sick, or more than five foot high. ’Tis impossible, therefore, that this passion can be oppos’d by, or be contradictory to truth and reason; since this contradiction consists in the disagreement of ideas, consider’d as copies, with those objections, which they represent. Here Hume seems to be saying that, because desire is not a state on which evidence bears—that is, because desires are not appropriately “consider’d as copies”—it follows that reason is simply silent on the issue of what to desire. Based on this passage, we might therefore think that Hume should deny that desires are judgment-sensitive attitudes. Desires are not judgment-sensitive attitudes because no sense can be made of the idea that there are reasons for or against having them. The desires on which we act must therefore be unmotivated. On the other hand, however, just a little bit later on Hume says this: I may desire any fruit as of an excellent relish; but whenever you convince me of my mistake, my longing ceases. I may will the performance of certain actions as means of obtaining any desir’d good; but as my willing of these actions is only secondary, and founded on the supposition, that they are causes of the propos’d effect; as soon as I discover the falsehood of that supposition, they must become indifferent to me. Here Hume seems to be saying that certain desires are sensitive to evidence. Consider my desire to procure fruit of a certain kind. As soon as I get evidence that that fruit is not an excellent relish, I lose my desire for it. Based on this passage, we might therefore conclude that Hume would agree that desires are judgment-sensitive attitudes and that the desires on which we act are therefore motivated. So should Hume agree that desires are judgment-sensitive attitudes, or should he disagree? Should he think that the desires on which we act are motivated or unmotivated? In order to answer these questions we need to make the all-important distinction between intrinsic desires, which are the “original existences” Hume speaks of in the first passage, and extrinsic desires, which are the “secondary” willings he talks of in the second passage. Extrinsic desires are amalgams of an intrinsic desire and a belief about how the world would need to be for the intrinsic desire to be satisfied. A desire to procure fruit of a certain kind is thus extrinsic, because it is an amalgam of some intrinsic desire—perhaps a desire to have relish with a certain taste—and a belief that having that particular piece of fruit would be a way of getting that sort of relish. Now suppose that I extrinsically desire a particular piece of fruit and then get evidence that the fruit does not have the taste that I intrinsically desire. If I am rational in the way I put my intrinsic desires together with my beliefs about how they might be satisfied, then my extrinsic desire for the piece of fruit will disappear. It will disappear because the belief component of the extrinsic desire will disappear. Extrinsic desires are thus judgment-sensitive attitudes, as Scanlon defines them, because, as amalgams of belief

Scanlon on Desire and the Explanation of Action


and desire, they are sensitive to reasons that bear on the truth of their belief component. They therefore count as “motivated desires,” in Nagel’s sense, because their motivating force depends on their being sustained by the beliefs that agents have about the reasons they have for believing what they believe about the way the world would need to be for their intrinsic desires to be satisfied. We can now see why Hume is committed to the conclusion that the secondary willings he talks about in the second of the two passages quoted above are judgmentsensitive attitudes. They are judgment-sensitive attitudes because secondary willings are extrinsic desires. It is, however, perfectly consistent with his being committed to this for him to be also committed to the conclusion that the intrinsic desires that partially comprise these extrinsic desires are not judgment-sensitive attitudes. This is the view of intrinsic desires to which Hume seems to commit himself in the first of the two passages quoted above. There he argues that, because those desires that he calls “original existences”—that is, because intrinsic desires—are not themselves beliefs, and because they have no beliefs as a proper part either, it follows that they are not the sort of psychological state that could be sensitive to evidence about the way the world is. Reasons do not bear on them. Hume thus commits himself to the conclusion that intrinsic desires are what Nagel would call “unmotivated” desires. They are unmotivated because their contribution to the motivating force of an extrinsic desire is not itself dependent on the motivating force of any belief. Let’s return to consider Scanlon’s objection to the claim that desires, understood as behavioral dispositions, have a substantial role to play in the explanation of action. To repeat, Scanlon’s objection is that, because desires are judgment-sensitive attitudes, it follows that when the desires on which we act are themselves produced by our beliefs about the reasons that there are to have them—that is, when they are motivated desires—it follows that these desires, considered as behavioral dispositions, are not themselves the explanation of what we do, but are rather part of what gets explained when we do what we do by the thing that really does the explaining: namely, our beliefs about our reasons. But if what we have just said is right, this objection is based on a total misunderstanding of the distinct, but complementary, roles played by intrinsic desire and belief in the explanation of action. In order to see that this is so, look at what happens when we map the distinction between intrinsic and extrinsic desires onto the standard story. According to the standard story, remember, desires and beliefs are both behavioral dispositions, and dispositions of both kinds are required for an agent to act. An agent must have the behavioral dispositions constitutive of his desiring that things be a certain way, and he must also have the behavioral dispositions constitutive of his believing that some basic action that he can perform will make things that way, and he must be instrumentally rational and put these two states together. We can restate all of this, but overlaying it with the distinction between intrinsic and extrinsic desires, as follows: the agent must both have the behavioral dispositions constitutive of his intrinsically desiring that things be a certain way, and he must have the behavioral dispositions constitutive of his believing that some basic action that he can perform will make things that way, and he must be instrumentally rational and put these two


Reason, Value, and Desire

states together so that he has the complex behavioral dispositions constitutive of his extrinsically desiring to perform the relevant basic action. When we restate the standard story in this way it turns out to be a mistake to think that we face stark alternatives: either agents act on motivated desires, or they act on unmotivated desires, but they do not act on both. For, according to the restated version of standard story, whenever an agent acts he acts on both a motivated desire (for he acts on the extrinsic desire that is the amalgam of his intrinsic desire and his belief about the basic action whose performance will make the world the way he intrinsically desires it to be) and he also acts on an unmotivated desire (for he acts on the intrinsic desire that is a proper part of that amalgam). It is therefore true that, whenever an agent acts, there is a desire on which he acts that is produced by his beliefs about the reasons that there are to have that desire. The belief component of his extrinsic desire is, after all, produced by whatever the agent takes to be the reasons that there are for believing that the basic action in question will make the world the way he intrinsically desires it to be. But it does not follow from this that his action can be fully explained by his taking those considerations to be reasons. A distinct role still has to be played by the intrinsic desire with which the means-end belief that is supported by the considerations that he takes to be reasons must combine. Let’s recapitulate. Scanlon claims that once we see not just that desires are judgment-sensitive attitudes, but that they are sustained by beliefs about the reasons that there are to have them, we see that desires, where these are understood as behavioral dispositions, play no crucial role in the explanation of action. Instead they are just a part of what gets explained by what really does the explaining when we act, namely, our beliefs about our reasons. As we have just seen, however, this objection fails. For although the extrinsic desires on which we act are both judgment-sensitive attitudes and motivated desires, they are themselves just amalgams of intrinsic desires and means-end beliefs, where these are both conceived of as behavioral dispositions playing distinct but complementary roles, and where, for all that has been said, the intrinsic desires that are components of our extrinsic desires are neither judgment-sensitive attitudes nor motivated desires. To put the point somewhat paradoxically, the fact that we always act on (extrinsic) desires that are both judgment-sensitive attitudes and motivated is perfectly consistent with our always acting on (intrinsic) desires that are neither judgment-sensitive attitudes nor motivated.

3. a revamped version of scanlon’s objection

Might Scanlon rerun his objection wholly in terms of the judgment-sensitivity of intrinsic desires, where these are understood to be behavioral dispositions? In other words, might he suggest that the fact that the intrinsic desires on which we act, understood as behavioral dispositions, are themselves produced by our beliefs about the reasons that there are to have them, entails that they too are motivated desires, and hence that they are no part of the explanation of what we do either? Instead, might he suggest, our intrinsic desires, so understood, are just another part of what gets

Scanlon on Desire and the Explanation of Action


explained in our doing what we do by the thing that really does the explaining, namely, our beliefs about our reasons for having the intrinsic desires that we have? Before considering the merits of this revamped version of Scanlon’s objection, it is worthwhile pausing for a moment to highlight just where, if this is what Scanlon’s objection really amounts to, it turns out that he disagrees with Hume. On this way of interpreting him, Scanlon disagrees with Hume about whether specifically intrinsic desires are judgment-sensitive attitudes: that is, states that a rational agent will have or not have depending on what he takes to be the reasons that there are for having them. Hume says that they are not; Scanlon says that they are. However it is important to remember that Hume gave an argument for his view that intrinsic desires are not the sort of psychological state for which there can be reasons. As he sees things, remember, for considerations to constitute reasons for having attitudes with a certain content, those attitudes would have to be appropriately “consider’d as copies.” Or, to put the point slightly differently, the considerations in question would have to constitute evidence for the truth of the proposition that expresses the content of those attitudes. The only psychological state for which there can be a reason is therefore either a belief pure and simple, or a complex psychological state, like an extrinsic desire, that has belief as a component. The question Scanlon must answer is where this argument goes wrong. As I understand it, Scanlon thinks that this argument has a false premise. He thinks that it is not true that all reasons are evidence. Scanlon tells us, for example, that “ . . . [a]ny attempt to explain what it is to be a reason for something seems to me to lead back to the same idea: a consideration that counts in favor of it. ‘Counts in favor how?’ one might ask. ‘By providing a reason for it’ seems to be the only answer.” In Scanlon’s view, the concept of a reason is therefore coeval with the idea of a consideration that counts in favor, and this pair in turn is primitive, permitting no further explication. There are therefore reasons for belief—that is, considerations that count in favor of believing—and, because what these considerations count in favor of is believing, these reasons are indeed evidence. But there are also considerations that count in favor of intrinsically desiring, where these are reasons in the very same sense of “reason” in which reasons are reasons for belief—both are considerations that count in favor—but, because the attitude that these considerations count in favor of is not believing, these considerations, which are reasons, are not evidence. Is Hume or Scanlon right about the nature of reasons? In favor of Hume’s view, note that we can say more than just that there are considerations that count in favor of believing. We can also say how it is that those considerations count in favor. As I understand it, this is what we try to do when we give substantive theories of confirmation and statistical reasoning. We attempt to say how it is that certain evidence bears upon the degree to which you should believe something. A good question to ask is whether we can say anything remotely similar about how the way in which the considerations that count in favor of certain intrinsic desires count in favor of them. If not, then Hume seems to have the better of the argument. To suppose that there are reasons for intrinsically desiring, much like reasons for believing, would be to admit that there is a glaring disanalogy between reasons for believing and reasons for


Reason, Value, and Desire

intrinsically desiring, a disanalogy so glaring as to cast doubt on there being reasons for intrinsically desiring. Derek Parfit, who also thinks that intrinsic desires are judgment-sensitive attitudes, thinks that we can at least begin to say something completely general about reasons for intrinsically desiring. He thinks that it is the fact that certain things have the intrinsic natures that they do that counts in favor of intrinsically desiring those things. For example, if there is a reason for intrinsically desiring enjoyment, then that reason is constituted by the fact that enjoyment has the intrinsic nature that it has. Or if there is a reason to intrinsically desire that we are able to justify our conduct to others, then that reason is constituted by the fact that being able to justify our conduct to others has the intrinsic nature that it has. But plausible though this may seem to be at first blush, it is important to realize that the suggestion does not help all that much. For the question now just becomes whether we can say how it is that the fact that certain things have the intrinsic natures that they do makes that fact about them count in favor of intrinsically desiring things with that intrinsic nature to a certain extent, whereas the fact that other things have the intrinsic natures that they don’t similarly count in favor of intrinsically desiring them to that same extent. Absent an answer to this question, the claim that there are reasons to intrinsically desire certain things and not others simply won’t be credible. This bears on an issue about which I wish to remain officially neutral here: namely, Scanlon’s suggestion that in addition to desiring, understood as a behavioral disposition, playing little or no role in the explanation of action, it plays almost no role in the justification of action either. For if Hume turns out to be right, and we can make no sense of there being reasons to intrinsically desire certain things rather than others, then it will turn out that intrinsic desires do indeed play a crucial role in the justification of action. For the only apt question to ask about reasons for action will be whether or not there are reasons for an agent to extrinsically desire to act in certain ways, where all such reasons will be conditional on the presence of some relevant intrinsic desire that the agent has, where this is understood as a behavioral disposition, and where acting in the way in question will satisfy that intrinsic desire. (Or, somewhat more accurately given that rationality seems to be on the side of an agent’s being disposed to behave in ways that make things turn out as he would like them to be, all such reasons will be conditional on the presence of some relevant intrinsic desire, where this is understood as a behavioral disposition that in turn squares with the agent’s dispositions to like it when things turn out to be certain ways.) This will be the only apt question to ask because all such reasons will be reasons to believe that some basic action will satisfy some such intrinsic desire. The focus of the discussion thus far has been on the coherence of the idea that there are reasons for intrinsically desiring to act in certain ways. Let us now put such doubts as we might have about the existence of such reasons to one side, and return to consider the revamped version of Scanlon’s objection, the version that takes that idea for granted. The revamped version of Scanlon’s objection is that when the intrinsic desires on which we  act, where these are understood as behavioral dispositions, are themselves produced by our beliefs about the reasons that there are to have them, these

Scanlon on Desire and the Explanation of Action


motivated intrinsic desires are not part of the explanation of what we do, but are rather just another part of what gets explained in our doing what we do by the thing that really does the explaining: namely, by our beliefs about the reasons that there are for having the intrinsic desires that we have. How convincing is this objection? Imagine two variations on the example of John given earlier. John+ is maximally reasonable. He only has intrinsic desires for things when he believes that there are reasons for having those intrinsic desires. His unreasonable twin brother John—, though he has the same intrinsic desires as John+, has intrinsic desires that are not based on his beliefs about reasons. In the terms outlined at the beginning, this means that both John+ and John—have the very same dispositions to behave in certain ways. In the possible worlds in which their means-end beliefs are the same, they therefore do exactly the very same things. Now let us suppose that both John+ and John—have an intrinsic desire to illuminate the room and a belief that they can do so by flicking the switch—that is to say, they have not just the same intrinsic-desire-dispositions, but also the same means-end-belief-dispositions—and let us suppose further that they both act accordingly. The only difference between them lies in the fact that John+’s action is the upshot of a motivated intrinsic desire, whereas John—’s is not: in other words, John+’s disposition to illuminate the room is the causal upshot of a belief, whereas John—’s is not. Does John—’s intrinsic desire play a role in the explanation of his action that John+’s does not play in his? That seems quite incredible on the face of it. Of course, what is true is that John+ would not have done what he did had he not had the beliefs he has about the reasons that there are for having his intrinsic desire to illuminate the room. But this does not show that John+’s intrinsic desire doesn’t play the same role in explaining what he does that John—’s plays in explaining his, because when we imagine John+’s having different beliefs about the reasons that there are for having intrinsic desires, we thereby imagine him having different intrinsic desires as well. That’s a consequence of our counterfactualizing under the supposition that John+ is maximally reasonable. So how might we determine whether their intrinsic desires do or do not play the same crucial explanatory role? Note that because John+’s intrinsic desires, understood as behavioral dispositions, are distinct psychological states from his beliefs about his reasons, we can ask whether, if John+ had had the very intrinsic desires that he has, he would have done exactly what he did, whether or not his intrinsic desires had been the product of his beliefs about the reasons that there are for having them. And the answer to this question is clear. He would have done exactly what he did. This is what is shown by the behavior of John+’s unreasonable twin brother John—. Nor is this surprising. For in each case their behavior simply tracks their behavioral dispositions. What else would we expect? But if John+ would have done exactly what he did so long as he had the intrinsic desires that he has, where these are understood as behavioral dispositions, then it would seem to follow that his having the intrinsic desire that he has to illuminate the room, so understood, must play a crucial role in explaining his doing what he in fact did, the very same role that John—’s intrinsic desire plays in explaining his doing what he did.


Reason, Value, and Desire

So where does the revamped version of Scanlon’s objection go wrong? It goes wrong in illicitly moving from the premise that the ultimate explanation of an agent’s action, when he acts on an intrinsic desire that is in turn explained by his beliefs about the reasons that there are for having that intrinsic desire, is not the intrinsic desire, but is rather the belief about the reason, to the conclusion that that intrinsic desire plays no role in the explanation of his action. This move is illicit because, as we have just seen, even when the ultimate explanation is the belief about reasons, a crucial role still needs to be played by the agent’s intrinsic desire. John+’s belief about the reasons that there are for intrinsically desiring to illuminate the room itself explains his illuminating the room only because it gives rise to that intrinsic desire, where the intrinsic desire is a behavioral disposition, as it is that disposition that gives rise to John+’s action. This is why John+ would have done exactly what he did, so long as he had the intrinsic desires that he has, even if his intrinsic desires had not been the product of his beliefs about the reasons he has for having those intrinsic desires.

4. scanlon on desire “in the directed-attention sense”

Later on in the first chapter of What We Owe to Each Other, Scanlon characterizes a subclass of desires that he calls desires “in the directed-attention sense.” Because these desires are also supposed to be capable of explaining action, and yet are not behavioral dispositions, their existence might also seem to create problems for the idea that desires, understood as behavioral dispositions, play a crucial role in the explanation of actions. Let us therefore consider what Scanlon has to say about desires in the directed-attention sense to see whether they really do pose such a problem. Scanlon works towards his account of desire in the directed-attention sense via a discussion of Nagel’s example of thirst. Suppose I am thirsty. What does this involve? . . . In addition to the dryness in my throat, the future pleasure brought about by drinking, and my judgment that this pleasure is desirable, there is the fact that I feel an urge to drink. But when we focus on this idea of a mere urge to act, separated from any evaluative element, it does not in fact fit very well with what we ordinarily mean by desire. Here we may consider Warren Quinn’s example of a man who feels an urge to turn on every radio he sees. It is not that he sees anything good about radios’ being turned on; he does not want to hear music or news or even just to avoid silence; he simply is moved to turn on any radio that he sees to be off. Quinn’s point is that such a functional state lacks the power to rationalize actions. . . . But as he also points out, although we may sometimes have such urges, the idea of such a purely functional state fails to capture something essential in the most common cases of desire: desiring something involves having a tendency to see something good or desirable about it. This is clear from the example of thirst. Having a desire to drink is not merely feeling impelled to do so; it also involves seeing drinking as desirable (because, for example, it would be pleasant). The

Scanlon on Desire and the Explanation of Action


example of the urge to turn on radios is bizarre because it completely lacks this evaluative element. I might seem to be saying here that there is no such thing as an unmotivated desire. Taken in Nagel’s sense this would entail that all desires arise from prior evaluative judgments of some kind, a claim which seems clearly false. What I am claiming, however, is not that all desires arise from prior judgments but rather that having what is generally called a desire involves having a tendency to see something as a reason. In this passage Scanlon rightly points out that the idea of mere dispositions to behave in various ways, in the light of one’s beliefs, “fails to capture something essential in the most common cases of desire.” He thinks that what we need to add, in order to capture the most common cases, is “having a tendency to see something as a reason.” But is that really what we need to add? As was pointed out above (§1), the psychological state that we ordinarily call “desire” comprises not just behavioral dispositions, but also a disposition to be affectively oriented towards the world’s being the way we are disposed to make it. We would normally like it if the world were that way. This state must not be confused with the thinking that it would be good if the world were that way, or with having a tendency to see reasons, because there is plainly a conceptual distinction between being disposed to like it if things were a certain way and thinking it would be good if they were that way, or with having any tendency to see any reasons for anything: we may be disposed to like it if things were a way that we believe is bad, or that we have no tendency to think that any reason could justify their being. Moreover, psychologists posit a psychological subsystem corresponding to the disposition to be affectively oriented towards the world’s being a certain way, a subsystem distinct from that corresponding to the disposition to make it that way, in order to explain the behavior of not just mature human beings (who might well have a tendency to see things as reasons), but also the behavior of rats, monkeys, and infant humans (who are plainly incapable of having such a tendency). Even conceptually unsophisticated agents, agents who have no concept of a reason, normally adjust their degree of motivation to accord with the degree to which they like it when the world turns out the way they are disposed to make it. What all of this suggests, to me at any rate, is that what is most obviously missing from someone like Quinn’s radioman, who just has behavioral dispositions, is this affective orientation towards the outcome of his turning on radios. What’s missing is his being such that he would like the world to be the way he is disposed to make it, an affective orientation towards the objects of his behavioral dispositions that he could share with rats and monkeys and infant humans. Perhaps Scanlon would agree with this. But he might then insist that there is an additional difference between desire in conceptually sophisticated human beings and desire in rats and monkeys and infant human beings, and that this is what he is trying to capture with his idea of desire in the directed-attention sense. This too is missing from Quinn’s radioman. In ordinary cases in which conceptually sophisticated human beings have desires, he might say, they do not just have behavioral dispositions together with dispositions to have an


Reason, Value, and Desire

affective orientation towards things being a certain way, but they also have an additional psychological state, the state of being disposed to see something to be a reason. As he puts it, when he finally provides us with his full-dress proposal: I am claiming . . . that having what is generally called a desire involves having a tendency to see something as a reason. . . . Even if this is true, however, this is not all that desire involves. Having a desire to do something (such as to drink a glass of water) is not just a matter of seeing something good about it. I might see something good about drinking a glass of foul-tasting medicine, but would not therefore be said to have a desire to do so, and I can even see that something would be pleasant without, in the normal sense, feeling a desire to do it. Reflection on the differences between these cases leads me to what I will call the idea of desire in the directed-attention sense. A person has a desire in the directedattention sense that P if the thought that P keeps occurring to him or her in a favorable light, that is to say, if the person’s attention is directed insistently toward considerations that present themselves as counting in favor of P. The trouble with this full-dress proposal, however, is that it makes no sense in Scanlon’s own terms. To repeat, Scanlon tells us that the only things for which reasons can sensibly be asked or given are judgment-sensitive attitudes. Here, however, he tells us that desire in the directed-attention sense is a matter of having a tendency to think that there are reasons for propositions. Presumably he is speaking loosely when he says this, so the question we have to ask ourselves is which judgment-sensitive attitude he has in mind. Since when p is a consideration that counts in favor of q, that sounds like the claim that p supports the truth of q, the answer might appear to be belief. But of course, this cannot be what Scanlon intends. He surely does not think that having a desire in the directed-attention sense that p is a matter of having one’s attention directed insistently to considerations that support the truth of the claim that p. What he intends is rather that having a desire in the directed-attention sense that p is a matter of having one’s attention directed insistently to considerations that favor. . . . favor what? Desiring that p? This would make a desire in the directedattention sense that p a strange sort of self-referential attitude, an attitude the like of which I have to confess I can make little sense. Intending to bring p about? That would make it impossible to have a desire in the directed-attention sense that p when you believe that there is no way of bringing p about. The answer Scanlon should give, of course, is that a desire in the directed-attention sense that p is simply a tendency to have your attention drawn towards considerations that present themselves as being either reasons to be disposed to have some positive affect, or to like it, when p (this is why it is possible to have such a desire while believing that there is no way to bring p about) or reasons to be so disposed behaviorally to bring p about (this is why what Scanlon calls desire in the directed-attention sense is not a strange sort of self-referential attitude: the tendency to see reasons to be disposed to bring p about is a different psychological state from the disposition to bring

Scanlon on Desire and the Explanation of Action


p about). The trouble, however, is that as soon as we understand desires in the directedattention sense in these terms it becomes clear that they too will succeed in explaining actions only if they first of all bring about desires understood as behavioral dispositions. The tendency will play its role in explaining action only if the behavioral disposition plays its. The upshot is that even if we agree with Scanlon that there is a distinctive class of desires in the directed-attention sense, as he characterizes them, their existence creates no problem at all for the idea that desires, understood as behavioral dispositions, play a crucial role when it comes to the explanation of action. conclusion

As I said at the beginning, Scanlon tells us that he has “become convinced that insofar as ‘having a desire’ is understood as a state that is distinct from ‘seeing something as a reason,’ it plays almost no role in the justification and explanation of action.” But as we have seen, even when psychological states distinct from desires, so understood, do play a role in explaining actions, they either do so by combining with such desires— this is how means-end beliefs get to play a role in explaining actions when they combine with intrinsic desires—or they do so by way of explaining such intrinsic desires or means-end beliefs. In both cases, desires understood as behavioral dispositions, and means-end beliefs so understood too for that matter, play a crucial explanatory role in the production of action. Nor should this be surprising. For, to repeat the message conveyed earlier on, what makes certain bodily movements actions is the fact that they are caused in the right way by intrinsic desires and means-end beliefs, where both are understood as behavioral dispositions. notes 1. T. M. Scanlon, What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998), 18. 2. See, for example, David Hume, A Treatise of Human Nature (Oxford: Clarendon Press, 1740/1968); Carl G. Hempel, “Rational Action” in Readings in the Theory of Action, ed. Norman S. Care and Charles Landesman (Bloomington: Indiana University Press, 1961), 281–305; Donald Davidson, “Actions, Reasons, and Causes” in his Essays on Actions and Events (Oxford: Oxford University Press, 1980), 3–20. 3. Scanlon What We Owe, 39. 4. The first four paragraphs in this section draw on my “The Standard Story of Action: An Exchange,” in Causing Human Action: New Perspectives on the Causal Theory of Action, ed. Jesús H. Aguilar and Andrei A. Buckareff (Cambridge, Mass.: MIT Press, 2010). 5. See Arthur C. Danto, “What We Can Do,” Journal of Philosophy (60) (1963): 434–45. 6. See my “The Humean Theory of Motivation,” Mind (96) (1987): 36–61. 7. Robert Stalnaker, Inquiry (Cambridge, Mass.: MIT Press, 1984), 15. 8. For an attempt to explain what it is for an agent to exercise his capacity to be rational in a way that is consistent with the standard story see my “Rational Capacities,” Weakness of Will and Varieties of Practical Irrationality, ed. Sarah Stroud and Christine Tappolet (Oxford: Oxford University Press, 2003), 17–38.


Reason, Value, and Desire

9. For more on this see my “The Explanatory Role of Being Rational,” in Reasons for Action, ed. David Sobel and Steven Wall (New York: Cambridge University Press, 2009), 58–80. 10. For more on the idea that action requires the differential explanation of behavior by desire and means-end belief see Christopher Peacocke, Holistic Explanation (Oxford: Oxford University Press, 1979). 11. See also Galen Strawson’s discussion of the Weather Watchers in chapter 9 of his Mental Reality (Cambridge, Mass.: MIT Press, 1994). In my “Galen Strawson and the Weather Watchers,” Philosophy and Phenomenological Research (58) (1998): 449–54, I attempt to explain how even these desires might be behavioral dispositions, albeit where the connection that they have with behavior may be very nonobvious. My current view is supposed to be agnostic on whether such desires have even nonobvious connections with behavior. 12. This seems to be the view of the psychologist Kent Berridge, a view endorsed and built upon in the philosophical literature by Richard Holton. See Kent C. Berridge, “Wanting and Liking: Observations from the Neuroscience and Psychology Laboratory,” Inquiry (52) (2009): 378–98; Richard Holton, Willing, Wanting, Waiting (Oxford: Oxford University Press, 2009). For earlier presentations of these ideas see Lloyd Humberstone, “Wanting, Getting, Having,” Philosophical Papers (19) (1990): 99–118; and Peter Railton, “Kant Meets Aristotle Where Reason Meets Appetite”, in Argument & Analyse, Proceedings of the 2000 Congress for the Gesellschaft fur Analytische Philosophie, ed. C.U. Moulines and K.-G. Niebergall (Paderborn: Mentis, 2002): 275–293. 13. See the later discussion of Warren Quinn’s radioman (§4), a character Quinn discusses in his “Putting Rationality in Its Place” in his Morality and Action (Cambridge: Cambridge University Press, 1993). 14. For more on this see Philip Pettit and Michael Smith, “Freedom in Belief and Desire,” Journal of Philosophy (93) (1996): 429–49. 15. Scanlon, What We Owe, 50–55, 363–73. 16. Ibid., 20. 17. Ibid., 20–21. 18. Ibid., 20. 19. Ibid., 21. 20. Ibid. 21. See Michael Bratman, Intentions, Plans and Practical Reason (Cambridge, Mass.: Harvard University Press, 1987); Holton, Willing, Wanting, Waiting. 22. Scanlon, What We Owe, 37. 23. Ibid. See also Thomas Nagel, The Possibility of Altruism (Princeton, N.J.: Princeton University Press, 1970). 24. Hume, Treatise, 415. 25. Ibid., 417. 26. See my “Instrumental Desires, Instrumental Rationality,” Supplement to the Proceedings of the Aristotelian Society (78) (2004): 93–109. 27. Scanlon, What We Owe, 17 28. Judith Jarvis Thomson, Normativity (Chicago: Open Court Publishing, 2008), 130. 29. Derek Parfit, On What Matters. 30. For my own preferred way of making sense of the idea of there being reasons for intrinsically desiring certain things to a certain extent see my “Beyond the Error Theory,” A World Without Values: Essays on John Mackie’s Moral Error Theory, ed. Richard Joyce and Simon Kirchin (New York: Springer, 2010), 119–39.

Scanlon on Desire and the Explanation of Action


31. In the terms in which Scanlon himself puts it in his Appendix on Bernard Williams’s distinction between internal and external reasons, this would be to suppose that all reasons have “subjective conditions” (Scanlon, What We Owe, 367). 32. This point is further developed in my “The Possibility of Philosophy of Action,” in Human Action, Deliberation and Causation, ed. J. Bransen and S. Cuypers (Dordrecht, The Netherlands: Kluwer Academic, 1998), 17–41. 33. Scanlon, What We Owe, 39. 34. Ibid., 38–39. See also Quinn’s “Putting Rationality in its Place” and Nagel’s The Possibility of Altruism. 35. Berridge, “Wanting and Liking.” 36. Scanlon, What We Owe, 39. 37. See again my “The Possibility of Philosophy of Action.” 38. Scanlon, What We Owe, 18. 39. I would like to thank Ruth Chang, Sebastian Muders, and R. Jay Wallace for helpful comments on earlier versions of this essay.

This page intentionally left blank

part two Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

This page intentionally left blank

5 Of Metaethics and Motivation the appeal of contractualism

Pamela Hieronymi

in 1982, when T. M. Scanlon published “Contractualism and Utilitarianism,” he noted that, despite the widespread attention to Rawls’s A Theory of Justice, the appeal of contractualism as a moral theory had been under appreciated. In particular, the appeal of contractualism’s account of what he then called “moral motivation” had been under appreciated. It seems to me that, in the intervening quarter century, despite the widespread discussion of Scanlon’s work, the appeal of contractualism, in precisely this regard, has still been under appreciated—even though Scanlon makes what he once called “moral motivation” central, throughout. Perhaps others simply do not find attractive what Scanlon and I find attractive. This sort of flat disagreement is possible. However, I suspect it is not widespread. I suspect, rather, that the appeal of Scanlon’s contractualism has been somewhat obscured both by the precision by which he stated it and by surrounding, ancillary, issues. I thus suspect the disagreement less flat and more interesting than often recognized. My aim, here, is to do my best to draw out and make vivid the appeal of contractualism. I will then examine the state of the disagreement. I will begin by considering two questions Scanlon thinks must be addressed by any moral theory, which he once called “the question of subject matter” and “the question of motivation.” After introducing these questions, I will spend time on the second, because it is contractualism’s answer to this second question that, Scanlon believes, provides contractualism with its appeal. I will then return to the first question, of subject matter—which will, by that point, have been revealed as not really distinct from the question of motivation, as Scanlon understands it. However, it is as an answer to this first question that Scanlon’s theory is most often criticized. I will examine a few of 101


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

the most popular criticisms and try to display why, once we have understood Scanlon’s project, they do not find their target. Because I suspect that even those who disagree with Scanlon’s account of the subject matter will find attractive his answer to the “question of motivation,” I will close by asking whether it is possible to wed Scanlon’s attractive answer to the question of motivation to an alternative answer to the question of subject matter—in other words, to an alternative theory of morality. Not without both difficulty and sacrifice, will be my answer. In particular, not without sacrificing the central place given, in Scanlon’s account, to what might be called freedom of conscience.

locating scanlon’s questions

In his early article Scanlon sets out two questions that, he thinks, must be addressed by any moral theory: the first he calls “the question of subject matter,” the second “the question of motivation.” The question of subject matter is familiar from standard discussions in metaethics. He introduces it, saying There is such a subject as moral philosophy for much the same reason that there is such a subject as the philosophy of mathematics. In moral judgments, as in mathematical ones, we have a set of putatively objective beliefs in which we are inclined to invest a certain degree of confidence and importance. Yet on reflection it is not at all obvious what, if anything, these judgments can be about, in virtue of which some can be said to be correct or defensible and others not. This question of subject matter, or the grounds of truth, is the first philosophical question in both morality and mathematics. I believe it would be a mistake to try to find a precise formulation of this question. Rather, Scanlon here gestures, broadly, at a general and recognizable area of inquiry, which he labels “the question of subject matter.” To expand: almost everyone believes, pretheoretically, that murder and cruelty are wrong and that you ought not to gain through the deceit or exploitation of others. But if asked what grounds these judgments—if asked why murder and cruelty are wrong, or why you ought not to gain through the deceit or exploitation of others—people are often at a loss for an answer, and what answers they do provide vary wildly. Some will appeal to the commands of God, some to the badness of pain, some to the dignity of persons or the excellences of the soul, some to the importance of autonomy, or to what they would want if in the other person’s shoes. Others think these pretheoretical judgments simply express basic moral facts, and that further explanation is both unnecessary and impossible. So, even though, pretheoretically, we make certain judgments that we think both correct and important, it is remarkably unclear, upon reflection, what, if anything, makes them correct. It is also not clear exactly what holds them together, in the class

Of Metaethics and Motivation: The Appeal of Contractualism


of “moral” judgments. And yet, at least pretheoretically, we think certain moral judgments are correct and others are mistaken. Scanlon’s first question, of subject matter, asks what these judgments are about, such that some of them are correct and others incorrect. It is obvious enough why one might think this area of inquiry represents “the first philosophical question in both morality and mathematics.” We will return to it in the next section. For now, our focus will be on the question of motivation. Scanlon’s second question, unlike the first, is not a familiar part of standard metaethical discussion. However, appreciation of this fact may have been hindered by the fact that he can seem to present it as continuous with what I will call the “traditional” question of motivation. I will therefore start with this traditional question, to highlight the contrast. As noted by others, much of twentieth-century philosophical ethics was occupied with attempting in some way to accommodate or appreciate what was widely taken to be an obvious truth about morality: that moral claims, or moral judgments, or moral demands, whatever else they are, must have some sort of foothold in the will or in the motivations or the psychology of either the person to whom they apply or the person who makes the judgment. They must be “essentially action guiding,” or “prescriptive” or “normative” in some special but obscure sense beyond that in which the instructions of a cookbook, the rules of a game, or the standards by which musical performances are evaluated are action-guiding, prescriptive, or normative. Call the claim that moral judgments or demands must find a foothold in the will or psychology of each person to whom they apply the internalist thought. The traditional question of motivation takes this thought seriously and asks how it could be realized: how does morality manage to secure a foothold in our will or psychology? As noted, Scanlon can seem to introduce his “question of motivation” as in some way continuous with this traditional question. But, in fact, Scanlon simply sets the internalist thought aside. So, in “Contractualism and Utilitarianism” he says, bluntly but sensibly: an adequate philosophical theory of morality . . . need not, I think, show that the moral truth gives anyone who knows it a reason to act which appeals to that person’s present desires or the advancement of his or her interests. I find it entirely intelligible that a moral requirement might correctly apply to a person even though that person had no reason of either of these kinds of complying with it. Whether moral requirements give those to whom they apply reasons of some third kind is a disputed question which I shall set aside. But what an adequate moral philosophy must do . . . is to make clearer to us the nature of the reasons that morality does provide, at least to those who are concerned with it. . . . It must make it understandable why moral reasons are ones that people can take seriously, and why they strike those who are moved by them as reasons of a special kind of stringency and inescapability. Scanlon thus dispenses with the traditional question of motivation, as something a minimal moral theory need not address. He replaces it with a much more fundamental question: What any moral theory must do, he says, is to help those who are concerned


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

with morality understand why the reasons it provides are so important, or what, exactly, their importance is. This more fundamental question remains absolutely central throughout Scanlon’s work, though it appears in a number of different guises. Perhaps the most precise and illuminating, if not the most colloquial, formulation focuses on moral failing and asks “what reason the fact that an action is wrong provides [one] with not to do it.” More colloquially: What is so bad about wrongful action? Or, better: why avoid morally wrongful action, as such? Or, to put it roughly but perhaps more vividly: Just what is the big deal about morality? Scanlon brings his question into further focus by considering what he calls “Pritchard’s dilemma.” H. A. Pritchard was an intuitionist; accordingly, he believed that the reason to do your duty is simply that it is your duty. No more can be said. Pritchard famously argued that moral philosophy rests on a mistake insofar as it looks for some further reason to do one’s duty. Any further reason, he thought, would be an extramoral reason, and so would be (as I would put it) one reason too many. By providing an extramoral reason, you would have failed to answer the question you meant to answer. Instead, you would have provided an ulterior motive, and so you would have changed the subject. The first horn of what Scanlon calls “Pritchard’s dilemma” is the mistake identified by Pritchard: by appealing to extramoral reasons, you change the subject. But Pritchard himself, by adopting intuitionism, falls afoul of the second horn of the dilemma Scanlon attaches to his name. Insisting, with Pritchard, that nothing more can be said about the reasons for doing your duty—that nothing more can be said about why doing one’s duty is so important—seems remarkably unsatisfying. Moreover, digging one’s heels in, just here, draws certain skeptical concerns that seem to require an answer. More than one prominent philosopher in the last half-century has suggested that the particular sense of importance we attach to avoiding immorality, as such, is something of a humbug—a bit of psychological conditioning we bring with us from childhood or from religious training. And, of course, Thrasymachus and Nietzsche have their own ideas about what gives such demands their special sense of stringency and inescapability. So it seems something more must be said about the importance or significance of morality, to avoid the charge that we are simply in the grip of a kind of taboo, superstition, or scheme of control arranged by those in or out of power. Scanlon’s “question of motivation” inquires between these horns. He wants to understand what more can be said about the reason to avoid wrong action, but he is not looking for a reason in addition to the wrongness of the action. Rather, he is looking to better understand the reason provided by the fact that an action is wrong—the reason provided by the wrongness itself, as such. So he asks his question, “what reason [does] the fact that an action is wrong provide [one] with not to do it”? Asked of other domains, the answer is relatively clear. We might ask what reason to avoid an action is provided by the fact that the action is imprudent. Start with a candidate prudential demand: I ought to floss my teeth each night. Whatever force this flossing imperative carries seems to derive from the importance of avoiding gum disease: this prudential directive carries no more, and no less, significance than that of

Of Metaethics and Motivation: The Appeal of Contractualism


that bit of well-being it is meant to promote. (Fodder, this, for the consequentialist.) If we now ask about the reason-giving force of prudential imperatives, as such, we seem to arrive at the importance of one’s well-being, in general. So, to follow Scanlon’s formulation, the reason to avoid an action provided by the fact that the action is imprudent is the fact that the action will, in some way, compromise one’s own well-being. We could continue to ask this question for different kinds of imperatives, and, in many cases, we can readily give plausible, candidate answers. The force of demands of strategy, as such, seems to be given by the importance of achieving your aim; the force of demands of grammar, as such, by the importance of communicating; the force of aesthetic demands, by the importance of creating or living among things of beauty. In fact, it seems, in each case, we discover the force of an imperative by considering the significance of its violation. We discover the reason to avoid violating a given kind of imperative by considering what is lost when such an imperatives is transgressed or what of importance such imperatives protect, promote, or create. When we turn to morality, however, it seems surprisingly unclear what to say. What is the reason provided by the fact that an action is wrong? What is the importance of avoiding wrongful action, as such? What, exactly, is the big deal about morality? Again, a surprising—indeed, shocking—variety of answers have been proposed. Some claim that, by violating moral requirements, you are failing to live an excellent human life. Others claim that, by violating moral demands you are, in some way, frustrating your own good—perhaps you are failing to realize your potential, or marring your own perfection or the harmony possible in your soul. Still others have it that you are violating the commands of the Creator, or the dignity of human life. Theories that Scanlon calls “formal” claim that, by violating a moral requirement, you are guilty of something like a contradiction, or of failing to make sense. Scanlon takes his toughest opponent to be the utilitarian, on whose account the reason to avoid wrongdoing, as such, is that wrongdoing, as such, fails to bring about the most well-being (or violates rules that, if followed, bring about the most well-being). The variety of answers is staggering. You might expect us to have a better handle on the importance of morality. In surveying the variety of answers given, it is easy to feel that each contains some important truth. Yet each also seems to Scanlon off-target, or at least incomplete, as an answer to his question. Each of these accounts, he thinks, fails to capture something of central importance to morality. Focusing on utilitarianism, Scanlon considers Peter Singer’s article on famine and says But when I feel convinced by Peter Singer’s article on famine, and find myself crushed by the recognition of what seems a clear moral requirement, there is something else at work. In addition to the thought of how much good I could do for people in draught-stricken lands, I am overwhelmed by the further, seemingly distinct thought that it would be wrong for me to fail to aid them when I could do so at so little cost to myself. It is this further, seemingly distinct thought about wrongness that Scanlon thinks remains unaccounted for by consequentialism, and, I suspect, by any view other than


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

contractualism. Scanlon thinks that, if we could understand, or better characterize, this seemingly distinct thought about wrongness, we would be able to answer his basic and yet surprisingly difficult question: we would have a better understanding of the reason provided by the fact that an action is wrong.

answering scanlon’s question of motivation

We can now turn to Scanlon’s own answer to his question of motivation (which will quickly return us to the question of subject matter). Scanlon arrives at his answer largely by reflecting on his own sense of what is lost or violated, in moral failing, and attempting thereby to identify what is missing in the competing answers that seem off-target or incomplete. His answer is captured in his own contractualist theory, according to which, “an action is wrong if and only if any principle that permitted it would be one that could be reasonably rejected by people [who were moved to find principles for the general regulation of behavior that others, similarly motivated, could not reasonably reject].” This contractualist formulation can be difficult to take in. We will consider it in some detail, in a moment. But note, first, that the underlying, core idea is familiar from political theory: we imagine ourselves as both legislators and citizens, creating the principles by which we will then govern ourselves. We imagine that we are symmetrically situated, that each of us has a veto, and that we must come to some kind of reasonable agreement. The principles of morality, as Scanlon understands it, are the principles that we would agree to, in this contractualist situation. They are thus the terms of self-governance adopted by those who recognize each other as having a symmetric standing to determine the terms of their mutual self-governance. They are, we might say, the principles that would be agreed to in a Kingdom of Equals, each of whom is committed to living in a kind of harmony with the rest and so accords to each one a symmetric standing in determining the terms of his or her own self-governance. On such a view, the significance of moral failing is that, in doing wrong, you have violated the terms that would recognize the symmetric standing of each to determine the terms of our self-governance. Thus, to act wrongly is, roughly, to fail to accord others that kind of standing, and so to fail to accord others a certain form of respect. It is this form of respect which Scanlon thinks goes missing in alternative theories, and it is this which he believes provides contractualism with its under-appreciated appeal. I will now examine Scanlon’s contractualist formula in a bit more detail. I will start with a toy example: It would be wrong, I assume, for you to stomp on my foot for fun. Why is it wrong? One wants to say, with the utilitarian, “Because it causes me pain.” And surely, whatever else we say, this must not turn out to be incorrect— that much of the consequentialist position must be preserved. For Scanlon, “because it causes me pain” not an incorrect answer, but it is incomplete. And surely this, too, must be right. After all, my dentist causes me pain on a fairly routine basis, but does nothing wrong thereby. So the mere fact that you cause me pain does not account for the wrongfulness of your action, even though that fact about pain should

Of Metaethics and Motivation: The Appeal of Contractualism


show up somewhere in the story. The consequentialist thinks it shows up as one among many other facts about what the action causes, or tends to cause—filling in these other facts will justify my dentist, but not you. Scanlon thinks, though, that filling in the further story as the consequentialist does will not account for the distinctive importance of moral failing—it will not provide a satisfying answer to his question of motivation. As I understand it, here is how the further story goes, for Scanlon: the fact that your action (or, better, actions such as yours in circumstances such as ours) causes me (someone in my position) pain provides me (anyone in my position) with grounds to reject any principle that would allow actions like yours, in circumstances like ours. But no one has grounds for rejecting a principle that would forbid actions like your in circumstances like ours. Thus, your action violates a principle that no one could reasonably reject, in Scanlon’s contractualist situation. Thus it is wrong. We should press: Why is it that no one has grounds for rejecting a principle that forbids your action? Although I think my pain provides me (or, those in my position) with reasonable grounds for rejecting any principle that allows your action, you might think that your amusement provides you (or, those in your position) with grounds for rejecting any principle that forbids your action. Scanlon answers that, if we assume that we are all committed to finding principles that everyone can agree to be governed by, so long as everyone is committed to finding and being governed by such principles, then (given the circumstances and the various interests at stake) my rejection is reasonable, whereas yours is not. But why is this? What makes my rejection reasonable and yours unreasonable? For Scanlon, the reasonableness of rejection is measured against the aim of finding principles that we can all agree to be governed by, given that we are all committed to finding such principles. But Scanlon acknowledges that this aim does not force a determinate answer in the way many of his interlocutors would like or hope. In determining the reasonableness of a rejection, we have to make what Scanlon sometimes calls a “substantive moral judgment.” Consider this judgment, in our toy example. Importantly, the judgment that your rejection is unreasonable is partly, but only partly, grounded on the fact that, in our circumstances, avoiding pain is somehow more important than being amused. The fact that pain trumps amusement, in this case, plays an important role. However, the question of whether your rejection is reasonable is not simply the question of whether pain trumps amusement, now dolled up in contractualist guise. The question does not simply reduce to weighing the benefits and burdens of the immediately involved parties. One must also consider, for example, the consequences and social significance of adopting, as a principle for the general regulation of behavior, a principle disallowing foot stomping, and these consequences or this significance may overturn a judgment based only on the benefits or burdens of the immediately effected parties. The reasonableness of rejection also depends on whether some alternative, perhaps finer-grained, or differently conditioned, principle would be preferable (e.g., your action is disallowed only if I told you in advance that I do not like my feet crushed), where its being preferable could again turn on the consequences and significance of


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

the alternative principle. Finally, it is worth noting that a reasonable rejection is grounded only in reasons that are, in Scanlon’s terms, both “personal” and “generic.” So, whether a rejection is reasonable does not merely turn on the relative weights of the immediate burdens and benefits to the immediately involved parties. Rather, thinking reasonably about whether to reject a principle requires both thinking as one subject to or protected by that principle—thinking, so to speak, as a citizen under it—and thinking about the significance and effects of adopting such a principle for the general regulation of behavior—thinking, so to speak, as a legislator. When Scanlon says the reasonableness of rejecting a principle is constrained by the aim of finding and being governed by principles that are acceptable to each, he means that it requires thinking both as citizen and as legislator, while recognizing the need to come to a reasonable agreement with others, each one of whom has symmetric standing in determining the terms of our self-governance. Beyond this largely structural account, however, Scanlon does not explain how to arrive at substantive judgments about when the rejection of a principle is reasonable. He simply admits that the idea of “reasonable” is one with “moral content.” He even allows, as grounds for rejection of a principle, robustly moral considerations, such as fairness. This open-ended appeal to reasonableness leaves many discomfited. I will return to worries about it. For now, I want to finish our consideration of Scanlon’s answer to the question of motivation and how his account captures the distinctive thought about wrongness that he found missing in Singer’s article. For Scanlon, the wrongness of an action is not explained simply by appeal to its bad effects, and, accordingly, the reason to avoid wrongness, as such, is not given simply by its bad effects. Rather, the wrongness of an action must be explained by evoking a further fact: the action violates principles that must be accepted by anyone who is committed to living on terms acceptable to each. This further fact provides with the distinctive reason to avoid wrongdoing, as such: by acting wrongly, you have acted in a way that not only neglects the interests of those you have wronged, but that also denies their standing to (partly) determine the terms on which we each shall live. You have, thereby, acted in a way that fails to accord them a certain form of respect. So, according to Scanlon, wrongdoing has its own distinct significance, one that is importantly quite other than causing suffering, failing to achieve human excellence, violating the commands of God, or failing to avoid error or make good sense as a rational creature. You have, as Scanlon sometimes puts it, violated the terms of a relationship of mutual regard, the terms on which a kind of mutual recognition is possible, and so you have put yourself in a different relation to your fellows.

of metaethics and motivation

It will not have been missed that, in displaying Scanlon’s answer to the question of motivation, we strayed into the territory of the question of subject matter. That is, in displaying Scanlon’s account of the reason to avoid wrongdoing, I had to provide his

Of Metaethics and Motivation: The Appeal of Contractualism


answer to the question of what it is for an action to be wrong. We thus strayed because, understood as Scanlon understands it, the question of motivation is not a merely psychological question: it does not ask what impulses, dispositions, desires, or sentiments can be relied upon to motivate moral action. Rather, it asks what important and distinctive reasons to avoid an action appear among the facts that constitute its wrongfulness. Scanlon has thus connected his question of motivation with the question of subject matter. In fact, he tends to support his answer to the question of subject matter largely by appeal to the attractive answer it provides to the question of motivation. Return, then, to the question of subject matter as I originally explicated it. Most of us have certain pretheoretical moral convictions, about, say, murder, cruelty, exploitation, and deceit, in which we have a high degree of confidence. However, if asked why, for example, killing people in certain circumstances is wrong, or why cruelty is objectionable, people give widely varying answers. Given the wide variety of answers, it can seem puzzling what grounds these important pretheoretical convictions, what makes some of them correct and others incorrect. It might also seem puzzling why they should be grouped together as a class. Scanlon’s account both groups these convictions into an understandable class and preserves a role for the answers people are inclined to give. Wrong actions violate the principles that must be accepted by anyone committed to finding principles acceptable to everyone who shares this commitment. More important, many of the answers that people give, when asked about their moral conviction, have a place in Scanlon’s story: they will appear as grounds for the rejection of principles. One could, and often should, appeal to such facts as pain or dignity when rejecting principles. So we have encountered two more points in favor of Scanlon’s view: not only does it provide a satisfying portrayal of the importance or moral failing, it also unifies the subject matter of morality—or, at least, a central part of it—and locates, within that subject matter, the role of many of the disparate considerations we pretheoretically thought belonged within it. Still, many people have been unhappy with Scanlon’s account, especially as an answer to the question of subject matter. We have already encountered one reason, in fact the most popular reason, for the unhappiness: Scanlon does not provide much guidance in determining whether a given rejection is reasonable. He thus seems content to rest his theory on what seems to many an insufficiently articulated or structured base of moral intuitions. Some think that, because of this reliance on an open-ended account of what is reasonable, the view is in some way empty or circular. Others think that, once we fill in the needed story about reasonableness, it will turn out that the distinctively contractualist appeal to reasonable rejection does no real work. These objections seem to me mistaken, and considering why they are mistaken can help us to better understand Scanlon’s position. Consider, first, the objection that Scanlon has rested his theory on an admittedly moral base (of moral claims, facts, intuitions, or forms of reasoning). Call this objector the reductionist. The reductionist insists that we specify non- or premoral (perhaps


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

“nonnormative” or “natural”) facts and forms of reasoning, which will deliver the truths of morality. Though the reductionist might be a contractualist, she will insist that both the grounds for the rejection of principles and the account of when rejection is reasonable be explicable in non- or premoral terms. Scanlon doubts that we will be able to reduce or explain moral judgments by appeal to purely “premoral,” “nonnormative,” or “natural” facts. These doubts rest on his belief in what he calls the “holism” of moral judgments. When considering the possibility of reduction, Scanlon takes welfare as the likeliest “premoral” candidate, and says, . . . it is misleading to suggest that when we are assessing the “reasonable rejectability” of a principle we must, or even can, set aside assumptions about other rights and entitlements altogether. . . . Suppose, for example, that we are considering a principle defining our obligations to those in need. This would seem to be a case in which considerations of welfare are most likely to be predominate. But in order to be in a position to aid someone, an agent must be entitled to dispose of the resources that are needed, and must be free from any obligation that would prevent him or her from acting in the way required to give aid.  .  .  . So in order to understand the scope of the proposed principles (the range of action it might require) we need to presuppose a framework of entitlements. What this illustrates is that a sensible contractualism, like most other plausible views, will involve a holism about moral justification: in assessing one principle we must hold many others fixed. This does not mean that these other principles are beyond question, but just that they are not being questioned at the moment. By highlighting the way in which even this moral reasoning rests on moral assumptions, Scanlon hopes to undermine the motivation for the reductive ambition. A second, and extremely common, charge is that Scanlon’s view is somehow empty or circular. Scanlon claims that an action is wrong if it is in violation of principles no one could reasonably reject, given the aim of finding principles that no one could reasonably reject, given that aim. But the idea of “reasonable” is left unspecified. Scanlon allows that it is constrained by appeal to the aim of finding principles that others, similarly motivated, could not reasonably reject—but that obviously makes reference to the very idea it was meant to constrain. So it seems to some that the view is empty or circular and that the question of whether it would be unreasonable to reject the principle must just be the question of whether it would be wrong to reject the principle. Then the theory is simply serving as a conduit for our pretheoretical intuitions. In charging such emptiness or circularity, one has to be careful. We have seen that Scanlon allows—in fact insists—that in determining whether rejecting a principle is reasonable one makes a substantive moral judgment. He does not hope to ground wrongness in premoral facts. Nor, even, is his primary aim to provide a theory that will generate correct moral principles from more basic or more secure moral claims. Rather, his aim is more “metaethical:” he hopes to provide an account of what wrongness is, or of the facts that constitute wrongfulness. He claims that the wrongfulness of an action or

Of Metaethics and Motivation: The Appeal of Contractualism


attitude is constituted by the fact that it is in violation of principles that could not be reasonably rejected in the contractualist situation. To show the view empty or circular, then, one would have to show that we cannot arrive at a judgment about when a principle can be reasonably rejected—that we can make no determination on that question—except by considering (pretheoretically) whether that which it disallows is wrong. But this Scanlon denies. Even though he allows that the notion of reasonable has “moral content,” and even though he espouses a holism about the moral (indeed, about the “normative”), he does not leave the notion of “reasonable” as impoverished as it would have to be to vindicate the charge that he has left the view empty or somehow reasoned in a circle. Rather, as we have seen, he believes that whether a principle can be reasonably rejected depends on a wide range of considerations—for example, the symmetric standing of each, the burdens and benefits to those immediately affected, the consequences or significance of adopting a principle for the general regulation of behavior, and the possibility of adopting alternative principles. We can think about whether rejection of a principle is reasonable by thinking about such considerations, without simply relying on our pretheoretical intuitions about whether the actions disallowed by the principle would be wrong. Indeed, I believe we have some fairly strong, pretheoretical convictions about what would be reasonable to reject in Scanlon’s contractualist situation. And, if these judgments closely track our pretheoretical judgments about when an action is wrong, this would not show the view empty or circular; it would rather be evidence that the view is correct. Moreover, it does seem that, when thinking about cases that one finds difficult (assisted suicide, perhaps—or choose your own example), this form of reasoning can put pressure on pretheoretical convictions about whether an action is wrong. But if this form of reasoning is able to put pressure on our pretheoretical ideas about whether an action is wrong—pressure that might lead one to question, not the theory, but rather one’s intuitions—then it cannot be simply channeling our pretheoretical convictions about which actions are wrong. The above objections miss their mark by attributing to Scanlon ambitions that he does not harbor. He is not aiming for a reductive account, nor does he hope to generate specific moral principles from more basic or secure moral claims. Again, his aim is more metaethical: he hopes to better understand wrongness, or what it is for an action to be wrong. A final kind of objector questions whether Scanlon has succeeded in his own ambition. This objector insists that Scanlon’s appeal to reasonable rejection, or justifiability to each, is not doing the work; it is redundant. The objector thinks the hard work of determining which actions we have overriding reason to perform or avoid will be done by whatever facts will provide the grounds for rejection of the principles supposedly governing the action. So wrongness turns out to be what Scanlon calls a “buck-passing” notion—it simply indicates the presence of other reasons. Thus, Scanlon’s central idea of reasonable rejection is not what constitutes an action as wrong: rather, whatever facts make a principle one that no one could reasonably reject will also, according to this objector, make actions that violate it wrong. The appeal to reasonable rejection of principles is otiose or redundant.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

This final objector actually makes three different claims, calling for different replies. The objector claims, first, that claims about which principles must be accepted in the contractualist situation will not change what we have most reason to do. The view is, using a term of A. J. Julius’s, “nonproductive.” Second, the objector claims that wrongness is a “buck-passing” notion. And, finally, the objector takes issue with Scanlon’s most central claim, by insisting that an action is constituted as wrongful, not by the fact that a principle prohibiting the action could not be reasonably rejected in the contractualist situation, but rather in some other way. Consider, first, the claim about buck-passing. Though some have used this term to characterize the third charge, above, this is not Scanlon’s usage. I will follow his usage. A “buck-passing” notion, for Scanlon, is one that provides no reason of its own, but rather merely points out, indicates, or labels the presence of other reasons. So, Scanlon’s question of motivation (“what reason is provided by the fact that an action is x?”), asked of a buck-passing notion, would return a null answer. Scanlon thinks value is a buck-passing notion. If we ask what reason is provided by the fact that a thing is valuable, Scanlon thinks the answer is: no reason is provided by the fact that it is valuable, as such. Rather, the fact that something is valuable simply indicates, or labels, the fact that there are other, specific, reasons to treat it in some specific way—to promote, protect, preserve, or pursue it. But, as we have seen, Scanlon denies that wrongness is buck-passing: the fact that an action is wrong provides an additional, distinctive reason to avoid it, namely, the action violates the terms that accord to each standing to determine the terms of our mutual self-governance—it violates the terms of respectful relations. Scanlon calls the fact that an action is wrong a “higher-order reason” to avoid the action—higher-order, because it depends on other, “lower-order” reasons, namely, those that provide grounds for the rejection of any principle allowing the action. Nonetheless, the fact that an action is wrong adds to the reasons to avoid it in a way that the fact that it is valuable does not (according to Scanlon) add to the reasons to pursue or promote or preserve it. So, to claim that wrongness is a buck-passing notion is just to deny what Scanlon finds appealing about his view. It is to be engaged in what I called, at the beginning, the flat disagreement. We can now see an initial answer to the claim that Scanlon’s view is nonproductive: even if it is—even if its recommendations are equivalent to those of a view that did not appeal to agreement about principles—the appeal to agreement about principles provides an additional, and distinctive, reason to avoid wrong action. But I doubt that the view will be nonproductive. Recall that, to arrive at the judgment that a principle cannot be reasonably rejected, we do not simply consider the benefits and burdens to the immediately involved parties; we also have to consider both the effects and the significance of adopting that very principle, and we have to consider whether a different principle might be superior. But these considerations might well change our verdict about whether we have most reason to perform a given action. If these further facts about the adoption of principles for the general regulation of behavior can change the verdict, then it seems that the contractualist view is productive.

Of Metaethics and Motivation: The Appeal of Contractualism


The final claim made by the final objector is the most pressing. The objector claims that the fact that an action violates principles that no one could reasonably reject, given the relevant aim, does not constitute the action as wrong. Rather, an action is made wrong in some other way. Of course, having been made wrong in some other way, an action might also, therefore, be disallowed by principles that no one could reasonably reject, given the aim of finding such principles. The objector might allow that this provides a distinctive reason to avoid the action. But the objector denies that the fact that the action is disallowed by unrejectable principles is what constitutes its wrongfulness. As Judith Thomson puts the point: For my own part, I cannot bring myself to believe that what makes it wrong to torture babies to death for fun (for example) is that doing this “would be disallowed by any system of rules for the general regulation of behavior which no one could reasonably reject as a basis of informed, unforced, general agreement.” My impression is that the explanation goes in the opposite direction—that it is the patent wrongfulness of the conduct that explains why there would be general agreement not to allow it. Scanlon considers this kind of objection, in detail, in several places. I will focus on his direct response to Thomson, in a footnote, where he says, “[t]he contractualist formula that Thomson quotes is intended as an account of what it is for an action to be wrong. What makes an action wrong are the properties that would make any principle that allow it one that it would be reasonable to reject (in this case, the needless suffering and death of the baby).” Scanlon here, in effect, draws attention to the fact that his is a “two-level” view, in which wrongness provides a “higher-order” reason. For an action to be wrong, according to Scanlon, is for it to be in violation of principles that no one could reasonably reject, etc. But, for an action to be wrong, there must be other, strong, “lower-order” reasons that count against it—other reasons that provide winning grounds for rejecting any principle that would allow the action. And in Thomson’s case there surely are. So, Scanlon need not deny that the unspeakable horribleness of torturing babies is what, in an important way, makes the torture wrong. He does, however, need to insist that the unspeakable horribleness, or the needless suffering and death, makes it wrong by providing grounds to reasonably reject any principle that would allow it. This, he says, is what it is for the action to be wrong. One could easily feel this dissatisfying. One might think that the action has been displayed as wrong, in fact as horribly wrong, well before we arrive at the thought that the horribleness, or the needless suffering and death, provides grounds for the rejection of any principle that allows it. The appeal to (as I have put it) what could be willed in the Kingdom of Equals seems, in this case, quite beside the point—a perverse kind of overkill. Such appeal is simply not needed, one might think, to establish the wrongfulness of an action likely to induce a violent visceral reaction in the morally vital. At this point, though, we must tread carefully. Everyone agrees that there are things that are so horrible that they must not be done, and Thomson has certainly put her


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

unswerving finger on one of them. Scanlon has run into trouble, it seems, by denying that it is the horribleness, alone, that constitutes the action as wrong. But, of course, to say that an action is unspeakably horrible, or that it causes pain to someone helpless, is not yet to say that it is morally wrong, or even that it must not be done (consider medical procedures performed in emergency circumstances). Unless we are content to rest with some form of intuitionism, or with the thought that certain actions are simply taboo, we will want to know what makes certain actions, not just unspeakably horrible, but also wrong. And notice that any theory that attempts to give a more articulated account of wrongness will inevitably leave behind our initial sense that anything that is unspeakably horrible, is also, for that reason alone, morally wrong. So this, by itself, cannot be a fair objection. Perhaps the dissatisfaction stems from the thought that Scanlon’s account provides, at least in cases like Thomson’s, an unattractive picture of the motivations of the moral person. That person has now been given, one might think, one reason too many: the horribleness of the torture now serves simply as input into a concern with something like good citizenship. But this objection misunderstands the view. Scanlon’s “question of motivation” is not a question about the psychology of moral agents. It is a question about the significance of moral failing—about the reason provided by the fact that an action is wrong. Further, and very important, although Scanlon does claim that the wrongfulness of an action is “normally decisive” reason to avoid it, he need not insist that its wrongfulness is always the most salient or pressing or, even (now moving beyond psychology) the most important reason to avoid it. To illustrate, it might help to consider how Scanlon’s “question of motivation” would be answered, if asked of demands of etiquette (say “please” and “thank you;” eat with your mouth closed; reply promptly to invitations; use the bread dish to the left; upon making an introduction, provide enough information to facilitate conversation). What is the reason to avoid violations of etiquette, as such? Here is a candidate account: by violating a demand of etiquette, you will upset certain conventionally established social expectations. Your transgression may, of course, also have other kinds of significance: you may also have shown ingratitude, been uncharitable, caused discomfort to those around you, or shown disrespect. However, what gives unity to the demands of etiquette, as a class, (on this candidate account) is simply that violating these demands upsets a certain range of conventionally established social expectations. So the reason to avoid a violation of etiquette, as such, is simply the importance of avoiding the upset of such expectations. Note that, on this account, the importance of satisfying the demands of etiquette, as such, typically will not be nearly as great as the importance of satisfying a particular demand of etiquette (one whose violation would, for example, also show ingratitude or badly inconvenience someone). So, the reason to avoid transgressions of etiquette, as such, typically is not the most important reason to do that which a particular demand of etiquette prescribes. This would explain why demands of etiquette, thought of in a general way, seem relatively unimportant, even though particular violations of etiquette can be very important.

