Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought 0198847874, 9780198847878

Citation preview

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

Facing Up to Scarcity

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

Facing Up to Scarcity The Logic and Limits of Nonconsequentialist Thought BARBARA H. FRIED

1

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

3

Great 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 of Oxford University Press in the UK and in certain other countries © in this volume Barbara H. Fried 2020 The moral rights of the author have been asserted First Edition published in 2020 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2019951998 ISBN 978–0–19–884787–8 Printed and bound in Great Britain by Clays Ltd, Elcograf S.p.A. Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

For my boys

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

Contents Preface Acknowledgments

1. Introduction: Facing Up to Scarcity: The Limits of Nonconsequentialism

ix xiii

1

2. Facing Up to Risk

22

3. What Does Matter? The Case for Killing the Trolley Problem (or Letting it Die)

40

4. Can Scanlonian Contractualism Save Us from Aggregation?

58

5. Tortious Harms

84

6. Can Contractualism Be Saved?

106

7. Is Nozick a Libertarian?

129

8. Rawls, Risk, and the Maximin Principle

149

9. The Unwritten Theory of Justice: Rawlsian Liberalism versus Libertarianism

155

10. Left-Libertarianism

176

11. Wilt Chamberlain Revisited: Nozick’s Justice in Transfer and the Problem of Market-Based Distribution

197

12. “If You Don’t Like It, Leave It”: The Problem of Exit in Social Contractarian Arguments

211

13. The Case for a Progressive Benefits Tax

235

Bibliography Index

251 263

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

Preface In the late 1990s, I published a book on the intellectual history of the Progressiveera law and economics movement.¹ The movement, which sprang up in response to the rise of laissez-faire constitutionalism in the US in the late 1880s, remained an important force in legal and economic thought through the late 1930s, when the Supreme Court, faced with Roosevelt’s threat to pack the Court, abruptly changed course. The great divide in legal, economic, and philosophical thought in those years was, as it is now, between those who believed that the primary obligation of the state was to advance the aggregate welfare of its citizens, and those who believed that individual rights significantly constrain the state’s legal or moral entitlement to do so. In the course of untangling the disagreements in a historical context, I became interested in the arguments in their own right. I went on to write several articles critiquing the internal logic of contemporary libertarianism. While doing so, I recognized that many of the problems I had identified in libertarianism were shared by left-leaning versions of nonconsequentialism that have dominated Anglo-American moral and political philosophy since the late 1960s. At a certain point, I decided that it was in intellectual bad faith not to say so, just because the policy conclusions of the latter were politically more congenial to me than those of the former. Over the next several years, I published a number of articles critiquing the logic of Rawlsianism, Scanlonian contractualism, trolleyology, social contractarianism, and the corrective justice literature in torts. Ten of the chapters in the book have previously been published in different forms. I welcome the opportunity to publish them together, along with three new chapters, because the central ideas that animate them are tightly interconnected, and the whole, I believe, is greater than the sum of its parts. I have edited all of the previously published Chapters, some significantly, to eliminate redundancies and make explicit the connections between different Chapters and different literatures. In the two decades since the first of the articles included in the book was written, I have moved much closer to full-blown consequentialism myself, at least with respect to matters of public policy. This is in part because I don’t believe there is a plausible alternative to address most of the collective problems we face, and in part because I find the moral imperative of welfarism more compelling than many of the moral imperatives driving nonconsequentialist thought. But I do not mean

¹ The Progressive Assault on Laissez Faire (Harvard University Press, 1998).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

x



the book to be an argument for consequentialism by default. Like any theory, it will have to stand or fall on its own merits. I have chosen to focus on the internal problems with nonconsequentialism for two reasons. First, in my view, that is where the biggest gaps in contemporary moral and political philosophy are. Since the resurgence of nonconsequentialist moral and political philosophy in the 1960s, an enormous amount has been written about normative disagreements between consequentialists and nonconsequentialists, and among nonconsequentialists themselves. Relatively little has been said about two problems that are common to all nonconsequentialist theories, neither of which directly engages normative disagreements. One is posed by moral psychology, the other by the internal logic of nonconsequentialism itself. A large and growing empirical literature in moral psychology has documented the irrationality, manipulability, and non-generalizability of many of the ‘moral’ intuitions that drive lay nonconsequentialist thought. Those findings pose a central challenge to nonconsequentialist philosophers, who share many of those intuitions: to what extent do those intuitions reflect moral reasons, as distinct from psychological or emotional ones? Addressing that challenge is, I believe, critical to the future vitality of nonconsequentialism as a distinctly moral enterprise. But aside from a few passing remarks, I do not explore it here. The problem I take up is with the internal logical coherence and operationalizability of nonconsequentialist principles. The near-hegemony of nonconsequentialist thought in Anglo-American moral and political philosophy over the past fifty years has insulated its key tenets—above all, that it is immoral to trade off one person’s most fundamental interests for another’s—from critical examination, not as to its normative appeal but as to its logical coherence in a world of scarcity. With respect to moral philosophy, the subject of the first part, I argue that nonaggregative principles cannot, as a logical matter, solve the central challenge facing all forms of social cooperation: how to reconcile the prima facie legitimate but conflicting interests of different people. I argue that the voluminous literature on tragic choices amassed over the past half-century has deflected that challenge by focusing on stylized hypotheticals that typically involve pairwise comparisons, closed classes of claimants, and determinate outcomes. Those restrictions, rarely met by real-world social choices, make nonaggregative principles appear to go through, and with results that appear to be morally acceptable. Relaxing any of those restrictions will drive nonconsequentialism closer to aggregation. Relaxing them all will drive it to aggregation full-stop. To put it another way, the plausibility of nonconsequentialism as a set of moral principles with general application depends upon our not seeing the world whole, in ways that are antithetical to the most basic commitments of nonconsequentialism itself. With respect to political philosophy, the subject of the second part, I argue that all nonconsequentialist philosophies, from Rawlsianism to libertarianism, necessarily depend

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi



xi

on aggregation to go through. In addition, I suggest, most are pulling some very thick conclusions out of very thin and often undefended premises. The second reason I have chosen to focus on the internal problems with nonconsequentialism is a pragmatic one. Lay moral intuitions are overwhelmingly nonconsequentialist. Anglo-American moral philosophy, which has been dominated by nonconsequentialist thought since the 1970s, shares many of those intuitions. In the policy space, those intuitions have backfired badly. An examination of nonconsequentialism in its more formalized, academic, form I hope will help elucidate why such intuitions, however appealing they may seem in the abstract, cannot provide solutions to the hard policy questions we actually face.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

Acknowledgments Over the many years that the chapters in this book were in the making, I have accumulated a substantial number of debts. It is a pleasure to be able to acknowledge them here. My memory is not up to the challenge of naming all the people who have been important interlocutors on different chapters. With apologies to those whom I have omitted, I would like to thank the following: Joseph Bankman, Sam Bankman-Fried, Charles Beitz, Paul Brest, Joshua Cohen, Dick Craswell, John Donohue, Ronald Dworkin, Johann Frick, Steven Goodman, Bob Gordon, Tom Grey, Dirk Hartog, Dan Ho, Aaron James, Frances Kamm, Mark Kelman, Lewis Kornhauser, Rahul Kumar, Chris Kutz, Loren Lomasky, David Luban, Rob MacCoun, Jon Mandle, Liam Murphy, Thomas Nagel, Martin O’Neill, Shepley Orr, Derek Parfit, Philip Pettit, David Reidy, Arthur Ripstein, Debra Satz, Seana Shiffrin, Bill Simon, Hillel Steiner, Peter Vallentyne, Leif Wenar, Allen Wood, and Sharif Youssef. A number of the foregoing disagree with much of what is written here, and I thank them for their willingness to engage seriously with the project nonetheless. Thanks also go to Ethan Herenstein for his stellar job copy-editing the manuscript. Two chapters in the book were written during my time as a Fellow at the Center for Advanced Study in the Behavioral Sciences at Stanford in 2009–2011. The Center is a close to idyllic place to work, and I am grateful to Iris Litt, the staff of the Center, and Stanford University for making my time there possible and productive. Drafts of chapters in the book were presented at numerous faculty workshops, lectures, and conferences over the years. I am grateful to the organizers and participants for the opportunity to present my work and their generous engagement with it. An incomplete list includes faculty workshops at Toronto, Yale, Stanford, University of Pennsylvania, University of Michigan, Cornell, Harvard, Columbia, NYU, University of Arizona, Queens University (Kingston, Ontario), and University of San Diego law schools; Conference on “Reappraising ‘Anarchy, State, and Utopia’ ” at Kings College London; Symposium on Human Rights and Welfare Economics at Harvard University; NYU Colloquium in Law, Philosophy and Political Theory; Berkeley Political Theory Workshop; Stanford Political Theory Workshop; American Studies Workshop at Princeton; UC Irvine Legal Philosophy Colloquium; Center for Ethics and Public Affairs at Tulane University; Law and Philosophy Seminar at Harvard Law School; Philosophy Department, Queens University (Kingston, Ontario); Law and Philosophy Workshop at Georgetown Law Center; Safra Lecture at Harvard Kennedy School of

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

xiv



Government; Symposium on Morality and the Law at Universidad Torcuato Di Tella, Buenos Aires; Conference on Political Philosophy and Taxation at University College, London; Tobias Simon Distinguished Lecturer, Florida State Law School; Symposium on Political Theory and Private Law, University of Virginia; Law and Philosophy Workshop, University of Chicago; and the 2015 Bell Distinguished Lecture in Law at the College of Wooster. Ten of the chapters have been published previously in somewhat different forms. I thank the publishers and editors for permission to reprint revised versions here. Chapter 2 is forthcoming in Fried, B. H., “Facing Up to Risk,” Journal of Legal Analysis, Volume 10, published by Oxford University Press: https://academic.oup. com/jla. Chapter 3 is reprinted from Fried, B. H., “What Does Matter? The Case for Killing the Trolley Problem (Or Letting It Die),” Philosophical Quarterly, Volume 62, Issue 248, pp. 505-29, Copyright © 2012, doi: 10.1111/j.1467-9213.2012.00061. x by permission of Oxford University Press: https://academic.oup.com/pq. Chapter 4 is reprinted by permission from Springer Nature: Fried, B. H., “Can Contractualism Save Us From Aggregation?,” Journal of Ethics, Volume 16, Issue 1, pp. 39-66, Copyright © 2011, Springer Science+Business Media B.V., doi: 10.1007/s10892-011-9113-3. Chapter 5 is reproduced from Fried, B. H., “The Limits of a Nonconsequentialist Approach to Torts,” Legal Theory, Volume 18, Issue 3, pp. 231–62, Copyright © 2012, Cambridge University Press, doi: 10.1017/S1352325212000183, reprinted with permission. Chapter 7 is reproduced from Fried, B. H., “Does Nozick Have a Theory of Property Rights?” in: Bader, R. A. and Meadowcroft, J. (eds.), The Cambridge Companion to Nozick’s ‘Anarchy, State, and Utopia’, pp. 230–52, Copyright © 2011, Cambridge University Press, doi: 10.1017/CCOL9780521197762.013, reprinted with permission. Chapter 9 is reprinted with permission from Fried, B. H., “The Unwritten Theory of Justice: Rawlsian Liberalism versus Libertarianism” in: Mandle J. and Reidy D. (eds.), A Companion to Rawls, pp. 430–49, Copyright © 2014, John Wiley & Sons, Inc., doi: 10.1002/9781118328460. Chapter 10 is reprinted with permission from Fried, B. H., “Left-Libertarianism: A Review Essay,” Philosophy and Public Affairs, Volume 32, Issue 1, pp. 66–92, Copyright © 2004, by Blackwell Publishing, Inc., doi: 10.1111/j.1467-6486. 2004.00005.x. Chapter 11 is reprinted with permission from Fried, B. H., “Wilt Chamberlain Revisited: Nozick’s ‘Justice in Transfer’ and the Problem of Market-Based Distribution,” Philosophy and Public Affairs, Volume 24, Issue 3, pp. 226–45, Copyright © 1995, Wiley Periodicals, Inc., doi: 10.1111/j.1088-4963.1995.tb00030.x. Chapter 12 is reprinted with permission from Fried, B. H., “ ‘If You Don’t Like It, Leave It’: The Construction of Exit Options in Social Contractarian Arguments,”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi



xv

Philosophy and Public Affairs, Volume 31, Issue 1, pp. 40–70, Copyright © 2003, Wiley Periodicals, Inc., doi: 10.1111/j.1088-4963.2003.00040.x. Chapter 13 is reprinted from Fried, B. H., “The Case for a Progressive Benefits Tax,” in: O’Neill, M. and Orr, S. (eds.), Taxation: Philosophical Perspectives, pp. 147–66, Copyright © 2017, doi: 10.1093/oso/9780199609222.001.0001, by permission of Oxford University Press: https://global.oup.com/academic. I am grateful to Oxford University Press for their enthusiasm for the project, and to my editor, Peter Momtchiloff, for his support and patience. My biggest professional debt is to Stanford Law School and Stanford University, where I have been fortunate to make my intellectual home for the past thirty years. Stanford has given me the extraordinary luxury to follow my intellectual interests where they have taken me. My colleagues over the years have made the law school a very happy place to be, as have the several deans I have worked under. Special thanks go to Paul Brest and Larry Kramer, who were beaux ideals as deans and remain treasured friends. Finally, I would like to acknowledge a significant intellectual debt to Joe Bankman and our sons, Sam and Gabe. When Sam was about fourteen, he emerged from his bedroom one evening and said to me, seemingly out of the blue, “What kind of person dismisses an argument they disagree with by labelling it ‘the Repugnant Conclusion’?” Clearly, things were not as I, in my impoverished imagination, had assumed them to be in our household. Restless minds were at work making sense of the world around them without any help from me. In the years since, both Sam and Gabe have become take-no-prisoners utilitarians, joining their father in that hardy band. I am not (yet?) a card-carrying member myself, but in countless discussions around the kitchen table, literally and figuratively, about the subject of this book, they have taught me at least as much as I have taught them. More importantly, they have shown me by example the nobility of the ethical principle at the heart of utilitarianism: a commitment to the wellbeing of all people, and to counting each person—alive now or in the future, halfway around the world or next door, known or unknown to us—as one. This book is for all my boys: Joe, Sam, Gabe, and Matt.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/2/2020, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

1 Introduction Facing Up to Scarcity: The Limits of Nonconsequentialism

[Ideal Kantian legislators] recognize one another as ends in themselves, with dignity above price . . . [and] unconditionally attribute a worth to persons that cannot be quantified and is not subject to trade-offs.¹ [F]or everything, we have to give up something else, and we are taught to set the advantage we gain against the other advantage we lose, and to know what we are doing when we elect.² When John Rawls published A Theory of Justice in 1971, utilitarianism was the game to beat in political philosophy, and Rawls made no bones about his intention to beat it. In his words, Justice as fairness . . . offers, I believe, an alternative to the utilitarian view which has for so long held the preeminent place in our moral philosophy. I have tried to present the theory of justice as a viable systematic doctrine so that the idea of maximizing the good [meaning aggregate welfare] does not hold sway by default.³

Three years after the publication of A Theory of Justice, Robert Nozick’s Anarchy, State, and Utopia appeared, reinvigorating libertarianism as a serious contender in academic philosophy. For all their political disagreements, Kantians and libertarians are united in their opposition to utilitarianism, and in particular to its methodological commitment to allow the hardships of some to be offset by the greater good of others

¹ Thomas E. Hill, Jr., “A Kantian Perspective on Moral Rules,” Philosophical Perspectives 6 Ethics (1992): pp. 295–6. ² Oliver Wendell Holmes, Jr., “The Path of the Law,” Harvard Law Review 10, no. 8 (Mar. 1897): p. 474. ³ John Rawls, A Theory of Justice, Rev. ed. (Cambridge, Mass.: Belknap Press, 1999), p. 513. For similar sentiments from other prominent Kantians, see Philippa Foot, “Utilitarianism and the Virtues,” Proceedings and Addresses of the American Philosophical Association 57, no. 2 (Nov. 1983): p. 273; T. M. Scanlon, “Contractualism and Utilitarianism,” in Utilitarianism and Beyond, edited by Amartya Sen and Bernard Williams (Cambridge: Cambridge University Press, 1982), p. 103. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

2  through some form of aggregation.⁴ In Joseph Raz’s words, nonconsequentialism must stand or fall on “the ways in which numbers do and do not determine the outcome of moral conflicts.”⁵ For classic liberals or libertarians, opposition to aggregation is grounded in a stringent version of rights that presumptively trumps any benefits others would realize from their curtailment. To quote the most famous contemporary articulation of that view, “Individuals have rights and there are things no person or group may do to them (without violating their rights).”⁶ For Kantians, opposition to aggregation is grounded in the principle that every individual has a worth that cannot be sacrificed for the greater good. In Jay Wallace’s words, [I]t is virtually a paradigm of moral thought to allow the grave objections of a single individual to defeat a project or plan that would otherwise be expedient for many; a willingness to defer to the standpoint of individuals in such cases seems to be required by their standing as ends in themselves (as Kantians would put the point).⁷

The essays collected in this book take stock of the nonconsequentialist project in two key areas. The first, the subject of the first part, concerns the scope of our individual and collective duties not to harm (or duties to rescue) others.⁸ Under an appropriately broad definition of harm, that duty encompasses most of the restrictions imposed on individual choice and collective decision-making in the secular liberal state. For libertarians who embrace the minimal (nightwatchman) state, it encompasses them all. The lion’s share of attention in the nonconsequentialist literature over the past fifty years has been given over to devising an alternative to aggregation to resolve the scope of that duty. Contributors have focused almost exclusively on a small set of hypotheticals (trolley-type problems) in which the reader must choose between sacrificing A and sacrificing B. I examine that literature in detail in Chapter 3. But my larger interest is in the negative pregnant implied by nonconsequentialists’ ⁴ In his Foreword to the 2013 edition of Anarchy, State, and Utopia, Thomas Nagel traces the common origins and common assumptions that bound Nozick’s version of libertarianism to the Rawlsian strain of liberalism unfolding contemporaneously. Robert Nozick, Anarchy, State, and Utopia, Rev. ed. (New York: Basic Books, 2013), pp. xi–xviii. I return to this matter at length in Chapter 10. ⁵ Joseph Raz, “Numbers, With and Without Contractualism,” Ratio 16, no. 4 (Dec. 2003): p. 360. ⁶ Nozick, Anarchy, State, and Utopia, p. xix. ⁷ Jay Wallace, “Scanlon’s Contractualism,” Ethics 112, no. 3 (Apr. 2002): p. 469. ⁸ The “duty to rescue” poses many of the same problems as the “duty not to harm” (indeed, for those who do not attribute any moral weight to the distinction between act and omission, they are identical). For ease of exposition, I refer only to the latter, unless the “duty to rescue” is implicated in the hypothetical under discussion.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



