from rationality to equality · from rationality to equality james p. sterba 1. 3 great clarendon...

From Rationality to Equality

This page intentionally left blank

From Rationalityto Equality

James P. Sterba

1

3Great Clarendon Street, Oxford, OX2 6DP,United Kingdom

Oxford University Press is a department of the University of Oxford.It furthers the University’s objective of excellence in research, scholarship,and education by publishing worldwide. Oxford is a registered trade mark ofOxford University Press in the UK and in certain other countries

# James P. Sterba 2013

The moral rights of the author have been asserted

First Edition published in 2013

Impression: 1

All rights reserved. No part of this publication may be reproduced, stored ina retrieval system, or transmitted, in any form or by any means, without theprior permission in writing of Oxford University Press, or as expressly permittedby law, by licence or under terms agreed with the appropriate reprographicsrights organization. Enquiries concerning reproduction outside the scope of theabove should be sent to the Rights Department, Oxford University Press, at theaddress above

You must not circulate this work in any other formand you must impose this same condition on any acquirer

British Library Cataloguing in Publication Data

Data available

Library of Congress Cataloging in Publication Data

Data available

ISBN 978–0–19–958076–7

Printed in Great Britain byMPG Books Group, Bodmin and King’s Lynn

Contents

Preface vi

1. Introduction 1

2. The Historical Connection to Immanuel Kant 10

3. From Rationality to Morality 23

4. Recent Critics of the Rationality-to-Morality Argument 63

5. Alternative Justifications for Morality 86

6. From Liberty to Equality 101

7. Recent Critics of the Liberty-to-Equality Argument 162

8. Alternative Justifications for Welfare and Equality 190

9. Conclusion 211

Selected Bibliography 220Index 227

Preface

Most contemporary moral and political philosophers would like to have anargument showing that morality is rationally required. In this book, Iprovide just such an argument and further show that morality, so justified,leads to substantial equality.The argument of this book has been a number of years in the making,

with earlier versions presented before many audiences, large and small, thelargest being my Presidential Address for the American PhilosophicalAssociation (Central Division) in 2008. Challenges to my argument multi-plied, following that address. Subsequently, two mini-conferences ofpapers on my argument were held at Notre Dame in 2009 and 2010.Other challenges to my argument were raised at the American Philosoph-ical Association Meetings and the Cato Institute in 2010 and 2011. As onewould expect, the argument has been modified and developed in light ofthese challenges.Obviously, if the argument of this book were ever widely accepted and

practically applied, it would radically change the societies in which we live.Of course, that would only happen if enough readers, like yourself, cameto endorse it. So the crucial question, at the moment, is: What do youthink of it?I wish to thank Bielefeld University and Zentrum für interdisziplinäre

Forschung (ZIF), in particular, Martin Carrier, for providing the perfectsetting for completing a final draft of the book during the summer of 2011and for moving it through production during the summer of 2012 withthe help of Martin Barr and Sarah Parker at Oxford University Press. I amalso grateful to Peter Momtchiloff, the Senior Commissioning Editor forPhilosophy at Oxford University Press, for his suggestions and assistance atcrucial stages in the process of writing the book. In addition, I wish tothank Oxford University Press, Cambridge University Press, WestviewPress, and Environmental Ethics for permission to reprint from previouslypublished material. Most importantly, however, I wish to thank mypartner, Janet Kourany, without whose constant support and encourage-ment, as only a fellow philosopher can provide, this book would surely nothave been possible.

1

Introduction

It is generally recognized that in today’s society academic philosophershave very little impact on moral and political decision-making. Forexample, in contrast to members of other disciplines and professions,philosophers have very rarely in our times been called upon to serve asadvisers to governors, labor leaders, presidents, prime ministers, or evendictators. To some extent, this is because philosophers have not, untilrecently, directed their attention at the practical issues that daily concernour moral and political leaders. But just as importantly it is becausephilosophers have done so little to resolve the fundamental conflictsbetween opposing moral and political ideals of our times. In this book,I will try to improve upon the status of my profession just a bit by offeringa justification of morality and further by showing how morality so justifiedleads to a demand for substantial equality.Most contemporary moral philosophers would like to have an argu-

ment showing that morality is rationally required, but given the history ofpast failures to provide a convincing argument of this sort, most contem-porary moral philosophers have simply given up any hope of defendingmorality in this way. Thus, I belong to a relatively small group of contem-porary moral philosophers that includes Stephen Darwall, BarbaraHerman, Christine Korsgaard, Allen Wood, and others, who think thatit is possible to provide such an argument for morality. I also argue that theconception of morality so justified can be shown to lead to a demand forsubstantial equality, thereby linking myself to a much larger group ofmoral and political philosophers, including Ronald Dworkin, AlisonJaggar, Will Kymlicka, Charles Mills, Thomas Pogge, Larry Temkin,and others, who think that a commitment to morality leads to broadlyegalitarian requirements, although the view I defend is more egalitarianthan the views defended by most of these other philosophers.

The Historical Connection to Immanuel KantMy project of trying to argue from rationality to morality and then frommorality to equality is inevitably associated with the moral and politicalphilosophy of Immanuel Kant. Kant too, especially in Part III his Ground-work of the Metaphysics of Morals, was hoping for a justification of morality interms of rationality. In Chapter 2, I take up Kant’s attempt to justifymorality and argue that he relies too heavily on a commitment to anincompatibilist conception of freedom, failing to recognize that additionalgrounds are required if morality is to be shown to be rationally preferableto egoism. Nevertheless, I maintain that Kant would have welcomed thefurther justification of morality I provide in this book given that it supportshis overall philosophical perspective in moral and political philosophy.The second part of my argument that begins with morality, in particular,

a libertarian morality, and attempts to derive therefrom a requirement ofsubstantial equality connects up with Kant’s political philosophy, whichhas standardly been given a libertarian interpretation. I show, however,that this interpretation of Kant’s work is mistaken and that his politicalphilosophy really has a welfare liberal foundation. Both liberty and welfareare important for him. Accordingly, Kant surely would have welcomedmy argument from libertarian premises to welfare liberal, and even egali-tarian, conclusions as a way of completing his own project in moral andpolitical philosophy.

From Rationality to MoralityIn Chapter 3, I set out my justification of morality. It is based on the claimthat the principle of non-question-beggingness favors morality overegoism, where morality is understood to be a nonarbitrary compromisebetween self-interested and altruistic reasons. A crucial step in this argu-ment for morality is to cast the basic conflict with egoism not as a conflictbetween morality and self-interest, but rather as a conflict between altruismand self-interest. I argue that while previous defenders of morality under-stood correctly that moral reasons could not be compromised with self-interested reasons, they failed to recognize that this is because moralreasons are already the result of a purportedly nonarbitrary compromisebetween self-interested and altruistic reasons. To ask that moral reasons beweighed against self-interested reasons is, in effect, to count self-interested

2 FROM RATIONALITY TO EQUALITY

reasons twice—once in the compromise between egoism and altruism thatconstitutes a conception of morality, and then again, assuming moralreasons are weighed against self-interested reasons, and this double-counting of self-interested reasons would be clearly objectionable from anon-question-begging standpoint. Thus, while previous defenders ofmorality intuitively knew that moral reasons could not be compromisedwith self-interested reasons, they were unable to conceptually back up andsee how morality itself can be represented as a compromise betweenaltruism and self-interest, and, for that reason, they failed to recognizemy approach to defending morality.1 In this chapter, I also argue thatmorality can be non-question-beggingly justified by appealing to premisesthat the egoist (and the altruist), in fact, share with the moralist.

Recent Critics of the Rationality toMorality Argument

As one might expect, my argument from rationality to morality has notbeen without its critics. The argument in its present form owes a great dealto past critiques of its earlier incarnations, in particular, to critiques pro-vided by Karl Ameriks, Robert Audi, Annette Baier, Kurt Baier, MartinCarrier, Stephen Darwall, Michael DePaul, Jamie Dreier, William Fitzpat-rick, Bernard Gert, Alan Gewirth, Elizabeth Harman, Barbara Herman,Alison Hills, Brad Hooker, Richard Joyce, Christine Korsgaard, SharonLloyd, Alastair MacIntyre, Alastair Norcross, Charles Pigden, BetsyPostow, Jeffrey Reiman, Robert Shaver, Peter Singer, Holly Smith,Michael Smith, David Solomon, Michael Thompson, among others.Some of these critiques were fatal in intent, so that if I have survived themat all, as I think I have, I have done so only through the argumentativeequivalent of justified self-defense. Other critiques were simply attemptingto alert me to perceived weaknesses in my argument that required shoring

1 My justification of morality also bears a loose analogy to the contemporary solution ofFermat’s last theorem. Fermat’s last theorem could not be solved until it was viewed in acertain way in light of more recent developments in mathematics. Likewise, my justificationof morality involves seeing the problem of justifying morality in a different way as emergingout of a conflict between egoism and altruism and mediated through the use of a moredeveloped account of non-question-beggingness that appears to have been unavailable toKant.