Of Metaethics and Motivation: The Appeal of Contractualism


Scanlon’s account of wrongness can take this form, in cases like Thomson’s. Although he insists that the reason provided by the fact that an action is wrong is important, and that it is normally decisive and normally takes priority in cases of conflict, Scanlon does not claim it is always the most important, salient, or pressing reason, in a given case. In fact, he can claim that someone who was moved, primarily, by the fact that torturing babies is wrong, in his narrow sense, would be monstrous—because that person, though concerned with wrongness, per se, would be completely out of touch with the reasons that make the action wrong. So, if one finds Scanlon’s account unsatisfying, in light of Thomson’s case, it cannot be because one feels Scanlon’s distinctive reason to avoid wrong action, as such, is not always the most important reason to act or should sometimes take a back seat. Scanlon agrees with this. If one is still moved by Thomson’s example, one likely thinks that Scanlon’s account is somehow off-base or off-key as an account of the distinctive reason to avoid wrong action. One would then be raising, against Scanlon, the objection he raised against Singer: With Thomson’s case in mind, one might say, “when I recognize this clear moral requirement, there is something else at work . . . in addition to the thought that the action is in violation of principles that no one could reasonably reject, I am overwhelmed by the further, seemingly distinct thought that the action is wrong.” This is, of course, a possible position—it is another kind of flat disagreement: one simply fails to share Scanlon’s sense of the distinctive importance of morality. One could, for this reason (or, for other reasons), fail to be convinced by Scanlon’s account of the subject matter. Those who find themselves thus disagreeing face the task of articulating their own answers to Scanlon’s questions. In voicing this flat disagreement, one is invited to propose or defend an alternative moral theory.

hijacking the appeal of contractualism?

Interestingly, though, even those who disagree with Scanlon’s answer to the question of subject matter often recognize the attraction of Scanlon’s answer to the question of motivation; they are not engaged in the initial flat disagreement. Rather, they agree that something like respect for each person, or treating others as ones to whom justification is owed, is central among our reasons to avoid wrong action. So I would like to close by considering something Scanlon does not consider explicitly (to my knowledge): What would it take to hijack Scanlon’s attractive answer to the question of motivation and wed it to an alternative account of the subject matter? So suppose one thought that the significance of wrongdoing is, at least in part, captured by the fact that one has acted in a way that one could not justify to others. It might seem that one could then claim that wrongness—what it is for an action to be wrong— is something prior to and independent of being unjustifiable to others, but, nonetheless, any action that is wrongful is also, and therefore, unjustifiable to others, and that this fact about justifiability does, as Scanlon claims, provide the (or at least, a) distinctive reason to avoid wrongdoing. (This seems to be the position Thomson hoped to take,


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

in making her objection.) If such an account were possible, then the appeal of contractualism could be had by another, perhaps less difficult, less abstract, or more determinate account of the subject matter. This would be a considerable blow to contractualism. To illustrate, while keeping matters simple, suppose that an action is wrong because it is in violation of the requirements laid down by our benevolent and just Creator and that one such requirement is that thou shalt not give false testimony. Thus, doing so is wrong. It might seem to follow that doing so is also, and therefore, unjustifiable to others. Thus, it might seem, if you give false testimony, you have not only offended against the commands of God, but you have also done something that does not show due regard to others as ones to whom justification is owed. Thus it might seem that even this overly simplistic theistic view can help itself to Scanlon’s attractive answer to the question of motivation. Notice that hijacker moves from “this action fails to meet some independently specified standard” to “this action is wrong” to “this action is unjustifiable to others,” while insisting that violating the proposed standard is what makes the action wrong. The hijacker thus attempts to tie Scanlon’s attractive account of the distinctive significance of moral failing to her own account of wrongness by exploiting the strong, intuitive connection between what is wrong and what is unjustifiable to others. In fact, given this strong, intuitive connection, it might seem that any theory that succeeds in telling us what makes actions wrong will be guaranteed to have, at its disposal, Scanlon’s account of the distinctive reason to avoid wrong action. Unfortunately for the hijacker, things are not so easy. You cannot simply move from “this action violates my specified standard” to “this action is morally wrong” to “this action is unjustifiable to others,” if, by “morally wrong” you mean nothing other than “violates my specified standard,” and by “unjustifiable to others” you mean to capture Scanlon’s attractive claim that, in wronging others, we have failed to accord them a certain standing or a certain form of respect. Most broadly, this is because not every standard puts respect for others on the line—and certainly not, more narrowly, the specific form of respect at issue in Scanlon’s account. Start with the broader, basic, crucial, but far-too-often overlooked point: we do not plausibly owe it to one another to do everything well. I take this to be a pretheoretical starting point, true for any plausible interpretation of what we owe to each other. Although there are standards of good mathematical reasoning, of good hygiene and personal health, of musical accomplishment and of athletic performance, we do not plausibly owe it to others to satisfy them. And, although a poor performance with respect to such a standard might cause upset to someone who cares deeply about it, the poor performance does not plausibly wrong him or her. Scanlon would account for these pretheoretical intuitions about what we owe to each other by appeal to claims about what would be reasonable in his contractualist situation: if we accord to each symmetric standing to determine the terms of our mutual self-governance, and if we are committed to finding such terms, then we would recognize that certain standards—even certain correct and important standards— could be reasonably rejected as the terms of our mutual self-governance. It would be reasonable to reject them on grounds of, say, taste, ability, priorities, or liberty.

Of Metaethics and Motivation: The Appeal of Contractualism


So, it is important to note that Scanlon’s notion of “justifiability to others” is much narrower than one might have thought. An action (or attitude) might be “unjustifiable” with respect to some correct and important standard, without being unjustifiable to others. In fact, Scanlon’s notion of “justifiable to others” is even narrower than a natural, plausible notion of “justifiable” that one might associate with Scanlon: it would be natural and plausible to think that an action is justifiable if it is supported by the balance of reasons and unjustifiable otherwise. But this is not Scanlon’s notion of justifiability to others: again, I can fail to do what the balance of reasons requires, in completing my logic homework, choosing my attire, or cooking my dinner, without violating terms of conduct that we all must agree to be governed by, if we accord each one symmetric standing to determine what those terms will be. To be justifiable to others, in Scanlon’s narrow sense, is to be justifiable by principles that must be agreed to in his contractualist situation. This narrow notion of justifiability to others is associated with a specific form of respect, and this form of respect gives contractualism its appeal. Again, according to contractualism, the significance of moral wrongdoing lies in the fact one has violated the principles that recognize the standing of each to partially, symmetrically, determine how one shall act—one has violated the terms that would be agreed to in the Kingdom of Equals. The hijacker finds this account of the significance of moral wrongdoing attractive and hopes to claim it for her theory. But she cannot do so simply by showing that some standard is plausibly identified as the “moral” standard (or that the standard leads to a good or decent human life, or even that it is rationally required or inescapable for rational agents). It must also be shown (or, must be independently plausible) that the standard would be ratified by contractualist reasoning. Only so is its violation unjustifiable to others in a way that compromises the form of respect Scanlon has identified. One might now reply, bluntly, that a poor performance with respect to the standards mentioned (e.g., musical, grammatical, rational) does not wrong another, does not violate what we owe to one another, because these are not the correct moral standards. But, one might continue, it seems intuitive and uncontroversial to claim that, if an action is in fact morally wrong, then it would be unreasonable to reject any principle prohibiting such an action as the basis for our mutual self-governance. So, once we grant that some prior, independent standard is the correct moral standard, then, one might think, it is sure to be ratified in the contractualist situation, and Scanlon’s attractive answer to the question of motivation is sure to follow. Thus, one might think, we are assured that that the true moral theory, whatever it is, will hijack Scanlon’s attractive answer to the question of motivation. There are two replies. First, we need to make sure that the previous point has been adequately appreciated. To summarize: you cannot simply move from (a) “this action violates my specified standard” to (b) “this action is morally wrong” to (c) “this action is unjustifiable to others,” if you are claiming that your standard is itself what constitutes the action as morally wrong and if, by “unjustifiable to others” you mean, not simply “unjustified with respect to my specified standard,” but rather, what Scanlon means, in securing his attractive answer—roughly, “principles disallowing this action


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

cannot be reasonably rejected in the Kingdom of Equals.” You can’t simply move from (a) to (b) to (c), because we have some pretheoretical convictions about what we could and could not insist upon, in the contractualist situation. So, if the hijacker accepts Scanlon’s attractive account of the significance of moral failing, that account provides a desideratum for her theory: the hijacker must show that her proposed, admittedly important, standard is also, plausibly, something that would be ratified in the contractualist situation. Thus, it is not so much that the correct moral theory will hijack Scanlon’s attractive answer to the question of subject matter, but that it must. This is a task it must complete. Not every theory—not even every otherwise plausible theory—will succeed in this task. Consider, first, the overly simplistic theistic theory. Even granting that we owe it to our benevolent and just Creator to live in accord with that Creator’s decrees, it is not at all clear why, in owing it to God, I also owe it to you to do so, or why I have wronged you in violating those decrees, as such. Of course, if I give false testimony about you, I have probably harmed you, and that harm may provide grounds to reject any principle permitting your action—but we must put this aside. We are asking why I have wronged you, not in harming you, but rather in violating the requirements of God, as such. But it seems that a principle requiring that we each live in accord with the decrees of the Creator could be reasonably rejected, in the Kingdom of Equals, on grounds of liberty (or, perhaps, on the ground that the establishment of such principles would muddy one’s own personal, spiritual devotion). Even granting that impiety is a serious mistake, it is not clear why it is any more reasonable for others to require that I honor my Creator than it is for them to require that I care for my body. More would need to be said, to make this plausible. I believe much the same can be said about accounts that rely on the constraints or demands of rationality or rational agency (whether they be consequentialist views like Sidgwick’s or Kantian views): even if we grant the requirements of rationality or rational agency are as these accounts claim, it is not at all clear why I owe it to you, or to anyone else, to live up to them, as such. Again, we do not owe it to one another to always be rational or to rightly respond to every reason. So, even granting that the standards of rationality are important—even that they are inescapable for creatures like us, or that they are generated by or generate the only real source of value in the universe (be that God, or pleasure and pain, or practical reason)—does not seem to secure the claim that we owe it to one another to live in accord with them or that they could not be reasonably rejected as a basis for our mutual self-governance, on grounds of, for example, ability, liberty, or privacy. Other accounts may fare better. Suppose the hijacker proposes, as the moral standard, respect for persons (or for rational nature, or for humanity). It seems plausible both that we owe it to each other to treat one another with respect and that we could not reasonably reject, as providing the principles of our mutual self-governance, the standards provided by respect for persons. Moreover, it is plausible that this concern will dominate, in the contractualist situation (i.e., it is plausible that, where respect comes into conflict with other values or goods, respect will take priority, and that, where it might seem that respect must be sacrificed, it turns out that it was not

Of Metaethics and Motivation: The Appeal of Contractualism


respect, after all, but some kind of pride or concern for self-image). Thus, this alternative seems to be in a good position to hijack Scanlon’s attractive answer to the question of motivation. The question is whether such an account provides a genuine alternative to Scanlon’s view. To make clear how this alternative account differs from Scanlon’s, we need to know how or whether treating persons with respect (for their humanity, or for rational nature) differs from treating them as ones who have partial, symmetric standing to determine the terms of our mutual self-governance on the basis of generic and personal reasons. It seems to me an open question whether these differ, one that could be answered only by giving more content or specificity to the notion of respect for persons. So, the first reply to the blunt objection points out that, once we grant that any correct moral theory will include Scanlon’s attractive account of the significance of moral failing, we have placed a significant constraint on moral theory. But there is another reply: even if this constraint is satisfied, even if the hijacker is as successful as she could possibly be, she would not have captured all that is attractive about Scanlon’s view. This is because, although the hijacker insists (or argues) that some independent standard is the true moral standard and so will be ratified in the contractualist situation, Scanlon thinks, to the contrary, that there is no other, independent, standard waiting to be ratified. Rather, as Scanlon puts it, “justifiability is basic.” This claim provides a remaining, important appeal to Scanlon’s theory, an appeal that will be sacrificed by any hijacker. To elaborate: the hijacker hopes to wed some alternative account of what makes an action wrong—what she takes to be the true moral theory—to Scanlon’s attractive answer to the question of motivation. As we have seen, she can do this either by choosing a moral standard that is, on its face, something that would plausibly be ratified in the contractualist situation or else by providing an argument that her standard will be so ratified. In contrast, Scanlon simply notes, in effect, that it is plausible that we owe it to each other (in some pretheoretical sense) to grant to one another standing to partially determine the terms of our mutual self-governance, so long as such standing is exercised consistently with each the standing of each to do the same. Further, and crucially, Scanlon thinks we owe only this to one another. That is to say, we do not, in constructing these moral principles, appeal to any other, prior or independent, moral standard (though we may appeal to moral principles established in some other iteration of the holistic contractualist method). Rather, Scanlon identifies, as the moral standard, whichever principles no one could reasonably reject, if we were all committed to finding such principles. Again, as he puts it, justifiability is basic. His is, so to speak, a minimalist account. An upshot of this minimalism is that any more specific standards, ideals, and pictures of human flourishing and goodness are, for Scanlon, either subordinated to or incorporated into the project of finding and abiding by the principles that must be accepted by all, in the contractualist situation. Although these other ideals or other kinds of value have a place in the morality of right and wrong, what place they have will be determined by asking what principles no one could reasonably reject (considering


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

personal and generic reasons). The overriding concern is to find reasonable terms on which we can get along, given our competing interests and ideals. Other concerns, standards, or ideals are either subordinated to or incorporated into that project. Some will find this unattractive. In fact, anyone who is committed to a contrasting picture, not just of the good or ideal life for humans, but of the good or ideal and therefore moral life for humans, should not find Scanlon’s view appealing. Anyone who believes, not only that some other standard provides an appealing picture of the good or excellent living that we would all do well to adopt, but also that we owe it to one another to adopt that standard—that the alternative standard, itself, rather than our need to find some mutually acceptable way to live together, should command agreement in the contractualist situation—will find themselves in disagreement with Scanlon. (Likewise for those who think that there is a prior and independently specifiable account of what is just or respectful.) And, of course, this is just what the hijacker was hoping for: an independent standard that would command agreement in the contractualist situation. Others will find this minimalism one of the view’s attractions—an attraction that cannot be had by the hijacker. In fact, I think the minimalism has at least two favorable results. The first, which Scanlon sometimes highlights, is what will seem to some (though certainly not to all) a kind of metaphysical minimalism—whatever degree of minimalism might be had by a nonreductive, constructivist theory. Moral principles are the product of our capacity for rational self-governance, our symmetric standing, and our need to get along, taking into account our various and competing interests. Second, and perhaps less often noticed, because the contractualist principles must take into account the (personal, generic) reasons arising from competing ideals, standards of human flourishing, or accounts of the good life—because even those who take such ideals seriously nonetheless must find the principles unreasonable to reject—the result, presumably, will be a set of principles that maximizes the liberty of each consistent with the liberty of others, and so preserves—to as great an extent as possible, given the commitment to finding mutually acceptable terms of selfgovernance— freedom of conscience. Within the constraints provided by the commitment to find mutually acceptable terms, the view remains otherwise neutral in its conception of the good life. The attendant liberty of conscience will be sacrificed in any account grounded in some prior, independent ideal or standard of morality— where recognition of the appropriateness or importance or inescapability of the independent standard secures agreement in the contractualist situation. What will be sacrificed, then, in even a successful hijacking Scanlon’s answer to the question of motivation, is the appeal of old-fashioned, modern, liberalism.


I hope here to have portrayed, with some vivacity, the appeal of contractualism. I hope also to have cleared some confusion about Scanlon’s so-called “question of motivation” and to have shown many of the standard criticisms of his view miss their

Of Metaethics and Motivation: The Appeal of Contractualism


target by misunderstanding its broadly (though not traditionally) metaethical ambitions. Finally, by considering what I take to be a serious criticism of the view, put forward by Thomson (among others), I hope to have shown how disagreement about contractualism is more interesting than is often noticed. Although those who disagree may hope to secure its appeal for their own theories, doing so requires some work, and, I argued, even if that work is successfully discharged, the alternatives still sacrifice something like liberty of conscience. We can note, in closing, that Scanlon’s view shares with Kant’s the result that moral facts are “practical:” they are facts about what people can (reasonably) choose. They are practical in another sense: they are chosen with a view to what it would be like to live in accordance with them. They are, nonetheless, facts, and facts we can easily invest with a great deal of importance. (Whether they are “natural” facts seems to me an unclear and so unhelpful question.) So, in the end, morality does, for Scanlon, have a connection to the will. But it is not to the will of each individual to whom it applies, as the internalist thought would have it. It is rather that moral facts are facts about possible agreements between reasonable people who share the aim of living with one another on terms that accord each symmetric standing to determine those terms. These may after all be facts of a rather queer sort. But it is not hard to see why we give them the kind of importance we do. notes 1. John Rawls, A Theory of Justice (Cambridge, Mass.: Belknap Press of Harvard University Press, 1971). 2. “Despite the wide discussion which [A Theory of Justice] has received .  .  . the appeal of contractualism .  .  . has been underrated. In particular, it has not been appreciated that contractualism offers a particularly plausible account of moral motivation.” T. M. Scanlon, “Contractualism and Utilitarianism,” in Utilitarianism and Beyond, ed. Amartya Sen and Bernard Williams (Cambridge: Cambridge University Press, 1982). 3. I had originally hoped to address, for this volume, Scanlon’s recent work on permissibility and blame (T. M. Scanlon, Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, Mass: Belknap Press of Harvard University Press, 2008).). But to do so, I felt it important first to explain the relation between Scanlon’s account of wrongness and his “question of motivation” (which is really a question about reasons for acting), because it might seem puzzling to find Scanlon, who supports his contractualism in large part by pointing to the appealing answer it gives to the question of motivation, now separating the question of whether an action is permissible from an examination of the individual agent’s reasons for acting. However, this paper does not reach to Scanlon’s recent work. I hope, though, that it will provide some help in reading the recent work, by clarifying Scanlon’s question of motivation and its role in supporting contractualism. 4. T. M. Scanlon, “Contractualism and Utilitarianism,” in The Difficulty of Tolerance (Cambridge: Cambridge University Press, 2003), 125. 5. These problems are sharpened by the fact that, in addressing them, moral philosophy— arguably like mathematics, but unlike the physical or social sciences—does not seek to accurately describe or explain an observable world. It is one thing to accurately describe the moral beliefs and practices of a given group of people, or to explain how or why those beliefs and


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

practices came to be accepted. Both projects belong to social science. It is another thing altogether to explain why people ought to live a certain way, or why certain practices are, in fact, unjust or morally abhorrent. 6. See, e.g., Stephen Darwall, Alan Gibbard, and Peter Railton, “Toward Fin De Siécle Ethics: Some Trends,” Philosophical Review 101(1) (1992). 7. How to motivate this thought? Perhaps one thinks that the instructions of a cookbook are action-guiding, prescriptive, or normative only to those who have the end of making a given dish, that the rules of a game are so only for those playing, and that the standards of musical performance are so only for those who care about music. One might think that moral judgments, in contrast, are to be action-guiding, prescriptive, or normative for everyone (or, for all rational creatures, or all humans). C. L. Stevenson thought it a plain desideratum on any account of the meaning of “good” that “‘good’ . . . must be ‘magnetic,’” that is, “a person who recognizes X to be ‘good’ must ipso facto acquire a stronger tendency to act in its favour than he otherwise would have” (Charles L. Stevenson, “The Emotive Meaning of Ethical Terms,” Mind 46 [1937]: 16). Christine Korsgaard characterizes a more recent version of the thought this way: “If I judge that some action is right, it is implied that I have and acknowledge, some motive or reason for performing that action. It is part of the sense of the judgment that a motive is present: if someone agrees that an action is right, but cannot see any motive or reason for doing it, we must suppose, according to these views, that she does not quite know what she means when she agrees that the action is right” (Christine Korsgaard, “Skepticism About Practical Reason,” Journal of Philosophy 83(1) [1986]: 9). (She says, in her own voice, “Practical reason claims, if they are really to present us with reasons for action, must be capable of motivating rational persons” [11]. And, of course, she takes moral claims to be practical reasons claims.) Thomas Nagel says, “a normative requirement on action must have correspondingly strict motivational backing. If ethics is to contain practical requirements, motivation theory, specifically the theory of rational motivation, must contain results that are similarly inescapable . . . A satisfactory explanation [of the basic principles of ethics] must account for the motivational force appropriate to requirements on action” (Thomas Nagel, The Possibility of Altruism [Princeton, N.J.: Princeton University Press, 1970], 4–5). To read this as a statement of the thought at hand, I am assuming that the “practical requirements” of morality apply to all, and that, similarly, the motivational force must appear in all. See also J. L. Mackie, Ethics: Inventing Right and Wrong (Penguin Books, 1977); Michael Smith, The Moral Problem (Oxford: Blackwell, 1994). For the seminal argument in favor of internalism about practical reasons, generally, see Bernard Williams, “Internal and External Reasons,” in Moral Luck (1981). 8. Scanlon, “Contractualism and Utilitarianism,” 127. 9. T. M. Scanlon, What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998), 4. Note that Scanlon (on page 3) characterizes this question as question his book is most primarily meant to answer. 10. See ibid., 150. 11. See H. A. Pritchard, “Does Moral Philosophy Rest on a Mistake?” Mind 21 (1912). 12. See, e.g., Phillipa Foot, “Morality as a System of Hypothetical Imperatives,” Philosophical Review 81(3) (1972); G. E. M. Anscombe, “Modern Moral Philosophy,” in Ethics, Religion and Politics (Minneapolis: University of Minnesota, 1981); Bernard Williams, Ethics and the Limits of Philosophy (Cambridge, Mass.: Harvard University Press, 1985). Note that the skeptical posture, too, drives one toward a psychological answer to the question of motivation.

Of Metaethics and Motivation: The Appeal of Contractualism


Scanlon instead looks for an answer that provides a reason in what he calls “the standard normative sense.” 13. Cf. Plato, Republic, ed. C. D. C. Reeve, trans. G. M. A. Grube (Indianapolis, Ind.: Hackett, 1992); Friedrich Nietzsche, On the Genealogy of Morals, trans. Walter Kaufmann (New York: Random House, 1969). 14. Scanlon labels this the “triviality” horn, which has led some to think Scanlon falls afoul of it, by presenting an “empty” view. I hope my presentation makes clear what the concern is and how Scanlon avoids it. 15. Scanlon, What We Owe, 4. 16. Thus, in asking the analog of Scanlon’s question, about imperatives of prudence, it seems we arrive at that at which the imperatives, as a class, aim to promote in one or another way. Some people will think the importance of moral demands must follow this same pattern: they must aim to promote a certain (kind of) good, and the importance of moral demands, as such, will be found by investigating the good they promote, as a class. In fact, many have a hard time imagining the importance or significance of a demand or imperative could be understood in any other way. But this is not clear that all must be so understood, and Scanlon will provide a different account of the relation between the particular reasons to avoid particular actions and the reason provided by the larger class. I will return to this below, and illustrate it with the example of etiquette. 17. Scanlon, “Contractualism and Utilitarianism,” 138. See also Scanlon, What We Owe, 152. 18. Imagine the parallel objection: “In addition to the thought of how much harm I would do to my own life or soul, I am overwhelmed by the further, seemingly distinct thought that it would be wrong.” Or, “In addition to the thought of how poor it would be, as a piece of human activity . . .”Or, “In addition to the thought of how impious it would be . . .” Or, “In addition to the thought that it would make no sense, as a piece of willing . . .” 19. Cf. the discussion of guilt in ibid., ch. 6. See note 29, below. 20. Ibid., 4. 21. “Roughly,” because Scanlon now thinks that whether an action is permissible does not depend on whether there was, in fact, disrespect or ill will in the mind of the actor. Rather, it depends on whether the action violates the principles, the establishment and following of which would show respect. See Scanlon, Moral Dimensions. 22. For less toy-like examples: it would also be wrong, typically, for you to say belittling things about me, or to repeatedly draw attention to ways in which I fall short, or to become uncooperative whenever I openly disagree with you. Note that the “for fun” in my toy example will now be interpreted by Scanlon as picking out features of one’s situation, not properties of the intention with which one acts. So, it would be impermissible for you to stomp on my foot when you would stand to gain nothing from it but amusement. If you did so in such a circumstance, by accident, not in order to secure your own amusement, what you did would still be impermissible. However, you would not be blameworthy. See ibid. 23. One might think this answer is mistaken: your action is wrong, not because it causes me pain, but because it violates my bodily rights. I would take any such objection seriously. Still, it seems that any account of the scope and limit of bodily rights will make reference, in some way, to goods and burdens such as pain. (Yet another alternative would say that, in taking pleasure in my pain, you display highly objectionable attitudes. But again, we need to know why these attitudes are objectionable.)


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

24. Scanlon provides some examples of the consequences adopting a principle: “If . . . I lived in a desert area and were obligated to provide food for strangers in need who came by my house, then I would have to take account of this possibility in my shopping and consumption; and if I am not entitled to photocopy articles at will when they turn out to be useful to my course, then I have reason to order a more inclusive anthology . . .” He then provides some examples of the significance of a principle: “Our need for privacy . . . is not met simply because, as a matter of fact, other people do not listen to our phone calls and go through our personal files. In order to have the benefits of privacy we need to have assurance that this will not happen, and this [assurance] is something that general acceptance of a principle can provide. . . . The fact that others recognize reasons to restrain themselves so that I may be free from observation and inquiry when I wish to be is important in defining my standing as an independent person who can enter into relations with others as an equal. [If these principles were not recognized] this would crucially alter my relations with other people, and even my view of myself. (Principles defining my distinctive rights over my own body—rights to say who can even touch it, let alone claim its parts for other purposes—are an even clearer example.)” Scanlon, What We Owe, 203–204. For another example in which considering the general principle might overturn the initial judgment, see the example (offered for a different purpose) of dangerous public projects, at ibid., 236. 25. I will not elaborate on what it is to be personal and generic. Cf. ibid., 204 and 19. 26. See ibid., 212. This occurs in the context of a broader discussion aimed at addressing a circularity objection. 27. You have, he argues separately, violated their value as rational creatures. See ibid., 103– 107. (See also, importantly, 170–71.) 28. Again, the recent work puts questions of permissibility at one remove from questions of the actual motivation of the particular agent. 29. I believe that his account of guilt and blame, of the “force” of a negative moral appraisal, provides an illuminating lens through which to see Scanlon’s entire project. See ibid., 267–77. Cf. also what he calls the “remorse test” in T. M. Scanlon, “Wrongness and Reasons: A Reexamination,” in Oxford Studies in Metaethics, ed. Russ Shafer-Landau (Oxford: Oxford University Press, 2007). 30. So he says, on the third page of Scanlon, What We Owe, “I begin by offering a characterization of the reason-giving force of such judgments [judgments of right and wrong], and then take that characterization as the basis for an account of their subject-matter.” (These projects are carried out in chapters 4 and 5.) 31. It is, no doubt, a slightly revised class from what our pretheoretical convictions may have expected. But, as we will see later, this is not a problem, but rather an indication that the theory is doing some work. 32. One might also have a shallower unhappiness: that Scanlon has not provided guidance in determining which actions are wrong. I hope I have made clear that this is not Scanlon’s ambition. I believe this is the unhappiness found in an otherwise commendable review by Jonathan Hughes and Stephen De Wijze, “Moral Contractualism Comes of Age,” Res Publica 7(2001). The complaint also seems to be made by Gauthier, who finds the view to yield an insufficiently determinate outcome. See David Gauthier, “Are We Moral Debtors?” Philosophy and Phenomenological Research 66(1) (2003). 33. Some might confusedly think that we must provide such a premoral base to avoid circularity. Scanlon will point out that avoiding circularity requires avoiding reliance on the notion

Of Metaethics and Motivation: The Appeal of Contractualism


of wrongness, but not on other moral notions. See Scanlon, What We Owe, 216. His account is holistic, but not viciously circular. 34. Theories that might be thought to be of this form are on offer: Perhaps principles cannot be reasonably rejected if they lead to maximal utility, overall, or if these are the principles that maximize expected utility for each position, or if they are what it would be rational choose given one’s self-interest, or given the aim of maximizing one’s primary social goods from behind a veil of ignorance. 35. His point here is of a piece with a more general view, which one might characterize as a holism about reasons. 36. Scanlon, What We Owe, 216. 37. Scanlon considers the distinct claim that utilitarianism is a “theorem” of contractualism, at Scanlon, “Contractualism and Utilitarianism,” 137ff., and Scanlon, What We Owe, 217–18. 38. It is sometimes hard to distinguish, in the secondary literature, the claim that the view is empty or circular—that, in giving an account of what wrongness is, Scanlon illicitly relies on our pretheoretical ideas (or that our intuitions about what is unreasonable to reject in the contractualist situation are indistinguishable from our intuitions about what is wrong)—from the claim that the view is, as people sometimes put it “redundant”—that what constitutes an action as wrong are the facts that make a principle unreasonable to reject, in the contractualist situation, rather than the fact that the action violates a principle that could not be rejected in that situation. The emptiness/circularity charge leads to and supports the redundancy charge. The charge of circulatory is made, most uncharitably, by Colin McGinn, “Reasons and Unreasons,” New Republic, May 24, 1999, 35. I believe it can also be found in Robert Merrihew Adams, “Scanlon’s Contractualism: Critical Notice of T. M. Scanlon, “What We Owe to Each Other,” Philosophical Review 110(4) (2001): 565–68; Alan Gibbard, “Reasons to Reject Allowing,” review of Scanlon’s What We Owe to Each Other, Philosophy and Phenomenological Research 63(1) (2003): 172; and David Sosa, “T. M. Scanlon, What We Owe to Each Other: A Big, Good Thing,” Noûs 38(2) (2004): 375. The charge is mentioned, but not endorsed by Thomas Nagel, “One-to-One,” London Review of Books, February 4, 1999. All of these objections are anticipated throughout Scanlon, What We Owe, especially chapters 4 and 5. Scanlon addresses them again in T. M. Scanlon, “Reply to Gauthier and Gibbard,” Philosophy and Phenomenological Research 66(1) (2003); Scanlon, “Wrongness and Reasons: A Reexamination.” 39. Thanks to Nishi Shah for helpful conversation on this point. A concern about a regress is raised by Gauthier, “Are We Moral Debtors?” 166. 40. “Pretheoretical,” although, admittedly, embedded in the theory—in the contractualist situation, as described by Scanlon. The convictions are embedded in the theory, but not derived from the theory. 41. One might, then, think that the open-ended appeal to “reasonable” is merely promissory: something to be filled in later by a more complete theory. I take its the openness to be an attraction of the view: we are left to determine, in evolving historical circumstances, what is reasonable. 42. A. J. Julius, “A Lonelier Contractualism,” (in progress). 43. This last is sometimes called the “redundancy objection” (and sometimes, unhelpfully, the “buck-passing” objection). (The aspect of it concerned with the order of explanation is sometimes called the “Euthyphro objection”). I will focus, in the text, on Thomson’s formulation of the objection. (In another place she states it in a way that makes “redundancy” an obvious label: Judith Jarvis Thomson, The Realm of Rights [Cambridge, Mass.: Harvard University Press, 1990], 188 n.5.) This objection has been raised by many (and is often hard to distinguish


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

from the emptiness or circularity objection from which it draws strength). It was raised early by Phillip Pettit, The Common Mind (Oxford: Oxford University Press, 1993), 299–302. It was raised again (though neither endorsed nor pursued) in Simon Blackburn, “Am I Right?” New York Times, February 21, 1999. It has more recently received quite a bit of attention: Philip Stratton-Lake, “Scanlon’s Contractualism and the Redundancy Objection,” Analysis 63(1) (2003); David McNaughton and Piers Rawling, “Can Scanlon Avoid Redundancy by Passing the Buck?” Analysis 63(4) (2003); Philip Stratton-Lake, “Scanlon, Permissions, and Redundancy: Response to McNaughton and Rawling,” Analysis 63(4) (2003). Michael Ridge addresses the objection in what seems, at first, to be a different way than I: he claims that Scanlon restricts the grounds for rejection to reasons that are “agent-relative.” See Michael Ridge, “Saving Scanlon: Contractualism and Agent-Relativity,” Journal of Political Philosophy 9(4) (2001); Scanlon, “Contractualism and the New and Improved Redundancy Objection,” Analysis 63(4) (2003). His answer is rejected (and the objection made) by Joseph Raz, “Numbers, with and without Contractualism,” in On What We Owe to Each Other, ed. Philip Stratton-Lake (Oxford: Blackwell, 2004), 57–60. The issue is nicely discussed in Tamra Frei, “The Redundancy Objection, and Why Scanlon Is Not a Contractualist,” Journal of Political Philosophy 17(1) (2008). My full reply unfolds in this section and the next, where I consider what positive position the objector might occupy. 44. See Scanlon, What We Owe, 98. 45. In fact, Scanlon hopes to use the fact that his view generates the correct class of answers as evidence for it. If it turned out that his view was coextensional with others, he could not make use of that line of reasoning. I owe thanks to A. J. Julius for helpful conversation on this matter. 46. See, again, the examples at note 24. 47. Thomson, Realm of Rights, 30, n. 19. 48. Scanlon, What We Owe, ch. 4, 5; Scanlon, “Contractualism and Utilitarianism”; Scanlon, “Wrongness and Reasons: A Reexamination.” 49. Scanlon, What We Owe, 391. 50. There is a further complication, in this case, about whether a baby can stand as a party to the contract. Scanlon addresses the question of the “scope” of the contract at ibid., 177–87. 51. Cf. Scanlon’s discussion of the “semantics” of “wrong” in “Wrongness and Reasons: A Reexamination.” See especially pages 10 and 15, where he says that the most minimal sense of wrongness is the idea that something “mustn’t be done.” 52. Perhaps to say that an action causes needless pain is enough to show that it must not be done—but there are other things that must not be done, which do not cause needless pain, and we would like to understand the relation between these, as well as the content of “needless.” 53. Thanks to Barbara Herman for helpful conversation on this point. 54. Cf. Scanlon, “Wrongness and Reasons,” 7–10, where Scanlon notes that a moral person will “most often” not be thinking about the wrongness of an action, per se, though its wrongness might play a “backstop” role. (One might be confused Scanlon’s claims that the fact that an action is wrong usually takes “priority,” but that claim concerns cases in which morality competes with other concerns. See Scanlon, What We Owe, 160–68.) 55. Note that demands of etiquette, so understood, are quite different than demands of prudence, in that the avoidance of upset social expectations need not be understood as what each demand or imperative in this class aims to promote. Arguably, the demands of etiquette aim or serve to promote a variety of goods: expressions of gratitude, ready topics of conversation,

Of Metaethics and Motivation: The Appeal of Contractualism


facilitation of event planning, etc. Arguably, there is no one (type of) thing that each member of the class can be said to aim at, in the way that each of the demands of prudence aim at one’s own well-being. (I am obviously denying that they each aim to promote something like social harmony. That seems to me to be something they produce, in fact, but not something they aim to promote in the way that the demands of prudence each aim to promote something that could be abstractly characterized as an aspect of well-being. This requires elaboration.) If this is so, the demands of etiquette provide an alternative to the model provided by prudence: although the class of prudential demands all aim to promote one’s well-being, the class of demands of etiquette does not achieve its unity in the same way. Those who resist the temptation to account for the force of all imperatives on the model of prudence often appeal, not to demands of etiquette, but rather to a quite different kind of imperative: the hypothetical imperative. I believe the temptation to focus on the hypothetical imperative is motivated in part by the internalist thought: this is a “formal” requirement of rationality, and so, one might think, one that all rational creatures are bound by. (There is much interesting recent discussion as to whether the hypothetical imperative is, in effect, a buck-passing notion. See, e.g., Niko Kolodny, “Why Be Disposed to Be Coherent?” Ethics 118[2008]; Michael E. Bratman, “Intention, Practical Rationality, and Self-Governance,” Ethics 119[3] [2009]). However, I think that, for many purposes, etiquette would provide a more interesting contrast. 56. Scanlon does not simply rest content with this disagreement. He would argue that his account not only captures the distinctive importance of morality but also better accounts for the “shape” of our actual moral convictions. Scanlon, What We Owe, 5. 57. Or, perhaps, he has not considered since the original article, where he was content to rest “on a qualified skepticism”? 58. As we have seen, the claim that a rejection is reasonable is, for Scanlon, also a “pretheoretical” claim—even though it is embedded in Scanlon’s theory. It is, however, different than the pretheoretical claim that an action is or is not wrong. 59. (And, for many standards, it is doubtful that there are personal reasons to insist on their satisfaction.) Of course, we often owe it to others to do what a given standard prescribes, for other sorts of reasons—e.g., we owe it to those who depend upon us to look after our health. And, of course, a poor performance with respect to one of these standards could be made wrongful, by making a promise or commitment to a person or group, to live up to the standards (if you are a member of a community of faith, perhaps, or of a team). But absent such overlay of additional commitment, a poor performance with regard to any of these correct and laudable standards is not, itself, plausibly something that we could reasonably insist upon, in determining the terms of our mutual self-governance. (Note that the appeal to such a mutual commitment is the foundation of, rather than an overlay upon, Scanlon’s account of the standards.) We do not, then, owe it to one another to live up to these standards, as such. 60. See, e.g., Julius, “A Lonelier Contractualism.” Cf. Scanlon, “Wrongness and Reasons: A Reexamination,” 9. 61. Notice: this is dangerously close to simply granting that Scanlon’s account of wrongness is correct. 62. Thanks to Mark Greenberg for helpful conversation on this point. 63. Again, it is also doubtful that there is a personal reason to require others to honor God. But see the next note. 64. One might try claiming that we each bear the image of God, and so we owe it to one another, as bearers of God’s image, for whatever reason we owe it to God. If this thought


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

could be made out, it might allow one to argue that this standard should be ratified in the contractualist situation. 65. This is another place at which focus on the internalist thought has distorted our subject matter: it has identified morality with inescapable constraints, rather than with the constraints we owe it to one another to abide by. These seem to me different classes. As a result, theorists have hoped to secure the claim that you have a reason to act morally, and so have attempted to locate morality in rationality. They have thus secured “essential prescriptivity” or inescapability, but, I think, have lost what is distinctive about moral requirements. 66. If it could be argued that they could not be rejected due to personal, generic interests of someone or some class of people, then we would have generated a principle of Scanlon’s sort. 67. Thanks to Seana Shiffrin for this suggestion. 68. Scanlon addresses this briefly. See Scanlon, What We Owe, 103–107. (See also, importantly, 170–71.) 69. These other accounts might be seen not wholly independent, but as providing a prior, underlying rationale for the features of Scanlon’s account: the symmetric standing of each and the restriction to personal and generic reasons. Perhaps this underlying rationale will appeal to the kind of creature we are (persons, or embodiments of rational nature). I think there is reason to avoid being so explicit about the metaphysics of our morals, as I hope will become clear below. 70. See Scanlon, What We Owe, 5 and 189. It is a confusing thing to say: after all, for something to be justifiable is for there to be some story that can be given, in its defense—so how can justifiability be basic? I hope what follows illustrates what Scanlon has in mind. (In particular, what he has in mind when he says that justifiability provides “the most general characterization of [morality’s] content” [189].) 71. Where Korsgaard reads Kant claiming, of the moral maxim, “All that it has to be is law,” Scanlon says, in effect, “the only thing moral principles must be is unrejectable by others with the relevant commitment.” Some might therefore call Scanlon’s a “formal” account, but this is not how Scanlon himself uses the term. See Christine Korsgaard, “Morality as Freedom,” in Creating the Kingdom of Ends (Cambridge: Cambridge University Press, 1996), 166. 72. Morality takes priority. For a model of how a competing ideal or concern will be both incorporated into and subordinated to the concern to live in ways acceptable to each, see Scanlon, What We Owe, 160–66. There he shows both how another ideal (his example is friendship) can play a role in shaping the principles of morality and how morality might play a role in shaping that other ideal. Cf. Scanlon’s own discussion of the relation of these ideals to his minimalism, at Scanlon, “Wrongness and Reasons: A Reexamination,” 18–19. 73. In his 2009 Locke Lectures, Scanlon has elaborated a bit on his metaphysical minimalism. 74. As we have noted, this sort of liberalism requires that other ideals be either incorporated into or subordinated to the goal of living on mutually acceptable terms. The liberty of conscience is not total, nor should the claim of neutrality be overplayed. 75. This essay has been a long time in the writing and owes many debts. It has benefited from extensive conversation and/or written comments from Mark Greenberg, Barbara Herman, A. J. Julius, Brent Kious, Rahul Kumar, Seana Valentine Shiffrin, and Julie Tannenbaum, as well as from audiences at University of California–Riverside, the Southern California Philosophy Conference, Arizona State University, the University of Western Ontario, Princeton University, and the OSU-Maribor-Rijeka Conference, Evaluating Agents. Thanks are also due to Stephen White, for research assistance. Finally, immeasurable thanks are owed to T. M. Scanlon, both for his careful and insightful work and for his extremely generous advice and support over many years.

6 Contractualism on the Shoal of Aggregation Rahul Kumar

1. scanlon’s original presentation of contractualism presents it as a sketch of a plausible alternative to any form of utilitarian, and more broadly, consequentialist, account of moral reasoning. That ambition is evident throughout the developed statement of the view in What We Owe to Each Other, mostly plainly in the requirement that the only reasons that are relevant to a principle’s assessment are those that an individual could appeal to on his own behalf as an objection to the principle. As Scanlon says, “this feature is central to the guiding idea of contractualism, and is also what enables it to provide a clear alternative to utilitarianism and other forms of consequentialism,” one he does not want to give up. The appeal of this feature of the account, which I will refer to as the “individualist restriction,” is obvious in light of its aspiration to be a successful alternative to any form of consequentialism: it rules out as irrelevant to a principle’s assessment the kinds of aggregative considerations that consequentialists take to be the basic materials of sound moral reasoning, those having to do with the potential aggregate value of a state of affairs or the implications of a proposed principle for a group of individuals. Certain questions, therefore, that can be framed in consequentialist terms, such as the justifiability of harvesting a healthy person’s organs into order to redistribute them to a large number of others who will die without them, cannot even be articulated within the contractualist framework. The now familiar worry about the individualist restriction, of course, is that it is overkill. In not allowing any kind of aggregative consideration to count as relevant, contractualism renders itself unable to make sense of the grounds of certain intuitively plausible judgments concerning what it is permissible to do that seem to be best made sense of by appeal to that very type of consideration. 129


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

On the face of it, this is not obviously an objection. The job of an account of moral reasoning is not, after all, to just systematize our considered moral judgments. Rather, it is to help us better understand their grounds. It is not an objection to such an account, therefore, that the characterization it offers of the basis of a certain domain of considered judgments provides grounds for concluding that some of those judgments are mistaken and need to be revised (consequentialists make revisionist claims all the time). What makes the seeming inability of the contractualist account to make sense of the relevance of aggregative considerations worrying is that it is easier to accept that the account is flawed than it is to accept that, for instance, the duty to rescue others when one can do so at little cost to oneself (one that falls squarely within the moral terrain contractualism is concerned with) does not require that, for example, one save the many rather than the one when one can do so at no additional cost to oneself. Scanlon is well aware of this worry: the contractualist justification he offers for why there is a duty, ceteris paribus, to save the greater number is meant to illustrate how judgments concerning cases in which aggregative considerations seem to be relevant can be plausibly defended without any such appeal. But serious doubts have been raised about both its soundness and whether, even if the argument is sound, it can be helpfully generalized to shed light on the relevance of aggregative considerations in cases whose structure does not at all resemble one of an indivisible good to be allocated and several symmetrically situated potential beneficiaries. Many, for instance, will find intuitive the view that, when faced with a choice between locating a research facility that develops volatile chemicals in a large center where, if there is an accident, millions will be made moderately ill and locating it in a sparsely populated area, if there is an accident, most people in the area will be made ill and some will certainly die, the only permissible course of action is to locate it in the sparsely populated area. But it’s hard to see how the relevance of the sheer number who stands to be harmed in this kind of case is to be squared with the individualist restriction. Scanlon calls the individualist restriction a feature of the contractualist account that he is reluctant to give up. But given the damage the restriction appears to do its overall plausibility, it is reasonable to wonder whether such reluctance is justified. Perhaps, as Derek Parfit has strenuously argued, it is not. According to Parfit, it in no way compromises the commitments constitutive of contractualism as a distinct account to allow as relevant to the assessment of a principle’s reasonable rejectability its potential implications for groups of individuals. Revising the account so as to eliminate that aspect of the restriction that disallows such appeals would in fact enable it to better account for the relevance of aggregative considerations that cannot be accounted for using the resources of the view as it now stands. The case for abandoning the individualist restriction in its present form seems both clear and compelling. In the first part of this discussion I offer some reasons for thinking that it is nevertheless mistaken. There is, I believe, a rationale for the restriction’s role in the account that is independent of concerns about what might be justifiable if aggregative considerations are allowed to count as relevant in moral reasoning, a rationale that appeals to contractualism’s central animating ideas. This does not, of

Contractualism on the Shoal of Aggregation


course, challenge the argument that dropping the restriction would enable the contractualist framework to do a better job of doing justice to the intuitive relevance of aggregative considerations in certain cases. But if the rationale for the restriction in the first place does not have to do with blocking the counterintuitive implications of always admitting as relevant aggregative considerations, that just will not be the right kind of reason to justify revising the contractualist account in a way that drops the restriction, either partially or wholly. Arguing that the individualist restriction is harder to uproot that than has been generally acknowledged may not, however, to be doing contractualism a favor. The restriction, after all, appears to be what stands in the way of doing justice on contractualist grounds to the relevance of aggregative considerations when it is intuitively plausible that they are relevant. Defending the centrality of the restriction in the contractualist account appears to leave the view effectively dead in the water as a plausible account of reasoning about what we owe to each other. In the second part of the essay, I offer partial grounds for rejecting this conclusion. Contractualism’s critics have not, in my view, shown that the individualist restriction really is incompatible with making sense of the intuitive relevance of aggregative considerations. The most frequently discussed kind of case, in which there is a choice between saving one or saving many, can, I believe, be shown to be straight forwardly amenable to analysis in contractualist terms. The difficult cases that the discussion ought to focus on have quite a different structure. I will not be able to do more than gesture at these cases in this discussion; what I have to offer here is, more than anything, a ground clearing exercise that, I hope, will bring some of the difficult issues that need to be tackled into view. 2. Scanlon calls the individualist restriction a feature of contractualism that is central to its guiding idea. According to Michael Otsuka, the thought here is that valid moral principles fixing how it is permissible to relate to others must be justifiable to each person. What motivates the demand that valid principles meet this kind of unanimity test is the importance of averting “the tyranny of the majority that would arise if moral principles could be collectively imposed by the combined force of the reasonable consent of the greater number over the individually much stronger but far less numerous reasonable objections of a small minority.” The individualist restriction, as he sees it, is the corollary of this basic demand. Were it not a feature of the contractualist framework, we could not, Otsuka holds, be assured that principles defensible on contractualist grounds will not license the kind of treatment of individuals that principles justified simply by appeal to considerations of aggregate benefit and burden are thought to at least potentially permit. As the guiding idea of contractualism, as he sees it, is thought to have to do with the protection of the sovereignty of the individual, dropping the restriction from the account would be self-defeating. Derek Parfit locates the rationale for the restriction in similar considerations. The individualist restriction, he writes, is given some support by one of Scanlon’s most appealing ideas, that of justifiability to each person. Because we are asking which are the principles that no one


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

could reasonably reject, we must consider each person’s grounds for rejecting some principle, and we can plausibly claim that these grounds are provided by this principle’s implications for this person. The idea that moral principles must be justifiable to each person would not, however, be enough, he argues, to justify the individualist restriction were it not buttressed by the thought that the restriction blocks a familiar kind of utilitarian reasoning that can be used to justify treating individuals in intuitively unacceptable ways in order to secure fairly small benefits for a great many people. The restriction blocks the reasoning that supports this kind of conclusion in what Scanlon suggests is “an intuitively appealing way. It allows the intuitively compelling complaints of those who are severely burdened to be heard, while, on the other side, the sum of the smaller benefits to others has no justificatory weight, since there is no individual who enjoys these benefits.” Parfit ultimately rejects this diagnosis of why utilitarianism lends itself to the intuitively unacceptable conclusions, as having to do with its allowing benefits and burdens that accrue to different individuals to be summed, as mistaken. And as the individualist restriction is a response to this mistaken diagnosis, it should, he argues, be dropped from the contractualist framework. He and Otsuka agree, however, that the ultimate motivation for the individualist restriction lies in what Parfit calls contractualism’s anti-utilitarian protective aim. That contractualism does not lend itself to supporting the kinds of intuitively problematic conclusions familiar from discussions of utilitarianism is certainly an important part of its appeal. But is the circumvention of the possible implications of the tyranny of the majority the rationale internal to the theory for the restriction? If we consider aspects of contractualist thinking concerning why complying with morality’s demands matters to those who do care, a case emerges, I believe, for concluding that it is not. Responding to something’s value, Scanlon holds, is a matter of seeing that certain properties of the thing provide one reasons to respond to it in particular ways. Proper appreciation of the value of a rare malt whiskey, for instance, is a matter of recognizing the reasons to savor its complexities as well as the reasons one has reason not to use it as paint thinner, or use it for a cocktail by mixing it with root beer. Appreciating the value of a human’s life, or the value of the person whose life it is, is a matter of recognizing him as a creature who has reasons, reasons to want to continue to live and reasons to want life to go better, and has a “capacity to select among the various ways there is reason to want a life to go, and therefore to govern that life in an active sense.” In particular, contractualism holds that what regulation of one’s conduct in a way that is respectful of the distinctive value of human life requires is being guided in one’s practical thinking, in whatever type of situation one finds oneself in, by a principle for the general regulation of that type of situation that is justifiable to anyone on grounds he cannot reasonably reject. What the connection is between respect for the value of persons as rationally self-governed creatures and being guided in one’s practical thought by principles

Contractualism on the Shoal of Aggregation


justifiable to any other on grounds no one can reasonably reject is not obvious. It comes more clearly into relief when we distinguish between the thought that (a) conduct or a principle that is justifiable to others is so because it is justified and the contractualist thought that (b) a principle or conduct that is morally justified is so because it is justifiable to any other (assuming that person also values his conduct being justifiable to others) on grounds no one can reasonably reject. The appeal to the importance of principles being justifiable to others is an appeal to a substantive view about what moral justification requires. Central to this view is that a principle justifiable to another on grounds that he cannot reasonably reject is one that he has reason to license others to be guided by in their practical deliberations. What is involved in assessing whether a principle is one that individuals have reason to license one another to be guided by is helpfully illuminated by considering whether such a principle could be the object of a hypothetical agreement. It invites us to think of a valid principle for the regulation of a certain type of situation as one that is worked out through a process in which all those within the moral domain (anyone, on this view, with a capacity for rational self-governance) participate as co-deliberators, each with an equal voice, in working out what the principle for the regulation of the type of situation in question ought to be. In evaluating candidate principles, each compares the implications for himself of each candidate principle being the principle regulating the type of situation in question. But the assessment by each of which, if any, candidate principle he has reason to agree to is not based solely on which principle’s implications are most favorable for him. Because each cares about the justifiability of his conduct to any other, each takes into account in his own assessment of whether a candidate principle is one that he has reason to agree to both his own reasons for favoring or objecting to the principle and the reasons of other individuals for objecting to or favoring it. Convergence on a principle, one that each, as assessed from his own point of view, has reason to agree to, is arrived at through a process, one might say, of individuals comparing the strength of one another’s objections to different principles. The converged-upon principle will be one of which it is true that the most serious objection that can be pressed from a particular point of view against it is not as strong as the objections, from other points of view, to any other candidate principle. To say that a principle no one could reasonably reject is one that individuals license one another to be guided by in their practical thinking is to say, therefore, that the principle is one that any person, assessing it from his or her own point of view, has reason to agree to for the regulation of the type of situation in question. By being guided by such principles in how one relates to others, one’s conduct conforms to what respect for the value of others requires. For to be guided by principles that others authorize or license is to recognize others in one’s practical thought as reason-assessing, self-governing creatures, capable of assessing reasons and justifications concerning principles regulating how it is permissible for individuals to relate to one another. Recognizably Kantian, the thought here is that the dignity of an individual lies in the fact that he or he stands to the moral law as both subject and legislator.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

Admittedly, this is just a sketch of what the argument is connecting respect for the value of human life and being guided by principles justifiable to any other on grounds no one could reasonably reject. A full defense of the argument for the connection would require a great deal more discussion. What matters for present purposes, however, is that this line of argument does not support contractualism in the form Scanlon presents it, with the individualist restriction. What it supports is the broadly contractualist thought that assessment of a principle’s justifiability requires engaging in a thought experiment in which a principle’s implications are considered from a number of distinct individual standpoints. In this thought experiment, each is recognized as a “co-deliberator” in the determination of what the principles are that are to govern how it is permissible for individuals to relate to one another, in the sense that a valid principle must be one that, from each standpoint, there is decisive reason to agree to. Each standpoint can be said, on this picture, to wield the hypothetical power to veto a proposed principle if it is the case that, as assessed from that standpoint, the balance of reasons does not favor acceptance of that principle. This leaves open the question of whether there is a clear rationale for the specific contractualist claim that the only considerations relevant for assessing the justifiability of a principle to another as one that he has reason, as assessed from his own point of view, to accept are those not ruled out by the individualist restriction. For it is compatible with, for instance, allowing groups of individuals to appeal to the implications of a proposed principle for them. Allowing this type of consideration as relevant to the assessment of a principle’s reasonable rejectability would still leave intact the connection that is central to the contractualist account between respect for the value of human life and being guided in one’s practical thought by principles no one can reasonably reject. What this suggests, therefore, is that whatever the rational for the individualist restriction might be, it does not flow in any straightforward way from a concern to regulate oneself in a way that is consistent with respect for the value of other human beings. Contractualism does not, however, identify the importance of complying with morality’s requirements (for those who take doing so to be important) with a concern to live on terms with others of mutual respect for the value of one another as persons. Rather, what accounts for morality’s motivational pull is the value of the relationship that holds between oneself and any other being a good relationship of its kind, one Scanlon calls a “relationship of mutual recognition.” For the relationship one stands in to others to be characterizable as one of mutual recognition is, on Scanlon’s view, to stand in solidarity, or moral community, with our fellow creatures. The idea here is helpfully illustrated by an analogy Scanlon offers to help illumine the character of the relationship that all rational beings stand in to one another, that of parents and children. That you stand in a relationship of parent to a child may just be constituted by a certain set of biological facts, quite independent of any attitudes you happens to have. In virtue of these facts, it can be said that there are certain duties you owe your child. What is dependent on your attitudes and conduct is whether one is successful in living up to the norms constitutive of being a good parent; if you do not do so, though you stand in a relationship of parent to the child,

Contractualism on the Shoal of Aggregation


that relationship will not contribute to your life and will not contribute to the child’s life to the extent that it might otherwise. The contractualist thought is that though the mere relational fact of being fellow rational beings is a sufficient basis for holding that we owe certain things to one another, it is also the basis for holding that all rational beings stand in a relationship to one another, one that can be valued for its own sake because there are standards governing how one ought, ideally, to behave and feel governing this type of relationship, and whose value can better or worse realized partly as a result of one’s own conduct and attitudes. In particular, contractualism holds that to take it to be important that one’s conduct be justifiable to any standpoint for which such conduct being permissible may have implications is to take anyone whose standpoint it is (or could be) to be one with whom one has a certain kind of relationship that one has reason to value, one that that can contribute to one’s life in the way other valuable relationships. “Could be” because it is important that one’s conduct be justifiable not just to those now living for whom it could have implications, but also those who have been or will be in the future. This is a striking claim about which there are good grounds to be skeptical. A proper assessment of its plausibility goes beyond what can be pursued here. My own inclination is to think that the view that the idea that all rational beings—past, present, and future—stand in a relationship to one another may not be as implausible as it might at first seem. One reason for thinking this follows Scanlon’s remark that charges of injustice or immorality that appeal to reasons for thinking that our institutions, policies, choices, or conduct are not justifiable to others lead us not only to think that we may have wronged others, but to feel estranged from them and guilty to them. That such sentiments are warranted in our relations with non-human animals is unclear. On one way of looking at the matter, feelings of estrangement from, and guilt to, another is warranted by a failure to conduct oneself in a way justifiable to that person on grounds he cannot reasonably reject. One might, as Scanlon suggests, quite reasonably wonder why it is important to take non-human animals that lack the capacity for rational self-governance to be creatures that can be wronged in this way. Recognition as one who can be wronged is to be recognized, after all, as having a certain status, that of a rationally self-governing being. Arguably, we can make sense of the grounds for it being wrong to cause many types of non-human animal to suffer needlessly without attributing to them any such status. One way the importance of recognition as a being capable of rational self-governance might express itself is in content of the duties owed to a person qua one who has a capacity for rational self-governance. The wrongness of certain kinds of treatment only makes sense if one treats a person capable of rational self-governance in this way. But it could also express itself through the recognition of the fact that those capable of rational self-governance are sensitive, and therefore vulnerable to the expressive and symbolic aspects of how they are treated because they are sensitive to the reasons of others for doing what they do. Non-rationally self-governing animals are not capable of this sensitivity and so are not vulnerable to what Philippa Foot calls “second order evil in human life, meaning for instance the misery that comes from the consciousness of being disregarded, lonely, or oppressed.”


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

This suggests, very tentatively, a way of understanding the thought that all capable of self-governance, at all times and all places, stand in a relationship to one another. It is because a rationally self-governing being is naturally disposed to care about not just the implications for him of what others do, but what their reasons for their conduct say about their attitudes towards him, that such beings are vulnerable to one another in a way that is characteristic of many kinds of personal relationship that we value. Like the relationship of parent and child, persons have certain obligations to one another simply in virtue of standing in that relationship to one another. Genuinely valuing the relationship and making a success of it, necessary for the value of the relationship to contribute to the lives of both parties to it, requires that, in both the case of a parent and that of the basic moral relationship between persons, one regulate both one’s conduct and one’s attitudes in ways that the standard for success in that type of relationship requires. Might the reasons that could be offered to establish a principle as justifiable to another in the broad sense diverge in an important way from those that could be offered in defense of a principle’s justifiability to another as being a principle that is consistent with valuing a relationship of mutual recognition holding between oneself and another? A suggestion that there is a difference here, and how it might be understood, is implicit in certain remarks that Nagel makes in his seminal article, “War and Massacre.” Nagel’s thought there is that hostility directed towards another establishes a personal, “I thou,” relationship between oneself and the object of one’s hostility. The hostility may be justified, but whether it is justified turns on whether there is an adequate justification for what one contemplates doing that can be addressed specifically to the victim of one’s hostility to whom one stands in this kind of relationship. Nagel sees this requirement as a significant source of constraint on how one might justifiability relate to another. By way of illustration of why this is so, he says If one abandons a person in the course of rescuing several others from a fire or a sinking ship, one could say to him ‘You understand, I have to leave you to save the others.’ Similarly, if one subjects an unwilling child to a painful surgical procedure, one can say to him, ‘If you could understand, you would realize that I am doing this to help you.’ One could even say, as one bayonets an enemy soldier, ‘It’s either you or me.’ But one cannot really say while torturing a prisoner, ‘You understand, I have to pull out your finger-nails because it is absolutely essential that we have the names of your confederates’; nor can one say to the victims of Hiroshima, ‘You understand, we have to incinerate you to provide the Japanese government with an incentive to surrender.’ Nagel’s thought here is hard to make particularly precise, but it can be taken to be something along the following lines: to see the relation between oneself and the object of one’s proposed hostility as personal is to understand the moral justification of one’s conduct as requiring an adequate reply to a challenge to the permissibility of one’s conduct posed by one’s victim in personal terms—“how can you do this to me.” An appropriate response to this kind of challenge is one that is appeals to similarly personal

Contractualism on the Shoal of Aggregation


considerations; to appeal to more impersonal considerations would be to fail to acknowledge the character of the challenge, and would not, therefore, engage the challenge in a way that aims to directly address the challenger. It is this idea that Nagel may have in saying that one could say to another “I have to leave you to save others,” whereas one cannot really say “I have to torture you to get the names of your confederates.” The first appeals to the way in which each of the others could press a personal challenge to doing other than saving each of them, a response whose relevance to his challenge that the one left behind cannot deny given the basis of his challenge. If, on the other hand, the confederates in question are something like fellow conspirators in a plan to rob a bank, appealing to the importance of gaining their names is a comparatively impersonal consideration, one whose relevance as a response to his more personal challenge the intended victim would be hardly unreasonable to deny. The contractualist picture of moral justification has deep affinities with the picture Nagel sketches. It portrays reflection on the justifiability of relating to another in a certain way as a matter of asking oneself what one could imagine saying to another in the face of his challenge to the permissibility of your relating to him in that way. The character of the challenge envisioned here is personal, as the imagined challenger is one to whom one stands in a particular kind of relationship, one of mutual recognition. What matters, therefore, is not just that the permissibility of relating to the challenger in the way in question be justifiable to him, but that it be justifiable on terms that that are consistent with continuing to stand in a relationship of mutual recognition to him. The individualist restriction answers to this concern by restricting the basis of potential challenges to a proposed principle to those of its implications that might matter to an individual because of their importance to his life. Any other basis of a challenge would have to do with something an individual might take to be important, but not important because of its importance to his life, and would not, therefore, bear on the question of whether his life has been appropriately recognized in the way that mutual recognition requires. It is for this reason that meeting the imagined challenge requires appealing to what a principle that did not permit relating to another in this way would mean for the life of another individual. What makes the challenger’s objection to being related to in a certain way personal is that it is based on what this being a permissible way to relate to another means for his life. By meeting this complaint by pointing to the more serious implications for the life of another of it being impermissible to so relate to another, one both recognizes the personal character of the objector’s complaint and offers a response whose relevance to his objection the challenger cannot deny (assuming he values standing in a relationship of mutual recognition to other individuals). A response that appealed to a collective burden, on the other hand, would not be one whose force the objector could not deny, as the possibility of some individuals having to bear a collective burden is not one that is inconsistent with standing in a relationship of mutual recognition with any other. It does not follow, after all, that because several individuals stand to be burdened to a certain extent, and that, taken together, their collective burden is very great, that they form a point of view that can be understood as one to which justification is owed and from which challenges could be issued.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

An appeal to a collective burden need not be made by a collective. An individual could make an impersonal appeal to the bad effects of a principle for a large number of people as a reason for wanting to reject a proposed principle. But this would not establish that the collective burden that might obtain as a result of individuals being licensed to be guided by that principle is grounds for reasonably rejecting that principle. All it would show is that one (not very strong) objection to the principle is that some might be burdened by having something that they are concerned about not happen, which is the implication of the collective burden obtaining for the person appealing to it as the basis of his objection to the proposed principle. The rationale for the individualist restriction is normally thought to have to do with concerns about what might prove to be morally justifiable if considerations of aggregate well-being are taken into account as relevant. What I have argued for so far is that this is not quite right; though the implications of the restriction for what turns out to be justifiable on the basis of the contractualist framework may contribute in certain ways to contractualism’s appeal, its place in the theory is motivated by an understanding of morality (or, at least that aspect of it contractualism theorizes) as deeply interpersonal that is at its heart. It is “deeply interpersonal” in that it treats moral concern as fundamentally having to do with valuing a relationship of mutual recognition holding between oneself and any other, and it is because the morally motivated person values standing in such a relationship to all and any other that he is concerned, in his moral thinking, about the way in which another might challenge on personal grounds the permissibility of being related to in a proposed way. Such a challenge presents the possibility of a rupture in his relationship to, and estrangement from, other individuals, and that is something he is concerned, as a morally motivated person, to avoid. What he values is standing in unity with his fellow creatures rather than having to hide from their (perhaps imagined) gaze, or finding ways to avoid any thought of them as a way of avoiding pangs of guilt. 3. The conclusion of the last section was that the individualist restriction is more tightly bound up with certain ideas that animate contractualist thinking than those who advocate its elimination from the account appear to recognize. Revising it so as to drop the restriction would require fairly foundational changes to other aspects of the account. That does not mean there might not be a strong case for doing so if that would significantly advance the realization of contractualism’s overarching theoretical aim of being a plausible alternative to any form of consequentialism. Whether it would would require surveying more cases than I can cover in this essay. But to start, I want to consider one substantive reason that has been frequently advanced for thinking the individualist restriction has to go: that the relevance of aggregative considerations cannot be made sense of on contractualist grounds in even the simplest case, such as the following: Rocks: You are in a boat, equidistant from two rocks. On rock A, there are six people. On rock B, there is only one person. The water is rising, and there is only time to get to one of the rocks and save its occupant(s), before water engulfs both rocks.