3

obsessive attention to trolley-type problems: if we can figure out a way to save these freaks and sports of nature from aggregation, everything else will take care of itself. It won’t. Virtually all collective choices we make require us to trade off one person’s interests against another’s. That is the universal tragedy of the social world. The essentially optimistic premise on which nonaggregation rests—that tragic choices between the fundamental interests of different individuals are the exception and not the rule—cannot tell us what to do about it. In pressing the futility of nonconsequentialists’ opposition to aggregation, I do not mean to deny that nonaggregative principles can and do play a significant role in society. Whenever we refuse to give weight to certain kinds of preferences— of rapists to rape; of vigilantes to take justice into their own hands; of racists, homophobes, misogynists to refuse to hire qualified members of disfavored groups; of state actors to censor unwelcome speech; of the public to curtail a disfavored individual’s right to a fair trial—our motivation for doing so is typically nonaggregative, and the decision rule that results is at least partially nonaggregative. To put it in more conventional philosophical terms, rights can and frequently do operate as a partial side constraint on permissible tradeoffs.⁹ But the voluminous and still growing philosophical literature on harms to others has been concerned principally not with conduct that is deemed per se impermissible (rape, murder, abrogation of due process rights) but rather with conduct that is, in Johann Frick’s words, “intuitively permissible” but that imposes (a risk of) harm on innocent others.¹⁰ One of the principal sources of confusion in both the legal and philosophical literatures over the past half-century has been the failure to keep the two categories distinct. My argument in the first half of the book is limited to the latter. Under an appropriately broad definition of harm, the latter category includes almost every noncriminal activity we choose to engage in daily (e.g., driving, riding a bike, reroofing our house, practicing the piano in an apartment building without soundproofing, living a life that generates a non-zero carbon footprint). It also includes the vast majority of risk-regulation decisions routinely made by the government: setting standards governing pollution, occupational safety and health, product safety, nuclear energy, etc.; allocating budgets among competing areas of need and competing claimants. ⁹ I am setting to the side some deep questions about whether the scope of the rights we protect against utilitarian objections is itself determined by rule-utilitarianism. For present purposes, I am happy to stipulate that some fundamental norms in society can be explained and defended only in nonconsequentialist terms. In addition, every comprehensive moral theory, including utilitarianism, is vulnerable to the critique that it smuggles in other norms at critical junctures. For a helpful discussion of this point, see Mark Kelman, “Hedonic Psychology and the Ambiguities of ‘Welfare,’ ” Philosophy and Public Affairs 33, no. 4 (Sept. 2005): pp. 391–412. ¹⁰ Johann Frick, “Contractualism and Social Risk,” Philosophy and Public Affairs 43, no. 3 (Summer 2015): pp. 205–11.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

4  In the first part, I argue that nonconsequentialists have not only failed to articulate a viable alternative to aggregation in this domain; they are doomed to fail. By fail, I do not mean that the alternatives on offer are morally unattractive (although many strike me as such). I mean that they cannot logically solve the problems they are being asked to solve, because they bring the wrong tools to the job. To be clear, I am not suggesting that the duty not to harm others can never be cashed out in nonaggregative terms in this domain. Of course it can. Should we ever find ourselves faced with the choice whether to redirect a trolley from one track to another, thereby saving five lives at the cost of one, we can in fact decline to redirect it, numbers be damned. But the cases we can resolve without resort to aggregation are an infinitesimal fraction of the tragic choices we face every day. They clearly do not add up to a viable, systematic alternative to aggregation. Whether they add up to enough to justify the extraordinary attention that has been devoted to them over the past fifty years I leave to nonconsequentialists. Nonconsequentialists’ failure to see this, I argue, can be traced to several recurring features of the literature. The first is that the bulk of philosophical attention has been trained on a small set of carefully constructed hypotheticals that require us to choose between two alternatives that affect small and ostensibly closed classes of claimants, ignoring both the implications of scaling up the proposed solution to all like cases and the opportunity costs entailed in choosing either alternative. The second is that nonconsequentialists who bar aggregation at the front door frequently smuggle it in through the back, in defining the scope of a right or exceptions to a right. As a result, ostensibly nonaggregative principles are operationalizable only because they have an illicit assist from aggregation. Finally, and most significantly, the literature has effaced the problems that decisionmaking under uncertainty poses for all nonaggregative principles.¹¹ I take up the last of these issues first, in Chapter 2 (“Facing Up to Risk”). In the balance of the first part, I consider all three issues in the context of different parts of the nonconsequentialist literature. Chapter 3 (“What Does Matter? The Case for Killing the Trolley Problem (or Letting It Die)”) looks at the high-stakes tragic choices literature in philosophy (hereinafter “trolleyology”). Chapter 4 (“Can Scanlonian Contractualism Save Us from Aggregation?”) examines Scanlonian contractualism. Chapter 5 (“Tortious Harms”) looks at the corrective justice literature in tort law. Chapter 6 (“Can Contractualism Be Saved?”) looks at Scanlonian contractualists’ recent attempts to bring decision-making under uncertainty within the contractualist framework. ¹¹ In decision theory, “risk” refers to a gamble with known odds and “uncertainty” refers to a gamble with unknown odds and (often) an unknown range of outcomes. For the most part, the distinction is not relevant to my arguments here, and for rhetorical convenience, I generally use the terms interchangeably, to describe all situations in which the decision-maker does not know what the ex post consequences of a choice will be at the moment of choosing. Where the distinction is relevant, I note it.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



5

In the second part of the book, I turn to political theory, looking at the principal strains of nonconsequentialist thought over the past fifty years. In Chapter 7 (“Is Nozick a Libertarian?”), I start with Robert Nozick’s libertarian manifesto, Anarchy, State, and Utopia. Chapter 8 (“Rawls, Risk, and the Maximin Principle”) takes a closer look at Rawls’s derivation of his two principles of justice, with particular attention to the Difference Principle. Chapter 9 (“The Unwritten Theory of Justice: Rawlsian Liberalism versus Libertarianism”) examines the kinship between Rawls’s argument for his two principles of justice and Nozick’s argument for libertarianism, and the kinship of both to utilitarianism. Chapter 10 (“Left-Libertarianism”) looks at efforts, loosely collected under the banner of leftlibertarianism, to wed a libertarian conception of self-ownership to a radically egalitarian redistributive scheme. Chapter 11 (“Wilt Chamberlain Revisited: Nozick’s Justice in Transfer and the Problem of Market-Based Distribution”) examines Nozick’s case against taxation of market income. Chapters 12 and 13 take up social contractarian arguments on the right. Chapter 12 (“ ‘If You Don’t Like It, Leave It’: The Problem of Exit in Social Contractarian Arguments”) takes a critical look at the various contractarian thought experiments offered to argue against income redistribution by the state. Chapter 13 (“The Case for a Progressive Benefits Tax”) argues that benefits taxation, the traditional redoubt of opponents of redistributive taxation, could as plausibly be invoked to support it. A few points of clarification. First, throughout the essays in the book, I use the terms “nonconsequentialism” and “deontology” interchangeably to refer to any moral theory (neo-Kantian, modern-day contractualist, libertarian) in which individual prerogatives are taken to trump an aggregative solution in a substantial domain of human conduct. For my purposes, the differences among strains of nonconsequentialism are indeterminate or immaterial. I use the term “nonaggregation” to refer to any decision rule concerning harm to others that evaluates conduct at least in part without regard to the numbers of people put at risk or the probabilities that harm will result. To the extent that nonconsequentialists hold hybrid views, the only part of interest to me here is the part ostensibly governed by nonaggregative rules. The rest has already been ceded to aggregation. Second, with very few exceptions, the essays in this book offer an internal critique of nonconsequentialism. That is to say, I accept arguendo the foundational moral commitments of the various schools of nonconsequentialist thought I discuss. My concern is with the logical coherence of those commitments and how they get cashed out into rules for action. Third, to anticipate what I expect will be a common objection, I do not mean to argue by indirection for utilitarianism or any other form of aggregate consequentialism. Every comprehensive moral theory faces serious challenges in defending

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

6  its choice of foundational principles and in moving from foundational principles to operationalizable rules. Consequentialism is no exception. Accepting arguendo its goal of maximizing aggregate (expected) outcomes, a number of fundamental questions remain: What costs and benefits do we care about? What metric do we use to measure them? How do we commensurate unlike goods and bads? Do we have good reason to believe that our real-world measurements of expected outcomes are more reliable than noise? It may turn out, as many critics of consequentialism have argued, that cost/benefit analysis and other formalized methods of aggregation are sufficiently unreliable that we would be better off giving up any pretense to scientism and just muddling through, with some procedural safeguards in place.¹² I’m skeptical of that extreme position, but it is not my purpose here to argue against it. In leaving these questions to others, I do not mean to deny their importance. With respect to my argument in the first part, however, they are irrelevant. Whatever the challenges facing aggregation, the burden of my argument is that virtually none of the difficult social choices we actually face can be resolved without it. Not so with respect to the issues explored in the second part. Here, we have a meaningful choice between consequentialist and nonconsequentialist approaches to redistributive justice. I have chosen to focus my attention on the problems facing nonconsequentialism, in part because that is where my own interests have taken me, and in part as a counterweight to the contemporary literature in political theory. The near-hegemony of left-liberal nonconsequentialism in Anglo-American academic philosophy over the past fifty years has, regrettably, insulated key tenets of the literature from skeptical examination. Rawlsianism is perhaps the most notable example here. The publication of A Theory of Justice in 1971 occasioned several foundational critiques from utilitarian¹³ and libertarian¹⁴ perspectives. Most utilitarian critics have moved on to other issues. Not so libertarians, for whom Rawls remains an important irritant and foil. But with very few exceptions, neither set of critiques has ever been taken seriously by Rawlsians. In the absence of robust internal dissent, Rawlsianism has achieved normalscience status in an uncomfortably large portion of contemporary political theory.

¹² For support for a procedural solution, see Henry Richardson, “The Stupidity of the Cost Benefit Standard,” in Cost Benefit Analysis: Legal, Economic, and Philosophical Perspectives, edited by Matthew D. Adler and Eric A. Posner (Chicago: The University of Chicago Press, 2001). ¹³ See Kenneth J. Arrow, “Some Ordinalist-Utilitarian Notes on Rawls’s Theory of Justice,” Journal of Philosophy 70, no. 9 (May 1973); John C. Harsanyi, “Can the Maximin Principle Serve as a Basis for Morality? A Critique of John Rawls's Theory,” Journal of Philosophy 69, no. 2 (Jun. 1975); David Lyons, “Rawls Versus Utilitarianism,” Journal of Philosophy 69, no. 18 (Oct. 1972). ¹⁴ Nozick’s extended critique of Rawls in Part II of Anarchy, State, and Utopia is the best-known of these efforts. It is also, in my view, the one part of the book that stands up to critical reflection.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

,   

7

No foundational theory, whatever its merits, warrants that degree of unquestioned fealty. It takes nothing away from the historical and intellectual importance of A Theory of Justice to say that this state of affairs has done no favors for contemporary political theory, and that Rawls’s basic framework is long overdue for serious and skeptical reexamination by those who are working within it. I would argue that the same is true for the now-dominant ethos of personal responsibility and personal blame in both the noncriminal and criminal realms. And it is true for the vast nonconsequentialist literature on permissible harms to others, in view of the fact (of which I hope to persuade readers) that the literature is stuck in a logical dead end. My primary motivation for collecting these essays in one place is to make the case for such a reexamination, and to suggest some of the directions in which it might fruitfully go.

1. First, Do No Harm In a secular, liberal democracy, the primary task of any moral or legal regime is to manage scarcity. I mean scarcity not in its ordinary language sense, denoting some absolute level of deprivation, but in the sense in which economists use it: to denote any situation in which the demand for a good exceeds its supply, with the consequence that we cannot satisfy all competing demands for it.¹⁵ In this sense of scarcity, virtually all goods in society are necessarily scarce, either because the material resources needed to produce them are finite (e.g., health, product safety, national defense) or the social opportunities to enjoy them are finite (e.g., activities we wish to pursue that threaten to harm or otherwise interfere with the pursuits of others). If a good is not scarce in this sense (breathing the air around us, thinking our private thoughts), in a secular, liberal society, there is no reason—legal or moral—to regulate access to it. The central task for any moral or legal system is to decide how to trade off the conflicting claims of different individuals or groups to scarce social resources. Many of the tradeoffs we make are no doubt suboptimal, measured by any normative criteria. But lines must be drawn somewhere, and wherever we draw them, we will be putting an implicit price on human life and trading off one individual’s fundamental interests for others’, or for the lesser interests of the many. Libertarian and Kantian nonconsequentialism, both committed to the view that such tradeoffs are prima facie impermissible and generally avoidable, cannot tell us how to make them.

¹⁵ More precisely, it denotes any situation in which, if a good were free, people would consume more of it than is available. This qualification is meant to set aside the market solution to scarcity: raise the price of a good until demand no longer exceeds supply.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

8  How could nonconsequentialists have concluded otherwise? The answer, I suggest, is that they have misunderstood or effaced the critical junctures at which such tradeoffs routinely get made. The first of those junctures is in specifying the content of protected values. Nonconsequentialist arguments generally start with broad, foundational principles like self-ownership, property rights, autonomy, the right to be treated as “an end in oneself,” and then derive rules to govern particular decisions from those broad principles through a process of deduction, or reflective equilibrium, or intuition. In the case of harm to others, that conceptual approach typically yields a clear duty—do not harm others, do not cross their boundaries—from which exceptions are carved out for cases in which there are thought to be compelling interests on the other side. Implicit in that understanding of the world are two factual assumptions, both of which, I argue, are wrong. The first is that harm-in-fact is generally a one-way street—that is, in any case of harmful conduct, one party is causing harm and the other is being harmed. The second is that we can generally vindicate the right not to be harmed simply by prohibiting harm-in-fact. For legal academics, both assumptions were upended by the Legal Realist revolution in the first half of the twentieth century. Jettisoning a conceptual understanding of rights for a functional one, the Legal Realists argued that rights are best understood not as free-floating entitlements that attach to individuals, but as a complex network of reciprocal prerogatives and duties between individuals. Thus, when we say that Tom, by virtue of owning Greenacre, has the absolute right to strip-mine the land in a fashion that causes harmful minerals to leach out onto his neighbor’s land, what we are saying in operational terms is that Tom’s neighbors (along with the world at large) are legally powerless to stop him from doing so, even if the consequence of his actions is to make their own property less usable, at the extreme destroying all of its value. As Oliver Wendell Holmes, Jr., pointed out over a century ago, the reciprocal nature of harm holds true even when we all agree whose interests should be sacrificed—as for example when we must choose between Campbell’s desire to “gratif[y his] ill-will” by punching Rodriguez in the nose, and Rodriguez’s desire not to have his nose punched.¹⁶ In short, wherever we put the entitlement, we will as a matter of fact sacrifice one party’s interests for another’s. Fifty years after Wesley Hohfeld formalized this insight for legal audiences in his taxonomy of reciprocal prerogatives and disabilities, the economist Ronald Coase put the same point in economic terms in his famous 1960 article, “The Problem of Social Cost.”¹⁷

¹⁶ Oliver Wendell Holmes, Jr, “Privilege, Malice and Intent,” Harvard Law Review 8, no. 1 (Apr. 1894): p. 5. ¹⁷ Wesley Hohfeld, “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning,” Yale Law Journal 23, no. 1 (Nov. 1913): pp. 16–59; Wesley Hohfeld, “Fundamental Legal Conceptions as

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

,   

9

Thus, the determination of who wronged whom cannot be resolved by an unmoralized (factual) determination of who harmed whom. It can be resolved only on the basis of some normative commitments (implicit or explicit) that lead us to favor one side’s interests over the other’s. One unfortunate consequence of nonconsequentialists’ conflating factual harm and moral wrong is that they often fail to recognize that they have made a normative choice at all, let alone the grounds on which they made it. Second, when nonconsequentialists face up to the necessity of tragic choices, they generally do so in the context of hypotheticals that invite us to consider only a limited set of consequences: the harms and benefits to specified individuals from a one-off choice. As I suggested above, among the facts thereby effaced are the opportunity costs to offstage others from that same choice, and the implications of scaling up a decision rule to resolve all morally homologous cases. Is $10 million too much to spend to save Baby Jessica’s life? The paradigmatic moral thought that “human beings have dignity and not mere price” and that the ideal Kantian legislator “unconditionally attribute[s] a worth to persons that cannot be quantified and is not subject to trade-off”¹⁸ seems to require that we answer that question in the negative. But before we do, wouldn’t it be relevant to know how many other $10 million rescues the same principle would compel us to fund, or how many more lives could be saved if we redirected the $10 million to pay for prenatal care for the poor, vaccinate a million children in the developing world against a deadly infectious disease, or countless other measures that would save far more lives less newsworthy than Baby Jessica’s. In a world without scarcity, we wouldn’t have to choose; we could save everyone. But in the world in which we actually live, we have no choice but to choose among different lives to save. As long as moral reasoning proceeds by examination of one-off hypotheticals in isolation from the rest of the world, those questions will never be asked, let alone answered. The third critical factor effaced from much of the nonconsequentialist literature is decision-making under uncertainty (risk). In the typical “tragic choices” hypotheticals that populate the literature, we are asked to choose between two actions, the outcomes of which are stipulated with certainty ex ante. May we take out an enemy munitions factory, knowing that in doing so, we will kill 1,000 innocent civilians? May we (must we) save one person from a serious disease, at the cost of curing headaches for ten million? May we (must we) deprive a hundred million viewers of the World Cup of twenty minutes of viewing pleasure in order to save one person from temporary, extreme pain?

Applied in Judicial Reasoning,” Yale Law Journal 26, no. 8 (Jun. 1917): pp. 710–70; R. H. Coase, “The Problem of Social Cost,” The Journal of Law and Economics 3 (Oct. 1960): pp. 1–44. ¹⁸ Hill, “A Kantian Perspective on Moral Rules,” pp. 292, 296.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

10



As I discuss in Chapter 2, nonconsequentialists are hardly alone in relegating risk to the margins of moral inquiry. The singular focus on actions with determinate consequences has occasioned surprisingly little comment from philosophers themselves. Indeed, it is less an argumentative assumption of moral philosophy than a near-universal and unexamined practice. Derek Parfit, one of the few contemporary philosophers to address it, suggests that choice under certainty and choice under uncertainty pose “quite different moral problems.” While the former “is a central part of moral theory,” he argues, the latter is akin to a “nonmoral decision” that can be relegated to pragmatic considerations.¹⁹ Few nonconsequentialists have explicitly endorsed the view that determinate harms are a different or higher moral kind than probabilistic harms. But the casualness with which many have ceded decision-making under uncertainty to aggregation suggests it is widely shared. It is impossible to overstate the significance of this omission for the nonconsequentialist enterprise. Once we are choosing under conditions of epistemic uncertainty, any rational decision rule requires that we discount consequences by their probability (frequency) of occurring. In short, it requires that we let the numbers count. As I discuss in Chapter 2, every action imposes some risk of grave harm to others, however unlikely such harm is to come to pass, and no action is certain to eventuate in harm. If the numbers do not count, nonconsequentialists are left with two unacceptable alternatives. They can treat any risk of grave harm to another, however small the risk or however great the potential offsetting benefits, as impermissible, leading to moral gridlock—a result all nonconsequentialists acknowledge to be unacceptable. Alternatively, they can treat only those actions that are certain to eventuate in serious harm as impermissible. This alternative leads to moral gridlock as well, provided we measure harms on a population-wide basis and treat statistical harms as certain harms: if repeated enough times, most actions are statistically certain to inflict serious harm on at least one person. Typically, however, nonconsequentialists equate actual harms with completed harms. Under that interpretation, a rule that forbids only those actions that will actually harm others licenses us to act as we wish until it is too late to avoid their bad consequences. There is of course a third alternative: distinguish between permissible and impermissible impositions of risk based on the level of expected harm, taking into account the seriousness of potential harms, their probability of eventuating, and the size of the at-risk population. But this alternative is foreclosed to nonaggregationists, who have refused to let the numbers count. Faced with this impossible choice, many prominent nonconsequentialists on the left and the right have conceded that nonconsequentialism lacks the internal

¹⁹ Derek Parfit, On What Matters, vol. 1, (Oxford: Oxford University Press, 2011), p. 162.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

,   

11

resources to deal with risk, implicitly or explicitly ceding decision-making under uncertainty to aggregation. Given that all decisions in the real world are made under conditions of epistemic uncertainty, this is no small concession. How then have nonconsequentialists concluded that nonaggregative principles can play a significant role in regulating risk? The answer, I argue in the first part, can be traced to four recurring features of the literature.