INTRODUCTION 3

up. So, in these cases, I attempted to buttress the defenses of my argument, asseemed appropriate.As one might expect as well, my now refurbished and more developed

justification for morality has attracted a new set of formidable critics, inparticular, Charles Mills, Anita Superson, Candace Vogler, Russ Shafer-Landau, and Allan Gibbard. In Chapter 4, I consider and reply to thesenew critics. Like many of my past critics, these recent critics also look uponmy argument for morality, for the most part, with deadly intent. Either myargument is thought to fail to adequately respond to some opponent ofmorality, such as a new form of egoism (Mills), or a certain version ofmoral skepticism (Superson), or it fails to employ a normative standard indefense of morality (Vogler), or it is trying to do what cannot be done(Shafer-Landau), or it uses a style of argument that can then be turnedagainst it (Gibbard). Clearly, these are very serious critiques of my defenseof morality. In this chapter, I hope to show how my argument can, in fact,be strengthened and developed through responding to them.

Alternative Justifications for MoralityIn general, there are three kinds of justifications for morality that must betaken into account before my own justification for morality can be fullyendorsed.First, there is a need to consider the possibility of other Kantian justifi-

cations of morality. That is why even before setting out my own justifica-tion in Chapter 3, I first consider the attempts by Alan Gewirth andChristine Korsgaard to ground morality in consistency alone.Second, there is the need to consider Hobbesian justifications for

morality grounded in self-interest. So, in Chapter 5, I take up the justifi-cations for morality developed by Kurt Baier and David Gauthier. Theobvious advantage of such justifications is that they are clearly non-question-begging vis-à-vis egoism since they simply start with egoisticpremises.Third, there is the need to consider Aristotelian justifications for morality

grounded in personal flourishing or eudaemonia. At first glance, Aristote-lians have an obvious complaint against the justification ofmorality that I amdeveloping in this book. For them, the very idea of framing the justificationof morality as requiring a response to the egoist presupposes a Kantian


conception of a morality grounded in rationality rather than an Aristotelianconception of a morality grounded in personal flourishing or eudaemonia.To adequately respond to this objection, it must be possible to recast mycentral argument into an Aristotelian framework. And that is what I alsopropose to do in this chapter. In addition, I also consider a widely popularAristotelian conception of morality developed by Ayn Rand, because itstrongly clashes with the justification of morality that I defend in this book.

From Liberty to EqualityThe conception of morality that I non-question-beggingly defend againstegoism is, at this stage of my argument, anything but complete. Itsrequirements seem to be open to a libertarian, or a welfare liberal, oreven a socialist interpretation. While I argue that this conception ofmorality is quite useful because it succeeds in showing the superiority ofmorality over egoism, the conception still needs to be completed in orderto determine when its requirements are enforceable. I do this in Chapter 6.Here, I claim, it behooves us to start with the assumptions that areacceptable from a libertarian perspective, the view that appears to endorsethe least enforcement of morality, given that I propose to show that thisview requires a right to welfare, and that further, this right to welfare,which is also endorsed by a welfare liberal perspective, leads to thesubstantial equality of a socialist perspective. My strategy is to find conflictsof (negative) liberty within the libertarian perspective, and then argue thatwhen these conflicts are appropriately resolved, they favor an allocation ofliberty that supports a right to welfare. Since fundamental rights areuniversal rights for libertarians, I then argue that extending this right towelfare to distant peoples, and, particularly, to future generations leads tothe egalitarian requirement that we use up no more resources than arenecessary for meeting our own basic needs, thus, securing for ourselves adecent life but no more.2 For us to use up more resources than this,

2 Basic needs, if not satisfied, lead to significant lacks or deficiencies with respect to astandard of mental and physical well-being. Thus, a person’s needs for food, shelter, medicalcare, protection, companionship, and self-development are, at least in part, needs of this sort.It also should be pointed out that at least some of the goods required for the satisfaction ofpeople’s basic needs are participatory in nature. These goods relate primarily to the satisfactionof people’s basic needs for companionship and self-development and require a level ofpolitical participation that is necessary for a decent life. Fixing the appropriate level for a

INTRODUCTION 5

I argue, without a technological fix on hand, we would be guilty ofdepriving at least some future generations of the resources they wouldrequire to meet their own basic needs, thereby violating their libertarian-based right to welfare. In effect, recognizing a right to welfare, applicableto all existing and future people, leads to an equal utilization of resourcesover place and time.While my egalitarian conclusion does depend on extending a right to

welfare to distant peoples and future generations, roughly the same con-clusion can be derived in yet another way—by taking into account themoral status of all living beings. I present this alternative argument for myegalitarian conclusion in this chapter as well.

Recent Critics of the Liberty-to-EqualityArgument

My argument from liberty to equality has not been without its critics.Accordingly, the argument in its present form owes a great deal to pastcritiques of its earlier incarnations, in particular, to critiques provided byElisabeth Ashford, Allan Buchanan, Douglas Den Uyl, Gerald Doppelt,Julie Driver, Carol Gould, John Hospers, Erin Kelly, Eva Kittay, EricMack, Alistair Macleod, Ernest Partridge, Thomas Pogge, Douglas Ras-mussen, Holmes Rolston, Ruth Sample, Daniel Shapiro, John Simmons,Tara Smith, Philippe Van Parijs, Karen Warren, Mary Ann Warren, PaulWeithman, Peter Wenz, Iris Young, and Naomi Zack. These previouscritiques have enabled me to improve my argument in innumerable ways.In Chapter 7, I consider and reply to would-be devastating critiques thathave now been directed against the latest version of my argument by JohnChristman, Gerald Gaus, Allan Gibbard, Tibor Machan, and Jan Narve-son. Like myself, Christman and Gibbard are critical of the libertarian viewendorsed by Machan and Narveson. In contrast, Gaus defends his ownform of libertarianism. But all of them find common cause against myliberty-to-equality argument, although for different reasons. Christmanchallenges my willingness to join libertarians in taking liberty to be basic.

basic needs minimum will turn out to be easier to do on the view that I ultimately will bedefending if everyone is entitled to receive just that minimum and no more, other thingsbeing equal.


Gibbard claims that my view favors the rich too much. Gaus maintains thatmy view requires too much altruism. Machan claims that implementingmy view would make for bad law, and Narveson maintains that I fail torecognize that in conflicts between the rich and the poor, normally thepoor, not the rich, are the first harmers. Obviously, I have much torespond to in this chapter in order to maintain my liberty-to-equalityargument.

Alternative Justifications for Welfare andEquality

Now others have also attempted to provide a justification welfare andsubstantial equality. In Chapter 8, I consider attempts to do so by GillianBrock, G.A. Cohen, Ronald Dworkin, John Harris, Richard Miller,Richard Norman, Henry Shue, Jeremy Waldron, and Left-Libertariansas a group. While some of these philosophers have argued from explicitlynonlibertarian premises (Cohen in his last book), and one has tried to showthat the libertarian’s own view is absurd (Dworkin), most have simply triedto show, just as I have, that the libertarian ideal, properly interpreted, leadsto welfare or equality. In this chapter, I examine each of these attempts inturn, comparing them to my own argument from liberty to equality.

The Contrast with Rawls’s WorkNow one might wonder why the work of John Rawls, arguably the mostimportant political philosopher of the twentieth century, does not figureprominently in this book. The reason for this is that there are two mainparts to the overall argument of my book, an argument from rationality tomorality (to be set out in Chapter 3) and an argument from libertarianmorality to equality (to be set out in Chapter 6), and Rawls, unfortunately,was not engaged in developing either of these arguments.Early on in A Theory of Justice, Rawls set aside the question of how to

respond to the egoist as a question that he was not concerned to answer.Nor in arguing for his welfare liberal view does Rawls attempt to derivethe equality he defends from an ideal of liberty, or make any seriousattempt to win over libertarians, like his long-time Harvard colleague,

INTRODUCTION 7

Robert Nozick. Rather, Rawls simply sets out his welfare liberal view aspromising alternative to libertarianism. So it is Rawls’s distancing himselffrom the two main arguments of my book that has led me not to discussRawls’s work more than I do.Nevertheless I have connected my work in two important ways to

Rawls’s work. In Chapter 5, I show how the nonmoral standard of non-question-beggingness, which I use to justify morality, has its analogue inthe moral standard of the original position which Rawls uses to justify histwo principles of justice. In addition, in Chapter 7, I show that G.A.Cohen’s attack on Rawls’s theory as being insufficiently egalitarian is itselfinsufficiently inegalitiarian.Interestingly, new scholarship on Rawls (Weithman, 2011) has now

opened up another way to connect Rawls’s work with the main argu-ments of my book. As it turns out, although Rawls did set aside anyattempt to respond to the egoist at the beginning of A Theory of Justice, atthe end of the book, when Rawls was assessing the stability of his theory,he worried that his theory might not be stable enough to resist a challengecoming from a “self-interested perspective.” And this worry over thestability of his theory, as is well known, ultimately led to Rawls makinghis political turn in Political Liberalism. So the challenge of a self-interestedor egoistic perspective to Rawls’s theory actually had a more significantrole in the development of his theory than is usually thought. Moreover,given that Rawls never explicitly attempted to resolve the self-interestedchallenge to his view that he himself raised, the twofold argument of thisbook can be understood as a way of completing Rawls’s own work bothwith regard to the challenge coming from egoism and the challengecoming from libertarianism as well. Seen in this way, the project of mybook can happily be regarded as Rawlsian as well as Kantian.