Contractualism on the Shoal of Aggregation


Most believe that the rescuer has a clear obligation to save the six, and that this obligation is appropriately characterized as one that is owed to each of the six, such that it is the kind of obligation whose basis it is contractualism’s aim to illumine. That is, to decide to head toward rock B rather than rock A strikes many as objectionable because it wrongs those on rock A. The particular problem that this case presents for contractualism can be seen if we start with the simple idea that the rescuer has a duty to aid, and that because all the potential recipients of the aid are symmetrically situated and there are no relevant asymmetries between them, each person who stands in need of aid has a claim to the aid. Because all claims cannot be satisfied, one could, with good reason, conclude that no matter what one does, someone will be wronged, and that the best one can do in this kind of case is to satisfy as many claims as possible, though this is not a duty owed to anyone. This kind of approach to the case is not compatible with the contractualist account, for two reasons. First, contractualism treats all obligations as obligations owed to individuals. If there is an obligation to save the greater number then it must, on this view, be an obligation owed to each of the greater number that he be saved. Second, if it is owed to a person that he be saved, contractualism holds that this obligation is in virtue of a particular principle for the regulation of type of situation the person finds himself in. The “situation” is one in which there are several individuals with a legitimate and equally pressing interest in being saved, so any valid principle for the regulation of this type of situation will be one that mandates a particular procedure for the determination of which individual, or individuals, has a valid claim against the rescuer to be saved. A basic constraint on a valid procedure for the regulation of this type of situation is that each person’s interest in being aided be taken into account in settling who has a claim to aid. As a constraint on any procedure for settling which person has a valid claim to be aided, this is a weak requirement, one satisfied by utilitarianism. In the Rocks case, recognition of each person’s interest in being aided requires that the importance that each person may, with good reason, attach to efforts being made to save his life be counted in favor of any course of action that ought to result in his life being saved. What blocks a utilitarian justification for saving the greater number—say, a principle that required the rescuer to always save the greater number—is the requirement that any principle for the regulation of this kind of conflict situation must be justifiable to each person, and in particular, to the person not saved. A utilitarian explanation for why the greater number should be saved appeals to a consideration that the individualist restriction rules out as relevant for assessing the justifiability to the person not saved of his not being saved. A procedure for settling how the indivisible good is to be allocated that has, as an implication of its correct application, that in Rocks, the six have a claim to be save and the one does not, that is justifiable to the one not saved must, as I see it, satisfy two desiderata: it must involve (i) taking into account each life that is at risk as counting in favor of any course of action that is likely to result in his life being saved, and (ii) the


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

identification of an undefeated reason that each of those not saved has reason to acknowledge from his own standpoint (provided he is morally motivated) as a justification for his not having a claim against the rescuer to be saved. A plausible procedure that satisfies these desiderata is one that emphasizes importance of symmetry. What this means is easiest explained by example. So, consider a simple case of Jones in one boat and Smith in the other. Both are sinking, equidistant from the rescuer, and there is only time to save one of them. If the rescuer is to stand in a relation to each of them of his conduct being justifiable to each of them, he has reason to take into account in his decision Smith’s life and Jones’s life as each counting in favor of two different and mutually exclusive courses of action. Moreover, neither Smith nor Jones can claim that the rescuer owes it to him to do what is necessary to save his life, as either making such a claim must be justifiable to the other by appeal to a principle no one could reasonably reject. Smith’s reason for wanting to be saved is symmetrical to Jones’s reason for wanting to be saved, then, in the sense that the life of each is a reason to take the same type of action (doing what is necessary to save a person’s life) even though each of their lives supports pursuing distinct particular actions. The justifiability of either of them having a claim to be saved requires producing a consideration that each has reason to acknowledge as relevant from his own standpoint, at least insofar as each values his claim being justifiable to the other on ground that person cannot reasonably reject, as to why their situations are in fact asymmetrical, such that there is a reason that favors one of them having a claim to be saved, but not the other. By assumption, there is no such reason, so there is no principle justifiable to both Smith and Jones on grounds neither can reasonably reject as to why one of them, but not the other, has a claim against the rescuer to be saved. This conclusion is consistent with each having a different claim against the rescuer, that the rescuer does what is necessary to save someone. Only a principle that would justify one but not the other having a claim to saved has been argued to be reasonably rejectable. This line of thought plays an important role in the argument for why it is that a procedure for settling cases like Rocks that is justifiable on contractualist grounds will be one that directs the rescuer to pursue a course of action that results in his saving each of the greater number because he owes it to each to do so. That is, on this account, the greater number ought to be saved, not because they are the greater number, but because doing is a consequence of the rescuer’s obligation, owed to each of those who constitute the greater number, to do what is necessary to save his life. Consider, first, a two-person case, with just A and B in peril. What a procedure that no one can reasonably reject for determining who, on the basis of all relevant reasons, has a claim to be saved here requires is that (a) the most important competing considerations that are of roughly equal strength be taken into account; (b) competing considerations that speak to opposing courses of action be set aside (or neutralized) for purposes of the resolution of the conflict, and (c) a determination be made as to whether or not there is a relevant consideration that remains that counts in favor of one of the relevant courses of action rather than the other. If there is, it will be decisive in settling the question of which course of action ought

Contractualism on the Shoal of Aggregation


to be pursued. Call steps (a)–(c) the conflict resolution procedure. The rationale for each of these steps calls for elaboration. Step (a) requires making a judgment as to what the most important relevant considerations are that favor each of the available courses of action; for present purposes, I will assume that relevant considerations are just the lives at stake (relevant qualifications will be taken up further on in the discussion). The relevant considerations speak to two courses of action one could embark on: either head towards A’s rock, or head towards B’s rock. The situation is such that both A and B cannot be saved. Doing nothing is not a relevant course of action. The conflict that needs to be resolved here is one that arises because there is a duty of aid, such that all those whose lives are at stake at least have a claim against the rescuer that he do something to aid those in need of it; the question for the rescuer is whether one or the other has an exclusive claim to be aided, such that he, but not the other, will be wronged if not aided. Step (b), in which symmetrical considerations that support distinct and mutually exclusive courses of action are neutralized with respect to the resolution of the question of which direction the rescuer ought to go in, relies on the importance to each that if he is to make a claim to be saved, that claim must be justifiable to the other (or others) on the basis of a principle for the regulation of the type of situation they find themselves in that no one can reasonably reject. It is because each cannot justify to the other a principle that gives his, but not the other, a claim to be aided by the rescuer, that both have reason to accept as neutralized the importance to each of being saved as relevant considerations for determining whether or not either has an exclusive claim to aid. Step (c) may seem puzzling in A versus B type cases, in which there are no relevant considerations which speak in favor of saving one rather than the other. In such cases, a valid procedure will be one that authorizes the rescuer to take a decision as to which course of action to undertake. Whoever the rescuer ends up saving, then, will be lucky that the rescuer decided to favor his, rather than the other. Note, though, that the luck here is limited to the decision concerning which direction to go in. Once that decision is made, the rescuer has reason to take the steps necessary to save the person whom he has decided to save and not reconsider the matter (wasting time may only result in it being impossible to save anyone). The other may well be disappointed that he is not to be saved, but cannot complain of having been wronged. To be wronged he would have had to be entitled to more than just having the importance of his life at stake being given due consideration. That requirement has been met by there being a reason that he has reason to acknowledge as a justification as to why he has no claim against the rescuer that he, rather than the other person, be saved. Elizabeth Anscombe’s seminal discussion of this kind of case suggests a further condition, that (b) the decision of who to save not be made on inappropriate grounds. Her suggestion, as I understand it, is that the prerogative to decide is not a prerogative on any basis whatsoever. As she puts it, it is important the basis of the decision not signal some kind of “ignoble contempt” for the person not saved. Rather, the decision concerning how to proceed ought to be one that is expressive of respect for the seriousness


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

of what is at stake—individual lives. That requires that the rescuer not take his decision on the basis of considerations that are either independently objectionable, or just seem to trivialize what is at stake. For example, to decide to save one, but not the other, because one is wearing a green sweater and the other is wearing a red sweater is objectionable to the extent that such reasoning trivializes the fact that the decision to be made is a life-and-death decision. These considerations might be taken as militating grounds against the rescuer having a prerogative to just plump in favor of one course of action or the other, in favor of a coin-toss as a means for deciding between the competing courses of action in the two-person case. This would have nothing to do with giving each person an equal chance of being saved, but with the value of a lottery as a suitable random mechanism for decision that does not raise questions about either the recognition or the trivialization of the relevant considerations. The attitudes of the person making the decision cease to be of concern in the same way, as they have no bearing on the outcome of the toss. The difficulty with this argument is that the force of the intuition at the heart of it—that, for instance, there is something objectionable about saving one person rather than another because the one saved is wearing a T-shirt from a concert of your favorite band—trades on an ambiguity. It is certainly true that there is something objectionable about treating the T-shirt a person is wearing as a justification for saving that person, or perhaps grounds for holding that he in fact deserves to be saved in a way the one left behind does not. But simply choosing to save one person rather than another because of the t-shirt he is wearing need not express anything about the importance of what is at stake, as the fact that the person is wearing a certain t-shirt need not figure as a consideration that has any kind of justificatory relevance for the decision of whom to save. Consider now a three-person case: A on one rock, B and C on the second rock (the same facts concerning the tides apply in this case). Once again, the immediate decision that must be taken to resolve the conflict concerns which of the two directions the rescuer should direct his boat, and it is to that decision that the contractualist procedure for resolving conflict cases is to be applied. The relevant reasons concern the lives that are at stake. One relevant reason stands unopposed. That reason decides the conflict. If the reasons to save A and B are to be neutralized, that reason will be the fact that C’s life is at stake. If the reasons to save A and C are neutralized, that reason will be the fact that B’s life is at stake. Which pair of opposing reasons are neutralized does not change the conclusion that the remaining relevant reason is one that counts in favor of the rescuer directing his efforts towards the rock upon which B and C are perched. To avoid misinterpretation, it is important that this does not establish that rescue is owed to C, in the sense that he has a claim against the rescuer that he be saved. The rescuer’s duty, at this stage, is that he comply with the outcome of the procedure that he is authorized (by a principle no one can reasonably reject) to use to resolve conflict cases. This is a duty that he owes to all those who are possible direct beneficiaries of the good whose allocation is the subject of the conflict. In directing his boat towards the

Contractualism on the Shoal of Aggregation


second rock, therefore, the rescuer is simply complying with a duty owed to all those who have an immediate stake in the conflict’s resolution. B and C have a claim against his to be saved only once it has been settled on the basis of a fair procedure that he should head towards B and C, rather than A. Is A’s not having a claim to be saved justifiable to his on grounds no one can reasonably reject? It is if the justification for neutralizing both the relevance of his life as a reason for heading in his direction and the relevance of another’s life as a reason for heading in a mutually exclusive direction from the two person (A or B) case is sound. The A or B and C case introduces nothing new as far as the consideration A is owed in determining which direction the rescuer ought to head in. After all, the source of the conflict—there being reason for the rescuer to direct his aid efforts in two incompatible directions—is identical in the A or B and the A or B and C case. The only difference the later case makes is that it introduces a relevant and undefeated consideration in favor of heading in the direction of B and C. Whether or not the contractualist account employs what is often called “the method of pairwise comparison” is a difficult question. The above line of argument offers some reason for thinking that it does not. A resolution of the A versus B and C case using strict pairwise comparisons would suggest that the right way to proceed is for A to metaphorically confront B, neutralizing both, and for A to then go on to metaphorically confront C (again, neutralizing both), as A has initially only been neutralized with respect to B, not C. The A versus B and C case would then be structurally identical to the A versus B case, leaving the rescuer to exercise his prerogative to decide whom to save. The line of argument on offer, however, treats the reasons to save A, B, and C as to be addressed at once, through a single decision procedure that in principle takes into account each person’s reason once and only once. Being taken into account by the procedure, one could say, is like having a bucket of water: to have the reason for one’s life being saved taken into account as part of a fair procedure whose aim is to fix who is entitled to be saved is to have one’s bucket of water emptied. No further consideration is owed. A natural concern about the line of argument on offer for why the greater number ought to be saved is that it is sensitive to the natural pooling of individuals i.e., the fact that some are grouped together at one location. Let us say that A’s life being at stake is a consideration whose relevance to the question of who has a claim against the rescuer to be aided ends up being neutralized along with the relevance of B’s life to the determination of that question. C’s life ends up being the decisive relevant consideration that leads the rescuer to favor the course of action that results in B and C being aided. B and C benefit, therefore, from the good luck of being together at the same location, such that the rescuer ends up being in a position to save them both, which he does in response to the valid claim of each that he is saved. One might find it troubling that B and C end up being aided because they are lucky to have found themselves in a circumstance in which the rescuer can aid both if he aids one. It is not obvious, though, why this should be a troubling feature of the proposed procedure as long as those for whom the outcome of the procedure is not favorable do not constitute a permanent minority. A degree of luck in who ends up being benefited is in eliminable in this kind of case.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

Many will not, however, be persuaded by this response, on the grounds that, though a degree of luck in who is saved is ineliminable, what fairness demands is that each who the rescuer could save be give some chance of being saved. By giving each person some chance of being saved, the thought goes, we fully acknowledge the equal moral significance of each person. The difficult question, on some views, is whether everyone should be given the same chance (which would require flipping a coin in cases where there are only two possible directions to go in), or each should be given a proportional chance by holding a weighted lottery. The appeal of the lottery rests, I believe, on the following thought: If there were only one person whose life was at stake, and someone was in a position to rescue that person without himself having to make a serious sacrifice of any kind, the person in need of aid would have a claim against him to be rescued. The only thing that changes when there is more than one person in need of aid, and all are in a symmetrical position with respect to the rescuer is that each person’s equal claim against the rescuer to be saved cannot be satisfied, as there is not enough of the good in question to be distributed equally. The best that can be done in order to satisfy the requirements of egalitarian fairness (Otsuka’s term) is to resort to a second best solution, in which we fairly distribute chances of being saved, as a chance at aid, unlike aid in this case, is a divisible good. There are several points about this argument worth making, but here I will just note one. That is that it is not accidental that this kind of argument is usually cast in terms of each person having a claim, or right, to be saved, and what fairness requires is that equal claims or rights be satisfied or respected equally. Whether or not this claim about what fairness requires is defensible, the thought that each person has a claim or right to be saved in a situation where there are several at risk and not all can be saved is plainly suspect. To defend it, an argument is required to justify why it is that any individual in this type of situation is owed anything more than the importance to his of being saved be taken into account as part of a procedure that no one can reasonably reject for determining who, if anyone, has a claim against the rescuer to be saved. At the very least, what this shows is that the “egalitarian fairness” argument rests on more substantive premises than its presentation as “intuitive” suggests. Nothing here excludes, of course, the possibility of there being a direct argument on contractualist grounds for some kind of lottery procedure for settling cases like Rocks. The supposed contractualist argument for settling Rocks through some kind of lottery procedure goes like this: let’s compare two principles, one in which the many are saved and the one is left to die, and another in which each is given some chance (perhaps an equal chance) of being saved. The suggestion is that the person who is granted no chance of being saved by the first principle can reasonably reject that principle on the grounds that the second principle imposes on no one as great a burden as that which the first principle imposes on him. Scanlon rejects this line of argument, one that employs what he calls (following Derek Parfit) the “complaint model” of assessing the grounds for reasonably rejecting a principle. According to the complaint model, a principle is reasonably rejectable if no one would be as burdened by some alternative principle as the person most seriously

Contractualism on the Shoal of Aggregation


burdened by that principle. Flipping a coin in cases like Rocks, therefore, proves to be not reasonably rejectable, as no one would be justified in arguing that flipping a coin in such cases imposes a burden on him greater than that imposed on anyone by some alternative principle. The weakness of the complaint model lies in its failing to adequately capture the force of certain grounds for wanting to reasonably reject a principle. This is nicely illustrated by the Rocks case. Focusing just on the way in which different persons might stand to be burdened depending on what the principle is regulating conduct in that type of situation results in an important ground for reasonably rejecting a principle that tells the rescuer to just flip a coin to decide whom to save: that it authorizes the rescuer to reason in the same way whether or not the conflict is between one and one or one and two, and thus does not require that each life that is stake be acknowledged in the reasoning as to who has a claim to be saved. Such a principle is one that can be reasonably rejected, as each person whose life is stake has reason to want the rescuer to acknowledge his having a stake in how the rescuer proceeds. This objection is not one intuitively captured by the complaint model, which can only characterize it as a kind of serious burden. One response to this objection is that it may show that settling Rocks by a simple lottery is impermissible, but it is no objection to a weighted lottery procedure. In some ways, this might even seem to be the fairest procedure, because it gives each person whose life is at stake some chance of being saved. The reasoning here appeals to the thought that a weighted lottery has the virtue of recognizing each person’s life as favoring the course of action that will result in his life being saved and giving each person at least some chance of being saved. A procedure that gives each person at least some chance of being saved is one, presumably, that no one can reasonably reject as it imposes on no one the burden of having no chance of being saved. The alternatives, that of just saving the greater number or giving each exactly the same chance of being saved, are reasonably rejectable in its favor. The earlier contractualist argument for why each of the greater number have a claim to be saved relies on the thought that (a) each has been given the consideration due to his if his life’s being in peril is taken into account in deciding whom to save as a reason favoring the course of action necessary to save his life and (b) it is not decisive only when its force is neutralized by the another’s life being in peril counting in favor of an incompatible course of action. The argument for the weighted lottery suggests that this will not do: the neutralization of the force of the reason flowing from one’s life being in peril can be reasonably rejected in favor of the force of that reason being discounted through the lottery mechanism. Discounting is certainly not as bad for the one who ends up with at least some chance of being saved compared to the alternative of having no chance of being saved. Grant, for purposes of argument, both that it is not as bad to have some chance of being saved rather than no chance of being saved and that a principle giving each some chance would not be as bad for anyone as a principle that will lead, in Rocks, to someone having no chance of being saved. There is still a question as to what the bearing is of these considerations on reasonable rejectability of a principle for the


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

regulation of this type of situation. If the complaint model is assumed, their relevance is clear: what they show is that a principle directing the rescuer to decide via a weighted lottery is not as bad as any of the alternatives principles for the one who is most seriously burdened by it. Without this assumption, though, nothing immediately follows from the fact that a person who ends up with no chance of being saved would prefer (or it would be not as bad for him) that the rescuer decide how to proceed in a way that gives him some chance of being saved. To insist, therefore, that some kind of lottery procedure more fully recognizes the equal significance of each person’s life than one in which the reason to save a person’s life is decisive in the decision of whom to save unless its force is neutralized along with the force of an equivalent opposing reason is simply question begging. The problem that the lottery procedure encounters here is not surprising. Its true appeal, it seems to me, is in seeing it as a second-best solution to a problem: what matters to each is that he be saved, but because it is not possible to save each person, giving each a chance at being saved is arguably a fair solution to solving the problem of how to allocate the good of being saved. The importance to each of recognition, which has a natural home in an account like contractualism that attaches primacy not to outcomes but to the character of the relationship that holds between individuals, plays no role in this rationale for the lottery. It is not surprising, therefore, that the move from the proposed procedure to some kind of lottery procedure for deciding whom to save proves under motivated; its natural home is as part of an account that emphasizes the moral importance to individuals of securing certain kinds of outcomes, of a type to which contractualism, in emphasizing the importance of relationships rather than outcomes, stands opposed. This is not to say that the appeal of the lottery procedure for deciding cases such as Rocks is illusory. But it is to say that a defense of the lottery procedure for settling such cases needs to proceed in the context of defending a broader non-contractualist account of moral wrongness. It cannot be justified on contractualist grounds without downplaying or ignoring the importance of certain central aspects of the theory. 4. The claim of the last section is that if the rescuer heads towards the rock with one, rather than the one with many on it, it is not just that the rescuer has done wrong in an impersonal sense. Rather, he has wronged each of those who constitute “the many,” insofar as he would have owed it to each to save his had he complied with the procedure. There are at least two ways, though, in which it could be argued that the view on offer fails to do justice to certain central intuitions that cases like Rocks elicit. First, it seems to be an implication of this view that if the rescuer disregards the appropriate procedure for determining who has a claim to saved in this kind of case, saving the one rather than the many, he ends up, at least in one respect, wronging the one saved. That will strike many as implausible; what does the one have to complain about, after all? He’s been saved! That this is an implication of the present view is undeniable; that it is an implausible implication is arguable. Intuitively, what is implausible is that the person who is saved would actually complain to the rescuer that he has been wronged in being rescued. But because a person does not press a complaint, perhaps for good reason it

Contractualism on the Shoal of Aggregation


does not follow that the person does not have legitimate grounds for complaint. It matters to him, after all, not just that he be saved, but that his being saved is consistent with the recognition in the decision of who is to be saved of the legitimate interest in being saved of all those whose lives are at stake in the situation. The value of the procedure discussed earlier lies in its providing a way to understand how saving the greater number can be consistent with this kind of recognition of each person’s life; it is not simply instrumental. In this respect, saving the one rather than the many in the type of case at hand is relevantly similar to the flaunting of a fair hiring procedure by relying on nepotistic considerations. In that case, the person employed can (but may well not) complain of having been cheated out of the recognition of having won a fair competition, as what matters is not just the outcome. Likewise, in our rescue case, the procedure that leads to a person being saved has a significance for that person that is not exhausted by the importance for him of the outcome. Second, as Frances Kamm points out, “as the number of people in the larger group who outnumber those in the smaller group increases, there is a greater wrong done if we do not save the greater number.” Whether this intuition is in any way in tension with the present view depends on how it is understood. In one way, it is of course true that as the number that constitute the larger group left behind grows, there is a greater wrong done, but that is because there are simply more people wronged. A better way to understand the intuition to which Kamm draws attention, however, requires distinguishing questions of culpability, having to do with the seriousness of a person’s wrongdoing, or the blameworthiness of the wrong doing, from questions of liability, having to do with the costs or burdens a person may be asked to bear as a consequence of his wrongdoing. Is Kamm’s intuition best understood as focused on culpability or liability? Does the nature of the infraction change, such that the wrongdoing becomes a more serious kind of wrongdoing as the number rises of those whose interest in being saved has not been properly taken into account rises? Or is it that a person has more to answer for as a consequence of his wrongdoing as the number of persons wronged increases, such that the seriousness of the kind of wrongdoing remains the same, but the specific instance of wrongdoing becomes a more dreadful example of its kind? The proposed analysis is fully compatible with the second possibility, but not with the first. This is so because the contractualist analysis of wronging is primarily concerned with questions of culpability; the kinds of burdens that a person may justifiably be asked to bear as a consequence of his wrongdoing is a further substantive question. Contractualist resources can, therefore, be usefully deployed to clarify the basis of the kind of intuition Kamm has noted (though perhaps not in a way that Kamm would find acceptable). A more basic objection to the view, one that is thought by both Kamm and Parfit to constitute a conclusive reason for rejecting the individualist restriction, goes as follows: imagine we are faced with a choice of saving A, enabling him to live another forty years, or saving B and C, enabling each of them to live another two years. According to Kamm and Parfit, only A has a claim to be saved; the greater benefit we can secure for


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

A compared to what can be secured for either B or C is a reason for treating this as a case in which there is no conflict as to who has a claim to be saved. But this conclusion is less plausible as we add individuals to the side of B and C. If the choice is between benefiting A a great deal or benefiting each of B-Z a little bit, we should, they claim, benefit B–Z. The only way to make sense of this intuition, the argument goes, is to accept that the combined benefit that will be secured for each of B–Z is a stronger reason to save them than the reason to save A. A fortiori, the individualist restriction commits the contractualist account to implausible implications. Assessing the force of this objection is complicated, partly because, as stated, it is misleading. Whether or not there is stronger reason to benefit B–Z is, for purposes of challenging the plausibility of the individualist restriction, neither here nor there. It might be true, but the relevant question is whether B–Z all stand to be wronged if A, rather than each of them, is benefited because being benefited is owed to each of B–Z?  My sense is that they would not be. But that does not show that there is no reason to benefit B–Z rather than A, nor does it show that there is not decisive reason to do what wrongs A and benefit B–Z. Whether there is turns on the difficult general question about the relation between the values (and the reasons that flow from them) to which that aspect of morality concerned with what we owe to one another answers and other values and reasons that can plausibly be classified as moral. For this reason, the case in question is not one I see as challenging the individualist restriction, as the intuition to which it appeals draws attention to moral reasons other than those whose basis contractualism aims to illumine. Finally, in closing this section, it is worth noting that one of the points that those skeptical of contractualist resources being able to help us understand why the greater number are to be saved in this simple type of case is that Scanlon’s own treatment of the one versus many case appears to appeal to the combined force of the individual claims to be saved in order to justify to the one not saved why the rescuer’s duty is to save the greater number. This requires what Scanlon takes to be a modest, but admits is not uncontroversial, assumption that, in the context of deliberating about saving lives, reasons can be added together. Critics have charged that this move is not in keeping with the spirit of the individualist restriction and the particularly strong form of moral individualism it expresses. Whatever the merits of this charge, the view on offer here in no way relies on combining the force of reasons, so does not need to make the kind of axiological assumption required by Scanlon’s argument. 5. Rocks is an example of a type of case involving the fair allocation of an indivisible good and several individuals each of whom stands to benefit from it in the same way and to the same extent. I have argued that, in this type of case, there is a contractualist justification for doing what will result in the greater number being benefitted. But, if convincing, all the argument shows is that contractualist resources can be marshaled to make sense of intuitive relevance in one type of case. The number of persons at stake is intuitively relevant, however, in more than just this type of case. Recall, for instance, a case mentioned earlier concerning the choice of locating a chemical research facility in a large urban center or in a sparsely populated rural area:

Contractualism on the Shoal of Aggregation


intuitively, the facility ought to be located in the sparsely populated rural area, though the consequences of an accident are much worse for any of the local population than it is for any of those living in the urban center. Structurally, however, this type of case is different from that of Rocks: in that case, the implications for each of the rescuer’s choice are the same. But in this case, the potential burdens are not equal. And intuitively, it seems that it is the greater number, each stand to have to bear the lesser, though comparatively morally relevant, burden that can complain of being wronged if exposed to the risk that the chemical research facility creates. Though a detailed exploration of how to account for intuitions about this type of case using contractualist resources goes beyond what I can undertake here, it is worth briefly considering what kinds of resources might be marshaled for the task. First, it is relevant that the above chemical factory case is one that involves the imposition of a risk of harm rather than the imposition of harm. Those living in the rural area may, therefore, stand to suffer a greater harm if the risk materializes as harm yet be less likely to suffer harm than those living in the urban setting. This may have to do in part with a sense that those living in the rural rather than the urban setting have a better opportunity to avoid being harmed (it being easier, for instance, for them to get out of harm’s way). Second, though those living in the urban center will not each be as seriously harmed by an accident at the chemical research facility, the significance of their being harmed may not be exhausted by the harm each incurs. The proper functioning of a large urban center could be considered an irreducible, non-exclusive public good: one that cannot in principle be realized without a great many people doing their part and contributes to the lives of both those who participate in realizing it and those who do not. Appealing to the possible implications of a proposed principle for the availability or accessibility of a public good, which may supervene on the fact that a great number of individuals stand to be harmed, does not violate the individualist restriction in the way appealing to the mere fact of the number who stand to be harmed clearly does. Such a strategy might be used to understand the basis of an intuition to which Scanlon draws attention in his discussion of aggregation, that it might be wrong to save a life rather than saving millions from paralysis or blindness. Depending on how the details of the case are filled out, however, a further, third, strategy is available. Say that the one will be saved or millions saved from paralysis or blindness depending on how certain public resources are deployed by state officials. A case might be made on grounds of the duty of the state to concern itself with public health for state resources being deployed in a way that balances the importance to those subject to the state’s authority of having the resource in question deployed in a certain way with the number of citizens who will be benefitted from the resource being deployed in this way. A principle governing state action that is sensitive to the number of persons who stand to be affected is compatible with the individualist restriction as long as the justification of the principle relies not the fact that a great many stand to be benefited, but on an account that connects the interests of a citizen with the state attending to public health concerns in a way that is sensitive to the numbers of persons in question.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

6. At the heart of deontological or non-consequentialist ethical theory, as Christine Korsgaard puts it, is the thought that the “subject matter of morality is not what we should bring about, but how we should relate to one another.” In this respect, Scanlon’s contractualism is avowedly deontological, emphasizing the importance to the morally motivated of standing in a particular type of relationship to all others, one of mutual recognition. It is one thing, however, to note the general relational character of the contractualist outlook but quite another to keep the importance of the point in view when working through substantive moral questions in a contractualist mode. This is particularly true, I believe, in the case of the discussion of how to make sense, on contractualist terms, of the intuitive relevance of aggregative considerations. In thinking about the relevant kinds of cases, we are naturally drawn to focusing on the possible outcomes of different choices, and from there it is easy to find oneself thinking that what is of central moral importance is what outcome obtains. What I have suggested in this essay is that many of the challenges to the plausibility of the contractualist account in its current form that the intuitive relevance of aggregative considerations are thought to present are rooted in attaching an importance to facts about outcomes that is at odds with contractualist thinking. Careful attention to the relational character of the contractualist account, I have argued, allows us to both make better sense of contractualism’s rejection of the relevance of aggregative considerations and start to see how the seeming relevance of such considerations in certain cases can be explained using contractualist resources. This later task remains crucially incomplete; here, I have only looked in detail at how the seeming relevance of the numbers of persons at stake is to be explained in one type of case, only briefly gesturing at how their seeming relevance might be accounted for in other types of case. But hopefully enough has been said to show that it is far from obvious that worries about aggregation are as damaging to the plausibility of the contractualist account as they are often taken to be. notes 1. Parfit distinguishes between two aspects of what I am here calling the individualist restriction: that which limits relevant implications to those that can be predicated of an individual’s life and that which rules out appeals to impersonal values. In this essay, I take myself to be arguing for a justification for both aspects of the individualist restriction. 2. Michael Otsuka, “Saving Lives, Moral Theory, and the Claims of Individuals,” Philosophy & Public Affairs 34 (2006): 130. The classic discussion of the anti-majoritarian import of a unanimity requirement is Thomas Nagel’s essay “Equality,” reprinted in Thomas Nagel, Mortal Questions (Cambridge: Cambridge University Press, 1979). 3. Derek Parfit, On What Matters, Vol. 1. (draft of summer 2008). 4. T. M. Scanlon, What We Owe To Each Other. (Cambridge, Mass.: Harvard University Press, 1998), 105. 5. Joseph Raz suggests something similar: “Contractualism is not so much an account of moral reasons, as an account of moral aggregation. It is a thesis about the ways in which numbers do and do not determine the outcome of moral conflicts.” Joseph Raz, “Numbers, with and without Contractualism” Ratio 16 (2003): 346–67.

Contractualism on the Shoal of Aggregation


6. Scanlon, What We Owe, 105. 7. Something roughly along these lines is suggested by Derek Parfit, On What Matters, Vol. 1 (Oxford: Oxford University Press, forthcoming). 8. Scanlon, What We Owe, 162. 9. T. M. Scanlon, Moral Dimensions: Permissibility, Meaning, and Blame (Cambridge, Mass.: Harvard University Press, 2008), 139–41. 10. As Scanlon notes, “We would, ideally, like our moral relationship with others to be mutual. This relationship is fully realized when we are moved to act in a way that is justifiable others and this concern is also reciprocated” (Scanlon, Moral Dimensions, 142). 11. See, for instance, Jay Wallace’s contribution to this volume. Wallace forcefully challenges the idea that sharing a common property of being rational, or having the capacity to be guided by reasons, could serve as the basis for a relationship with another. Though I cannot pursue the matter here, it seems to me that a proper assessment of Wallace’s challenge requires consideration of whether or not sharing the property of being rational with another is something that one can appropriately valued as the basis of a relationship in the way, say, being a co-national might be properly valued. See also Samuel Scheffler’s probing discussion of the issue of a relationship-based morality in his “Morality and Reasonable Partiality,” in Equality and Tradition: Questions of Value in Moral and Political Theory (Oxford: Oxford University Press, 2010), 41–75. 12. Scanlon, What We Owe, 163. 13. Though the matter requires a fuller treatment than can be pursued here, I do not see this point as being at all in tension with Scanlon’s reasons for holding that attitudes are not non-derivatively relevant for permissibility. It would be if the only thing that mattered to the morally motivated person on the contractualist account is conducting himself within the boundaries of the permissible, but the suggestion here is that, in caring about the character of the relationship between him and others, moral concern goes well beyond a concern to be in mere conformity to the boundaries of the morally permissible. 14. Philippa Foot, “Rationality and the Virtues,” in Moral Dilemmas (Oxford: Oxford University Press, 2002), 168. 15. Nagel, Mortal Questions, 68. 16. Nothing here precludes recognizing the existence of “collective persons” such as corporations. But these are artificial persons, constructed and recognized (legally and morally) to the extent that they serve the interests of individuals. The interests of a collective person can thus arguably always be cast, for purposes of thinking about what it is morally permissible to do, in terms of the interests of the individuals served by collective person construction. 17. On the contractualist account, the morally motivated will of course be concerned not only to avoid acting impermissibly, but to avoid blameworthy conduct (a concern not just with the permissibility but also with the meaning of one’s conduct). 18. Assume no special ties, no permanent or standing minorities, etc. 19. The suggestion here is not that the rescuer needs to actually engage in any kind of deliberation concerning how to proceed, and certainly not that the rescuer needs to work through the procedure defended in this section. The point of the procedure, rather, is to clarify why it is justifiable to the one not saved on grounds no one can reasonably reject that that the balance of reasons favors saving the many. 20. As Kamm and Scanlon both point out, however, this is not a constraint without teeth, as it identifies simply flipping a coin in a case like Rocks as one that can be reasonably rejected on the grounds that the simple coin toss procedure is not one consistent with taking into account the importance to each person of being aided. Otsuka criticizes this argument,


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

arguing that if we toss a coin in the relevant type of conflict cases, it of course makes a difference to what is done whether it is a 1 versus 1 case or a 1 versus many case. But this seriously misconstrues the Kamm/Scanlon argument. The objection to the simple lottery is not that the coin toss makes no difference to what is done in the 1 versus 1 as compared to 1 versus many case; it is that the reasoning in virtue of which a particular course of action is justified is the same whether or not the case is a 1 versus 1 or a 1 versus many case. The problem with the simple coin toss, that is, is that it does not require taking into account each person’s interest in being aided in settling the question of who has a claim to aid. Otsuka does not acknowledge this objection. 21. It is important here that recognizing the value to a person of his life being saved be counted in favor of a course of action that is likely to result in his life being saved. The choice that the rescuer faces, therefore, is between courses of action, rather than potential outcomes. This preserves the non-consequentialist emphasis on the importance of what a person does. 22. One might think that something like a utilitarian principle—one that requires the rescuer to always save the greater number—could still be justified on contractualist grounds. One such argument for this conclusion takes seriously the importance of the fact that what is sought is principle for the regulation of the rescuer’s conduct in cases such as Rocks. This consideration, together with the assumption of no standing minorities, suggests that the right perspective to take in thinking about which principle is justifiable to each person is an ex ante perspective: the relevant principle should be thought of as the one that any individual has reason to want the rescuer to comply with assuming he does not know whether he is or is not part of the greater number. If we make the reasonable assumption that any given individual is more likely to be part of the group consisting of the greater rather than the smaller number who can be saved, we have a ready justification for a principle requiring the rescuer to pursue the course of action that will result in the greater number of lives being saved. The problem with this argument is that it offers no grounds for thinking that the fact that what is sought is a principle is a reason for why one who stands to be left if the rescuer is guided by a “save the greater number” principle could not reasonably reject this principle in favor of one that gave him at least some chance, and perhaps an equal chance, of being saved. Though there is a sense in which the fact that what is sought is a principle introduces a kind of ex ante perspective into the assessment of whether or not a principle can be reasonably rejected, it is not the kind of ex ante perspective required by this kind of argument. For more discussion of the connection between principles and an ex ante perspective, see my “Defending the Moral Moderate: Contractualism and Commonsense,” Philosophy & Public Affairs 28(4) (1999): 275–309. 23. Such a procedure may result in the same course of conduct that would have resulted had the rescuer simply aimed to do his best to comply with the reasons that apply to him, but rescuer’s justification will be a distinctly moral justification, insofar as it is one that explicitly concerns the justifiability to each person of reasoning about appropriate conduct. 24. In “Can a Nonconsequentialist Count Lives?” Philosophy & Public Affairs 31(1) (2003): 71–94, David Wasserman and Alan Strudler criticize arguments of the kind I am offering on the grounds that relevant considerations should be “weighed” against one another (91). In emphasizing the importance of symmetrical claims, I mean to the reject the use of the weighing metaphor. As the argument developed in the next paragraphs in no way relies on the “weight” metaphor, it is, I hope, not vulnerable to worries about lives being of infinite weight or incommensurable. Cf. S. Matthew Liao, “Who Is Afraid of Numbers?” Utilitas 20(4): 447–61. 25. This section revisits, and I take to be an improvement over, a similar argument I offered in “Contractualism on Saving the Many” Analysis 61(270) (2001): 165–70

Contractualism on the Shoal of Aggregation


26. Otsuka (in “Saving Lives”), criticizing an argument I advanced in an earlier essay on saving the greater number, argues that my position commits me to the view that in cases like that of A and B, it is an implication of my position that the rescuer has no reason to save anyone. An argument of this form is also advanced by Tyler Doggett, “What is Wrong with Kamm and Scanlon’s Argument against Taurek,” Journal of Ethics and Social Philosophy 3(3) (2009): 1–15. I agree the implication is absurd, but it is neither an implication of the position advanced in the early essay or this one, for the reason made explicit in this paragraph. 27. See Scanlon’s discussion of the rationality of decision. “Reasons: A Puzzling Duality?” in Reason and Value: Themes from the Moral Philosophy of Joseph Raz, ed. R. Jay Wallace, S. Scheffler, P. Pettit, M. Smith (Oxford: Oxford University Press, 2006): 231–46. 28. G. E. M. Anscombe. “Who Is Wronged?” in Absolutism and its Consequentialist Critics, ed. Joram Haber (Lanhan: Rowman & Littlefield, 1994), 159–60. 29. The value of the coin toss is to be understood, therefore, as a suitably impartial mechanism; its value as a device for distributing chances is not relevant. The difference between the two interpretations of the significance of a coin-toss is nicely illustrated by the following thought experiment, put forward by Kornhauser and Sager: Imagine a lottery is held in which all the participants believe the outcome to be completely random, such that all believe they had an equal chance of winning. In fact, the lottery was not random but rather was biased towards a particular outcome, though the bias has not been introduced by anyone, i.e., it is due to a mechanical failure. On the first interpretation, the outcome of the lottery is not a legitimate basis for deciding who should benefit. On the second interpretation, the outcome of the lottery is still a valid basis for deciding whom to benefit. For on the second interpretation, what matters is not that everyone gets an equal chance at being benefited, but that the decision not be tainted by any objectionable biases on the part of the benefactor. See L. Kornhauser and L. Sager, “Just Lotteries,” Social Science Information 27(4) (1988): 483–516. 30. Notice that though there are more considerations that favor one course over the other, this does not immediately establish that there is a stronger case for going in one direction or another. To presume this is to presume the additivity of reasons. As will become clear, there is no reason here to do so. For relevant considerations concerning additivity, see Shelly Kagan, “The Additive Fallacy,” Ethics (99) (1988): 5–31. 31. That method, as Nagel originally presents it (the “least unacceptable to the person to whom it is most unacceptable” rule for evaluating the burdens imposed on individuals by a principle) invites us to look at the potential outcome of a principle for each person who stands to be benefited or burdened by a particular principle. Though Scanlon’s 1982 presentation of contractualism in “Contractualism and Utilitarianism” suggests something like Nagel’s understanding of pairwise comparison as the way to evaluate the reasonable rejectability of a principle, it is not, I believe, the right way to understand evaluating the reasonable rejectability of a principle. To the extent contractualism, as I understand it, incorporates pairwise comparison (in Nagel’s sense), it is in its adherence to the individualist restriction. 32. Thanks to Andy Brook for the analogy. 33. An example of a permanent minority would, for example, be the black population of a dominantly Caucasian voting district, in which voting patterns repeatedly divide along the color divide. 34. Otsuka, “Saving Lives,” 113. 35. Insightful discussions of which is preferable are John Broome, “Kamm on Fairness,” Philosophy and Phenomenological Research 58 (1998): 955–61 and Iwao Hirose, “Weighted Lotteries in Life and Death Cases,” Ratio 20(1) (2007): 45–56.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

36. Scanlon explicitly rejects this possibility in his discussion of aggregation for reasons that are more or less those I offer in the main text. 37. Otsuka (in “Saving Lives”) asks “what grounds have Kamm and Scanlon to insist that the adjustment must be so radical—i.e., that in comparison with the two-rock case the first person’s chance of being saved must plummet from 50 percent to zero, while the second person’s skyrockets from 50 percent to 100 percent, which will also be the third person’s chance of being saved?” (117). Scanlon, at least, has no justification because it is only a version of contractualism presupposing the complaint model that in any way suggests that respect for the value of a person’s life is a matter of assigning chances. Otsuka does not explicitly invoke the complaint model, as he presents the thought that respect for the value of a person’s life in conflict cases like Rocks requires giving them some chance of being saved as a raw intuition that is not defended. At least invoking the complaint model provides some broader theoretical reason having to do with the norms of interpersonal respect for persons for accepting that intuition. 38. Thanks to Dave Langlois for helpful discussion of this point. 39. F. M. Kamm, “Owing, Justifying, and Rejecting,” in Intricate Ethics: Rights, Responsibilities, and Permissible Harm (Oxford: Oxford University Press, 2007), 480. Kamm suggests a way in which Scanlon’s proposal could be used to illumine this judgment; the assessment of the plausibility of that proposal is a matter that I will set aside. 40. Could it be that Kamm and Parfit are both attributing to contractualism some version of the complaint model? If so, this objection awaits some defense of why it is that contractualism is either committed or is most plausible if it incorporates the complaint model. Whether Scanlon suggested something like it in Scanlon (in “Contractualism and Utilitarianism”) strikes me as neither here nor there—what matters is what the most plausible version of the view looks like. 41. Thanks to Aaron James for helpful discussion of this point. 42. On this point, see Michael Otsuka, “Scanlon and the Claims of the Many Versus the One,” Analysis 60 (2000): 288–93. 43. What We Owe, 240. How much weight should be put on this intuition is unclear: those who are not blind or disabled systematically underestimate the quality of life available to those are. 44. “To later generations, much of the moral philosophy of the twentieth century will look like a struggle to escape from utilitarianism. We seem to succeed in disproving one utilitarian doctrine, only to find ourselves caught in the grip of another. I believe that this is because a basic feature of the consequentialist outlook still pervades and distorts our thinking: the view that the business of morality is to bring something about . . . [t]he subject matter of morality is not what we should bring about, but how we should relate to one another.” (Christine Korsgaard, “The Reasons We Can Share: An Attack on the Distinction between Agent-Relative and Agent-Neutral Values,” Social Philosophy and Policy 10[1993]: 24). 45. I am indebted to more people over the years than I can recall for helpful discussions of the issues discussed in this essay. Recently, I have particularly benefitted from comments and discussion with Michael Gibb, Aaron James, Dave Langlois, Alan Strudler, Nien-hê Shieh, and to audiences at York University and Carleton University for many good questions.

7 Immoral, Conflicting, and Redundant Promises Seana Valentine Shiffrin

thomas scanlon’s creative and stimulating work on promising figures prominently among the many valuable contributions he has made to recent moral and political philosophy. His work has served as a major source of inspiration for the lively discussion about promising that has flourished over the last two decades. I agree wholeheartedly with Scanlon’s nonconventionalism about promising but disagree with him about the content of the moral principles that underlie the promissory commitment. Here, I propose to pursue that disagreement by exploring a variety of neglected cases of promising that highlight some of the differences between expectation-oriented views of promissory commitment like Scanlon’s and rights-transfer views like my own. Scanlon’s theory posits that the purpose of promising is to give assurance about the promisor’s assertions of future performance. On Scanlon’s view, promises give assurance by creating expectations of performance in the promisee through the articulation of the promisor’s intention to perform when this intention is transparently articulated for the purpose of assurance; promises are binding because it is wrong intentionally and voluntarily to induce another’s expectations that one will act a certain way in order to assure that person but then to disappoint those expectations. Scanlon’s theory has faced criticism on two main fronts. First, some of us doubt that the aim to assure is as ubiquitous and as essential a motive for promising as he posits. For related reasons, we deny that the existence and the binding force of promises hinges primarily upon the expectations of the promisee. Second, some worry that expectation-based views of the moral force of promises confront a circularity—they are unable to explain how the articulation of a promise can, alone, without reliance on a convention, give rise to the sort of legitimate expectations moral agents should then respect. 155


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

The rather different idea that a binding promise involves a transfer of a right from one party to another to decide whether to act in a certain way serves as a nonconventionalist alternative to Scanlon’s expectation-based view. A rights-transfer theory does not make assurance or expectations central to the theory of promising, although it acknowledges that the aim to assure is common and important, and further, that morally significant expectations often arise in the promisee because of a promise. One variation of the rights-transfer theory, the existence view, holds that it is constitutive of a promise that it involves the valid transfer of a right from the promisor to the promisee. A distinct variation, the bindingness view, recognizes some speech acts as creating promises even if they do not involve a valid rights-transfer, but holds that a promise is binding only if it is created through a valid rights-transfer. Because of its relative simplicity and its greater plausibility to me, I will focus on the existence view. Others, who believe immoral promises are true promises but without moral force, may find that the bindingness view better fits their intuitions. On either variation, the salient feature of the promise—the rights transfer—may occur independently of whether the relevant promissory utterance persuades or should reasonably persuade the promisee of future performance. Hence, rights-transfer views avoid the difficulty that expectation-based (and reliance-based) views face in explaining why the promisee was justified in developing the expectation that the promisor would act as she said she would. Indeed, rights-transfer views have the strength that they can explain how a promise provides justification or warrant for a promisee’s developing an expectation of performance. Further, it seems as though many promises that do not generate expectations of performance may nevertheless be binding—as with promises to repay a creditor who believes the promisor will renege. The rights-transfer theory can explain, without circularity, why a promisor must perform without appealing to the wrong of disappointing expectations or of squandering others’ investments: it is because she has transferred her right to decide to act otherwise to someone else and so has no moral right to make a fresh choice at the present. The rights-transfer theory, however, faces some difficulties of its own that arise when one considers the rather neglected examples of immoral and conflicting promises, and, in particular, the virtually orphaned but fascinating case of redundant promises. I plan here to identify and explain the problems these cases pose, to flesh out a version of the rights-transfer theory in more detail, and to investigate what resources the theory has to address these problems. These resources come into sharper focus, I submit, when the formulation of the rights-transfer theory is further refined. Rather than conceiving of the right that is transferred in a broad way (as the right to decide to act otherwise), I will argue it is appropriate and illuminating to begin by conceiving of the relevant right as the right, between the promisor and the promisee, to be the one to decide whether to act, how to act, and on what grounds. As I will clarify, a promise may, in different circumstances and on different occasions, involve a more precise commitment to transfer a right to decide whether to act for a particular sort of reason. The cases I will discuss will help motivate why these refinements are necessary and significant.

Immoral, Conflicting, and Redundant Promises


In addition to attempting to flesh out the rights-transfer theory and to take it through some paces, I aim to make two other contributions: First, I hope to stimulate interest in these under-discussed categories of promising for their own sake. However, although they are fascinating cases, I hasten to acknowledge that they may be difficult cases that even plausible theories cannot dispatch with ease. Second, my effort to untangle the puzzle of redundant promises suggests a different, but potentially important, notion of partiality than usually occupies the spotlight in discussions of interpersonal relations and of promising, in particular. Whether this notion figures in the correct account of redundant promises or not, it is an aspect of partiality that deserves greater attention.

1. introduction to the problems

Although I will refine the formulation as the essay proceeds, the gist of the rightstransfer theory may first be grasped by putting it in its simplest, broad form. Put broadly, the rights-transfer theory claims that a binding promise effects a transfer from the promisor to the promisee, through the communicated intent to do so, of the promisor’s right to decide to act otherwise than how she promised to act. On the version of the theory I will explore, the existence view, promises themselves are constituted and made binding by these transfers. The transfer explains why the promisor is bound to perform: he no longer has a (moral) right to decide whether to do otherwise. It further explains why the promisee has the power to waive performance: the moral permission of the promisor to decide to act otherwise now lies in the hands of the promisee and the promisee may exercise this permission. The theory thereby explains the basic structure of the promissory relationship with satisfying simplicity. But this theory might be thought to be embarrassed by the three following sorts of cases: immoral promises, conflicting promises, and redundant promises. The theory seems to presuppose that the promisor had a right to perform or not to perform the promised action. So, first, how can it make sense of promises between criminal conspirators to perform criminal actions? How can it be that Mary the bank-robber can promise Joe, her conspirator, to neutralize the bank guard if Mary has no moral power to manhandle or murder the bank guard in the first place? Is Mary’s communication not a promise at all? Second, what about cases in which the activity promised is morally innocuous but one already has a duty not to perform it for some more contingent reason? If Alice promises to meet Bob for lunch and then subsequently promises to meet Carly for lunch at the same time, how can the rights-theory make sense of her second promise? It might be thought that Alice no longer has the right to decide whether or not to meet Bob for lunch; she is committed. So what right could she transfer to Carly? Is this like or unlike the criminal case? Finally, how does the rights-transfer theory account for redundant promises, that is, promises to perform an act that the promisor is already obliged, and knows he is


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

already obliged, to perform for other reasons? Suppose Karen knows that Roger—her dashing bounder of a friend—occasionally steals money from her purse when she leaves the room. She reminds him that this is outrageous behavior and asks him to promise not to do it again. Roger’s promise not to do so seems to make perfect sense even though he lacks any right to do otherwise. How can the rights-transfer theory of promises make sense of what seems like a perfectly intuitive judgment that standard promises are involved in each of these three cases—those of immoral promises, conflicting promises, and redundant promises? Further, how can it do so while also vindicating what seem (to me) to be the appropriate judgments about bindingness: the immoral “promise” is not binding, but the conflicting and redundant promises do exert moral force? Notice that an assurance theory handles these cases fairly easily, at least on the surface. Through their promises, the criminal gives assurance to the conspirator, the friend to her lunch dates, the petty thief to his acquaintance. The recipients now expect performance, in part, on the basis of these assurances. There are costs, however, to such smooth sailing. The assurance theorist then has to relinquish the idea that promises are pro tanto binding if he wishes to avoid the unpleasant consequence that the conspirator has an obligation or greater reason to perform an immoral action because one has promised to do it. This relinquishment may be thought, by some, to represent a strength and, by others, a weakness. And, having relinquished this idea, the assurance theorist must explain why some of these problem promises have moral force (e.g., the conflicting promises) while others do not. With respect to redundant promises, some may be inclined to think this assurance is not worth much. If one doubts the (known) duty will be performed in the first place, why should one place one’s trust in the assurance provided by a promise? Such skepticism is the point of impeaching witnesses by revealing their prior crimes. Jurors are supposed to think that because a person committed a theft and thereby showed a willingness to violate a known duty, their oath to give truthful testimony is unreliable. On the other hand, if one thought that there were different sorts of moral transgressions as well as different causes and related character traits giving rise to transgressions, the fact a person has acted badly in one respect (or is prone to do so) might not necessarily support an inference that she is entirely unreliable in other respects. Hence, the assurance provided by the promise might nonetheless be meaningful and assuring. In any case, regarding immoral, conflicting, and redundant promises as true promises seems quite intuitive. In particular, it seems intuitive that each conflicting promise is binding and that a wrong is done to either one of the parties who is stood up. Further (though it is a harder case), it seems as though the redundant promises exert some moral force. The assurance theory seems rather handily to accommodate this assessment and to make sense of why such promises might be made. Although I do not plan to vindicate each of these intuitive assessments, my task is to sketch a rightstransfer account of these cases that either fits the phenomenology of our judgments about them or, in the case of immoral promises, gives a convincing account of how the phenomenology might reasonably mislead.

Immoral, Conflicting, and Redundant Promises


2. immoral promises

I will begin with promises to engage in immoral conduct. I have in mind, as will be evident, conduct that is intrinsically immoral and that one has no autonomy right to perform, for example, theft, assault, murder, or fraud. To distinguish: some conduct may be all-things-considered immoral in a particular situation but not intrinsically immoral, for example, refusing to swim in one’s clothes may be generally permissible but immoral when a young child’s life is at stake. Other conduct may be intrinsically immoral yet fall within the scope of what one has an autonomy right to do, for example, failing to apologize when one should or acting stingily with one’s needy friend. Although such conduct is wrong, it is important to the meaning or significance of the performance of the corresponding moral conduct that there is an autonomy right to forbear. The sorts of promises I have in mind when I write of “immoral promises” are like those of a criminal, for instance, who promises a conspirator to rob a bank or to drive the getaway car. These may appear to be bona fide promises. We would not be surprised if her conspirator were angry if she failed to perform them. This phenomenon does not submit to a straightforward explanation on the rights account. The criminal has no relevant moral right to rob the bank or drive the getaway car, or to decide whether or not to do so; ipso facto, she has no relevant right to transfer to her conspirator through a promise. Indeed, I think the rights-transfer account should say that the criminal has not truly and successfully made a promise. Although she has presented something that wears the shell of a valid promise’s form, it lacks its substance. Although we often speak of promises between conspirators, perhaps it is not an overly counterintuitive form of revisionism to suggest that this is shorthand for something else. Suppose the criminal offers what appears to be a promise to rob the bank. Does she have any more reason to rob the bank than she did before she attempted to promise to do so? No. Her declaring that she has promised to rob the bank does not alter her moral status with respect to robbing the bank at all. It is not merely that she has some reason to rob it in virtue of the promise, but still a greater reason not to. Her utterance does not create any valid justificatory reason whatsoever, not even a pro tanto one, to rob the bank. The contrary view—that a true promise has been generated, providing the promisor with some reason to rob the bank, but one outweighed by the evil of robbing the bank—encounters the difficulty that, with respect to more minor evils than robbery, the evil of the promised act may not in fact outweigh the importance of fidelity. This view may thus have to acknowledge moral obligations to pilfer petty cash just because one promised to do so. This seems to me absurd. (Or are we really to say that the importance of property always outweighs the importance of fidelity?) One cannot, by declaration, bootstrap oneself into a moral obligation to perform an intrinsically immoral activity. I concede that it can be linguistically appropriate and convenient to say that the conspirators promised one another to rob the bank or to retaliate against a snitch. Indeed, in this essay, I make use of this convenience. This use does not strike me as


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

dispositive evidence of whether they have made what I call “true,” “valid,” or “bona fide” promises. In other contexts, we assign the noun associated with successes to invalid attempts, sort of like an honorific concession prize. For instance, statutes that are invalidated as unconstitutional are still called “statutes,” and sometimes are referred to as “law,” even when they retain none of the legal force of valid statutes and are treated as void ab initio. The same holds true of contracts whose formation defects (including having immoral or illegal acts as their subject) render them void. Courts will treat them as invalid attempts at contracts and as though they were never made. Still, courts will refer to these invalid agreements as void contracts, although they have none of the force or standing of a true, valid contract. I regard immoral promises analogously. We may use the noun “promise” linguistically to acknowledge that a certain sort of attempt was made, one that fails because of its substantive content and thereby does not have the characteristic moral and practical significance of a true promise. Although I deny that immoral “promises” are true promises and that they give rise to reasons to perform the “promised” activity, I recognize that the attempted promise may well alter the moral relationship between the attempted promisor and promisee. A satisfactory theory of immoral promises should accommodate and explain this fact. For instance, even though the promisor does not have reason, by virtue of her promissory language, to perform the immoral act that, ex hypothesi, she has no moral power to perform, her conspirator may be angry if she does not perform. How can we make sense of this anger and the real feelings of betrayal occasioned by breaches of this sort? In many cases, both parties incorrectly believe that a binding promise has formed. Sometimes, this is because they do not recognize that the object of the attempted promise is immoral. In many cases, though, conspirators are not self-deceived about the moral status of their plans. They may, however, believe that even binding immoral promises can be formed. Perhaps our language of promising strongly encourages this idea because we lack a clear term for an attempted promise that is not formed or is not binding because of its immoral character. In such cases of false moral belief, the party’s failure to perform involves many of the same character flaws and morally defective motives as those implicated by intentional omissions where the failure to perform itself would be immoral. This may go a long way toward explaining why these failures to perform evoke resentment and anger, although the omission to perform, considered by itself, is not immoral. I am also tempted to analyze these promises as forms of misrepresentation. In saying “I promise to rob the bank,” Mary misrepresents what rights she has and that she may legitimately transfer. The wrong done to the recipient lies not in the failure to perform, but in the initial misrepresentation of what she was empowered to commit to. This assessment of the wrong is, I believe, entirely consistent with the conspirator’s knowledge that Mary lacks the moral power to act in these ways. By analogy, a lie is still a lie and it still morally offends even when the recipient is aware that it is a lie and even when the speaker knows the recipient knows it is a lie. Think about some of the transparent excuses students give for absences or late papers. Even if your students believe that given the complexity of modern families you might entertain the possibility that they have four grandmothers, they do not really think you will believe that

Immoral, Conflicting, and Redundant Promises


all four of them died in the same term, each on the eve of an essay deadline. Or, consider the transgression committed by an obvious perjurer on the stand who knows his testimony will be roundly contradicted. Lies may be told without the intent to even attempt to persuade: Sometimes the lie is offered because one must say something and one wishes to avoid revealing or acknowledging the truth even if the fact of one’s evasion is transparent. My suggestion in the case of the attempted immoral promise is that the recipient is wronged by the pseudo-promisor’s fraudulent misrepresentation that the action is in that person’s moral power to perform, but not by the failure to perform as such. Had the action been performed, such anger might still have been appropriate had the promisee privately changed her mind about the desirability of performance. One can imagine a reformed criminal exclaiming to her conspirator “What were you thinking when you promised me to kill the bank guard?? Were we both out of our minds?” Such outbursts are unlikely, of course, but then again some lies are desired by their recipients. In some circumstances, the desirability and the transparency of the lie may partly mitigate the wrong of the lie, but they do not entirely eradicate it. The bare misrepresentation may suffice to make it a wrong. Further, the misrepresentation may encourage the recipient to engage in wrongful conspiratorial behavior by supporting tendencies to rationalize the permissibility of such behavior or to downplay its wrongfulness. In addition, in this way and others, third parties may be put at greater risk in virtue of the misrepresentation. My analysis thus provides an error theory of the phenomenological sense that attempted immoral promises are true promises. It provides an alternative explanation of how their breach can generate anger and other reactive attitudes. As I will now argue, it can also answer the other main argument given for the position that these are in fact promises. Altham and Anscombe argue that conspirators’ immoral promises must be true promises or else one could not make sense of the following case: Leonard pays Jack money in exchange for Jack’s undertaking to kill Harry. If Jack reneges, Leonard may demand the return of his money. But as Altham argues, “If there was not really a promise, [Leonard] would have nothing to invoke in demanding his money back.” The existence of a bona fide promise, however, is not necessary to explain the basis for Leonard’s demand. Although Leonard has no right to demand performance or compensation for the failure to perform, Leonard may demand restitution on two grounds that do not presuppose the validity of Jack’s attempted promise: First, most simply, Leonard may demand restitution on the grounds that he has given money to Jack and received nothing valuable (or negotiated for) in return although it was obvious that Leonard had no intention of making a gift. Second, in those cases in which Leonard operates in moral ignorance of the wrongness of what was “promised,” Leonard may demand restitution on the grounds that he gave Jack money on the basis of Jack’s misrepresentation that Jack had the (moral) power to render a desired service in return. Neither explanation invokes the idea that Jack actually promised. Further, neither explanation treats Jack’s duty to return the money as resting on Leonard’s disappointed expectation that Jack would perform. This latter feature represents an


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

advantage over an account that appeals to a breach of promise or a failure to make good on an assurance given by Jack. What about more complex, nested cases such as the following? Aida and Bert form a conspiracy to rob a bank. They consider whether to do it Tuesday or Wednesday. They decide on Wednesday. Aida, in a burst of insecurity, asks Bert to promise that he will rob the bank on Wednesday with her, rather than on Tuesday on his own, thereby leaving Aida out of the deal. Bert promises not to jump the gun. We might be inclined to say that if Bert sees the light and decides against robbing the bank entirely then Bert has not wronged Aida at all. What should we think instead if Bert decides he stands to reap higher rewards by cutting Aida out of the arrangement and robbing the bank on Tuesday? Doesn’t Aida have a complaint based on the promise that Bert made about Wednesday in particular (and not just a complaint that Bert and Aida mutually misrepresented to each other whether they had powers to rob banks at all)? Or take a more quotidian case involving the same characters: Aida, who turns out to be Bert’s domestic partner, calls Bert and asks him to get $20 on the way home. Bert replies that he won’t be able to both get to ATM and go to the gym as he’d planned, and that his interests in exercise are as or more important than Aida’s need for the cash. Aida urges Bert to take the money from the office’s petty cash drawer and to replace it tomorrow. I assume for the purposes of the example that this is wrong to do but that Bert promises to do so. Bert then shows up at home empty-handed simply because he was forgetful and not because he morally reconsidered. Does Aida really have no complaint at all? How should the rights-transfer theorist account for what seem like reasonable complaints the putative promisees would have in these cases? I am of two minds about these cases. On the one hand, those who receive (and especially those who elicit) immoral promises really have no reasonable stake in their fulfillment at all, whatever the reason for the failure to perform. On the other hand, it seems that Bert has, in some sense, mistreated Aida and that Aida has some complaint. I do not, however, think that complaint arises from any duty created by the attempted promise. This may be partly brought out by the fact that the complaint may seem to disappear if Bert’s failure to perform was motivated by reconsideration of the morality of the activity rather than by greed or negligence. Whatever complaint she has, then, must in some sense be rooted in Bert’s reason for not performing. My suggestion is this: Bert’s agreement to rob the bank on Wednesday rather than Tuesday or to steal the petty cash is not a binding promise because it either directly or indirectly presupposes a right to do what Bert could have no right to do, whatever his reason. But, Bert’s agreement may have generated some duty of consideration toward Aida even if it did not transfer an entitlement to her to demand performance. Suppose Aida made clear her ultimate purposes: she wanted the money in the first case to pay her rent and in the second case to buy cookies from the next door neighbor’s child. One might think that in making the agreement and deliberately attempting to cultivate expectations in Aida that her legitimate interests would be promoted, Bert took on some responsibility for attending to the legitimate interests behind Aida’s request. In making the agreement, Bert committed to performing some action to further Aida’s interests and to not allowing certain sorts of considerations

Immoral, Conflicting, and Redundant Promises


(e.g., personal inconvenience or greed) to be grounds for ignoring them. When Bert treats Aida’s expectations negligently or recklessly, he has violated that duty of consideration, although Bert would not violate it if he omitted to steal because he was motivated by the wrongness of the action; that reason need not represent any failure of care or concern for Aida’s interests but would represent the proper accounting for them relative to the action they together contemplated. Aida and Bert together share responsibility for contemplating and agreeing to use an immoral action to further Aida’s interests; when either reconsiders on moral grounds, neither does wrong. But when Bert fails for reasons orthogonal to the appropriate valuation of Aida’s interests in the circumstances, for example, for greed or through negligence, Bert may have violated a duty of consideration engendered by the agreement. This is not to say—in any sense whatsoever—that Bert ought to have waited until Wednesday or taken the petty cash. It is a duty that perhaps could only be satisfied, compatibly with his other duties, through moral reconsideration. But, the judgment that failure to perform for poor reasons may violate a duty of consideration may explain a sense that there may be some residual duties of remedy that apply here. Bert may owe Aida an apology in cases where he failed from negligence or greed. To remedy the omissions for those reasons, perhaps Bert should offer Aida some legitimate funds to help with Aida’s rent or, if Bert realizes his negligence on the way to the gym, perhaps Bert should stop at the ATM and skip his workout. These remedial actions are not owed in virtue of any expectations Aida had to the money, for those expectations were intimately connected to and infected by the use of illegitimate means. Rather, in making the agreement or in representing himself as transferring a right, Bert implicitly assumed some relational duties not to show disregard for Aida’s interests in certain ways or for certain reasons; it is therefore appropriate, as a remedy for that disregard, to demonstrate a more active form of consideration for those interests (even though the remedy, as with many remedies, may involve doing more than Bert initially would have had to do had he complied with his duties in the first place).