1.1 Conflating Ex Ante and Ex Post Epistemic Perspectives and the Problem of Hindsight Bias If we know with certainty what the consequences of available options will be at the time we must choose among them, an ex ante epistemic point of view (POV) collapses into an ex post one. That is the fanciful world presupposed by trolleyology.²⁰ Whenever different information is epistemically available to us ex ante and ex post—the case whenever we must choose under conditions of epistemic uncertainty—the two POVs diverge. When they diverge, the appropriate POV from which to evaluate actions or institutional arrangements depends on the question we are trying to answer. At one extreme, we can evaluate choices from an ex ante POV of radical uncertainty both as to the identity of the winners and losers and the extent and nature of the wins and losses. This is the epistemic POV enforced by the thick veil of ignorance adopted by Rawls—an appropriate one if the question on the table is: What institutional arrangements would people adopt were they evaluating choices from a purely impersonal POV? At the other extreme, we can evaluate choices from an ex post POV, from which perspective we know the consequences of a given choice, including the numbers and identities of winners and losers and the extent and nature of their wins and losses—an appropriate POV if the question on the table is, should we remediate the unfortunate outcomes of prior choices, and if so, how? Utilitarians and other consequentialists are generally committed to some version of an ex ante epistemic POV. That is to say, they evaluate proposed rules and actions in light of the best information available about expected consequences at the time of choice. Nonconsequentialists, in contrast, have staked their ground at both extremes and almost every epistemic POV in between. In general, however, they have opted for an ex post POV—deliberately in some cases, unconsciously in others. Some of the epistemic POVs nonconsequentialists have adopted seem apt for the moral problem under consideration. Many do not. Once again, it is hard

²⁰ It is also the very real world of statistical certainty, although, as I suggest above, focusing on that version of certainty is not going to help the nonconsequentialist cause, as it lands us back in moral gridlock.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

12



to overstate the significance of this mismatch. Judged from a normatively defensible epistemic POV, some of the hardest problems that have transfixed nonconsequentialists turn out to be non-problems, and many of the solutions nonconsequentialists have offered to real problems turn out to be non-solutions. In the former camp, I would put (inter alia) the “paradox” of moral luck, as well as the widespread puzzlement about how risky conduct that has not yet resulted in harm can be judged wrongful. In the latter camp, I would put Scanlon’s Greater Burden Principle, discussed in Chapter 4, and the massive literatures in both philosophy and law that assess harms to others through the lens of Kantian duties or libertarian rights. Other nonconsequentialists have not merely assumed that if an action injures others, the action must somehow have been wrongful; they have argued forcefully for that backward inference. As I discuss in Chapter 5, the argument is succinctly captured in the paradoxical imperatives that populate the Kantian literature on harm to others, to the effect that one has a duty not just to try not to harm others, but to succeed. If nonconsequentialists really mean what they say, in a world of epistemic uncertainty the imperative amounts to hindsight bias raised to high moral principle. It would be an exaggeration to say that nonconsequentialism reduces to hindsight bias. But for the substantial portion of the literature addressing harm to others, it is not that much of an exaggeration. In Chapters 3, 4, 5, and 6, I look more closely at the confusion between ex ante and ex post POVs in different strands of nonconsequentialism. Chapter 3 looks at the vast trolleyology literature. As suggested above, for the most part this literature has dealt with decision-making under uncertainty by pushing it offstage, focusing instead on cases in which the ex ante and ex post epistemic POVs coincide (that is, where ex post outcomes are stipulated with certainty ex ante). The consequence is to embed hindsight bias in the “facts” themselves. Chapter 4 looks at Scanlonian contractualism. Scanlon, committed to a thicker version of the self than Rawls, populates his thought experiment with representative persons who know the different preferences and circumstances they are endowed with, and hence who face, ex ante, different expected consequences from any proposed decision rule. But having constructed an ex ante epistemic POV, Scanlon immediately abandons it, measuring the burden imposed on each representative person not by the expected outcome to her of a given alternative, but instead by what, peeking ahead, we ascertain to be the “weightiness of the burdens . . . for those on whom they fall.”²¹ (To put it another way, complainants are grouped by ex post outcomes, not ex ante prospects.) So, a principle that

²¹ T. M. Scanlon, What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998), pp. 236–7 (emphasis added). Derek Parfit adopts essentially the same position in “Justifiability to Each Person,” Ratio 16, no. 4 (Nov. 2003): pp. 368–90.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

,   

13

would be rational for a self-interested person to choose ex ante, based on everything she knows about her situation (including the probabilities of different outcomes), she may reasonably reject once she is endowed with the knowledge that she will turn out to be the big loser after all. Scanlon’s resulting decision rule not only does not reflect what his imagined reasonable persons would have agreed to ex ante under conditions of uncertainty; it cannot be implemented in the face of uncertainty. When confronted squarely with that reality in the examples he poses, Scanlon retreats to aggregation. In more recent writings, Scanlonian contractualists (and Scanlon himself) have conceded that in a world of uncertainty, an ex post epistemic POV cannot supply a guide to action, and have abandoned it in favor of a (largely) ex ante one. I examine the new ex ante contractualism in Chapter 6. While it avoids the logical impossibilities of the original version, I argue that it faces a number of serious, if not fatal, problems, some carried over from ex post contractualism and some of its own making. Chapter 5 looks at the corrective justice literature in torts, which comes closest to articulating the literal, paradoxical interpretation of “act so as not to injure others”: you have a duty to act in a fashion that will not cause X, with the corollary that if your actions do cause X, we can say in retrospect that you should have acted differently. The paradox is often invisible to its proponents, because it is viewed through the lens of hindsight bias, defanged sub silentio by converting the duty not to cause X into a duty to compensate the victim for X, or by qualifying the “duty not to harm” in a fashion that slips aggregation in through the back door.

1.2 Resort to Hybrid Models Another common strategy for dealing with risk has been to develop hybrid models that split the difference between aggregation and nonaggregation. One frequently suggested division is that we resolve choices under conditions of epistemic certainty with nonaggregative principles, and turn choices under uncertainty over to aggregation in some form.²² Scanlon, Derek Parfit, Michael Otsuka, and Jules Coleman, among many others, have proposed versions of this solution. Peoples’ psychological tendency to treat a probability of 1.0 and a probability of .95 as discontinuous has been well-documented in the psychological literature, but the normative justifications for doing so are not obvious. Justified or not, as I suggest above, this “solution” saves nonconsequentialism by rendering it largely irrelevant.

²² See Chapters 3, 4, and 5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

14



Others have suggested splitting the difference through some form of threshold deontology: ignore the numbers until they get too large, at which point revert to aggregation.²³ Whatever the normative arguments for and against threshold deontology in general, like Scanlon’s Greater Burden Principle it cannot be applied to decision-making under uncertainty, where every choice has the possibility of producing unacceptable consequences and none is certain to. In such a case, we cannot assess the relative intolerability of outcomes without first accounting for the probability (frequency) of harm. At that point, we have let the numbers count in order to determine whether the numbers should count. Others have proposed splitting the difference based on whether the decision is an individual or a collective one: Individuals, choosing how to act, are strictly bound not to trade off the costs to one person against the benefits to another, but the state is free to aggregate in setting general rules for action. Government-house utilitarianism has its virtues (although the most obvious ones are utilitarian in nature). But why would we want to reach different answers to the identical question—for example, how fast may people drive through a crowded downtown?— depending on whether the state sets the speed limit for us or instead delegates the decision to individual drivers under a mandate to use “reasonable precautions”? As I discuss in Chapters 2 and 3, one explanation is that many nonconsequentialists believe that the risks posed by a given action differ, depending on whether we are evaluating a single act or a large class of such acts. The risks posed are identical; they are simply expressed in different terms. If a risk is assessed over a large population, as the government typically does (for example) in setting speed limits, it is generally stated in frequentist terms: for example, raising the speed limit from 55 miles per hour (mph) to 65 mph will result in one additional death for every million miles driven. If it is assessed with respect to a single car trip, it is generally stated in probabilistic terms: for example, if Maria drives 65 mph rather than 55 mph on her 10-mile drive to work today, the chance that she will kill someone increases by .00001 percent. Nonconsequentialists, however, have radically different intuitions about the right answer depending on which verbal formulation is used. If the incremental risk is expressed in frequentist terms and assessed over a large enough population that it produces at least one statistical death and if statistical deaths are treated as actual deaths, nonconsequentialists incline to the lower (55 mph) speed limit, because slightly inconveniencing millions of drivers seems a trivial price to pay to save a life. If the identical risk is expressed as the increased probability that Maria will kill somebody when she drives to work on Tuesday, nonconsequentialists view the higher (65 mph) speed limit much more favorably, because the .00001 percent incremental risk it would impose seems so trivial.

²³ See Chapter 4.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

,   

15

1.3 Relying on Secondary Factors to Make the Tradeoff When dealing with tragic tradeoffs (the realm of trolleyology), many of the canonical hypotheticals in the nonconsequentialist literature introduce factual differences, in addition to (the risk of) harm itself, that affect the average person’s instincts about the right choice. These include the act/omission distinction, whether we have specific rather than statistical foresight of the harm, and whether the identities of the ultimate victims are known. Many have questioned whether these additional factors sound in morality at all. I touch briefly on that question in the first part, but my primary concern is different. It is that the presence of these secondary factors confounds what exactly our intuitions are snagging on. If one removes them, leaving only the bare fact that the available choices all carry a risk of harm to unidentified others, many nonconsequentialists have no clear intuitions about the right answer, or conclude the problem must be turned over to aggregation. More recent contractualist literature has taken on the challenge of handling risk under nonconsequentialist principles. But as I discuss in Chapter 6, many of the decision rules that have been offered produce results that are indefensible from any normative vantage point, most of all one that seeks to treat each individual with equal respect.

1.4 Aggregation by Another Name When nonconsequentialist arguments appear to go through without counting the numbers, it is generally because the numbers have already been smuggled into the definition of rights or duties that cannot be traded off. The nonconsequentialist literature is filled with terms like “due care,” “reasonable precautions,” “good reasons,” “non-negligent conduct,” conduct in which the actor is not “at fault.” All of these terms purport to tell us what sorts of risky conduct it is permissible to engage in. But they are mere placeholders for criteria yet to be articulated. Is an epileptic driver whose seizures have been controlled by medication for thirty years exercising “due care,” acting “reasonably” or “non-negligently,” etc., when he sets out to drive on a given day? Is an auto manufacturer using due care when it designs an engine that has a one in a million chance of malfunctioning for every 500,000 miles driven? If not, why, and what more is required of the manufacturer? Until we answer these questions, what we have is a scheme that says, in effect, ex post harm to others gives rise to liability, except when it doesn’t. As I discuss in Chapters 3, 4, and 5, when nonconsequentialists have tried to operationalize words like “due care” or “negligence,” they have cashed them out in terms that differ little if at all from conventional cost/benefit analysis: sum up the aggregate (expected) benefits and costs to society from available alternatives, and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

16



pick the one that has the largest (expected) net benefit. The result is aggregation in two steps rather than one. Countless decisions are made every day that pose uncertain risks of potentially serious harm to unidentified others. Every time we get in a car, market a new drug, ride a bike down the street, decide whether or not to recall a product, or decide what level of precautions to take against earthquakes or potential terrorist attacks, we are taking our own and others’ lives into our hands. How we ought to go about making those decisions is of enormous social importance. If there is a viable alternative to aggregation, it needs to be put on the table so that it can become part of the public debate over risk regulation, and assessed relative to cost/benefit analysis and other versions of aggregation or a purely procedural solution. If (as I believe) there is not, nonconsequentialists have a responsibility to face up to that fact, and stop lending the imprimatur of morality to what amounts to nothing more than hindsight bias.

2. The Just State In the second part of the book, I turn to nonconsequentialist political theory, focusing on arguments for and against a redistributive state. In Chapter 7 (“Is Nozick a Libertarian?”), I start with an examination of Nozick’s libertarian manifesto, Anarchy, State, and Utopia. Nozick’s argument straddles the two parts of this book, with one foot in questions of individual morality and the other in political theory. This is hardly surprising, given the standard libertarian view that the principal job of the just (minimal) state is to protect individual rights from usurpation by others. Beyond that function, the powers of the just state are defined negatively: it may not interfere with individual rights, except perhaps to the extent necessary to fund its bare-bones operations (national security, courts, etc.), and it most emphatically may not engage in redistributive taxation. Nozick’s defense of individual rights, laid out in Part II of Anarchy, State, and Utopia, is in close conversation with the “duty not to harm” literature that I explore in Chapters 2 through 6. While his list of inviolable rights differs sharply from those of the nonconsequentialists discussed in earlier chapters, both arguments ultimately depend on the same mistaken conception of individual rights as largely nonoverlapping spheres, yielding the same logically erroneous conclusion that in protecting one person’s right (to be free from harm, to protect her property, to pursue her own life projects, etc.) we are rarely required to sacrifice another’s. In Parts I and III of Anarchy, State, and Utopia, Nozick turns his attention to state legitimacy and the reach of state power beyond its minimal nightwatchman function. Astonishingly, for the most prominent libertarian of our times, in Part I

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  

17

he abandons all bedrock libertarian principles, justifying the emergence of the minimal state on utilitarian grounds, and in Part III presents a Utopian vision of states as communities of majoritarian, collective self-determination. In Chapter 8 (“A Note on Rawls and Risk”), I take a closer look at Rawls’s arguments for using a leximin function (the Difference Principle) to distribute primary goods in society. In his rational choice mode, Rawls concedes that a leximin function would be rational only on the assumption that the parties to the social contract would be infinitely risk averse. He also concedes that assumption is generally implausible. But he famously argues it is rational in the unique circumstances posed by his Original Position. I challenge the reasons he gives in support of his argument, and suggest a deeper puzzle about his conclusion. Rawls’s veil of ignorance reduces the problem of social choice to the problem of individual choice under uncertainty. That is to say, whatever the representative person would choose in order to maximize her preferences is the same as what every other member of society would choose, yielding the aggregate utilitarian solution. Rawls acknowledges that his solution and the utilitarian solution will frequently converge. The question is, why would they ever diverge? I suggest that Rawls’s answer rests on a misunderstanding of the role of risk aversion in conventional utility functions. In the decades since Rawls’s A Theory of Justice and Nozick’s Anarchy, State, and Utopia appeared, a small cottage industry has sprung up dedicated to showing how little is required to turn Rawls into Nozick and Nozick into Rawls. In Chapter 9 (“The Unwritten Theory of Justice: Rawlsian Liberalism versus Libertarianism”), I offer my own contribution to the genre, identifying the various junctures at which, starting from the same core commitment to liberal individualism, Rawls and Nozick veer off in radically different directions. Something is driving these identical premises to opposite conclusions, but that something does not inhere in the concept of liberal individualism, and does not follow from either Rawls’s contractualist thought experiment or Nozick’s conception of rights as trumps. I offer another observation as well: how little is required to turn both Rawls and Nozick into utilitarians—indeed, how far they have already travelled down that road on their own. In Chapter 10 (“Left-Libertarianism: A Review Essay”), I look at the indeterminate relationship of left-libertarianism to right-libertarianism on the one side, and to egalitarianism on the other. Left- and right-libertarians both start with the Lockean postulates that (i) we own ourselves and by extension the fruits of mixing our labors with the commons (self-ownership), provided that (ii) we leave in the commons “as much, and as good” for others to use. From there, however, the two camps part company. Right-libertarians, reading self-ownership very broadly and the Proviso very narrowly—indeed, often reading it out of the picture entirely— not surprisingly conclude in favor of maximal private rights, and a minimal state in which virtually no distributive decisions are left to collective choice.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

18



Left-libertarians, reading self-ownership very narrowly and the Proviso very broadly—in some cases extending it to culture, language, and even the gene pool—not surprisingly conclude in favor of a distributive scheme at least as egalitarian as Rawls’s, and quite possibly more so. Once again, something is driving these identical premises to opposite conclusions, but that something is neither self-ownership nor the Lockean Proviso. Chapter 11 (“Wilt Chamberlain Revisited: Nozick’s ‘Justice in Transfer’ and the Problem of Market-Based Distribution”) takes up Nozick’s argument for an individual’s right to retain the full exchange value of her capital or labor, free from taxation or any other state encumbrances. Nozick rejects the standard libertarian argument that people have a right to retain the market value of their assets in virtue of the fact that it reflects their marginal contribution to society. In its place, he offers his principle of Justice in Transfer: if Susan, the owner of Goldacre, sells it to Carlos for $1 million, Susan is entitled to keep the $1 million because Carlos, who owned the $1 million, voluntarily transferred it to Susan. The argument is a familiar one in the context of gifts or bequests. But, I argue in Chapter 11, Nozick’s extension of it to market incomes is based on a false analogy. For a libertarian, the normative issue raised by gift transfers is whether Carlos has a right to give away what he owns without being subject to a transfer tax. Yes, says Nozick’s Justice in Transfer. The normative issue raised by market exchanges is different: whether Susan is entitled to keep, free from taxation, the appreciation in Goldacre’s value since the time she purchased it. Carlos’s willingness to buy Goldacre is irrelevant to resolving that issue. Chapter 12 (“ ‘If You Don’t Like It, Leave It’: The Construction of Exit Options in Social Contractarian Arguments”) and Chapter 13 (“The Case for a Progressive Benefits Tax”) take up social contractarian arguments from the right against the redistributive state. Like Rawls’s contractualism on the left, social contractarianism on the right is constructivist: the fair division of social surplus generated by political cooperation is whatever division would (could?) emerge from a fair bargaining process. As with Rawls, then, the key question for social contractarians is, what makes a bargaining process fair? In the realm of private contracts, libertarians are maximalists when it comes to protecting the personal endowments parties bring to the table, but minimalists with respect to every other aspect of the bargain. If you could have gotten a cell phone for half the price you paid at your local price-gouging electronics store had you only had transportation to a suburban mall or a computer to shop online, well, them’s the breaks. The seller was under no obligation to give you the benefit of a bargain you couldn’t obtain on your own. But when libertarians turn to the social contract, they routinely bleach out all of the costs the most fortunate members of society would realistically face were they to try to make good on their threat to take their marbles and leave. It is not obvious why the most

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

19

fortunate have a right to the benefit of a bargain that, in the real world, they could not negotiate on their own behalf. Why remove these background constraints on the hypothetical social contract, but leave all others—the distribution of natural talents and social privilege, background economic conditions, etc.—untouched? In Chapter 13, I focus on a real-world constraint on the wealthy’s ability to maximize their taxes-for-public-services deal that has received little attention in the philosophical literature. In geographically-defined polities, publicly provided goods come bundled not just with each other, but also with non-public goods tied to locale: friends, family, work opportunities, housing stock, aesthetics, culture, language, availability of quality health care, the weather. As a practical matter, polities that contain a disproportionate share of such amenities have the leverage to extract “voluntary” payments from wealthy taxpayers far in excess of either the cost of providing explicit public goods or their value to the individual taxpayer. Think Manhattan. If suppliers in the private market are entitled to leverage locational and other advantages, why not in the market for public goods? I suggest that were we to treat these real-world constraints on exit realistically when constructing the hypothetical bargaining situation, the social contractarian thought experiment could yield a wide range of distributive outcomes, but the libertarian ideal of a narrowly defined benefits tax is unlikely to be among them.