ConclusionMoral and political philosophy has always been a collective endeavor. It issomething we work out together in an ongoing discussion. Althoughsome of us are more engaged in that discussion than others, we are allengaged in it because its two central questions—Why be moral? and Whatdoes morality require?—are inescapable. Even the moral skeptic or theegoist, or the moral skeptic or the egoist in each one of us, has to engage


these questions, if only to try to show that we are not really required to bemoral because morality is not really rationally binding on us. Accordingly,my own work on these questions in this book is itself a product of a longhistorical discussion that goes back at least to Plato and to discussions thatI myself have had over the years with other philosophers, students, andpeople from all walks of life. Those discussions are especially evident in myresponses to recent critics in this book.3

Yet discussions in moral and political philosophy ideally should leadsomewhere. Hence, the goal of this book is to provide an argumentfrom rationality to equality that will help to resolve the fundamentalconflicts between opposing moral and political ideals of our times andthereby prepare the way for the peaceful implementation of its egalitarianconclusions.

3 Truth be known, the five recent critics of my argument from rationality to morality andthe five recent critics of my argument from libertarian morality to equality did not just turn upby happenstance. Rather, I actively recruited these well-known philosophers because theywere identified through their publications with perspectives that fundamentally contrastedwith my own. I utilized various venues in which I could get these philosophers to challengethe view that I am developing in this book. I then responded to their critiques and later gavethem a chance to revise their critiques in light of my initial responses, which they all did. AsI see it, my view is best understood, as well as usefully tested, through critiques, like these, andby my responses to them. For further elaboration of this way of doing philosophy, see myJustice for Here and Now (New York: Cambridge University Press, 1998), ch. 1.

INTRODUCTION 9

2

The Historical Connection toImmanuel Kant

Morality within the Limits of Reason AloneIn the history of moral philosophy, Immanuel Kant (1724–1804) is bestknown for his attempt to ground morality on reason alone. For Kant, thefundamental law of morality or categorical imperative is itself an a priorilaw of reason. As such, it relies on experience for neither its origin nor itsjustification. By insisting on a priori grounding for morality, Kant rejectedearlier attempts to ground morality on observable facts about human nature,such as that all humans desire happiness. Kant also rejected any attempt toground morality on religion or the existence of God. For him, morality canonly be grounded on reason alone, or more precisely, on practical reasonalone, which is the capacity of an agent to freely determine his her ownactions. As Kant put it in Religion within the Limits of Reason Alone:

So far as morality is based upon the conception of man as a free agent who, justbecause he is free, binds himself through his reason to unconditional laws, it standsin need neither of the idea of another Being over him . . . nor of any incentive otherthan the law itself, for him to do his duty.1

Now morality clearly depends on the assumption that it is possible for us toact on its requirements, that is, it depends on the assumption that we arefree. For Kant, this implied that actions are not determined by anythingother than our will, which itself is free. As Kant saw it, if our actions weredetermined by anything other than our free will, we would, as he oncesuggested, be like puppets, merely reacting to the forces that play upon us.Expanding on the same theme in hisCritique of Practical Reason, Kant wrote:

1 Immanuel Kant, Religion within the Limits of Reason Alone, trans. Theodore Greene andHoyt Hudson (New York: Harper & Row, 1960), Preface, first paragraph.

Suppose I say of a man who has committed a theft that this act, by the natural lawof causality, is a necessary result of the determining ground existing in the preced-ing time and that it was therefore impossible that it could have not been done.How, then, can judgment according to the moral law make any change in it? Andhow can it be supposed that it still could have been left undone because the law saysthat it should have been left undone? That is, how can he be called free at this pointof time with reference to this action, when in this moment and in this action hestands under inexorable natural necessity? It is a wretched subterfuge to seek anescape in the supposition that the kind of determining grounds of his causalityaccording to natural law agrees with a comparative concept of freedom. Accordingto this concept, what is sometimes called “free effect” is that of which thedetermining natural cause is internal to the acting agent . . . So one might call theactions of man “free” because they are actions caused by ideas we have producedby our own powers, whereby desires are evoked on occasion of circumstances andthus because they are actions brought about at our own pleasure; in this sense theyare called free even though they are necessary because of the determining groundsthat preceded them in time. With this manner of argument many allow themselvesto be put off and believe that with a little quibbling they have found the solution tothe difficult problem which centuries have sought in vain and which could hardlybe expected to be found so completely on the surface.2

As Kant saw it, if we attempt to ground morality in experience or the factsabout human nature, we are inevitably driven to endorse an inadequatecomparative conception of freedom, what we would call today, a compat-ibilist conception of freedom. This is because our experience and the factsabout human nature are presumably governed by deterministic laws. Sothat if we consider human nature just as we consider other objects ofnature or experience from a natural science point of view then, Kantthinks, we have to admit that the only laws that govern them are presum-ably deterministic laws of nature. We therefore must grant that we have noscientific basis for the assumption that we are free in anything other thanwhat Kant regarded as an inadequate compatibilist sense of freedom. Thatis why Kant was convinced that an empirical grounding could not providethe basis for an adequate conception of human freedom.Accordingly, in the third section of his Groundwork of the Metaphysics of

Morals, Kant tries to show how morality is possible only if a different kindof freedom is possible. Kant understood this different kind of freedom tohave both negative and positive aspects. Understood negatively, freedom

2 Immanuel Kant,Critique of Practical Reason, trans. Lewis White Beck (New York: Bobbs-Merrill, 1958): 96.

THE HISTORICAL CONNECTION TO IMMANUEL KANT 11

is a form of causality that is independent of “alien causes determining it.”For Kant, “alien causes,” were laws of nature. Yet while this freedom isindependent of the laws of nature, Kant did not see it as altogetherindependent of law. Understood positively, Kant claimed, this freedomhas its origin in a pure will that issues in laws of freedom. Moreover, theselaws that a rational agent gives to herself through her pure will are not validfor some rational agents but not others. For Kant, they are laws that arevalid for all rational agents, and thus they command ends that could bewilled by all rational agents.

Compatibilist FreedomYet while Kant’s defense of morality builds upon a rejection of a compat-ibilist conception of freedom, many philosophers have been unwilling togive up on the idea that such a conception of freedom is defensible andconsistent with morality.Many contemporary defenses of compatibilist freedom also offer a

“could have acted otherwise” analysis of free acts. According to suchanalyzes, to say that a person X could have acted otherwise than he didis to say that there was in the choice situation some physically possibleaction A such that, had X willed (or chosen or tried or tried hard, etc.) toperform A, he would have acted otherwise than he did. Thus, for example,to say that Harriet, hard-core smoker that she is, could have stoppedsmoking when her friend Freda did is to say that there was some specificaction (or series of actions), e.g., buying and using a set of graduatedcigarette filters, or joining Smokers Anonymous, or showing up forhypnosis, or the like, such that, had Harriet willed to perform that action(or series of actions), then she would have stopped smoking when herfriend Freda did. Notice that, according to this analysis, Harriet mightactually have tried to stop smoking when her friend Freda did, might havefailed to do so, and nonetheless might actually have had the ability, thenand there, to do so—as many ex-smokers will doubtless attest.Still, it has been objected that this sort of an analysis fails because its

analysans, but not its analysandum, will be satisfied by situations in whichX could not have willed to perform A. Thus, returning to hard-coreHarriet, there might have been some one particular action, such as buyingand using a set of graduated cigarette filters, such that, had Harriet willed to


perform that action, then she would have stopped smoking when herfriend Freda did. But it might also have been the case that Harriet couldnot have willed to buy and use such a set of graduated cigarette filters, that,hard-core smoker that she is, she could not have brought herself to will (ortry or choose) to do it. And if that were also the case, then we would wantto say that Harriet really could not have stopped smoking when her friendFreda did—that the analysandum of our example is false—even thoughthe proposed analysans is true. And this is, of course, to say that theproposed analysis of “X could have acted otherwise” is, as it stands,inadequate. But so will be any other analysis, any other “attempt to define‘can’ in terms of ‘will if ’.” For any such attempt will presumably introducea verb of which “she” may be made the subject; it will then be grammat-ically permissible to insert “cannot” between the subject “she” and theverb, and then it will be possible to describe a situation in which theanalysans is true and the analysandum is false.To deal with this problem, let us return to the analysis of “X could have

acted otherwise than he did” suggested earlier. Let us agree that theanalysans should include something like:

a. There was in the choice situation some physically possible actionA such that, had X willed to perform a at time t0 (t0 = or > t), hewould have acted otherwise than he did at t.