3. conflicting promises

I have been contending that what appear to be promises to engage in inherently immoral behavior are better understood as misleading statements that take the form of promises but, in fact, are not true promises. A similar analysis seems more difficult to sustain in the case of conflicting promises. I have in mind overlapping promises to lunch. Although it may be wrong for Alice to make a lunch promise to Carly given her prior promise to Bob, unlike the recipient of an immoral promise, Carly does have a right to demand performance. She is reasonably disappointed, morally, by Alice’s failure to perform as such, even once she learns of the earlier promise to Bob. This makes it harder to sustain any suggestion that Alice’s promise to Carly is not a true promise. But if the second promise, to Carly, is a true promise, how can the rights-transfer theorist make sense of this because one might think Alice’s promise to Bob has already transferred her right to decide with whom to have lunch? The difficulty is not unique


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

to the case of conflicting promises: it arises in many cases of promises to engage in behavior that is not intrinsically immoral, but that conflicts with other all-thingsconsidered duties established by the context, such as a promise to lend you money that all things considered I should give to Oxfam or a promise to support your proposal at a meeting, although really it is not the right one for our department and conflicts with my fiduciary obligations. These all seem as though they are true promises, although of course they may not be all-things-considered binding (because of the conflict, among other things). Explaining how they could be true promises under the rubric of the rights-transfer theory requires greater precision than I have so far offered about what right it is exactly that has been transferred from Alice to Bob when Alice promises Bob to meet him for lunch. My suggestion, which I will continue to refine once I turn to redundant promises, is that these promises do not involve the simple transfer of the right to decide to be in one place or another or to do some act or another. If the rights transfer were that sweeping, then the second lunch promise would resemble the attempted immoral promise just discussed. Instead, I suggest these promises involve the transfer of a more limited right than the broad right to decide whether to perform an activity or not. What is transferred is the specific right to be the one, between the two parties, to decide the matter. This understanding dovetails with what I have elsewhere argued is a crucial function of the power to promise—to provide a mechanism within a relationship by which autonomous parties may share and transfer their discretionary powers. Normally, between A and B, A has the moral authority to decide whether to follow through on A’s intention to lunch with B. When A promises to have lunch with B, A transfers that decisionmaking authority—between them—as to who may decide whether that intention will be implemented, while setting a default that it will and that they will lunch together. A’s promise to B does not, however, alter the balance of power between A and C. Between A and C, A remains the one with the moral authority to decide whether to follow through on her intentions. Hence, it is possible for A to transfer the decision-making authority between A and B to B and also to transfer the decision-making authority between A and C to C. So even if A has promised B to meet B for lunch, A may still make a valid promise to C to meet C for lunch. And, in a sense, both promises could be honored if either C or B “waives” the promise and allows A to make other plans. Of course, if the lunch dates are at the same time and both B and C elect for A to show up, then, necessarily, A will act in one case without right because A will act against a decision of a person who has rightful authority about the matter. But understanding the right that has transferred in this way has the advantage that it does not make the second promise a sham. It also explains why if B were to cancel, then A is bound to respect C’s authority and therefore, to show up and meet C. This result is harder to explain on the theory that the second promise is empty or a sham. Finally, this account helps to explain why, generally, a promise is not simply transferable. Suppose A promises to meet B for lunch at P1, T1. Typically, B may not transfer that promise to D, making A obliged to meet D. This is because, typically, all A transfers

Immoral, Conflicting, and Redundant Promises


to B is something specific about the relation between A and B, not a general power to make decisions on A’s behalf about A’s presence at a time. This way of understanding the structure of the content of the promise gibes with the prior analysis of immoral promises. Absent a promise, between A and B, A has the right to decide whether to meet B for lunch and so can transfer that right through a promise. A may act wrongly by making a similar, conflicting promise to C but that does not render the promise to C an immoral one per se because the content of what is promised does not involve intrinsically immoral activity. Hence, the conflicting promises are possible. Contrast the case of immoral promises. Between A and B, A does not have the moral right to decide whether to rob a bank or “take care” of the guard. Neither of them have any such decision-making power. Hence, neither may transfer rights over that decision to the other and so neither can promise the other to engage in that activity. Although for A to be responsible for robbing the bank, she must be an autonomous agent who is responsible for her decisions, she has no moral discretion about what decision to make. That is not true with respect to the subject matter of the promises that may conflict with other promises or with other duties. Despite A’s promise to B, prior to any promise made to C, A still retains the decision-making power between herself and C about where A should lunch. Given her promise to B, A would be wrong to decide other than to lunch with B, but absent a promise to C, it would be inappropriate for C to tell (versus advise) A where to lunch; between A and C, that decision still lies within the sphere of A’s discretionary powers. Likewise, although there may be a right answer about how much I should give Oxfam, it is within my discretion and power to try to identify that amount and arrange my affairs to so donate; ditto with the exercise of my judgment power with respect to departmental affairs.

4. redundant promises

Redundant promises offer perhaps the most challenging case for the rights-transfer theory. Although they have not received much philosophical attention, they are not all that atypical. Debtors issue fresh promises to repay debts that are still outstanding; wrongdoers promise never to commit the wrong again; the known grifter or criminal promises not to prey upon a friend. In my view, redundant promises cannot be analyzed in the same way as immoral promises, mostly because the analysis falls so flat phenomenologically. For, suppose one hazarded the hypothesis that these promises involve misrepresentation. The bounder who promises not to steal from your purse might be thought essentially to do nothing. He represents himself as having a right to decide whether to steal or not, a right he then transfers to you. The wrong would be in the making or purported making of the promise because it misrepresents what his degrees of freedom are. This hypothesis does not seem plausible, though. If the bounder nevertheless rifles through your purse after making the promise, it seems as though you have a distinct reason to feel betrayed by the theft, one that sits atop the background reason that your


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

purse should not be subject to trespass. The misrepresentation theory does not do much to accommodate this sense that the promise itself makes a difference. Nor will the analysis just offered in the case of the conflicting promises, as it stands, help. In the example at hand, it really is already in the promisee’s decisional power whether or not the acquaintance takes money from her purse. So it cannot be that the thief transfers that decisional power to the promisee. The appropriate analysis of redundant promises seems multi-fold, differing according to content and context. Communications that present themselves as redundant promises may be divided into at least three categories: faux promises, promises regarding imperfect duties, and promises regarding perfect duties (the most difficult category). a. Faux Promises

In some cases, the redundant promise is not in fact a true promise but is not a failed attempt or a misrepresentation either. In these cases, the language of promise is used to signal that the speaker acknowledges explicitly that the behavior in question is a true wrong. Sometimes, when a prior wrongdoer promises never to engage in that action again, what may be transpiring is less a promise than the issuance of an explicit recognition that the behavior is morally inapt and perhaps a complementary statement by the speaker that although she at one time was unwilling or unable to conform her conduct to ethical standards, she now recognizes those standards and plans to adhere to them. In other related cases, some utterances taking the form of redundant promises may be elicited or provided to show that the speaker understands how a general moral rule applies in this specific case. Imagine the following exchange: I am visiting another university and overhear a teaching assistant say that she is teaching in the common room at 1. Half an hour later, I see a graduate student entering the common room at 12:50 with his lunch bag in hand. Meaning to be helpful, I say “In case you don’t know, I heard that the common room is being used at 1 p.m. for an undergraduate section.” He says, in reply, “Don’t worry. I promise to finish my lunch by then.” One way to read this colloquy is that the graduate student does not so much promise as he confirms to me that he realizes that he has a duty to vacate the common room when it is reserved for another purpose and that he recognizes that duty applies to this situation. It is not clear that much more than that is happening. The promissory language is just used to signal his understanding of the moral situation. It may also communicate the intention to comply with the relevant duty. But one piece of evidence that it is not also a real promise is the reaction I might plausibly have were the common room not vacated by 1 p.m. If the section were not mine and the common room was not my special responsibility, it would be peculiar for me to regard him as having committed a transgression against me were he to linger nonetheless. Yet promises generally ground a specific personal complaint by promisees when breached. If it were my business, so to speak, we might then have a case of a redundant promise with respect to a perfect duty—the truly hard case—to which I shall turn in a moment after discussing the second category, that of redundant promises with respect to imperfect duties.

Immoral, Conflicting, and Redundant Promises


b. Imperfect Duties

Although some statements that take the form of redundant promises may not be promises at all, others seem to do more than just announce facts about the speaker’s moral awareness and good will. In many cases, the statement seems to empower its recipient with a new and special complaint if the speaker later acts in ways inconsistent with the statement’s contents. In some cases, what I suggest is going on is that an imperfect duty has been made more precise. I may have an imperfect duty to help my friends financially if they are in need, but on any particular occasion it may be permissible to decline a request. If, however, I promise, then although I am doing what I have a duty to do in some sense, in another, I am making the commitment much more precise and determinate. Before, my friend could complain if I never helped her; now she may complain if I fail on this occasion.

c. Perfect Duties

The most interesting cases are those in which one promises to do that which one already has a perfect duty to do, as in our signal case of the thief promising not to steal from an acquaintance’s purse. Here, the background duty is perfectly determinate and in need of no further precisification. What right is it that the thief transfers? We might say the thief transfers the right to decide otherwise in the counterfactual case in which the perfect duty no longer obtains, for example, on those occasions on which taking without permission becomes permissible. In this particular case, this may be a difficult counterfactual to get one’s mind around. In other cases of perfect duties, the circumstances that give rise to the perfect duty may alter in more familiar ways. For instance, I assume that in two-parent families that the parents together have a perfect duty to pick up their young children from school (or to make a satisfactory arrangement), and that each parent has a perfect duty to do so if the other is unavailable. Suppose B is unavailable. Then A, the other parent, has a perfect duty to pick up the children. A also specifically promises her child that she’ll pick the child up. If B suddenly becomes available, A could fulfill her perfect parental duty by having B agree to do it but A’s promise could not be discharged in the same way. Her promise obligates her to pick the child up despite the fact that B is now available. But there are cases in which the discharge conditions of the perfect duty and the discharge conditions of a promise may be the same, in which case the counterfactual oomph of the redundant promise would be ethereal. In the case of the promise not to pilfer, the perfect duty not to take money from another’s purse might be overcome. If there were a medical emergency, for instance, and an urgent need for funds for a quick-acting medicine, that circumstance might represent the limiting condition on the perfect duty; but so too it probably represents the limiting condition on the thief’s promise. So is the redundant promise here superfluous? I do not think so. Consider the content of the reaction to breach. Suppose no exigent circumstances obtain but our thief steals from the purse anyway, despite the promise. The victim’s reaction is likely and appropriately to be more personal in nature than it would be


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

absent the promise: Some particular betrayal has occurred that goes beyond the betrayal of the theft alone, had the promise not been made. Or, suppose the thief forbore and is asked by a fellow pick-pocket why the thief didn’t relieve the easy mark of her money. The thief sincerely replies, “Well, to be honest, the only reason was that I wasn’t sure I could pull it off without being detected.” If the promisee overheard this exchange, I can imagine her being reasonably, albeit uncomfortably, peeved not just because the thief’s report betrays the attitude that theft is permissible, but because it more specifically betrays an indifference to the particular wrong of stealing from her, given the promise. I can imagine her thinking that the thief should have abstained also and largely because he promised to do so. The promisee may reasonably feel put out that the thief treated her impersonally—as just any old possible victim—rather than specially, in a more personal way. Where the utterance taking the form of a redundant promise is not merely a disguised acknowledgment of a standing moral duty, it seems to alter the reason why a particular action is or is not to be performed. Prior to the promise, the purse is to be left alone for the same reasons, basically, that every purse is to be left alone: there is an impersonal duty not to interfere with others’ legitimate personal property. After the promise, the purse is to be left alone for a more personal, promisee-oriented reason. Through his promise, he alienates his prerogative to treat the purse-owner impersonally regarding this matter, as the thief would otherwise treat anyone else. Put in this way, the idea should be somewhat familiar. This is in part what we do when we befriend a person. We relinquish our power to treat the person as we would treat anyone else and we generate claims in them to expect that we will treat them as not just one among many but as having agent-specific reasons to expect certain behavior from us. Some of the behavior that is to be expected is unique: we will give greater weight to their needs and interests and we will perform certain actions we otherwise would not because they are our friends. This aspect of friendship and personal relationships, along with some of the common emotions associated with relationships, is often emphasized in the literature. But, the move from the impartial to the partial stance with respect to what actions to perform or what weight to give various reasons is not the only important shift of deliberative perspective involved in special relationships. Often the behavior called for by the relationship is not different than the behavior we owe to others. I am obliged to refrain from gratuitous slapping of strangers, friends, and annoying acquaintances alike. What is different with a friend is what reason I have to perform or refrain: my reason with respect to a friend should be my care or concern for that individual or perhaps the fact that we have a friendly relationship. What is different with the thief is that he agreed to take something about her perspective about the matter as authoritative and to abstain, in some sense, for her sake. This shift is not captured by a standard version of the distinction between impartiality and partiality. In some circumstances, I may have reason to give my friend’s interests greater weight than a stranger’s when they conflict, but that sort of reason does not do any work to explain the difference between the nature of the quotidian duty not to assault the friend and the quotidian duty not to assault the stranger. Both duties are agent-relative in the sense that is often meant—namely, I cannot assault

Immoral, Conflicting, and Redundant Promises


the friend or the stranger in order to prevent more assaults by others (or myself). Both duties are owed to the specific person and are not fungible. But, with respect to the friend, there is a deliberative difference: In some sense or another, my reason should refer to something quite specific about that person or our relation (without in any way suggesting that contrary behavior would be licensed with respect to other people). I think of this as the move from the impersonal to the personal. With respect to the stranger, I owe that particular person—as I owe everyone else, each as an individual—the sort of respect that obliges me to refrain from assault, etc.; with respect to the friend, I owe that person a certain sort of consideration and care that propels such restraint, a sort that I do not owe to all others each as individuals. Not all redundant promises initiate this movement from the impersonal to the personal. Some redundant promises may overlap with duties that already take a personal form. Suppose the single parent has a specific, personal, perfect duty to pick up her child after school. So a promise she makes on top of that to pick her child up cannot be analyzed in terms of a maneuver that substitutes a personal reason for an impersonal one. This is not the only such example—lawyers may promise clients to exert their best efforts or to honor their clients’ rights to decide whether to settle or not. These are already duties owed to specific individuals with whom a personal relationship has already been forged. Of course, the force of the promise may be just the counterfactual force mentioned before, but I suspect that something else, deliberative in nature, is going on at least in some cases. The idea that the promise involves a commitment to a deliberative stance is not in itself so unusual. Other familiar promises take a deliberative form, explicitly manifesting concern for the reasons on which a promisor will act. A promisee who is altruistic, or perhaps does not wish to owe a debt of gratitude, may ask her friend to promise that he will not perform some action, for example, see a particular movie, for her sake but rather, only to do so if it interests him. With respect to redundant promises, I suggest that what is being promised, implicitly, is to take on in one’s deliberation a greater proportion of the promisee’s perspective on the significance of the duty, how it is to be discharged, what sort of effort to exert to fulfill it, and how to assess whether exonerating circumstances hold. What do I have in mind? Take the standing duty to pick up your child from school. It is important to be timely. This means, often, interrupting desired tasks and even leaving earlier than will typically be necessary. But, nonetheless, one acts serviceably well with respect to this duty if one tries hard in light of known, likely hazards even if one does not do all that one can. If the school is ten minutes away in light traffic and traffic is light at the relevant time, one may have to leave twenty minutes in advance to protect against unexpected delay but one surely does not have to leave—every day—an hour and a half in advance although that would ensure being timely. You need not do all that you can to contain all possible risks from maturing. Now suppose you typically take one route and make it on time reliably. But your anxious child strongly prefers that you take the other route because it’s more direct and contends that that’s the proper way to be on time. Her reasons are fine ones but so are yours: Although your route is longer, there’s often less traffic, it is prettier and you have


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

become expert at gaming the traffic lights. It is reasonable for you to choose your route but still, there is a greater (albeit small) risk of being late with your route if the timing is inexact on a heavily trafficked day; perhaps there is a lesser risk on other sorts of days. If this mild conflict has been aired and is thereby in the background when you promise to be on time, I think that there is greater moral pressure for you to take the route your child prefers; she will have some real complaint if you are late because you took your own route (even if the choice absent the promise is a perfectly reasonable one and even if the delay really is not your fault). To return to the rights framework, with this sort of redundant promise, I suspect that—often at least—you transfer the right to deploy your exclusive judgment about how to fulfill one’s duty, which means to use, what risks of failure may permissibly be run, etc. You agree, in part, to take on some of the promisee’s perspective and judgment about how to approach the preexisting duty. I am tentative about this suggestion. In part, it is because it does not seem as though the promisee has a complaint if you do show up on time and used your own route (though I am honestly unsure about this). Perhaps she does have a complaint but it is one that it is petty to voice in the circumstances; maybe all she is entitled to is raised eyebrows followed by a concessive smile. If she has a right to complain, this suggests that the right transferred would be the right to use one’s exclusive judgment about how to comply with the duty, but if she has no right to complain, perhaps all that is transferred is the right to enjoy impunity if one breaches while using one’s own judgment, even without objective fault. I do not have much in the way of argument that this is exactly what happens and I do not think our intuitions or reactions are all that sharp in this domain. Nonetheless, I am attracted to a view with this structure because it seems as though something goes on even in these redundant promises, and that the relevant complaints the promisees could make have to do with the promisors not sharing power in the appropriate way in light of the promise. Because the relevant issue could not concern what action is performed, it will have to concern how or for what reason the action is performed. Although I find the degree of tentativeness in our understanding of redundant promises frustrating, in another way it is entirely apt. Accounts that revolve around assurance and the development or strengthening of expectations too handily address redundant promises (as well as immoral and conflicting ones), treating them as just other cases in which new expectations are generated or greater levels of confidence in existing expectations are warranted. These are difficult cases. Yet the assurance view does not recognize the redundant promise as an especially hard case (except in those cases in which the promisee already has a complete and full expectation that the promisor will perform on the basis of the preexisting duty). This oversimplifies what seems to be indubitably rugged terrain. The preliminary lessons I would draw from considering the surprisingly rich subject of redundant promises are that although they do submit to fruitful analysis within the rights-transfer framework, their structure is not always transparent. Sometimes, they only appear to be promises but are instead forms of assertion: for example, testimony acknowledging one’s primary duty. In other cases, they seem to have true promissory

Immoral, Conflicting, and Redundant Promises


power: rendering imperfect duties determinate. In yet further cases of promises to perform already perfect duties, they may commit the promisee to performance even if the limiting conditions of the primary, preexisting duty are met. But there are cases in which the limiting conditions of the preexisting duty and the limiting conditions of the overlapping promise are similar if not identical. In such cases, I suggest that the redundant promise still has moral force—what it involves is a commitment to deliberate in a certain way or to act as though one has deliberated in a certain way. One replaces the default method of deliberation (the impersonal perspective or one’s own perspective) with a method that gives greater prominence to the deliberative perspective of the specific promisee. This amounts to a distinctive form of partiality. Those are the main points about redundant promises that I have been trying on. Three larger themes emerge from them. The first is that the content of promises varies in ways rarely remarked upon. It is commonplace that different promises have different actions as their objects. But the subject of redundant promises highlights that the content of a promise may also vary on other dimensions, including with respect to one’s deliberative stance. Further, the promise’s content may vary depending upon not merely the words used, but also what other things are present and fixed in the moral environment. Second, some of these variations are quite subtle and inexplicit. The moral features of promises are, like many other moral actions, superficially familiar and usually partly accessible on the surface. Yet, many of their characteristics and nuances are not at our fingertips. This, I concede, is a controversial view that will flummox some and irritate or alienate others. Third, these suggestions about one’s deliberative stance—here, with respect to redundant promises and, earlier, in the discussion of nested immoral promises— reveal a larger schism between rights-transfer theorists about promises and some contemporary ethicists. If this analysis of redundant promises is correct and the right that is transferred involves the reasons one should or can act upon, then the dispute may implicate the issues of whether one’s motives may be relevant to the sort of action one performs, its permissibility and its moral nature. Some resistance to such views has been prominent in Scanlon’s most recent work. For a variety of reasons, I embrace a broader view of the relevance of motives to the nature of moral action. But, I will not here embark on a defense of the broad relevance of motives to the moral classification of actions. Rather, I will end by observing that the interweaving of these issues further underscores the philosophical richness of the topic of promising. notes 1. I am grateful to Scott Altman, Jorah Dannenberg, Scott Dewey, David Goldman, Mark Greenberg, Jeffrey Helmreich, Barbara Herman, Pamela Hieronymi, A.J. Julius, Gregory Keating, Niko Kolodny, Gabriel Rabin, William Rubenstein, Tim Scanlon, Vicki Steiner, Terry Stedman, Judith Jarvis Thomson, Gary Watson, Stephen White, members of a course at University of California–Los Angeles on Truthtelling and Promising, and audiences at Harvard University, Princeton University, the University of North Carolina, and the University of Southern California Gould School of Law, for help and constructive trouble.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

2. I will focus here on (and cite) Scanlon’s theory as presented in “Promises,” ch. 7 in What We Owe to Each Other (Cambridge, Mass.: Belknap Press of Harvard University Press, 1999). An earlier version of that article appeared as “Promises and Practices,” Philosophy & Public Affairs 19(3) (Summer 1990): 199–226. Another important article on the subject, “Promises and Contracts,” may be found in The Difficulty of Tolerance (Cambridge: Cambridge University Press, 2003). 3. Scanlon’s article came out the same year as Judith Jarvis Thomson’s important and influential work on the subject. Together, these two works reinvigorated the subject. Judith Jarvis Thomson, “Giving One’s Word,” ch. 12 in The Realm of Rights (Cambridge, Mass.: Harvard University Press, 1990). 4. Scanlon, What We Owe, 304. I will focus here on the view articulated in Scanlon’s first book, although I believe his position has evolved in two ways, both relating to the mental states of the promisor and promisee. The published view stresses the significance of the actual expectation of the promisee and the actual intention of the promisor to induce that expectation. I believe now that Scanlon endorses a more “objective” version of the position along the following, rough lines: a promise is a speech-act that gives warrant to the promisee to form an expectation of performance on the part of the promisor; it is wrong to provide such warrant insincerely (i.e., without the intention of performance) and it is wrong to provide such warrant and then act in a way that would disappoint the expectation that the promisee had reason to develop (assuming the promisee does not reject the promise). Whether or not it is one’s intention to give that warrant, if one acts in ways reasonably understood to give that warrant (e.g., by saying “I promise” in a nonjoking fashion), one is bound to respect the expectations one has given the recipient a warrant to form; likewise, what matters is not whether the expectation is formed by the promisee, but whether its formation would be reasonable. This more objective formulation still leaves open whether we should say that it would be reasonable for the promisee to form an expectation of performance in controversial cases such as the Profligate Pal in which the promisee has every reason to predict the promisor will not in fact perform. See Scanlon, What We Owe, 312–14, and my discussion in Seana Valentine Shiffrin, “Promising, Intimate Relationships, and Conventionalism,” Philosophical Review 117(4) (October, 2008): 486–92. If the reasonability of the expectation turns upon the objective likelihood of performance, then Scanlon’s theory still (implausibly in my view) entails the Profligate Pal is not obliged to perform. If the reasonability of the expectation does not turn on what it is reasonable to anticipate, then it seems less clear what work the appeal to expectations does in the theory. In any case, with respect to the issues discussed in this essay, I do not think much turns upon whether the promisor intends to assure or whether, through means other than a rights-transfer, the promisor acts in a way that may reasonably be taken to provide assurance. 5. See, e.g., Shiffrin, “Promising, Intimate Relationships, and Conventionalism,” 486–93; Stephen Darwall, “Contractualism, Root and Branch: A Review Essay,” Philosophy & Public Affairs 34(2) (March 2006): 211–14; Michael Pratt, “Scanlon on Promising,” Canadian Journal of Law and Jurisprudence 14(1) (January 2001): 146–49. 6. See, e.g., Niko Kolodny and R. Jay Wallace, “Promises and Practices Revisited,” Philosophy & Public Affairs 31(2) (April 2003): 119–54; John Deigh, “Promises under Fire,” Ethics 112(3) (April 2002): 483–506. 7. See, e.g., Shiffrin, “Promising, Intimate Relationships, and Conventionalism,”; David Owens, “A Simple Theory of Promising,” Philosophical Review 115(1) (January 2006): 51–77; Joseph Raz, “Voluntary Obligations and Normative Powers II,” Proceedings of the Aristotelian Society 46 (suppl.)(1972): 100–101; Gary Watson, “Promises, Reasons and Normative Action,”

Immoral, Conflicting, and Redundant Promises


in Reasons for Action, ed. D. Sobel and S. Wall (Cambridge: Cambridge University Press, forthcoming). Judith Jarvis Thomson’s view, that a promise involves giving a claim to, or creating a claim for (rather than transferring a right to), the promisee is closely related. I do not think she would describe herself as a rights-transfer theorist. Instead, she believes that promises, like other sorts of word-givings, give the recipient a claim—in this case a claim that one must do what one said one would do. (See Thomson, “Giving One’s Word,” 305.) I am more attracted to the characterization of promises as transferring rights rather than creating claims in the promisee that the content of what one has said is true and that it may be relied upon. The former more clearly accounts for the fact that when one promises, one must perform, whereas a mere claim that one do what one has given one’s word to do might be discharged by warning that one has changed one’s mind and will not perform after all or by compensating for any monetary losses associated with nonperformance. The idea of giving a claim does not seem, as obviously, to explain the default and distinctive requirement of performance in the case of promises. I discuss the significance of the obligation to perform as such at greater length in Seana Valentine Shiffrin, “Could Breach of Contract Be Immoral?” Michigan Law Review 107(8) (June 2009): 1551–68. 8. The problems I will discuss arise in one guise or another for both variants. The main difference between them is that the existence view holds that immoral promises are not true promises at all, whereas the bindingness view holds they are true promises, but without binding force. With respect to conflicting and redundant promises, both views hold that the relevant promises are valid and must grapple with the same issues about whether they exert moral force and why. Mutatis mutandis, both versions of the theory must resolve issues about what sorts of rights are transferred in order to grapple with the problem cases I identify. 9. See Scanlon’s discussion of this case and his denial that a binding promise forms in What We Owe, 321. I criticize Scanlon’s analysis of this case in Shiffrin, “Promising, Intimate Relationships, and Conventionalism,” 487–93. 10. An articulation of the problem appears in Thomas Pink, “Promising and Obligation,” Philosophical Perspectives 23(1) (December 2009): 405–406. 11. I focus on the case in which the promisor knows he has a preexisting duty. That case presents both the issue of what work the redundant promise could do and whether the redundant promise as such could make sense to the promisor and the promisee. Where the promisor does not believe he is under a preexisting duty, it is clearer what phenomenological and motivational work the promise would do for the promisor and therefore clearer what value the promise would have for the promisee. See Joseph P. DeMarco and Richard M. Fox, “Putting Pressure on Promises,” Southern Journal of Philosophy 30(2) (Summer 1992): 57; Pink, “Promising and Obligation,” 414. 12. For that matter, so may conventionalist accounts of promising. Conventionalist accounts, e.g., may interpret the convention of promising as recognizing immoral promises as true promises, but ones that do not create an obligation. See J. E. J. Altham, “Wicked Promises,” in Exercises in Analysis, ed. Ian Hacking (Cambridge: Cambridge University Press: 1985): 7, 13. Or they may claim that the convention does not recognize promissory language purporting to commit to immoral activity as promises at all. 13. Bernard Williams, Truth and Truthfulness (Princeton, N.J.: Princeton University Press, 2004), 89. Of course, the promise may seem more meaningful when the promisor does not realize she has a preexisting duty. 14. Jurisdictions differ about what sorts of crimes are admissible for purposes of impeachment. To impeach, some states only admit evidence of crimes involving dishonesty, such as


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

perjury or fraud. Other states, and the Federal Rules of Evidence, allow the admission of evidence of some other felony convictions (if more probative than prejudicial) on the grounds that they support an inference of poor character and, therefore, untrustworthiness. See Federal Rules of Evidence, R. 609(a)(1). The rationale is that the commission of a felony, whether involving dishonesty or not, shows a willingness to place one’s own interests over others and over the importance of respecting social norms; one might, therefore, do the same as a witness. See, e.g., Richard Green, “Evidence,” Southern Illinois University Law Journal 13(3) (Spring 1989): 649; David Sonenshein, “Circuit Roulette: The Use of Prior Convictions to Impeach Credibility in Civil Cases under the Federal Rules of Evidence,” George Washington Law Review 57(2) (December 1988): 281. Nonetheless, even the Federal Rules of Evidence privilege evidence of crimes of dishonesty over evidence of other crimes. Federal Rules of Evidence, R. 609(a)(2). Even when evidence of nonfraudulent felonies is admissible to cast doubt upon the veracity of testimony, character evidence is disallowed for the purpose of establishing any greater likelihood that the defendant is guilty of the crime with which he or she is charged. See generally Kenneth J. Melilli, “The Character Evidence Rule Revisited,” Brigham Young University Law Review 4 (1998): 1579. 15. This sort of distinction (between different moral transgressions and different character flaws giving rise to them) explains why impeachment evidence of felonies that do not involve dishonesty is sometimes excluded (even under the Federal Rules of Evidence). Hawaii, in developing its rule for witness impeachment, relied explicitly on this reasoning in Asato v. Furtado, 474 P.2d 288, 294–96 (Haw., 1970) (witness’s conviction for heedless and careless driving inadmissible for impeachment because it “bears no rational relation to a witness’s credibility”). The Supreme Court of Hawaii explained: We think that there are a great many criminal offenses the conviction of which has no bearing whatsoever upon the witness’s propensity for lying or truth-telling, and that such convictions ought not to be admitted for purposes of impeachment. This is true not only of minor offenses . . . but also of some major offenses like murder or assault and battery. It is hard to see any rational connection between, say, a crime of violence and the likelihood that the witness will tell the truth. Id., 294–95. See also Hawaii Rules of Evidence, R. 609 (a) (2006). Although the Federal Rules of Evidence permit evidence of prior crimes other than those of dishonesty, the admission of such evidence is not automatic as it is with crimes of dishonesty; instead, its probative value must outweigh its prejudicial value. Often, that burden is unmet. See, e.g., United States v. Begay, 144 F.3d 1336, 1338 (10th Cir., 1998) (witness’s prior drug conviction improper for cross-examination because insufficiently probative of dishonesty); Eng. v. Scully, 146 F.R.D. 74, 78 (S.D.N.Y., 1993) (prior felony conviction of murder not admissible for witness impeachment because character flaw indicated by murder not sufficiently probative of dishonesty. “Murder is not necessarily indicative of truthfulness, and the probative value of a murder conviction is substantially outweighed by the danger of unfair prejudice.”); United States v. Rosa, 891 F.2d 1063, 1069 (3rd Cir., 1989) (evidence of past fraud admissible but evidence of bribery inadmissible because bribery is “not the kind of conduct which bears on truthfulness or untruthfulness.”). 16. I acknowledge, however, that some people’s intuitions and considered philosophical positions differ. Altham takes the view that immoral promises must be promises and uses the example to cast doubt on efforts, such as John Searle’s, to derive “ought” from “is.” See Altham, “Wicked Promises,” 1–21; see also Margaret P. Gilbert, “Three Dogmas about Promising,” in Promises and Agreements, ed. Hanoch Scheinman (Oxford: Oxford University Press, forthcoming). Ironically, Searle does regard the immoral promise as a true promise, albeit one

Immoral, Conflicting, and Redundant Promises


whose reason-giving force is outweighed by the “evil nature of the promised act.” See John Searle, Rationality in Action (Princeton, N.J.: Princeton University Press, 2001), 194–95. Although I doubt immoral promises are true promises, my aim is not to vindicate or undermine Searle’s is/ought derivation. As I discuss in note 8, the bindingness version of the rightstransfer theory does recognize these as promises, albeit ones that are not binding because they do not actually transfer a right (and not because the moral force of the promise is outweighed by the significance of the immorality of the act). 17. Gary Watson develops a like-minded criticism in his “Promises, Reasons, and Normative Powers,” 20–28. But see Gilbert, “Three Dogmas About Promising,” who argues that immoral promises create obligations but contends that they are the sort of “obligations” that do not provide their possessor with a moral reason to perform. This position allows Gilbert to accommodate the linguistic familiarity of calling immoral promises true promises. The bindingness version of the rights-transfer view also has this feature. In both cases, though, divorcing obligations and promises generally from a necessary connection to moral reasons to perform seems too high a price to pay for this accommodation. 18. I have attempted to dislodge the general purported mystery about how promises could generate moral reasons just by declaration in “Promising, Intimate Relationships, and Conventionalism” with respect to typical promises to perform actions that are, ex ante, morally discretionary. I agree, however, that it would be beyond puzzling if by declaration we could create moral reasons to do things that, ex ante, are morally impermissible and lie outside our sphere of autonomy. 19. See, e.g., United States v. Stevens, 130 S.Ct. 1577, 1591 (2010) (refusing to uphold an “unconstitutional statute” on the basis of the government’s intention to implement it with discretion); Christian Legal Soc. Chapter of the University of California, Hastings College of the Law v. Martinez, 130 S.Ct. 2971, 3017 (2010) (reaffirming that “[a] law neutral on its face still may be unconstitutional if motivated by a discriminatory purpose”); Lemon v. Kurtzman, 411 U.S. 192, 198 (1973) (referring to “unlawful statutes”). 20. See, e.g., United States v. Mardirosian, 602 F.3d 1, 7 (2010) (reiterating that “contracts for illegal purposes are void as a matter of public policy”); Massachusetts Mun. Wholesale Elec. Co. v. Town of Danvers, 411 Mass. 39, 54–55, 577 N.E.2d 283 (1991) (“When adjudicating rights between parties to a contract that is void ab initio, courts treat the contract as if it had never been made.”). 21. See, e.g., Thomas Firestone, “Mafia Memoirs: What They Tell Us About Organized Crime,” Journal of Contemporary Criminal Justice 9(3) (August 1993): 201–204 (quoting, among others, Sammy “the Bull” Gravano’s testimony in which he explains that the Mafia lifestyle “didn’t seem wrong” and reporting that many Italian-American mobsters “were raised in communities in which becoming a gangster was seen as a legitimate and even desirable career path.”). 22. Organized crime groups standardly inculcate strong senses of loyalty and “honor” among members as well as a sense of duty to carry out the groups’ directives. See, e.g., Letizia Paoli, Mafia Brotherhoods (Oxford: Oxford University Press, 2003), 67–87; Francis A. J. Ianni, “The Mafia and the Web of Kinship,” Public Interest 22 (Winter 1971): 80–98; David E. Kaplan and Alec Dubro, Yakuza: Japan’s Criminal Underworld (Berkeley and Los Angeles: University of California Press, 2003), 8–17; Martin Booth, The Dragon Syndicates: The Global Phenomenon of the Triads (London: Doubleday, 1999), 158–61. 23. Jokes differ from transparent lies because the joke (as well as the intentional piece of fiction or other sort of entertainment) is transparently proffered as such. The lie is a representation the speaker does not believe but offers to the recipient as a representation to be treated as true; this is compatible with it being evident that the speaker does not believe its content.


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

Jokes and fictional material are not offered to be taken as true but rather are offered for another purpose. One reason why some jokes are morally questionable, or even cruel, is that the timing of the speech and the timing of the revelation of the purpose of the speech are far from coincident. 24. See Thomas L. Carson, “The Definition of Lying,” Nous 40(2) (June 2006): 289–90. Strangely, Carson’s example unnecessarily specifies that the perjurer believes she will not be convicted for perjury to underscore the perjurer is not intending to deceive. 25. There is more to say about why transparent lies are wrong. Part of the explanation lies, I think, in the fact that they generate epistemic burdens for the recipient in light of our standing dispositions to believe what interlocutors say and particularly to regard the testimony of friends and acquaintances in a charitable light and to treat their assertions with politeness and respect. Confronting transparent lies may generate forms of cognitive dissonance, forcing recipients to enact barriers against the operation of these dispositions or tempting listeners into wishful thinking and other sometimes dangerous forms of conflict and truth avoidance. Dealing with transparent lies also burdens the listener ethically: she must either behave inauthentically and act as though nothing has happened, or confront the liar and spark a confrontation. 26. I argue that moral misrepresentations, et alia, may be wrong independent of their consequences on particular occasions in “Lies and the Murderer Next Door,” unpublished ms. 27. Altham, “Wicked Promises,” 9; Elizabeth Anscombe, “On Promising and Its Justice, and Whether It Need be Respected In Foro Interno,” in Ethics, Religion, and Politics: Collected Philosophical Papers of G.E.M. Anscombe (Minneapolis: University of Minnesota Press, 1981), 16. 28. The principle of “unjust enrichment,” provides that a person who receives a benefit besides a gift from another party owes restitution thereto “if the circumstances of its retention are such that, as between the two persons, it is unjust for him to retain it”; see Restatement of the Law of Restitution (St. Paul, Minn.: American Law Institute, 1937), § 1 comment C. Such circumstances include cases where one party pays another in exchange for services or goods never received; see, e.g., Landeis v. Nelson, 808 P.2d 216, 219 (Wyo., 1991) (finding it unjust for franchisor to retain money paid by franchisee in exchange for franchise when the franchisee “received nothing in return”). This principle may apply even to payments for services or transfers that, like the case of an immoral promisor, the recipient has no legal duty or right to render. “A person who transfers something to another believing that the other thereby comes under a duty to perform the terms of a contract or becomes subject to other duties, is ordinarily entitled to restitution . . . if the obligation intended does not arise and if the other does not perform . . . ,” Restatement of the Law of Restitution § 47 at comment b. See also id. § 52. 29. Suppose Jack performs. On the restitutionary theories I have articulated, it would seem that Leonard could claim restitution anyway: for we cannot say, in a sober moral voice, that Leonard received true value for his money when Jack killed Harry (even though Leonard got what he wanted). Further, whether or not Jack performs, Jack misrepresented what rights he had to convey. I do not regard this as an embarrassment (although it is no priority of mine that Leonard be made whole). Jack had no business accepting money in exchange for his “promise” and it is not disturbing if the money he accepted is not rightfully his to keep. Such reasoning is not unheard of in criminal cases involving money paid for illegal enterprises, where the payor has been found entitled to full restitution for the amount spent, regardless of what she received in return; see, e.g., Watts v. Malatesta, 186 N.E. 210 (N.Y., 1933) (casual horse gambler entitled by statute to restitution for full amounts waged to unlawful bookmaker regardless of

Immoral, Conflicting, and Redundant Promises


amount won or lost). Even a gambler who knowingly transfers money to a professional bookmaker for the purpose of betting has been found to have “never parted with the title to his money”; People v. Stedeker, 67 N.E. 132, 133 (N.Y., 1903). One justification for such rulings is that where the money was paid for unlawful purposes, the recipient has no rights to it, for “from such activities there could arise no rights at all”; see Hofferman v. Simmons, 49 N.E.2d 523, 526 (N.Y., 1943) (professional gamblers not entitled to proceeds seized by police). 30. Both cases are versions of examples suggested by Jorah Dannenberg. 31. Although I do not believe Scanlon’s expectation account provides the correct account of what promises are or why they bind, I do agree with the insight that the deliberate cultivation of others’ expectations may have serious moral consequences, although they may not consist of obligations to perform what is expected. 32. Shiffrin, “Promising, Conventionalism, and Intimate Relationships.” 33. If the promise to C were a sham, the obligation to meet C would grow only from a concern for C’s interests, as described in the prior section, but C would not have the right to demand performance as such. C would only have a claim that her underlying interests are responded to, which may or may not require performance as such. 34. Contract law is, however, permissive of some sorts of assignments of rights to third parties, but the ability of a promisee to assign contractual rights to another without the promisor’s consent is quite limited in personal-service contracts and other circumstances in which the identity of the party to whom performance is owed might reasonably matter to the promisor. See Restatement (Second) of Contracts (St. Paul, Minn.: The Institute, 1981) §§ 317, 318; E. Allan Farnsworth, Farnsworth on Contracts (New York: Aspen Law & Business, 1990), § 11.10 (1990). 35. The issues they raise may also arise with respect to other moral phenomena. For instance, an apology, in some respect or another, seems to involve some commitment not to engage in the same wrong again. Because one’s original behavior was a wrong, one is already bound, by the underlying principle making it a wrong, not to engage in that activity. But, recommission of the wrong seems all the more galling after an apology has been made. This essentially raises the same issues as redundant promises do, although an apology is not, on its face, a promise. Jeff Helmreich explores these issues with respect to apologies in “More Than Words,” unpublished ms. 36. See, e.g., Restatement (Second) of Contracts, §§ 82, 86; Farnsworth on Contracts, § 2.8. 37. Compare the legal presumption against injunctions that are worded so as to instruct the defendant to obey the law in fairly general terms or in the actual terms of the relevant statute or legal provision. See Federal Rules of Civil Procedure, R. 65(d) (requiring an injunction to be specific in its terms); Burton v. City of Belle Glade, 178 F. 3d 1175, 1201 (11th Cir., 1999) (invalidating an order “not to discriminate in future annexation decisions”). The legal presumption does not seem to be based on the idea that an injunction to obey the law generally would inaccurately presuppose or suggest that, without it, agents were not so bound. Rather, it seems to rest on the idea that such injunctions do not supply much guidance and specificity to those who have already demonstrated their failure to respond appropriately to the law as it is already worded. 38. Promises seem to presuppose that the promisee has an interest (or might reasonably be thought to have an interest) in what is promised or, at least, in having the power to decide whether what is promised is performed. 39. Scanlon might say the promise alters the meaning of the action not to be performed; see T. M. Scanlon, Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, Mass.: Belknap


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

Press of Harvard University Press, 2008), 52–56. I disagree with him about this way of putting it but that disagreement should be pursued elsewhere. 40. See, e.g., Niko Kolodny, “Love as Valuing a Relationship,” Philosophical Review 112(2) (April 2003): 154, 160–61; Samuel Scheffler, “Relationships and Responsibilities,” Philosophy & Public Affairs 26(3) (July 1997): 189–209; Bernard Williams, “Persons, Character and Morality,” in Moral Luck (Cambridge: Cambridge University Press, 1981), 17–18. 41. Related ideas are developed in J. David Velleman, “Love as a Moral Emotion,” Ethics 109(2) (January 1999): 355–60, 366–72, and Daniel Markovits “Promise as an Arm’s Length Relation,” in Promises and Agreements, ed. Hanoch Scheinman (Oxford: Oxford University Press, forthcoming). Markovits suggests that the move from the impersonal to the personal, essential to love, is antithetical to promising. See Markovits at 8–13. I aim to show that promising, in some cases, involves this very move; if plausible, this compatibility dissipates one of Markovits’ arguments about the tension between promises and intimacy. 42. The move from the impersonal to the personal stance is not always appropriate in intimate relationships. For instance, if one’s child or friend is enrolled in one’s course, the reason one has to give her the grade she earns or to call on her when she raises her hand is exactly the same reason as to call on any other student. To inject the personal here at all is entirely inappropriate. 43. I am grateful to Terry Stedman for helpful discussion on this point. 44. For an even stronger emphasis on the promisee’s perspective, see Jorah Dannenberg, “Promising as Valuing,” ch. 4, in Promising as Paradigm (Ph.D. diss., University of California– Los Angeles, 2010). 45. See Scanlon, Moral Dimensions. Some of his prior work on promising did not reflect this view, as he acknowledges, id., 37–39. His original principle against manipulation condemns actions as impermissible manipulation if through the action, the actor intentionally leads a party to believe an exchange will be made when the eliciting agent has no intention of fulfilling his end of the bargain (What We Owe, 298). In later work, he says that if the action would lead the other party to suffer a serious loss, due to detrimental reliance, it would still be impermissible whether or not the eliciting agent’s intention is to cause this loss or to allow this loss to occur (Moral Dimensions, 39).

8 The Trouble with Tolerance* Angela M. Smith

§ 1. introduction

it has long been recognized that tolerance is an elusive concept. It is elusive for a number of reasons, not the least of which is that it is not at all clear what kind of thing the term “tolerance” is supposed to refer to. In the literature, one can find discussions of tolerance as a value, an ideal, a practice, a behavior, an attitude, a principle, and a virtue, among other things. And it seems that tolerance, understood in any of these senses, might be something possessed or exercised by persons, or something possessed or exercised by institutions. Thus, we talk of tolerant people, tolerant political systems, and tolerant organizations, and we assume that there is some obvious quality they all share. It is no wonder, therefore, that the concept of tolerance has given rise to so many apparent “paradoxes”: with so many different ideas of tolerance in play, it is not surprising that our attempts to make sense of the notion would encounter conceptual difficulties. My own interest in the idea of tolerance was initially sparked by T. M. Scanlon’s penetrating discussion of the topic in his essay, “The Difficulty of Tolerance.” One of the many things I admire about this article is Scanlon’s careful distinction between tolerance as a personal attitude and tolerance as a political practice, and his insistence that a truly tolerant society is one that embodies tolerance at both levels. That is to say, a truly tolerant society is not merely one that affords certain legal and political protections to its citizens, but one whose citizens themselves are committed to approaching the “informal politics” of social life with attitudes of tolerance (190). By a society’s “informal politics,” Scanlon seems to have in mind those social and cultural arenas in which we debate, as fellow citizens, what our society is and what it ought to be. 179


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

Achieving a level of tolerance in our informal politics, Scanlon notes, is “importantly and irreducibly, a matter of attitude” (190). It requires, among other things, that we appreciate that those with whom we disagree are just as fully members of our society as we are, and equally entitled to contribute, through their modes of living and public discourse, to the definition of our society (192–93). It requires that we approach our disagreements in a “spirit of accommodation” (198), and with an eye to finding areas of common ground upon which we can all agree. And it requires that we strive to contain even our deepest conflicts within a framework of mutual respect and recognition (193). Sustaining such attitudes toward those with whom we disagree over fundamental matters is not easy to do—it is frightening and it is risky; yet it is, Scanlon suggests, the key to realizing a genuine culture of tolerance. This emphasis on the significance of personal attitudes like open-mindedness, civility, and respect to the achievement of a genuinely tolerant society seems to me exactly right, and deeply important. What is less clear to me is whether a defense of these sorts of attitudes can really amount, as Scanlon supposes, to a defense of tolerance itself as “an attitude that we all have reason to value” (188). My puzzle here relates to a tension Scanlon himself describes in the introduction of his article. In some cases, it seems, being “tolerant” of another person is not really an admirable attitude to hold, but merely a “second best” alternative to attitudes of genuine respect and recognition. Those who are moved by racial or ethnic hatred, for example, may come to have attitudes of “tolerance” toward those whom they despise. Although it is good in such situations that tolerance keeps them from oppressing those whom they abhor, it would be much better if they were to rid themselves of these unjustified attitudes of disapproval altogether. Although Scanlon recognizes the existence of these “impure” cases of tolerance—where tolerance is “merely an expedient for dealing with the imperfections of human nature” (188)—he thinks there are also “pure” cases of tolerance, where tolerance is in fact the morally ideal attitude to hold toward other citizens in society. These will be cases in which “persisting conflict and disagreement are to be expected and are, unlike racial prejudice, quite compatible with full respect for those with whom we disagree” (188). In such cases, Scanlon suggests, adopting attitudes of tolerance toward others is not merely a “second best” response, but the essence of respectful relations with others in a pluralistic society. Although I find this idea appealing, it seems to me there is still a tension present even in these allegedly “pure” cases of tolerance that should give us pause. This tension is clearest when one considers things from the perspective of those who are the target of such attitudes. Few of us, I think, would be pleased to learn that other citizens in society hold attitudes of tolerance toward us or our practices. If a stranger were to inform me that he is tolerant of my failure to attend church regularly, for example, my inclination would be to say, “gee, that’s swell, but what gives you the moral authority to be tolerant of me?” There can be more than a whiff of both condescension and impertinence to such an attitude—even in allegedly “pure” cases— which sits rather uneasily with Scanlon’s suggestion that tolerant attitudes embody a form of respect.

The Trouble with Tolerance


Indeed, it is just this tension—between the apparent condescension involved in attitudes of tolerance and Scanlon’s defense of tolerance as an expression of mutual respect—that I want to explore in this essay. My main claim will be that what Scanlon has in fact given us is a compelling defense of tolerance as a virtue of character, rather than an attitude. Though these notions are sometimes used interchangeably in the literature, I will argue that it is important to keep them distinct. I will further argue that, perhaps surprisingly, the virtue of tolerance only rarely calls upon us to take up attitudes of tolerance toward our fellow citizens, and that in a morally ideal world such attitudes would not be necessary. What this means, if I am correct, is that there are no “pure” cases of attitudinal tolerance within the political sphere, in the sense that Scanlon has in mind. Taking up an attitude of tolerance toward a fellow citizen, I will argue, is always merely “an expedient for dealing with the imperfections of human nature” (188). When such an attitude is virtuous, however, the imperfections in question reside not in the nature of the one who is tolerant, but in the nature of the one who must be tolerated. My strategy will be as follows: In the next section, I will put forward a hypothesis about why the attitude of tolerance gives rise to resentment in certain contexts. Though I think my interpretation of what the attitude of tolerance involves is contestable, I will argue that it makes sense of our intuitions about the appropriateness of tolerant attitudes in a wide variety of cases. In section 3, I will look at Scanlon’s defense of tolerance, and explain why I think it is better understood as a defense of tolerance as a virtue rather than an attitude. I will go on to explain why someone with the virtue of tolerance, as Scanlon understands it, will not be disposed to adopt attitudes of tolerance toward her fellow citizens. Finally, in section 4, I will suggest that there may be one situation in which a person with the virtue of tolerance may feel called upon to adopt an attitude of tolerance toward a fellow citizen: namely, when that citizen refuses to accord to others basic forms of mutual respect and recognition. In such cases, I will argue, a tolerant person may justifiably regard her fellow citizen as an immature participant in their shared political life who deserves attitudes of tolerance, but not full moral respect.

§ 2. tolerance as an attitude

Let me begin, therefore, by considering what is involved in holding an attitude of tolerance toward another person. Scanlon himself describes the attitude of tolerance in this way: “Tolerance requires us to accept people and permit their practices even when we strongly disapprove of them. Tolerance thus involves an attitude that is intermediate between wholehearted acceptance and unrestrained opposition” (187). Tolerance, as an attitude, thus seems to involve at least three distinct elements: (1) a sentiment of strong (moral) disapproval toward a person and/or her practices; (2) a desire to interfere with or prevent that of which we disapprove; and (3) a judgment, on moral grounds, that we ought not to act on this interventionist desire. It is important that our reason for restraint in (3) be a moral reason, rather than a prudential or practical one. A person will


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

not count as having an attitude of tolerance if she refrains from interfering with others only because she lacks the power to do so, or because she fears legal sanctions if she does so. Such a person may be said to tolerate others, as a matter of practice, but she cannot be said to have attitudes of tolerance toward them. Were she to acquire the power, or lose her fear of sanctions, she would gladly interfere with the activities of others, and do so without moral qualms. Do these three elements fully capture what is involved in the attitude of tolerance? Many philosophers apparently think so. If we reflect more carefully upon Scanlon’s description of the attitude of tolerance, however, it should be apparent that this threepart analysis is inadequate. For one of the crucial aspects of Scanlon’s description is the notion of “permission”: tolerance, he says, requires us to permit practices of which we strongly disapprove. It is important to emphasize that Scanlon’s definition is not idiosyncratic in this regard: the claim that tolerance is a matter of “permitting,” “deliberately allowing,” or “refusing to prohibit” something of which one disapproves is standard in discussions of tolerance. What is less often noticed, however, is that talk of “permission,” by its nature, implies a presumption of legitimate authority. I am only tempted to use the language of “permission,” after all, when I think that I am in a legitimate position to withhold it. I feel comfortable asserting that I permit my students to bring coffee to my classes, for example, because I think I have the legitimate authority to proscribe this practice, if I so choose. It can make sense, therefore, to say that I have an attitude of “tolerance” toward my students with regard to this practice (assuming that I disapprove of it, for some reason, but refrain from acting on this disapproval). By contrast, I would feel quite uncomfortable asserting that I permit my colleagues to bring coffee to our faculty meetings, as this would imply, falsely, that I take myself to have some legitimate authority over their conduct in this domain. To say that I have an attitude of “tolerance” toward this behavior on the part of my colleagues, therefore, would strike me as equally inappropriate. It might be objected here that there is something fundamentally misguided in the idea that attitudes of tolerance embody a presumption of legitimate authority. After all, those who take up attitudes of tolerance generally do so precisely because they think it would not be morally legitimate for them to interfere with some practice of which they disapprove. Part of the content of their attitude is the judgment that they should refrain from interfering with others, on moral grounds. So how could it possibly be part of the content of this attitude that they do have such authority? In response, I think it is important to distinguish between the general judgment that one stands in a position of legitimate moral authority over another, and the particular judgment that some specific exercise of that authority would or would not be morally legitimate. My claim is that attitudes of tolerance, by their nature, embody the general presumption that one stands in a position of legitimate authority over another. This presumption is implied, I submit, by the idea that tolerance involves “permitting” something of which one disapproves. Even if one takes oneself to have such general authority, however, there is a further question about whether and when the exercise of that authority would be morally legitimate. So there is nothing inconsistent in the

The Trouble with Tolerance


thought that attitudes of tolerance embody both the general presumption of legitimate authority over others, and the specific judgment that one should refrain from exercising this authority, for moral reasons. If this is correct, then I think we need to add a fourth element to our description of tolerance, in order to capture the dimension of presumptive authority this attitude seems to involve. In addition to (1) a sentiment of strong (moral) disapproval toward a person and/or her practices and (2) a desire to interfere with or prevent that of which we disapprove, I think we must also add (3) the (perhaps implicit) presumption that we have some legitimate authority over the conduct of others in the domain in question, along with (4) a judgment, on moral grounds, that we ought not to exercise this authority to interfere with or prevent that of which we disapprove. Once we recognize the complexity involved in the attitude of tolerance, we can begin to see some of the different ways in which such an attitude can seem inappropriate or less than fully admirable in certain contexts. The first way, of course, is if the original sentiment of disapproval toward a person or her practices is itself patently morally objectionable. As Scanlon notes, disapproval motivated by racial or ethnic hatred does not seem to call for tolerance so much as the eradication of the underlying feelings of hatred. Attitudes of tolerance directed toward people one despises on the basis of their race, ethnicity, or any other ascriptive feature, although better than attitudes of unrestrained opposition, are clearly “impure” in the sense that they embody an unreasonable and morally offensive antipathy. I will assume that if the original sentiment of disapproval is morally objectionable, then the associated desire to interfere with or prevent that of which we disapprove is also morally objectionable. So this is a second way in which attitudes of tolerance can be less than fully admirable in certain contexts. If my “tolerance” toward another is simply a matter of holding in check beliefs and desires that are themselves morally repugnant, it is certainly better than “intolerance,” but it is far from a morally ideal attitude to take toward others in society. Philosophers who write about tolerance have tended to assume that if there is something morally suspect about the attitude in a particular case, it must have to do with the inappropriateness of either the first or the second element (or both). This seems to be Scanlon’s assumption, for example, when he urges us to distinguish between “impure” cases of tolerance—where it is just an “expedient” for dealing with objectionable attitudes of disapproval—and “pure” cases, where neither the disapproval nor the desire to interfere with others is obviously morally suspect, yet we need tolerance to keep us from interfering unduly in the lives of others with whom we reasonably disagree. In the “pure” cases, Scanlon suggests, tolerance is a perfectly appropriate—indeed, morally admirable—attitude to hold toward others. This assumption, however, strikes me as premature, for it overlooks the significance of the third element in attitudes of tolerance. That element, recall, is the presumption that we have some legitimate authority over the conduct of others in some domain. What I want to propose now is that often it is concern over this third element, rather than one of the other two, that justifiably gives rise to feelings of resentment in those who are the target of tolerant attitudes. And if this is the case, then even “pure”


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

cases of tolerance might not be as morally unproblematic as Scanlon supposes. Let me explain. Presumably, aside from cases of pure prejudice of the sort mentioned above, no one should be particularly bothered by the fact that some people strongly disapprove of them, their beliefs, or their practices. I strongly disapprove of the views of (some) Republicans, for example, and (some) Republicans undoubtedly strongly disapprove of my views, but I take it there is nothing particularly troubling about this sort of reciprocal disapproval. Nor, I think, is there anything particularly worrisome about the fact that each side has a desire to prevent the views of the other from becoming dominant in society. We may regard such a desire as part and parcel of believing that the other side’s view is deeply mistaken. Rather, it is the third element, the presumption that one has some legitimate authority over the conduct of others in some domain, that can reasonably give rise to resentment in certain contexts. The problem, as I see it, is that in many situations this presumption is itself morally offensive, insofar as it implies an inequality of status between the person who tolerates and the one who is tolerated. To see what I have in mind, we might consider what attitude it is appropriate for me to take toward my fellow citizens in this contentious presidential election year. When I hear self-described “conservative” voters interviewed on the news, I often become agitated by what I regard as their deeply misguided views about gay marriage, immigration, and national security, among other issues. I have a strong desire that their views not become socially dominant, and I am motivated to argue against these views whenever I have an opportunity to do so. Yet I also believe that my fellow citizens have the same rights as I do to express their views, to campaign hard for their positions, and to vote for the candidate they think will best promote their values. Suppose, however, that I were to try to communicate this complex set of attitudes to one of these conservative voters by telling her that I have an attitude of tolerance toward her and her political activities. “I am quite tolerant of you conservatives,” I say. “So please, go ahead and campaign and vote for whomever you like.” My guess is that this announcement would not go over well, and for good reason. For to say that I am “tolerant” of conservatives or their views seems to imply that I take myself to have some legitimate, albeit unexercised, authority over them. It suggests that I see myself as “permitting” them to carry on with their activities, despite my severe moral disapproval of them. But such a presumption, I would argue, is itself grounds for reasonable resentment. I would likewise resent the suggestion that conservatives are tolerant of me and my political activities, and for the same reason: such an attitude is offensive, insofar as it implies that they take themselves to have some (unexercised) authority over me and my conduct. What is surprising here, if this diagnosis of the case is correct, is that there is nothing particularly offensive about being the target of attitudes of severe opposition and disapproval. I know that conservatives disapprove of my views and think that I am deeply mistaken, and they know that I feel the same way about them and their positions. So far, this is just healthy, albeit robust, disagreement. As long as we each recognize that the other has a right to hold, express, and advocate for her views, there is no reason to resent these attitudes of disapproval and opposition. It is only when the

The Trouble with Tolerance


supposedly positive attitude of “tolerance” comes onto the scene that I think there are legitimate grounds for complaint. The problem is that once one side sees itself as “tolerating” the other, this implies the judgment that there is an inequality in our moral status qua citizens. Now there is no longer an assumption that we are symmetrically related in our civic opposition, but one side views itself as “allowing” the other to carry on with its mistaken activities. And it is this presumption, I submit, that is offensive. Let me mention one other example to try to bring out the distinctive kind of reasonable resentment I think can be triggered in those who are the target of tolerant attitudes. In 2003, a group of students at Duke University decided that they wanted to do something to counteract the impression that Duke was a homophobic campus. After much discussion, they decided the best way to do this would be to print up and freely distribute t-shirts with the slogan “Gay? Fine by me,” and to encourage members of the campus community to wear these shirts on a designated day. The campaign was a huge success, and it soon spread to other college campuses across the country, including Notre Dame, Boston College, and the University of New Hampshire. I think it would be fair to say that this was, at bottom, an effort to communicate attitudes of tolerance to the gay and lesbian members of these campus communities, and I am sure many of the people who wore such shirts saw this as a way of expressing attitudes of tolerance toward them. But when we think about how this message of tolerance might have been received by the gays and lesbians on campus, I think we can again see the problematic underside of tolerance. For this message seems implicitly to embody an offensive presumption on the part of the person wearing the t-shirt—namely, that he or she has some special authority (perhaps in virtue of his or her heterosexual status) to grant or withhold permission to others to be openly gay on their campus. I suspect that the presumptuousness of this message does not show up clearly in this case because homosexuality is still, unfortunately, such a contentious topic in our society. But the presumptuousness of the message would be clear, I take it, if the shirts had instead sported one of the following slogans: “Jewish? Fine by me”; “Black? Fine by me”; “Celibate? Fine by me”; or “Pro Life? Fine by me.” In all of these cases, it seems to me, it would be presumptuous for a complete stranger to imply that she stands in a relationship to others such that her permission or blessing is either required or sought by those who possess the property in question. Yet this is, I think, precisely the message that is conveyed when one adopts an attitude of “tolerance” toward the gay and lesbian members of one’s campus community by telling them that their sexual orientation is “fine by me.” I want to make clear here that my point does not rest on the claim that attitudes of disapproval toward homosexuality or homosexual practice are patently morally objectionable. My claim is not that this is an “impure” case of tolerance, in Scanlon’s sense— that is, a case in which there should not be any disapproval in the first place, so “tolerance” should not be required. For better or worse, this is a live moral debate in society today, and I will assume that this is a topic over which citizens can, reasonably, disagree. My point, rather, is about the internal logic of attitudes of tolerance: To have an attitude of tolerance toward another, I submit, is to see oneself as allowing or permitting a trait or practice of which one disapproves. And it is precisely this offensive


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

presumption that I think is conveyed by the t-shirt slogan—that the wearer is granting permission to gays and lesbians to be themselves. What I hope these examples bring out is that we may have different intuitions about the value of tolerance depending upon whether we are on the giving or receiving end of this attitude. Many of us congratulate ourselves for our attitudes of “tolerance” toward people and views of which we disapprove, thinking of such attitudes as a kind of moral achievement on our part. Yet when we consider how such an attitude looks from the receiving end, I think it often appears less like a moral achievement than an offensive presumption. If I am right that attitudes of tolerance, by their nature, embody a presumption of legitimate authority over others, then it would seem such attitudes are morally appropriate only in situations where such legitimate authority is indeed present. I do not mean to deny that this condition is met in a wide variety of situations. For example, I assume that parents have such authority over their nonadult children. A parent might reasonably, and nonoffensively, adopt an attitude of tolerance toward her rebellious teenage son, because it is well within her authority to demand different behavior from him. I also assume that bosses have some legitimate authority over the conduct of employees in their workplace, teachers have some legitimate authority over the conduct of students in their classroom, homeowners have some legitimate authority over the conduct of guests in their houses, and restaurant managers have some legitimate authority over the conduct of patrons in their restaurants. In each of these cases, it can be perfectly appropriate to say that the individuals in positions of authority are “tolerant” of the behavior of individuals falling within their sphere of authority. So I do not mean to be raising a general doubt about whether attitudes of tolerance are ever appropriate. My doubt, rather, is whether it is ever morally appropriate for one person, qua citizen, to adopt an attitude of tolerance toward another person, qua citizen. And my suspicion is that the answer to this is a qualified “no,” because our status, as fellow citizens, is supposed to be an equal one. To adopt an attitude of tolerance toward a fellow citizen is to imply that you deny this equal status, and take yourself to have some legitimate authority over his or her conduct. And this, in itself, can reasonably be regarded as morally offensive. My reason for qualifying the claim that it is never appropriate to adopt an attitude of tolerance toward a fellow citizen will become apparent in section four. But for now, I want to explain why I think even the qualified claim poses problems for Scanlon’s defense of tolerance as an attitude. The central problem is easily, though perhaps paradoxically, stated: Scanlon’s interpretation of the tolerant person’s attitude toward citizens in society with whom she disagrees seems to rule out the adoption of tolerant attitudes by her toward those very citizens. Here is how Scanlon describes the tolerant person’s attitude toward her opponents: Even though we disagree, they are as fully members of society as I am. They are as entitled as I am to the protections of the law, as entitled as I am to live as they choose to live. In addition (and this is the hard part) neither their way of living nor mine is uniquely the way of our society. These are merely two among the

The Trouble with Tolerance


potentially many different outlooks that our society can include, each of which is equally entitled to be expressed in living as one mode of life that others can adopt. If one view is at any moment numerically or culturally predominant, this should be determined by, and dependent on, the accumulated choices of individual members of the society at large. (192) He goes on to note that, “What tolerance expresses is a recognition of common membership that is deeper than these conflicts, recognition of others as just as entitled as we are to contribute to the definition of our society” (193). The tolerant person, Scanlon claims, regards her fellow citizens as having an equal entitlement to all of the same rights, protections, and privileges that are available to her in society. But then, if my analysis above is correct, the tolerant person cannot regard her fellow citizens with attitudes of tolerance, because that would imply that she does not really see them as having this equal entitlement; rather, she regards herself as having a special authority to determine what their entitlements shall be. I should make clear here that I do not take myself to be disagreeing with the substance of Scanlon’s position at all. What I am disagreeing with, or raising doubts about, is Scanlon’s claim that he is giving a defense of tolerance as an attitude, rather than a defense of the attitudes that a tolerant person would take toward her fellow citizens. It is easy to assume that these must come to the same thing, but I hope my discussion above has shown that they do not. Tolerant people, I am inclined to say, are not likely to take up tolerant attitudes toward their fellow citizens. This leaves us, therefore, with the question of what, exactly, is involved in being a tolerant person. Once we focus on tolerance as a virtue, rather than an attitude, I think it will become clear why Scanlon thinks tolerance is something we all have reason to value.