3. The Future of the Nonconsequentialist Project The essays in this book focus on the academic philosophical literature, and much of the analysis is internal to that literature. My ultimate concern, however, is not academic, in either sense of the word. Many of the nonconsequentialist principles examined here have implications for real-world policy-making that ought to give their proponents pause. Chief among them is the postulate with which I began: “[Ideal Kantian legislators] recognize one another as ends in themselves, with dignity above price . . . [and] unconditionally attribute a worth to persons that cannot be quantified and is not subject to trade-offs.”²⁴ That sentiment has no doubt inspired much good in the world. But many of the worst pathologies in contemporary American public policy can be traced, directly or indirectly, to our refusal to face up to scarcity and the ubiquitous tragic tradeoffs it necessitates. I include here our collective refusal to think or talk about the kinds of tragic tradeoffs between peoples’ most fundamental interests that we make daily, overtly, covertly, or by default; our inability to think systematically about risk—that is, about potential problems before they become actual ones with actual victims; our indifference to costs (in lives or

²⁴ Hill, “A Kantian Perspective on Moral Rules,” pp. 295–6.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

20



money) that are out of our immediate line of sight; our instinctual reaction, whenever a choice turns out badly, that the choice itself must have been wrong and the officials causally responsible for it morally to blame; the susceptibility of our moral intuitions to psychological framing, and with it our fatal attraction to governance by anecdote, which holds our judgments hostage to the manipulations of politicians and others with a strong interest in hijacking them. All of these lay reactions have close analogues in contemporary nonconsequentialist philosophy: the conviction that no action can be judged wrongful until it eventuates in harm; the conviction that certain harms and uncertain harms are different moral kinds, and that it is outside of—if not beneath—the job description of moral philosophers to think about the latter; the practice of honing moral instincts on one-off, bizarre hypotheticals in which the class of victims is closed and the available choices all have determinate consequences; and, most of all, the paradigmatic moral thought that we must “allow the grave objections of a single individual to defeat a project or plan that would otherwise be expedient for many.” The US is hardly alone in its pathological refusal to face up to the hard choices that scarcity necessitates, but among developed nations it sets the pace. That refusal permeates individual consciousness and politics, and severely constrains acceptable political rhetoric in this country. There is a reason, beyond mere rhetorical felicity, why Barack Obama’s 2008 presidential campaign slogan was “Yes We Can,” rather than “Sometimes We Can” or even “Usually We Can.” And there is a reason why, in defending himself against Republicans’ charge that the Affordable Care Act would result in Death Panels for Grandma, the one thing Obama couldn’t say is the one thing that is ineluctably true: however we fix the health care mess in this country, we are going to be rationing health care. If we refuse to pull the plug on Grandma, we will be pulling it on some other equally deserving but less visible individual. When the government gives in to the televised pleas of frantic parents to allow their 10-year-old daughter to jump the queue for a lung transplant, notwithstanding her lower probability of success, will any of the thousands of outraged viewers who lobbied the government on her behalf stop to wonder whom she jumped over, whether that person has children who love and need her, and whether she will now die instead? If nonconsequentialist philosophers do not mean to abet the disastrous realworld instantiations of their moral intuitions, I think they have some soulsearching to do. Philosophers may not have much influence on commonsense morality. But at the very least, they ought to be pushing back against its catastrophic failures, rather than encouraging and underwriting the habits of mind that produced them. This is strong language, but I think it is warranted. One final observation. After almost two decades of thinking and writing about the philosophical disagreements between utilitarians and nonconsequentialists, I am still unsure what the practical stakes of the disagreement are, at least from the perspective of Rawlsians and other Kantians in the left-liberal tradition. As long as

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

21

some form of instinctual “don’t tread on me” libertarianism remains a strong force in the American political landscape and leads to policy outcomes that are congenial to them, libertarians have no instrumental motivation to question the logical capacity of deontology to resolve the real-world problems we face. But the policy goals of Rawlsians and other left-liberal Kantians are, in my view, just a hair’s breadth from consequentialists’, properly understood.²⁵ Rather than making common cause with consequentialists, they have spent much of the last fifty years obsessing over their differences, real and imagined. For left-liberal Kantians, this seems to me a valuable opportunity squandered.

²⁵ For further discussion of this point, see Chapter 9.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

2 Facing Up to Risk In 1972, a new Ford Pinto stalled in the middle of a highway, causing the car behind to rear-end it. The force of the collision pushed the Pinto’s rear-mounted gas tank forward and into the differential housing, puncturing it. Fuel spraying from the punctured tank ignited inside the car, engulfing it in flames. One of the two occupants died from the resulting burns. The other, a 13-year-old boy, was severely disfigured for life, requiring multiple skin grafts over the following decade. In the lawsuit the heirs and survivors subsequently filed against Ford for defective product design, pretrial discovery revealed that before the Pinto ever went into production, Ford knew about the danger that a rear-end collision would explode the gas tank. Indeed, its own crash tests revealed the engine was vulnerable to puncture and explosion at crash speeds as low as 20 mph. Internal memos from Ford engineers to top management estimated that at a cost of $5 to $11 per car, the Pinto could be redesigned to decrease the probability of explosions significantly, saving an expected 180 lives. Ford management, estimating that the redesign would cost the company $137 million but save it only $49.5 million in projected damage awards from the incremental deaths, injuries, and property damage caused by the design defect, declined to make the change. The jury found against Ford, awarding the plaintiffs roughly $3 million in compensatory damages, and the boy an additional $125 million in punitive damages, an amount subsequently reduced on appeal.¹ In the years since, Ford’s conduct in the Pinto case has come to stand for all that is wrong with cost/benefit analysis, the decision procedure that Ford employed in deciding not to redesign the Pinto. Taking the generally received view of Ford’s behavior as accurate, it is easy to understand why Ford was widely regarded at the time as the poster child for bad corporate actors.²

¹ Grimshaw v. Ford Motor Co., 119 Cal. App. 3d 757, 808 (1981). ² It was alleged, among other damning facts, that Ford’s cost/benefit analysis was faulty; that the Pinto was significantly less safe than most other cars on the road, likely making its behavior legally negligent even without a faulty cost/benefit analysis; that the crash engineers at Ford strongly urged redesign, but were overruled by top management, which was notoriously indifferent to safety concerns; that Ford knew that federal safety regulations then pending would require it to protect the integrity of the gas tank in crashes up to 30 mph, but chose to wait until the regulations were adopted to comply with the standards; and that Ford withheld information about the risks for fear of bad publicity. In the years since the case was decided, some have questioned the accuracy of the received account. The best known of these revisionist accounts is Gary T. Schwartz, “The Myth of the Ford Pinto Case,” Rutgers Law Review 43, no. 4 (Winter 1991): pp. 1013–68. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

23

But the moral that many have drawn from the Ford Pinto case goes well beyond the particular facts of the case. Put simply, it is that trading off lives for any amount of money or other lesser goods is wrong. As we will see in subsequent chapters examining the contractualist, tort, and trolley literatures, in one form or another that sentiment underlies much of the vast literature on harm to others in Anglo-American moral philosophy over the past half-century. Its intuitive appeal is obvious. But in a world of epistemically indeterminate consequences—which is to say, the world in which we live—it cannot supply a general guide to action. This chapter is about why that is so, why nonconsequentialists continue to believe otherwise, and the social costs to all of us when the same belief asserts itself in politics and policy-making.

1. The Moral Terrain of Interpersonal Tradeoffs To elucidate how the nonconsequentialist literature has sidelined the problem of risk, I borrow Philip Tetlock’s psychological taxonomy of tradeoffs.³ Tetlock divides peoples’ interests into sacred and nonsacred values, which correspond (roughly) to the distinction in the nonconsequentialist literature between fundamental and nonfundamental interests. He identifies three different categories of tradeoffs, based on the interests implicated on each side: routine tradeoffs, taboo tradeoffs, and tragic tradeoffs. Routine tradeoffs are tradeoffs between two non-sacred values. Crucially, in the nonconsequentialist literature, such tradeoffs are taken to include not just those in which the underlying interests at stake on each side are nonfundamental (e.g., allocating tickets to an oversubscribed sporting event) but also those in which sacred values are implicated on one or both sides, but the probability of harm to those values is low (e.g., flight paths that cross over a residential neighborhood, thereby putting the population at risk in the highly improbable event of a crash). At least when judging actions from an ex ante perspective (that is, before we know whether the plane actually kills anyone), most nonconsequentialists have ceded routine tradeoffs to aggregation, either because they have concluded that the choice among the interests at stake does not pose a moral dilemma and hence can be handled by any noninvidious decision rule (including aggregation), or because they have reluctantly concluded that, in the case of decision-making under uncertainty, we have no choice but to aggregate. At the other extreme are taboo tradeoffs, which are tradeoffs between sacred and nonsacred values. Taboo tradeoffs describe the realm in which interpersonal ³ Philip Tetlock et al., “The Psychology of the Unthinkable: Taboo Trade-Offs, Forbidden Base Rates, and Heretical Counterfactuals,” Journal of Personality and Social Psychology 78, no. 5 (2000): pp. 853–70. I am grateful to Rob MacCoun for pointing me to Tetlock’s formulation.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

24

   

tradeoffs are forbidden on nonconsequentialist grounds, because the sacred value(s) on one side always trump the nonsacred value(s) on the other, irrespective of the numbers involved. T. M. Scanlon’s Greater Burden Principle rules out one set of taboo tradeoffs; libertarian property rights as absolute side constraints rule out another. Taboo tradeoffs are the principal redoubt of nonconsequentialism. As a result, nonaggregative principles have a significant role to play in regulating potentially harmful conduct only if the domain of taboo tradeoffs is itself large. The nonconsequentialist literature generally presumes that it is. In between routine tradeoffs and taboo tradeoffs lie tragic tradeoffs, which are tradeoffs between two sacred values (a life v. a life, a life v. ten lives, a life v. a lifetime of pain, etc.). May we kill one to save a hundred? May we privilege the survival of family members or friends over strangers? May we bomb enemy targets to protect innocent civilians, knowing that some number of other innocent civilians will be collateral damage? May we (must we?) prefer known victims to unknown ones? In T. E. Hill’s words, cases like these are “notoriously hard” to resolve by appeal to the principle that “one must not trade or sacrifice the dignity of anyone for ‘more dignity’ for others,” because both of our available choices will result in serious harm or death to one of the interested parties. In such a situation, we cannot “fully respect the dignity of all.”⁴ Over the past fifty years, an enormous amount of high-octane intellectual talent has been devoted to figuring out whether these hard cases can be resolved by nonaggregative principles, or instead have to be ceded to some form of aggregation. In Chapters 3, 4, and 5, I take up the different solutions nonconsequentialists have proposed for handling these carefully curated tragic tradeoffs. My concern in this chapter is with the negative pregnant that justifies the attention lavished on them: once nonconsequentialists figure out what to do with tragic tradeoffs, the other two categories will take care of themselves. (In the case of routine tradeoffs, we are permitted to aggregate; in the case of taboo tradeoffs, we are not.) This view of the moral terrain, I suggest, gets things exactly backwards. The high-stakes, hypothetical tragic choices that populate the nonconsequentialist literature—do we save one from death or 100,000 from paralysis—may be difficult to resolve without resorting to aggregation. But the tragic choices we actually face in life, in which the outcomes of available choices are all indeterminate, are impossible. As a result, far from offering a comprehensive alternative to aggregation, nonaggregative principles can operate only at the margins—indeed, arguably outside the margins entirely—of the real-world choices we face, relegating the literature on tragic choices to a luxury indulgence of our nonaggregationist instincts. Thus, nonaggregative principles in the domain of harm to others are viable in the minimal sense of operationalizable only as long as they are confined to a very ⁴ Thomas E. Hill, Jr., “A Kantian Perspective on Moral Rules,” Philosophical Perspectives. Vol. 6: Ethics (Chicago: University of Chicago Press, 1992): p. 293.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

25

small and rarified set of tragic choices. And they will seem to present a viable, systematic alternative to aggregation only as long as attention is confined to that set. In my view, that blinkered perspective describes most of the vast nonconsequentialist literature over the past fifty years on our duty not to harm (or duty to rescue) others. How then could it be that most nonconsequentialists (neo-Kantian, contractualist, and libertarian alike) have concluded otherwise? The answer, I suggest, lies in two features of the nonconsequentialist literature that have led philosophers to overestimate vastly the potential domain of nonaggregative principles. The first is the disproportionate amount of attention given to simple trolley-type tragic choices. As I discuss in more detail in Chapter 3, while such choices are notoriously hard—indeed, impossible—to resolve without trading off one person’s vital interests for another’s, in the stylized versions that populate the nonconsequentialist literature they can be resolved by reference to other nonaggregative criteria (a maximin decision rule, doing v. allowing, intending v. foreseeing harm, contributory negligence, identifiable v. nonidentifiable victims, etc.). They can, that is, if we do not inquire beyond the stipulated boundaries of the hypothetical. The second and in my view far more important explanation is that nonconsequentialists have failed to take seriously the problem of decision-making under uncertainty (risk).⁵ For a long time, nonconsequentialists ignored the problem of risk entirely. In recent years, a small number have turned their attention to it, but (as I argue in Chapter 6), have yet to show that nonaggregative principles offer a viable, systematic solution to it. That failure is no small thing. Outside of criminal conduct and (arguably) warfare, the vast majority of injuries to others are the accidental consequence of socially productive activities.⁶ The balance of this chapter is devoted to the problem of risk.

2. The Problem of Risk As Sven Ove Hansson has remarked, “Throughout the history of moral philosophy, moral theorizing has for the most part referred to a deterministic world in ⁵ Risk and uncertainty have different technical meanings. Roughly, risk refers to gambles with known odds, and uncertainty to gambles with unknown odds. For present purposes, that distinction is irrelevant, and for ease of exposition I use the terms interchangeably here. ⁶ Although criminal conduct often entails harms that, judged ex ante, are close to certain to result, it does not raise the problematic interpersonal tradeoffs that trolleyology is designed to help us think through, because we put no value on one side of the tradeoff: the liberty of individuals to engage in criminalized conduct. Problems of just warfare are arguably a closer real-world analogue to trolley problems. I take no position on the usefulness of trolleyology in thinking through moral dilemmas in this arena. For a skeptical view, see David Luban, “Risk Taking and Force Protection,” in Reading Walzer, edited by Yitzhak Benbaji and Naomi Sussman (New York: Routledge, 2014); David Luban, “Unthinking the Time Bomb,” in Global Basic Rights, edited by Charles R. Beitz and Robert E. Goodin (New York: Oxford University Press, 2009), pp. 181–206.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

26

   

which the morally relevant properties of human actions are both well-determined and knowable.”⁷ Given that the consequences of our actions are always epistemically uncertain ex ante, nonconsequentialists’ fixation on choices with specified, determinate outcomes seems to require some explanation. In his well-known article, “Conspiracy and the Paranoid Style: Causality and Deceit in the Eighteenth Century,” Gordon Wood supplies a historical one, which traces back to the seventeenth century. The mechanistic conception of the physical world produced by the Scientific Revolution, Wood argues, became the paradigm in which the enlightened analysis of all behavior and events now had to take place. Cause was something that produced an effect; every effect had a cause; the cause and its effect were integrally related. Such thinking created a new world of laws, measurements, predictions, and constancies or regularities of behavior all dependent on the same causes producing the same effects.⁸

Causal relationships that lacked such regularity and predictability, it was thought, did not deserve the name of science. In Samuel Stanhope Smith’s words, chance was “only a name to cover our ignorance of the cause of any event.”⁹ Applied to the realm of human action, the mechanistic story located the first mover in human intentions and motivations, from which the action and its consequences followed deterministically. In Wood’s words, Although the new science tended to remove man from the center of the physical universe, at the same time it brought him to the center of human affairs in ways that even classical and Renaissance thinkers had scarcely conceived of. It promised him the capacity to predict and control not only nature but his own society, and it proceeded to make him directly and consciously responsible for the course of human events.¹⁰

God still hovered in the background as the theoretical First Mover. But God was notably absent from the only causal story most people cared about: the story of cause and effect in everyday human affairs. ⁷ Sven Ove Hansson, “Risk and Ethics,” in Risk: Philosophical Perspectives, edited by Tim Lewens (London: Routledge, 2007), p. 30. ⁸ Gordon S. Wood, “Conspiracy and the Paranoid Style: Causality and Deceit in the Eighteenth Century,” The William and Mary Quarterly 39, No. 3 (Jul. 1982): p. 413. I am grateful to Sharif Youssef for bringing to my attention the relevance of Wood’s thesis to mine. ⁹ Samuel Stanhope Smith, Princeton Lectures on Moral and Political Philosophy, vol. II (Trenton, N.J.: 1812), pp. 22, quoted in Wood, “Conspiracy and the Paranoid Style,” p. 414. In Wood’s words, “Just as devout Puritans believed that nothing occurred without God’s providence, so the liberalminded believed that nothing occurred without some person willing it. Earlier, men had sought to decipher the concealed or partially revealed will of God; now they sought to understand the concealed or partially exposed wills of human beings.” Wood, “Conspiracy and the Paranoid Style,” p. 420. ¹⁰ Wood, “Conspiracy and the Paranoid Style,” pp. 412–13.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

27

The Enlightenment’s mechanistic account of cause and effect in the physical universe has given way over the past two centuries to more complex and probabilistic understandings of causation. But in the social world, the mechanistic account remains very much alive. In modern analytic moral philosophy, it is lodged in two foundational assumptions, the first explicit, the second not: our willing an action is the (morally relevant) cause of the action, and the consequences of our actions follow deterministically from the actions themselves. The first of these assumptions supplies the moral undergirding for the Kantian notion of a perfected will, and with it the congeries of condemnatory responses that reign supreme in contemporary American politics and common-sense morality: blame, retribution, guilt, etc. If we are not the cause of our own intentions—if they are caused by God, or social forces, or biological ones—then of what moral interest is the quality of our willing or the ends we seek to bring about? The Enlightenment view of the individual will as an uncaused cause may eventually buckle under mounting evidence to the contrary from neuroscience, genetics, etc., but it has so far managed to ward off similarly powerful evidence of social determinism.¹¹ My concern here is with the second of these assumptions: that the consequences of our actions follow predictably and deterministically from the actions themselves. In contrast to the first assumption, which has been debated endlessly in contemporary philosophical arguments over free will v. determinism and compatibilism v. incompatibilism, philosophers have said little about the second. Indeed, the central role given to actions with (ex ante) well-determined and knowable consequences is less an argumentative assumption of moral philosophy than a near-universal and unexamined practice. Those who have explicitly defended moral philosophers’ singular focus on conduct stipulated to have (ex ante) determinate consequences have generally done so on one of two mutually inconsistent grounds. The first is that the right way to handle risk can be derived from the right way to handle determinate harm, because the imposition of risk is itself a completed harm or other wrong, or because the wrongness of risk is parasitic on the wrongness of the harm it threatens. Hence, by resolving the latter we resolve the former as well.¹² The second is that choices made under conditions of epistemic certainty are a different moral kind from those made under conditions of epistemic uncertainty, and that only the former are the proper subject of moral theory. The latter are taken to implicate less lofty considerations. Both arguments for the proposition that risk is unproblematic, in my view, get things backwards. The moral principles nonconsequentialists have extracted from their extended engagement with trolley problems, far from solving the problem of ¹¹ For further discussion, see Barbara H. Fried, “Beyond Blame,” Boston Review, July/August 2013. ¹² I return to this argument in Chapter 5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

28

   

risk, are viable only as long as they are not extended to garden-variety risks. And if classic trolley problems present a difficult choice for nonconsequentialists, garden-variety risks present an impossible one. In On What Matters, Derek Parfit offers a rare, explicit defense of the “different moral kind” argument: In trying to answer [what acts are right and what wrong], it is best to proceed in two stages. We can first ask which acts would be wrong if we knew all of the morally relevant facts . . . . After answering these questions, we can turn to questions about what we ought morally to do when we don’t know all of the relevant facts. These questions are quite different, since they are about how we ought to respond to risks, and to uncertainty. As in the case of non-moral decisions, though these questions have great practical importance, they are less fundamental. These are not the questions about which different people, and different moral theories, most deeply disagree. Given the difference between these two sets of questions, they are best discussed separately. So I shall often suppose that, in my imagined cases, everyone would know all of the relevant facts. We can then ask what we ought to do in the simplest, fact-relative sense.¹³