And let us agree, given the above discussion, that the analysans must alsoinclude something like:

b. X could have willed to perform A at t.

But here we should not regard the latter as unanalyzable. Indeed, weshould allow that the latter amounts to something like:

X would have willed to perform A at t0 ifi. The facts of the choice situation that were reasons motivating X toact had been relevantly different from what they were at t0; and/or

ii. X’s values had been relevantly different from what they were at t0

where “the facts of the choice situation that were reasons motivating X toact” means those facts of X’s environment that were both reasons foracting and were taken by X to be reasons for acting; and where “relevantlydifferent facts”means physically possible facts that would provide sufficientreasons to motivate X to prefer to do A without altering the basic oppor-tunities that X justifiably believes obtain in the choice situation, and


“relevantly different values” means physically possible values that likewisewould have been sufficient to motivate X to prefer to do A withoutaltering the basic opportunities that X justifiably believes obtain in thechoice situation. Again, let us agree that the problem pointed out earlierrequires us to further include in the analysis:

c. The facts of the choice situation that were reasons motivating X toact could have relevantly different from what they were at t0; and/or

d. X’s values could have been relevantly different from what theywere at t0.

But, again, let us not allow that the latter are unanalyzable. Indeed, thelatter can be analyzed as follows:

i. The facts of the choice situation that were reasons motivating X toact could have been relevantly different from what they were at t0;that is to say, they would have been relevantly different if still earlierfacts had been relevantly different; and these earlier facts could havebeen relevantly different, that is to say . . . and/or

ii. X’s values could have been relevantly different from what theywere at t0; that is to say, X’s values would have been relevantlydifferent from what they were at t0 if*. the facts of the choice situation that were reasons motivating

X to have the values that he did at t0 had been relevantlydifferent from what they were at t0, so as to motivate X toprefer relevantly values; and these facts of the choice situationcould have been relevantly different from what they were at t0,that is to say, they would have been relevantly different if stillearlier facts had been relevantly different; and these earlier factscould have been relevantly different, that is to say . . . ; and/or

**. the facts of previous choice situations that were reasons mo-tivating X to acquire the values that he had at t0 had beenrelevantly different from what they were, so that X wouldhave acquired relevantly different values by t0; and these factsof previous choice situations could have been relevantly dif-ferent from what they were, that is to say, they would havebeen relevantly different; and these still earlier facts could havebeen relevantly different, that is to say . . .

At no time, therefore, must we settle for an “absolute,” unanalyzable“can” “grinning residually up at us like the frog at the bottom of the


beer mug,” as J.L. Austin suggested would be true of all compatibilistanalyzes.3

Still, it might be objected that people not infrequently claim to know,sometimes on very plausible grounds, that they or others “could haveacted otherwise than they did.” Yet according to the analysis of freedomoffered here, a potentially infinite number of general conditions have to besatisfied before any statement of the form “X could have acted otherwisethan he did” is true. And that being the case, how could anyone everknow that such a statement is true? But is there really any problem here?To say that a potentially infinite number of general conditions have to besatisfied before any statement of the form “X could have acted otherwisethan he did” is true is not to say that one must know that an infinitenumber of conditions have been satisfied before one can plausibly be saidto know that the statement is true.4 Nor is it to say that any one of theseconditions, or the evidence relevant to determining whether or not it hasbeen satisfied is just as significant as any other for deciding whether or notit has been satisfied is just as significant as any other for deciding whetherthe statement of the form “X could have acted otherwise than he did” istrue. In short, it does not follow from the proposed analysis of freedomthat any particular kind or amount of evidence is needed to warrant a claimof the form “X could have acted otherwise than he did.” And this seems tobe as it should be. For the question of the kind and amount of evidencethat is needed to warrant a claim of the form “X could have actedotherwise than he did” seems to be just the kind of question whose answerwill very properly vary from context to context, and cannot, therefore, bedecided in the abstract, in conjunction with a general analysis of “X couldhave acted otherwise than he did.” At any rate, this analysis cannot befaulted for not providing such an answer.5

As we have seen, Kant rejects compatibilist accounts of freedom asinadequate. For him, they attempt to explain free acts as acts that are

3 See Roderick Chisholm, “J.L. Austin’s Philosophical Papers,” Mind (1964): 20–5.4 For example, one can know that a scientific theory is true without knowing that all the

empirical implications of that theory are true.5 For a fuller defense of this compatibilist analysis of free action, see James P. Sterba and

Janet A. Kourany, “How to Complete the Compatibilist Account of Free Action,” Philosophyand Phenomenological Research 41 (1981): 508–23. Harry Frankfurt has propounded a well-known argument that acts can be free even when the agent could not have acted otherwise.In the last section of the paper just citied, Kourany and I argue that Frankfurt is mistaken onthis point; free acts are acts where the agent could have acted otherwise.


internally and immediately caused in or by the agent, but in a way that atsome point connects up to causes outside the agent both at the presentmoment and in the past. However, Kant does not seem to recognize thatcompatibilists can offer accounts of free acts, like the one that I have justpresented, according to which agents could have acted otherwise thanthey did. So it is unclear whether Kant would have taken compatibilistaccounts of freedom more seriously if he had recognized that they couldbe developed in this way. In any case, Kant spends little time discussingsuch accounts before turning to his own incompatibilist account of free-dom in order to explain the origin and justification of morality.

Incompatiblist FreedomNow central to Kant’s incompatibilist account of freedom is his claim thatpure wills, or rational agents through their pure wills, are uncaused causes.Taking pure wills or rational agents to be uncaused causes gives them anindependence from the facts about human nature that are part of theempirical world in which we live. Thus, taking human agents to beuncaused causes secures for Kant the negative aspect of his conception offreedom. But how does taking human agents to be uncaused causes securethe positive aspect of his conception of freedom? More specifically, whywould uncaused human agents choose the moral law or any kind of lawfor that matter? Why would they even choose to act in a law-like mannerrather than arbitrarily or lawlessly? Nothing about human agents specifiedsimply as uncaused causes tells us anything about what they will do, orwhat law, if any, they would will for themselves.A parallel problem arises in Plato’s Euthythro. There Socrates raises this

sort of question: Are actions right because God willed them or does Godwill actions because they are right? Now if we go with the first alternativeand assume that God is just an all-powerful uncaused cause, nothing more,nothing less, then clearly anything is possible. God could will what wepresently think are good actions to be right in a consistent, law-likemanner, like the rule that we should help our neighbor in need, or Godcould will what we presently think are bad actions to be right in aconsistent, law-like manner, like a rule to torture people for the fun ofit; or God could arbitrarily will sometimes what we consider to be goodactions to be right and sometimes will what we consider to be bad actionsto be wrong in no consistent pattern. Needless to say, these consequences


of the first Euthythro alternative have not made that option very attractive.Socrates clearly rejects it.Now it would appear that a similar unattractiveness afflicts Kant’s

account of rational agents as uncaused causes. Their willing need not resultin any law at all, let alone a law that we would think deserves to be called amoral law. Even if we just assume that human uncaused causes wouldchoose to act in a law-like manner, rather than lawlessly—clearly anassumption Kant appears to make—it doesn’t follow that they wouldnecessarily adopt for themselves something that we would call a moral law.Of course, they could adopt Kant’s categorical imperative in, say, its first

formulation. They could will that we:

Act only on the maxim which you can at the same time will to be auniversal law of nature.

But they could also adopt instead the basic principle of universal ethicalegoism:

Each person ought to do what best serves their own self-interest.