§ 3. tolerance as a virtue

When we describe someone as a “tolerant person,” I take it we are attributing to her a certain character trait, understood as a complex set of dispositions to think, feel, and act in certain ways and for certain sorts of reasons. I assume, further, that we usually intend such a description to be complimentary: “She’s such a tolerant person!” is rarely, if ever, put forward as a criticism. This suggests that, as applied to people, tolerance is best understood as a virtue of character. Our question, therefore, is how we should understand the virtue of tolerance, and what sorts of attitudinal and behavioral dispositions we should expect from someone who possesses such a virtue. We might begin by considering a claim made by John Horton, in his discussion of tolerance as a virtue. Horton writes: The virtue of tolerance should include more than forbearance in not acting [sic] restrictively toward those who act objectionably; it should also include not having an excessive and inappropriate range of objections. The tolerant person


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

is not a narrow-minded bigot who shows restraint; he or she is not someone with a vast array of prejudices about others’ conduct but who nonetheless heroically restrains him- or herself from acting restrictively toward them. The restraint involved in toleration is not exclusively of action but also of judgment. The tolerant person is not too judgmental toward others. In becoming less judgmental, a person becomes more tolerant. Horton notes that part of what it means to be a tolerant person is that one is not overly judgmental toward others. Here we see an important difference between tolerance as a virtue and tolerance as an attitude. Tolerance as an attitude requires disapproval (for one can only have an attitude of tolerance toward something of which one disapproves). But one does not necessarily become more tolerant, as a person, by disapproving of more things. If that were the case, then the most tolerant person would be “the narrow-minded bigot” who disapproves of everything, but “heroically restrains” herself from acting on her prejudices. And this is certainly not what we have in mind when we think of a tolerant person. To be clear, Horton’s claim here is not that the most tolerant person is the one who has no objections to anything. One will not count as having the virtue of tolerance if one is indifferent to such things as child abuse, genocide, or rape. One is not being overly “judgmental” in regarding people who engage in such activities with horror and repugnance, and one will in fact be less than virtuous if one fails to recognize the moral viciousness of such behavior. Horton’s point, rather, is that one cannot really be said to have the virtue of tolerance if one has an excessive and unreasonable range of things that one finds objectionable. I think more needs to be said, however, about why the tolerant person is not overly judgmental of others. For it is one thing to be nonjudgmental because one simply does not care about other people and their practices, and another thing to be nonjudgmental because one has made efforts to try to understand and appreciate ways of life that differ from one’s own. A truly tolerant person, I would argue, is someone who has a disposition to learn about the differing social, cultural, and religious practices of others, particularly those that she finds initially alien or threatening. And such a person will be inclined to withhold judgment about these practices until she has had an opportunity to discover what it is that others find of value in them. This is not to say that a tolerant person will never conclude that another’s beliefs, practices, or ways of life are objectionable—it is a serious mistake to equate tolerance with relativism, or with uncritical multiculturalism. But a tolerant person will be less likely to object to other people or their practices simply out of ignorance or fear of difference. A person with the virtue of tolerance, therefore, will not be excessively judgmental of others, and will have dispositions to associate with and learn from people who differ from herself. The fact that a tolerant person will be open-minded in this way, however, does not mean that she will not have strong moral convictions of her own, convictions that may well conflict with those of other citizens in her society. So our next question is: How will a person with the virtue of tolerance approach deep disagreements with her fellow citizens?

The Trouble with Tolerance


It is here that I think Scanlon’s account—understood as a defense of tolerance as a virtue rather than an attitude—comes into its own. What Scanlon rightly stresses, in a way that I have not seen emphasized before, is that the virtue of tolerance requires more than simply a willingness to live with beliefs or practices that we may find deeply offensive; it requires, in addition, a willingness to live with the ever-present threat that the speech and conduct of others might lead our society to evolve in ways that we find deeply unwelcome. This is a frightening thing, because our ability to pursue ways of life that we find worthwhile can be deeply affected by how the prevailing customs, values, and practices of our society evolve and change. A tolerant person, Scanlon suggests, is one who recognizes this risk, yet is firmly committed to seeing her fellow citizens as equal members of society with an equal entitlement to contribute to its definition. This commitment is put severely to the test when we encounter citizens whose vision of what our society is and what it ought to be differs sharply from our own. To take just one of Scanlon’s examples, recent debates over the teaching of “creation science” in schools can be understood as just one front in a much larger battle over the question of what role religion should properly play in our public political culture. The extreme heat of these debates stems, in large part, from the fact that the partisans on each side feel deeply threatened by the underlying vision of society championed by the other. In such a context, it can be extremely difficult to maintain an attitude of respectful disagreement toward one’s opponents. In a passage of admirable candor, Scanlon confesses to encountering just such a difficulty in the regulation of his own attitudes. He writes: I firmly believe that “creation science” is bogus and that science classes should not present scientific theory and religious doctrine as alternatives with similar and equal claim to the same kind of assent. I therefore do not think that it is intolerant per se to oppose the creationists. But I confess to feeling a certain sense of partisan zeal in such cases, a sense of superiority over the people who propose such things and a desire not to let them win a point even if it did not cost anyone very much. In the case of science teaching, there is a cost, as there is in the case of school prayer. But I am also inclined to support removing “In God We Trust” from our coinage and to favor discontinuing the practice of prayer at public events. (196) This passage provides some helpful clues about how Scanlon thinks a person with the virtue of tolerance would approach deep disagreements with her fellow citizens. In the remainder of this section, I will flesh out some of these suggestions, and explain again why I do not think a distinctive attitude of tolerance has a natural place in Scanlon’s vision of the tolerant person. The first thing to note is that Scanlon clearly does not think that the virtue of tolerance requires us to be doubtful or skeptical about the correctness of our own position when we come into conflict with our fellow citizens. As he points out, it is not a mark of intolerance to believe firmly that one’s own position is correct and that the other


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

side’s is mistaken. If tolerance required us to give up or adopt a stance of skepticism toward our most firmly held beliefs whenever they came into conflict with those of others, it would not have a claim to be regarded as a virtue. The second thing to note, however, is that even if one is confident in the correctness of one’s own position, a tolerant person will not regard her opponents with attitudes of superiority, contemptuousness, or scorn. I think this is one of the most difficult, and most underappreciated, qualities of a tolerant person. It is a quality that is rarely evident in the partisans on either side of “the culture wars” in the United States, for example, which is part of what makes the debates in this area so ugly and polarizing. Liberals have a tendency to portray their opponents as backward, uneducated, and downright stupid; conservatives have a tendency to portray their opponents as selfish, licentious, and downright morally wicked. The problem with these attitudes is that they do not leave any room for the possibility of mutual respect in the face of deep disagreement. Citizens who view one another through these pejorative stereotypes cannot possibly engage in meaningful dialogue and debate over the future of their shared society. As Scanlon notes, such attitudes force us to see one another as “just rival groups contending over the same territory” (193). A person with the virtue of tolerance, therefore, will make every effort to distinguish between her attitude toward her opponents’ views and her attitude toward her opponents themselves (197). So far as possible, she will not allow her perhaps deep disapproval of her opponents’ views to lead her to demonize or cast aspersions on the character of her opponents themselves, and she will continue to accord them basic attitudes of respect and goodwill. I will qualify this point somewhat in the next section, but in general a tolerant person will maintain these attitudes of respect and goodwill toward those with whom she reasonably disagrees. There is one final feature of the tolerant person’s approach to disagreement that Scanlon gestures at in the passage quoted above. This final feature is what Scanlon later describes as “a spirit of accommodation, a desire to find a system of rights that others (all those within the broad reach of the relation ‘fellow citizen’) could also be asked to accept” (198). As Scanlon hints at in the passage above, and explicitly acknowledges later, it is this spirit that he fears may be lacking in his advocacy of removing references to God from our coinage and prohibiting prayer at public events. When one believes strongly in a principle—such as the separation between church and state, or a woman’s absolute right to choose whether to terminate a pregnancy— it is hard not to regard any departure from that principle as a threatening and unacceptable compromise, which should be vigorously and vehemently opposed. Scanlon’s suggestion is that a truly tolerant person will resist such forms of uncritical inflexibility, and will honestly and sincerely reflect on the question of whether, and if so where, she and her opponents might find areas of common ground. This kind of open-mindedness and spirit of accommodation is of course highly discouraged in our polarized political climate, and those who display such attitudes are often regarded as traitors in the causes they support. Nevertheless, I think Scanlon is right to see this disposition to look for areas of common ground as a significant and admirable feature of the tolerant person.

The Trouble with Tolerance


It seems clear, therefore, that tolerance, as a virtue, involves far more than merely refraining from interfering with people and practices that we find objectionable. It is a complex character trait that involves dispositions to judge, feel, and act in certain distinctive ways when confronted with beliefs and practices that differ from our own. I have identified four of these dispositions, but there are undoubtedly others that could be added to the list. In general, though, I have argued that a tolerant person will: 1. Not be excessively judgmental, but will also not hesitate to condemn practices that violate the legitimate moral rights and expectations of others. 2. Actively seek out opportunities to learn about beliefs and practices that differ from her own, and will withhold judgment about these beliefs and practices until she has a good understanding of why others find them of value. 3. Not abandon her own fundamental convictions whenever they conflict with those of others, but will also be cognizant of the fact that on many questions there is room for reasonable and respectful disagreement. As such, she will not hold attitudes of contempt or scorn toward those whose views differ from her own. In addition, she will regard her fellow citizens as equal members of society, who are just as entitled as she is to contribute to its definition. 4. Approach her disagreements with others in a spirit of accommodation, and will be honest, with herself and others, about what forms of compromise she can and cannot reasonably reject. One thing it is worth noting here is that it is not obvious where tolerance, as an attitude, would fit into a tolerant person’s stance toward her fellow citizens. Although the virtue of tolerance certainly requires certain attitudes (e.g., respect, open-mindedness, and goodwill, among other things), these are not themselves attitudes of tolerance. Given the tolerant person’s commitment to seeing her fellow citizens as equal members of society, with an equal entitlement to participate in the shaping of it, it would be odd for her to regard herself as “permitting” or “allowing” them to play a role in this process (as I am assuming an attitude of tolerance would imply). In the final section, however, I want to consider the possibility that there may be one circumstance in which a person with the virtue of tolerance may see herself as called upon to take up an attitude of tolerance toward a fellow citizen.

§ 4. tolerance in a nonideal world

In section two, I argued that attitudes of tolerance, by their nature, embody a presumption of legitimate authority over others, and that this is why such attitudes can be offensive when directed toward our fellow citizens. Insofar as we take ourselves to stand in a relation of equality with our fellow citizens, it seems inappropriate to hold attitudes toward them that imply an inequality of moral status. In the last


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

section, I suggested that a person with the virtue of tolerance would not be disposed to adopt such attitudes toward her fellow citizens, for precisely this reason. I must confess, however, that I have been working with a rather idealistic (some might say Pollyannish) picture of societal disagreement. Although I think it is true that if all of us had the virtue of tolerance, and approached our disagreements in the way described above, there would be no place for attitudes of tolerance toward our fellow citizens, the sad fact of the matter is that many of us do not approach our disagreements in this way. Some people regard others in society as worthless and inferior, and express their disdain through the language of hatred and exclusion. In such cases, it is not merely that the view being expressed is objectionable; it is that one group of people regards another as not worthy of a place in society at all. It is in precisely these sorts of cases that Scanlon claims it is essential that we distinguish between “one’s attitude toward what is advocated by one’s opponents and one’s attitude toward those opponents themselves: it is not that their point of view is entitled to be represented but that they (as fellow citizens, not as holders of that point of view) are entitled to be heard” (197). But this, I think, may be asking too much of the tolerant person, and may be granting too much to those who would deny others an equal place in society. It is perfectly appropriate to invoke this distinction when one’s opponent herself recognizes and is respectful of the equal membership of others in society. In such cases, drawing a distinction between the message and the messenger seems very much to the point. But I am not at all sure why we should regard those who explicitly deny the equal status of others in society as having an entitlement to be heard, given that they clearly do not recognize a similar entitlement in others. The entitlement to be heard, I am inclined to say, contains an implicit reciprocity condition: One can legitimately claim it only if one recognizes a reciprocal entitlement in others. I need to be very careful here to avoid misunderstanding. I am certainly not claiming that people who express attitudes of hatred, contempt, and exclusion toward other members of society should not be seen as having the same First Amendment rights to speak as everyone else. Our question here is not about what laws and protections regarding speech should exist, but about what attitude a tolerant person should take toward those in society who refuse to acknowledge the equal moral status of some of their fellow citizens. Although it may be true that a tolerant person will regard such people as having an entitlement to speak, as a matter of law, I am not convinced that, as Scanlon puts it, she must regard them as having an entitlement to be heard, as a matter of principled respect for them as fellow citizens. I would suggest that this latter entitlement, which we might understand as an entitlement to be taken seriously as a participant in what Scanlon calls “the informal politics of social life,” should not be viewed as unconditional. A tolerant person might justifiably withhold attitudes of principled respect from those who are unwilling to accord similar attitudes of principled respect to others in their society. To say that a tolerant person might justifiably withhold the attitude of principled respect from such citizens, however, does not yet tell us what attitude she should take toward them in its stead. And it is here, I think, that we may find a place for the

The Trouble with Tolerance


distinctive attitude of tolerance. If my analysis above is correct, citizens who are committed to denying the equal status of other members of society actually forfeit their entitlement to be taken seriously as participants in the informal politics of social life. In virtue of their own disrespectful speech and behavior, they in fact bring about a diminution of their own moral status vis-à-vis other citizens in society, for whom the entitlement to be taken seriously remains in full force. And it is precisely because they have introduced this inequality of moral status, I submit, that they may be legitimately subject to attitudes of tolerance. For if I am right that the problem with taking up an attitude of tolerance toward a fellow citizen in normal circumstances is that it falsely implies the existence of an inequality of moral status, then it would seem that this problem is not present in cases where such an inequality genuinely exists. This suggests that there is one case in which a tolerant person may be justified in adopting an attitude of tolerance toward a fellow citizen. What I have in mind here is something like the following: When dealing with citizens who themselves deny the equal membership of others in society, a tolerant person will not regard them as having an entitlement to be taken seriously in the informal politics of social life, but she may still choose to grant them a respectful hearing, nonetheless. In so doing, she could be described as holding an attitude of tolerance toward them: though she has the legitimate authority to refuse to take such citizens seriously as interlocutors in societal debates, she might refrain from exercising this authority, for moral reasons. In particular, she might refrain from exercising this authority in an effort to model for them how it is appropriate for a citizen to relate to other citizens in society with whom she deeply disagrees. Just as a parent might adopt an attitude of tolerance toward a teenage son who is making unreasonable demands—say, by engaging him in respectful dialogue rather than simply dismissing his demands out of court—so, too, a tolerant person might adopt an attitude of tolerance toward citizens who unreasonably refuse to acknowledge the equal membership of others in society. By continuing to engage in respectful dialogue with such people, despite their utter lack of entitlement to such treatment, a tolerant person demonstrates what it takes to be a citizen who is deserving of attitudes of principled respect. There is an element of condescension in this tolerant attitude, to be sure: a tolerant person will certainly see herself as granting to such citizens a privilege to which they are not, strictly speaking, entitled. But in cases of this sort, I think this element of condescension is perfectly warranted. Indeed, I am inclined to say that a tolerant person, in such a circumstance, is justified in viewing her fellow citizen as something closer to a citizen apprentice, who should be accorded attitudes of tolerance, rather than principled respect, until she has developed a more mature understanding of the demands and responsibilities of citizenship. I think James Baldwin gives beautiful expression to what this kind of attitudinal tolerance involves in a passage from his book, The Fire Next Time. Writing to his nephew James on the 100th anniversary of the Emancipation, he says: Please try to be clear, dear James, through the storm which rages about your youthful head today, about the reality which lies behind the words acceptance and


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

integration. There is no reason for you to try to become like white people and there is no basis whatever for their impertinent assumption that they must accept you. The really terrible thing, old buddy, is that you must accept them. And I mean that very seriously. You must accept them and accept them with love. For these innocent people have no other hope. They are, in effect, still trapped in a history which they do not understand; and until they understand it, they cannot be released from it. And he goes on to say, those innocents who believed that your imprisonment made them safe are losing their grasp of reality. But these men are your brothers—your lost, younger brothers. And if the word integration means anything, this is what it means: that we, with love, shall force our brothers to see themselves as they are, to cease fleeing from reality and begin to change it. For this is your home, my friend, do not be driven from it; great men have done great things here, and will again, and we can make America what America must become. There is no denying the element of condescension in Baldwin’s vision here. White people are portrayed as “innocent, lost younger brothers” who need to be rescued from their blind and self-destructive ignorance. Yet I take it there is also no denying the justice of this portrayal in the context of 1962 America. And what is striking about the passage is Baldwin’s insistence that African-Americans not turn away from their White oppressors in disgust, but rather embrace them, lovingly, as fellow citizens who have lost their way. Despite White America’s persistent and brutal denial of equal membership to Blacks, Baldwin still regards his oppressors as fellow citizens—indeed brothers, albeit lost, younger ones—in a single, shared society. This is a remarkably tolerant attitude to hold toward those who have consistently denied one’s own moral worth and humanity. And yet, despite how admirable Baldwin’s tolerant attitude was in the context of the time, we might still wish that he had never faced circumstances in which he felt called upon to adopt such an attitude toward his fellow citizens. It is for this reason that I said, at the beginning, that I do not think there are any “pure” cases of attitudinal tolerance, in the sense that Scanlon has in mind. Tolerance, I have argued, is always a “second best” attitude to hold toward a fellow citizen, because it implies an inequality of moral status that is in tension with the view of others as equal members of society. This is obviously true in those “impure” cases where tolerance simply holds in check sentiments of prejudice or hatred that we should not have in the first place. But it is equally true, I think, in cases where tolerance is itself a morally admirable response to the prejudiced and hateful attitudes of others. In both cases, we have reason to wish that the underlying attitudes that make tolerance necessary did not exist. In the meantime, however, we should be thankful for the presence of tolerant attitudes of both sorts, for they still have an enormously important role to play in the decidedly nonideal world in which we in fact live.

The Trouble with Tolerance


notes * I am greatly indebted to Sandra Reiter for enormously helpful discussion of many earlier versions of this essay. I also received extremely useful written comments from Jason Benchimol, Michael Blake, Jeremy Fischer, Rahul Kumar, Sally Markowitz, and Cathy Yu, and I also thank all of them as well as Ken Clatterbaugh, Paul Glezen, Sarah Goering, and Jean Roberts for equally stimulating conversations on this topic. Finally, I am grateful to the audiences at Willamette University, Washington and Lee University, and Indiana University for their many helpful comments and suggestions. Though I fear several of the people listed here disagree profoundly with my understanding of the attitude of tolerance, they have all been marvelous exemplars of the virtue of tolerance in their interactions with me. 1. For a nice overview of why tolerance is such a philosophically elusive concept, see David Heyd’s Introduction in his edited volume Toleration: An Elusive Virtue (Princeton, N.J.: Princeton University Press, 1996), 3–17. For a thought-provoking discussion of how we should understand the relation between practices or attitudes of toleration, on the one hand, and the virtue of toleration, on the other, see Bernard Williams’ contribution to this same volume, “Toleration: An Impossible Virtue?” 18–27. 2. Michael Blake has pointed out to me that the so-called paradoxes of tolerance arise not so much from this plurality of referents, but from the internal content of the attitude of tolerance, which requires us to hold attitudes of approval and disapproval simultaneously. Although I agree that many of the apparent paradoxes associated with tolerance stem from this internal tension, I think that some others stem from a lack of clarity about what the term “tolerance” refers to. In particular, I think the failure to distinguish between tolerance as an attitude and tolerance as a virtue can lead to the paradoxical claim that the most tolerant person is the one who disapproves of everything but who refrains from acting on her unreasonable set of antipathies. 3. T. M. Scanlon, “The Difficulty of Tolerance,” in Toleration: An Elusive Virtue, ed. David Heyd, 226–239. This essay is reprinted in Scanlon’s collection The Difficulty of Tolerance: Essays in Political Philosophy (Cambridge University Press, 2003), 187–201. Page numbers in the text refer to the reprinted version. 4. In “ What Tolerance Is” Ethics 115 (2005): 68–95, Andrew Cohen notes that tolerance as a moral attitude and tolerance as a moral virtue are different, but goes on to “treat these as equivalent” in his discussion (76). Bernard Williams also slips between talking about tolerance as an attitude and as a virtue in “Tolerating the Intolerable,” in Philosophy as a Humanistic Discipline, ed. A. W. Moore (Princeton, N.J.: Princeton University Press, 2006), 126–34. 5. Unless otherwise specified in the text, my specific concern here is with attitudes of tolerance taken by one citizen toward another citizen, or what we might call attitudes of civic tolerance. As I suggest later, there may well be instances of “pure” tolerance in other domains. 6. There is some disagreement over the question whether it is appropriate to speak of “tolerance” only in cases where one has a strong moral objection to some thing, or whether it is equally appropriate to talk of “tolerance” in cases involving merely dislike or distaste. For a defense of the former view, see Peter Nicholson, “Toleration as a Moral Ideal,” in Aspects of Toleration: Philosophical Studies, ed. John Horton and Susan Mendus (London: Methuen, 1985); for a defense of the latter view, see Baroness Warnock, “The Limits of Toleration,” in On Toleration, ed. Susan Mendus and David Edwards (Oxford: Oxford University Press, 1987). I wish to remain neutral on this question. 7. To mention just a couple of examples: In “The Instability of Tolerance,” George P. Fletcher claims that “tolerance presupposes a complexity of two sentiments: the first, an impulse to


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

intervene and regulate the lives of others, and the second, an imperative—either logical or moral—to restrain that impulse,” in David Heyd, Toleration, 158. Likewise, in “Tolerating the Intolerable,” Bernard Williams writes, If we are asking people to be tolerant, we are asking for something more complicated than [the elimination of attitudes of disapproval]. They will indeed have to lose something, their desire to suppress or drive out the rival belief; but they will also keep something, their commitment to their own beliefs, which is what gave them that desire in the first place. There is a tension here between one’s own commitments and the acceptance that others may have other and perhaps quite distasteful commitments. This is the tension that is typical of toleration, and the tension which makes it so difficult (in Susan Mendus, ed., The Politics of Toleration in Modern Life (Durham, N.C.: Duke University Press, 2000); reprinted in Bernard Williams, Philosophy as a Humanistic Discipline (Princeton, N.J.: Princeton University Press, 2006), 127. 8. D. D. Raphael describes tolerance as a matter of “deliberately allowing or permitting a thing of which one disapproves” in “The Intolerable,” in Susan Mendus, ed., Justifying Toleration (Cambridge University Press, 1988), 139. Unlike most writers on the topic, Raphael goes on to note that one can meaningfully talk of allowing or permitting only if one is “in a position to disallow,” and that if you are to claim any credit for your tolerance, “you must think that you have a right to exercise your power” (139). But then he makes the surprising claim that the judgment that something is morally wrong by its nature “implies that the something may be properly prevented,” so that the view of oneself as having a legitimate standing to interfere is, as it were, guaranteed by the judgment of moral wrongness. But this strikes me as exceedingly implausible. I may believe that adultery is morally wrong, for example, without thinking that I have any legitimate standing to prevent it. 9. John Horton claims that “the core concept of toleration is the refusal, where one has the power to do so, to prohibit or seriously interfere with conduct that one finds objectionable.” In “Toleration as a Virtue,” in Heyd, Toleration, 28. 10. One might think that it implies only that one has the power to interfere with others, but not necessarily the legitimate authority to do so. If we reflect carefully on the language of permission, however, I think we will find that it embodies a presumption of legitimate authority. It may be, however, that some people regard the mere existence of power as sufficient to establish legitimate authority. 11. I am grateful to Michael Blake for pressing me on this issue. 12. For the history of this movement, see 13. I am inclined, myself, to put sentiments of disapproval toward homosexuals into this category, because these sentiments often seem to be motivated solely by feelings of hatred and fear of difference. But because, sadly in my view, there is no consensus yet in our society that these sentiments are “patently morally objectionable,” I will leave open the possibility that attitudes of tolerance toward homosexuals might qualify as instances of “pure” tolerance of the sort Scanlon has in mind. Scanlon himself is less clear than one might have hoped in delineating what cases he sees as falling into the “pure” and “impure” sides of the tolerance ledger. It may be that when it comes to debates over public policies regarding homosexuals (such as laws protecting their equal rights in employment and housing, and allowing civil unions or marriage), we can distinguish between those who attempt to put forward legitimate public reasons for the policies they advocate and those who can only cite the alleged sinfulness of homosexuality in defense of their views. But there are obvious difficulties in drawing this distinction, and it would take me too far afield to pursue the question further

The Trouble with Tolerance


here. My suggestion that we might appeal to “public consensus” to determine which sentiments qualify as “patently morally objectionable” at a given time draws its inspiration from Cheshire Calhoun’s sensitive discussion of a related issue in her article “The Virtue of Civility” Philosophy & Public Affairs 29(3) (summer 2000): see esp. 272–73. 14. I should make clear here that I do not in any way doubt the good intentions or the sincerity of the students who conceived of this campaign, and it may be that it has, overall, done much more good than harm in making gay and lesbian students feel welcome on certain college campuses. My worry is simply that the message itself could be interpreted in a way that is offensive, insofar as it implies that certain members of the campus community (i.e., the heterosexuals) have the authority to determine whether it is “okay” to be gay on that campus or not. 15. Sally Markowitz has suggested to me that the discomfort many of us feel with the notion of “tolerance” may in fact be a function of its distinctive genealogy, as arising in a particular kind of pluralistic society in which certain social groups are dominant and others marginalized. Dominant groups, of course, get to determine what is, and is not, to be tolerated. Even when this leads to good things—the expansion of rights and privileges to previously oppressed groups—the logic of dominance and marginalization remains and reminds us of these systematic social inequalities. If this is correct, then there will always be a tension involved in the attempt to use the language of tolerance to describe a relation between (alleged) equals. I am grateful to Sally Markowitz for very helpful discussion on this point. 16. I may be “tolerant” of people smoking in my house, even though I disapprove of it and wish they would not do so. It is appropriate to describe my attitude in this case as one of “tolerance,” because I have the general authority to determine whether smoking is allowed in my own house. But I would not say that I am “tolerant” of people smoking on the street, as my authority over people’s smoking behavior does not extend into the public sphere. I am grateful to Lad Sessions for this example. 17. A restaurant manager might reasonably hold an attitude of tolerance or intolerance toward the use of cell phones in his restaurant, for instance, because he has the authority to determine what his cell phone policy shall be. I think it would be exceedingly odd, however, to say that I can reasonably take up attitudes of tolerance or intolerance toward a fellow diner’s cell phone use. Even if I shoot dirty looks at the cell phone user, make loud comments about the rudeness of using cell phones in restaurants, or approach his table and ask if he would take his conversation outside, I would not describe any of these activities as my failing to “tolerate” his cell phone use. As a fellow diner, I may urge him to consider the interests of others in the restaurant, but I have no legitimate authority either to grant or refuse him permission to continue talking on his cell phone. Talk of tolerating or not tolerating his behavior, therefore, strikes me as inappropriate. A restaurant manager, by contrast, could legitimately say to a patron, “I don’t tolerate cell phone use in my restaurant.” I realize that some people may have different intuitions about a case like this. Our use of the terms “tolerance” and “toleration” is quite expansive, and I can see how one might think that any decision to interfere with behavior of which one disapproves counts as a failure to tolerate it, and any decision not to interfere with behavior of which one disapproves counts as an attitude of tolerance toward it. What I am urging, however, is that we distinguish between cases of interference (or refusal to interfere) that are grounded in a presumption of legitimate authority, and those that are not. This is essentially a matter of the attitude with which we interfere (or refuse to interfere) with others. Do I see myself as refusing to permit


Ethical Themes: Contractualism, Promissory Obligation, and Tolerance

my fellow diner to talk on his cell phone, or as making a plea that he consider how his behavior is affecting others? I interfere with his behavior in either case, but it seems to me it matters a great deal what attitude I hold in doing so. The former attitude embodies a presumption of unequal status and addresses a demand that my fellow diner might reasonably resent. The latter embodies a presumption of equal status and addresses a plea that my fellow diner might refuse, but need not regard as presumptuous. 18. Michael Blake has objected to me that this misconstrues the true nature of the tolerator’s attitude. The tolerator’s thought, he suggests, is not that “I have legitimate authority over you and I’m not going to use it,” but rather “We (together) do indeed have the power here to coerce you, but because we respect your mistaken beliefs, we’re not going to use that power— because it isn’t legitimate to do so.” Although this may be correct, I think the basic concern still stands: one side is claiming a privileged status as a member of a “we” that has the general but unexercised authority to control the activities of the other. The implication is that the other is not also a member of this “we,” and to that extent is not an equal member of the society. I am grateful to Michael for pushing me to reflect further on this point. 19. This is not to say that people cannot be criticized for being “overly tolerant” in certain circumstances. But I take it such people would not normally be regarded as “tolerant people” but as “indulgent people.” 20. Horton, “Toleration as a Virtue,” 38. 21. For an interesting discussion of this aspect of Scanlon’s account, see Paul Weithman’s book review of The Difficulty of Tolerance in Ethics 114(4) (2004): 836–42. 22. One example of this is that many liberals now refer to Mike Huckabee, the evangelical Christian presidential candidate from Arkansas, as Mike Hickabee. 23. Here we might note the titles of three of conservative commentator Ann Coulter’s books against liberals: Slander, Treason, and Godless. Thanks to Michael Blake for the Coulter references. 24. It could be argued that Justice Stephen Breyer embodied this spirit of accommodation in his approach to the two Ten Commandment cases considered by the Supreme Court in 2005. In his book The Nine (New York: Doubleday, 2007), Jeffrey Toobin writes that Breyer, in effect, “told politicians to stop erecting provocative religious monuments, with the understanding that old ones could stay” (304). Breyer reasoned that one of these Ten Commandment monuments had been in place for decades without complaint or incident, whereas another was mounted recently with specific and provocative religious intent. Only the latter, therefore, could reasonably be said to be in violation of the Constitution’s separation of church and state. Strong defenders of the Establishment Clause no doubt viewed this as an unprincipled compromise, but it could also be viewed as an attempt to strike a fair and reasonable accommodation on church/state issues. (Whether it is appropriate for  Supreme Court Justices to approach cases in this spirit is, of course, another matter entirely.) 25. What I have in mind here is similar to the view Scanlon defends in his recent work on blame. What he calls the “normal moral relationship” with others involves certain default dispositions, such as dispositions to help others with their projects when this can be done at little cost, dispositions to take pleasure in others’ successes, and dispositions to hope that things go well for others. When a person’s attitudes indicate that she does not have reciprocal dispositions of basic goodwill toward others, however, Scanlon suggests that these default dispositions can appropriately be suspended toward that person. See Moral Dimensions : Permissibility, Meaning, Blame (Cambridge, Mass.: Harvard University Press, 2008), 143–45. Rahul Kumar has helpfully

The Trouble with Tolerance


suggested to me that many of the claims I am making about the virtue of tolerance might also (and perhaps more fruitfully) be cast in terms of the distinction Scanlon draws between “meaning” and “permissibility” in his recent work. What is valuable about tolerance, in Scanlon’s view, is that it expresses an attractive ideal of relations amongst co-citizens. The truly tolerant person is one who not only conforms to what the entitlements of others require (acts permissibly), but one who does so for certain reasons (thereby expressing a particular meaning through her actions). The tolerant person expresses through her actions that she properly values living with others on terms of mutual respect and recognition. I think there is something deeply right about this suggestion, though I do not have the space to explore it further here. 26. A tolerant person may choose to do this, but is certainly not required to do so. And in some circumstances, the view being espoused might be so hateful that it would be inappropriate for a tolerant person to grant the other a respectful hearing. I take it part of what is involved in having the virtue of tolerance is that one is able to determine when it is and is not appropriate to grant such a hearing. 27. James Baldwin, The Fire Next Time (New York: Vintage International, 1962). 28. Ibid., 8–9. 29. Ibid., 9–10.

This page intentionally left blank

part three Political Themes: Conservatism, Justice, and Public Reason

This page intentionally left blank

Attributed to Ludwig Wittgenstein: “I don’t mind what I eat, as long as it’s always the same.” Uttered by Morrie Cohen, and heard by me, many times, from around 1950 to 1985: “It’s not like it was.”

9 Rescuing Conservatism a defense of existing value 

G. A. Cohen

0. hegelian prelude

hegel says that “Spirit” achieves freedom when the subject finds itself in its own object, so that “it is at home with itself in its own otherness as such.” This essay explores modes of finding oneself in the other. I am not here interested in the characterization of that condition as freedom, not because that is an unimportant aspect of Hegel’s claim—it is, after all, his claim—but just because I fry other fish here. The conservatism that I defend is Hegelian to this extent: in each of three cases that I shall distinguish, namely, that of accepting the given, of valuing the valuable, and of valuing the valued, the subject is at peace with the object. For me, it is a pregnant moment in the New Testament when Jesus, awaiting his arrest in the Garden of Gethsemane and foreseeing the toils to come, cries out “Oh, Lord, take away this cup,” but then corrects himself: “but not my will, Lord, thine.” The motif is abandonment of striving, of seeking a better state, and instead going with the flow, as do the lilies of the field, which are at peace with the world, and therefore with themselves. There is a connection, as yet to be mapped, between the conservatism that I defend, and cherish, and the Gethsemane idea. 1. introduction

I have for decades harbored strongly conservative, that is, strongly small-c conservative, opinions, on many matters that are not matters of justice, and I here mount an exposition and the beginnings of a defense of what I believe to be my widely, although perhaps not universally, shared, conservative attitude. 203


Political Themes: Conservatism, Justice, and Public Reason

I do not have conservative views about matters of justice. Conservatives like me want to conserve that which has intrinsic value, and injustice lacks intrinsic value— and has, indeed, intrinsic disvalue. I also promote the distinct claims of what I shall call personal value, but I do not endorse the personal valuing of injustice. (I shall say something in section 6 about the relationship between small-c conservatism and large-C Conservatives, many of whom are indeed devoted to conserving injustice.) I am a kind of conservative not only in that I have the strong small-c conservative attitude that I shall describe, but also in that I endorse certain conservative factual assessments according to which a lot of valuable things have been disappearing lately. I join the ranks of the complainers down the ages who say: “Things ain’t what they used to be.” Do not suppose that, because that lamentation is perennial, it is misplaced. Anti-conservatives say, “But people have always said that things are getting worse,” and anti-conservatives mean thereby to convey that the conservative lamentation expresses an illusion. But it is entirely possible that at any rate certain kinds of things have always been worse than they were before. A wise Hungarian was once asked how things were going for him. He replied: “How are things going for me? Oh, you know, things are about average. Not as good as yesterday, better than tomorrow.” In fact, I think lots of good old things are being lost and lots of good new things are arriving. It is the conservative disposition, in the present sense of “conservative,” to lament the first fact more than nonconservatives do. But, and as I will explain in section 3, it does not follow that a person who is conservative in the present sense welcomes the second fact—that lots of good new things are arriving—less than nonconservatives do. In any case, there will be no defense of my conservative factual assessments in what follows. Please bear in mind throughout that I am trying here to describe, in an attractive light, one kind of conservative disposition, that is, my own. It is indeed my own disposition, and if I did not have it I would not have been motivated to write this essay, but I think that this disposition of mine is not an eccentric one: I think that everyone who is sane has something of this disposition, even if the people that I call conservatives here have a sturdier form of the disposition than others do. But let me add that I do not claim that the word “conservative” is more closely associated with the particular disposition that I shall describe than it is with every other disposition. The only claim that I make about the word “conservative” is that it applies without strain to the attitude that I shall describe, whatever the relationship of that attitude may be to conservatism more broadly understood: there is some attention to that theme in subsection (ii) of section 3 below. This is my first foray into this territory, and, if it seems to you that I am making some big mistake somewhere, or indeed throughout, then do not conclude that you are failing to see something. The probability is greater that it is I who am failing to see something. Making big mistakes is something that even the best philosophers do at early stages of research in philosophy. Therefore, a fortiori, it is something that I do, at least at early stages of my work on a theme. A final preliminary remark. In this essay I seek to identify some conservative truths that bear on choice and that, so I think, are too often neglected in our practical

Rescuing Conservatism


deliberations and that are certainly widely neglected in our philosophical deliberations about our practical deliberations. But please do not expect me to say to what extent our practice should honor the truths that I hope to expose, in comparison with other truths the honoring of which sometimes conflicts with honoring what I perceive to be certain conservative truths. Philosophers often have something novel to say about what, as it were, ingredients should go into the, as it were, cake even when they can say nothing about the proportions in which these ingredients, or values, are to be combined, across different cases: not because that is not important, but because the problem simply does not yield to general recipe making. Philosophers sometimes end their articles by saying this sort of thing: it is a task for future work to determine the weight of the consideration that I have exposed. Yet nobody ever gets around to that further work. Many wish they could, but nobody knows how to do it. Although philosophers cannot produce the weighing that is necessary in any practical discussion, their disposition to notice things in ordinary experience that other people miss means that they can nevertheless make a contribution to an immediately practical question. They can contribute by identifying a value that bears on choice and that is being neglected. Consider an analogy. A bunch of us are trying to decide which restaurant to choose. Suppose everybody talks a lot about how good the food is in various restaurants, how much it costs, and how long it takes to get there. Someone, hitherto silent, is uneasy. She feels that we have been leaving something out of account. Then she realizes what it is: “Like, nobody,” she says, “is considering the décor!” This person has made a significant contribution to our practical discussion. But we should not expect her now also to say exactly how important a restaurant’s décor is compared to the other things that matter when we are choosing a restaurant.

2. keeping valuable and valued things as they are, and accepting the given

Kenora Rainy River College is a Canadian liberal arts college (of my invention) that was founded before the First World War. From its foundation, it set itself the purpose of providing training for undergraduates that would be different from and better than what they would receive as part of a university that also has a graduate school. At Kenora there are no doctoral programs, and Kenora is proud of its undergraduateserving character. Some young new professors challenge the wisdom of the absence of doctoral programs. They urge that the institution could and would do more for education, including for the education of its own undergraduates, if it also trained graduate students. The presence of  more advanced students would stimulate the undergraduates, and the caring tutorial arrangements for the latter that are now in place could be preserved. The new professors urge that the mission of the College, which is to educate undergraduates, would be better realized under the suggested extension of the College’s remit, and that the wider purpose of serving education would also be advanced.


Political Themes: Conservatism, Justice, and Public Reason

Older professors resist the proposal for change. They look for flaws in the case that the new professors present. But one older professor presents grounds of resistance of a different kind. She says: I do not disbelieve that undergraduate education would be promoted by the proposed changes. Rather, I challenge a presupposition of the young professors’ argument, which is that everything that we justifiably decide to do may and must be justified as conducive to some good that our decision might produce. For, in addition to the consideration of what good we might do, which must of course affect our decisions, there is also the consideration of what we are, of our identity, and we may legitimately have regard to our desire to preserve that identity. I believe that it belongs to the identity of Kenora Rainy River, to our central organizing self-conception, that we are an institution that caters only to undergraduates. And I believe that it is worth preserving that status not only because it satisfies the legitimate desires of many of us but also because Kenora is a valuable social creation, partly because of what makes it different from otherwise similar social creations. As a valuable social creation, it merits preservation, and a radical enough transformation would induce both deformation of our identity and, with that, a loss of (some of) the distinctive value that the College embodies. We should not put in train a change that might cause or begin too radical a transformation of the College that we rightly love. Note that to believe that Kenora should be preserved in its distinctiveness is not to believe that it is the only beautiful flower in the garden. It is not to believe that an academic institution that sustains a more maximizing attitude to educational output does not also have its own distinctive value. On the contrary: the old professor’s conservative attitude is at least de facto congenial to variety, because much variety reflects accident, and conservatism is constitutively friendly to the results of accidents (because of its tenderness towards already existing value and therefore to whatever value merely happens to exist: I shall say more about that in section 3 below). Two arguments are ingredient in what the professor says in favor of Kenora keeping itself (more or less) as it is, in the stated respect. One is that it is the legitimate desire of its members to preserve their particular corporate identity. The other is that Kenora is a valuable creation that everyone, therefore, and not just the College’s members with their particular desires, has reason to wish to see preserved, with its distinctive value intact. The two arguments correspond to a distinction between two ways of valuing something other than as a pure function of the amount or type of value that resides in it. In the first way of valuing that I have in mind, a person values something because of the special relation of the thing to that person. In the second way, a person values something as the particular valuable thing that it is, and not merely for the value that resides in it, but not, in this second case, because of her own special relationship to the thing in question. “Particular valuing” is discussed in section 3, and “personal valuing” in section 4.

Rescuing Conservatism


As I said, both of these valuings, valuing the personal and valuing the particular, as such, value a thing other than solely on account of the amount and type of value that resides in the thing. But that is so for contrasting reasons in the two cases. It is true of valuing the personal, because what is then valued need not have any intrinsic value, or, if it does have intrinsic value, it need have only little, in comparison to the amount of personal value that it has. And it is true of valuing the particular because, even though the particular indeed gets its value, in the first instance, from the intrinsic value that it has, our valuing of it, the particular, is not merely a valuing of the intrinsic value that it has, but also a valuing of it, the particular itself. Its personal value for them and its particular value for anyone are not the only considerations that should govern the decisions of its professors about the future direction of Kenora. But it must be wrong to omit them from consideration. Because the College is a valuable human creation, it is not right to treat it as a mere means for the production of good results, as we do if we ask only what is the best that can be got out of it, or the best that can be made of it, just as that is not the only question that we should ask about a human being; and there is additional reason for it not to be the only question when the human being is oneself (or when one is part of the human creation in question). Just as you may love somebody because of who and what they are, rather than just for the value of what they produce and for the value of what they instantiate, so you may love a loveable institution because it is the institution that it is and it possesses the character that it has. So if you seek to set the agenda for an institution, you must ask not only what its goals are and should be, and how it may best achieve them, but also what it, the institution, is. And you have, once again, additional personal reason to do so, reasons of specifically personal value, when you, collectively, constitute the institution in question. Conservative conviction, as I understand it, thereby exhibits a bias in favor of retaining what is of value, even in the face of replacing it by something of greater value (though not, therefore, in the face of replacing it by something of greater value no matter how much greater its value would be). But there is a third idea, beyond preservation of the (intrinsically) valuable and the (personally) valued, in conservatism that warrants notice here, namely, the idea that some things must be accepted as given, that not everything can, or should, be shaped to our aims and requirements; the attitude that goes with seeking to shape everything to our requirements both violates intrinsic value and contradicts our own spiritual requirements. Given things that are of value ought to be retained, but, beyond that, we need some things to be given quite apart from whether they have value other than because they are given. It is essential that some things should be taken as given: the attitude of universal mastery over everything is repugnant, and, at the limit, insane. (Note that the idea that some givens warrant preservation other than because they are (otherwise) valuable is consonant with what is said in the penultimate paragraph of this section about keeping bad human features.) That we must accept some givens, not any and all givens, but plenty of givens, was well illustrated by my daughter Sarah Cohen’s wise reaction to the title of Jonathan


Political Themes: Conservatism, Justice, and Public Reason

Glover’s book of 1984, What Sort of People Should There Be? Sarah, then ten years old, noticed the book and its title on my desk and exclaimed: “That’s div! They should be like us.” Certain things are to be accepted from nature, and that includes aspects of ourselves. That wisdom is part of the wider conservative thought that certain things are to be taken as they come: they are not to be shaped or controlled. Consider this allegory. Quite far along a certain continuum there sits a man who is surveying his own fleshly parts, that is, those of his parts which are still made of flesh, which includes some of his brain-flesh parts, and he is replacing defective bits of his flesh by perfect artificial substitutes, made out of whatever best serves, such as silicon, tungsten, reprocessed dung, and so forth. The man has been doing this for some time, and a lot of him is already artificial. That is surely a ghastly scenario. But it conveys the attitude that many people manifest towards human frailty and deficiency, or, more accurately, the attitude that they defend. That phrasing is more accurate because no one really has the artificial man’s attitude in practice: in practice, everybody is conservative to some degree. One commentator on the foregoing objected that the example is biased, because the deficient fleshly parts are replaced in the example by undeficient, unfleshly, artificial parts, and, to test for the presence of the conservative attitude that I extol, we should consider the replacement of bits of our bodies by artificially produced bits of flesh in particular, which have been genetically modified to match our own, save for the elimination of the undesired imperfections. I reply that it is hard to see why, in the absence of any conservative impulse, we should prefer flesh to other materials that ex hypothesi do whatever we had expected flesh to do. The alternative scenario is indeed less ghastly, but precisely because it at least retains the medium (i.e., flesh) in which our capacities have traditionally been embodied. The given flesh is indeed replaced, but the attitude of preserving the given is more satisfied when flesh, rather than silicon, does the replacing: thereby the given form of embodiment is retained. There is a particular kind of slippery slope in these matters that commands attention, and I do not mean one that relates to the difficulty of drawing a line between what we (now) find acceptable and what we (now) find unacceptable. Suppose we face no such difficulty at all, and we think that the right limit to plasticization is given by precisely this amount and type of plasticization. But we also know that, once thus plasticized, people would favor more plasticization, and, perhaps, unending plasticization, as each successive plasticization renders people insensitive to the extra plasticness of the next one. Then we have strong reason to bring about less plasticization than we think optimal (though not, of course, than we think it optimal for us to bring about, all things considered). The problem that we face is not that we do not know where the line should be, but that, if we draw it at that point, then they, or, for that matter, we, later, will draw the line improperly elsewhere, given what their, or our, desires and tolerances will consequently have come to be. Partly because of fear of such a slippery slope, I am wary of certain forms of stemcell research even as a means of finding a cure for horrible diseases. A culture in which such research is routine practice will smile on the next stage of the use of living beings

Rescuing Conservatism


for research in the way we, who may find current forms of stem-cell research acceptable, do not. We then have strong reason to prevent that successor culture from supervening. We have reason not to bring into being people for whom what we regard as acceptable practice is the accepted practice because they may go beyond us in what they find acceptable, perhaps beyond their existing practice to just the extent that we went beyond our once existing practice, and, therefore, further than we would have gone and than we think anyone should go. If there is such a propensity in human nature, to go just that bit beyond where we are now in a dangerous direction, then we might owe it to the future of humanity, though not to the desires of what threaten to be future humans and super-humans, not to start the journey. And do not say, “If your child’s life depended on it. . . .” There are all kinds of awful things that I would not otherwise dream of doing that I might do if my child’s life depended on it. When people say: “If you had cancer . . . ,” one can sometimes reply: “Yes, of course, that might unbalance my judgment.” Making people imagine that they are in dire straits in order to cause them to agree with something is an attractive resort for those whose arguments are not (otherwise) strong. I said that I was wary of certain forms of stem-cell research: I did not say that I would prohibit them. My philosophical opponent is not one who says, “Yes, there’s something unattractive about using human life like this, but when you consider the gains. . . .” My philosophical opponent is rather one who says, “There’s nothing wrong with using these life-forms in this way, and, since the gains are so great. . . .” The word “that” is subject to different interpretations when it appears in the protest, “How can you deny us a cure for x for the sake of that?” “That,” here, might be thought by the speaker to denote something that is nothing, but it might also be thought by her to denote not nothing but something that is not enough. Although we exhibit a difference of judgment, I might have no philosophical quarrel with the speaker who means the latter. If I want us to continue as we are, do I want us to retain our negative features? What if a genetic manipulation could, for example, eliminate envy? One might fear, of course, that with the elimination of envy there would also disappear certain possibilities of virtue: most straightforwardly, the virtue of rising above one’s own envy, but also, perhaps, less evident ones, that are more secretly tied to envy, or to its possibility, in the economy of the human psyche. But even once we have set such collateral considerations aside, I would not want to eliminate all of our bad features. I conjecture that that is partly because the negative traits are part of the package that makes human beings the particular valuable creatures that we personally cherish, and are therefore worth preserving as part of that package, but it is also partly because we court vertigo if we seek to place everything within our control. I proceed to distinguish between particular valuing (section 3) and personal valuing (section 4). Correspondingly, one might distinguish between the reason to preserve human beings—that they are creatures that exhibit a certain form of value, and our (additional) reason to do so, which is that they are us. But beyond those two themes, which we can call preserving the valuable and preserving the valued, a third has obtruded, which will not be revisited in the sequel, and that we can call preserving


Political Themes: Conservatism, Justice, and Public Reason

the given (as such). I set it aside not because it is unimportant but because it is too deep to explore here. 3. conserving the valuable versus conserving value (i) Conserving What Has Value: The Nature of the Thesis

The conservative propensity that I defend in this section is to preserve particular intrinsically valuable things, as such. There is a distinct and relatively well-known case for preserving things that are personally valuable, and I pursue that case in section 4. That well-known case has strong sociopsychological components, but they play no role in the more metaphysically phrased defense of preserving what is valuable that I offer in the present section. The recommended solicitous attitude to existing valuable things (precisely) does not reduce to endorsing the purposes that the things serve, or the principles that they exemplify, which might be better served or exemplified by things that do not exist and which could be created at little cost, but by which we should nevertheless not replace existing valuable things. (I do not mean that we should not replace them no matter what the cost of forgoing a superior alternative may be: the bias that I assert in favor of existing value is not absolute). Note that, even if it is worth keeping something only if it satisfies some purpose or principle, it does not follow that the weight of the reason for keeping it is limited to the extent to which it satisfies that purpose or principle. The conservative impulse is to conserve what is valuable, that is, the particular things that are valuable. I claim that we devalue the valuable things we have if we keep them only so long as nothing even slightly more valuable comes along. Valuable things command a certain loyalty. If an existing thing has intrinsic value, then we have reason to regret its destruction as such, a reason that we would not have if we cared only about the value that the thing carries or instantiates. My thesis is that it is rational and right to have such a bias in favor of existing value, that, for example, if you happily replace a fine statue by a merely somewhat better one, the production of which requires destruction of the original statue, then you mistreat the now destroyed work as (so to speak) having had the merely instrumental value of being a vessel of aesthetic value. A salient, though not the only, alternative to conserving what is valuable is to maximize value, but conservatives are resolved to conserve the valuable at the expense of maximizing value: what we distinctively value are the particular bearers of value. The conservative attitude that I seek to describe is a commitment to the conservation of what has value. “Conservation of what has value” is, indeed, the canonical phrase here, and not “conservation of value.” For if we take the phrase “conservation of value” in the standard way, on the model, that is, of “conservation of energy,” or “of matter,” a conservation of value policy is entirely unconservative (in the present sense of “conservative”). Conservation of what has value is not in that sense, in the conservation of energy sense, conservation of value, for you lose no value, value itself is conserved, when you destroy something valuable and replace it by a thing of the same value. The conservative disposition is not to keep the value rating high but to keep the things that now contribute to that rating.

Rescuing Conservatism


Valuing something is in some respects like loving it, whether or not either is a species of the other. Consider, then, two old popular songs about love, here used as means to make the distinction I need, or one structurally parallel to it. First, there is You, You, You, an Ames Brothers’ hit of 1953: You, you, you I’m in love with you, you, you, I could be so true, true, true, To a girl like you, you, you. If I were the girl addressed by the Ames Brothers, I would beam during the first three lines of their stanza but my beam would give way to a frown when I heard the fourth line. Sure, I want to be loved because of my qualities, because of what I’m “like,” but once my lover makes that move from my attractive qualities to me, I want it to be me that he loves, and not just the qualities. I do not want to think that, if someone with the same qualities now came along, then he would be indifferent between staying with me and switching to her, or that he would be even truer to someone who had even better qualities than I have. I do not want it to be inscribed in my lover’s love for me that I am a good illustration of my particular virtues, and if my qualities decline, then I want love not to alter when it such alteration finds. I want to be loved not as a good bearer of my virtues, but as the bearer of them that I am, or, indeed was. By contrast with the Ames Brothers, Olivia Newton-John starts and finishes in the right way. She does not go from “you” to “a girl”: she makes the necessary move from “a” to “the.” She begins with, “You better shape up, because I need a man,” but, when she finds one who has shaped up, he becomes “the one that I want, the one that I wa-ant (ooo, ooo, ooo).” You might think that it does not distinguish conservatives that they want to conserve what has value, because everybody wants to conserve what has value. I have two responses to that objection. First, although, so I would myself insist, it is indeed true that everybody has some desire to conserve what has value, it distinguishes conservatives of the present stripe that they want to conserve what has value more than other people do. Second, and more importantly, and, even though it is true that everyone wants to conserve what has value, not all who reflect on value and our relation to it notice the truth that I am defending: Many philosophers, and most nonphilosophers when they philosophize, ignore the truth in conservatism that everyone who is sane recognizes and honors in practice, to some degree. Among the philosophers that I have in mind are utilitarians, who purport to see nothing wrong with destroying value, if more value results. To seek to maximize value is to see nothing wrong in the destruction of valuable things, as long as there is no reduction in the total amount of value as a result. Unlike the conservative, the utilitarian is indifferent between adding to what we have now got, at no cost, something that has five units of value, and adding something worth ten units of value at the expense of destroying something worth five. The utilitarian says: “Let us have as much value as possible, regardless of what happens, as a result of that policy, to existing bearers of


Political Themes: Conservatism, Justice, and Public Reason

value: they do not matter, as such.” Conservatism sets itself against that maximizing attitude, according to which the things that possess value, by contrast with the value they possess, do not matter at all. The Kenora Professor’s resistance to a radical change in the nature of her College does not depend on a belief that what would thereby supervene would be less valuable than what would thereby be lost. Conservatism is an expensive taste, because conservatives sacrifice value in order not to sacrifice things that have value. We keep the existing particular valuable things at the expense of not making things in general as valuable as they could be made to be. Value, one might provocatively say, is not the only thing that is valuable: so are particular valuable things. And the two desiderata sometimes need to be traded off against each other. A nonutilitarian pluralist value-maximizing consequentialist, who believes in maximizing the aggregate of irreducibly different kinds of value, might regret destroying one kind of value as part of a project which produces more value overall. But, once again, and given that she is indeed a value-maximizing consequentialist, she cannot regret its destruction as such, as opposed to the nonappearance of anything that has that kind of value within the set of things that are there when day is done: She has no more reason to regret the destruction of something which has a specific kind of value than she has to regret no such thing being there now, and/or no such thing having been there in the first place. Value-maximizing consequentialism stands in especially sharp contrast with conservatism as I have defined it, but a view need not be maximizing to be unconservative. Maximizing consequentialism, as understood here, is only one species of conception in which the bearers of value, as opposed to the value they bear, do not count as such, but matter only because of the value that they bear, and are therefore, in a deep sense, dispensable: That is the essential, and canonical, contrast: all such conceptions are rejected by the principle that valuable things warrant conservation. Thus, if one may be a “sufficientarian” with respect to the promotion of value (as opposed to, very differently, with respect to distributive justice), if, that is, it is possible to think that all that matters is that the world be made good enough, then conservatism in the present sense opposes that, too, as a sole normative principle about value. It follows from the essential contrast that, as Michael Otsuka has shrewdly observed, a Cohen-conservative, although not a (comprehensive) value-maximizer, might yet be a maximizer, in a certain local sense: [W]e might have a [Cohen-]conservative who thinks we should, above all else, maximize the quantity of preserved value by, for example, destructively throwing one of Michelangelo’s Slaves into the path of the oncoming trolley so that it causes the trolley to come to a halt before it builds up enough speed to run over and destroy the other five Slaves. Indeed, what reason could he have not to do it? The doctrines of doing and allowing and of double effect have purchase with respect to the treatment of human beings, but maybe not to the treatment of things. Or if, as I contemplate in subsection (v)

Rescuing Conservatism


below, there are deontic obligations not to destroy things that have no rights (i.e., obligations that carry no corresponding obligation to prevent their destruction), then, even so, my argument does not stand on such deontic ground, as is, I hope, made clear in the said subsection. The conservative propensity is to conserve, to not destroy, and, therefore, to not replace, even (within limits) by something more valuable. But it is not a propensity against creation of new things, save, perforce, when their creation requires or causes a destruction of existing value. There is no conservative objection, in my sense of “conservative,” to things of new kinds, as long as they leave the old things intact. A conservative is not against the creation of a Hockney painting, but she resists destroying a mediocre Filippo Lippi to make the creation of a Hockney possible, artistically superior though she might think the Hockney is bound to be. I said in section 1 that the mark of a conservative is that she laments the loss of existing good things more than a nonconservative does. It does not follow, and it is not necessarily true, that a nonconservative welcomes the good new things more than a conservative does. You can consistently adore Byzantine icons partly because of their oldness and Frank Gehry architecture partly because of its newness. When conservatives lament what has been lost, say, for example, traditional craftsmanship, when they lament, William-Morris-like, that now everything is made by machines, then some anti-conservatives scornfully reply that the world we have lost was a world of poverty and toil, and that conservatives are irresponsible romantics. But the proposition, which I have no reason, here, to deny, that humankind is a net beneficiary of modernization, is not a reason for not lamenting what has been lost, unless you are a nonconservative evaluator. A conservative can believe that what rises from the ashes is the greatest building ever and that it was right to build it, yet still feel distraught that the old building was destroyed. A conservative can also believe that the culture of a people would be much richer if its present culture were replaced by a different once, and yet resist that destruction, not only because the people are attached to their culture (a matter on which the discussion in section 4 below bears), but also just because of the culture’s existing value. (ii) Some Limitations of My Thesis

Let me now state some limitations of my thesis, considered as an attempt to formulate a central conservative thought. Note, first that the thesis does not manifestly endorse longevity as such. The special claim that I have defended is of the value that exists, regardless of how long it has been around. Even if the picture was painted only five minutes ago, there’s a reason not to destroy it in order to use its pigment to produce a better one. Richard Christian characterizes a preference for what I call, here, “longevity,” as embodying a thesis about the behavior of value, which says that it increases over time. He says that the stated thesis is wholly independent of my own central claim that existing value is to be favored, which, he illuminatingly says, is a thesis not, like the longevity thesis, about the behavior of value, but about the correct response to value. If Christian’s independence claim were true without qualification, then that would be a pity, because it would mean that


Political Themes: Conservatism, Justice, and Public Reason

there is a lack of connection between the conservatism defended here and more traditional conservative ideas, and that would make my distinctive claim less interesting than it would otherwise be. The question of whether the two themes are in some way connected requires further study. (To be sure, we have reason, if I am right, to keep on retaining what is valuable, and longevity will be a by-product of doing so, but that does not stamp value on longevity itself.) Let me note a further limitation (beyond its possible indifference to longevity) of the position that I am defending. Consider the second epigraph to this essay: “It’s not like it was.” The remark expresses a familiar attitude that we would normally call “conservative,” namely a certain bias in favor of reversing recent change. The position that I defend in the present section cannot account for that bias, and indeed it rather runs against it, because reversing recent change means destroying what now exists. It may, of course, be metaphysically impossible to reverse recent change by (literally) bringing back into existence a particular thing that has gone out of existence. But you can sometimes bring into existence a true replica of what existed, or a feature or set of features that used to obtain but no longer do. And some desire to do so is certainly part of what we should normally consider the conservative attitude. The desire can be defended on grounds related to the theme of section 4, which is that familiar things and usages that are lodged in our lives have a personal value for us that makes them worth keeping, and worth “restoring,” through a replication of their features, when they have been destroyed recently, and when we are therefore still oriented to a life that includes them. But the conservatism defended in this section seems to offer no reason for restoring recently disappeared features. (iii) Objections (a) The Regret Objection

I say that if, despite the gain in value that necessitated its destruction, we regret destroying the old building, that is because we have a certain bias in its favor: We are, up to a point, disposed to retain it rather that replace it by something better. But, so it might plausibly be said, we also regret not building the new more valuable building when we decide to retain the existing one, so why should we not also infer that we have a bias in its favor? Reply: It does not follow logically from the first regret that we have the stated bias. But I say that it is impossible to make sense of the first regret unless we value what exists “beyond its intrinsic value,” unless, so to speak, we “overvalue” it, in the relevant sense. By contrast, the best explanation of the second regret requires positing no bias in favor of the unrealized future more valuable thing: the regret is simply explained by the increase forgone in intrinsic value itself. There is no “overvaluing” bias here. (b) Lippi-Hockney

Return to the Lippi/Hockney example. I said that a conservative does not oppose new creation, as long as the new creation leaves the old creation intact. But someone might object: If people look at a Hockney, there is that much time spent not looking at a

Rescuing Conservatism


Lippi. So new creation can imply replacement, and my conservatism therefore cannot be so sanguine about new creation. But then I am in trouble, because I surely do not want to prohibit or restrict access to Hockneys. But the propounder of that argument mistakes the nature of my conservatism with respect to valuing the particular. For I do not celebrate, in the first instance, our experience of the valuable things, but, instead, the valuable things themselves. And the only way that the value inherent in the Lippi can be destroyed, the particular value that I have in mind, is by destroying the Lippi. Surrounding it by a distraction of Hockneys does not do that. Someone might protest: If it has value independently of its contribution to our experience, doesn’t it follow that it would be good to preserve a work of art even if it  were no longer to be perceived, and even in a perceiverless world? And aren’t those consequences absurd? In order to address that pair of questions, let us distinguish between the case of a world of blind people and the case of a world of no people, or other relevant perceivers. In the first case I think the blind people could value the fact that their world contained such beauty, even though no one could appreciate it. So it might indeed follow from my position that it is good that unperceived aesthetic value exists. But I do not find that embarrassing. And if it also follows from my position that something could have aesthetic value even in a wholly perceiverless and conceiverless world, then some will no doubt want to get off the bus there, but I would ride on even then. If we had 1,000 paintings and we knew that 100 will never be looked at again, but we did not know which ones they were, would we say: “If only we knew which are the hundred that no one will ever look at. Then we could harmlessly destroy them, and use the space that they now occupy for something useful”? If more space were imperative, would we just as soon throw those never-again-to-be-experienced hundred paintings into the fire as we would some nondescript logs of wood? Note that the questions that were posed for my view three paragraphs back may also be posed, mutatis mutandis, for a subset of my opponents, namely, those who believe that value (itself) is as intrinsic as I believe it to be, but not that existing bearers of value enjoy a special status. And the answers to those questions will be essentially comparable to the ones that I gave. It is, accordingly, inappropriate to press skepticism about intrinsic value in the particular context of this essay’s polemic. (c) The Irrationality Objection

Many would say that attachment to the value of particular things, in opposition to maximizing value, or, indeed, to value satisficing, or to any view that fails to put a premium on existing value, is irrational. Why have less value (or, for example, less than enough) when you can have more (or at least enough)? Well, ex hypothesi, the attachment is not justified by the contribution that the thing makes to the general value account. But it does not follow that the antimaximizing (anti-etc.) stance is irrational, and I do not believe that it is: It is an error nourished by much academic work, especially in economics, but no doubt the error is of deeper than merely academic origin—it is an error to think that being


Political Themes: Conservatism, Justice, and Public Reason

rational about value requires the merely abstract accounting of it that denies the value of particular things as such. And even if you think that I am wrong to call that an error, even if you think that there is never a reason to honor an embodiment of value, as such, I am sure that you cannot think of a noncircular argument for that conclusion: for to what could you appeal, to judge between the claims of abstract value and the claims of particular value? (Some are quick to suppose that favoring existing value must reflect a fear that what would replace it would be less good, in universal terms. I cannot refute that hypothesis. But I see no reason to believe it, unless it has already been groundlessly decided that a preference for existing value is irrational and therefore in need of some such explanation as that the preference reflects anxiety about loss of abstract value.)