This explanation raises more questions than it answers. Insofar as what is wrong with (potentially) harmful conduct is the (potential) harm itself, it is not apparent why “certain” and “uncertain” harms would raise “quite different” moral problems. All acts involve consequences that are (ex ante) more or less uncertain. Trolley problems, in which the consequences are stipulated to be known with certainty ex ante, simply represent the limit case at one extreme.¹⁴ Given that we are dealing with factually continuous phenomena, why would we think they raise morally discontinuous problems? Why is the difference between (say) a 95 percent probability and a 100 percent certainty as to any of the morally relevant facts (the likelihood that harm will result, the identity of the victim(s), etc.) a difference in kind rather than a difference in degree, and a very slight one at that? In addition, assuming that the two do present quite different moral problems, it is not apparent why we should regard the moral problems raised by uncertain harm as “less fundamental” than those raised by certain harm. As Parfit acknowledges, it cannot be due to their relative practical importance. While certainty is king in the world of trolley hypotheticals, in the real world the consequences of our acts are always uncertain ex ante. This is true even of harms that are intended (in the strong sense of desired or the weak sense of foreseeable). If I point a loaded gun at your head and pull the trigger, I am overwhelmingly likely to kill or seriously injure you, but I am not certain to do so. The gun could misfire; ¹³ Derek Parfit, On What Matters, vol. 1 (Oxford: Oxford University Press, 2011), p. 162. ¹⁴ I examine trolley problems in detail in the next chapter.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  

29

I could have forgotten to load it; the bullet could be deflected by a metal plate in your skull. So also, if I divert the trolley, I may believe I will thereby save five from certain death at the cost of one life, but I can never be certain, ex ante or (in many cases) ex post. Perhaps diverting the trolley will cause it to tip over before it reaches the one; perhaps if I had done nothing the five would have seen the trolley in time and moved out of the way. A fortiori, what is true of knowingly inflicted harms is true of accidental harms. Thus, from an ex ante perspective, the problem of all harm, accidental or not, is the problem of risk.¹⁵ In what sense then does risk present a “less fundamental” problem? Parfit’s comment suggests two possible answers. One is that the solution to the problem of risk cannot be of fundamental moral interest, because, as of the moment we must choose how to act, we lack a “morally relevant fact” bearing on the wrongness of the act: what its actual consequences will be. If this is the claim, it is problematic on a number of levels, but I focus on just one here. Whether the actual consequences of an act are in fact morally relevant to its rightness or wrongness needs to be established. If what we seek is an (ex ante) action-guiding norm, they decidedly are not. And until their relevance is established, the argument begs the question. (I return to this issue below.) Alternatively, Parfit can be read to suggest not that the problem of risk is unimportant, but that the answer is uncontroversial, at least as compared to trolley problems. (“These are not the questions about which different people, and different moral theories, most deeply disagree.”) If so, I am uncertain what he has in mind, but perhaps it is the fact that many nonconsequentialists have concluded that risk must be ceded to aggregation. But those who have given up on risk have not given up on nonconsequentialism, as they assume that a substantial domain of harm-producing conduct remains for nonconsequentialist principles to operate on, even after risk is removed. I turn to that assumption now.

3. Risk and Aggregation The problem that risk poses for nonaggregation is easily demonstrated. Suppose you are planning to drive to the movies. You always drive carefully (holding in suspension for the moment what exactly we mean by “carefully”). But you know that even driving carefully, there is a 1 in 50,000 chance that in the course of the trip you will seriously injure or kill someone. If you could know with certainty ¹⁵ As David Luban has noted, this is true as well of the moral dilemmas of war that are often taken to be the real-life analogues to trolley-type hypotheticals: whether it is permissible to torture others to obtain information that will (“certainly”) save innocent lives, or to kill enemy noncombatants to save the lives of one’s own combatants or citizens. See Luban, “Risk Taking and Force Protection,” and “Unthinking the Time Bomb.”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

30

   

that, of the tens of thousands of times that you are going to drive over your lifetime, today is the day when driving will result in serious injury or death to another, then you could simply stay home today. Doing so entails a personal sacrifice, to be sure—you’ll have to put off going to the movie. But it is hardly a tragic one, and it is one that aggregationists and nonaggregationists alike would make. This is the seductive fantasy world into which trolleyology invites us: a world in which all tradeoffs can be made by pairwise comparisons of the actual harms that will result, allowing us to ignore all the other instances of identical, risky conduct in which, as the odds would have it, no one ends up being harmed. But in a world in which you cannot know ex ante which (if any) of the tens of thousands of times that you drive over your lifetime you will end up injuring someone, the chance that you take with others’ lives today is the same chance you take every other day that you get in your car and drive somewhere. As a result, whatever we decide about the permissibility of your driving today will apply equally to all instances of driving. That reality leaves nonconsequentialists with two unacceptable alternatives, whether they are proceeding under the maximin decision rule adopted in Scanlon’s Greater Burden Principle or the typical libertarian prohibition on boundary crossing. If the mere possibility that an act will result in death or grave harm to others is sufficient to rule it out, then no act is permissible (except perhaps when necessary to save others from harm more severe than the proposed act threatens to impose).¹⁶ On the other hand, if the mere possibility that an act, undertaken with due care, could result in serious harm to others is not sufficient to rule it out—if harm must be absolutely certain to eventuate on this particular occasion—then in a world of indeterminate consequences, all actions are permitted until it is too late to avoid their consequences. Writ large, the same two unacceptable choices present themselves when we make the same moral calculus over a large number of identical acts—for example, in choosing rules of general application to govern whether and under what conditions people may drive. The only difference between the two cases is the mode of presentation. In large-number cases, the risk is conventionally stated in frequentist rather than probabilistic terms. Thus, rather than asking whether a 1 in 20,000 chance that an average driver will injure or kill someone when she drives X miles today is too great a risk to allow her to take, we ask whether the statistical certainty that 1 out of every 20,000 car trips of X miles will result in serious injury or death is too great a harm to allow the general population to inflict by driving. If we refuse to let the numbers count when the choice is posed in frequentist terms, we end up with the same two unacceptable alternatives. If we regard ¹⁶ Scanlon acknowledges the absurdity of this result. But he believes it can be avoided without having to discount harms by their probability of occurring (aggregation), if we allow probabilities to be factored into the definition of the reasonable precautions that “the agent has to take to avoid causing harm.” What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998), p. 209. As I discuss below, this maneuver simply relocates aggregation; it doesn’t diminish its role.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  

31

statistical certainty as morally equivalent to absolute certainty (whatever we take the latter to mean), then driving is never permissible, because over the course of a year, however prudently we drive, collectively we are certain to kill or injure tens of thousands of people. On the other hand, if statistical certainty isn’t enough—if we have to know with absolute certainty that a particular instance of driving is going to result in grave injury or death before we can rule it out—then we cannot say anything about the permissibility of driving at all, or under specified conditions, until the harm has eventuated. Consider the Ford Pinto case again. As I stated at the outset, the moral many have drawn from the Pinto case is that it is unethical to allow some people to die or suffer serious injury simply because it would cost “too much” to prevent it. But what exactly does that mean Ford should do? However safe a car design, for some amount of money it could be made safer. And however much money Ford spends on making it safer, it can never eliminate the possibility that a design feature will be the but-for cause of someone’s death. Does that mean that Ford may not manufacture and sell cars? If not, what does it mean? Suppose that by investing $5 billion in additional safety features, Ford could reduce expected deaths or serious injuries by .01 percent. Should Ford (must Ford) make that investment? If so, how about $50 billion? $500 billion? In a world of finite resources, we have to draw the line somewhere. I don’t think any nonconsequentialist would disagree with that, and I imagine all would draw the line considerably short of $500 billion, given that such a sum—which in the end consumers will bear—would price most consumers out of the car market and hijack scarce societal resources that could save thousands if not millions of additional lives if deployed in a more cost-effective manner. If the nonconsequentialist response is, “then save the millions of others too,” we will in short order bankrupt the world. But wherever we draw the line, we will knowingly be allowing Ford to increase the number of preventable deaths “merely” to save money (to be used for some other, generally offstage, purposes). The rule that determines where to draw the line therefore cannot come from the paradigmatic nonconsequentialist thought that “human beings have dignity and not mere price,” and that they have an “unconditional[] . . . worth that . . . is not subject to trade-offs.”¹⁷ The most telling evidence that nonaggregative principles cannot deal with garden-variety risk has come from nonconsequentialists themselves. Some have been explicit about that incapacity. Consider this, from Robert Nozick: It is difficult to imagine a principled way in which the natural-rights tradition can draw the line to fix which probabilities imposed unacceptably great risks upon

¹⁷ Hill, “A Kantian Perspective on Moral Rules,” pp. 292, 296.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

32

    others. This means that it is difficult to see how, in these cases, the natural-rights tradition draws the boundaries it focuses upon.¹⁸

Or this, from John Rawls: In discussing how to reconcile natural duties (which include “the duty not to harm or injure another”) when they “come into conflict, either with each other or with obligations,” Rawls reluctantly concedes that with utilitarianism off the table, “I do not know how this problem is to be settled, or even whether a systematic solution formulating useful and practicable rules is possible.” Happily, he concludes, he needn’t answer the question in order to derive the general principles governing the basic structure of society.¹⁹ Rawls’s dismissal of the utilitarian solution is brief but telling. He considers and rejects use of the “utilitarian principle to set things straight,” not because it isn’t a viable solution, but because it would very quickly lead to utilitarianism full stop: “Requirements for individuals so often oppose each other that this would come to much the same thing as adopting the standard of utility for individuals”—a solution that he has already ruled out on the grounds that it leads to “an incoherent conception of right.”²⁰ Scanlon sounds a similar note in explaining his choice to adopt an ex post view of the problem of harm to others in What We Owe to Each Other: “I did that because I was concerned that the introduction of ex ante benefits and burdens would open the door to aggregation of a kind that I aimed to avoid.”²¹ Many more have folded implicitly, by qualifying our prima facie unconditional duties (not to sacrifice the one for the good of the many, not to cross others’ boundaries) in terms that sound in deontological values but get cashed out as aggregation.²² This is surely a case in which covert tools are not good tools. If concepts like “undue risk” or “reasonable precautions” have an operational meaning distinct from aggregation, it needs to be put on the table so it can be assessed side by side with aggregation. If they don’t—if (as I believe) they boil down aggregation manqué—then we can cross that purported disagreement between nonconsequentialists and consequentialists off the list and move on. But as Rawls foresaw with respect to harm more generally, given that virtually everything we do poses at least some remote risk of seriously compromising the fundamental interests of others, ceding risk of harm to aggregation comes at a very high price for the nonaggregationist project. Every time we jog down a city street, ¹⁸ Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 75. ¹⁹ John Rawls, A Theory of Justice (Cambridge, MA: Belknap Press, 1971), pp. 114, 339–40. For other representative statements ceding the problem of risk to aggregative techniques, see Jules Coleman, Risks and Wrongs (Oxford: Oxford University Press, 2002), p. 210; Joel Feinberg, Harm to Others (Oxford: Oxford University Press, 1987), pp. 190–3. ²⁰ Rawls, A Theory of Justice, p. 339. ²¹ T. M. Scanlon, “Reply to Zofia Stemplowska,” Journal of Moral Philosophy 10, no. 4 (2013): p. 510; Scanlon, What We Owe to Each Other, pp. 204–5. ²² For further discussion, see Chapter 5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

33

or decide to release a prisoner early, or decide not to breach patient confidentiality to warn someone else that she may be at serious risk, however prudent our decision, we are taking others’ lives in our hands. As a result, ceding risk to aggregation saves nonaggregative principles by rendering them largely irrelevant. Not surprisingly, then, those nonconsequentialists who have grappled with risk have sought ways to keep it in the nonaggregationist fold. I have already mentioned two: (i) smuggle aggregation in by the back door through qualifications placed on our absolute duty not to harm others (e.g., except when we act with “due care”); and (ii) choose among different persons’ fundamental interests based on factors that sound in deontological values (e.g., statistical v. identified persons, proximity, doing v. allowing) but not the value that ostensibly underwrites our duty not to harm: that every person has a worth “that cannot be . . . subject to trade-offs.”²³ A third approach has been to seek a meta-principle that provides nonaggregative justifications for embracing aggregation—for example, the retreat to reasons in the Kantian literature; reliance on what Nozick describes disparagingly as “a utilitarianism of rights” before proceeding to deploy it himself to derive the just minimal state.²⁴ All three of these strategies to rescue risk from aggregation have been discussed in the literature, some extensively so, and I return to them in later chapters. For the balance of this chapter, I want to focus on a fourth strategy that has played a much larger but generally unnoticed role in the nonaggregationist literature: recharacterize ex ante uncertainty as certainty through what amounts to hindsight bias. It would be an exaggeration to say that the case against aggregation is built on hindsight bias. But it would not be that much of an exaggeration. Strip away hindsight bias, and it is hard to avoid the conclusion that nonaggregative principles cannot supply a viable, systematic alternative to aggregation.

4. Risk as Seen through Hindsight Bias The term “hindsight bias” refers to our psychological propensity to believe, without adequate evidentiary support, that the actual (ex post) consequences of our choices could have been predicted ex ante.²⁵ Monday morning quarterbacking is one colloquial description of the phenomenon. ²³ For a trenchant critique of Scanlon’s attempt to distinguish between certain and uncertain harms on this basis, see Alastair Norcross, “Contractualism and Aggregation,” Social Theory and Practice 28, no. 2 (Apr. 2002). For further discussion of this issue, see Chapters 3 and 4. ²⁴ Nozick, Anarchy, State, and Utopia, p. 14. For an excellent summary of many of the arguments, see Larry Alexander and Michael Moore, “Deontological Ethics,” Stanford Encyclopedia of Philosophy, November 21, 2007, http://plato.stanford.edu/entries/ethics-deontological. ²⁵ More precisely, it refers to the propensity to conclude, ex post, that actual consequences had a greater probability of resulting than the evidence that was available ex ante would have supported.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

34

   

The standard tragic choices problems in the nonconsequentialist literature— trolley problems and their many cognates—obviate the need for explicit hindsight bias by building perfect 20/20 hindsight into the definition of the problem. Because we are endowed ex ante with perfect knowledge of the ex post consequences of our actions, ex ante and ex post epistemic perspectives converge. Ex ante, we know that act X will produce Y consequences. Ex post, when it in fact produces Y consequences, we have learned nothing new. As suggested above, that convergence allows nonconsequentialists to indulge in the illusion that the costs (for example) of strictly applying the maximin rule adopted in Scanlon’s Greater Burden Principle or treating rights as trumps will be limited to those few occasions in which serious harm will result. But in a world of epistemic uncertainty, the two perspectives come apart by definition. Yes, the ex post consequences of our choices can yield important information that allows us to update the prudence of similar choices going forward. In some number of cases, the consequences can shed light on the ex ante prudence of the particular decision that produced them. But our reactions to bad outcomes are, alas, not that discriminating. They generally take the much cruder and logically indefensible form of concluding that if something bad has happened and some human agent is a but-for cause of it, the agent violated her duty not to harm others, and hence is to blame (in the moral, not causal, sense). But in a world of epistemic certainty, we cannot choose to act so as not to harm others. We can, of course, act so as to minimize the chances that we will harm others, given the other objectives we are trying to achieve. But doing so necessarily requires that we discount harms and benefits by their probability (frequency) of occurrence. That is to say, it requires that we let the numbers count. Our impulse to reevaluate conduct in light of its consequences is hardly surprising. We live in the world that is, not in alternative worlds that might have been much more likely ex ante to come to pass but did not. When one is a but-for cause of some tragic consequence, it is almost impossible not to run the film backwards and think, if only I had gone to the store an hour later, none of this would have happened. That retrospective recharacterization of the tragic choice we actually faced—how to trade off the convenience of driving against the remote possibility that on any occasion we drive, including this one, we will harm uncountable and unidentifiable others—offers the illusion that the choice needn’t have been tragic at all. After all, would it really have been a great inconvenience to go to the store an hour later? That is to say, it invites us to reconstruct every choice as a trolley problem, in which we can choose with perfect knowledge of the consequences of our available choices. That illusion is based on simple hindsight bias. As I suggested above, faced with the choice we actually make daily—whether to drive at all, and if so, with what level of precaution—all we know is that there is a remote possibility that this will be the day when, driving prudently, we will nonetheless kill someone. Of course,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

35

that remote possibility gets factored into our rational ex ante choice. But it is only one factor, and in weighing it against others, we rightly discount it by its low probability of occurring. It manifestly does not prove dispositive for the tens of millions of Americans who drive on any given day, or the government charged with deciding under what conditions they are permitted to do so. One can find numerous statements in the nonconsequentialist and consequentialist literatures to the effect that the consequences of our actions determine whether we acted wrongly. Many of those statements appear to be throwaway comments or misstatements, or are using right and wrong in an unconventional sense, and I am not concerned with them here.²⁶ I am concerned only with cases in which hindsight bias actually drives the author’s determination whether an act was, as of at the moment chosen, permissible or not. The problem of retrospective reevaluation of an act in light of its consequences has received significant attention from philosophers in one context in the deontological literature: the paradox of moral luck. In the classic moral luck case posed by Bernard Williams, a bus driver sets out one day, following his accustomed route and driving with due care (meaning, with the level of care society expects of him). A young child darts out in front of the bus. The driver, who does not see her and could not have seen her until it was too late to stop, hits and kills the child.²⁷ ²⁶ A number of eminent consequentialists/utilitarians, including G. E. Moore, J. C. C. Smart, and Peter Railton, at times appear to commit themselves to the identical form of hindsight bias, equating “right action” with that “action which does in fact produce the best results.” It is inconceivable to me that they really mean what they seem to say. Utilitarianism is a practical philosophy that cares only about optimizing consequences. As Smart states, “the utilitarian criterion . . . is designed to help a person, who could do various things if he chose to do them, to decide which of these things he should do” in pursuit of that end. J. J. C. Smart, “An Outline of a System of Utilitarian Ethics,” in Utilitarianism, For and Against, edited by J. J. C. Smart and Bernard Williams (Cambridge: Cambridge University Press, 1973), p. 46. That is to say, it is a guide to action. The best rule-ofthumb for optimizing actual consequences is to optimize the expected consequences of one’s action, given available information about the probability and (dis)utility of possible outcomes at the moment one must act. In contrast, the injunction to choose those actions that will produce the best consequences is, in a world of epistemic uncertainty, no decision rule at all. Notwithstanding loose talk to the contrary, I believe that Moore, Smart, Railton, and other selfstyled “actual consequences” consequentialists (Railton’s term) are, at the end of the day, using the term “right action” in a very different sense: not to describe how an actor should have acted, but to describe how close a given action comes to achieving the desired end (optimizing utility/welfare). Smart, using “right v. wrong” to denote the latter, deploys “rational v. irrational” to denote the former: the rational action is that “action which is, on the evidence available to the agent, likely to produce the best results,” whereas the right action is the one that “does in fact produce the best results.” So, “what is rational is to try to perform the right action, to try to produce the best results”; Smart, “An Outline of a System of Utilitarian Ethics,” pp. 46, 47. In that usage, “right action” describes ends and “rationality” describes means. ²⁷ The original version of the hypothetical comes from Bernard Williams, “Moral Luck,” in Moral Luck (Cambridge: Cambridge University Press, 1981), pp. 20–39. In some versions, the hypothetical is altered to make the driver a little bit careless (e.g., he takes his eyes off the road for a split second) but with bad consequences that are massively out of proportion with that moment of carelessness. That change does not remove the central problem of outcome luck; it simply changes the degree of luck involved. For an astute discussion of these issues, see Jeremy Waldron, “Moments of Carelessness and Massive Loss,” in Philosophical Foundations of Tort Law, edited by David G. Owen (Oxford: Clarendon Press, 1995), pp. 387–408.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

36

   

If asked ex ante (that is, before the child was hit) whether the bus driver was acting as we would have him act, the average person would presumably say yes, because the driver was driving as we would have him drive. But if asked the same question ex post, after things turn out badly, the average person (or so it is assumed) would conclude that the driver acted wrongly after all. What are we to make of these apparently contradictory evaluations of an identical act? In his response to Williams, Thomas Nagel argues that they do not create a logical paradox, because they are responding to different evaluative criteria.²⁸ The ex ante judgment reflects a rational calculus of how we would have people act, based on the probabilistic information available to them at the time they act. It is, in short, an action-guiding norm. The ex post “judgment” expresses the sorts of negative emotional reactions (regret, guilt, remorse, anger, blame, etc.) we all have when things turn out badly, through no one’s fault. Our natural propensity to feel guilt or anger for the bad consequences of (all things considered) prudent choices may be unfortunate, in the sense that it guarantees much unhappiness in life to no good purpose. But it does not pose a logical paradox. It simply reflects the fact that our rational and emotional lives operate on different planes and respond to different stimuli. But clearly many neo-Kantians and corrective justice enthusiasts mean what they say when they say that the consequences of our actions determine whether we should have acted as we did.²⁹ In Richard Epstein’s words, “[T]he first task of the law of torts is to define the boundaries of individual liberty . . . . [T]he liberty of one person ends when he causes harm to another. Until that point he is free to act as he chooses, and need not take into account the welfare of others.”³⁰ In the moral luck literature, the conflation of ex ante and ex post epistemic perspectives is on the table for all to see and argue over. In contrast, when the same logical impossibility is housed in the imperative to “act so as not to harm others,” it typically slips by undetected.