Needless to say, these alternative “laws” are quite different. Kant’s cat-egorical imperative is understood to be symmetrically action-guiding, thatis, each agent is to act according to the imperative and others are at leastnot to interfere in the agent’s so acting. By contrast, the principle ofegoism is asymmetrically action-guiding, that is, while each agent shouldact according to the egoistic principle, others when their interests conflictare to attempt to stop the agent from so acting. A useful analogy here, onewe will discuss in more detail later, are the “oughts” of competitive gameswhere what one player ought to do, opposing players ought to try to stophim from so acting. In any case, what we can see here is that both theprinciple of egoism and Kant’s categorical imperative have the form ofultimate practical laws. So that even if we could discover an argument thathuman uncaused causes would opt for laws over lawless activity, Kantwould still need an additional argument showing that his categoricalimperative should be preferred to the law of egoism.

Morality and EgoismIs there such an argument? Well, we could ask why would a rational agentendorse the law of egoism? The honest answer appears to be she wants to


do what best serves for her own self-interest, but she also is committed tobeing rational at least in the minimal sense of being consistent, and thisrequires her to admit that if it is rational for her to pursue her best self-interest then is equally rational for others to do the same. It is this that leadsto an endorsement of the law of egoism, while giving it the asymmetricallyaction-guiding interpretation we just considered. Moreover, what thisshows is that it is facts about human nature, not those about our mereinclinations, but rather those about the concern we all have for own self-interest together with a concern for consistency that is the underlyingrationale for the law of egoism. So couldn’t Kant argue that this shows thatthe law of egoism has an empirical grounding that is absent from hiscategorical imperative?Yet is Kant’s categorical imperative really so different from the law of

egoism in this regard? Clearly, acting on Kant’s categorical imperative doesnot serve just the interests of the agent, although it does do that to somedegree. Rather, it serves the interests of all those who would be affected bythe law in a way that purports to be reasonable to all. As ChristinaKorsgaard puts it, the categorical imperative aims at a “good [that] is aconsistent harmonious object shared by all rational beings.”6 I think this isright. But note what constitutes this shared good is given to us by factsabout human nature, not those that are about our mere inclinations, butthose that are about what really is for the good of ourselves and otherstaken together. It is these facts that are worked into the formulation ofKant’s categorical imperative, just as the facts about what is for the self-interested good of one agent opposed to all others are worked into theformulation of the law of egoism.So both the law of egoism and Kant’s categorical imperative help

themselves to facts about human nature in formulating their universallaws. Accordingly, if we want to support morality over egoism, we willneed a further reason for doing so. Kant was aware of this problem. As hesaw it, an analytic appeal to rationality alone would not serve to defendmorality.7 While the defender of morality could claim to be fully rational,so could the defender of egoism. Kant’s hope was that by wedding

6 Christina Korsgaard, “Kant’s Formula of Humanity,” Kant-Studien 77(2) (1986):183–202, 196.

7 Immanuel Kant,Groundwork of a Metaphysics of Morals, trans. by H.D. Paton (New York:Harper & Row, 1956): 440.


morality to an incompatibilist account of freedom, he would be able togive morality a foundation that was unavailable to egoism or to those otheraccounts of morality that were based on facts about human nature. Wehave now seen that Kant’s incompatibilist account of freedom, his view ofhuman agents as uncaused causes, does little to support, and does more toundermine (since uncaused causes can will anything and everything)Kant’s own attempt to defend morality. It is only when Kant goes on todefend his categorical imperative by appealing to facts about humannature, such as the good we all can share together, that his account ofmorality begins to get a hold on us, even while it still fails to show thatmorality is rationally required. Notice too that this appeal that Kant comesto make to facts about human nature to justify morality accords well with acompatibilist conception of freedom that is all about facts as to how humanagents could have acted otherwise.According to some interpreters, after theGroundwork, Kant himself gave

up on trying to show that morality is required by rationality.8 Maybe Kantcame to realize that his incompatibilist account of freedom did nothing toestablish that morality is rationally required. In any case, in this book,I attempt to provide just the sort of justification of morality that Kant hasreason to welcome, albeit one that does not rely on an incompatibilistconception of freedom, as Kant’s does, with its problematic view ofrational agents as uncaused causes.9

Kant and Libertarian MoralityTurning to Kant’s views on liberty, welfare, and equality, there is astandard interpretation that begins with Wilhelm von Humboldt and

8 See John Rawls, Lectures in the History of Moral Philosophy (ed.) Barbara Herman (Cam-bridge, MA: Harvard University Press, 2000): 253–72; Karl Ameriks, Kant’s Theory of Mind(Oxford: Clarendon, 1982), ch. 6.

9 Another possibility, suggested to me by one of the reviewers of this manuscript, is thatjust as Kant assumed that we have theoretical knowledge and only sought to show how suchknowledge is possible, so Kant can be interpreted as also just assuming that we have practicalknowledge, including knowledge of moral requirements, and then only seeking to show howmorality or moral knowledge is possible. Interpreted in this way, Kant would not be seekingto justify morality at all. He would only be seeking to show how morality, assuming that it isjustified, is possible. But if this is all that Kant sought to establish with respect to morality, thenthe completion of my project of showing how morality is justified would also successfullycomplete Kant’s project of showing howmorality is possible as well. My argument would justdo all of what Kant, on this interpretation, wanted to do and more.


runs through Friedrich Hayek, Robert Nozick, and Michael Oakeshottthat sees Kant as decidedly libertarian.10 The main grounds for this inter-pretation is found in Kant’s Doctrine of the Right where he maintains thatthe state’s legitimacy derives from the need to guaranteed mutual externalfreedom to its members.11 So, it would seem that in a Kantian state,coercive power could only be used to secure external freedom—just aslibertarians maintain.As further support for their interpretation, libertarians cite Kant’s explic-

it and repeated rejection of welfare or happiness as a grounds for legisla-tion. According to Kant, it is not the business of the state to provide for thewelfare or happiness of its citizens but rather to conform its practice to theprinciples of right (i.e., freedom). As Kant put it, “A government estab-lished on the principle of benevolence toward the people, like that of afather toward his children—that is, a paternalistic government . . . is thegreatest despotism thinkable.”12 In addition, like libertarians, Kant main-tains that the equal dignity or worth of each human to be “quite consistentwith the greatest inequality in terms of the quantity and degree of theirpossessions.”13

Now opponents of this libertarian interpretation of Kant rely heavily onthe following passage from the Metaphysics of Morals:

To the supreme commander there belongs indirectly, that is insofar as he has takenover the duty of the people, the right to impose taxes on the people for its ownpreservation, such as taxes to support organizations providing for the poor, found-ling homes and church organizations.The general will of the people has united itself into a society which is to maintain

itself perpetually; and for this end it has submitted itself into a society which isto maintain the state in order to maintain those members of the society whoare unable to maintain themselves. For reasons of state the government is thereforeauthorized to constrain the wealthy to provide the means of sustenance to those whoare unable to provide for even their most necessary natural needs. The wealthyhave acquired an obligation to the commonwealth, since they own their existence

10 See AllanWood,Kantian Ethics (New York: Cambridge University Press, 2008), ch. 11;Alexander Kaufman, Welfare and the Kantian State (New York: Oxford University Press,1999), chs. 1–3.

11 Immanuel Kant, The Metaphysical Elements of Justice, trans. John Ladd (New York:Bobbs-Merrill, 1965): 315–16.

12 Immanuel Kant, Practical Philosophy, (trans. and ed.)MaryGregor (Cambridge:CambridgeUniversity Press, 1999): 291.

13 Ibid. 292.


to an act of submitting to its protection and care, which they need in order to live,on this obligation that state now bases its right to contribute what is theirs tomaintaining their fellow citizens.14

Clearly, the passage comes down strongly in favor of the state’s providingaid to the poor, foundling homes, and church organizations.Nevertheless, proponents of the libertarian interpretation see this pas-

sage as legitimating the provision of welfare to the poor only when failingto do so would threaten the very existence of the state.15 No doubt, failingto provide welfare to the poor might threaten the very existence of thestate under certain extraordinary conditions, but it is hard to see howfailing to provide for foundling homes or church organization would everhave such an outcome. So clearly Kant must have envisioned somebroader justification for state action here. There is also reason to thinkthat Kant’s expressed opposition to the state’s pursuit of welfare or be-nevolence was grounded in his opposition to paternalism. Needless to say,a justification for providing for the basic needs of the poor does not have torest on paternalism. If the poor do not want to receive aid, they can be freeto reject it.An even stronger case for rejecting the libertarian interpretation of

Kant’s views is based simply on what is implied by the state’s grantingequal liberty rights to all its citizens. Thus, Kant thought that it wasn’tenough for citizens to just formally have equal rights; they must also beable to enjoy them.16 And this for the poor, in particular, would requirethe provision of an adequate welfare, without which their “equal (nega-tive) liberties” would be of little value.Nevertheless, the case I will be making against the libertarian view in

this book does not depend on utilizing the distinction between havingliberties and enjoying them, important though that distinction is. Rather,I will be arguing that a rationally and morally defensible account of equalnegative liberty, all by itself, requires a societal right to welfare.Yet what about extending such a right to distant peoples and future

generations? Obviously, the practical basis for seriously implementing suchan extension of welfare rights did not exist in Kant’s time. Moreover, thevery idea that through the excessive consumption of the Earth’s resources

14 Ibid. 468.15 Kaufman, Welfare and the Kantian State, ch. 1.16 Immanuel Kant, Metaphysical Elements of Justice, 306.


we could generally threaten the existence and basic well-being of futuregenerations probably did not even occur to many people until the twenti-eth century. Still, Kant was well aware of unjust wars and practicesbetween the European nations of his time. In addition, the Europeancolonial conquests had been long under way by Kant’s time. All of whichraised serious problems of corrective justice, some of which still persist tothe present day. However, in Perpetual Peace, Kant limited himself toproposing a set of principles under which the peaceful coexistence ofnations would be achievable.17 In this book, I will attempt to go furtherby directly addressing the question of the welfare rights of distant peoplesand future generations. It is clearly a question that we can no longer excuseourselves from thinking seriously about. On this, Kant, if he were aroundtoday, would most suredly have agreed.