(d) Wiggins’s Rosebush

David Wiggins has proposed a counterexample to my view, namely, a rosebush that stood in his garden, which served no external purpose, and which he therefore valued intrinsically, but one towards which, he says, he did not exhibit the bias that I say goes with valuing something intrinsically. When Wiggins was dismayed to see that his gardener had cut down the rosebush, he was fully assuaged when the gardener said that she would shortly replace it by a more beautiful one. I can think of three saving responses to the Wiggins counterexample, but I am not sure which, if any, is sound. First, one might conclude that Wiggins must have valued the bush not for its intrinsic value but merely for the pleasure it gave him. But I think he would protest that the stated response is untrue to the phenomenology of his experience: He would say that it gave him pleasure because he recognized its value. Second, one might conjecture that, despite his protest to the contrary, Wiggins did have a bias in favor of his rosebush, but not a very strong one, so that any regret at its destruction was swamped by the good news that a better one was coming. And, third, one might hold that the bias that I argue for is appropriate only once a certain threshold of intrinsic value has been passed, and that the value of a single rosebush falls below that threshold. The second and third responses might be said to gain a certain plausibility from the reflection that Wiggins’s rosebush example is more arresting than would be an otherwise parallel one in which a complete and luxuriant garden is replaced by an even better one. (e) McHose’s Blocks of Wood

I wish to notice an interesting counterexample to my thesis that is due to Brad McHose. A sculptor contemplates producing a wooden statue, one that he can fashion in either one of two ways. One way involves assembling unpainted wooden blocks into the form of the statue and then painting them. Prior to being painted, the block formation, let us suppose, is a statue that has intrinsic value. Thus painting it “destroys” something of intrinsic value, albeit in order to create something of (we can suppose)

Rescuing Conservatism


greater intrinsic value (the painted statue). But the sculptor can also produce the same painted statue without destroying anything of intrinsic value by painting the blocks prior to assembling them, because in this different case she does not (we may stipulate) produce anything of intrinsic value until the painted statue has been assembled. If there is a reason to preserve things that have intrinsic value, and to regret destroying such things, then the sculptor has a reason, if she is going to produce a painted statue, to paint the blocks in advance of assembling them. But it seems odd to think that she has such a reason, that is, to paint the blocks first for the sake of avoiding the (production and) destruction of something of intrinsic value in the course of producing the painted statue. Thus it seems that there is not a general reason to preserve things that have value. Perhaps the right reply is to identify a concept of privileged value creation, in which valuable things that are created for the sake of creating things of greater value are exempt from the normal protection of valuable things. Consider what we should say if the sculptor dies before she has painted the assembled objects. Would it be good, would it be permissible, to paint them fifty years later, when the work in question could not so readily be regarded as merely provisional? A more radical reply to McHose challenges his example. One might conjecture that the equanimity with which we view the “destruction” of the midprocess work of art is that we do not really accept the stipulation that it is a complete and (fine) work of art. Suppose, contrastingly, an elegant miniature version of a projected building for some reason needs to be destroyed in order that we may produce the (much more valuable) building itself. Would that not be a pity, or certainly more of a pity than the “destruction” of the unpainted blocks? And the aptness of the scare-quotes in that last sentence points to a further weakness in the McHose case: Do we really believe that to paint a statue is to destroy it, especially when, as is ex hypothesi true here, its value is thereby enhanced? But many more distinctions have to be made. One might, for example, distinguish different ways of destroying a Tintoretto and, correspondingly, different attitudes to its destruction: by burning it in order to keep Picasso warm enough to do his painting, or by cutting it up and incorporating it into a Picasso, with the Tintoretto pigments still discernible.

(iv) Valuing the Particular and Incommensurability

Notice that the irreplaceability of which I have spoken of some valuable thing A by some more valuable thing B does not reflect incommensurability between the value of the A thing and the value of the B thing. Irreplaceability and incommensurability are manifestly different things, but in some of the literature on incommensurability the two are routinely conflated. So, for example, it is said to show the incommensurability of friendship and money that one would not give up a friendship for money. But although it may be true that one should not give up a particular existing friendship for money, it does not follow, and it is not in all respects and in all contexts true, that the values of money and


Political Themes: Conservatism, Justice, and Public Reason

friendship are incommensurable. I might have to choose between a job in Tulsa and a job in Milwaukee, and I might reason that in Tulsa I will make more money but in Milwaukee I will make more friends, and I might, noncorruptly, with my soul intact, consider those desiderata on some sort of common scale and decide which is more important to me. But no one without moral defect gives up an existing friendship for money, or, indeed, and with no hesitation, for another friendship that promises to be more valuable even from the point of view of what is deepest in the value of friendship. To be sure, I may have good reason to move to another city, and I may anticipate that it will then be difficult to continue to interact with a certain current friend, but that is not (quite) giving up a friendship for money: I would not say to that friend, “I am sorry, but we cannot (in any sense) be friends any more, since I am moving to Tulsa.” What I can say is: “It is unfortunately likely that our friendship will come to an end,” and that is different. Predicting the demise of a friendship is not choosing to conclude it (e.g., for a gain in money).

(v) Valuing the Particular and Deontology

Is my thesis “just” a generalization of the familiar deontological truth, which is, pressed by critics of utilitarianism, that it is wrong to kill a sad person if a happy person can thereby be produced? There is, of course, a structural affinity here. But the stated familiar point is normatively overdetermined. For there is, first, the point that human beings have a right not to be destroyed, by contrast, perhaps, with cathedrals, sunsets (and colleges?). So binding is that right that there is, as it were, little room left on the normative billboard for the undeniable truth, which is indeed a special case of my claim, that the value embodied in a human being warrants protection in that embodiment. Nor would one happily say, in respect of the more familiar point, that it justifies a mere bias in favor of existing human value. Now, restrictive deontological obligations might not reflect rights on the part of the items whose treatment they govern. A deontological bar on destroying something of value need not acknowledge a right that the thing in question possesses. Tom Hurka has helped me to see that the bias that I defend against destroying what is of value must nevertheless be distinguished from deontological constraints, in the contemplated wide sense of “deontological.” The paragraphs that close this subsection are an edited version of his remarks. The view under defense here gives priority to the preservation of existing goods over the creation of new ones. A familiar deontological view gives priority to not oneself destroying goods, or more generally, to not oneself choosing against them, over promoting goods. (Causing evils gets a treatment analogous to that of destroying goods.) These views yield the same verdict in the principal case that I have discussed, where one has the option of destroying one good in order to replace it with another slightly better good. But the views come apart in more complex cases. Imagine that you can either prevent an existing good from being destroyed by natural forces or create another slightly better good. My view says that you should preserve the existing good, but the deontological view does not: it treats preserving and

Rescuing Conservatism


creating as two forms of promoting and allows you to create the new good. It may do this even when the first good is threatened with destruction by some other person’s wrongful act, because what it prioritizes is your not destroying goods, not not destroying in general. For a possible case on the other side, imagine that someone else is in the process of creating a new good and that I could intervene to sabotage his project, preventing his good from being created and thereafter creating another and greater good myself. My view does not forbid this, because I am not thereby destroying an existing good that has priority. But the deontological view may forbid it, because the sabotage is aimed directly at preventing a good from being realized. (vi) Unanswered Questions

Here are some good questions that were raised by Kimberley Brownlee and that I do not address: Might conserving something require repairing it, or even modifying it? What is more conservative, to repair the old pot with new clay, or to let it crumble? Does Cohen-conservatism permit cultivation and development of that which has value? Also neglected here is the truth that (sometimes?) cultural values can persist only if they are allowed to change: the matter is discussed with great subtlety by Samuel Scheffler in his “Immigration and the Significance of Culture” (106 et seq.). The last two paragraphs raise issues that require more investigation than they get in this essay, regarding the question of what categories of things can qualify as vessels of value. Two examples of such different categories are, I think, material objects, such as trees, and processes, such as a tree’s growth, decay and death. And note that there can be conflicts not only, as in Otsuka’s Slaves case (see subsection (i) of section 3 above), between saving some particular material objects and saving other ones, but also between protecting a process, like the self-renewing activity of a forest, and protecting objects, like trees, that the process operates on, and destroys. At the moment, I see no problems specific to such conflicts, problems, that is, that make the conflicts relevantly different from conflicts between preserving different valuable things within a single category of things. Consider, in this context, deliberately ephemeral works of art. Consider, further, that it is in the nature of some valuable things to disappear, and it is therefore impossible to say that they should be preserved because of their intrinsic value: examples would be bonfires, and maybe shooting stars. But there is an application of my conservatism even to such things: for one would not destroy a bonfire before its time (to fizzle out) has come, in favor of what one knew would be a better bonfire. Tim Scanlon has observed . . . that the degree of bias toward the existing, and regret that is not assuaged by replacement by another thing of the same kind (another friendship or another [college] or whatever) depends on and derives from considerations internal to the value in question. That is, it depends on the particular attitudes and intentions that valuing such a thing requires us to have and respect (=take seriously).


Political Themes: Conservatism, Justice, and Public Reason

And, at the limit, the value of something might be of such a sort that the bias, although more or less strong in other cases, quite disappears. If that is right, then perhaps the most that one can say is that some things warrant the bias, and that the bias means rejecting maximization and its cognates. The reduced claim is: There exist valuable things in respect of which a conservative bias is apt. This existentially quantified claim is, of course, less interesting than the universally quantified one that I have tried to defend, albeit with an openness towards qualification and modification. But one cannot say that the residual point is of no, or of little interest. For it is equivalent to a denial of the inference that goes from: A would be better than B, therefore we should replace A by B. And many take that inference for granted, and apply it in practice, and, so I have argued, it is an importantly mistaken inference. Also, the reduced claim invites what might be an interesting research program, into what forms of value demand preservation and what forms do not. (vii) Beyond Existing Particulars

I wish to notice a pregnant observation on a draft of the present essay that is due to Anca Gheaus. Her observation promises and threatens (if I may lapse, temporarily— I assure you—into Hegelian) to heb auf the main theme of the present section. Gheaus suggests that, just as one may “recognize the special importance of particular embodiments of value and thus resist change in order to preserve these embodiments,” so one may sometimes “welcome change for the sake of change” and thus “accept the loss of some value in order to create something new. Like the conservative attitude, this one is at odds with” value maximization. Fashion (things being in and out of it, that is) seems to illustrate the point. As Gheaus observes, “the two attitudes can co-exist in the same person, at the same time, with respect to different things.” One can want (to keep) something because it exists, but one can also want (to get) something because it does not exist, and in both cases the focus is on the particular, whether it be actual or prospective, and against a merely abstract accounting of value. If Gheaus is right, my defense of what I call conservatism is not impeached, but that which I defend is now seen to be an instance of something more general, namely of the claims of particular valuable things over value in general. So, for example, we see the bias in valuing particular things over abstract value in the typical (not universal, to be sure) aspiration of the artist, who is aiming not to produce a certain mass or type of value but a particular valuable thing. In a post-Gheaus reworking of the theme of this essay, the emphasis on particular valuable things that now exist would be relaxed. Needless to say, the whole matter requires further study. And that further study must address a case put by Michael Otsuka. Otsuka first argues that I am committed to a bias in favor of things that will exist over things that could be brought into existence. Suppose, he says, that you know that some wonderful thing will come into existence in the future, and you also know that you could set in train a process that will lead to its destruction and replacement by something superior the like of which could not be created without the said destruction.

Rescuing Conservatism


If you indeed set the “improving” process in train, then you would be “mistreat[ing] the [to-be] destroyed [thing] as having the merely instrumental value of being a vessel of aesthetic value” (see the third paragraph of subsection (i) of section 3 above). So future existents, and not just present ones, can, on my view, be subject to the same abuse, and the same respect, now, as present ones. The distinction between the mass and type of value and particular value still obtains, but the present/future distinction disaligns with the distinction between particular and abstract value. But this unavoidable (if Otsuka is right) application of my view to future particulars is strongly challenged by Otsuka’s asteroids example: Note that the bias in favor of actually existing things that you defend has the following implication regarding future things that are not yet actual: Suppose that you come to learn that, after the extinction of human beings (and any other creature capable of enjoying beauty), a most beautiful patch of wilderness (far more beautiful than Yosemite or anywhere else in the world) will come into existence. It will, however, not last long, as a small asteroid will land on it, thereby transforming it into a nondescript, barren crater. I agree with you that you would have reason to ensure that this beautiful patch is preserved by now destroying this asteroid. Now suppose that you come to learn, in a variant of the above scenario, that the asteroid will hit this spot on earth before the beautiful patch would have begun to emerge, thereby rendering this spot inhospitable to the emergence of anything beautiful. If, however, you destroy this asteroid, then the beautiful patch will come into existence. It is an implication of your view that you would have less reason to destroy this asteroid in this scenario, as such destruction would not preserve an actually existing thing of value. Rather, it would instead allow for the coming into existence of a thing of such value. I do not find the thought that you would have less reason a compelling one. This brilliant example must serve, for the time being, as a springboard for further reflection. 4. personal value

I have a pencil eraser (or, in British English, a “rubber”) that I have used ever since I became a lecturer forty-six years ago. When I got it, it was a cube, but now it is a sort of sphere, and although it is small, most of it is still there. It is not because I make very few mistakes that most of my eraser is still there, but because (a) I do not use pencils very much; (b) it takes only a little bit of rubbing to eliminate a mistake; and (c) I do not notice all of my mistakes. I would hate to lose this eraser. I would hate that even if I knew it could be readily replaced, not only, if I so wished, by a pristine cubical one, but even by one of precisely the same off-round shape and the same dingy color that my eraser has now acquired. There is no feature that stands apart from its history that makes me want to keep this eraser. I want my eraser, with its history. What could be more human than that? If I had a paint brush that once belonged to Monet, I would value it as such, that is, because it once belonged to Monet, and if it were removed and replaced by a more


Political Themes: Conservatism, Justice, and Public Reason

elegant brush, one that belonged to Manet, by a well-wisher who knew that I think Manet a greater painter than Monet, and that I care about elegance in brushes, then I would know that the well-wisher would have failed to understand something, and I do not mean that he would have failed to understand the sanctity of private property. He would have failed to understand that wanting to keep what one (already) has is a special sort of wanting. It is not true that I am attached to the particular thing because I value attachments, and this particular old thing serves well as something to be attached to. Rather, I am attached to the thing itself, as such, and not for any general reason, such as the general reason that it is good to have attachments (which it certainly is). Th us, if my would-be benefactor had said: Oh, but I know that in time you will grow even more attached to the Manet brush, I should remain unmoved. It is the Monet brush that I want, not some quantity of attachment to something or other. Consider the resistance to suburban supermarketization, on behalf of urban neighborhood shops. People who seek to protect neighborhood shops point to their many advantages, to the many purposes that they serve so well: local effects like providing access to provisions for old and infirm people, a meeting place that stimulates community, eccentric product lines, and also wider effects like reducing motor traffic, and opening up opportunities for people from deprived minorities to rise on the social scale, through self-, and even other-, employment. But we deceive ourselves if we think that it is only because they deliver specifiable economic and social benefits that we cherish our local shops. It is not only the purposes that they serve that justify our resistance to their destruction. It is also because in all their vagariously caused uniqueness they are part of a social and cultural landscape to which we belong. I do not want you or, for that matter, me, to knock down part of my house and replace it by something bigger and better, even if it truly is bigger and better. I want my home to remain what it is and I want my neighborhood to remain what it is. (Of course, if either were really lousy, I would welcome, I might crave, something new. But if each is quite good, or good enough, then I may want them substantially as they are, despite their deficiencies on a generalization of the criteria that would make me judge them to be unacceptable if they were a lot less good, by those criteria, than they actually are.) Both the economic market and state planning tend towards the destruction of the particular, and therefore towards the destruction of things that are both impersonally and personally valuable. (By “tend” I mean “tend,” not “always do.”) To be sure, some people are happy to pay over the odds, that is, to pay more than what they buy is worth to them, for stuff from the local shop, in order to help keep it going. But that’s clearly countermarket behavior: game-theoretically speaking, these benign buyers are suckers. And some planners might accept its being there as a reason to keep something there. But usually they will prefer a general consideration, something that the thing does well, or even the general consideration that a majority want it kept there—which is not, of course, the reason for keeping it there that the members of that majority themselves have. The logics of the market and of planning tend against the truth that people want particular valuable things, not just satisfaction of general desiderata. If everything is

Rescuing Conservatism


added to or subtracted from the environment in neglect of the value of particularity and variety as such, then everything will tend to be the same everywhere, because serving the same requirements: every quarter will have its Starbucks. But market choice is relentlessly general in the grounds of its preference, so market mania is deeply anti-conservative and Thatcherist Toryism is a great betrayal of conservatism (about which I shall say more in section 6). If you want everything to be optimal, nothing will be good. And some things have to just be, they have to just be there, if anything is to be good. We are attached to particular things because we need to belong to something, and we therefore need some things to belong to us. We cannot belong to something abstract. We do not keep the cathedrals just because they are beautiful, but also because they are part of our past. We want the past to be present among us. We do not want to be cut off from it. We rejoice in our contact with the culture of our past. We of course value our particular past in the respectful way that we value any past culture, but we also value it in a more personal way. We want to be part of what Edmund Burke (famously) called “the partnership not only between those who are living, but between those who are living, those who are dead, and those who are to be born.” When people ask me how I can bear to think of my destiny, which is death, I tell them that I find great consolation in the American poet Carl Sandburg’s line: “Something began me and it had no beginning; something will end me and it has no end.” Does its possession of personal value give us a reason not to destroy some thing, other than the reason that doing so would distress the person or people for whom it has that value? I am inclined to say “yes,” but that is not so evident as it is where the value is not (merely) personal but intrinsic. 5. paradoxes of change

If we are conservatives in my sense, then we have reason to change slowly because we have reason to be what we are and to carry on with what we have: we cannot simply erase our background and replace it by something better. A conservative regulation gives life continuity. We cannot reinvent ourselves, or our language, or anything that really matters, every day according to what our resources now are and what our opportunities now are. We cannot keep everything “under review.” Conservatives often make a point about the unforeseeable consequences of large and rapid change, a point about looking before leaping, and about the difficulty of seeing what’s coming, and so forth. But that l’imprévu point is not my point: I recommend changing slowly for no such strategic reason. We have reason to change slowly even when we know very well what is coming next (and after that). In the brilliant climax of his 1980 critical notice of two books by Anthony Kenny (Will, Freedom and Power and Freewill and Responsibility), Ted Honderich wrote: There is also the sad fact that our author, who might be thought to have an obligation in these matters, has joined those users of English who take concepts to include some that are voted for, driven, lived in, and installed. You can’t do any of


Political Themes: Conservatism, Justice, and Public Reason

that to a concept, and you can’t morally justify one either. . . . No doubt the time will come when you can, but we needn’t rush toward it. I have ruminated on the implications of that wise remark many times in the nearly thirty years that have passed since I read it. They include these: that (1) the first stage of linguistic evolution, or, rather, of one pervasive kind of linguistic evolution, is linguistic error; but that (2) once the error has become common enough, it ceases to be an error, and it may become even required usage, so that an older usage now qualifies as merely fussy, or at any rate quaint; but that (3) we who are attached to what is now correct usage may reasonably seek to decelerate the process of change, by avoiding deviation ourselves, which we ought to do anyway, quite apart from our particular tastes, because deviations are errors. (Maybe, because deviations are errors, there should not be any linguistic change of this type, even when it is beneficial in the end. Maybe, if it is wrong for it to start happening, it is wrong for it to happen.) There are, of course, parallels in nonlinguistic behavior and style, in what is “the done thing.” Do the thing that is not done often enough, and it becomes the done thing. But, returning to language, we must here remember the point made in section 3 above, which is that the disposition to protect the old can coexist, as it does in my own case, with a welcoming attitude to those new things that do not displace the old things. There is nothing wrong with expansion of the vocabulary stock, or with neologisms, like “like,” as such. (Like, I really, like, like, “like.”) The Honderich resistance that I endorse is against making errors that are in time self-legitimating and that thereby destroy old meanings. 6. conservatism and the political right

Wanting to conserve what has value does not imply wanting to conserve exploitation and injustice, since they lack value. Wanting to conserve what has value is consistent with wanting to destroy disvalue. Of course, something that is unjust can have value, and even in a fashion that is linked to the very thing that makes it unjust. Certain practices of deference, for example, which embody injustice, also facilitate forms of exchange that are deeply mutual in a distinctive way. But you can be both egalitarian and conservative by putting justice lexically prior to (other) value (whether or not—a further question—one could say that something that merits destruction because it is unjust might nevertheless be of great value, all things considered). (Lack of in-) justice, such egalitarians might say, is the first virtue of things. This deference to justice does not trivialize the “we want to conserve what is valuable” thought. For conservation of the valuable still sometimes trumps value maximization (and so forth) and it needs something as momentous as justice to trump the (nonjustice) value that we seek to conserve. I do not say that I am myself so uncompromising an egalitarian, so lexically projustice. I am not sure that we should regret the production of all the wonderful material culture that we have inherited and that was produced at the expense of gross injustice. Of course we

Rescuing Conservatism


regret that it was produced in that way, but would we be happier if it had not been produced, on the assumption that there would have been no non-unjust way of producing it? To be sure, if one is prepared to compromise justice, then one will be more prepared to do so, ceteris paribus, if one has the conservative attitude that I recommend. For the bias towards existing value will render an injustice more tolerable (than it is for a nonconservative compromiser) when this injustice cannot be eliminated without eliminating intrinsically good things. Note, further, that wanting to conserve the valuable, even wanting to keep things that have increased in value as a result of tradition, is not “wanting to keep things as they are,” as such, where that means “wanting to preserve the status quo, the existing social and political situation.” That is too undiscriminating a characterization to be the name of a proper object of conservative concern. And, in order to preserve valuable things in the status quo, we might have to revolutionize our situation. To keep the pictures safe, we might have to tear down the charming but damp and draughty old museum and build a new one. One thing Karl Marx said about the socialist revolution was that that revolution was necessary to preserve the fruits of civilization against the ravages of capitalism. Classical British Conservatism combined conservatism with a small “c” with preservation of wealth and inequality: The rich man in his castle The poor man at his gate God made them, high or lowly, And ordered their estate. (“All Things Bright and Beautiful,” 1848, by Cecil Alexander.) The combination of conservatism with wealth and inequality was relatively easy to sustain in a precapitalist society, but, when inequality became capitalist inequality, the combination of conservatism with wealth and inequality became untenable, among other reasons because capitalism so comprehensively transforms everything, including itself: in the phrase of the Communist Manifesto, under capitalism, “all that is solid melts into air.” As Marx also said, “all earlier modes of production were essentially conservative,” by contrast with capitalism. When the rich morphed, fully, into capitalists, the British Conservative Party became the anti-conservative market party. As a matter of history, the bottom line of Conservatism with a capital “C” turned out to be not conservatism with a small “c” but preservation of wealth and inequality. With fierce international competition, conserving old ways is too costly to the maintenance of wealth. And with historical working class gains in place, small-c conservatism becomes a buffer against inequality. For the sake of protecting and extending the powers of wealth, big-C Conservatives regularly sacrifice the small c-conservatism that many of them genuinely cherish. They blather on (as Prime Minister John Major did) about warm beer and sturdy spinsters cycling to church and then they hand Wal-Mart the keys to the kingdom. They are thereby in tune with the propensity of capitalism, which is to maximize a certain kind of value, in sovereign disregard of the value of any things.


Political Themes: Conservatism, Justice, and Public Reason 7. apology

I have been working on this essay (intermittently) for a few years, and I have received an abundance of excellent criticism of it, some of which has been reported above. The essay was initially bold and simple, but it soon lost its pristine character. Incorporating criticisms, and my responses to them, turned it into a patchwork product. Forty-five years ago I heard Bernard Williams praise a philosopher’s book because its text revealed that the author “did his best thinking off the page.” Well, much of my own thinking on the present topic, be it best (the thinking, that is) or otherwise, has been set here before you, on the page. In the Popperian terms that Bernard Williams might have accepted to express the point, there has here been too much context of (would-be)-discovery and not enough context of justification. But that is because festschrifts have deadlines, and this one is upon me, and what you have before you is simply what I have been able to do so far. I would have liked to give Tim something more finished, but I know that he will, with his usual ample generosity, understand. notes 1. I have a large debt to David Lloyd-Thomas, whose fascinating unpublished writing on particular value nudged me towards the present topic, and who has commented helpfully on drafts of this essay. I also thank John Baker, Annette Barnes, Marshall Berman, Chris Bertram, Akeel Bilgrami, Michael Bratman, Geoff Brennan, Harry Brighouse, Kimberley Brownlee, Richard Christian, Josh Cohen, Michael Cohen, Michèle Cohen, Angela Cummine, Paul David, Jon Elster, Cécile Fabre, Jeremy Farris, Dov Fox, Michael Freeden, Michael L. Friedman, Anca Gheaus, Kerah Gordon-Solmon, Kent Greenawalt, Alan Hamlin, Barbara Herman, Tom Hurka, David Kaplan, Tsehai Khemet, Rahul Kumar, Brad McHose, Alistair Macleod, John McMurtry, José Marti Marmol, Mike Martin, Calvin Normore, Kieran Oberman, Michael Otsuka, Avia Pasternak, Piero Pinzauti, Amélie Rorty, Michael Rosen, Tim Scanlon, Sam Scheffler, Shlomi Segall, Seana Shiffrin, Lucas Stanczyk, Hillel Steiner, Jenny Szende, Patrick Tomlin, Jeremy Waldron, David Wiggins, Andrew Williams, Susan Wolf, Chris Woodard, and Arnold Zuboff. 2. Which is the value that something has in itself, independently of its consequences, including its consequences for human utility and edification. Not everybody believes that there exists intrinsic value, within the meaning of the foregoing sentence, but my principal opponents in this essay, whom I confront in section 3 below, share my belief in its existence. 3. It certainly sometimes has extrinsic value: slavery, which is an injustice, helped to build the (intrinsically valuable) pyramids. 4. For a particularly (albeit predictably) intelligent elaboration of such skepticism about conservative factual assessments, see Bernard Williams, “Modernity and the Substance of Ethical Life,” in his In the Beginning Was the Deed (Princeton: Princeton University Press, 2005). 5. Except insofar as the new things might crowd out the old: it is a shame when Wren’s City churches are surrounded by ugly skyscrapers. But note too that new things sometimes (albeit more rarely) enhance the presentation of old things. 6. That is, not who we are, which is our identity in the austerely metaphysical sense, but what we are, which is our identity in the vaguer but very important sociocultural sense that I try to formulate in the italicized words in the next sentence of the text. 7. The importance of personal valuing, in the stated sense, was brought to my attention by David Lloyd-Thomas. In a number of unpublished writings that were the original stimulus to

Rescuing Conservatism


the present reflections, he called what I am calling “personal value” particular value: I think “personal value” is a more appropriate name for it. 8. See the eraser example at the beginning of section 4 below. 9. The second disjunct in that alternation is added because I mean to reject not only the treatment of the College as merely an instrument for certain purposes but also the treatment of it as merely a vehicle of (even noninstrumental) value, as opposed to something that is also itself valuable. See, further, subsection (i) of section 3, below. 10. The constraint is normative, not causal: I do not mean that the given is immovable, that we must accept it in the way that we must accept the law of gravity, or (as some think) the Laffer curve. For another case in which a conservative normative truth might tend not to be discerned because it gets conflated with a structurally similar causal truth, see the second paragraph of section 5 below. 11. For a sensitive treatment of how unchoosing acceptance of the specifically cultural given may be reconciled with the liberal valuation of autonomy and choice, see Samuel Scheffler, “Conceptions of Cosmopolitanism,” in Boundaries and Allegiances (Oxford: Oxford University Press, 2001), esp. 126ff. As Scheffler says, the liberal “state merely disallows the coercive enforcement of cultural affiliations; it does not offer a voluntaristic theory of their ultimate moral import” (128). 12. London: Penguin Books, 1984. 13. I think this word is obsolete. So far as I was able to tell, it meant “doltish.” 14. As the point is sometimes expressed: we should not “play God.” Jonathan Glover is mistaken when he opines that “When the objection to playing God is separated from the idea that intervening in this aspect of the natural world is a kind of blasphemy [against God Himself— GAC], it is a protest against a particular group of people, necessarily fallible and limited, taking decisions so important to our future” (ibid., 47). But one may also object on a ground more fundamental than the latter and at which I gesture in the sentence to which the present footnote is attached. To be sure, many, and almost certainly Glover, will regard that “ground” as obscure and confused, but let us acknowledge that impasse, and not obliterate the distinctive conservative view of the given, which is expressed eloquently enough by David Wiggins: If we cannot recognize our own given natures and the natural world as setting any limit at all upon the desires that we contemplate taking seriously; if we will not listen to the anticipations and suspicions of the artefactual conception of human beings that sound in half-forgotten moral denunciations of the impulse to see people or human beings as things, as tools, as bearers of military numerals, as cannon-fodder, or as fungibles; if we are not ready to scrutinize with any hesitation or perplexity at all the conviction (as passionate as it is groundless, surely, for no larger conception is available that could validate it) that everything in the world is in principle ours or there for the taking; then what will befall us? Will a new disquiet assail our desires themselves, in a world no less denuded of meaning by our sense of our own omnipotence than ravaged by our self-righteous insatiability? (Wiggins, Sameness and Substance Renewed [Cambridge: Cambridge University Press, 2001], 242). 15. I allow here, what might be challenged, that the manufactured flesh is flesh. Nicer minds than mine might protest that it ain’t flesh if it didn’t come to be the way (natural) flesh does. 16. A slippery slope of this structure can also give nonconservatives pause. If, for example, such a slope in respect of judgments about the time after which abortion should be permitted leads to tolerance of infanticide, then it leads to an evil on which conservatives have no monopoly of aversion.


Political Themes: Conservatism, Justice, and Public Reason

17. See, further, section 5, below. 18. Franklin Roosevelt once said about right-wing Latin American dictators: “Well, yes, they’re bastards, but they’re our bastards.” And I say, ad (some of) our negative characteristics: “Well, yes, we’re bastards, but we’re our bastards.” 19. For highly suggestive remarks about the given in general, see Michael J. Sandel, The Case against Perfection (Cambridge, Mass.: Harvard University Press, 2007), 26ff. et passim. 20. That is, anything at all, whether or not it counts as a thing in the narrow sense in which, e.g., a process, or a feature, does not. See subsection (vi) of section 3, below, for some remarks on valuable processes. 21. The value/what has value distinction is a generalization of the distinction between exchange-value and what has it that I use in the course of my (so I believe it to be) demolition both of the labor theory of value and of the claim that the said theory supports a charge of exploitation. In some contexts, “value” and “what has value” are just different stylistic variants. In the present context the difference between them makes all the difference. (See G. A. Cohen, “The Labour Theory of Value and the Concept of Exploitation,” Philosophy & Public Affairs 8 [1979]: 338–60, reprinted with revisions, in History, Labour and Freedom [Oxford: Oxford University Press, 1988].) 22. It follows that, unless we are to say that utilitarians have no values, we must interpret what Samuel Scheffler says here creatively, in order not to judge it to be false: “ . . . it is difficult to understand how human beings could have values at all if they did not have conservative impulses. What would it mean to value things but, in general, to see no reason of any kind to sustain them or retain them or preserve them or extend them into the future?” If we take Scheffler literally, the answer to his question is easy: it would mean, for example, to be a utilitarian. (The passage appears in Scheffler’s magnificent essay, “Immigration and the Claims of Culture,” Philosophy & Public Affairs 35 [2007]: 105.) 23. Michael Otsuka, private communication, August 2005. 24. To be sure, a thing’s being especially valuable because it is old is not the same as its being especially valuable because it exists. I acknowledge the important distinction between oldness and existence three paragraphs hence. 25. Private communication. 26. Something might first be valuable because it is new, gradually lose that value, and pari passu gain the value of longevity. Its peak value would depend on the sort of thing that it is. (Maybe it is possible for me to be delighted both that the Seagram’s Building is only forty years old and that it is as much as forty years old.) 27. This paragraph was inspired by astute comments from Patrick Tomlin. 28. I am, indeed, inclined to doubt, in all prephilosophical innocence, that this describes a metaphysical possibility, but Calvin Normore advises me that there is a controversy in the philosophical literature as to whether a particular can go out of existence and then come back into it. 29. I adverted earlier (see note 5) to the negative effect of surrounding Wren’s churches by ugly skyscrapers. Michael Otsuka, whose example that is, suggests that “surrounding a Wren church with skyscrapers does seem to devalue the value inherent in the Wren church and not merely our experience of that value.” But I do not agree that it does that, unless we conceive Wren’s creation not as the church itself but as the church-within-a-particularcontext, with some motifs in it that are intended to resonate well with its surroundings. Nothing that I want to claim would be affected by that expansive reading of what the church is.

Rescuing Conservatism


Although—so I insist—the Hockneys circumscribing the Lippi do not devalue it, I need not deny that new things can devalue existing valuable things. As Marx asked, what is “Jupiter compared with the lightning conductor, and Hermes compared with the Credit Mobilier?” And new things can also enhance the value of, or even confer value on, existing things. 30. The position defended above, that aesthetic value is inherent in aesthetically valuable things, is consistent with an account of aesthetic value as a secondary quality, one, therefore, which is constitutively apt to stimulate or enable experiences of certain kinds. I here neither endorse nor reject that account. 31. In their interesting article on “The Reversal Test,” in Ethics 116 (2006): 656–79, Nick Bostrom and Toby Ord scout what they call the “status quo bias.” Caveats in the second paragraph of their article exclude its application against this one. 32. It was presented at Stanford University on November 12, 2007. 33. Joseph Raz, Morality of Freedom (Oxford: Oxford University Press, 1986), 346–53. 34. In a letter to me dated May 23, 2007. 35. Cf. the discussion of the McHose example, at subsection (iii)(e) of section 3, above. 36. Private communication, March 2008. 37. Private communication, July 11, 2008. 38. Private communication, July 18, 2008. Otsuka adds: “Since I don’t find it a compelling thought, I’m led to wonder whether there is any justified bias in favour of the preservation of existing value as such. Perhaps the lesson we should take away from reflection on this case is that any such bias must be explained in terms of a less interesting claim regarding the personal value that existing things tend to have.” 39. People sometimes ask whether the present essay has any practical relevance. Well, it has at least this much: politicians and planners who are persuaded by it would realize that it is not enough to justify a replacement for something that it would be better than what it replaces. 40. And every conservative diatribe like this one will badmouth Starbucks. In fact, Starbucks has things that recommend it, but I am regretting sameness, which is to say, lack of variety, not how bad each of the members of the repetitive set is. To regret that there’s only one kind of flower in the garden is to say nothing about how beautiful that flower is. 41. Reflections on the Revolution in France, 1790. I know that everybody quotes that: I’m not trying to be original here, just trying to hit the nail on the head, which Burke did. 42. This is not the Neurath-like idea that, at any given time, some things must be taken for granted whereas others are evaluated against the stable background that the accepted things provide. Within that idea, one chooses what the stable background will be, just as, in Neurath’s analogy, one chooses which parts of the boat to replace and/or repair and which to keep as they are for the time being. I am saying the different thing that, at any given time, some matters are beyond our power to reconsider, salva stabilitate. 43. For a fine treatment of it, see Geoffrey Brennan and Alan Hamlin, “Analytic Conservatism,” British Journal of Political Science 34 (2004): 675–91. 44. Mind 89 (1980): 133. I make no comment on whether Honderich is right, or was right in 1980, about the word “concept.” I praise not the substance but the form of his admonition of Kenny. It might seem ironical that an intransigent opponent of conservatism (see his Conservatism [London: Penguin Books, 1991]) should have given so powerful a formulation to an essential conservative thought. But maybe it is not really ironical, because, so Honderich says on pages 1–2 of the book, his topic is not small-c conservatism. On the other hand, Honderich might indeed be thought to be taking a swipe against small-c conservatism when he writes that “if Conservatism were at bottom a defence of the familiar, in


Political Themes: Conservatism, Justice, and Public Reason

the plain meaning of the term, we should have a mystery on our hands, the mystery of how an egregious idiocy could have become a large political tradition” (ibid., 2). To be sure, it would be an egregious idiocy to defend everything that is familiar, but it would also be an egregious idiocy to reject, for that reason, a characterization of conservatism according to which a conservative is more disposed to defend the familiar than others are. 45. I propose a resolution of the tension between the Marxist commitment to advancement of productive power and the Marxist commitment to those at whose expense that advancement occurs in Section V of “Peter Mew on Justice and Capitalism,” which uses materials assembled in my History, Labour, and Freedom, 303–4. 46. I owe that point to Michael Otsuka. 47. Compare the questions raised by Brownlee questions that are reported in the first paragraph of subsection (vi) of section 3, above. 48. As John McMurtry points out to me, my comment here may represent a misappropriation of Marx, because Marx might have meant that what was to be preserved was, the level of development of the productive forces, and that is a nonconservative desideratum, because of its indifference to the value of the particular. 49. Which Marshall Berman has made more famous by using it as the title of his magnificent book. 50. Capital, vol. I, ch. 15, section 9. 51. Aristotle distinguished between the usefulness of things and their abstract economic value and he disparaged the desire to accumulate the latter, a desire that has no shape and therefore no limit: see his remarks in Politics, bk. I, chs. 8–10. If I were more continental, and therefore less continent, than I am, I would speculate that the maximizing stance in axiology is a projection onto the realm of value in general of the capitalist attitude to value, which is “Accumulate! Accumulate! That is Moses and the prophets” (Karl Marx, Capital, vol. I, ch. 24, section 3). 52. Or rejected: I think Bernard Williams did not admire Popper.

10 Global Political Justice and the “Democratic Deficit”* Charles R. Beitz

a familiar argument for the reform of global institutions goes something like this. Once upon a time, the boundaries of states were more-or-less coextensive with the reach of their governments’ decisions. A state’s citizens and those affected by its government’s policies were mostly the same. The formal structure of international order consisted primarily of agreements among states for fairly narrow and specific purposes. There were not many transnational actors of economic or social significance and those that existed were effectively regulated by the states in which they operated. Then came globalization. For various reasons and to varying extents, governments in a “globalizing” world have the capacity through various kinds of policies to affect the circumstances of people outside their borders. These external effects are frequently inescapable for those on whom they fall and can be highly adverse. International organizations have grown in number and influence but are controlled for the most part by elites and lack mechanisms that render their operations accountable to the individuals who are affected by their policies. Transgovernmental networks of officials are increasingly able to exert independent influence in domestic level policy making. Transnational actors have proliferated—particularly multinational firms and nongovernmental organizations (NGOs)—and have developed operating strategies that enable them to escape effective regulation by individual governments. As a result of all this, the citizens of states, including democratic states, have less control than in an earlier age over the conditions of their lives. They suffer from the external effects of decisions taken by other governments as well as from the effects of decisions of international and nongovernmental transnational actors not subject to effective public regulation. The familiar argument I have in mind contends that this is unjust. People whose interests are put at risk by the decisions of others should have a share of control over 231


Political Themes: Conservatism, Justice, and Public Reason

these decisions. The global political order conspicuously fails to conform to this precept. Institutions should be reformed or restructured so that individuals whose interests are substantially affected by political decisions are provided with procedural means of exerting leverage over the mechanisms by which these decisions are made. The reforms usually called for consist of some mixture of changes at the level of international organizations (e.g., reform of voting procedures that would afford equal representation in international decision-making bodies for individuals affected by the decisions made) and at that of individual states (e.g., establishment of extraterritorial voting rights for outsiders potentially affected by a government’s decisions). We might call this combination of diagnosis and prescription the “democratic-deficit view.” I do not suggest that anybody has actually held the view in precisely this form, but some writers have been in the neighborhood. There are at least three common responses to the democratic-deficit view. Some say that the magnitude of the institutional changes that would be necessary to give all individuals a share of control over the decisions that affect them are so great as to be practically unattainable. It would be unrealistic—or utopian, in the disparaging sense—to seek global democracy in any form worthy of the name. Another response is that democracy does not scale to the global level. There are too many obstacles—the issues requiring resolution are too remote from people’s lives to motivate informed participation, there are too many differences in the political cultures of the world’s societies to afford common ground for deliberation, we cannot imagine any basis of representation that would be fair to individuals while recognizing the importance of states. This kind of response is consistent with the first, but it makes a deeper point. The first response holds that global democracy, whatever its institutional details, is too ambitious a goal for the present, whereas the second holds that global democracy would not function well even if its institutional forms could somehow be brought to exist. The third conventional response—and perhaps an inference from the second—is that any institutions of global scale with sufficient authority to govern effectively, including any formally democratic ones, would in practice operate in ways that would frustrate their purposes. Kant’s rejection of world government has this character. As Rawls puts it, any such government “would either be a global despotism or else would rule over a fragile empire torn by frequent civil strife as various regions and peoples tried to gain their political freedom and autonomy.” These three responses converge in holding that it would be a mistake to make global democracy a proximate political goal. However, all are compatible with the thought that the absence of mechanisms of effective popular control over decisions that have transnational consequences is a form of political injustice. (To say the responses are compatible with this thought is not, of course, to ascribe the thought to the writers mentioned.) They take the central question to be what is realistically or practically achievable as a remedy, given that a sufficient democratization of global institutions is either beyond reach or likely to be self-defeating. I do not want to suggest that this question is unimportant. I believe, however, that the temptation to accept the democratic-deficit view’s diagnostic premise while disputing its prescriptions for political reform should be resisted. In its exclusive focus on

Global Political Justice and the “Democratic Deficit”


the formal relationship between those affected by political decisions and the mechanisms through which these decisions are reached, the view embodies a mischaracterization of the conditions under which we regard regimes as democratic and a truncated understanding of the reasons why we should care about them. I believe that much of what seems unrealistic or impractical in some proposals to eliminate the global “democratic deficit” is traceable to deficiencies of the critique. It is not that the idea of a global democratic deficit is empty. I take it as given that we are living through a momentous shift in the world political economy whose various components are usually grouped under the heading “globalization,” and that the existing world political structure is inadequate to manage the consequences of this shift fairly and efficiently. However, the idea of a global “democratic deficit” can convey an excessively simple view of the reasons for concern about the existing structure and focus attention on strategies for reform that do not realistically address the deficiencies. Once we grasp the complexity of these reasons we can see that the range of institutional remedies is broader and perhaps more achievable than it may have appeared.


The form of the familiar view I want to consider might be understood as applying the medieval maxim, quod omnes tangit ab omnibus approbetur (“what touches all should be approved by all”) to the contemporary global political system. This maxim originated as a principle of Roman private law requiring that no change in rights be effected without the consent of each party in interest. This is a strong requirement and it was often interpreted less restrictively, as the rule that “all must be given a hearing and defense of their rights” before a judicial determination of a suit could be made. The private-law maxim was brought into the political realm in Edward I’s writ summoning the clergy to the Parliament of 1295. In that context its significance was that delegates sent to Parliament should come with the authority to take decisions binding the communities they represented; if, for example, they were to consent to a tax, their consent should obligate the members of their communities. The maxim was eventually glossed as the principle that the imposition of a law should have the consent of those whose interests are “touched” by it. Like the private-law maxim, this is a strong requirement, and its most common interpretation was also less restrictive: Decisions did not necessarily require the consent of all but rather the participation of all or of their representatives in a proceeding in which everyone’s interests could be articulated and negotiation could take place about the best resolution of the matter at hand. Within modern democratic thought “what touches all” is typically invoked in either of two contexts. One is the context of institutional reform within a society. The existing political institutions exclude some members of the society from participation in decision making (perhaps they are denied the franchise) and a demand for inclusion in the decision-making group (the “demos”) is grounded on the claim that all those affected by the institutions should have a say in their operation. The other context involves fixing the boundaries of decision-making institutions. Here the question is


Political Themes: Conservatism, Justice, and Public Reason

sometimes taken to be “how to constitute [the demos] in the first place.” But it is difficult to make sense of constituting the demos “in the first place.” In practice we do not encounter choices involving the formation of a demos de novo; practical choices usually involve the prospect of change or reform in already-established arrangements— secession, for example, or unification. From a sufficiently abstract perspective the two contexts may therefore seem to converge: In both cases what appears to be at stake is revision of existing arrangements so as to include people in the decision-making group who had previously been excluded from it. But this degree of abstraction can be misleading. In the first case, the boundaries of the community are taken as given and the question is which of those within the community should be included in the demos. In the second case, the boundaries of the community themselves are in question: directly, if the demand is for inclusion of outsiders, or indirectly, if it is for compensatory measures like extraterritorial voting rights. Partisans of the democratic deficit view make both kinds of appeal to “what touches all”: the first kind, when they advocate reform of existing international institutions, and the second kind, when they advocate measures giving outsiders control over domestic-level laws and policies that have external effects. But the difference between the two contexts is important; we should not assume unreflectively that a principle that would be reasonable in the first would also be reasonable in the second. In one strain of democratic thought something like “what touches all” is taken to be foundational. We see this, for example, in an argument given by Jeremy Waldron for majority rule: he thinks its status as a requirement of political justice follows directly from “what touches all” together with an acceptance of the value of political equality. Waldron holds that “what touches all” is “entirely unexceptionable.” He is right to suggest that it corresponds to a widely accepted idea but I do not believe it is unexceptionable. On reflection, it seems to be incompatible with the actual structure and functioning of regimes we ordinarily regard as democratic. Moreover, according to the most plausible view of its moral basis, it is not a foundational principle at all but rather an intermediary one of limited application. It is not even clear how the maxim should be understood. Let me comment first about its interpretation and then turn to the reasons we might hesitate to endorse it as a basic requirement of political justice. The maxim in question holds that all and only those affected (“touched”) by the outcomes of a decision-making process are entitled to a say in the process. There are at least three dimensions of this formulation that need clarification. The first involves the respect in which one must be affected (“touched”) by a decision in order to be entitled to a say in it. The important distinction is between causal and deontic views. On a causal view, one is “touched” when a decision affects one’s interests (or, in the simplest sense of affecting an interest, when the decision makes it more or less likely than it would otherwise be that one gets something one wants). On a deontic view, one is “touched” when a decision changes one’s rights or obligations. Because there are various ways my interests can be affected by a decision without changing my rights and obligations, these are clearly distinct. Most writers adopt the first view (sometimes without noticing the alternative), though it might be argued that the second is more consonant with the maxim’s origins in the law. The second question concerns the scope of “all affected.”

Global Political Justice and the “Democratic Deficit”


Does it comprise those actually affected by a decision or those possibly affected? The question arises because the composition of the affected group may depend on the content of the decision. Although the “actualist” interpretation has a superficial attractiveness, it should be clear that if we take an interest in the maxim because we think it may be helpful for purposes of institutional design or reform, such an interpretation would be incoherent because it would require us to anticipate decisions not yet made. A “possibilist” interpretation would avoid this. Here is the final interpretive issue. Suppose we distinguish between “institutional” and “abstract” variants of the maxim. Waldron suggests a version of the first, which is more literal (and, I believe, more faithful to the traditional idea): “what touches all should be decided by all.” (To remove a remaining ambiguity we should say that “decided by all” means that each member should have a share of control over the outcome or, more precisely, a chance to be decisive over it; a capacity to exercise influence through informal means is not sufficient.) Rawls employs a form of the abstract variant: “What touches all concerns all.” The difference between the two variants is not simply cosmetic. The institutional interpretation bears directly on the structure of decision procedures by identifying those entitled to a share in control over outcomes—namely, all those (who might be) “touched” by them. The abstract interpretation has no immediate bearing on procedures; it pertains to the substantive inputs to the judgment that should take place in reaching a public decision. It says something like this: If your interests might be affected by a decision, then you have a right to expect that they be taken into account by those responsible for reaching the decision. The abstract formulation, but not the institutional one, might be satisfied in a perfectly benevolent despotism or in a “decent consultation hierarchy” or, perhaps more plausibly, in a pluralistic structure in which various types of decisions are vested in a mixture of fora, some representative of those affected and others not. The most rigorous interpretation of the maxim derives from combining the “causal,” “possibilist,” and “institutional” alternatives we have distinguished. This gives us the precept that “all those whose interests may be affected by any of the possible outcomes of a decision-making process are entitled to exert (a share of) control over the process.” It is worth noting how far-reaching this precept might be: Because it applies to anyone whose interests might possibly be affected, it takes in not only those whose interests would be affected by the decisions actually taken but also those whose interests would have been affected by decisions that might have been but, in fact, were not taken. This is a very demanding precept. Notwithstanding, for those inclined to regard the maxim as foundational in democratic theory—“democratic fundamentalists,” we might say polemically—this is likely to be the preferred interpretation (as it is, for example, for Goodin).


The trouble is that in its fundamentalist form the maxim states a requirement that is incompatible with institutional arrangements that most people are likely to regard as unobjectionable from the perspective of democratic commitment.


Political Themes: Conservatism, Justice, and Public Reason

It might be said that small-scale democratic systems—common rooms, for example, or clubs or, arguably, the idealized model of direct democracy sometimes associated with classical Athens—comport or could be made to comport with the fundamentalist interpretation of the maxim. But it requires a good deal of ingenuity to argue that the same can be said of the decision-making processes of modern states. For one thing, the political agenda is seldom limited to issues that affect all of the members (and certainly not to issues that affect all equally). Issues affecting proper subsets of the membership or affecting some members more than others routinely become matters of public decision. In a democracy that respects political equality these are resolved through processes that give as much influence to those not affected at all or affected only in minor ways as to those centrally concerned. In this way “what touches all” might be said to be violated through overinclusion. Moreover, in many large-scale systems, including virtually all modern administrative states, some issues are resolved through processes over which there is either no or severely attenuated popular control. The clearest illustration is administrative law, in which entire categories of issues are delegated to agencies whose decisions are not subject to the retrospective control of any elected body. We might say this is one kind of underinclusion. Judicial review by unelected courts is another. If “what touches all” were a fundamental principle of political justice, we would have to regard both over- and underinclusion as incompatible with democratic commitment. The fact that most people (or, anyway, those not in the grip of a theory) do not do so suggests to me that it is a mistake to take the maxim to be morally basic. Of course, this impression might be wrong. But it is difficult to find reasons why we should regard “what touches all” as having any greater significance than that of a middle-level precept of limited applicability. Recall our earlier distinction between “institutional” and “abstract” interpretations of the maxim. Although plainly distinct, these formulations are not necessarily unrelated. A possible view is that the relationship is instrumental: We improve the chances that everybody’s interests will be taken into account by entitling each person to exercise leverage, directly or through representatives, over the relevant decision. In many contexts this may be highly plausible. What should be observed about such a view is that it is a contingent matter whether the relationship holds. Whether decisions reached through an inclusive participatory process are actually likely to yield outcomes responsive to the interests of those affected depends on how effectively each person uses the opportunities defined by the rules of the process and on the distribution of various categories of interest across the population. These facts were recognized by Mill, for example. His view about the social circumstances in which a participatory regime would function effectively was almost certainly incorrect in its details. But the important point for us is that if we accept “what touches all” because we believe that entitling each person to exercise a share of control is the best way to ensure that each person’s interests will be taken into account, then we do not hold that maxim as a basic principle of political morality. It is intermediary. The fundamental principle, according to the view we are entertaining, would be the substantive requirement that political decisions should take account of all affected interests.

Global Political Justice and the “Democratic Deficit”


Someone who regards the maxim as fundamental needs to tell a different story about its basis. Perhaps the most plausible account arises as a by-product of an argument about procedural fairness. Consider a simple (though by now anachronistic) case once suggested by Brian Barry. Imagine a railroad car that has not been designated either “smoking” or “no smoking.” Each passenger either wants to smoke or objects to others smoking. All the other cars are full. The doors are closed. Everyone wants a decision. Barry argues that when there are “no presumptions as to merits,” the obvious solution is to adopt a decision procedure that gives everyone equal leverage over the outcome: each person casts one vote and the majority rules. Barry describes majority voting as the “natural” solution to the decision problem. He might be interpreted as proposing a strategic rationale for majority voting, but it is easy enough to see how someone might think this solution morally appealing as well: By allowing the decision to be determined in a way that gives each preference equal weight, the majority procedure treats each person’s preference as equally important. In this way it shows equal respect for each person affected. In the absence of agreement about substantive standards, it might even seem that there is no other basis for judging a decision procedure to be fair. This argument adds an egalitarian element to the basic idea of “what touches all”: It says not only that what touches all should be decided by all but also that each member of “all” should have equal leverage in the decision-making mechanism. The meaning and basis of the egalitarian constraint raise a series of issues that are mostly extraneous to our present concerns. Leaving these issues aside, the question is what the argument suggests about the significance and status of “what touches all.” I believe the answer is this: Any procedure that excludes some of those affected from a share of control over a decision fails to show a due respect for those excluded. Exclusion embeds in the constitution of public life the idea that those excluded do not matter. As Rawls writes, the excluded, like disenfranchised slaves, are “not treated as sources of claims . . . they are not recognized as persons at all.” Inclusion, on the other hand, does treat agents as sources of claims, which is to say, it treats them as persons whose interests deserve consideration in the making of decisions that might affect them. (On this construction, whether people’s claims deserve equal consideration is a further question.) This argument comes closer than the argument we considered earlier to treating inclusion in decision making as having fundamental rather than only derivative significance. It appeals to an interest in the public status associated with the roles defined by a procedure. I once described this as an interest in “recognition” and argued that such an interest has first-order importance in reasoning about the design of institutions.  I still believe this to be true. The difficulty is that an argument of this form falls short of giving us “what touches all.” For one thing, an appeal to the interest in recognition does not settle the question of the range of decisions in which people are entitled to participate. We see this, for example, in the fact that in well-ordered democratic states the interest in recognition can be satisfied notwithstanding that authority over certain kinds of political decisions may be assigned to officials who are not retrospectively accountable to the people. (Again, think of administrative law.) Exclusion from these


Political Themes: Conservatism, Justice, and Public Reason

decisions need not convey a lack of recognition. There is also a more basic point. Imagine a society in which the power to decide is concentrated in a very small group, there is a procedure of consultation in which all voices can be heard, and partly for this reason decisions do in fact respect the important interests of most people. There are no insular minorities or excluded groups, so when decisions occasionally favor some people over others this can plausibly be regarded as a passing event, unlikely to lead over time to systematic disadvantage. Such a system is plainly incompatible with “what touches all” because it excludes most people from sharing in control over decisions that touch them, but it may nevertheless respect people’s interests in recognition. Whether in fact it does so depends on the background political culture that gives content to this interest and the performance of the society’s institutions. Even setting aside the difficulty about the range of decisions made subject to popular control, “what touches all” once again seems to be an intermediary rather than a fundamental principle.


Let me step back and try to describe the position these comments suggest. First, to summarize: We often think of democracy in a kind of shorthand as “rule by the people.” “What touches all” may seem to derive from or even to restate this idea. The picture usually in the background is hydraulic: democratic institutions are complex mechanisms that accept expressions of the popular will as inputs and transfer the force of these expressions to concrete processes of legislative and executive decision making. The people are said to “rule” when the outputs of these processes are determined by the inputs. If one accepts this picture, then it would be natural to regard “what touches all” as a fundamental principle of political morality. However, as we have seen, the considerations that might seem most plausibly to underlie such a belief fall short of justifying it. Our earlier reflections illustrate that the hydraulic picture of democracy conveyed by or, anyway, often associated with the traditional maxim is superficial. We need a more elaborate conception. To see why, it will help to recognize that the idea of democracy can be applied at several different levels of social organization. The idea is expressed differently at each level. With apologies for being telegraphic, let me try to distinguish three of these. First, democracy can be a property of decision-making procedures. As a first approximation, to say that a procedure is democratic is to say that its outcomes are determined by the expressed political preferences of those authorized to participate in it. But this is only a first approximation. Not every procedure that satisfies this description is democratic. As Barry puts it, it is not sufficient that the decision-making process “is de facto affected by the preferences of the citizens but not in virtue of any constitutional rule.” There must be some institutional connection between the expressed political preferences of the participants and the outcome of the procedure. The decision-making procedure should be arranged so that each participant has an

Global Political Justice and the “Democratic Deficit”


opportunity to be decisive over its outcomes. In an egalitarian democratic procedure, each participant’s opportunity to be decisive should be the same. (Some people might wish to say that it is part of the meaning of democracy that its procedures are egalitarian; a system in which opportunities to be decisive are unequal is not democratic. I believe this confuses things: One might have reasons to want a system to be democratic in my procedural sense from which it would not follow that the system should be egalitarian. Think, for example, about Mill’s defense of plural voting.) Let us call this idea “procedural democracy.” Second, democracy can be an attribute of a political constitution—that is, of an assignment of decision-making authority to various agents, each with its own procedural norms, together with a system of political liberties that define and protect the processes of political participation. These liberties typically consist of more than those that constitute the system of competition for public office itself; they also define the communicative environment in which various forms of public deliberation and political contestation take place, framing the formal process of political competition and providing in various other ways for the equal public status of democratic citizenship. Consistently with these provisions the constitution may also restrict the scope of legislative authority and delegate certain classes of decisions to offices not responsible to the people. When we describe such a regime as a “constitutional democracy,” we convey a judgment whose normative basis takes into account a broader range of considerations than those pertaining to the manner in which the regime’s procedures enable the political preferences of citizens to be elicited and aggregated to yield decisions. It recognizes, on the one hand, that the expression and aggregation of political preferences is the end-point of a process of agenda formation, negotiation and deliberation that have their own distinctive norms. It also recognizes, on the other hand, that this process typically operates in parallel with, or perhaps one should say it is embedded in, a larger process of political contestation—a “public sphere”—in which individuals and groups seek to advance their interests and attract support for competing political principles against a background of shared political experience. This process, too, requires a normative framework. It is not an error of classification to describe either type of norm as “democratic.” The reason is that, underlying both elements of a democratic constitution, there is an idea of popular sovereignty, of a group of people governing themselves: the constitution organizes a complex public process through which the body of citizens directs and controls the disposition of the power of the state. It is myopic to identify this process with the institutional mechanism for registering and aggregating preferences for office-holders or policies. Finally, democracy can be an attribute of a society as a whole. The idea of a “democratic society” is not simply that of a society with a democratic constitution; in Rawls’s phrase it is “a conception of society as a system of cooperation among equal persons.” As Joshua Cohen observes, such a conception has two related elements: each member is recognized as being entitled to equal respect regardless of social position and the basis of this entitlement is understood to be each person’s possession of a capacity for a sense of justice. To be sure, these elements are expressed in the political constitution itself, but they are also expressed elsewhere in the social order: in particular laws


Political Themes: Conservatism, Justice, and Public Reason

and policies (e.g., protections of equal opportunity), in the social-economic structure and in the society’s political ethos. The idea of a democratic society cannot be reduced without remainder to features of a constitution and still less to features of the main procedures of political decision-making. It is more encompassing than the hydraulic conception registers. When we think about domestic-level societies, we see that the ideas of a democratic procedure, a democratic constitution and a democratic society are related. People differ in their understandings of the relationships, but it seems clear that the content of our ideal of a democratic society will influence our conception of a democratic constitution and that this in turn will shape our view of the details of democratic procedures and their place in the larger structure. For example, as I observed earlier, the egalitarian elements of procedures of election and representation provide a form of public recognition of the equal status of citizenship fundamental to the idea of democratic society as a fair system of social cooperation. Likewise, constitutional protection of minority rights limits the authority of the legislature in order to guard against an otherwise-predictable form of inequitable treatment under majoritarian procedures, and so serves as a condition of the fairness of democratic institutions. These cases also illustrate a further element of complexity. In both cases, the line of thought connecting the idea of a democratic society with the pertinent elements of a democratic constitution must rely on various empirical premises about the social background. It is a contingent matter whether any particular procedural inequality will offend the interest in recognition; some such inequalities, taken as elements of a larger constitutional structure, may be in this respect benign. (Consider, e.g., the indifference of most U.S. citizens to the inegalitarian system of representation in the upper house of the Congress.) And the nature and importance of a constitution’s protection of minority rights will depend on the composition of the society’s population and the history and content of the political culture. This dependency on more-orless general features of the social background should of course not be surprising: most theories of democracy have judged constitutional schemes at least in part by the results, broadly construed, that they are likely to produce, and any judgment about likely results would have to take account of the relevant empirical conditions. This is true, for example, of the theories of both Mill and Rawls. An unsystematic inventory of the criteria of results found in these views include (a) the tendency of political decisions to take account of and respond to people’s most important interests, (b) their tendency to respect people’s basic liberties, (c) the system’s contribution to a public political culture in which citizens treat each other with respect as self-standing sources of claims, (d) its success in fostering an environment of effective contestation and responsible political deliberation, and (e) the system’s capacity to promote political stability, brought about particularly by its ability to induce acceptance of its procedures for succession in office as legitimate. For now I do not claim any special authority for the items on this list; I mean only to show how much is concealed about the content of the idea of a democratic society by a conception of democratic political justice that fixes on the procedures that elicit and aggregate individual political preferences.

Global Political Justice and the “Democratic Deficit”


Finally, I note that the quasi-instrumental structure of democratic justification and the resulting dependency on features of the social background has a practical reflection in the historical fact that movements for the reform of political institutions typically have not emerged primarily as efforts to bring about a well-defined ideal system of government. More often they have been responses to perceptions of injustice in a society and its political economy. Reform of political institutions has been seen as a necessary condition and perhaps a vehicle for changes in the economy and society required by considerations we might best classify as ones of social justice: the impetus has been remedial or prophylactic and the adequacy of various particular reforms has been seen to depend on their chances of success in avoiding or rectifying injustice.


I draw two lessons from these observations for the subject of global political justice. First, as I hope will be clear, the democratic-deficit view depends on a conflation of the ideas of procedural and constitutional democracy. The problem to which that view is a response is a problem about the global analog of a state’s constitutional structure— about the proper arrangement of decision-making institutions, the distribution of authority among them, their internal norms, the scope of participation in them, and the like. But the conception of democracy invoked to solve this problem is procedural. In the domestic case, we acknowledge that a democratic constitution may embrace procedures for certain types of decisions that are not themselves democratic. Which procedures should be employed and which not and, among the former, how the procedures should be regulated by the larger constitutional structure, are determined by considering the results (again, broadly construed) that would be produced by the alternative procedural configurations available under the prevailing background circumstances. We judge these results in the context of a view of the aims that a democratic society seeks to achieve through its constitution. So a judgment that a particular process—say, that for making public laws—should be democratic in the procedural sense would be the result of a nontrivial chain of reasoning. The democratic-deficit view short-circuits this chain of reasoning by appealing directly to an idea of procedural democracy. But this confuses two different levels of thought. This is significant because in the global case an argument for democratic decisionmaking procedures cannot rely with much assurance on empirical considerations parallel to those on which we rely in the domestic case. As I observed earlier, one condition of the acceptability of democratic procedures is, roughly speaking, that there be a reasonable expectation that their outputs will respect people’s most important interests. For example, there should be reason to expect that the procedures would not result in the oppression of popular minorities. I noted that the reasonableness of that expectation in ordinary cases typically rests on various empirical premises concerning the social background. Even if one holds to a view of global political society as having the potential to become a “system of cooperation among equal persons,” the inference to a principle requiring decision-making processes to be procedurally democratic would


Political Themes: Conservatism, Justice, and Public Reason

face formidable empirical challenges. The background premises ordinarily taken for granted in the domestic case—for example, about the motivation and knowledge of citizens, the density and effectiveness of the institutions that compose the public sphere and the backdrop of shared political culture that provides structure and reference points for public deliberation and contestation—are not nearly as compelling when transported to the global level. So there is less reason for confidence that democratic decision procedures in global institutions would produce broadly acceptable outcomes (e.g., that they would respect the interests of minorities). There are two points of qualification. First, the empirical contrasts between the domestic and the global levels are ones of degree and there is no reason to deny that circumstances at the global level might change in ways that would narrow the contrasts. Still, reflection about these contrasts suggests that they are substantial at present and that the changes that would be required at the global level to make the argument plausible will not easily be achieved (not least because they would depend on complex social and institutional changes at the domestic level in many of the world’s societies). Second, the skepticism I have expressed draws on a model of global democracy in which those responsible for making decisions at the global level are held finally accountable to the individuals affected by their decisions (rather, for instance, than to state-level governments). One might imagine, for example, a popularly elected global assembly, or perhaps elected special-purpose bodies differentiated functionally. This seems to me consistent with what I earlier labeled democratic fundamentalism: Its concern is that “those affected” should have a share of control over the decisions that affect them. If one shifts, say, to a federalist model in which global institutions are held accountable to the governments of states (and states are presumed to be democratic . . .), then the skeptical position might seem less persuasive because then the argument for global democracy (or democratic global federalism) would not need empirical premises parallel to those I mentioned in the domestic case. It would need different premises, taking governments as their subjects, whose plausibility is hard to gauge. But the point to be made is that the shift away from a model of accountability to individuals to one that relies on state governments or other intermediary agencies to represent the interests of individuals at the supranational level is also a shift away from the conception of democracy that underlies the democratic deficit view. The working out and defense of such a model of global institutions would more closely resemble the complex, quasiinstrumental reasoning I described earlier than a straightforward application of “what touches all.” There is also a second lesson. As I noted, part of the argument for embedding the value of political equality in the constitution of a democratic state is that doing so is a means of giving public recognition to the status of citizens as equal cooperating members of society. It conveys a communal acknowledgement of the equal individual worth of each citizen. I believe that the significance we attach to people’s interests in recognition helps to explain why we believe that opportunities for political participation should be open to all. It is essential to take seriously the public nature of this recognition. What is important is not simply that the constitution embeds equality by defining

Global Political Justice and the “Democratic Deficit”


a system of equal political liberties for all citizens. These liberties describe a public role accessible and, as we might say, visible to all. By occupying this role and observing its norms in my interactions with others who also occupy the role and observe its norms, I demonstrate my recognition of the others as equal participants in public life. The interest in recognition is satisfied through a reciprocal undertaking of a public role against a background that makes the social significance of each citizen’s conduct intelligible to the others. The question is whether there is any analogous dynamic in prospect at the global level. The answer depends in part on the direction in which one anticipates that democratic reform of the global structure might occur. If, for example, one envisages reducing the global democratic deficit by implementing some form of representation by population size in existing international organizations, it seems unlikely that individual interests in recognition would be affected much, if at all, because the basis of representation in these fora would continue to be by state. What is contemplated is essentially an adjustment of state-by-state voting weights. Establishment of a directly elected global parliamentary body might be different, because it would involve defining a new public role for individuals as global citizens. But the significance of this role and the seriousness with which it is likely to be taken would depend on various further features of the world, including the authority of the global parliamentary body itself, the effectiveness of whatever governance institutions it supervises and the nature of the political subunits of global society. If, as some writers suggest, we can plausibly imagine a reconstitution of the global political order roughly on the model of the European Union (EU) with its parliament and with political subunits which are all constitutional democracies, then perhaps the exercise of global citizenship could have a significance for individuals in some ways analogous to the exercise of citizenship within democratic states. But simply to state the antecedent conditions is to see how remote the possibility is. Even if the possibility were not so remote, we might ask whether the interest in recognition has the same importance at the global level that it has within states. As I have stressed, in contrast to the interest in avoiding substantively objectionable outcomes, the interest in recognition is importantly social: It depends for its satisfaction on reasonably transparent public institutions and on the public political culture of a society. If we imagine a world composed of societies whose institutions are arranged so that the interest in recognition is satisfied at that level, then its simultaneous satisfaction at the global level might seem less important. If so then there may be more room for instrumentally desirable variances from a rigorous application of “what touches all” in designing or reforming global institutions.