²⁸ Thomas Nagel, “Moral Luck,” in Moral Luck, edited by Daniel Statman (Albany, N.Y.: State University of New York Press, 1979), pp. 57–71. A similar point can be made about the doctrine of strict liability in torts—that is, the requirement imposed in certain circumstances that the injurer compensate the injured whether or not the injurer was at fault (acted negligently, etc.). It is not easy to explain why, if someone acted exactly as we would have had him act, corrective justice requires that he compensate the victim for the resulting harm. Why him and not some equally innocent bystander, or society at large? But the moral requirement that he do so does not create a moral paradox. It simply adopts different normative criteria for evaluating conduct and the obligation to compensate those injured as a result of it. ²⁹ For further discussion of this point, see Chapters 4 and 5. It should be noted that Kant himself believed that the rightness or wrongness of an action does not depend upon its consequences. It depends instead on whether the action complies with the “maxim” guiding it. In this one respect, Kant shares more with expected-value consequentialists than with contemporary Kantians, although the content of their respective “maxims” is obviously very different. ³⁰ Richard Epstein, “A Theory of Strict Liability,” Journal of Legal Studies 2, no. 1 (Jan. 1973): p. 204 (emphasis added).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   ? 37 As I discuss in Chapter 5, among non-libertarians, the imperative to “act so as not to harm others” has led to some very odd conclusions, including the second of the two unacceptable alternatives discussed above: we cannot say anything about the permissibility of the act until it actually results in harm to others. In John Goldberg and Benjamin Zipursky’s formulation, The wronging [meaning harming] of one person by another is the very essence of . . . [acting wrongfully], and until such an event happens, there is no occasion to inquire whether an actor can or should be held to have acted wrongfully by violating a moral or legal obligation of conduct.³¹

Among libertarians, the hindsight bias built into the duty not to harm others operates differently, because libertarians generally don’t engage the permissibility of conduct from an ex ante epistemic perspective at all. Typically, they enter the (hypothetical) scene ex post—that is, only after they conclude, based on the “fact” that someone’s “boundary” was crossed, that a rights violation has occurred. Most never address the question of how an agent, operating in a world in which every action carries some risk of harm to others, should go about deciding ex ante how to act. As noted above, Nozick’s response in a moment of candor was that deontology could not answer that question. He is right. But then, how does it come to be that rights theorists can offer a confident judgment on the same question ex post—that is, conclude that what you did was wrong after all, on deontological grounds, because in the event it ended up violating another’s absolute right to her person or property? The explanation, once again, is simple hindsight bias.

5. Why Should We Care? The discussion up until now has focused on the logical problems facing nonconsequentialist principles in a world of epistemic uncertainty. But nonconsequentialists’ refusal to face up to risk and the tradeoffs it inevitably entails produces significant real-world costs that ought to worry all of us. In our private lives, most low-probability risks we run will never ripen into harm, because of the one-off nature of our decisions. Not so in the public policy realm, where the rules we choose often govern millions or hundreds of millions of events over the long term. As a result, even very low-probability harms are overwhelmingly likely to come to pass at some point. If we parole enough prisoners, one of them will turn out to be Willie Horton. If we develop a vaccine for AIDS that has a one in a ³¹ John C. P. Goldberg and Benjamin C. Zipursky, “Tort Law and Moral Luck,” Cornell Law Review 92, no. 6 (Sept. 2007): pp. 1123, 1138.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

38

   

million chance of triggering a fatal reaction and we administer it to enough people, someone is going to die from it. We can often reduce those risks by greater precautions. But at a certain point such precautions become prohibitively costly—either in dollars spent relative to the reduction in risk achieved or in new risks the precautions themselves create.³² And generally speaking, no amount of precaution will eliminate all risk. If the world’s most conscientious child welfare agency has to supervise thousands of children from dysfunctional families, over the long run some child under its watch is going to die from abuse or neglect. The question is, how should we understand and respond to those bad outcomes when they inevitably come to pass? We know how the average citizen responds: You (government officials, Ford Motor Company) had a duty to keep us safe, and you failed to do so. (As one commentator dryly remarked, “Seldom do we hear a company that was responsible for a deadly accident justify the loss of lives by saying that it was the result of a decision which, in terms of its effects, produced far more good than harm.”³³) And politicians and policy-makers know how the average citizen will respond. The result is to drive politicians and other public employees to channel enormous resources into preventing the high-visibility bad outcomes for which they know they will be blamed. Several years ago, negotiations to ease California’s budget crisis and prison overcrowding by early release of nonviolent offenders fell apart over exactly this problem. As one Republican legislator acknowledged with refreshing candor: “If we let someone out early, and that man commits a crime, the Assembly members are worried that that will come back to haunt them like the old famous Willie Horton ads.”³⁴ Similar concerns explain why child welfare agencies responded to the rash of high-visibility deaths of children in foster care in the 1990s by redirecting almost all of their resources to children they thought might be at some risk of death, ignoring the thousands of others that were facing serious but (they believed) nonlethal threats of abuse and neglect. The same desire to avoid ex post blame in the public and private sectors drives a substantial portion of our health care expenditures, investments in homeland security, and countless other major public policy decisions. Our propensity to think that if our choices have bad consequences the choices themselves must have ³² One tragic example arose in the wake of the Haitian earthquake in 2010. Responding to one sketchy relief effort by a group of American Baptists, the authorities halted all evacuations of sick and injured children as relief workers scrambled to obtain documentation that would prove they were not taking the children out of the country illegally. The New York Times reported that in the first week alone, an estimated ten Haitian children died or became seriously ill as a result of not being able to be airlifted out of Haiti. http://www.nytimes.com/2010/02/09/world/americas/09airlift.html?scp=1&sq= haiti%20children%20airlift&st=cse. For another widely discussed example of the seeming perversity of many risk reduction efforts, see Jonathan Wolff, “Risk, Fear, Blame, Shame and the Regulation of Public Safety,” Economics and Philosophy 22 (2006). ³³ Hansson, “Risk and Ethics,” p. 32. ³⁴ Carol Pogash and Solomon Moore, “California Officials Fear Abduction Case May Hurt Efforts on Parole,” New York Times, August 30, 2009, http://www.nytimes.com/2009/08/31/us/31abduct.html.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   ? 39 been bad may be immune to reason. As a result, the only way policy-makers may be able to curb its influence is by manipulating the psychological salience of bad outcomes or obscuring the factual chain of causation that produced them. But surely, before we reach that conclusion, we should decide whether it ought to be curbed—whether, that is, outcome-determinative nonconsequentialists are wrong to think that if only we act prudently enough, we can somehow avoid inflicting seriously bad outcomes on some individuals as the price of realizing more trivial benefits for the many. The full-blown consequentialist will say, of course they’re wrong. It is tragic when reasonable actions have bad consequences, but it is nobody’s fault. In a world of scarcity (in the economist’s sense), whatever we do has potential costs to someone. The best we can do is to act in a way that we expect will minimize aggregate costs relative to aggregate benefits, however we calculate or weight them, and, if the costs to individual victims are serious enough, remediate them ex post on welfarist grounds. If nonconsequentialists really believe that no amount of money is too much to require car companies to spend to lower the risk of fatal accidents, or that it would be immoral to permit driving for the mere convenience of others if some number of innocent pedestrians will die as a result, then they should say so, and be prepared to live with the economic consequences of those positions. If, on the other hand, they understand that in a world of uncertain consequences, any possible rule of conduct imposes tradeoffs among competing and often fundamental interests, then they have a moral obligation to acknowledge that fact, explain how they would make those tradeoffs, and how, if at all, their solutions differ from conventional aggregation. We may not be able to change the widespread intuition that bad consequences imply bad conduct; it may be an irreducible part of what it means to be human, and may need to be accommodated in some fashion in public policy simply in virtue of that fact. But if the intuition is wrong, philosophers ought to be the ones saying so most clearly, and doing what they can to counteract the pernicious public policy consequences of all of us at least half-believing otherwise.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

3 What Does Matter? The Case for Killing the Trolley Problem (or Letting it Die)

For the past forty years, the philosophical literature on the duty not to harm others has developed around a set of (now canonical) hypotheticals that require us to choose between harms to different people or groups of people: trolley problems, forcible transplants, Bernard Williams’s hypothetical of Jim and the Indians, John Taurek’s dilemma of whether to save the one or save the five, scenarios that implicate the doctrine of double effect (e.g., the munitions bombing cases), etc. The hypotheticals typically share a number of features beyond the basic dilemma of tragic tradeoffs between two (groups of) persons with conflicting fundamental interests. The implicated parties are all on stage, identified, and in close proximity to the agent. The causal chain between act and harm is direct and apparent. The agents are all individuals (as opposed to institutions), who are being asked to make a one-off choice without considering the consequences for off-stage others or the plausibility of scaling the decision rule to deal with all like cases. Finally, the consequences of the available choices are determinate and known at the moment of choosing. Although the occasions and mechanisms for making such tragic choices vary, for ease of exposition I refer to hypotheticals that share most or all of these secondary features as trolley problems, and the pursuit of right answers to these dilemmas as trolleyology. The intellectual hegemony of the trolley problem has shaped contemporary nonconsequentialist thought in a number of unfortunate ways. First, as I argued in Chapter 2, it has resulted in nonconsequentialists devoting the bulk of their attention to an oddball set of cases at the margins of human activity, while largely ignoring conduct that (outside of the context of criminal activity and warfare) accounts for virtually all harm to others: conduct that is prima facie permissible (mowing a lawn, fixing your roof, driving a car down a city street) but carries some uncertain risk of accidental harm to generally unidentified others. Proceeding by a kind of common-law case method that treats every problem as a new problem, trolleyology has produced a morally unconvincing patchwork of ad-hoc rules and exceptions that must be yet further qualified when faced with the least perturbation of the facts. The result, to borrow A. R. Ammons’s lovely phrase, is a “too adequate relationship” between moral principles and hypothetical facts, and between moral principles and desired results. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  ? 41 Second, the hermetic focus on trolley problems has led trolleyologists to misdiagnose what is going on in the problems themselves. The trolley literature has meticulously analyzed how our intuitions about harm to others change as we move from one trolley-type setup to another (life v. life; life v. lives, life v. limbs; act v. omission; upstream v. downstream harms; etc.). But once we abstract away from the secondary features common to all trolley problems and are left only with fundamental interests that cannot all be protected, many nonconsequentialists have no clear intuitions about the right answer, or conclude that the choice has to be ceded to some form of aggregation. This suggests that what is driving people’s intuitive resistance to aggregation in trolley problems is not third-party harm/ harm tradeoffs per se but rather the secondary features peculiar to trolley problems. Maybe our penchant to care much more about certain-to-result harm to identified victims than a risk of harm to as yet unidentified ones is defensible in moral terms. But if so, the terms will be very different from those typically invoked to explain the “right” result in trolley problems: our duty not to harm others. And if it is not defensible, trolleyology more likely belongs in the descriptive domain of moral psychology than the normative domain of nonconsequentialist moral philosophy. Third, the refusal to venture outside of the carefully constructed boundaries of each trolley problem has led to a moral myopia that blots out most of the world. Thomas Nagel trenchantly pressed this issue with respect to Nozick’s principle of the inviolability of human beings.¹ The problem is no less acute in politically more liberal, Kantian-inflected, notions of inviolability. Fourth, trolleyologists generally treat the posture in which trolley problems “happen” to present themselves—and in particular what on-stage players “happen” to know about the consequences to them of adopting different available courses of action—as an exogenous variable that defines the moral problem to be solved but is itself morally neutral. That view is central to the contractualist version of trolleyology, as I discuss in Chapters 4 and 6. It is also difficult if not impossible to defend, and has opened the entire trolley enterprise to the charge of begging the question. To put the point in the strongest terms, if, on further thought, nonconsequentialists concluded that the appropriate epistemic point of view from which representative individuals should formulate their objections to a candidate principle is before they know how things will turn out for them under that principle, then virtually the entire trolley literature becomes morally irrelevant, and nonconsequentialists will be driven to adopt principles that diverge little if at all from standard aggregation.²

¹ Thomas Nagel, “Libertarianism Without Foundations,” Yale Law Journal 85, no. 1 (Nov. 1975): p. 140. ² For the argument that ex ante contractualism leads to aggregation, see Larry Alexander, “Lesser Evils: A Closer Look at Paradigmatic Justification,” Law and Philosophy 24, no. 6 (Nov. 2005), p. 617

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

42

  ?

Derek Parfit’s 2002 Tanner lectures, published in On What Matters along with commentaries by Barbara Herman, T. M. Scanlon, Susan Wolf, and Allen Wood, are a notable contribution to the annals of trolleyology. In the course of hundreds of pages, Parfit offers and interrogates dozens of hypotheticals to tease out nonconsequentialists’ moral intuitions about when it is permissible or morally required to sacrifice some in order to save others from harm. Every one of those hypotheticals involves choices among consequences deemed certain to befall identified, close-at-hand, fictitious persons—that is to say, every one is (in my sense of the term) a trolley problem.³ Parfit himself is hardly a nonconsequentialist, and his aim in On What Matters is not to defend nonconsequentialism in general or trolleyology in particular. It is rather to show that, even starting from the inhospitable premises of trolleyology (inhospitable, that is, to consequentialism), Kantians and Kantian contractualists will be driven to consequentialist conclusions. Parfit is also responsible, more than any other person, for putting the moral claims of as-yet unidentified victims on the philosophical map, starting with his exploration of the non-identity problem in Reasons and Persons.⁴ While the moral complexities the non-identity problem surfaces are different from those pushed off-stage by trolleyology, the two are not wholly unrelated. For all these reasons, it would be odd to describe Parfit himself as a trolleyologist, and I do not mean to do so here. And yet, as I discussed in Chapter 2, even Parfit has succumbed to the allures of trolleyology. For Parfit’s immediate purposes, the detour through trolleyology may be at worst a waste of time, and at best a tactically shrewd belt-and-suspenders operation to bring along those Kantians who are unpersuaded by arguments from the ex ante epistemic point of view. But taking a longer-term view, I think it is a tactical error for consequentialists to engage trolleyologists on their own terms (i.e., trolley problems) without registering objections to those terms. As I suggested above, the common-sense intuitions that trolleyology traffics in—how could it be right to throw the fat man off the bridge to stop the trolley, shoot

n. 23. I discuss when ex ante contractualism is likely to diverge from conventional aggregation in Chapter 4. ³ Well, not quite. One of the examples Parfit gives of duties we owe to as yet unborn persons involves a piece of glass negligently left in the woods, and stepped on ten years later by a 5-year-old child. Derek Parfit, On What Matters, vol. 1 (Oxford: Oxford University Press, 2011), pp. 217–18. This is clearly a case of risky conduct with uncertain consequences and as yet unidentified potential victims. But Parfit avoids most of the difficulties that arise in regulating risk by choosing conduct that anyone would recognize as negligent and describing it as such. As a result, the “right” answer—the conduct is impermissible—is essentially stipulated up front. The only remaining moral task is to resolve the non-identity problem—that is, how to deal with the claims of an as-yet nonexistent victim. ⁴ Derek Parfit, Reasons and Persons (New York: Oxford University Press, 1984). Indeed, as discussed below, Parfit relies on the claims of future, as yet unidentified, persons in On What Matters to derive consequentialist conclusions from the nonconsequentialist premises of trolleyology.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

43

one innocent person to save five, leave a workman to suffer electric shocks for an hour so that soccer fans world-wide can watch the World Cup uninterrupted?— are arguably psychological rather than moral in nature, and have at best an extremely limited domain. But the visceral power of those intuitions lends the nonconsequentialist principles extracted from trolleyology a surface moral plausibility they have not earned. In short, the right position for consequentialists to take with respect to trolley problems is to just say no to the lot of them.⁵ All of this is to say that my disagreements with Parfit are largely a friendly family quarrel. The real object of criticism here is committed trolleyologists. Nevertheless, On What Matters provides a useful text to flesh out these concerns, and to make the case for nonconsequentialists’ killing the trolley problem, or at the very least letting it languish from neglect while they redirect their attention to the kinds of tragic tradeoffs we confront and make daily.

1. Minimizing the Prevalence of Tragic Choices Allen Wood devotes a substantial portion of his commentary in On What Matters to decrying the outsized role of trolley problems in nonconsequentialist philosophical arguments. I share many of his objections, including to the fantastical nature of the dilemmas trolley problems pose; the absence of contextual information that in real life changes the moral complexion of tragic choices; and the unrealistic stipulation that the outcomes of all available choices are known with certainty ex ante.⁶ But we have very different—indeed, opposite—worries about the ways in which preoccupation with trolley problems has distorted nonconsequentialist thought. That disagreement points to a more profound disagreement about the inevitability of interpersonal tradeoffs, and the concomitantly limited relevance of Kantian notions of individual inviolability in any imaginable social world. Wood believes that obsessing on trolley problems has led nonconsequentialists to overestimate the occasions on which we must, unavoidably, choose between one person’s life, health, or ability to pursue her own projects, and another’s. I believe that it has led them to underestimate those occasions, dramatically. There is no logical contradiction here. Both of these distortions can—and I believe do— coexist. But the latter is, to my mind, the far more troubling one.

⁵ Since I originally wrote this sentence six years ago, it has proved more difficult than I had hoped to say no to the lot of them. Trolleyology is so deeply entangled with the nonconsequentialist project on harm to others that if one refuses to engage trolley problems, one refuses to engage that literature. ⁶ Allen Wood, “Humanity as an End in Itself,” in On What Matters, vol. 2, edited by Derek Parfit (Oxford: Oxford University Press, 2011), pp. 69–70.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

44

  ?