17 Kant, Practical Philosophy, 311–52.


3

From Rationality to Morality

Now to defend or justify morality, it would be helpful to show that moralityis grounded in rationality. This requires not just showing that morality issimply rationally permissible because that would imply that egoism andimmorality were rationally permissible as well.1 Rather, what needs to beshown is that morality is rationally required, thus excluding egoism andimmorality as rationally permissible.2 Unfortunately, the goal of showingthat morality is rationally required has been abandoned by most contem-porary moral philosophers who seem content to show that morality issimply rationally permissible.3 No doubt most contemporary moral philos-ophers would like to have an argument showing that morality is rationallyrequired, but given the history of past failures to provide a convincingargument of this sort, most contemporary moral philosophers have simplygiven up any hope of defending morality in this way.4 Here, in contrast,I hope, maybe foolishly, to provide just such a defense of morality and to

1 Although I mention egoism and immorality separately here, egoism, for me, is still oneparticular type of immorality. I further understand the moral–immoral classification (wheremoral includes morally permissible as well as morally required) to be exhaustive. Accordingly,I take the amoralist to be just a particular kind of immoralist, not too different, if different atall, from the egoist. Of course, the egoist and the amoralist both deny the force of moralreasons, but when they act on their denials they also become immoralists, i.e., people who actimmorally.

2 While egoism is an ethical perspective because it provides norms about how one shouldbehave, it is not what I would regard as a moral perspective because it never requires a personto sacrifice her overall interest for the sake of others. Yet even when egoism is seen in this wayas an ethical, not a moral, perspective, the egoist can still be regarded as immoral when she failsto conform to requirements of morality.

3 John Rawls is typical here, as is Thomas Nagel. See Rawls’s A Theory of Justice (Cam-bridge, MA: Harvard University Press, 1971): 136; Nagel’s The View from Nowhere (NewYork: Oxford University Press): 200ff. See my earlier discussion of how my argumentcontrasts with Rawls’s work.

4 Ibid.

show further how morality so justified leads to a demand for substantialequality.

Alan GewirthNow the most ambitious attempts to justify morality have tried to showthat endorsing egoism or immorality is somehow inconsistent. In oneversion of this argument, offered by Alan Gewirth, our desires for thefreedom and well-being necessary to achieve our purposes, when univer-salized, as consistency requires, are said to lead to our endorsing a moralright to freedom and well-being, which in turn requires us to reject egoismand immorality.5

The central premises of Gewirth’s argument for morality can be sum-marized as follows:

1. All agents regard their purposes as good according to whatevercriteria are involved in their actions to fulfill them.

2. Therefore, all agents must affirm a right to the freedom and well-being necessary to achieve their purposes.

3. All agents must affirm such a right on the basis of simply beingprospective, purposive agents.

4. Hence, all agents must affirm that every prospective, purposive agenthas a right to freedom and well-being.6

Gewirth claims that the universalized right affirmed in the conclusion ofhis argument is a moral right, that is, a right that is action-guiding for therightholder and for others as well, a right that implies at least that othersought not to interfere with the exercise of that right. Such rights aresymmetrically action-guiding because they are action-guiding both forthe rightholder and for others as well.Nevertheless, the success of Gewirth’s argument depends on the im-

possibility of interpreting the universalized right in his conclusion asanything other than a moral right. Unfortunately for Gewirth’s argument,

5 Alan Gewirth, Reason and Morality (Chicago: University of Chicago Press, 1978).6 Ibid. chs. 1 and 2; Alan Gewirth, “The Rationality of Reasonableness,” Synthese 57

(1983): 225–47; “From the Prudential to the Moral,” Ethics 95 (1985): 302–4; “Why ThereAre Human Rights,” Social Theory and Practice 11 (1985): 235–48; “Ethics and the Pain ofContradiction,” Philosophical Forum 23 (1992): 259–77.


another interpretation is possible. According to this interpretation, auniversalized right can be deduced from the premises of his argument,but it is a prudential right, not a moral right. This interpretation is plausiblebecause Gewirth maintains that the right referred to in premise (3) isprudential,7 and the universalization of a prudential right can be under-stood to be a prudential right, albeit a universal one.8

Now what distinguishes a prudential right from a moral right is that aprudential right is action-guiding for the rightholder only and not forothers, and so it does not imply that others ought not to interfere withthe exercise of that right. Accordingly, when I say that I ought, pruden-tially, to pursue my freedom and well-being, I have to grant that others,similarly situated, ought to do so as well—that everyone has the samejustification as I have for behaving self-interestedly—but I don’t have togrant that I should help or not interfere with their pursuit of freedomand well-being or that they should help or not interfere with my pursuitof freedom and well-being, as a symmetrically action-guiding moralright to freedom and well-being would require.Prudential rights are also analogous to the “oughts” found in most

ordinary cases of competitive games—cases that we otherwise would havethought conform to the requirements of practical reason. For example, infootball a defensive player may think that the opposing team’s quarter-back ought to pass on a third down with five yards to go, while notwanting the quarterback to do so and indeed hoping to foil any suchattempt the quarterback makes. Or, to adapt an example of Jesse Kalin’s, ifyou and I are playing chess, at a certain point in the game I may judge thatyou ought to move your bishop and put my king in check, but thisjudgment is not action-guiding for me. What I in fact should do is sitquietly and hope that you do not move as you ought. If you fail to makethe appropriate move and, later in the game, I judge that I ought to putyour king in check, that judgment, by contrast, would be action-guidingfor me. So, prudential or self-interested oughts are asymmetrically action-guiding, just as the oughts of competitive games are asymmetrically

7 For Gewirth’s claim that the right in premises (3) is prudential, see “Replies to MyCritics” in Gewirth’s Ethical Rationalism edited by Edward Regis (Chicago: University ofChicago Press, 1984): 205–12.

8 No doubt some will find Gewirth’s notion of a “prudential right” to be something of anoxymoron, but in his account the notion is clearly defined to be the equivalent of a prudentialas opposed to a moral ought.

FROM RATIONALITY TO MORALITY 25

action-guiding. Universalizing prudential or self-interested oughts, there-fore, as consistency demands, only leads to generalized asymmetricallyaction-guiding oughts; it does not lead to the symmetrically action-guiding oughts that constitute morality. Given then that the universalright to freedom and well-being in the conclusion of Gewirth’s argumentcan plausibly be interpreted to be a prudential right, Gewirth’s justifica-tion of morality cannot succeed, because it depends on the impossibilityof interpreting the universal right in the conclusion of his argument asanything other than a moral right.In responding tomy critique, Gewirth claims that my failure to recognize

the strength of his particular argument is rooted in our different accounts ofprudential and moral rights.9 For Gewirth, prudential rights aim to supportthe interests of the agent or speaker, whereas moral rights aim to support ortake favorable account of the interests of at least some persons other than thespeaker or the agent. For me, prudential rights are asymmetrically action-guiding in that they are action-guiding for the rightholder only and notfor others, and so they do not imply that others ought not to interfere withthe exercise of those rights. I also think that these rights are analogous tothe asymmetrical “oughts” found in most ordinary cases of competitivegames. By contrast, moral rights, I claim, are symmetrically action-guiding.Thus if I have a moral right to my life, then that right should guide both myactions and the actions of others; in particular, it should lead others not tounreasonably interfere with my life.Gewirth thinks that these two accounts of moral and prudential rights

do not coincide because, on his account, a (moral) right may guide aperson’s actions with the aim of supporting both his own interests and theinterests of others. But this is true on my account as well. This is because inclaiming that moral rights are symmetrically action-guiding, guiding bothoneself and others, I never held that the reasons why they are action-guiding for oneself are always just self-interested. For example, I canexercise my right to life, and frequently do, for the benefit of others aswell as myself. Nevertheless, it still holds that my right to life, unlikeprudential rights and the oughts of competitive games, is action-guidingfor others as well as myself because others have at least a prima facie duty