The practical aim of a theory of global political justice is to offer guidance for the reform and possibly the reconstruction of the structure of the world political order, importantly including the principal supranational political and economic institutions.


Political Themes: Conservatism, Justice, and Public Reason

In this section I want to consider briefly an alternate route by which one might be led to believe that “what touches all” should control the choices we make about institutional reform in the present, notwithstanding the doubts I have expressed so far. I observed earlier how remote it seems that the global political realm could be brought to satisfy the various social background conditions necessary to secure the analogy with the democratic state (or for that matter the EU). But someone might say that remoteness is not impossibility and that there is a point in trying to envision what a robustly democratic global institutional order would be like, even though achieving such an order would require surmounting great political obstacles. Borrowing Rawls’s distinction, such an effort would be a project in global “ideal theory” and would be a necessary preliminary to thinking systematically about the reform of our existing institutions. The latter project belongs to “nonideal theory.” It stands in a teleological relation to ideal theory: as John Simmons puts it, “nonideal theory requires us to work toward achieving the perfectly just basic structure that would result from strict compliance with the ideal theory of justice.” Nonideal theory is in this way “transitional:” “A good policy in nonideal theory is good only as . . . a morally permissible part of a feasible overall program to achieve perfect justice, as a policy that puts us in an improved position to reach that ultimate goal.” Now suppose for the sake of argument that the best ideal theory of global political justice requires the global basic structure to be (constitutionally) democratic and that within such a structure the institutions whose policies have the most pervasive impacts on people’s lives should be governed (procedurally) democratically. The line of thought I have in mind argues as follows. Because nonideal theory directs us to work toward achieving the ideal, our efforts at procedural reform in the present should aim to bring our world into a progressively closer resemblance to the ideal. Therefore we should seek to implement democratic procedures in our most consequential global institutions wherever this can feasibly and permissibly be done. This gives something like “what touches all” a toehold in the political ethics of the actual world, even if the wider social and institutional environment does not yet satisfy conditions analogous to those found within democratic societies that inform our acceptance of democratic procedures at that level. This position in nonideal theory is only as persuasive as the ideal theory from which it derives, and I have simply stipulated that an ideal theory of global justice requires establishing something like a global democratic polity that would bear comparison with domestic-level democratic societies. This claim lies at the heart of dispute about global political justice and the stipulation might well be challenged. But what I want to say here is that, even if it were accepted, the inference proposed in the last paragraph would be insecure. One reason is that, on the most plausible interpretation of the relationship between ideal and nonideal theory, it is not necessarily the case that the best thing to do in the nonideal world is to take whatever steps are open to us to bring that world into a progressively closer resemblance to the ideal. The controlling aim, as Simmons puts it, is to put us in an improved position to achieve the ideal. In general we cannot assume that a reform that brings the world into a closer resemblance to the ideal will necessarily put us in a better position to achieve it. For example, suppose, as

Global Political Justice and the “Democratic Deficit”


some people believe, that preferential treatment for racial minorities is a feasible and potentially effective means of achieving the ideal of a society in which there is fair equality of opportunity for all. Preferential treatment is incompatible with equality of opportunity in the present. The “progressive resemblance” view would therefore disallow preferential treatment: it makes society less like the ideal. But by hypothesis this would put us in a worse position to achieve the ideal. The position we are considering also faces a more serious problem. Under circumstances of substantial injustice where a complex and partly unknowable series of events must take place in order to bring about ideal justice, we may not be in a position to say whether any particular reform measure will make ideally just institutions either more or less accessible from the present. There is too much uncertainty, even if we allow that there is some uncertainty in most every political choice. What then? Taking nonideal theory to be transitional in the sense described earlier, I think we must say that although the principles of ideal theory may help us to identify injustices in the present and perhaps even to distinguish the more severe injustices from the less, those principles exercise no determinate sway over our present choices. There is no reliable way of knowing which of the choices open to us will improve the chances of achieving the ideal. It would be a mistake, of course, to infer that political choices in the nonideal world are therefore not open to moral appraisal at all or that there are no moral standards by which they should be guided. We may face choices in the nonideal world that cannot reasonably be decided in the perspective of a controlling ideal, but we may nevertheless aspire to act for the better. What we ordinarily cannot do under the conditions I described is to import ideal principles into the nonideal world and hold that they should control our choices in this world. This shows that the alternate route I described does not lead to something like “what touches all,” at least not in the rigorously straightforward interpretation we have been considering. Instead, it points towards a perspective on practical choices about institutional reform compatible with the instrumental or remedial understanding of political justice I gestured at earlier. I do not believe there is any very informative shorthand description of this perspective. Confronted with the possibility of acting towards reform of existing decision-making institutions, we seek to achieve improvement across several dimensions which are unlikely to be systematically related and whose relative importance is likely to vary with the context. Thus, for example, we may aim to create better incentives for agents to align their conduct with the purposes of the institution and its norms, to induce the provision of information needed to assess the performance of these agents, and to stimulate horizontal sharing of information about issues arising in multiple venues. We may seek to elicit feedback from those affected by decisions reached and to create incentives for agents to take account of it. We may try to devise procedural safeguards to protect against or remediate objectionable outcomes, particularly when they are predictable under typical circumstances. Various institutional mechanisms that might be capable of achieving such goals are imaginable. I cannot discuss them here. But simply to observe the range of aims shows that reflection about principles for the reform of global governance must be more complicated than the democratic-deficit view might suggest. It must face the


Political Themes: Conservatism, Justice, and Public Reason

prospect of conflict among various desiderata for acceptable procedures under conditions in which there is no clear priority rule. These observations apply mainly to the reform of the decision-making procedures of existing institutions and regimes. But the competences of existing institutions are limited. Often the point of appeals to “what touches all” is to argue that these competences should be extended or new institutions established for types of conduct not now subject to effective international regulation (for example, greenhouse gas emissions or humanitarian intervention). The question posed is whether the scope of global governance should be expanded to allow those affected by the conduct of some class of agents to bring this conduct under a form of political control in which those affected can participate. How might the perspective I have sketched influence a reply to questions of this form? Because that perspective is instrumental and highly sensitive to context, any brief reply would have to be abstract. The main point is that two issues are combined in the question as I have formulated it. One concerns the proper scope of political authority; the other, the mechanism by which this authority should be controlled. We should not assume that the considerations that determine a resolution of the first will also determine a resolution of the second. Thus, for example, one might favor establishment of a global climate change regime on the grounds that a continued increase in atmospheric greenhouse gas concentrations will have very widespread bad effects without also holding that such a regime should be governed in a way that entitles all affected by it to a voice or a vote. In general, institutional design should respond to an array of desiderata whose simultaneous satisfaction, particularly under highly non-ideal circumstances, may not be possible. In particular, considerations of effectiveness, cost and inclusiveness may conflict. The regime most likely to be effective in regulating the undesirable consequences of some type of conduct may not be the most inclusive. In that case we face a need for compromise among conflicting desiderata analogous to that which we observed in the case of reform of existing decision-making procedures. Practical judgment will depend heavily on an appreciation of features specific to the case informed by a more general grasp of the politics of international cooperation. It is unlikely to be determined by general principles of political morality.


Because these last remarks are fairly abstract, I want to conclude by considering briefly how the perspective I have suggested might influence the way we think about two ideas for global governance reform that have currency today. Consider first the idea that global governance processes should be made more accountable. There has been disagreement about how this idea should be interpreted. A central question is whether accountability mechanisms should make decision makers accountable to those affected by their decisions (i.e., the “decision takers”) rather than to other parties such as an institution’s principals or third parties like NGOs. The democratic deficit view answers in the affirmative as a matter of principle. If one were to

Global Political Justice and the “Democratic Deficit”


take the broader perspective I have gestured at, one would reason more instrumentally, asking which way of configuring the mechanisms for accountability would likely be most successful in alleviating the most serious of the injustices to which the institution in which the procedures operate is vulnerable. If, for example, it turns out that accountability mechanisms that empower third parties to act on behalf of decision takers are more effective in protecting the latter’s interests, it would be a mistake to resist on the grounds that such mechanisms offend against political justice. More generally, there may be more space for reliance on NGOs, or expert commissions, or stakeholder consultations than a concentration on redressing a perceived democratic deficit would allow. To illustrate, consider the international standard-setting process for state-level regulation of banking practices and payment systems. Among other aims, bank regulation seeks to control the extent to which the activities of individual banks subject other agents to financial risk. Because the global capital market is highly integrated, regulatory failures in one state can rapidly impose external costs on agents elsewhere. The coordination of bank regulation is therefore an important function of global governance. The main provider of this coordination at present is one of three committees hosted by the Bank for International Settlements in Basel—the Committee on Banking Supervision (or “Basel committee”), established by the governments of the ten largest economies (the “G10”) and now enlarged to include representatives of seventeen states. Some observers argue that this committee’s failure to adopt sufficiently rigorous banking standards contributed to the recent global recession. The Basel committee is composed of experts nominated by governments but it functions as an expert body with a considerable degree of insulation from domesticlevel political authorities. It has no independent legal standing and no authority to make law; instead, it seeks to coordinate bank regulatory policy within states through various informal means aimed at influencing regulators. Although its membership is limited, its standards have acquired global authority, in part because the major international financial institutions expect adherence to the standards as conditions for support. The last point means that the Basel committee should be an easy target for a proponent of the democratic deficit view: it is a clear case of a governance institution whose activities are potentially of great importance for the well-being of agents who are excluded from participation in the decisions that produce them. It might be argued as a matter of political justice that the global coordination of bank regulation should be reorganized in such a way that those authorized to carry out this function can be held accountable by those affected by their decisions. If, however, we consider the matter from the instrumental perspective suggested earlier, we might take a different view. The justification of a reform program would require an identification of the undesirable consequences brought about by the existing process, a description of an alternative to the existing process that could reasonably be regarded as achievable and an argument that the alternative, if implemented, would be less harmful than the existing process without generating other results that would be even less desirable. In the case of international coordination of bank


Political Themes: Conservatism, Justice, and Public Reason

regulation, one recent analysis proposes that the existing informal process be replaced by a regime based on a negotiated global treaty containing provisions that render the main policy decisions accountable to a global financial governance body composed of specialist representatives of the states parties to the treaty. Its aim is to generate better outcomes for the individuals ultimately affected by bank regulatory policies, whereas the means proposed consist of measures that aim to make the regulatory regime more effectively accountable to officials of states. The argument is entirely instrumental and relies on the empirical premise that accountability to specialist state officials within a treaty-based (or constitutionalized) structure is the best available means of protecting the interests of individuals. Assuming arguendo that the premise is correct, this does not seem objectionable as a matter of democratic principle. A similar analysis applies to the idea that the global democratic deficit could be narrowed by establishing a popularly-elected global assembly. This idea has arisen in connection with the United Nations (UN) governance system. Some people believe that a system based on a General Assembly with representation by state and a onestate, one-vote decision rule, together with a Security Council with a great-power veto, is objectionably undemocratic. They argue that an elected “second chamber” would help rectify the deficit. For example, Stanley Hoffmann takes it to be a “point of principle” that the system needs “far greater democratization” and proposes an “Assembly of People’s Representatives” that would “introduce unofficial voices into the global debates.” Which voices would be heard in such an assembly and how, if at all, it would influence the governance of the UN or of global institutions more generally plainly depend on the details of the reform proposal, in particular on the authority of the body, its size and the provisions for election of members. Hoffmann’s proposal is typical in suggesting that the popular assembly would initially have only “powers of general recommendation” but anticipating that it would gain more substantial authority in time. The proposal’s advocates appear to be motivated by a version of the democraticdeficit view. For the reasons given earlier, I do not believe we should find the argument persuasive. But it is worth asking if there are considerations internal to democratic theory other than those associated with that view that bear on an assessment of the proposal. Off hand there seem to be at least two. First, to some extent the appeal of a popular second chamber is that it would improve the quality of global governance. (I say “global governance” rather than “the governance of the UN” because some advocates believe a second chamber might exercise oversight over, or at least take cognizance of, the activities of other international organizations as well.) The mechanism is one of accountability: the hypothesis must be that by making international decision-makers accountable to such a forum, an incentive would be established to take more seriously the consequences of decisions for those affected. It is not clear, however, how effective this form of accountability might be, given the assembly’s lack of authority over decision makers and the various reasons for doubting that a global representative assembly could function well in the absence of the background conditions discussed earlier. If the aim is to align the incentives of decision

Global Political Justice and the “Democratic Deficit”


makers with the interests of decision takers, then, as before, there may be other and possibly more effective ways to do so, perhaps functionally differentiated and established closer to the point of decision. (Indeed, this may illustrate one danger of the democratic deficit view: it may focus reform efforts in the wrong place.) The other and perhaps more interesting consideration is the hope that a global popular assembly might help make global governance more “transparent.” The meaning and significance of this may be obscure. Transparent procedures are ones that operate “in the open”—that is, in ways that can be monitored by those who are not parties to the procedures. Transparency is achieved by establishing mechanisms that make reliable information about the conduct of the procedures available to nonparticipants with an interest in monitoring them. But the amount and type of information made available and the range of those to whom it is made available can vary, so transparency can come in differing forms and degrees. Moreover, it can answer to differing aims. For example, transparency is sometimes valued as a means to minimize institutional corruption; the requirement to disclose information about how a decision has been reached is a disincentive for officials to act for private gain. Although corruption is a concern in global (as in other) institutions, the aim more closely associated with the idea of democracy is that of achieving a shared, public understanding of the standards according to which decisions have been reached. Transparency with this aim would seek to create incentives for decision-makers to reveal the reasons for their decisions and, inter alia, to make available the information on which these decisions rest. When a process is transparent in this way, nonparticipants can see that decisions have been reached according to the appropriate standards or principles and in a manner that was adequately informed by the pertinent facts. One reason to value transparency is that it may increase the chances that decisions will be accepted as legitimate and complied with voluntarily. Another is that it brings the grounds of decisions into the public realm where they can be challenged and contested. In both respects, the transparency of decision-making processes aims to make it possible for the application of norms to become a subject of public concern, or as we might say, to constitute a kind of public reason. Would a global people’s assembly actually contribute to achieving greater transparency in global governance decisions? The question is speculative, but there are reasons for doubt. Having only the power to recommend, the assembly would not itself have decision-making authority, so whatever transparency it might achieve would have to come about through some process that makes possible the review of decisions made by other agents. Because the assembly would not have the range of powers typically disposed by a national legislature, it would lack leverage to establish incentives and disincentives for those agents that would be necessary to render their decision-making processes transparent. And the assembly would have to operate in a decentralized global institutional order in which authority for many of the most consequential decisions rests outside the UN system—in autonomous international financial institutions, for example, or in network structures like the Basel committee—so even the power to recommend would have to be exercised across a jurisdictional gap.


Political Themes: Conservatism, Justice, and Public Reason

The growth of a global public sphere in which decisions can be contested on the basis of publicly shared norms seems essential to any aspiration for democratic reform in global governance. It would be a part of a global analogue of the idea of a democratic society. But, to speculate again, a more realistic strategy for advancing this goal may be to exploit rather than resist the fact that the world political order lacks the institutional machinery of the state. This would mean concentrating on functionally specific regimes and institutions and developing mechanisms appropriate to their aims to render their decision-making processes more transparent. One domain in which one sees intimations of such a development is that of global administrative law. Another is that of global human rights, where a rich discursive practice is developing consisting of a body of shared public norms and a range of legal and political mechanisms for implementing them. In both domains the contribution to an emergent global public reason will come about, to the extent it does, otherwise than through the activities of an institution modeled on a domestic-level parliament. I would not want to leave the impression that there is nothing to the idea of a global democratic deficit. The main point is that if there is something objectionable about global institutions that lack the trappings of democratic political institutions as we know them in domestic society, it is not simply that those affected by their decisions lack a share in control over those decisions. We need to know what harms those affected are exposed to by a decision-making process and why we should believe that changes in the process aimed at introducing forms of democratic accountability would alleviate the harms. We also need to know if some other means of reforming institutions would secure the substantive end of avoiding or alleviating these harms more effectively. Any adequate theory of global political justice would address these questions somehow. The trouble with the simple theory of a democratic deficit is that it deflects attention from these questions rather than concentrates on them. notes * I am grateful to Samuel Freeman, Robert Keohane, and Stephen Macedo, and to audiences at workshops at Cornell University and the Monash University Center in Prato, Italy for comments on various ancestors of this essay. Although his name does not appear elsewhere in these footnotes, I want to express my very deep thanks to Tim Scanlon for his philosophical advice and personal friendship ever since I was a graduate student. Anyone who knows him will see the influence of his thought in this essay. My life as a scholar would not have gone as it has without the inspiration of his ideas and example. 1. “Some mixture” because to some extent the two levels of reform may be substitutable (for example, domestic level decisions likely to have external effects might be brought within an international regulatory scheme). 2. For example, David Held, “Cosmopolitanism: Globalisation Tamed?” Review of International Studies 29 (2003): 466; David Held, “Democratic Accountability and Political Effectiveness from a Cosmopolitan Perspective,” Government and Opposition 39 (2004): 372–76; Daniele Archibugi, The Global Commonwealth of Citizens: Toward Cosmopolitan Democracy (Princeton, N.J.: Princeton University Press, 2008), esp. ch. 2.

Global Political Justice and the “Democratic Deficit”


3. This position is suggested by Ruth W. Grant and Robert O. Keohane, “Accountability and Abuses of Power in World Politics,” American Political Science Review 99 (2005): 29–30, 33–34. 4. Robert A. Dahl, “Can International Organizations Be Democratic?” in Democracy’s Edges, ed. Ian Shapiro and Casiano Hacker-Cordón (Cambridge: Cambridge University Press, 1999), 30–32. 5. John Rawls, The Law of Peoples (Cambridge, Mass.: Harvard University Press, 1999), 36. For Kant’s view, Immanuel Kant, Toward Perpetual Peace: A Philosophical Sketch (1795), in Practical Philosophy, trans. and ed. Mary J. Gregor. The Cambridge Edition of the Works of Immanuel Kant (Cambridge: Cambridge University Press, 1996), 336 [Ak. 8:367]. 6. One finds a suggestion of this position in one strand of Kant’s thought. Ibid., 328 [Ak. 8:357]. 7. Gaines Post, Studies in Medieval Legal Thought: Public Law and the State (Princeton, N.J.: Princeton University Press, 1964), 170. For the Roman origins, see Justinian, C. 5, 59, 5, 2 (“quod omnes similiter tangit, ab omnibus comprobetur”). 8. From the perspective of a monarch in desperate need of tax revenue, this seems to have been more a matter of practical necessity than political conviction. Michael Prestwich, Edward I (London: Methuen 1988), 466–67. 9. C. H. McIlwain, Constitutionalism and the Changing World: Collected Papers (New York: Macmillan, 1939), 145. 10. Post, Studies in Medieval Legal Thought, 237–38 and Brian Tierney, “Freedom and the Medieval Church,” in The Origins of Modern Freedom in the West, ed. R. W. Davis (Stanford, Calif.: Stanford University Press, 1995), 86–88. 11. This word is used in several ways in the literature. Some writers take it to refer to the group of individuals entitled to participate in making political decisions; others, to (something like) the political community or polity. It is important that these are distinct ideas. For clarity in the text I adopt the first usage, though in fact I believe we need a more elaborate conception. There is a helpful discussion in Christian List and Mathias Koenig-Archibugi, “Can There Be a Global Demos? An Agency-Based Approach,” Philosophy & Public Affairs 38 (2010): 76–110. 12. Robert E. Goodin, “Enfranchising All Affected Interests, and its Alternatives,” Philosophy & Public Affairs 35 (2007): 40–41 n. 1. 13. Jeremy Waldron, Law and Disagreement (Oxford: Oxford University Press, 1999), 114. 14. There is a small literature devoted to the analysis of this maxim. I rely especially on Frederick G. Whelan, “Preface: Democratic Theory and the Boundary Problem,” in Liberal Democracy (Nomos XXV), ed. J. Roland Pennock and John W. Chapman (New York: New York University Press, 1983), 13–47 and Goodin, “Enfranchising All Affected Interests, and Its Alternatives,” which inter alia provides a good guide to the previous literature. See also Gustaf Arrhenius, “The Boundary Problem in Democratic Theory,” in Democracy Unbound: Basic Explorations. Stockholm Studies in Democratic Theory (Stockholm, Sweden: Filosofiska Institutionen, Stockholms Universitet, 2005), 14–28; Hans Agné, “A Dogma of Democratic Theory and Globalization: Why Politics Need Not Include Everyone it Affects,” European Journal of International Relations 12 (2006): 433–58; and David Miller, “Democracy’s Domain,” Philosophy & Public Affairs 37 (2009): 201–28. 15. Perhaps not “and only,” but for our purposes it does not matter. 16. Ludvig Beckman defends one form of the deontic view in “Democratic Inclusion, Law, and Causes,” Ratio Juris 21 (2008): 348–64. Goodin holds that such views produce a “fatally underinclusive” principle (“Enfranchising All Affected Interests, and its Alternatives,” 50). But


Political Themes: Conservatism, Justice, and Public Reason

the judgment that the underinclusion is “fatal” seems circular: the principle is objectionably underinclusive only if it is already accepted that the relevant notion of being “affected” is causal. 17. As Goodin explains. Ibid., 52–55. 18. Waldron, Law and Disagreement, 114. Earlier, however, he writes that when a “problem affects millions, then a respectful decision procedure requires those millions to listen to one another and to settle on a common policy in a way that takes everyone’s opinion into account” (110). This is compatible with the abstract variant. 19. John Rawls, A Theory of Justice, rev. ed. (Cambridge, Mass.: Harvard University Press, 1999), 205. 20. Goodin puts it more radically: “we will have to give a say to anyone who might possibly be affected by any possible decision arising out of any possible agenda.” “Enfranchising All Affected Interests, and its Alternatives,” 55. This is the formulation he endorses. 21. Careful readers will note that this is ambiguous. Say that an issue is a set of alternative outcomes—say, issue X consists of alternatives A and B. A decision is a choice among these alternatives. Suppose the decision actually taken is A. One sense of a decision that might have been but was not taken is B. But decision-making agents are typically authorized to make decisions concerning some range (possibly an unrestricted range) of issues—for example, X, Y, Z—and to exercise some degree of control over which issues come up for decision. Suppose the agent places X and Y on its agenda. Another sense of a decision that might have been but was not taken is the resolution of issue Z. The formulation in the text should be read to embrace both of these. 22. Though not everyone thinks so; David Miller defends a more restricted form of the maxim in “Democracy’s Domain.” 23. As Waldron has it (Law and Disagreement, 114). 24. Marc Fleurbaey and Harry Brighouse take this to be objectionable and propose a method by which it might be avoided in “Democracy and Proportionality,” Journal of Political Philosophy 18 (2008): 1–19. 25. As Daniel Weinstock summarizes the point: “In the pursuit of the realization of their citizens’ interests, actually existing democracies have come up with a dizzying array of undemocratic and indirectly democratic institutions and mechanisms aimed at securing people’s vital interests against the epistemic and motivational shortcomings of directly democratic institutions.” “The Real World of (Global) Democracy,” Journal of Social Philosophy 37 (2006), 19. 26. The seminal work is Richard E. Stewart, “The Reform of American Administrative Law,” Harvard Law Review 88 (1975): 1667–813. There is a discussion in Henry S. Richardson, Democratic Autonomy (Oxford: Oxford University Press, 2002), ch. 1. 27. John Stuart Mill, Considerations on Representative Government, in Essays on Politics and Society II, ed. J. M. Robson [Collected Works, vol. 20] (Toronto: University of Toronto Press, 1977), ch. 4. 28. The formula, “should take account . . . ,” leaves it open how account should be taken and, presuming that account-taking involves comparing the impact of a prospective decision on a range of interests, the basis upon which the comparison should be made. These are substantive questions of political morality. Part of the criticism of the institutional version of “what touches all” is that it deflects attention from these questions. 29. Brian Barry, Political Argument (London: Routledge & Kegan Paul, 1965), 312. See also Barry’s reconsideration of the case in “Is Democracy Special?” Philosophy, Politics, and Society,

Global Political Justice and the “Democratic Deficit”


Fifth series, ed. Peter Laslett and James Fishkin (New Haven, Conn.: Yale University Press, 1979), 161–72. I consider this case in Political Equality: An Essay in Democratic Theory (Princeton: Princeton University Press, 1989), 75–77. 30. Compare Waldron’s defense of majority rule in Law and Disagreement, 114ff, and the earlier discussion of voting and respect, 108–10. 31. I discuss them in part I of Political Equality. 32. John Rawls, Political Liberalism (New York: Columbia University Press, 1996), 33. 33. In Political Equality, ch. 5. For a similar view developed in greater detail, see Thomas Christiano, The Constitution of Equality (Oxford: Oxford University Press, 2008), ch. 3. 34. John Rawls’s idea of a “decent consultation hierarchy” is such a system, but it need not be taken as paradigmatic. The Law of Peoples (Cambridge, Mass.: Harvard University Press, 1999), sec. 9. 35. Compare Joshua Cohen, “For a Democratic Society,” in The Cambridge Companion to Rawls, ed. Samuel Freeman (Cambridge: Cambridge University Press, 2002), 91–103. The distinctions here differ from Cohen’s. 36. Barry, “Is Democracy Special?” 25–26. 37. Well discussed in Philip Pettit, “Democracy, Electoral and Contestatory,” in Designing Democratic Institutions (Nomos XLII), ed. Ian Shapiro and Stephen Macedo (New York: New York University Press, 2000), 105–44. And see Jürgen Habermas’s illuminating account of the importance of the public sphere in modern democratic politics in, e.g., The Structural Transformation of the Public Sphere, trans. Thomas Burger (Cambridge, Mass.: MIT Press, 1989). 38. We do not need an elaborately romantic conception of a “people” to make sense of this idea. See Samuel Freeman, “Original Meaning, Democratic Interpretation, and the Constitution,” Philosophy & Public Affairs 21 (1992): 3–42. 39. Rawls, A Theory of Justice, 336. 40. Cohen, “For a Democratic Society,” 96. 41. See, for example, Simon Caney’s description of cosmopolitan political institutions in Justice Beyond Borders (Oxford: Oxford University Press, 2005), ch. 5, esp. 160–64. In his view a democratic framework for global politics should render global institutions accountable to persons, not states. The argument for this view combines right-based and instrumental considerations, so it would be inaccurate to characterize the view as a form of democratic fundamentalism. 42. For this formulation see Jack Lively, Democracy (Oxford: Blackwell, 1975), 134–35 and the discussion in my Political Equality, 109–10, 113–14. 43. That is, it might be the result of a search for principles that would characterize a fully just world order under the assumptions of “strict compliance” with the principles under “favorable” historical and social conditions. Rawls, A Theory of Justice, 216. I leave aside some important qualifications. See A. John Simmons’s penetrating discussion in “Ideal and Nonideal Theory,” Philosophy & Public Affairs 38 (2010): 5–36. 44. Ibid., 18, 22. 45. Thanks to Samuel Freeman for the example. And see Simmons, “Ideal and Nonideal Theory,” 24–25. 46. Simmons writes: “Then, presumably, we must simply muddle through the best we can . . . .” Ibid., 24. On the indeterminacy of nonideal theory under likely empirical circumstances, see Joel Feinberg, “Duty and Obligation in the Non-Ideal World,” Journal of Philosophy 70 (1973): 263–75.


Political Themes: Conservatism, Justice, and Public Reason

47. See, for example, the provocative and illuminating discussion of “deliberative polyarchy” beyond the state in Joshua Cohen and Charles Sabel, “Global Democracy?” N. Y. U. Journal of International Law and Politics 37 (2005): 763–97. 48. Climate change may be a case in point; the Kyoto regime, which is highly inclusive by contemporary standards, has not been notably effective. I am grateful to Robert Keohane for conversation about this, and to Robert O. Keohane and Kal Raustiala, “Toward A PostKyoto Climate Change Architecture: A Political Analysis,” in Post-Kyoto International Climate Policy: Implementing Architectures for Agreement (Cambridge: Cambridge University Press, 2009), 372–402. 49. Here I am agreeing with Grant and Keohane, who write that “effective accountability at the global level will require new, pragmatic approaches.  .  .  .” “Accountability and Abuses of Power in World Politics,” 34. Unlike them, however, I would not regard such approaches as a departure from democratic principles. 50. There is an accessible discussion in Kern Alexander, “Global Financial Standard Setting, the G10 Committees, and International Economic Law,” Brooklyn Journal of International Law 861 (2008–9): 861–81. 51. Alexander, “Global Financial Standard Setting,” 880–81. 52. For example, by formulating model policies, promulgating “best practices,” establishing principles for cooperation among central banks and maintaining and supporting a network of state-level officials. There is an illuminating discussion, placing the Basel committee in a wider framework of “global network governance,” in Anne-Marie Slaughter, A New World Order (Princeton, N.J.: Princeton University Press, 2004), ch. 1. 53. Kern Alexander, Rahul Dhumale, and John Eatwell, Global Governance of Financial Systems: The International Regulation of Systemic Risk (Oxford: Oxford University Press, 2006). ch. 5, esp. 172–73. 54. The assumption is only arguendo. Although constitutionalizing the Basel process would likely render it more responsive to the interests of its decision-takers, there is a further question about the character and norms of the broader global discursive environment in which it would operate. It is a challenge to conceptualize the institutional or regime structures this would require. Cohen and Sabel’s reflections about “global anomalous administrative law” are instructive. “Global Democracy?” esp. pt. IV. 55. He also suggests establishing a “consultative assembly” of representatives of NGOs and corporations. Stanley Hoffmann, “World Governance: Beyond Utopia,” Daedalus 132 (2003), 33. He recognizes the obstacles to such a reform (Ibid., 34–35). For similar proposals see Johan Galtung, “Imagining Global Democracy,” Development and Change 35 (2004): 1073–79, Richard Falk and Andrew Strauss, “Toward Global Parliament,” Foreign Affairs 80 (January–February 2001): 212–20 and Boutros Boutros-Gali, “The Missing Link of Democratization” (June 9, 2009), available at (consulted August 19, 2010). 56. Galtung proposes that the people’s assembly be composed of one delegate per million people, using states as constituencies; evidently he believes this is the minimum that would afford effective representation (“Imagining Global Democracy,” 1074). Today the world’s population is about 6.9 billion, so the body would have about 6,900 members. For comparison, the largest national legislature appears to be the National People’s Congress of the People’s Republic of China, whose eleventh session had 2,987 members (See englishnpc/Organization/node_2846.htm, consulted August 19, 2010). Thanks to Christopher Achen for conversation on this point.

Global Political Justice and the “Democratic Deficit”


57. Hoffmann, “World Governance: Beyond Utopia,” 33. Cf. Boutros-Gali, “The Missing Link of Democratization” (“Over time its authority and powers could evolve.”). 58. Here, for example, is Boutros-Gali: “Applying democratic principles to international institutions must be an essential component of any reform of global governance.  .  .  . A direct link between global institutions and the people on the spot needs to be established.” Ibid. 59. One example is a proposal to establish a special procedure for authorization of preventive uses of military force. It aims to create incentives for states to obtain ex ante authorization by and to provide ex post information to an impartial international body. Allen Buchanan and Robert O. Keohane, “The Preventive Use of Force: A Cosmopolitan Institutional Proposal,” Ethics and International Affairs 18(1) (2004): 1–22. 60. Boutros-Gali, “Missing Link of Democratization.” 61. There’s an NGO for that: Transparency International, a “global coalition against corruption” ( 62. See in particular the illuminating discussion of the prospects in Cohen and Sabel, “Global Democracy?” esp. 794–97. The now-canonical source is Benedict Kinsbury, Nico Krisch and Richard B. Stewart, “The Emergence of Global Administrative Law,” Law and Contemporary Problems 68(3–4) (Summer/Autumn 2005): 15–61. 63. There is a summary description in chapter 2 of my The Idea of Human Rights (Oxford: Oxford University Press, 2009).

11 Establishment, Exclusion, and Democracy’s Public Reason Joshua Cohen

the first amendment to the U.S. Constitution bans laws “abridging the freedom of speech,” as well as laws respecting an establishment of religion or prohibiting its free exercise. No simple theory can explain the contours of the speech clause: the many varieties of speech that merit protection, or the nature of those protections. But one important rationale for the First Amendment’s stringent protections of speech has to do with democracy. Alexander Meiklejohn’s Free Speech and Its Relation to Self-Government presents the classic democratic interpretation of the free speech clause. A central purpose in assuring free expression, he says, is to establish conditions required for popular self-government, for the supremacy of the people over the government. The free speech right is not best understood as limiting democracy in the name of “a Law of Nature or of Reason,” or as simply restricting it to protect individuals from intrusive government regulation. Instead, that right is also democracy-enabling. It “springs,” Meiklejohn said, “from the necessities of the program of self-government.” Popular self-government depends on open public discussion and reflection, on the “thinking process of the community.” Free speech protections secure this process of reasoning together from “mutilation,” at least at the hands of government. In so doing, they help to enable “politically equal” citizens to form collective judgments and exercise ultimate authority. At least since Meiklejohn’s influential argument, democratic themes have played an important (though by no means exhaustive) role in interpreting the First Amendment’s prohibition on laws “abridging the freedom of speech, or of the press.” Considerations about the preconditions of democracy have not played a comparably central role in interpreting the First Amendment’s religion clauses. But they should. Although religious freedom also has many roots, and no simple story can 256

Establishment, Exclusion, and Democracy’s Public Reason


explain the contours of its appropriate protections either, there is an important democratic strand in the case for ensuring religious freedom. To be sure, this case is different in its details from the democracy-constituting argument about free expression. But it is also about establishing conditions that enable the members of a political society, with their conflicting fundamental convictions, to reason together as political equals. I will focus here on the establishment clause (although a comparable story can be told about free exercise). The religion clauses, I will argue, are not only about protecting religious conviction, conscience, and conduct from intrusive government regulation. They also are about the inclusion of equals in the enterprise of self-government. That inclusion presents a central challenge to pluralist democracies, but also lies at the heart of their moral promise. To focus my discussion of this general theme, I will explore the idea that the establishment clause expresses, among other things, a concern about religious endorsements. The core idea is that the First Amendment is hostile to laws and public policies that can reasonably be understood as conveying an official endorsement of a particular religion or of religion generally. The policies that arguably convey such endorsement range widely, from symbols and prayers, to congregationalist (default-majoritarian) churchproperty regulations, allocations of public funds to religious organizations, school curriculum decisions, and zoning rules. Although concerns about official religious endorsements are hardly new, Justice O’Connor underscored the constitutional issue in her concurrence in Lynch v. Donnelly (1984), a public symbols case about a Nativity scene in a Pawtucket holiday display. As it happens, O’Connor denied that the display amounted to an official endorsement, and thus denied that it was an objectionable establishment. But along the way, she urged that official endorsements are troubling, and stated the objection: that they make “status in the political community” turn on religious affiliation. They distinguish favored “insiders” from unfavored “outsiders.” I will call this the Endorsement-Exclusion Thesis. According to this thesis, religious endorsements are troubling because they exclude. The Endorsement-Exclusion Thesis has not commanded the unalloyed enthusiasm of legal scholars or judges. Still, it captures something important about religious establishments. Interpreted along the lines I will propose, the Endorsement-Exclusion Thesis suggests two democratic themes at work in our reflection on what a religious establishment is and why it is of concern. Resistance to exclusionary endorsements responds, first, to a concern about usurpation. Religious conceptions are outlooks of comprehensive scope comprising fundamental convictions (and associated practices). An endorsement inappropriately takes a collective position on these matters. It announces our position on a subject on which, because of the nature of those convictions, no such collective position is possible (except by restricting the bounds of the political community). Rejecting exclusionary endorsements helps, therefore, to constitute a political community of equals, each recognized as having the independent powers of judgment associated with popular self-government. Moreover, second, hostility to exclusionary endorsements helps (because of the same features of religious convictions that I just mentioned) to enable codeliberation among political equals on a terrain of public reason.


Political Themes: Conservatism, Justice, and Public Reason

My discussion has six parts. First, I distinguish endorsement from three other familiar concerns about religious establishment. Second, I sketch the EndorsementExclusion Thesis, and interpret the civic exclusion it contemplates and condemns as exclusion in the space of reasons. Third, I clarify the Endorsement-Exclusion thesis. I will say that we have an endorsement of the relevant kind when an informed observer has good reason to believe that a collective decision was taken because of religious convictions; in particular, the convictions provide the most plausible rationale for the policy, given the context of the policy and the specifics of its design. Fourth, I ask why we should accept the Endorsement-Exclusion thesis. To focus the question, I ask why, if religious endorsement excludes, the same thing is not true for any controversial endorsement? To answer, I present, fifth, an account of religion. I emphasize that religious views express conceptions of the sacred that give adherents practical reasons that are comprehensive in scope and fundamental in substance. Sixth, drawing on this conception, I explain why concerns about endorsement have particular force when it comes to endorsements of religious views. Such endorsements usurp independent judgment and exclude from the space of reasons. A ban on endorsements, in contrast, helps to constitute equal standing and democracy’s public reason. I conclude with a comment on the importance of inclusion in the space of reasons.

i. three establishment issues

Establishment is not about just one thing, not simply a concern about coercion (shared with the free exercise clause), or about anti-discrimination (shared with the equal protection clause). So before getting to endorsement, I want to locate the specific concerns by distinguishing them from three other establishment issues. First, liberty. Classical religious establishments—the ones that especially concerned Locke and Madison—included forced endorsement of an official confession; required church attendance; bans on practices associated with religious worship; payment of clergy and support for churches from tax revenues; compulsory prayer. These establishments raise issues about the liberty of conscience and worship associated with acting on determinate religious convictions, as well as the freedoms needed for the formation of those convictions. Coercive constraints on fundamental convictions and the worship associated with those convictions certainly raise religious establishment concerns. But those concerns extend beyond coercion, even in a very capacious sense, understood to comprise psychological, peer-pressure coercion (what Justice Scalia once disparaged as merely Freudian, not robustly Blackstonian). Putting a cross on city hall, for example, would be an unacceptable establishment. In an effort to treat coercion as the exclusive establishment concern, we might call this “coercion by proselytizing” (imagine that the cross is donated, so no tax revenues are used to pay for it). But instead of stretching coercion beyond recognition, it seems best to say that the trouble is endorsement. Justice Scalia said precisely that when he observed that the constitution condemns:

Establishment, Exclusion, and Democracy’s Public Reason


[G]overnment-sponsored endorsement of religion—even when no legal coercion is present, and indeed even when no ersatz, “peer-pressure” psycho-coercion is present—where the endorsement is sectarian, in the sense of specifying details upon which men and women who believe in a benevolent, omnipotent Creator and Ruler of the world are known to differ (for example, the divinity of Christ). Scalia offers no defense of this parochial sense of “sectarian.” But my interest for now is confined to observing that concerns about religious establishment extend beyond coercion. What else is at stake? A second concern is about protecting religion from politics. Thus religion is arguably damaged by entanglements that make it politically accountable, say, as condition for the receipt of funds, or when its ideas are subjected to the “unhallowed perversion” required for public consumption, as when a rabbi is given a pamphlet with guidance on how to present religious ideas at a middle school graduation ceremony. When church and state are intimately related, Locke observes, the church “is for the most part more apt to be influenced by the court than the court by the church.” Thinking of the Anglican Church post-1538, he observes “how easily and smoothly the clergy changed their decrees, their articles of faith, their form of worship, every thing, according to the inclination of those kings and queens.” To be sure, the Tudors and Stuarts unhesitatingly backed their regal inclinations with uncomplicatedly Blackstonian coercion. Nevertheless, Locke is pointing to an establishment concern that does not reduce to coercion. To take contemporary examples, we might worry about the fate of religiosity when officials prepare benedictions by committee to ensure a colorlessly nonsectarian civil religion, or decide on the precise forms of neutering or architectural arrangement that suffice to render displays of religious symbolism publicly acceptable (whether accompanying a crèche with three reindeer makes the crèche acceptable), or (in the spirit of Lord Eldon’s rule on implied trusts) make church-property rules founded on determinations about which faction has departed from Church doctrine, or investigate whether groups have convictions that suit them to receive public resources, or suitably segregate their unsuitable convictions from their suitably public-spirited conduct. A third reason for concern about establishment is to protect politics. Thus religious conviction has been a source of or at least vehicle for arriving at, or at least for powerfully expressing in vivid form, ethical insight, sometimes with large and constructive political implications. Think of abolitionism, with its evangelical roots; or the American civil rights movement, the just war tradition, or arguments about life’s beginning and end. We do not want to flatten the terrain of ethical reflection through entanglements that turn religious conviction into vehicle for political contest, or yet another occasion for an administrative colonization of the lifeworld. In addition, religion can damage politics by fueling social conflict. In a political society with strong religious convictions, the nonestablishment hope is that you can diminish the severity of that conflict by reducing its stakes. By eliminating the possibility of a confessional state, or imposing a high bar to establishing large prizes


Political Themes: Conservatism, Justice, and Public Reason

for which conflicting convictions may compete, you may transform the conflict between and among religious confessions. If religious symbols can appear in public places, we invite conflict about whose symbols will appear; if resources can be distributed through religious providers, we invite political conflict over which providers, about whether there is religious favoritism in designating providers, and about what secular conditions the providers must meet in order to be acceptable recipients of public resources; and if endorsements are available, it is hard to resist the temptation to compete for one. Finally, whatever other purposes it may have, politics is at least in part about addressing collective problems that resist address through separate individual efforts. The entanglements associated with religious establishment arguably distract from this pragmatic, problem-solving aspect of politics, and may focus political attention and conflict on arguably less constructive questions, ones less susceptible to human remedy.

ii. endorsement

A fourth establishment issue—and this brings me to my main subject—has been a central theme in the Supreme Court’s establishment decisions for the past twenty-five years, beginning with Justice O’Connor’s concurrence in Lynch v. Donnelly. The case, as I mentioned earlier, was about a crèche and larger holiday display in Pawtucket. O’Connor focused on the troubles with an official endorsement of religion. Endorsement was one of two ways—excessive entanglement was the other—by which government could commit the fundamental establishment sin of unacceptably linking political standing to an “adherence to a religion.” What animates the hostility to regulations and policies that can be understood as official endorsements of religion? Justice O’Connor originally presented the idea as a way to understand the animating ideas behind the purpose and effects prongs of the Lemon test. A statute only satisfied this test, thus did not raise establishment concerns, if it had “a secular legislative purpose,” if its “principal or primary effect” was one that “neither advance[d] nor inhibit[ed] religion,” and if it did “not foster an excessive government entanglement with religion.” The intuitive idea behind O’Connor’s endorsement test is that the government may not make a policy that either sends (as a matter of intent) a message of endorsement—of favoring a certain religion or religion over non religion—or that an informed and reasonable citizen (e.g., a reasonable adherent of another faith) will receive as a message of endorsement. Why not? The central normative claim is that it endorsement excludes: thus the Endorsement-Exclusion Thesis. A religious endorsement announces (in intent or in effect) the official position that a certain religion or family of religions or religiosity itself is favored (or disfavored) by the political community, and thus assigns a favored status to those who hold the favored views. This announcement is said to make religion “relevant . . . to status in the political community.” The endorsement sends the message that adherents are insiders, and nonadherents are outsiders. In short, the

Establishment, Exclusion, and Democracy’s Public Reason


Endorsement-Exclusion Thesis is that endorsement itself—by making religion relevant to status in the political community—is directly injuriously exclusionary. Relevant to status of what sort? O’Connor says “status in the political community,” but that statement requires interpretation. Endorsements are not classical political exclusions (not, that is, deprivations of rights to speak, associate, vote, hold office) nor is it clear that the interests of the outvoted are being discounted, and that the ban on endorsements protects, in an equal protection spirit, against the special vulnerability of some group(s). So it does not look like there is an exclusion from the status of citizen. Is the point instead that an endorsement announces that some groups (e.g., Protestants) are sociocultural insiders and others are sociocultural outsiders? Or that endorsements establishing ranks and orders with their privileges, thus more feudal than democratic, with democracy understood as a society of equals? As an alternative, I suggest that endorsement conveys a civic exclusion, but that the exclusion is not from the “space of rights and interests,” or from the dominant culture, but from the space of reasons. A collective decision has been taken which some members reasonably take to be unsupported by any suitable reason at all. Suppose we say that democracy is not simply about ensuring political rights and representing interests, but is also in part about reasoning together, on a common ground that we appropriately expect others to occupy. Suppose we find something compelling in the idea of a public reason of democracy. Then we will think that exclusion from the space of reasons is a kind of civic exclusion, and a matter of democratic concern. I will return to these issues later. The Endorsement-Exclusion Thesis has played an especially prominent role in cases involving public symbols: crèches, crosses, menorahs, benedictions, and huge tablets donated by the Fraternal Order of Eagles displaying the Ten Commandments as part of a promotional scheme for Cecil B. DeMille’s eponymous film. But the significance of these cases transcends the work of an exterior decorator, concerned about the proximity of crèche to Santa’s cheerful helpers. The Endorsement-Exclusion Thesis nicely captures the importance.

iii. some questions

That said, the Endorsement-Exclusion Thesis may prompt hesitation or skepticism. A first concern is about the legally manageability of “endorsement.” This concern is reinforced by the context-dependency of judgments about whether there has been an endorsement (see below, p. 262). Concerns about manipulability are especially serious if you interpret endorsement, as Justice Scalia does, as a matter of adopting a policy that makes a person “feel excluded.” But the endorsement concern is very different, less about feelings than about reasonable objections of informed observers, observers informed by an understanding of a legal text, legislative history, implementation, and relevant constitutional values. The issue is not whether a person is made to feel bad, or even whether it would be reasonable for a person to feel bad, or made to feel excluded.


Political Themes: Conservatism, Justice, and Public Reason

Consider, for example, the 2005 Pennsylvania decision (in Kitzmiller v. Dover Area School District) rejecting a Dover, Pennsylvania school board requirement about teaching Intelligent Design (ID) in high schools. Judge Jones used the endorsement test. And in applying it, he asks us to consider a deeply informed observer, someone who knows something about natural selection, is aware of the previous history of litigation on teaching “scientific creationism,” knows that the leading ID textbook (Of Pandas and People) originally used the phrase “creation science”— which was eventually subject to a global search-and-replace with “Intelligent Design”—and appreciates that phenotypic changes sometimes have nonselectional explanations (e.g., pleiotropy, genetic drift, architectural constraints). We are asked to consider whether such a deeply informed observer (perhaps a deeply informed high school student) would reasonably construe the “evolution is a theory” disclaimer adopted by the Dover School Board as a religious endorsement. We can safely assume that no one in the School District—anyway, no students in the high school—actually met these exacting standards, and thus no one regarded it as an endorsement. Still, the school board may have enacted an exclusionary endorsement even if no one felt excluded, and even if those who did feel excluded did so for unsound reasons. The issue is not about feelings, but about whether it is reasonable to believe that the regulation or policy depends for its adoption or implementation on a particular religious view, in this sense: that that religious view provides the most plausible rationale for adopting the regulation or policy (with its specific contours) in the setting in which it was adopted, and it is therefore hard to see why the regulation or policy would have been adopted had it not been a way to express the religious convictions, including convictions about the special importance of religiosity. The Dover School Board had no reason to do what it did, at the time and in the way in which it made its decisions, except to express a religiously founded opposition to neo-Darwinism. The point is not that the endorsement makes a person feel excluded, but that the person rejects the views that an informed observer (whether opposed or in favor) reasonably supposes to endorsed by enacting the policy, endorsed in that they provided the most plausible rationale for it. Understanding endorsements this way, the large cross is a problem. Sunday closing, probably not. Crèches are less clear. They do, we should remember, symbolize the incarnation, the spirit’s being made flesh, the central event for the Christian faith. But they may be presented in a larger context that is designed to remind of the holiday season, rather than the Incarnation, and the range of convictions in the community. The idea is that an endorsement is understood to exclude in these cases because endorsement is a matter of doing something on the basis of a consideration that some members of the community take to provide no reason at all: those who do not accept the divinity of Christ will see no reason to publicly symbolize it. What about the cases—from Mueller v. Allen to Zelman v. Simmons-Harris (upholding an Ohio school voucher plan)—that appeal to the idea that when a flow of benefits to religious schools (and perhaps other institutions) is mediated by “true private choice,” then the “circuit between government and religion” is cut? Concerns

Establishment, Exclusion, and Democracy’s Public Reason


about public endorsement have no grip because the flow of funds reflects individual decisions, not a collective choice. This appeal to private choice as neutralizer and thus endorsement-nullifier is too quick. To be sure, many of the issues raised by these cases are not best treated as questions of endorsement (or lack of same), but of the kinds of entanglements that are damaging to religion and politics (the second and third concerns noted in my earlier comments on Establishment issues.). Still, judgments about endorsement need at least to pay attention to the substance of a policy and the effects that can reasonably be expected to result, and not to unreflectively assume that the results issue from independent choices and cannot be attributed to a collective purpose. A second objection to the endorsement test is that it rests on a misguided personification. Policies, you might say, are the result of a process, a product or interest group bargaining or of party competition and legislative bargaining. The resulting policies are not endorsements because they are not the decisions of an agent, thus not the expression of the single will that endorsement demands. Think of a bargaining game. The bargainers may all accept the result, even if no one endorses it. To be sure, the objection continues, policy choices have winners and losers. The winners may share religious convictions and make policy based on those convictions. But the losers retain their rights, and can still advance their interests. They might win next time: this year tree and menorah, next year crèche and crescent. Politics is a competition among a plurality of persons and groups. You have grounds for complaint that you are excluded if you are denied the right to participate, or if your interests count for nothing. But no basis for complaint about exclusionary endorsements: no basis because no endorsement; no endorsement because no agent. No agent, no endorsement; no endorsement, no exclusion. Whatever our larger view about permissible personification, this objection is misplaced here. To see why, consider again the distinction between three kinds of inclusion (and correspondingly three kinds of exclusion): inclusion in the space of rights, the space of interests, and the space of reasons. When a policy is made that makes sense to adopt only on the basis of considerations that have no weight for some of the equal members of a democratic society, for which they reasonably judge there to be no reason at all, then the policy is in a perfectly good sense exclusionary, even if their rights and interests are taken into account in the process. They are not simply outvoted. Instead, the process produces a result that is justified only if there is a reason for it, and we can always say that the process yields an endorsement of what justifies the result. And if they judge there to be no such reason (though they are aware of course that others think there is a prima facie reason), then, for them, the regulation or policy to lack any rationale: it is adopted on the basis of considerations that they judge not to be reasons at all. That is the essential point. No extravagant assumptions about permissible personifications in a pluralist democracy are needed. This point about exclusion from the space of reasons brings me to the third (and most fundamental) objection: that the concern about endorsement and resulting exclusion expressed in the Endorsement-Exclusion Thesis is vastly too powerful because it suggests that any policy conveys a kind of exclusion. Every policy, I have


Political Themes: Conservatism, Justice, and Public Reason

said, conveys an endorsement of the beliefs, goals, and principles that serve to justify the policy (or at least of some set of justifying beliefs, goals, and principles, assuming that there is more than one). But every collective decision also has some opponents. If every endorsement conveys an exclusionary message about the civic status of those who reject the justifying beliefs, goals, and principles, if every endorsement is an exclusion of those who reject what has been endorsed, then every policy disagreement is transformed into an excluding insult. Democracy is all denigration, all the time.

iv. three cases

How might we resist this unhappy conclusion? To focus the discussion, I will distinguish three (highly stylized) cases of disagreement, which I will call pure policy disagreement, political-philosophical disagreement, and religious disagreement. In the case of pure policy disagreement, I agree with the goal of a policy—the goal, for example, of increasing employment—but disagree with the means, which relies more heavily than I think it should on spending rather than tax reductions. So we have a disagreement on factual beliefs. Assume now that the government endorses (through its policies) the beliefs of the spending party. It seems extravagant to say that those who reject those beliefs that animate and justify the policy that gets adopted are excluded, that they are treated as outsiders. The disagreement is profoundly consequential, but all parties endorse the animating goals. Moreover, they can see how factual beliefs about how to advance those goals are likely to differ among reasonable people. Think of adopting the policy as the conclusion of a public argument, rather than the product of a political process. There is no basic premise in the argument—either about the goals or the means for achieving them—that those who reject the policy fail to understand or see how someone might endorse, or think is irrelevant to public policy. Moreover, they see that some policy is needed. Even though opponents may reject the factual beliefs about how to increase employment—perhaps they do not share the priors of the proponents—the premises are all within reach, even for those who reject them. In this first case, we have both endorsement and disagreement, but no exclusion. The second case is political-philosophical disagreement. Consider a classical-liberal/ egalitarian-liberal disagreement about, for example, distributive fairness, issuing in a disagreement about whether tax policy should be used to benefit low-income earners. Suppose the party in power acknowledges the philosophical disagreement, goes forward with the policy justified by a more egalitarian liberalism, and reminds opponents that “we just had an election.” Adopting the policy, we endorse a philosophy. Are those who reject the philosophy sent a message of civic exclusion? Of course they retain their political rights and can win the next election. But that is also true of the people who receive the allegedly exclusionary religious message. Instead, we need to say something like what was said in the policy disagreement case. Assume for the sake of argument that we have a population who all accept the basics of democracy, with its idea of members as free and equal, with independent judgment and owed appropriate treatment, with due regard for their rights and interests. In the

Establishment, Exclusion, and Democracy’s Public Reason


case of political-philosophical disagreement, all the premises that serve to justify the policy (thus arguably endorsed in adopting it) are within reach for all who meet this condition. Those who reject the conclusions (perhaps because they endorse a more classical liberalism) think that the best understanding of members as equals, with different ideas of what is right, competing interests, including conflicting moral and religious ideas, is not expressed in the winning philosophy, the more egalitarian form of liberalism. Nevertheless, that latter philosophy represents an effort to express in a political outlook those democratic ideas; the underlying ideas are fully intelligible; moreover, those ideas are relevant to public policy. Those who reject the philosophy but accept the basic elements of democracy, can see that the rationale for the view that they reject rests its appeal on values and forms of argument that they accept. (The same of course can be said if the classical liberal view wins.) In the third case, the endorsed ideas are religious. Why is there now an endorsement-exclusion issue? In each of the three stylized cases, we have the same conjunction of official endorsement and disagreement or rejection by some. What is special in this third case. What, beyond the conjunction of endorsement and disagreement, implies exclusion? Before getting to the answer, I need to say something more about religion.

v. religion

In his Elementary Forms of the Religious Life, Durkheim says “A religion is a unified system of beliefs and practices relative to sacred things, that is to say things set apart and forbidden—beliefs and practices which unite into a single moral community called a Church, all those who adhere to them.” The main elements of this illuminating conception are its emphasis on the collective aspect of religion, the importance it assigns to sacredness, and its nontheism. Consider these in turn. Collective. Durkheim emphasizes the collective aspect of religion, associated with a church. But he also offers a very minimal, generic characterization of a church, as a collection of people who adhere to a shared set of beliefs and practices relative to the sacred. Of course, in some cases (e.g., Roman Catholicism, and other hierarchical churches, including Eastern and Oriental Orthodox, as well as semi-hierarchical churches such as Assemblies of God and Presbyterianism) the church has a much more elaborate collective structure than a loose knit community of adherents. Moreover, that structure may (consider again the case of Catholicism or Episcopalianism) not be a matter simply of history and tradition, but instead reflect central religious beliefs and practices, which define the church as a sacramental union and “corpus mysticum.” Sacred. An idea of the sacred is also fundamental, and there are four main elements. First, sacredness is a special kind of importance. Sacred importance is intrinsic, not conferred, at least not humanly conferred, not a product of desire, or choice, or social convention. It is a feature to be recognized and appreciated for its significance. Moreover, the intrinsic importance is very great, exalted, thus not to be traded off, negotiated over, or trespassed upon. That special importance might be expressed in an idea of


Political Themes: Conservatism, Justice, and Public Reason

supreme, overriding obligations (e.g., when the religion endorses a theistic theory of sacredness) with a lawgiving deity. But a notion of obligation need not be central. In general terms, then, to regard something as sacred is to suppose that there are reasons for treating it as having an exalted, intrinsic importance: reasons with an especially commanding importance. Second, sacredness pertains to both “beliefs and practices,” and practices and beliefs relative to the sacred are often intimately related. So, for example, suppose I believe that some object (person, place, practice) is sacred, thus of intrinsic and exalted importance. In this case, the belief itself tells me that it needs to be treated in special ways, and that the relevant practices are important, required, necessary. The sense of importance is not, as in the case for example of ethno-cultural identities, an add-on attitude; importance is ingredient in the content of the convictions. The practice is not simply habitually or conventionally associated with the religion, but designated by the convictions as of special importance. Consider, for example, a mandatory time of rest, covering of the head or face, posture of supplication, consumption of wine, or wafer, or peyote, or manner of killing and preparing food. Third, we have not simply discrete sacred things but a “unified system,” a more or less coherent, integrated view, with wide (perhaps encompassing) implications over a life. Thus a religious view is comprehensive. This idea plays an essential role in Rawls’s characterization of religions—along with some secular moral views, including the liberalisms of Kant and Mill, as well as an Aristotelian perfectionism—as comprehensive doctrines: comprehensive in covering life’s course (not confined to the political), doctrinal in their high degree of internal integration. It is also central to the view, first suggested by Judge Arlin Adams of the Third Circuit in Malnak v. Yogi, subsequently adopted in four circuits, that one of the main features of religion is its claim to a “comprehensive truth.” This focus on comprehensiveness has, in lower courts, served to distinguish religious outlooks from veganism (in California), as well as from the Church of Marijuana, the “New Age concepts” expressed in San Jose’s display of Quetzalcoatl (the Aztec Plumed Serpent), and the revolutionary naturalism of Frank Africa’s MOVE. Joining this dimension of comprehensiveness to the two previous points, we have a practical orientation, governed by reasons for belief and conduct that are of great importance, capacious in its coverage, integrated in its content. And fourth, the importance of a conception of the sacred is perhaps that it answers to a concern that human life be redeemed, redeemed by assigning life’s common activities—work and family, birth and death, love and suffering, friendship and intellectual reflection—a larger significance, a significance larger than their mundane social functions and their responsiveness to our desires, needs, and embodiment. You might say sanctification rather than redemption, but that characterization is too concrete, too parochial for the purpose at hand, because sanctification suggests a conception of holiness that is not essential to the idea of the sacred. Religious ideas surely have other roles as well: providing a theodicy (Weber), a proto-science (Quine), a vehicle for imaginatively expressing ethical demands (Spinoza), or intuitions of unity (Schleiermacher), but the idea of the sacred arguably has this more constructively redemptive importance as well.

Establishment, Exclusion, and Democracy’s Public Reason


Nontheistic. Unlike Locke, or Madison, or the Supreme Court prior to its 1965 Seeger decision on conscientious objection, Durkheim’s characterization was not focused on Protestant Christianity, but on an effort to understand, as he put it, “primitive and simple religion.” Although theism might be part of a theory about the sacred, it need not be, and it is sacredness rather than the underlying theory that accounts for it that is important. Finally, in characterizing religious convictions, I have emphasized their content, not their psychological or motivational strength. To be sure, their content—focused on the sacred—defines them as important. They announce their own importance both in the kinds of reasons they provide and in the surrounding doctrine about the roots of sacredness. But importance is not the same as psychological or motivational strength, as measured by a desire’s intensity or a belief’s resistance to change. This emphasis on content—the weightiness of reasons associated with the sacred, and their comprehensiveness—follows the line of thought in Malnak v. Yogi and the cases that have drawn on it. The account of religion in the Seeger case, on conscientious objection, is not so clear, but can be interpreted along the lines I am suggesting. There, the Supreme Court explained religion in terms of its role or “place” in the life of an adherent. More precisely, the issue was how to understand the religious exemption from military service under the Universal Military Training and Service Act, which defined a belief as religious just in case it was “in a relation to a Supreme Being involving duties superior to those arising from any human relation.” In part to avoid conflicts with the free exercise clause, the Court said: We believe that under this construction, the test of belief “in a relation to a Supreme Being” is whether a given belief that is sincere and meaningful occupies a place in the life of its possessor parallel to that filled by the orthodox belief in God of one who clearly qualifies for the exemption. Where such beliefs have parallel positions in the lives of their respective holders we cannot say that one is “in a relation to a Supreme Being” and the other is not. I am taking that “parallel position” or “place” not to be psychological—not about the strength of convictions or their persistence—but deliberative. The place or position (as suggested by the phrase “duties superior to those arising from any human relation”) is to provide a comprehensive and fundamental guide to conduct, a guide whose importance, as understood by the adherent, is explained by the content of the convictions. vi. democracy

Returning now to the central question: why does the conjunction of endorsement and disagreement imply a troubling exclusion in the case of religion, but not in the cases of policy disagreement and political-philosophical disagreement? I have two answers, drawing on different connections with democracy.