Wood’s argument that obsessing on trolley problems has led us to exaggerate the necessity for tragic choices is straightforward. By stipulating that we face a tragic choice and focusing attention solely on how we ought to choose—whether to flip the switch to kill one or instead let five die, whether to rescue five strangers on one rock or one’s own child on another—trolley problems encourage readers to regard such tragic choices as unavoidable. But in real life, Wood argues, most tragic choices could have been avoided if the relevant individuals had taken adequate preventive measures at an earlier moment in time, for example, by building safer trolleys, putting up better signage to warn passersby, erecting fences to “prevent[] anyone from being in places where they might be killed or injured by a runaway train or trolley.”⁷ By pushing those earlier decisions off-stage, Wood argues, trolleyology has pushed them off the philosophical agenda as well. Out of sight is out of mind in academic discourse as much as in other aspects of life. Wood acknowledges that some tragic choices are unavoidable. But, he suggests, most of those involve “extreme and desperate situations in human life” such as war, anarchy, pestilence, famine, or natural disaster. In contrast, when faced with quotidian decisions like allocation of health care services, if we find ourselves deliberating whom to save and whom to let die, it is in all likelihood because we have made a “voluntary decision . . . to turn health care, or even human life as a whole, into something horrible and inhuman, something like war, that ought never to exist.”⁸ Without doubt, many of the tragedies that occur in the developed world are easily preventable. Indeed, a case could be made that the most serious problems facing contemporary American society can be traced to our collective failure to invest prudently in prevention (using “prevention” broadly to include all ex ante investments to improve ex post outcomes) and instead waiting to remediate those outcomes after the fact. But the fixation on trolley problems hasn’t contributed to that systemic failure except perhaps indirectly, by leading people to misunderstand the nature of the choices we face when we do invest proactively in prevention. In contrast, nonconsequentialists’ obsession with trolley problems has led many to underestimate radically the occasions in life in which we are forced to make tragic choices. Given that trolley problems are about nothing but tragic choices, the claim that obsessing on them has led philosophers to underestimate the need to make such choices no doubt seems paradoxical. I don’t think it is. By presenting tragic choices only in “extreme and desperate,” indeed freakish, circumstances, trolleyology has inadvertently led both authors and consumers of that literature to regard tragic choices themselves as rare and freakish in

⁷ Wood, “Humanity as an End in Itself,” p. 74. ⁸ Wood, “Humanity as an End in Itself,” pp. 79, 80.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

45

nature. But they are neither. They are ubiquitous and for the most part quotidian, and they typically result not, as Wood suggests, from “human vulnerability to nature, and . . . human wickedness”⁹ but from the finite nature of the resources we depend on to realize our projects in the world. This returns us to the problem of economic scarcity discussed in Chapters 1 and 2. However wealthy a society is, however many doctors it has trained, however many procedures it underwrites and public health projects it has undertaken, at some point it cannot put more resources into addressing the health needs of one citizen without leaving unaddressed even more pressing needs of another. In a country as rich as the United States, we could significantly improve outcomes just by allocating our existing health care budget more sensibly. But there will always be more we could do for the sick if money were no constraint. And on the prevention side, the needs are virtually limitless (medical research, prevention of diabetes, obesity, drug use, and poor cardiovascular health, eradicating infectious diseases, etc.) Long before we reach the point where an additional dollar invested in health care and prevention would yield no incremental benefit in lives saved, we will (rightly) conclude that investing that additional dollar in, say, education, social workers, police, mitigation of climate change, etc., would have a greater positive impact on peoples’ lives. In a world of finite resources, we have to draw the line somewhere. I do not imagine any nonconsequentialist would disagree with that, and some might well draw it short of the $3.4 trillion the US spent on health care in 2016. But wherever we draw it, someone will die sooner than they had to because we did not invest a few more dollars in a cure for cancer, treatment of diabetes, etc. In that sense, we will unavoidably be making “tradeoffs between the deepest interests of different people.”¹⁰ Moreover, the way we make those tradeoffs typically violates the one principle that almost all nonaggregationists agree on: it is wrong to cause death or serious harm to one person in order to avoid more trivial harm to (or realize more trivial benefits for) others, no matter how numerous those others are. As Michael Ridge put it, “innocent lives always dominate convenience.”¹¹ But this principle routinely produces the “wrong” answer in garden-variety conduct that imposes a risk of harm on others. Take for example the permissibility of constructing a sports stadium, knowing that, however great the precautions taken, there will remain some irreducible risk of death to innocent passersby from falling debris. If a passerby does die as a result of falling debris, that harm will typically be much more serious than the harms thereby prevented (loss of recreational enjoyment to spectators, loss of revenues to players and management, etc.). Thus, under the

⁹ Wood, “Humanity as an End in Itself,” p. 81. ¹⁰ Wood, “Humanity as an End in Itself,” p. 79. ¹¹ Michael Ridge, “How to Avoid Being Driven to Consequentialism: A Comment on Norcross,” Philosophy and Public Affairs 27, no. 1 (Jan. 1998): p. 55.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

46

  ?

principle that “innocent lives always dominate convenience,” the project would clearly be ruled out. But no one, trolleyologists most certainly included, believes it would be wrong to build a stadium in any case in which there is a non-zero chance of serious injury or death to passersby.¹² It is easy to see why. Given how few choices in real life read to the philosophical (or lay) mind as trolley problems, if nonaggregationists stuck to their guns in every one of them—we may not kill one even to save 100,000 from the loss of a limb; we must rescue Jones from an hour of extreme pain, even though doing so will deprive a hundred million viewers of the pleasure of watching a World Cup match on TV—life would go on pretty much as before. In this very practical sense, nonaggregative solutions to trolley problems, whether or not they strike most people as morally persuasive, will at least strike them as feasible, for the same reason that the conventional Duty of Easy Rescue strikes most people as feasible: because only rarely will anyone have to alter his daily life to meet that duty.¹³ (To put it another way, Fiat justitia ruat coelum is an easier imperative to live by when we know the heavens are in fact not going to fall.) Not so in the case of garden-variety risks. If the numbers do not count in choosing between greater and lesser harms under conditions of uncertainty, such that the mere possibility of severe harm to even one person precludes actions with expected lesser benefits to millions, life as we know it would pretty much grind to a halt. Acknowledging the unacceptability of that result, many nonconsequentialists have conceded, explicitly or implicitly, that when it comes to risk, we are all aggregationists.¹⁴ As I discussed in Chapter 2, that concession comes at a very high cost to the nonaggregationist project. Others have argued that nonaggregative principles do have a role to play in resolving risk, but the principles are different from those that apply in the case of trolley problems. The precise contours of those principles, however, have yet to be spelled out, and it remains to be seen whether they can be made both viable and clearly distinct from conventional aggregation. I am doubtful, for reasons I discuss at length in the succeeding chapters. But for now, the only point I wish to urge is that the answer to that question is far more important to the future vitality and relevance of the nonconsequentialist project than anything more to be learned from trolleyology. ¹² For Scanlon’s acknowledgment of that apparent inconsistency and his justification for it, see Scanlon, What We Owe to Each Other, pp. 235–7. For the argument that they are not in fact reconcilable, see Alastair Norcross, “Comparing Harms: Headaches and Human Lives,” Philosophy and Public Affairs 26, no. 2 (Spring 1997): pp. 136–67. ¹³ The Duty of Easy Rescue states that if you can save someone at a trivial cost to yourself, you must save her. ¹⁴ As I suggested in Chapter 2, this may also be what Parfit has in mind in saying that “[t]hese are not the questions about which different people, and different moral theories, most deeply disagree.” Parfit, On What Matters, vol. 1, p. 162.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  

47

2. Misdiagnosing Trolley Problems The single-minded focus on trolley problems has not just led trolleyologists to give short shrift to the problem of risk; it has also done them no favors in understanding trolley problems themselves. Once again, the claim may seem paradoxical, but I don’t think it is. That many of the principles extracted from trolleyology do not produce the “right” answer if applied beyond trolley cases suggests that the principles, at a minimum, are stated too broadly, and more likely are misstated. In particular, two other factors present in all trolley problems are very likely operating as (unacknowledged) limiting features: the harms in question befall victims who are identified and generally proximate to the actor. Can these limitations be defended? It is beyond dispute that the identifiability and proximity of the victim are emotionally and psychologically salient to most people in judging the permissibility of (potentially) harmful conduct. The crucial question for trolleyologists is whether they are morally relevant, and if so, why.¹⁵ Peter Singer’s 1972 article, “Famine, Affluence and Morality,”¹⁶ forced that question in the context of the Duty of Easy Rescue. The debate he started continues to this day.¹⁷ In Singer’s words, if the rationale for the Duty of Easy Rescue is that “if it is in our power to prevent something very bad from happening, without thereby sacrificing anything morally significant, we ought, morally, to do it,” why should that duty be limited to one-off heroic rescues of proximate victims?¹⁸ Why does it not extend to the obligation of moderately affluent Westerners to make what are, for us, relatively trivial financial sacrifices to help alleviate mass starvation in the developing world? One answer, of course, is that doing so would eliminate one of the chief advantages nonconsequentialism is thought to have over utilitarianism: that it does not demand an implausible degree of other-regardingness from each of us in our private lives. But, as Elizabeth Ashford has argued with respect to Scanlon’s version of the Duty of Easy Rescue, that answer seems to save morality by reducing it to whatever we are willing to do because it costs us little.

¹⁵ I don’t mean to endorse the distinction between rational and emotional responses. But I take it to be basic in some form to what is meant by rationality, duties, or rights in Kantian and deontological frameworks. ¹⁶ Peter Singer, “Famine, Affluence, and Morality,” Philosophy and Public Affairs 1, no. 3 (Spring 1972). ¹⁷ For trenchant arguments in the Singerian vein, see Elizabeth Ashford, “Utilitarianism, Integrity, and Partiality,” Journal of Philosophy 97 (2000): pp. 421–39; Elizabeth Ashford, “The Demandingness of Scanlon’s Contractualism,” Ethics 113 (2003): pp. 273–302. For various efforts to defend (contra Singer) the moral relevance of identifiability and proximity in formulating the Duty of Easy Rescue, see Parfit, On What Matters, vol. 2, p. 211; Sarah Miller, “Need, Care and Obligation.” Royal Institute of Philosophy Supplement 80 (2006): pp.141–3; Richard W. Miller, “Beneficence, Duty and Distance,” Philosophy & Public Affairs 32, no. 4 (2004): p. 379; James Lenman, “Contractualism and Risk Imposition,” Politics, Philosophy, and Economics 7, no. 1 (2008): p. 116. ¹⁸ Singer, “Famine, Affluence, and Morality,” p. 231.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

48

  ?

Notwithstanding the equally critical role that identifiability, proximity, and (most of all) certainty have played in delimiting the Duty Not to Harm, few nonconsequentialists have acknowledged that role and even fewer have tried to justify it. While the challenge posed by risk has gotten more attention of late in the nonconsequentialist literature, most of that attention is directed at solving the myriad problems that follow from distinguishing categorically between harms that are certain and uncertain to occur, rather than defending the moral significance of the distinction to begin with. As I suggest in Chapter 2, many nonconsequentialists, including Parfit himself, seem to take the moral significance as self-evident.¹⁹ It is not. Given the decisive role that certainty, along with identifiability and proximity, plays in generating the “right” answer to trolley problems and the critical role that trolley problems in turn play in formulating the Duty Not to Harm, establishing that these factors have a moral and not just psychological or emotional basis is essential to the nonconsequentialist project. The first step is to recognize the decisive role that all three factors do in fact play in trolleyology. And that recognition will come only when trolleyologists leave behind the world of trolley problems for the world of uncertain risk of harm to statistical and/or remote victims, and observe systematically which of their intuitions survive the journey.

3. Taking Trolley Problems as We “Find” Them An unstated premise of trolleyology is that every trolley problem, however bizarre the stipulated facts. is equally appropriate grist for the philosophical mill, and that the best articulation of our Duty Not to Harm is whatever set of principles can accommodate (make cohere) our intuitions about the largest range of trolley problems presented. I believe that premise is impossible to defend. It disposes of two core normative questions sub silentio as if they were matters of fact: whose interests ought to be taken into account, and what is the appropriate epistemic perspective from which to assess them?

3.1 Whose Interests Count? In the standard trolley problem, the only persons whose interests count are the on-stage players that populate the hypothetical: those who will live or die as an immediate consequence of the choice, and (if the moral agent is a character in the hypothetical) sometimes the agent as well. Among the many others with stakes in

¹⁹ Parfit, On What Matters, vol. 1, p. 162.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     “” 

49

the decision who are routinely ignored are (i) family members, bystanders, and other interested third parties currently alive who will be directly affected by the choice made; (ii) unidentified individuals whom we could have saved if we had diverted the same resources to their rescue; and (iii) all the people who would be affected if we scaled up the solutions to one-off trolley problems to apply to all morally analogous choices. The first of these omissions could be corrected, albeit at the cost of great empirical uncertainty about the “right” answer to many trolley problems. I return to this problem in Chapter 6. The second and third, in contrast, are essential to the project of trolleyology. If forced to take into account the wider opportunity costs of making either of the two choices presented in a standard trolley problem, or forced to scale up the more implausible conclusions trolleyology generates, most trolleyologists would quickly turn consequentialist. Many of these issues have been discussed in the nonconsequentialist literature in the context of threshold deontology, the Paradox of Rights, agent-centered prerogatives, the Doctrine of Doing v. Allowing, and the Duty of Easy Rescue, among other topics.²⁰ I add one general observation here. The challenge that scaling-up poses to nonaggregative decision rules has, if anything, been understated in the literature, because of the prominent role given to life/life or life/lesser harm tradeoffs. However unappealing it may be to sacrifice 100,000 in order to save one, it is possible to do so, because there is no limit to the number of people we could kill, let die, or let live severely compromised lives (assuming we have no intention of caring for the latter). To put it another way, killing people or letting them die is relatively cheap, if not free. But the moment we have to pay to indulge the paradigmatic moral thought that no person should be sacrificed for the lesser interests of others—as, for example, we must do if we really believe that no amount of money is too much to spend to save a person’s life—the thought cannot be indulged very far before nonconsequentialists must, and will, fold.

3.2 From What Epistemic Perspective Should a Moral Agent Choose? The information that people are endowed with concerning their ex post fate under a proposed principle swamps in importance all other “facts” that influence their preferences. If we ask whether hypothetical person X could reasonably agree to ambulances being permitted to speed whenever it is the case that five patient lives would thereby be saved for every one pedestrian killed, we are likely to get one ²⁰ For an excellent overview of this terrain, see Alexander, Larry, and Michael Moore. “Deontological Ethics,” Stanford Encyclopedia of Philosophy, November 21, 2007, http://plato.stanford.edu/ entries/ethics-deontological.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

50

  ?

answer (no) if X knows she will be the one killed and a different one (yes) if, as far as she knows, her odds of turning out to be one of the five are five times greater than her odds of turning out to be the one. Which answer is the relevant one, morally speaking? The response, implicit or explicit, from many nonconsequentialists is both: whatever information we happen to have about our ex post fate under a proposed principle at any given moment is (in Parfit’s words) a “relevant, reason-giving fact” that must be taken into account in assessing whether, as of that moment, we could reasonably accept that principle.²¹ Treating the epistemic point of view of the characters that populate a trolley problem as an exogenous fact that frames moral analysis but is not itself subject to it has produced no end of confusion in the trolley literature. The examples here are legion.²² Consider two of the variants on Taurek’s famous dilemma that Parfit and Scanlon present in On What Matters.²³ Taurek’s original hypothetical asks us to consider a variant on the following set-up: A would-be rescuer (Green) is faced with the choice of saving five people (White, Blue, Yellow, Red, and Black) on one rock or one person (Orange) on another. The question Taurek poses is: If Green chooses to save the five rather than Orange based solely on the numbers, will he thereby violate Orange’s Kantian right to equal respect? Scanlon, like virtually every nonconsequentialist who has considered the hypothetical (other than Taurek himself), concludes the answer is no. At least when choosing between identical harms, both of which fall on the same side of the act/omission divide, Scanlon argues the numbers should count as a tiebreaker. In one of the variants that Parfit and Scanlon consider (Lifeboat II), Orange is Green’s child. As a consequence, Green now has a strong personal preference to save Orange rather than the five. Both Parfit and Scanlon conclude that, contra Lifeboat I, it is now permissible for Green to save Orange at the cost of the five, because over the long run, allowing each of us to be partial to our children in this way would make life go significantly better for almost all of us. In Parfit’s words: The[] good effects [of saving five rather than one] would be massively outweighed by the ways in which it would be worse if we all had the motives that such acts would need. For it to be true that we would give no such priority to saving our own children from harm, our love for our children would have to be much weaker. The weakening of such love would both be in itself bad, and have

²¹ Parfit, On What Matters, vol. 1, p. 356. The ambulance example comes from Frances Kamm, Intricate Ethics (Cambridge, Mass.: Harvard University Press, 2007), pp. 29–30, 274–5. I return to the example in Chapters 4 and 6, and interrogate the widespread view among nonconsequentialists that one has a right to renege on prior agreements once one learns of one’s own fate under them. ²² I return to this issue in later chapters. ²³ John C. Taurek, “Should the Numbers Count?,” Philosophy and Public Affairs 6, no.4 (Summer 1977): pp. 293–316.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     “” 

51

many bad effects. Given these and other similar facts, the optimific principles would in many cases permit us, and in many others require us, to give strong priority to our own children’s well-being.²⁴

Fair enough. But how does one get to the optimific solution from nonoptimific (Kantian/contractualist) premises? As Parfit notes, the easy answer is to consider the question ex ante, from behind a veil of ignorance: count only the preferences people would have before they possessed any individualized knowledge about the odds they would turn out to be Green, Orange, or one of the five.²⁵ From that position of equal ignorance, if everyone makes the same calculation about the expected value of partiality in their lives as Parfit has made on behalf of the average person, everyone will prefer a rule that allows Green to be partial to his own child. (This just illustrates the general point that where everyone is endowed with the same preferences and equal ignorance, ex ante contractualism converges with the optimific solution.) In developing his own theory, however, Scanlon rejected ex ante contractualism on principle, insisting that representative complaints must be weighted in accordance with the actual (ex post) harms that will ultimately befall actual victims.²⁶ But weighting complaints in accordance with actual (ex post) harm produces the “wrong” answer in Lifeboat II (Green may not choose to save his own child). Ignoring for the moment Green’s agent-relative preference to save his own child, the now self-identified White, Blue, Yellow, Red, and Black and the now selfidentified Orange each face certain death if Green’s choice doesn’t go their way. This gives each of the six a complaint (certain death) of equal qualitative strength. This is so, even if each of the six, judging the matter from an ex ante position of equal ignorance about the consequences to his life of adopting one rule rather than the other, would have chosen to let partiality trump. At that point, presented with identical complaints from each of the six who would be facing certain death, Scanlon’s tiebreaker principle would kick in to let the numbers count, once again yielding what Scanlon and Parfit both take to be the “wrong” answer: Green must save the five rather than his own child, Orange. Adding in Green’s preference would seem to change this result only if we are allowed to aggregate Green’s reasons to save Orange with Orange’s reasons to want to be saved, and count the sum as one individual’s reasons. How exactly we could justify aggregating Green’s and Orange’s reasons under Scanlon’s stipulation that only “individual reasons” count is not obvious. Moreover, even if we could justify it, Orange clearly wins the day only if we ignore the preferences of all the off-stage loved ones of the five who would die as a result of that choice. If we assume that at least one of those off-stage loved ones has as strong a preference to ²⁴ Parfit, On What Matters, vol. 1, p. 385. ²⁵ Parfit, On What Matters, vol. 1, pp. 349–50. ²⁶ I explore Scanlon’s justification for that position at length in Chapter 4.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