9 See Alan Gewirth, “The Rational Justification of Morality Revisited,” in Social andPolitical Philosophy: Contemporary Perspectives, (ed.) James P. Sterba (London: Routledge,2001): 71–6.


not to interfere with my life. So contrary to what Gewirth claims, ouraccounts of moral and prudential rights do coincide after all.Now, Gewirth thinks his argument from rationality to morality works

because when an agent universalizes a prudential claim that she ought(prudentially) to have—or (prudentially) has a right to—freedom andwell-being, and thus asserts that everyone ought to have (or has a rightto) freedom and well-being, the agent is thereby committed to “uphold-ing the interests of other persons as well as herself.” But while an agent,through universalization, does certainly make a claim about others as wellas herself (the claim clearly refers to everyone), the claim still does notcommit the agent to “upholding the interests of other persons as well asherself,” as long as the ought or right in the universalized claim is inter-preted as asymmetrically action-guiding, in a manner analogous to theoughts of competitive games. What this shows is that while consistencyand universalizability do require an egoist to admit, at least to herself, thateveryone else has the same justification as she has for behaving egoistically,they do not require her, in my terms, to endorse symmetrically action-guiding oughts or rights or, in Gewirth’s terms, to “uphold the interests ofother persons as well as herself.”

Christine Korsgaard

Still, another version of the argument to justify morality by consistencyalone advanced by Christine Korsgaard maintains that our reasons foraction must be public in the way that languages are public and that thispublicity requirement is inconsistent with egoism and immorality.10

10 Christine Korsgaard, “The Sources of Normativity,” in The Tanner Lectures on HumanValues (Salt Lake City, University of Utah Press, 1992): 20–112. In her e-mail comments onan earlier version of this critique of her view (2/3/08), Korsgaard says that she does not intendher argument here to be an appeal to consistency, as I characterize it, but rather argues thategoism fails a publicity requirement for reasons analogous to the way “private languages” fail apublicity requirement for languages. But in “The Sources of Normativity,” Korsgaard doessay, “The idea of a private language is inconsistent with the normativity of meaning,” (p. 95)and I take this to also mean “inconsistent with the publicity of meaning.” So I would havethought she is also claiming that just as private languages are inconsistent with the publicity ofmeaning, so egoism is inconsistent with the publicity of reasons. And, of course, what I amarguing here, is that objectionable though the egoism may be on other grounds, the view isnot inconsistent with a publicity requirement because its reasons are public analogous to theway that languages are public, or to the way that the “oughts” or reasons of competitivegames are public, even though they are not public in exactly the same way that moral reasons


According to this argument, egoism is, in fact, a myth, no more possiblethan private languages are possible. But while it is surely the case that anynormative ideal must be public in the sense that it is communicable toothers, don’t egoistic and immoral ideals satisfying this requirement?Consider the egoistic ideal in its general form according to which

everyone ought to do what best serves his or her overall self-interest.11

From Plato to the present, we find numerous attempts to defend theconsistency and reasonableness of this ideal. Throughout this time, phi-losophers have discussed this egoistic ideal and communicated with oneanother about it. Moreover, in practice, egoists would also be willing tocommunicate their reasons or interests to others who have overlapping orcompatible reasons or interests in order to secure for themselves thebenefits of coordination in joint endeavors.But what about those occasions when the reasons of egoists conflict

with the reasons of others? Surely then egoists will not want to communi-cate their reasons to those with whom they are in conflict so as not to loseout. It is just here that egoists will want their reasons to be kept private. Inthis respect, egoists differ sharply from those who are committed tomorality. Those committed to morality usually want to communicatetheir reasons to those with whom they are in conflict in the hope that amorally acceptable resolution of the conflict can be achieved.12 So we canagree that egoistic reasons have a private dimension to them that moralreasons lack. Nevertheless, even when egoists are striving to keep theirreasons private, those reasons still remain public in the sense that they arecommunicable to others. In such cases, their reasons can also be found outeven when egoists are striving to conceal them.In this respect, such egoistic reasons are again analogous to the reasons

found in competitive games. Players in football are usually trying toconceal the particular reasons they have for being in certain formations,just as players in chess are usually trying to disguise the particular reasonsthey have for making certain moves. Nevertheless, in such cases, the

are public. But, of course, Korsgaard wasn’t arguing that egoistic reasons fails a publicityrequirement because they are not public in exactly the same way that moral reasons arepublic!

11 This general form is usually called universal ethical egoism. There is another form ofegoism, individual ethical egoism, that I will take up later.

12 Those committed to morality, however, may want to hide their reasons from those whoare not similarly committed so as to avoid being taken advantage of by such individuals.


players can be found out, as when an offensive lineman in football inad-vertently signals a running play by the way he lines up to block.It might be objected that the “oughts” of competitive games, unlike the

“oughts” of egoism, are contained with a higher normative structure that isitself governed by the symmetrically action-guiding “oughts” of morality.This is true.13 But for an analogy to be useful, the analogues need not beidentical in every respect. However, for an even closer analogue to theasymmetrically action-guiding “oughts” of egoism, think about the asym-metrically action-guiding “oughts” used by players who systematically andsuccessfully cheat to win games they would otherwise lose.What this shows is that egoism meets the reasonable demand of being

consistent with the publicity requirement of languages and competitivegames by being communicable to others, even though it does not meet thestronger publicity requirement of morality of usually wanting its recom-mendations to be communicable to others. Meeting the stronger publicityrequirement of morality would render the practice of egoism self-defeating in just the same way that it would render the practice of manycompetitive games self-defeating. But this fact could only count againstthe practice of egoism if it also counts against the practice of competitivegames, which it surely does not.Still, Korsgaard wants to show that egoism must be rejected because we

cannot (in consistency) escape having a moral identity.14 She wants toshow that we cannot fail to identify ourselves as just someone, a person,one person among others who are equally real. So she asks us to consider asituation where she is being tormenting by a stranger. She asks hertormentor, “How would you like it if someone did that to you?” Hertormentor, whom we can presume believes herself to be an egoist, re-sponds, “Someone doing that to me, why that would be terrible! But then

13 Christine Korsgaard makes a similar point in Self-Constitution (Oxford, Oxford Univer-sity Press, 2009): 195. There she claims that in competitive games a shared good is pursued bypitting people against each other within the limits of certain rules. But at the same time that ashared good is being pursued in competitive games another nonsharable good—winning—isalso being pursued, and it is with respect to this nonsharable good that the asymmetrical“oughts” of competitive games hold. Similarly, egoists also can recognize shared goods onthose occasions when their interests are fully compatible with the interests of others. Gener-ally, it is only when their interests conflict with the interests of others that egoists engage inobjectionable behavior.

14 The Tanner Lectures on Human Values, Lecture 3, 93ff.; The Sources of Normativity,(Cambridge: Cambridge University Press, 1995), Lecture 4, 132ff.


I am me after all.” According to Korsgaard, this response is not intelligiblespeech because in this situation her tormentor cannot fail to see Korsgaardas another person like herself equally real.Now suppose we grant that Korsgaard is right in that we must think of

ourselves as one person among others who are equally real. Would it stillbe possible then for one of us—a would-be egoist—to intelligibly say whatKorsgaard’s hypothetical tormentor said to her?Clearly, to be an egoist is to have an overriding concern for one’s own

interests. An egoist who endorses egoism in its general formwould also haveto acknowledge that given that one’s interests are overriding for oneself, theinterests of everyone else would be similarly overriding for them.15 For suchan egoist, “oughts” here, like Gewirth’s “prudential rights,” are asymmet-rically action-guiding, that is, they require each person to do what bestserves his or her own interests, but they do not require anyone to do whatbest serves the interests of anyone else unless that also happens to best servehis or her own interests. Like Gewirth’s notion of a prudential right, theegoist’s “oughts” are also analogous to the oughts found in most ordinarycases of competitive games. For example, in baseball a catcher may thinkthat the runner on second ought to try to steal third, while not wanting therunner to do so and indeed hoping to throw the runner out if she tries todo so. Or, in a cross-country race, I might know that a competitor oughtto charge “Cardiac Hill” if he is going to have any chance of beatingme in a5K race, while still hoping that he just takes the hill at his regular pace,mistakenly thinking that he can later make up the distance between usby out sprinting me in the last leg of the race. So the egoist’s “oughts” areasymmetrically action-guiding in just the same way as the “oughts” ofcompetitive games are asymmetrically action-guiding.Thus, when Korsgaard’s tormentor says, “Someone doing that to me,

why that would be terrible! But then I am me after all.” She is speakingabout her overriding concern for her own interests. She is not denying thatKorsgaard is another person; she is just saying that from her egoistic pointof view, her interest in tormenting Korsgaard is totally overriding. More-over, egoists can actually be helpful to others, or at least not hinder them,in situations where their interests are not at stake. Egoism does not requirethat you always disregard or go against the interest of others; it only