Political Themes: Conservatism, Justice, and Public Reason

One answer concerns a usurpation of individual judgment associated with endorsement. Suppose the state endorses a religious view. It then takes a position on an outlook of comprehensive scope that comprises in a relatively integrated way ideas about the sacred and about conduct and attitudes toward what has sacred importance: thus overarching ideas that give a basic orientation to life. One concern about endorsement, about taking a public position—in effect announcing a position as our position on these issues—is that these matters, comprehensive and fundamental, are the proper domain of individual reflection and judgment, not matters on which collective judgment is appropriate, on which we can be said to have a position. Taking a collective position on the right way to live, on the proper conception of the sacred, or that a conception of the sacred is itself important, usurps the judgment of equal members of a democratic society. More precisely, it either usurps their judgment or excludes them. It usurps when those who do not share the officially endorsed view are still understand as members of the people, with equal standing. In this case, a collective judgment is made in their name, as one of us. But their own fundamental convictions may well require that they reject any judgment on fundamentals as made or spoken for them, even if they agree with the judgment. The collective judgment excludes if the people is understood to include only those who can accept the collective judgment as their own. In the policy and political-philosophy cases, in contrast, the endorsed judgments are not and are not presented as parts of comparably comprehensive and fundamental outlooks. This concern about usurpation may prompt two kinds of objections. First, the claim that endorsements usurp may seem a little overwrought. But the account of religion provides grounds for concern. Consider the case of public symbols. Symbols are not isolated images, but crystallizations of larger outlooks. Of course, sometimes a symbol is just a symbol. Think, however, of prohibitions on graven images, or Hadith-based condemnations of depictions of the Prophet. Those prohibitions reflect troubles about idolatry, not my focus here. But they suffice as reminder about the potential importance of symbols. Assuming comprehensiveness, a religious endorsement embraces an outlook that covers, in an integrated way, more or less all of life. Moreover, its coverage expresses a view whose practical requirements—whether basic obligations or weighty demands of other kinds—reflect a distinctive conception of the sacred, with its associated intrinsic and special importance. A symbol of the mysterious union of the incarnation, for example, points to a vastly larger outlook. That outlook addresses the relationship of human and divine, human redemption, the vastness of our failing (sick, fallen, dead, lost, closed, captive, prisoners, slaves, St. Gregory said), the comparable vastness of the compensatory sacrifice, and the prospects of a reconciliation (and emphasizing symbols of the incarnation, rather than the arguably less hopeful symbols of the theologia crucis, is itself a matter for concern). A second objection is that the usurpation objection may appear to rest on a secular morality of autonomy, liberalism as a philosophy of life, the idea that the unreflective life is not worth leading, and that the only proper authority over conduct is a person’s own judgment, reflectively considered. But no such general thesis about the importance of autonomous choice underlies the complaint about usurpation. Instead, the

Establishment, Exclusion, and Democracy’s Public Reason


idea of usurpation of judgment is more concrete and political. It is more concrete in that it draws on the features of religious convictions I have noted earlier, and more political in that it draws on two related features of the conception of members of a democratic society: that we are political equals and independent judges of political values. The status of equal and independent judge means that collective judgment must not extend to the comprehensive and fundamental values that constitute a religious outlook. Not because autonomy takes precedence. The concern flows instead from special nature of religious convictions, which are assumed to have the distinctive importance and practical comprehensiveness associated with the sacred. Given that special content, the plurality of conflicting religious ideas cannot be taken seriously as the basic and comprehensive convictions of the equal members of a democracy—which means that they cannot be taken as the adherents themselves must take them—if they are submitted to collective judgment. The second answer is about reasons, in particular about democracy’s public reason. If usurpation raises issues of equality and independent judgment, this second answer draws on the further idea that collective decisions in a society of equals are to be supported by reasons. In the case of the religious ideas, however, there is no basis for the claims about reasons that we saw operating in the cases of policy disagreement and political-philosophical disagreement. First, there is no case for the claim that people who endorse the values associated with democracy—an idea of equality, of equal regard for the judgments and interests of members—will find the religious ideas of others intelligible. As a personal matter, I find that I simply am not sure that I understand religious claims well enough to say that I disagree with them. Here, I mean to sweep widely, to include claims about creation, predestination, incarnation, transubstantiation, reincarnation, special relationships to particular groups, dictation of sacred texts in human languages, or angels in America, in particular upstate New York America. Treated literally, they seem hard to believe. Responding to Jaspers, Bultmann says “He is as convinced as I am that a corpse cannot come back to life or rise from the grave, that there are no demons and no magic causality.” We can instead interpret them, in the traditions of Spinoza, Kant, Schleiermacher, Feuerbach, and the social gospel tradition as imaginatively vivid expressions of ethical, anthropological, or existential ideas. With Spinoza, for example, we can say that they give an imaginative gloss to a cognitive-moral core, glossing the core with local color, a color designed locally to make the moral considerations more intelligible and attractive, attractive and intelligible because more imaginatively vivid and motivationally immediate for particular cultures and communities. Though tempting, this cognition-plus-imaginative-gloss strategy does not do justice to the sometimes-powerful resistance to such “demythologizing” reinterpretation, a resistance shared by members of different faiths, as well as those who have fallen from faith. When Haines, the patronizingly imperial Oxonian at the Martello Tower, asks Stephen Daedalus “You are not a believer, are you? I mean a believer in the narrow sense of the word. Creation from nothing and miracles and a personal God,” Stephen Daedalus responds: “There’s only one sense of the word, it seems to me.” A theological literalism expressed as well in Daedalus’s subsequent comment: “ ‘You behold in me,’ Stephen said with grim displeasure, ‘a horrible example of free thought.’ ”


Political Themes: Conservatism, Justice, and Public Reason

These reflections on difficulties of understanding are not founded on a theory of cognitive significance, nor are they meant as criticism: you might even think of them as a confession. Indeed, adherents of some faiths treat their own willingness to stand fast in the face of incomprehension—or the merely negative comprehension of apophatic theology—as expressing a proper humility in the face of a grandeur greater than which none can be conceived. Instead, my aim is simply to describe the fathomless mutual incomprehension that belongs to this territory, an incomprehension that does nothing to undercut the case for protecting comprehensive and commanding convictions. As Chief Justice Burger said, writing for the Court in Thomas v. Review Board (upholding compelled exemptions), “religious beliefs need not be acceptable, logical, consistent, or comprehensible to others in order to merit First Amendment protection.” And second, suppose there is some mutual comprehension. Suppose committed adherents of different outlooks think they understand one another, enough at any rate to know that they genuinely disagree, not simply verbally and imaginatively. Still, they will typically not see a case that can be made for those ideas—or for resolving differences between and among competing views—from a common ground of conviction that they can reasonably expect, in the way that they can see how someone might reasonably endorse a more or less egalitarian view from a common ground of political values. Consider disagreements, all of which have had large ethical-practical resonance, about the nature of the trinity (filioque procedit or per Filium procedit), the nature of Christ (dyophysite or monophysite), solafideism, pre- and posttribulation millennialism, original sin and its Pelagian and semi-Pelagian opponents, about the theologies of incarnation and cross, how the sacred redeems, the possibility of an historically unfolding revelation of a timeless message and a corresponding “development in doctrine” (aggiornamento), or the appropriateness of submission as a basic religious attitude. Let us say that members of a democracy are politically reasonable when they accept democratic values and practices (an idea of persons as equals, with powers of independent judgment), and also acknowledge the difficulties of arriving at agreement on fundamentals among people who accept democratic values and who exercise their judgment conscientiously. Thus they acknowledge what Rawls has called the “burdens of judgment,” and reject the idea that disagreement on comprehensive fundamentals is a basis for reproach. So politically reasonable people will have competing religious and secular moral views, and remain in a state of fundamental disagreement—perhaps mutual incomprehension—on what they regard as fundamentals, their respective comprehensive conceptions and practices regarding the sacred (if they have such practices and conceptions at all). Political reasonableness, thus understood, is a normative idea that belongs to political life, in particular, the political life of equals in a democracy. The thesis that politically reasonable people disagree is not a claim about the limits of reason, according to some philosophical theory of reason, whether empiricism or a natural law rationalism. It is about the demands it is appropriate for us to make on one another as equal members, attentive to the complexities of reflection on fundamentals.

Establishment, Exclusion, and Democracy’s Public Reason


Suppose now that a policy can reasonably be understood as a religious endorsement. Then we do have a kind of exclusion, an exclusion that operates in the space of reasons. Though all retain their rights and can advance their interests, some see no reasons at all for the policy. It treats some members as outsiders: they reject the considerations that justify the policy, though they are not unreasonable (not in the sense defined) for rejecting them. Endorsement excludes because it conflicts with the ideal of public reason, which requires that political justification proceed on a shared terrain of argument—shared by people who regard one another as free and equal, despite irreconcilable doctrinal disagreements and disagreements about justice. This concern may seem to overlook an important symmetry: if you endorse, you exclude; if you reject the endorsements, you exclude. Thus, in the public symbols arena, either way, crèche or no crèche, rabbi or no rabbi at the graduation, we have an endorsement, an endorsement either of a religion or a secularist rejection of religion, thus a corresponding exclusion. It is politics, with winners and losers. Another neutralist bluff, another unmasking. The hard work of demystification is never done. This symmetry claim is doubly misguided. One difficulty is that it fails to understand that there may well be grounds within some religious views themselves for restricting public justification to reasons that are on common ground, thus for resisting endorsements. As a logical matter, the point is undeniable: starting from religious axioms, you may be able to derive, as a theorem, the proposition that, when it comes to political relations and questions, shared standards of reasoning are typically called for, for the purposes of justification. But let us not simply stand on logical ceremony. Catholic natural law theory, after Vatican II’s Declaration of Religious Freedom, uses an axiom of human dignity and a diagnosis of modern circumstances—a pluralism that newly reveals the scope of human possibility—as a basis for treating the political domain as a space suited to common reason. The human dignity associated with creation in God’s image demands no less, or so it seemed after a profound and profoundly controversial “development in doctrine” within a message of transcendent validity. And you can, arguably, tell a similar story from within Islam, founded on the principle of human vicegerency, an opposition to coercion in matters of religion, and a critique of idolatry. An additional difficulty is suggested by reflection on secular moral liberals who believe that human beings have a dignity beyond all price, that the unchosen life offends human dignity, that cognitive reflectiveness condemns conventional religiosity, and that cognitive and moral maturity are at odds with the submission to text and authority associated with fundamentalism and traditionalism. These views should not be expressed in public symbols either: we should not replace the rabbi with a high school version of Kant’s idea that moral rightness is about self-legislation, not heteronomy, even if the lawgiving other is God. The concern about religious endorsement is not an invitation to a comparably sectarian secularism. But we should also not forget that secular moral liberals do not say in political justification nor offer in officially endorsed symbolism, nor ought they to say or offer, everything that take to be true about religiosity. To be sure, theistic salvation religions are distinctive. Still, democracy’s public reason requires a mutual self-restraint, a self-restraint that is more mutual and more nearly symmetrical than we commonly appreciate.


Political Themes: Conservatism, Justice, and Public Reason vii. conclusion

With exclusionary endorsements, we have usurpations without support in democracy’s public reason. Now ordinarily, when collective decisions are made that meet this description, we have force without reason. But at least in some cases of endorsement (e.g., about public symbols) there is no force. Because there is no coercion, and perhaps no distributive effect, the lack of support in acceptable reasons can seem inconsequential. In my discussion of usurpation, I raised a concern about seeming overwrought. That concern generalizes: being outsider to an exclusionary endorsement does not exactly rise to the level of being burned at the stake, thrown in prison, or even taxed to support some “great and abominable church.” But an exclusionary endorsement is objectionable, and the objection is serious. It is about the conditions of democracy, as a political association of equals in which justification ought to be inclusive in the space of reasons, not simply in the space of rights and interests. This point about democracy generalizes, but may have particular force in some democracies. Here, we do well to bear in mind Lincoln’s observation about the United States. This country, he observed, was not born from the blood of an ethnos, but conceived in an abstract idea. And it was dedicated not to a tradition, but to a comparably abstract proposition. As Lincoln suggests, its fragile unity is philosophical, not constitutional: four score and seven years ago, not three score and sixteen. Moreover, Lincoln’s concern about fragility and the possibility of enduring extended from this nation to “any nation” conceived in an idea and dedicated to a proposition. In a democracy rooted in principles, a concern about inclusion in the space of reasons is not marginal. It is foundational. notes 1. I am grateful to Nannerl Keohane, Larry Kramer, Martha Nussbaum, Thomas Nagel, Eric Osborne, Philip Pettit, Richard Pildes, Eric Posner, Samuel Scheffler, Seana Shiffrin, Jeffrey Stout, and Kathleen Sullivan for instructive comments on earlier drafts, and to Osborne for wonderful research assistance. I also thank Sullivan and the students in our Religion and the Constitution seminar (Winter 2009) for helping me to understand the religion clauses and appreciate their complexities, the Berkeley-Stanford Political Theory group for very helpful suggestions on a preliminary sketch of the ideas, and Josiah Ober for forcing me to rethink my unreflectively theistic, obligation-centered conception of religious convictions and their importance. I presented earlier versions of this essay as the Ralph Gregory Eliot Lecture at Yale University Law School, to workshops at University of Chicago Law School and New York University Law School, as the James A. Moffett ’29 Lecture in Ethics at Princeton University, and to the September Group. I am grateful for all the advice I received on all these occasions. 2. I first appreciated the force of this point when I read Scanlon’s “Freedom and Expression and Categories of Expression” (“Categories”). Although Scanlon’s 1972 article on the subject, “A Theory of Freedom of Expression,” has been much more widely discussed, the view in “Categories” presents a much more plausible because much richer structure. Scanlon’s own intellectual development, in particular the evolution of his distinctive form of contractualism, had much to do with the effort to develop a more compelling theory of free expression. For the essays on free expression, as well as a brief intellectual autobiography, see T. M. Scanlon, The

Establishment, Exclusion, and Democracy’s Public Reason


Difficulty of Tolerance (Cambridge: Cambridge University Press, 2003), Introduction, and chaps. 1 and 5. One of the broad lessons that I took away from thinking about “Categories,” and one of the uncountably many things I have learned from Scanlon, is that philosophical reflection on basic liberties should be attentive to the complexities that emerge when legal systems aim to protect those liberties. Although the argument about religious establishment that I present in this essay focuses on just one strand expressed in the ban on religious establishments, I have followed the earlier lesson that I learned from Scanlon and thus have been closely attentive to legal-constitutional terrain in thinking about the problem. 3. Alexander Meiklejohn, Free Speech and its Relation to Self-Government (New York: Harper & Brothers, 1948), 26. 4. Ibid., 25, 105. 5. See also, among many others, John Hart Ely, Democracy and Distrust (Cambridge, Mass.: Harvard University Press, 1980); Cass Sunstein, Democracy and the Problem of Free Speech (New York: Free Press, 1995); Owen Fiss, The Irony of Free Speech (Cambridge, Mass.: Harvard University Press, 1996); Stephen Breyer, Active Liberty (New York: Knopf, 2005), 39–55. For discussion of Meiklejohn’s impact on the interpretation of the speech clause, see Anthony Lewis, Make No Law (New York: Random House, 1991), 154–55. My own views about free expression are not entirely Meikeljohnian. In “Freedom of Expression,” I presented an account of freedom of expression on which stringent protections are not confined to political speech, understood as inputs to political deliberation. See Philosophy, Politics, Democracy (Cambridge, Mass.: Harvard University Press, 2009), ch. 4. In chapter 7, “Democracy and Liberty,” I explained how this expansive conception of free speech protections can itself be understood in terms of an ideal of deliberative democracy, with its requirements of reason-giving among equals under conditions of reasonable pluralism. 6. For the general line of argument, see Joshua Cohen, “Democracy and Liberty,” in Philosophy, Politics, Democracy, ch. 7. 7. Lynch v. Donnelly, 465 U.S. 668 (1984), at 692. 8. Ronald Dworkin expresses a concern about usurpation of judgment in his discussion of religious liberty in Is Democracy Possible Here? (Princeton, N.J.: Princeton University Press, 2006), ch. 3, esp. 71. His “ethical individualism” ties the concern about usurpation to his account of personal responsibility, one of the essential elements of human dignity. In my discussion here, I try to avoid relying on such a foundational ethical conception. See below, 23–24. 9. For an interpretation of the religion clauses that emphasizes equal protection, antidiscrimination themes, see Christopher L. Eisgruber and Lawrence G. Sager, Religious Freedom and the Constitution (Cambridge, Mass.: Harvard University Press, 2007). The account of religious liberty (especially of exemptions) in Brian Barry’s Culture and Equality: An Egalitarian Critique of Multiculturalism (Cambridge, Mass.: Harvard University Press, 2002) is substantively similar in spirit, though Barry’s treatment has an unfortunately dismissive tone, and is not specifically about the U.S. Constitution. 10. For a helpful account of the landscape of Establishment Clause values, see Kent Greenawalt, Religion and the Constitution, Vol. II: Establishment and Fairness (Princeton, N.J.: Princeton University Press, 2008), esp. chaps. 1 and 10. 11. John Locke, A Letter Concerning Toleration, in John Locke, Two Treatises of Government and A Letter Concerning Toleration, ed. Ian Shapiro (New Haven, Conn.: Yale University Press, 2003); James Madison, “Memorial and Remonstrance,” available at founders/documents/amendI_religions43.html. 12. Lee v. Weisman, 505 U.S. 577 (1992).


Political Themes: Conservatism, Justice, and Public Reason

13. Ibid. 14. Thus the Court’s statement in Lee v. Weisman: “The State’s role did not end with the decision to include a prayer and with the choice of a clergyman. Principal Lee provided Rabbi Gutterman with a copy of the ‘Guidelines for Civic Occasions,’ and advised him that his prayers should be nonsectarian. Through these means the principal directed and controlled the content of the prayers. Even if the only sanction for ignoring the instructions were that the rabbi would not be invited back, we think no religious representative who valued his or her continued reputation and effectiveness in the community would incur the State’s displeasure in this regard. It is a cornerstone principle of our Establishment Clause jurisprudence that ‘it is no part of the business of government to compose official prayers for any group of the American people to recite as a part of a religious program carried on by government,’ Engel v. Vitale 370 US 421, 425 (1962), and that is what the school officials attempted to do.” See Ibid., 588. 15. Letter Concerning Toleration, 231. 16. The Supreme Court condemned Lord Eldon’s Rule in Watson v. Jones, 80 U.S. 679 (1871). 17. See Jürgen Habermas, “Pre-political Foundations of the Democratic Constitutional State,” in Habermas and Joseph Ratzinger, Dialectics of Secularization, esp. 40–52 (San Francisco: Ignatius Press, 2007). 18. This concern to avoid “religiously based social conflict” is the central theme of Justice Breyer’s discussion of the dangers of Establishment in his forcefully worded dissent in Zelman v. Simmons-Harris, 536 U.S. 639 (2002), 717–29. 19. Lynch v. Donnelly, 465 U.S. 668 (1984). 20. Lemon v. Kurtzman, 403 U.S. 602, 612–13 (1971). 21. Eisgruber and Sager, Religious Freedom and the Constitution. 22. A suggestion that Martha Nussbaum made in correspondence. 23. See the discussion of public symbols cases in Kathleen Sullivan and Gerald Gunther, First Amendment Law, 3rd ed. (New York: Foundation Press, 2007), 571–97. 24. Tammy Kitzmiller et al. v. Dover Area School District et al., (400 F. Supp. 2d 707). 25. Percival Davis and Dean H. Kenyon, Of Pandas and People: The Central Question of Biological Origins, second edition (Dallas: Haughton Publishing, 1993). 26. Zelman v. Simmons-Harris, 536 U.S. 639 (2002). 27. Ibid., 652. 28. Emile Durkheim, The Elementary Forms of the Religious Life, trans. Joseph Ward Swain (New York: Free Press, 1965 [1915]), 62. 29. Charles Taylor is critical of William James for offering an excessively Protestant, individualist description of religious experience, and thus for downplaying a constitutively collective aspect of religious life. See Varieties of Religion Today (Cambridge, Mass.: Harvard University Press, 2002), 22–26. 30. John Rawls, Political Liberalism (New York: Columbia University Press, 1996), 12–13, 37, 78. 31. 592F.2d 197 (3d Cir. 1979) (per curiam) at 209. 32. Friedman v. Southern Cal. Permanente Medical Group, (2002) 102 Cal.App.4th 39 [125 Cal. Rptr.2d 663]; United States v. Meyers, 95 F.3d 1475 (10th Cir. 1996); Alvarado v. City of San Jose, 94 F.3d 1223 (9th Cir. 1996); Frank Africa v. Commonwealth of Pennsylvania, 662 F.2d 1025 (3rd Cir 1981). 33. This element of the conception of religion has no echo in Durkheim’s view. I am indebted to Jeffrey Stout for emphasizing this point. 34. United States v. Seeger, 380 U.S. 163 (1965). 35. Durkheim, Elementary Forms, 37. See Ibid., 44–50 on the need to avoid including an idea of “divinity.”

Establishment, Exclusion, and Democracy’s Public Reason


36. United States v. Seeger, 380 U.S. 163, 164 (1965). 37. See Catechism of the Catholic Church (New York: Galilee Trade, 1995), 457. 38. Karl Jaspers and Rudolf Bultmann, Myth and Christianity: An Inquiry into the Possibility of Religion without Myth, trans. Norbert Guterman (New York: The Noonday Press, 1958), 60–61. 39. Benedict de Spinoza, Theological-Political Treatise, trans. Michael Silverthorne and Jonathan Israel (Cambridge: Cambridge University Press, 2007), 9, chaps. 4–6. 40. James Joyce, Ulysses (New York: Random House, 1961), 19–20. 41. Thomas v. Review Board, 450 US 707 (1981), 714. 42. John Rawls, Political Liberalism, 54–58. 43. On the importance of “development in doctrine” to the ideas adopted at Vatican II, particularly for reinterpreting the value of toleration, see John O’Malley, What Happened at Vatican II? (Cambridge, Mass.: Harvard University Press, 2008). 44. For a sketch, see my “Minimalism about Human Rights: The Most We Can Hope For?” Journal of Political Philosophy 12(2) (2004): 190–213. 45. The phrase comes from the Book of Mormon, 1 Nephi 13–14. The precise referent (Judaism? Roman Catholicism? Hellenized Christianity?) is a matter of interpretive disagreement. For discussion, see Stephen E. Robinson, “Nephi’s ‘Great and Abominable Church,’ ” Journal of Book of Mormon Studies 7(1) (1998): 32–39.

12 The Significance of Distribution Aaron James

in matters of distributive justice, we take it to be important how benefits and burdens are distributed among different people. But what, precisely, is important about this? In particular, what, from the point of view of justice, is ultimately at stake in what distributions come about? T. M. Scanlon has been coy about what his moral contractualism might imply for justice. Yet the theory bears on this question of stakes. The significance of distribution then lies not in what distributional patterns come about per se, but in independently valuable relations of recognition. In recent years, many egalitarians (e.g., many luck egalitarians) have proceeded as though a distribution (of goods, resources, opportunities, capabilities, or welfare) can be just (or fair) by its very nature, in and of itself. The basic aim of the theory of distributive justice, on this now standard approach, is to say what this intrinsically just distributional pattern is (equality? priority for the worse off ? everyone having enough? something else?). Scanlon’s theory implies that this cannot be right: a distribution, taken as such, cannot be owed, and so cannot be justice. Or at least this follows given the platitude, due to Aristotle, that justice is, by nature, giving each his or her due. The platitude tells us that to distribute justly is simply to give to each individual what he or she is due or owed, as determined by an independent conception of what this is. According to Scanlon’s independent conception of “what we owe to each other,” no individual can be owed a distribution across persons, as such. We are at most each owed our respective shares—only what we can reasonably ask for on our own behalf. Any reason to favor one distribution of shares over another must flow, not from the intrinsic or impersonal value of a pattern, but from the various reasons—what Scanlon calls “personal reasons”—that different individuals have in their respective positions or standpoints. 276

The Significance of Distribution


Few if any defend luck egalitarianism in thoroughgoing personal terms. As I will explain, a fully personal argument cannot, in any case, be straightforward: it must be grounded in specific contexts of justification, in light of their varying personal reasons, and admit of generalization only as such contexts allow. Luck egalitarianism is then difficult to defend, in its usual form, as a more or less general answer to the question of what distributive justice involves. If Scanlon’s theory implies that distributions, as such, are not what ultimately matters, it also tells us what is finally at stake: the significance of distribution depends on independently valuable relations among people. We are to treat others as they are owed, by acting only in ways we could justify to them, because this sustains a valuable “relation of mutual recognition” with them (W, 162). If distributive justice is indeed an issue of what we owe to others, as Aristotle’s platitude suggests, then it, too, is relational: what is ultimately at stake in how benefits or burdens are distributed across persons is whether or not we, the distributing agent, stand in a relation of recognition with each of the different individuals involved. We see the intended priority of relations over outcomes, in an intuitive way, in the following two sets of (Strawson-inspired) examples—one set from interpersonal morality, one set from social justice. In both of two interpersonal morality cases, hot coffee spills from my cup into your lap, leaving you burned and inconvenienced. In a first case, I have not been particularly careful, gesticulating in a way I knew might cause a spill. In the second variation, I take due care, but am bumped from behind; the spill reflects no fault of mine. In both cases the worldly outcome, aside from how I have governed myself, is by stipulation exactly the same. Yet there remains an important difference in my moral relation to you, because of what the outcome does, or does not, reflect about me. In the first case, I have been careless in a way I could not justify to you; I stand, as I will put it, unjustified by or with respect to you. In the second case, because I am not at fault, there is no change in our moral relation—though your pants are no less hot for that reason. Now consider a society (or any organized group) and a similar contrast. Suppose some people have much less than others. In a first variant, the condition of those disadvantaged is a product of common institutions or practices which could be feasibly reformed at a small cost, leaving the disadvantaged much better off. Yet those who collectively sustain and shape the common structures ignore or never find out about this possibility; they have what are, for them, more pressing concerns. In a second version of the case, exactly the same distributional state of affairs is created by a natural disaster, which has no cause in human activity, and which persists despite all possible efforts of prevention and mitigation. Although in both cases the worldly outcome (aside from the way the society has governed itself) is by stipulation exactly the same, there remains an important moral difference in the social relation people bear to one another. In the first version, inaction is unjustifiable to the disadvantaged, and so the common relations stand unjustified by them. In the second version, because the society or group is not at fault—because it displays no failure of self-governance— it stands as justified by the disadvantaged as before disaster struck—even if they are materially no better off for that reason.


Political Themes: Conservatism, Justice, and Public Reason

In each set of cases, the contrasting extenuating circumstances (in which the spill was pushed, and the disadvantage borne of natural causes) highlight how a valuable relation with others is both intuitively independent from mere worldly outcomes and what is of ultimate importance. I take Scanlon’s theory to develop this thought in the following way. Scanlon says the motivational starting point of his theory, the idea of “justifiability to others,” can be characterized in terms of a valuable relation we bear to others—a relation of being in “unity with our fellow creatures,” in Mill’s borrowed phrase. The relation (which Mill mishandled) is implicit in Scanlon’s beautifully ambiguous name for the domain of his concern, “what we owe to each other.” The name refers at once to a relation one person can have with another one person, as well as to a relation which each person can bear to all others. According to what I will call the basic relational orientation of Scanlon’s theory, the relation we bear to a particular other is primary: We bear the moral relation to all others in the same way we bear it to a single individual. We bear the moral relation, not to everyone, but to every one. Scanlon calls the relation we bear to others “recognition,” but without clarifying the exact role of this characterization. Does it merely explicate the moral relation’s distinctive form of value? Or does it have a deeper explanatory role? I suggest the latter. Once we flesh out what it is to “recognize” another, in one’s motivation, we find a general rationale for Scanlon’s restriction to personal reasons. Roughly, if recognition is to put us in the moral relation with each other, taken separately, then it can only traffic in personal reasons. For personal reasons—reasons which refer to the people who have them—motivationally register persons as distinct individuals. Impersonal considerations, by contrast, contribute nothing to this recognitional function, and so are irrelevant, as such, to what we owe to each other. The same goes, then, for justice: we have to keep the grounds of distribution personal. Only distributions grounded in personal reasons recognize people for the distinct individuals they are; only such distributions are fit to establish the relation of unity with our fellow creatures that justice and morality is ultimately all about. In sum, Aristotle’s platitude suggests the following two-premise argument. The first premise is that distributive justice is by nature part of what we owe to each other. The second premise is that distributions cannot, as such, be owed. It follows that distributions, as such, cannot be just. The relational foundation of Scanlon’s theory explains this second premise in positive terms: distributions do not, as such, establish a relation of moral recognition with each person taken separately. And the first premise can be confirmed by examples that show that, as much as outcomes matter in distributive justice, relations are what are ultimately at stake.

what can be owed: individualism

Scanlon’s theory strongly overdetermines the conclusion that distributions, as such, cannot be owed. As I explain presently, this follows from any of several general constraints on owing, which narrow the field of relevant considerations before we apply the well-known contractualist formula and judge what people can “reasonably reject.”

The Significance of Distribution


Scanlon takes the constraints to apply because they are part and parcel of the special moral domain he seeks to describe: To wholly ignore them is simply to change the subject, from the topic of what we owe to each other to something else. The question is then why the domain should be understood in Scanlon’s way. I will suggest the answer lies in the basic relational orientation of his theory. According to Scanlon’s individualism, we are to determine what is owed only by considering individuals taken one by one. This constraint has both formal and substantive elements. The formal element is a general ban on the direct aggregation of benefits or burdens across persons. In considering whether a burden to one person is justifiable, we compare only the benefits or burdens to another one person, and then another one person, and so on, serially, until everyone is considered. What is justifiable can only be a function of the strongest reasons arising from some such position; there is simply no room to directly consider the sum of benefits across persons (or to use any other direct aggregative function). But what reasons are relevant when different positions are being considered? Scanlon’s substantive answer is what might be called his personalism. Grounds for reasonable rejection include personal reasons, which “have to do with the claims and status of individuals in certain positions” (W, 219). Direct appeals to impersonal reasons are disallowed. The two elements of individualism require separate justifications. The formal ban is supposed to give intuitively correct answers where aggregative views such as utilitarianism falter (e.g., Jones the technician, stuck in the World Cup transmitter, cannot be sacrificed for the sake of modest gains to millions [W, 235]). It is also meant to indirectly allow appropriate aggregation, for nonaggregative reasons. None of this entails personalism. For all the formal ban says, equality of distribution might be justified on impersonal but nonaggregative grounds. To be sure, the grounds of objection cannot be simply that such a distribution is intrinsically just or intrinsically owed, because this is precisely what is in question: a reasonable objection cites reasons why a distribution is just or owed. Yet the formal ban does not preclude objections to an unequal distribution on the simple ground that it is a bad thing to have happen from an impersonal point of view. Personalism does rule this out: the only relevant grounds for reasonably rejecting a principle are personal reasons, specifically, personal reasons which ground objections one could raise on one’s own behalf. Some such reasons concern features of distributions. I can object “he has more than me” in a personal key, on my own behalf, so long as the objection refers to the fact that I have too little, that I am stigmatized as a result of having less than others, that I am likely to be controlled or dominated by those who have more, that my self-respect is at stake, that I have helped to create the goods in question, and so on. What is essential is that the consideration adduced is explicitly or implicitly self-referential: it is a reason for me, as a matter of its agentrelative form, but also in some way concerns my interests or status, as a matter of its substantive content. The mooted objection must be self-referential, the idea goes, if it is to express an objection on my own behalf. Personalism is nevertheless consistent with several important forms of relativity. First, the content of a justified principle can be essentially comparative. Rawls’s


Political Themes: Conservatism, Justice, and Public Reason

difference principle, for example, refers to relative positions of different people (the “worst off ” or “least advantaged”), though its grounds are ultimately personal reasons concerned with each social cooperator’s interest in a greater rather than a lesser share of the social product. Second, even at the level of justification, what is owed is always comparative as such in the sense that we can never assess someone’s personal benefits or burdens without looking at how others fare (W, 195–96). Third, impersonal distributional ideals such as equality of distribution can be indirectly significant. One can at most give, on one’s own behalf, personal (implicitly self-referential) reasons to be allowed to promote them. Yet their impersonal value can strengthen one’s personal claim to this (W, 221). Fourth, Scanlon allows objections to principles “simply because they arbitrarily favor the claims of some over the identical claims of others: that is to say, because they are unfair” (W, 216). So, for example, suppose a principle gives people unequal shares, depriving some of “benefits and opportunities they have reason to want.” A less advantaged person can expect a relevant difference between his or her situation and those better off. If there isn’t one—if there are no relevant personal reasons on the side of those who get more—he or she can object of arbitrary treatment, on personal grounds. These forms of relativity do not imply that a distribution as such can be owed. What is owed depends on a judgment of reasonableness, which is simply a judgment of sufficiency: a judgment that certain objections are, by comparison to other relevant objections, sufficient to rule out proposed principles. The class of relevant objections is in turn constrained by personalism: They must cite personal reasons, which, owing to their self-referential nature, express objections on someone’s behalf. A distribution can never be owed, taken simply as a distribution, simply because it can only be owed in virtue of some such personal objection. But why not define the class of “personal reasons” more inclusively? That is, one might allow direct objection to distributional states of affairs, without self-reference, and define the class of objections “on one’s own behalf” to include these personal reasons. When I object that “he has more than me,” then, I object on my own behalf, based on a personal reason I have that disfavors the unequal distribution, simplicter. Might we then say, given such an objection, that a distribution, as such, is owed after all? We cannot say this, however, even if we accepted the proposed inclusive definition. What is owed always depends on the relative force of various personal reasons, arising from various standpoints. The proposed inclusive definition simply puts a further personal reason into the mix. It does not imply that other possible countervailing reasons are insufficient. The suggested inclusive interpretation nevertheless raises the important question of what is at stake in accepting Scanlon’s self-referential characterization of personalism, rather than some such modification. Scanlon mainly offers a bookkeeping argument: the overall project is supposed to vindicate the working self-referential conception by allowing us to capture and perspicuously analyze any and all considerations of moral significance. It is open for argument whether it does that. Even if we grant its adequacy on this count, however, we may also want a more substantial rationale, grounded in the moral phenomena. I suggest one below, derived from the idea of recognition.

The Significance of Distribution


what can be owed: further constraints

If individualism means that distributions, as such, cannot be owed, the following features of Scanlon’s theory also imply this. Capacity-sensitivity: One is only owed something when an agent is capable of regulating his or her conduct as regards that something, given what he or she can understand, plan for, and act on, on the basis of “commonly available information about what people have reason to want” (i.e., “generic reasons”). (W, 204) So no one can be owed a distribution as such. For the existence of a given distribution may or may not be within the power of any individual or collective agents, due to contingent limits on their regulative capacities. Agent-relativity: One can be owed something only if some particular kind of agent has reason him or herself to regulate his or her conduct as regards that something, that is, reason which is not equally reason for other agents to so regulate their conduct, and not equally reason to get them to so act. So a distribution as such cannot be owed, whatever other value it may have. If it were owed, every agent would instead have reason to promote it as far as he or she could, by acting on their own or by spurring others to action. Normativity: One is only owed something by agents who have normally conclusive reasons for action (i.e., conclusive absent extenuating circumstances) as regards that something. So a distribution taken as such cannot be owed. For it will not normally supply conclusive reasons for action; it will normally have value, if at all, only as one value to be balanced against other values (even if it is conclusively required in certain situations, given further reasons). A final feature draws from each of the aforementioned constraints. Context-sensitivity: The different types of agent, capacities, and generic interests are potentially so various that no single regulative principle is likely to be defensible as normally conclusive for them all. So no distribution, as such, is likely to count as owed, independently of a specific context of justification. In conjunction with individualism, these features substantially constrain distributive argument. We cannot then use the common methodology which simply (i) answers Amartya Sen’s question “equality of what?” (welfare? social goods? resources? opportunities? capabilities?), (ii) formulates an appropriate criterion of distribution (equality? priority for the worst off ? sufficiency?), perhaps defending it against alternative criteria, and (iii) concludes that the result constitutes distributive justice. This methodology is at best incomplete. Step (iii) assumes that there are no further constraints on justification to consider. This is fine if we assume distributions, as such, can be intrinsically just: there is then no further question of scope—no further question whether the proposed argument assumes a certain context of justification, and, if so, what exactly that context is. If Scanlon’s constraints apply, however, then no


Political Themes: Conservatism, Justice, and Public Reason

distribution is intrinsically just, and the context of justification is crucial. Until we have a credible sense of what agents are addressed, their capacities, the range of relevant generic interests, and how a regulative principle could nevertheless be normally conclusive across some range of cases, abstract answers to the questions “What is equality?” and “Equality of what?” do not establish even a presumptive criterion of just distribution. The abstract ideal might be a worthy value, but not what is owed. Take luck egalitarianism as an example. One does presumably have some personal reason to somehow be compensated for one’s undeserved bad luck, if only because this would make one better off. Yet the extent to which others can be asked to act on this reason will depend on what is in a given agent’s regulative powers and what any proposed measures will cost them. If there are perhaps situations in which no one could reasonably reject being asked to take compensatory measures of certain sorts, there are also surely many situations in which people can raise reasonable objections—for example, the cost of compensation is high, success is unlikely, and the unlucky person is relatively well off. So long as such cases are not entirely peripheral, and especially if they are common, luck egalitarianism cannot take any very general form. Its defensible scope will have to be restricted to the contexts lacking in such countervailing personal considerations. Rawls’s theory, by contrast, starts closer to the ground, in a specified institutional setting. Principles are tailored to a society’s basic structure, to a list of socially created primary goods, to more or less normal cooperators, and to the general realities of life, motivation, and society. This narrow focus makes it credible to regard the resulting limits on inequality as normally conclusive demands—as not simply valuable but owed. What egalitarian principles lose in breadth of application, because of contextsensitivity, they gain in normative force.

relational foundations

If Scanlon’s constraints are significant, the question is why we should accept them. Is there some organizing rationale why they come all together or not at all? Or can we pick and choose? Why not accept the framework save individualism, as Derek Parfit suggests? Or why not ban direct aggregation (unlike Parfit) and allow other impersonal considerations, rejecting only personalism? Again, someone might then reasonably reject an unequal distribution on the grounds that it is bad state of affairs from an impersonal point of view. Scanlon’s general answer is that individualism is part and parcel of what the domain of what we owe to each other is like. This is less obviously true as compared to the other constraints on owing. Normativity does pretty clearly follow simply insofar as right and wrong are at issue. To judge an act of injury or neglect to be morally wrong is to judge that there is normally conclusive reason not to do it. For reasons of “ought implies can,” capacity-sensitivity plausibly follows as well: the only conclusive reasons will be reasons agents are able to keep track of, plan for, and act on. Agent-relativity is also natural, though perhaps still in need of explication, insofar as it is notoriously

The Significance of Distribution


hard to capture principles of right and wrong by appeal to impersonal value. And context-sensitivity might be seen to follow, though still less straightforwardly, as an elaboration of what all of these prior constraints mean for justificatory methodology. Still, why individualism? Even if the formal ban can be motivated given familiar problems with utilitarian aggregation, why personalism? Why the restriction to personal considerations, to the exclusion of impersonally valuable distributions? Scanlon says individualism, including personalism, is “a central feature . . . that I would not want to give up” (W, 229). But this is not of course to say why it is correct. Scanlon does provide a clue in his comments on the question of whether to preserve a natural treasure such as the Grand Canyon, for impersonal reasons. He writes: . . . it is important to bear in mind the limited range of the part of morality we are trying to characterize. The contractualist formula is meant to describe one category of moral ideas: the requirements of “what we owe to each other.” Reasons for rejecting a principle thus correspond to particular forms of concern that we owe to other individuals. By definition, impersonal reasons do not represent such forms of concern. They flow from the value of those objects themselves, not (at least in the first instance) from anything having to do with my relation to other people. (W, 219, emphasis added) As Scanlon does not elaborate in great detail in What We Owe to Each Other, his commitment to personalism, as expressed in these remarks, seems mainly to be an expression of his overall project. The project is not to show why we have to keep things personal, but that we can. This is significant in itself. Recall the ambiguity in the name “what we owe to each other.” The name at once denotes two relations: a relation one person can bear to another one person, which the one can in principle bear in return, and a relation any one person can bear to all others. A fundamental goal of Scanlon’s theory is to make credible the intuitively attractive thought that the former relation is primary: the relation each bears to all is a relation any two can bear, pair-wise, to each other. Specifically, these relations are of the same basic kind, in the sense that both depend on the same kinds of considerations. So insofar as the basic one-to-one relation is personal, in a sense which excludes impersonal considerations as irrelevant, the relation we bear universally and serially to all others is personal as well, in that same sense. And this is just personalism. Because impersonal reasons are then excluded, aggregation or distribution across persons can only be required, if at all, on the basis of personal reasons. Scanlon’s big idea is that, because his contractualist theory meets a range of theoretical desiderata—it accounts for the motivational basis, the subject matter, and the content of morality, as it is ordinarily understood—we can find credible the idea that morality is at bottom relational, in at least the above abstract sense. Still, the question remains: can we make good sense of the basic pair-wise relation which grounds the whole project? Scanlon has further explicated the relation we bear to persons that we do not bear to the Grand Canyon. He writes:


Political Themes: Conservatism, Justice, and Public Reason

. . . recognizing your value as a rational being is a matter of responding properly to the reasons we have that flow from your reasons, that is to say, from the reasons you have “on your own behalf.” (Nothing similar can be said about the Grand Canyon, which does not have any reasons.) Although we may have reason not to fill up the Grand Canyon, our reason not to break someone’s legs is in part that he or she has personal reasons, him or herself, not to be so injured. What we may or may not do is thus a matter of what is justifiable to that person, in the light of those personal reasons. This is helpful, but at most the start of a general rationale for personalism. This says only why recognizing the value of person must include reference to his or her personal reasons, not why impersonal reasons should be excluded. Why doesn’t recognizing your value require responding to your reason not to have your legs broken and to the fact that this is an impersonally bad thing to have happen? We have a rationale for personalism only when this kind of further impersonal consideration is ruled out. I will now suggest one such rationale by fleshing out the personal nature of Scanlon’s motivational starting point, the idea of “justifiability to another.”

the moral relation

The justifiability of an action to a person is strictly speaking a relational property of actions, a relation between an action and a person. But a relation between persons does seem to be what Scanlon takes to be ultimately at stake in matters of right and wrong. He writes: [T]he pain of guilt involves, at base, a feeling of estrangement, of having violated the requirements of a valuable relation with others. . . . [W]hat is particularly moving about charges of injustice and immorality is their implications for our relations with others, our sense of justifiability to or estrangement from them. . . . [W]hen we look carefully at the sense of loss occasioned by charges of injustice and immorality we see it as reflecting our awareness of the importance for us of being “in unity with our fellow creatures.” (W, 162–63) But what is the single underlying “valuable relation with others” at issue here? Scanlon does explain how the unjustifiability of our conduct to others changes how they can appropriately relate to us: they may properly resent us, be unwilling to remain friends, or simply feel uneasy about being asked to shake our hand. This is, however, only to say that our conduct enables or undermines various specific ways of relating, not that there is a single, underlying “valuable relation with others” at stake. The question then is, What would be a basic moral relation of this kind? As I understand it, the relation in question includes a single basic relation between persons—the moral relation, if you will—and two forms of relational status or standing. The basic relation holds between any two people with certain capacities of rational

The Significance of Distribution


assessment, specifically, capacities which enable, at least in theory, a justifying exchange of reasons for acceptance. Such people not only bear a common property of rationality, they bear a relation of encounterablity. In principle, the two could have a certain face-toface encounter: the one could, in theory, be asked for reasons why there was sufficient reason for his or her conduct, despite foreseeable consequences for the other, and the other could, in theory, accept or reject the sufficiency of those reasons. You ask why I missed our appointment, I explain, and you accept, or reject, my putative justification. When this idea of encounterability is simply out of place, because the thing in question (e.g., a tree, the Grand Canyon) lacks the capacities needed to even in principle have the justifying exchange, one simply has no such relation or standing in relation to that something (even if some other standards of conduct concerning it still apply). There are two forms of status, or two ways of “standing,” in this basic encounterability relation. I am in good standing when all of my actions are justifiable to you—when my reasons would win your acceptance (assuming you are informed, well-motivated, and reasonable). Even if we never have the justifying exchange, I thereby stand justified with respect to you, or (for short) justified by you (much as you might say “fine by me”). I am not in good standing, perhaps to the degree my total conduct falls short, if something I have done or failed to do, which might in some way have affected you, could not be justified to you if you asked; you would not accept the sufficiency of my reasons. Even if you never do ask, I thereby stand unjustified by or with respect to you, to some degree. (Though we bear the basic moral encounterablity relation to everyone, one’s standing as justified by another need not be mutual; the other may or may not stand justified by one.) So the moral relation is thus a relation between persons. Might some conception of its personal nature explain why impersonal considerations must be excluded from the exchange of reasons for acceptance? I suggest the following conception, which has four parts. The first is the 1. Attributabiltiy Condition: How people stand in the moral relation depends only on events that are attributable to one or both of those people. This excludes all purely impersonal occurrences in the world (e.g., earthquakes) as irrelevant. It is not to say what personally attributable events do matter. Thus we may add the 2. Subjective Recognition Thesis: How one stands in the moral relation to a person i depends only on whether or not one displays the subjective regard necessary to recognize i, as reflected in the considerations operative in one’s deliberation or action. Everything thus depends on whether we recognize others in our motivating grounds. But what does “recognizing someone” require? According to the 3. Distinctness Condition: One recognizes a person i, in the sense relevant to one’s moral relation to i, only if the considerations operative in one’s


Political Themes: Conservatism, Justice, and Public Reason

deliberation or action register i as an individual, that is to say, as distinct (i) from the impersonally described and evaluated course of events and (ii) from other persons. In order to exclude impersonal considerations, then, we need only add the further claim that 4. Only motivation by personal considerations registers a person as distinct from the impersonally described and evaluated course of events and from other persons. Motivation by impersonal (or non-person-referential) considerations fails to register persons as distinct in one or both of the required senses, (i) and (ii). In other words, personalism follows from (2), (3), and (4). In short: only personal reasons motivationally register persons as distinct individuals, and such motivation is all that counts.

the attributability condition

The Attributablity Condition, (1), is not strictly necessary for the above reasoning. Its role is to motivate concern with recognition. The argument in its favor generalizes from the coffee case. In such cases, a bad outcome changes my relation to you only insofar as it reflects on or is attributable to me. We generalize to the irrelevance of all outcomes by firmly denying the possibility of “moral luck”: we add, that is, that it makes no difference what outcome, good or bad, actually comes about. The painful spill might have been averted by stroke of good luck. Still, if my conduct was faulty (because I was reckless), our relation will not be as it was before. In other words, the moral relation holds between us as persons. Any exchange of justification for acceptance which bears on that relation therefore cannot take purely impersonal occurrences into account. Now, strictly speaking, that is not to exclude impersonal considerations from the content of personal motivation; in theory, one might still owe it to others to care about and be moved to promote impersonally valuable states of affairs. Yet this thought now lacks a clear rationale. Why, from the point of view of what we owe to others, should we be concerned with and motivated by the thought of certain impersonal events, as such, if the occurrence of those events is wholly irrelevant to our standing with respect to one another?

the subjective recognition thesis

Impersonal concern will of course seem natural if there is nothing else to put in its place. The Subjective Recognition Thesis provides a plausible alternative: one is to recognize others in the considerations operative in one’s deliberation or action.

The Significance of Distribution


A person is fully recognized, we may say, when we give significance and appropriate weight to his or her interests in our motivating grounds. This has a familiar meaning in face-to-face interactions involving conversation and coffee—I notice, and give sufficient weight to, foreseeable costs to a person I can independently identify. It is less clear what “full recognition” might entail as regards distant strangers we cannot know of in any particularity. How does one recognize everyone in a world of several billion people? Scanlon accounts for limitations in our conceptual, epistemic, and other regulatory capacities by framing moral argument in terms of “generic” reasons—reasons we know people have by generally available information. So insofar as I govern myself in a way which is responsive to the generic interests of persons in your position (by following the appropriate principles, with an understanding of their point and generic grounds), I can in a general sense be said to fully recognize you— even if you in your particularity never come before my mind. The relevant general sense of “recognition” is therefore quite distinct from those associated with special causal or social relationships, for instance, picking someone out of a crowd, identifying someone as the same individual one saw before, granting permission to speak or vote within a deliberative body, or knowing someone in all his or her individuality, as a lover might expect of his or her beloved. Indeed, because recognition of the relevant general sort cannot depend on any such specific way of “recognizing” persons as particular individuals, it must be “impersonal” in a fairly straightforward sense. This does not, however, imply impersonal concern for the good of people or humankind—that is, impersonal motivation of the sort we seek to rule out. This is because personal motives, even in a highly generic form, implicitly or explicitly refer to the individuals who have them. They can thus be said to relate us to those different individuals, rather than to humanity or people at large. If I take the pain my hot coffee will cause you, or a headache you happen to have, to be not just a bad state of the world, but bad for you, according your personal reasons, then reference to you is part of my motivating grounds. This remains true when the grounds are quite generic, concerned only with persons in your position. I think, “one suffers in that condition; it is bad for one,” making no reference to your particular sensitivity, distinctive personality, or other distinguishing features. Nevertheless, the grounds assume, by virtue of the kind of grounds they are, that you are a particular individual who has the personal reasons in question, whether or not anyone else does. I thus see pain as significant because it is someone’s pain, rather than a mere impersonally bad occurrence in the world. And I am not simply registering persons in pain in the aggregate, but rather the individual and several persons under the generic description—that is, each and every person like you for whom the condition or sensation in question hurts. In this albeit indirect way, you are registered as distinct from impersonally desirable events in the world, such as the preservation of the Grand Canyon or the reduction of headaches or pain overall, as well as from other persons, who may have different personal reasons. I implicitly pick you out from both the world’s impersonal goings on as well as the crowd.


Political Themes: Conservatism, Justice, and Public Reason the distinctness condition

In other words, personal motivation meets the Distinctness Condition. If “full recognition” is hard to characterize in general terms, conditions (i) and (ii) are bare minimum requirements for recognition of persons as distinct individuals in the most basic sense. If persons are not registered as distinct from the impersonally described and evaluated course of events, we do not count as related to persons as distinct from impersonal objects such as the Grand Canyon. And if persons are not registered as distinct from other persons, we do not count as directly related to each and every person, taken separately. If this is right, we finally have our rationale for personalism: impersonal (or otherwise non-person-referential) motivation is irrelevant to the moral relation, because it makes no contribution to minimal recognition in the specified senses; it fails the Distinctness Condition. To see why, consider first a case of aggregation. Suppose I spill the hot coffee on your lap, not negligently, but solely for the sake of net impersonal gains to other people. (Perhaps the resulting spectacle will cause several bystanders to talk and make a profitable business arrangement that will benefit millions of consumers.) If Kant would say that I am treating you as mere means, not as an end in yourself, the present version of the objection is that I do not fully recognize you as a distinct person. Most important, the problem here is not simply the way your interests have been balanced against the interests of others, a problem that might justify only the formal ban on aggregation. For the idea of recognition also limits the sorts of considerations that may be put into the balance in the first place. If I am only careful not to spill my coffee because the world is an impersonally better place when pain is minimized, this does not register you as distinct from the impersonal course of events, even if the ban on aggregation is respected. If my only thought is that pain is a bad thing to have happen, I have yet to recognize you, the person in pain, as distinct from any way pain might happen to come about or be reduced in the impersonally described and evaluated history of the world. I see the pain, but have yet to recognize the person whose pain it is. To be sure, pain is necessarily someone’s pain. Even if this is true and known, however, this is insufficient for recognition as part of the content of one’s motivating grounds. One can acknowledge the descriptive fact that pain must be someone’s pain without having any motivating thought about its normative significance. One can even be moved by the badness of pain, knowing that the pain must be someone’s, and yet not be moved by the fact that it is someone’s pain. This fact might be as incidental as the known fact that the person has a beating heart, or has spent most of his or her life relatively close to the surface of the earth. Recognition of the person, within one’s motivating grounds, is only achieved when we add personal considerations: we have to add the motivating thought that the pain is bad for a certain reason—because it is bad for the someone whose pain it is. The motives of actual people of course often mix personal and impersonal considerations. The present claim is that any recognition of others we achieve is explained by the personal motives. If impersonal motives are insufficient taken by themselves, this

The Significance of Distribution


shows that they have nothing to add. Why would they suddenly become relevant simply because a wholly different, personal form of motivation was introduced?

recognition in distribuiton

If all of this is right, we can now say more precisely why distributions, as such, cannot be owed: they do not provide a basis for recognizing persons as distinct individuals, in the sense that our moral relation to each person requires. To see this, suppose that, in resolving some distributive problem, we distribute things equally among the people involved. Suppose that our motivating basis is simply that equality of distribution is a good state of the world from an impersonal point of view. (Specifically, our basis is the thought that this reason alone grounds reasonable rejection of any principle that allows inequality). It follows that we do not place value on individuals’ having their respective shares, for their own sakes, for their personal reasons. In that case, however, the chosen distribution is simply a way of bringing about a valuable pattern in the world. And this is just to say that the people involved are not recognized as distinct from the role they happen to have in the larger impersonal world. We can of course add, as a further claim, that distributive shares are also valuable for the different persons, for their personal reasons. But, as above, recognition is then achieved only because personal considerations have been introduced. If the equal distribution is indeed motivated only by impersonal considerations, we also cannot say that the people are recognized as distinct from each other. People can be viewed as distinct from the larger impersonal world and yet not as distinct individuals. Humanity might, for instance, be seen to have a single utility function, the maximization of which happens to require equal distribution (e.g., because of diminishing marginal utility). Yet, much as Rawls famously claimed, to distribute equally on this impersonal basis is not to fully recognize persons as distinct from each other. The formal ban on direct aggregation partly captures this thought: it leaves no (relevant) standpoint of evaluation beyond the imagined objections of different, numerically distinct persons. Personalism is also needed, however. Again, without personalism, we can introduce nonaggregative impersonal considerations from any given standpoint, for example, as grounds for equality of distribution. But suppose, as above, that such grounds provide our only motive for distributing things equally. In that case we are not moved on behalf of anyone. This does not display recognition of the fact that distinct persons are involved. For the distribution is by stipulation not on behalf of a particular someone, in the sense that it is motivated by his or her personal reasons in its favor. Nor, then, can we be supposing that every one has some personal reason to favor the distribution. So we cannot say, in that sense, that equal distribution is on behalf of, and expresses recognition of, each and every one. Yet we do have to say that the distribution is “on behalf of everyone,” in some nondistributive sense. After all, if people weren’t involved, no distribution problem would come up. The distribution, then, is on behalf of everyone but no one. The different individuals involved nowhere figure into the picture; they are not recognized as individuals, as distinct from one another. We can of


Political Themes: Conservatism, Justice, and Public Reason

course add reference to their respective personal reasons, but, again, recognition is then achieved only because personal motivation has been introduced.

some objections

Before turning to what all of this means for political philosophy, I should consider two objections which highlight the intuitively attractive basic relational orientation I mean to be developing. A first objection is that the above conception of recognition is too specific: it suffices for recognition, one may say, simply that no one could reasonably complain of one’s motivating grounds. Hence, if impersonal considerations are allowed as sufficient bases for reasonable objection, being motivated by them would suffice for recognition. Personalism would of course rule that out, but what is wanted here is a rationale for personalism from an idea of recognition. This shows that my proposed characterization of recognition assumes and merely develops the basic relational orientation presently under consideration. Again, we are trying to explicate a relation one can bear to all that is of the same basic kind that we can bear to a single individual person. If we broaden “recognition” as our objection proposes, however, it becomes unsuitable for this role. It is not clear that any individual person is recognized, as such, if impersonal grounds are allowed. According to the proposed view, it suffices for recognition if my motivating grounds reflect principles no one could reasonably complain of. But here we should ask, “Suffices for recognition of whom?” If someone is imagined to complain, it would not necessarily be because that person has a complaint on his or her own behalf. For all we have said, the grounds would be equally available to everyone, and the objection on behalf of no one. The corresponding motivation might then count as recognizing “everyone” in a general nondistributive sense. But we cannot yet say that it recognizes any given person. We cannot yet say we have characterized a relation we bear to all that is of the sort that we bear to a particular person, that is, a relation we bear separately to each and every person, taken as a distinct individual. A second, quite different objection suggests that recognition can instead be seen as only one component of the moral relation. In that case, individuals could be recognized by the acknowledgment of their respective personal reasons, but impersonal reasons might otherwise bear on what people are owed. A partial reply is that certain impersonal considerations (e.g., of aggregation) undercut recognition. This is only a partial reply, because it does not rule out the more moderate view which bans aggregation and merely rejects personalism, allowing impersonal concerns with distribution. Acknowledgment of the personal value of each having his or her share could then do the job of recognizing each person, whereas equality of distribution might be owed because of its impersonal goodness. In reply to this more modest view, we should recall the challenge posed by the Attributability Condition, now supplemented by a personal conception of recognition. The Attributability Condition by itself implies that purely impersonal events, whether distributive or otherwise, are irrelevant to the moral relation that we bear directly to

The Significance of Distribution


different individuals. But why then should we care about those events, at least as far as our relation to others is concerned? Shouldn’t we simply be concerned with the different individuals, and care about worldly events only insofar as they personally matter for one or another person?

relational theories of justice

I now turn, at last, to what all of this means for political philosophy. I have already described the inadequacy of the Sen-inspired methodology of justification. Abstract answers to the questions “What is equality?” and “Equality of what?” do not warrant conclusions about what is owed or required by justice. The case for any principle of distribution must be made in terms of personal reasons, in a context-specific, inductive way. We have also now said why a fully personal argument is necessary: only personal reasons are relevant to the relation of recognition that is chiefly at stake when moral questions of just distribution come up. In other words, a relational theory of justice must be correct. It is important that Scanlon’s theory does not imply any particular relational theory, however. We can still take any of several views about who exactly the “justice relation” relates. According to an interpersonal relational theory (of justice generally, or of a certain justice concept), the justice relation holds between each individual of the world and each other individual. Special social relationships—families, nations, international organizations— raise special opportunities for action as well as special concerns, but do not alter the basic question of what this direct justice relation between all individuals requires. According to a social relational theory (of justice generally, or of a certain justice concept), by contrast, the justice relation holds between individuals and the organized social relationships they are a part of, or substantially affected by, which may or may not include everyone. Rawls, for example, justifies his claims about domestic distributive justice, and the reasoning appropriate for their justification, based on an independent conception (“model conception”) of what a society is. Elimination of inequalities is owed to each member, not because this is necessarily owed among all human persons, but because this is a condition of realizing a just social relation of the sort a domestic society involves (a well-ordered, fair scheme of social cooperation, among free and equal moral citizens). Although Scanlon’s individual-oriented theory may seem to imply or strongly support the former, interpersonal view, this is not in fact the case. Indeed, the pluralistic nature of Scanlon’s framework is especially congenial to social relational views. Because different contexts can require different principles, there is nothing arbitrary or inappropriate in failing to assess all human activity by a single general distributional standard, contrary to Rawls’s critics. At the level of interpersonal relations, principles of voluntary exchange can be different from principles of mutual aid, even if both have distributive effect. At the level of social relationships, domestic justice can be different from justice in the complex social practices and institutions that organize the global scene. And if we are so pluralistic within levels, we can be pluralistic


Political Themes: Conservatism, Justice, and Public Reason

across levels as well. If collective life can be said to raise its own distinctive and equally fundamental questions of justification, then principles for social forms do not need to be somehow derivable from principles justified at the interpersonal level, but rather justified for those social forms themselves. Rawls, for example, is a nonreductivist in that sense: principles of social justice, seen as principles for collectively sustained social practices and institutions, are neither less fundamental nor derivable from principles of interpersonal conduct, and indeed the latter are at best a secondary or independent matter. Why might collective life raise its own distinctive and equally fundamental questions of justification? The question is barely coherent within a consequentialist framework. Scanlon’s framework supports one plausible answer provided only modest further assumptions about the conditions of political life. We need only have generic interests at stake in the governance of something other than our respective individual selves, and thus a fresh collective occasion for asking what the recognition of each affected person requires. Thus we may say that political philosophy’s concern with distributive justice arises in part because we find large-scale patterns of distribution largely beyond any given individual’s control. Because such patterns cannot be said to personally reflect on those individuals, in the sense required by the Attributability Condition, they are not, on a large scale, a matter of justifiability among persons, as such. Yet they cannot simply be left to fate. As is often (but not always) the case, a large-scale pattern can be attributed to certain large-scale social relationships. Especially when such relationships are collectively governed, whether through centralized politics or by decentralized means, the collective then holds a distinctive form of power over each individual caught up in the distributional machine. Distributive justice, on this conception, is achieved only when the relationship’s organization, and resulting distributive pattern, is such that each such individual is fully recognized. The question of what we owe to each other, in other words, is not then about what each owes to each, but rather about what we, collectively, owe to each.

the priority of relations over outcomes: further examples

I have now explained what Scanlon’s theory entails if we situate justice within what we owe to each other. In the remainder of my discussion, I return to the large question of why justice should be so situated in the first place. I offer two reasons for this. First, unlike an alternative reading of Aristotle’s platitude, situating distributive justice within what we owe to each other directly explains the great significance we attach to distributive justice. Second, the resulting priority of relations over outcomes is confirmed by numerous examples akin to our initial natural disaster case. If Aristotle’s platitude was supposed to link justice to what we owe to each other, it does not have to be interpreted in this way. On an alternative interpretation, justice is by nature giving each his or her due only in the following sense: each is (by nature) owed his or her appropriate place in a pattern of distribution. Nothing in the nature

The Significance of Distribution


of justice then precludes a distribution from being required as such. If justice requires that reward be distributed according to merit, then each is due the place in the pattern of reward that he or she deserves. Or if justice requires a distribution free from undeserved misfortune, where a person’s misfortune is deserved only insofar it results from his or her own choices, then each is due his or her place in a pattern of this kind. An adequate conception of distributive justice needs to explain why the issue is normally of great if not overriding importance to us. Situating justice within what we owe to each other explains this in a direct way: what is a stake is a moral relation of corresponding significance. By contrast, the alternative reading of Aristotle’s platitude just suggested does not directly explain the significance of distributive justice, if it explains it at all. The proposal is of course that justice can require a certain distributional pattern. But the question is then why the pattern should be seen to have the great importance we assign to justice. Nor is it enough to say that people are due or owed certain things. If one’s due is simply having one’s place in a certain pattern, whatever it is, and whatever the pattern’s value, no essential reference to the value of that pattern is assumed. If the distributional pattern has little or no value, it is hard to see why it should be of great importance, as a matter of justice, for each to have his or her place in it. And even if one adds that the properly specified distributive pattern is indeed valuable, as an independent claim, this will not necessarily meet the relevant explanatory burden. The account of the pattern’s value may or may not capture the general importance that distributive justice has for us; everything depends on what sort of value is involved. The argument against situating justice within what we owe to each other might instead appeal to examples. One might argue that certain distributions are so obviously unjust, in and of themselves, that my argument shows only that justice cannot be part of what we owe to each other. As I will now explain, a range of examples suggests otherwise. Examples such as the natural disaster case suggest that the significance of distribution lies in the character of social relations rather than mere outcomes, even in cases of obviously unjust distribution. Consider an emergency situation involving a sinking ship. Knowing they have a moderate chance of rescue, but being fair people, the passengers adopt an impeccably fair procedure for distributing the different kinds of life preserving device on hand, in effect assigning different ex ante life prospects to each. It could well have been “every person for him or herself.” Because the passengers instead chose fair arrangements, however, they stand in a just social relation, however the different life prospects turn out. Everyone may die; everyone may be saved; or some may die whereas others are saved. If the procedure of distribution really was impeccably fair, the people stand justified with respect to each other whatever the eventual result. An emergency situation is hardly a model for what justice substantially requires of a governed society in normal conditions. Because the outcomes of regularized coordination can be to a considerable extent known and controlled, it will often be culpably negligent to simply leave them to fate. Here the independence of relations from outcomes is revealed mainly in extenuating circumstances, as in our natural disaster


Political Themes: Conservatism, Justice, and Public Reason

cases. Although the same unequal distribution of material goods obtains in the two societies, in the one case it does not reflect on, or is not attributable to, the society itself. Thus the society displays no failure of recognition. The outcome is unfortunate and the hand of fate has perhaps been unfair, but just as my moral relation to you is not undermined when my coffee is simply pushed onto your lap, there is no injustice in the social relations going on. In still further cases, it becomes clear that no untoward outcomes need come about. The society can simply have the wrong relation to a welcome outcome. In a variation on the natural disaster case, suppose that disaster never strikes despite overwhelming evidence to the contrary. If the negligence remains, say, because the evidence was played down or ignored, the society remains no less unjust. The fortunate outcome is not to its credit; it says nothing about the society’s relation to its members. The society’s members have not been fully recognized. Extenuating circumstances aside, we may often find it hard to believe that a society in the normal conditions of human life will lack control over distributive outcomes. We therefore expect certain distributions to actually come about, as a condition of just social relations. We may then find extreme poverty combined with great inequality obviously unjust, but not because we think such a distribution is, as such, obviously unjust, in and of itself. Indeed, any appearance of obvious intrinsic injustice fades when we consider the familiar test-case of two isolated societies. Suppose the one society is rich and goods and opportunities are fairly distributed, whereas the other is tragically poor, despite all best efforts and a fair distribution of what scarce goods and opportunities there are. Here the global state of affairs of severe poverty and great inequality seems neither just nor unjust if the societies cannot interact. Perhaps the inequality is regrettable, fate has been unfair, and the world is in one way bad, or otherwise objectionable. Yet it reflects nothing about the character of the social relations going on. There is no reason as yet to deny that everyone stands justified with respect to everyone else, in both their special relationships and their bare relations as persons. It is important that the essential thought here does not depend on the impossibility of interaction. If the societies could suddenly interact, but had yet to do so to any significant degree, the global distribution of goods or opportunities will still not necessarily reflect the character of international or transnational social relations, such as they are. The Attributability Condition is still not met. It is only when, given a fair amount of luck and time, meaningful patterns of interaction become established between our two societies or their members that we can begin to consider the extent to which the global distribution of goods or opportunities is genuinely within their power, and so reflective of or attributable to their social relations as opposed to the mere workings of fate. Any number of circumstances could prevent regularized interaction of this kind. Perhaps the societies have limited capacities for effective communication; perhaps they have radically different views of the world and its workings; perhaps they as yet lack mutual awareness of common problems; perhaps they fail to see stable coordinative solutions; perhaps they see too many solutions, as no one solution is yet salient enough to provide a “focal point” solution; perhaps there is

The Significance of Distribution


uncertainty about the degree of expected compliance and so the costs of participation; perhaps all of these factors have a role, to different degrees, in different contexts of possible interaction. Although such obstacles to cooperation often combine with ill will, distrust, and violent, manipulative, or exploitative action, they can in principle arise without anyone being at fault. The conditions for cooperation themselves can simply be wanting. I do take the current state of international relations to give rise to significant egalitarian demands, beyond humanitarian obligations to eliminate absolute poverty. The state system, the rule of international law, and the institutions of trade and lending governing the global economy generate significant principled limitations on both poverty and inequality across societies, even if they are of a more limited and varied kind than the requirements of domestic justice. I take this to be the case, however, because patterns of social coordination have arisen which generate special responsibilities, over and above the requirements of mere humanity. It is not enough that a world without poverty or inequality is desir