15 Again, the view of the individual ethical egoist will be taken up later.


requires that you do so when it would best serve our own interests. Egoistscan (consistently) recognize the reality of other persons and sometimesserve their interests without also granting them either equal rights or moralstanding.In addition, although Korsgaard in her example draws her tormentor

into a public discussion of her views, egoists tend to avoid such discussions.In public, egoists usually defend morality as strongly as anyone, because itis when other people are moral that egoists usually benefit most frombeing egoists. And they usually benefit most when they are not known tobe egoists but are thought to be moral. So egoists generally try to cover uptheir self-interested tracks, so to speak, and do everything in their power tofoster morality in others. But egoists are reluctant to speak their views inpublic. Korsgaard’s tormentor is not speaking “mere noise” to her. She issimply telling Korsgaard that from her perspective her own interests arethe only ones that count in this context. Obviously, this is not a veryconsoling message for someone who is being tormented, but it is anintelligible message nonetheless, especially when we imagine analogousexchanges in the context of competitive games. Thus, imagine a quarter-back saying to a defensive lineman who has just sacked him three times,“How would you like it if someone did that to you?” To which thelineman responded, “Someone doing that to me, why that would beterrible! But then I am a defensive lineman playing against you after all.”So unable to show that egoism cannot meet the publicity requirement oflanguage or that egoists can be reduced to talking “mere noise,” Korsgaardfails to provide the justification for morality over egoism that she isseeking.16 From the failure of both Gewirth’s and Korsgaard’s arguments,therefore, we can see that if we are to defeat the egoist or immoralist, wewill need to base our argument on more than consistency alone. Mycritique of Gewirth’s and Korsgaard’s justifications of morality is thattheir appeal to consistency alone is too thin a reed on which to support

16 There may be a way, however, to bring Christina Korsgaard’s work in line with thedefense of morality I am proposing in the next section. At the very end of her responses to hercritics in Sources of Normativity, Korsgaard remarks that perhaps the most essential rule ofreflection is “that we should never stop reflecting until we have reached a satisfactory answer,one that admits of no further questioning” (p. 258). If we interpret this rule as requiring anon-question-begging starting point, and thus taking into account the behavior of womenand focusing on altruism as well as egoism, then, it may be possible to see Korsgaard’s workand my own as pointing toward a common justification of morality.


a justification of morality. To properly justify morality, we need to gobeyond consistency and embrace non-question-beggingness as well.

My ArgumentSo let us begin then by imagining that each of us is capable of entertainingand acting upon both self-interested and moral or altruistic considerations,and that the question we are seeking to answer is what considerations itwould be rational for us to accept as reasons for action.17 This question isnot about what considerations we should publicly affirm as reasons foraction, since people will sometimes publicly affirm considerations asreasons for action that are quite different from those they are prepared toact upon.18 Rather, it is a question about what considerations it would berational for us to accept as reasons for action at the deepest level—in ourheart of hearts—since we are trying to answer this question as far aspossible without self-deception or hypocrisy.19

Of course, there are people who are incapable, by nature, of actingupon moral or altruistic considerations. For such people, there is noquestion about their being required to act morally or altruistically. Yetthe interesting philosophical question is not about such people but aboutpeople, like ourselves, who are capable of acting morally as well as self-interestedly, and who further, let us assume (usefully idealizing a bit) areaware of all the relevant moral and self-interested considerations they arecapable of acting upon and are seeking a rational justification for followinga particular course of action. It is important for my argument later that we

17 “Ought” presupposes “can” here. So unless people have the capacity to entertain andfollow both self-interested and moral reasons for acting, it does not make any sense to askwhether they ought or ought not to do so. Moreover, as I will make clear later, moral reasonsare understood here to include some altruistic reasons and some self-interested reasons. So thequestion of whether it is rational for us to follow self-interested reasons rather than moralreasons should be understood as the question of whether it is rational for us to follow self-interested reasons exclusively rather than some appropriate set of self-interested reasons andaltruistic reasons, other things being equal.

18 A self-interested consideration for X is a consideration favoring the self-interest of X,and a self-interested reason is a self-interested consideration that has met the appropriatestandards for motivation and justification. The same holds true for moral or altruistic consider-ations for X and moral or altruistic reasons for X.

19 What would count as a reason for action here would do so on both defensible internalistand externalist accounts of such reasons.


also include here knowledge of one’s relevant altruistic considerations,many of which are also moral considerations anyway.20

A Non-question-begging ArgumentIn trying to determine how we should act, let us assume that we wouldlike to be able to construct a good argument favoring morality overegoism, and given that good arguments are non-question-begging, weaccordingly would like to construct an argument that does not beg thequestion.21 So instead of trying to justify morality on grounds of consis-tency alone, as Gewirth and Korsgaard tried to do, I propose that we alsoappeal to the principle of non-question-beggingness. The principle ofnon-question-beggingness requires that we do not argue in such a waythat only someone who already knew or believed the conclusion of ourargument would accept its premises, or put more succinctly, that we notassume what we are trying to prove or justify.Now the question at issue here is what reasons each of us should take as

supreme, and this question would be begged against egoism if we pro-posed to answer it simply by assuming from the start that moral consider-ations are the reasons for action that each of us should take as supreme. Butthe question would be begged against morality as well if we proposed toanswer the question simply by assuming from the start that self-interestedconsiderations are the reasons for action that each of us should take assupreme. This means, of course, that we cannot answer the question ofwhat reasons we should take as supreme simply by assuming the generalprinciple of egoism:

Each person ought to do what best serves his or her overall self-interest.

We can no more argue for egoism simply by denying the relevance ofmoral considerations to rational choice than we can argue for altruism

20 It should also be noted that not all the considerations that people are capable of actingupon are relevant to an assessment of the reasonableness of their conduct. First, considerationsthat are evocable only from some merely logically possible set of opportunities are notrelevant; they must be evocable from the opportunities people actually possessed. Second,considerations that radically different people could have acquired are also not relevant.Instead, they must be considerations that people could have acquired without radical changesin their developing identities.

21 I am indebted to Jamie Dreier for a valuable discussion of this stage of my argument.


simply by denying the relevance of self-interested considerations to ra-tional choice and assuming the following general principle of altruism:

Each person ought to do what best serves the overall interest of others.22

Consequently, in order not to beg the question, we have no alternativebut to grant the status of prima facie reasons for action to relevant self-interested and moral or altruistic considerations and then try to determinewhich of these reasons for action we would be rationally required to actupon, all things considered. So hereafter I will be referring to theserelevant self-interested and altruistic considerations as reasons for action.Notice that in order not to beg the question here, it is necessary to back offboth from the general principle of egoism and from the general principleof altruism, thus granting the prima facie status to both relevant self-interested and moral or altruistic reasons for action. This is the only wayone can argue here such that we do not already assume what we are tryingto prove. From this standpoint, it is still an open question, whether eitheregoism or altruism will be rationally preferable, all things considered. Thestance we are taking here does, in effect, give considerations in favor ofone’s self-interest and considerations in favor of the interests of others thestatus of prima facie reasons, although either can still be ruled out asreasons, all things considered.

Psychological EgoismOf course, someone might try to defend egoism by claiming that it ispsychologically inevitable that despite appearances to the contrary, we allactually behave self-interestedly all the time. This is the thesis of psycho-logical egoism, and it is a surprisingly strong thesis.No doubt, we are well aware that some people we assumed were acting

primarily for the good of others later turn out to be significantly motivatedby self-interest. For example, a number of years ago, many people weresurprised and dismayed to learn of the very large salary being paid toWilliam Aramony, then president of United Way of America. After

22 I understand the pure altruist to be the mirror image of the pure egoist. Whereas thepure egoist thinks that the interests of others do not count for herself, except instrumen-tally, the pure altruist thinks that her own interests do not count for herself, exceptinstrumentally.


Aramony was pressured to resign and later convicted of fraud, a significantdrop in giving to United Way persisted for some time, with some peoplethinking United Way was no longer the type of public service organiza-tion they had once thought it to be.23

Psychological egoism, however, does not mainta

from rationality to equality · from rationality to equality james p. sterba 1. 3 great clarendon...

Documents