52

  ?

have his family member saved as Green has to save Orange, then we once again have a tie, relegating us to Scanlon’s tiebreaker principle, which again yields the “wrong” answer. So instead, Scanlon does what trolleyologists generally do, when adhering to the ex post perspective generates the “wrong” answer: he reverts, sub silentio, to ex ante contractualism, and with it to an aggregationist solution: A principle requiring us always to give the needs of strangers the same weight as those of friends and family members would be one that each of us could reasonably reject, because it would make impossible special relationships that we have strong reasons to want to have.²⁷

Parfit, in contrast, sticks to an ex post perspective in Life Boat II, but argues that Kantian contractualists can—indeed, must—reach the “right” (optimific) answer even from that perspective.²⁸ The argument is complicated, but the essential moves are these: (i) In most situations, each of us has sufficient, morally plausible reasons to act either selfishly or altruistically. (ii) But Kant’s Universal Law imposes a further requirement, in addition to the requirement that we have sufficient reasons to act as we do: we must also act in accordance with principles that everyone could rationally choose, if they were choosing principles that everyone could accept.²⁹ (iii) That further requirement will always be met by the optimific rule, because, while each of us may have good reasons to act selfishly, by stipulation we also have good reasons to choose those principles that will on the whole make things go better for everyone³⁰—that is to say, to act altruistically. (iv) In contrast, it will never be met by a principle that is significantly nonoptimific, because those who are disadvantaged by the principle more than trivially have neither self-interested nor altruistic reasons to accept it. (v) Thus, in Lifeboat II, while White, Blue, Yellow, Red, and Black have sufficient reason to prefer themselves, they also have sufficient reason to choose the optimific rule (Green may choose to save his own child), even at the cost of their own lives. In contrast, Green has no good reason to accept the nonoptimific rule that he must save the five in preference to his

²⁷ T. M. Scanlon, “How I am Not a Kantian,” in On What Matters, vol. 2, edited by Derek Parfit (Oxford University Press, 2011), p. 133. ²⁸ Parfit, On What Matters, vol. 1, pp. 382, 388–9. ²⁹ Parfit, On What Matters, vol. 1, pp. 401–2. ³⁰ Parfit, On What Matters, vol. 1, pp. 378, 379.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     “” 

53

own child, because it satisfies neither his altruistic nor his self-interested motivations. To generalize Parfit’s conclusion, whenever self-interest and the optimific rule diverge significantly, Kantians and Kantian contractualists should agree that the optimific rule always wins.³¹ But why should we assume that White, Blue, Yellow, Red, and Black have sufficient individual reasons to choose the optimific rule, even knowing that their own lives are at stake? As Parfit notes, a desire-based, subjectivist account of reasons will not yield that result, unless we impute an implausibly high degree of altruistic motivation to individuals.³² At the other extreme, an objectivist account of reasons that holds that “we always have most reason to do whatever would be impartially best” (“Rational Impartialism”)³³ will always yield that result. Indeed, it will yield the stronger result that the only principle we have sufficient reason to choose is the optimific rule. But it reaches the “right” (consequentialist) conclusion only by assuming it. Parfit’s actual route to the optimific result in On What Matters is different from Rational Impartialism, and subtler. Like Rational Impartialism, Parfit’s account of sufficient reasons is objectivist. But unlike Rational Impartialism, his version (a “wide, value-based objective view”) accepts that both partiality and altruism can supply morally sufficient reasons to act, depending on the circumstances. Parfit then argues that one circumstance in which all people would have a morally sufficient reason to sacrifice themselves for others, even if they ultimately chose not to, is when the stakes are high enough—if, for example, they would thereby save millions of other people. (Parfit is not making a claim here about peoples’ actual motivations, but it is a plausible description of those as well.) He then converts virtually every trolley-type problem into a high-stakes choice, by stipulating that “in the thought-experiment to which the Kantian Formula appeals,” all the on-stage players would weigh in the balance the consequences not just to each other, but to the “billions of people” who would be governed by a given principle, “both now and in all future centuries.”³⁴ That is to say, Parfit scales up the problem to motivate the consequentialist solution. Thus, in Lifeboat I (Taurek’s original example), Parfit’s Orange would think to himself, if we adopt a rule that says that someone in Green’s position should always save the larger number, “[t]hough I would die, my choice would indirectly save at least a million other people . . . . So even on . . . more egoistic views, I would have sufficient reasons to give up my life to save these very many other people.”³⁵

³¹ Parfit, On What Matters, vol. 1, p. 378. ³³ Parfit, On What Matters, vol. 1, p. 382. ³⁵ Parfit, On What Matters, vol. 1, p. 382.

³² Parfit, On What Matters, vol. 1, p. 381. ³⁴ Parfit, On What Matters, vol. 1, p. 382.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

54

  ?

And in Lifeboat II, Parfit’s White, Blue, Yellow, Red, and Black would think to themselves, if we adopt a rule that allows someone in Green’s position to be partial to his own child, we will die, but we will make it possible for billions of people from now through all eternity to be partial to their children in a way that will make all of their lives go significantly better. (Query why the self-sacrificers wouldn’t be entitled at the very least to offset the gains to billions of future people from adopting the optimific rule with the aggregate loss to (millions of?) future people whose lives (like theirs) will go better under the nonoptimific rule? Doing so will not change the optimific conclusion—by definition, optimific rules will always produce greater overall welfare than nonoptimific ones. But in many cases it might reduce the welfare gap between optimific and nonoptimific rules sufficiently to boot the self-interested, nonoptimific choice out of the “high stakes” camp.) A good case can be (and has been) made that both ex ante contractualism and Parfit’s ex post high-stakes altruism produce consequentialist results by smuggling in consequentialist premises. But for present purposes, the point I wish to make is different and more general: the “right” answer to any trolley problem depends entirely on how the problem is set up, and in particular what information we impute to the choosers at the moment of choice. That makes it imperative for trolleyologists to defend, normatively, the set-ups they choose. If they do not—if they take each hypothetical as it just “happens” to arise—they will end up with a set of principles that are confused at best, and indefensible at worst. In my view, that pretty much describes the state of trolleyology at the moment. What set-up should trolleyologists adopt in analyzing the sort of tragic tradeoffs posed by trolley problems? Insofar as they are seeking general principles to live by—the self-described project of trolleyologists—I think the only morally defensible epistemic point of view from which to ask what principles everyone would have reason to accept is before they know how things will actually turn out for them over the course of their lives if a given principle is accepted. It is also the only epistemic point of view that can yield anything approximating general agreement (unless, like Parfit, we allow self-interested motivations to count in theory, but ensure that when they lead to significantly suboptimal results they will never carry the day in fact). As I argue in Chapter 4, contractualists have concluded otherwise—have concluded, that is, that everyone is entitled to take into account whatever information they just happen to possess in whatever hypothetical factual scenario we just happen to place them—because they have conflated two questions: whether, in choosing principles to live by, people are entitled to know their own preferences, aptitudes, and general situation in life; and whether they are also entitled to know how things will actually turn out for them if a given principle is adopted. Parfit makes the same mistake in On What Matters, in assuming that “relevant, reason-giving facts” include both general information about who we are and the circumstances we occupy in life, and knowledge of whether we will turn out to be

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     “” 

55

“one of the few people on whom . . . great burdens would be imposed” if a given principle were chosen.³⁶ But general knowledge about ourselves and our circumstances in life is sufficient to generate thick, differentiated, representative individuals and surface the genuine disagreements among them. Whether individuals should also be endowed with knowledge about ex post outcomes is a separate question that is orthogonal to whether we are committed to thick selves: Green is Green before he knows that he is the one who will have to choose between saving his own child and saving five, and he is Green after he discovers it. The question is, which of Green’s two epistemic states ought to be determinative? That is a normative, not a factual, question, and whatever answer we give has to be defended on normative grounds. If we go with the earlier Green, our conclusions will tip strongly in the direction of the optimific solution. If we go with the later, at which point Green is presumed to know how things will turn out for him, we are back to moral gridlock. Neither of these outcomes is acceptable to nonconsequentialists, and so they have sought to split the difference, following the model of Kamm’s Ambulance cases. At time 1 (Kamm’s Ambulance I), when everyone is in a condition of equal ignorance, everyone will agree to the optimific solution. At time 2 (Kamm’s Ambulance II), when Green knows that he is the one who will die, Green is allowed to renege on the agreement under the Greater Burden Principle. Writ large, that move returns us to moral gridlock, in two steps rather than one. It is, however, never writ large. While these two-step choice situations play an outsized role in the philosophical literature, in real life we rarely have the opportunity to revisit the decision made in time 1 before it becomes irreversible. But suppose we did. It is one thing for everyone to agree in principle that ambulances should be permitted to speed whenever doing so will save five statistical patient lives for every one statistical pedestrian life thereby lost. But, many feel, it is quite another for a driver to adhere to that agreement when pedestrian Smith walks in front of the ambulance and the driver could avoid hitting her by applying the brakes. I agree. The question is, what kind of thing is it, what should we do about it, and how can we reconcile that response to our response in Ambulance I? For utilitarians, the answer to these questions is straightforward, at least in theory. Both choices should be resolved by the same moral criterion: pick the decision rule that maximizes expected welfare. In cases that come up in the posture of Ambulance I (ex ante decisions under conditions of uncertainty), the optimal rule will usually be to sacrifice the one to save the five. The optimal rule in Ambulance II is a more complicated calculus. Seeing Smith right in front of him and knowing that he could save Smith by hitting the brakes will change the

³⁶ Parfit, On What Matters, vol. 1, p. 356.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

56

  ?

social meaning of the earlier agreement for the driver. It will change it as well for immediate bystanders and the larger society, to the extent that they are aware of the nature of the choice that the driver faced and made. If it changes the social meaning enough, it could conceivably tip the optimific calculus, leading a utilitarian to conclude that anyone who ever finds herself in the position of the driver in Ambulance II either may or must hit the brakes, notwithstanding the general agreement in Ambulance I to the contrary. But for utilitarians, that qualification introduces no analytical or moral inconsistency. It simply recognizes that in some circumstances, following the “sacrifice one to save five” rule will occasion enough psychological discomfort and/or political instability that the optimific move is to allow an exception. My guess is that most nonconsequentialists, in the end, would strike roughly the same compromise as utilitarians: adopt an ex ante decision rule more or less by the numbers, but permit or require individual actors to deviate from the rule in cases in which adhering to it feels overwhelmingly wrong because of the immediacy of the consequences. This is in fact the compromise they have struck in dividing the universe of potentially harmful conduct into certain harms (trolley problems) and uncertain harms (risk), and resolving the former by nonconsequentialist principles and the latter by aggregative principles that differ little if at all from the optimific solution. Unlike welfarists, however, it is not clear that nonconsequentialists can account for both halves of that compromise by some unifying moral principle.³⁷ The candidate principle offered to date—that a policy that “is not acceptable at every time [to every person] is plausibly acceptable at none”³⁸—is not going to do the job. At most, it explains why we are drawn to let the pedestrian who finds herself trapped in Ambulance II renege on her assent to an optimific solution in Ambulance I (“Sure, she said that then, but look at her now”). But it does not explain why her assent in Ambulance I was good to begin with, since it was foreseeable from the vantage point of Ambulance I that whoever in that group ultimately turned out to be the pedestrian in some future Ambulance II would live at least long enough to regret her decision. And it does not explain why we ought to let the victim renege in Ambulance II. As much as she now finds unacceptable, from the vantage point of Ambulance II, her agreement to the optimific solution in Ambulance I, she would find equally unacceptable, from the vantage point of Ambulance I, her preference for a nonoptimific solution in Ambulance II. To put the problem in general terms: since our preferences with respect to a given policy will change with a change in our epistemic perspective, the principle that a policy that “is not acceptable at every time [to every person] is plausibly ³⁷ For the argument that they cannot, see Norcross, “Comparing Harms: Headaches and Human Lives.” ³⁸ Lenman, “Contractualism and Risk Imposition,” p. 117.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     “” 

57

acceptable at none” means that no policy will ever be plausibly acceptable. It also means that if a pedestrian’s imminent peril is manifest while the driver still has time to hit the brakes, her interests are dispositive, but if it is not discovered until the moment after, they are not counted at all. What moral principle could possibly justify this distinction? Finally, the proposition that every epistemic vantage point is of equal moral relevance is hardly self-evident. For the many reasons spelled out in this chapter, I think it is difficult if not impossible to defend. But at a minimum, it requires defense. Other nonconsequentialists have responded, in effect, “Of course the right results in Ambulance I and II cannot be explained by the same moral principle. But why would we think they should be, given that they present very different moral problems?” As my discussion in Section 1 suggests, that view seems hard to defend if the fundamental bad in both scenarios is the actual harm that could or will befall real people. Why would future harm to others present a moral problem that is different in kind, depending on whether it is deemed certain or uncertain to come to pass? If the fundamental bad posed by Ambulance II is not the impending harm itself but rather the secondary features introduced in the move to Ambulance II (the certainty that harm will result, the identifiability and proximity of the victim, etc.), then there is indeed no reason to assume that we should resolve Ambulance II in accordance with the same principles applied in Ambulance I. But in that case, trolleyologists face two further challenges. The first they share with utilitarians: to show how inconsistent decision rules applied to the same act depending upon the epistemic vantage point from which it is judged can be made to cohere into a socially stable regime. The second they do not: to show that the decision rules they would adopt from the vantage point of Ambulance II have a moral, and not merely psychological or emotional, basis. I return to this issue in the context of Scanlonian contractualism in Chapters 4 and 6.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

4 Can Scanlonian Contractualism Save Us from Aggregation? Over the past half-century, contractualist theories have played a prominent role in the assault on utilitarianism in general, and aggregation in particular. In contrast to the other nonconsequentialist arguments I examine in the first part, contractualist constraints on aggregation derive not from the first-order duties and corresponding rights of individuals, but instead from the requirement that whatever rights and duties we agree to must be acceptable to each individual subject to them. In traditional consent-based versions of contractualism, constraints on aggregation are built into the requirement of unanimous (hypothetical) agreement in a suitably conditioned bargaining situation. In justification-based models of contractualism, which judge the reasonableness of a principle by the reasonableness of the reasons offered in support of it, constraints on aggregation are built into the sorts of reasons that do and do not count. Notwithstanding the formal differences in modes of justification, all three approaches tend to lead to the same conclusion in the hands of the same person. This is hardly surprising, given that the suitability of the bargaining situation and the reasonableness of reasons offered to justify a given principle are judged by the normative commitments that drive first-order morality. Nevertheless, contractualists’ rhetorical stress on individual hypothetical assent pushes into the foreground epistemic problems raised in earlier chapters. In this chapter, I look at the role of contractualist arguments in contemporary moral theory, focusing on T. M. Scanlon’s What We Owe to Each Other.¹ In the second part of the book, I examine the role of contractualist arguments in political theory, looking at Rawls on the left and James Buchanan and other social contractarians on the right. In Scanlonian contractualism—the starting place for most contemporary leftleaning contractualists in moral theory—the only complaints that count for or against a proposed principle are those held by some hypothetical (representative)

¹ T. M. Scanlon, What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998). For Scanlon’s original, slightly different, version, see “Contractualism and Utilitarianism,” in Utilitarianism and Beyond, edited by Amartya Sen and Bernard Williams (Cambridge: Cambridge University Press, 1982), pp. 103–29. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

          

59

individual (Scanlon’s “individual reasons” restriction). The relative gravity of different individuals’ complaints is judged through pair-wise comparisons, without regard to the numbers of others who hold similar complaints (that is, without aggregation). I argue here that Scanlon’s original formulation of contractualism cannot supply a viable alternative to aggregation, for many of the same reasons that the noncontractualist theories discussed in Chapters 3 and 5 cannot. Some of the principles proposed appear to be aggregation under a different name. The rest are clearly distinct from aggregation, but they are viable only as long as they are restricted to the oddball set of hypotheticals around which the literature has been built. They cannot be scaled up to resolve the more general problem of potentially harmful (that is, risky) conduct. Over the past few years, a number of Scanlonian contractualists, including Scanlon himself, have modified their position in response to criticisms, including some of those offered below. I take up the modified versions in Chapter 6. There, I argue that while they save contractualism from its most obvious problems, they do so at the cost of turning contractualism aggregationist. A couple of preliminary comments. First, for current purposes, I accept the moral plausibility of a mixed theory—one that (in Stephen Perry’s formulation) holds not that all risks cannot be run, but that “certain risks cannot be run, so that consequential justification must take place within permissible deontological bounds.”² The questions I’m interested in are: (i) what are the permissible bounds that contractualists are proposing; (ii) do they follow from contractualist commitments; (iii) are they viable, in the minimal sense that they dictate coherent criteria for action; and (iv) would such criteria produce morally plausible results. Second, as with the versions of nonconsequentialism discussed in the rest of the first part, I am concerned only with the permissibility of acting in a way that imposes (a risk of) harm on others. I do not address the ultimate victim’s right to compensation or some other form of redress from the party who caused the harm. How we respond, ex post, to the harm we have caused is an independently important question that contractualist principles may shed light on, but it is a different question.

1. The Challenge for Contractualists: Finding a Middle Way between Ex Ante and Ex Post Epistemic Perspectives on Harm Many kinds of information affect the sorts of potentially harmful conduct people would find it acceptable for others to engage in: the general risks associated with ² Stephen Perry, “Responsibility for Outcomes, Risks, and the Law of Torts,” in Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge, Mass.: Cambridge University Press, 2001), p. 78.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

60

      ?

such conduct, aspects of the individual’s own situation that might put her at greater or lesser objective risk than others, and her own subjective preferences (over possible outcomes, about risk-taking, etc.). All of this information together gives each individual a rough idea of her own expected outcome under a candidate principle. As in earlier chapters, I will refer to decisions made from this epistemic standpoint as the “ex ante POV.” Scanlonian contractualists’ commitment (contra Rawls and Harsanyi) to individual reasons formulated by differentiated, situated selves leads them, plausibly, to endow representative individuals with all these forms of knowledge. In Rahul Kumar’s words, “Valid moral principles are principles that must be justifiable to each person (provided she is appropriately motivated), from her own point of view, with no informational restrictions.”³ But as I noted in Chapter 3, they typically endow each representative with one other piece of information: how she will actually fare under a proposed principle. As in earlier chapters, I will refer to decisions made from the epistemic perspective that endows each representative individual with knowledge of how things will actually turn out for her under a proposed principle as the “ex post point of view” (ex post POV). Ex ante and ex post POVs should not be confused with the temporal position individuals occupy, relative to the moment they must commit to a principle or a course of action based on it. If individuals just “happen” to know the consequences to them of accepting a given principle before they must choose whether to accept it—the epistemic POV typically adopted in trolleyology, as I discuss in Chapter 3—from a temporally ex ante position such individuals will “naturally” have access to what I am calling an ex post POV. In real life, however, people never know with absolute certainty the ex post consequences of acting on a given principle when they must choose whether to act on it. As I discussed in previous chapters, once a hypothetical choice scenario incorporates naturally occurring uncertainty, the only way for contractualists to endow someone in a temporally ex ante position with an ex post epistemic POV is, in effect, to allow him to peek ahead in time to see how things will turn out for him if the principle is adopted. That is to say, they must allow him to view an impending choice through hindsight. An ex post POV should be understood to refer interchangeably to these two ways of acquiring knowledge of the ex post consequences of adopting a given principle before it is acted on: by having those consequences stipulated to be certain ex ante, or by allowing potentially affected individuals to peek ahead to ex post outcomes. From the start, Scanlon staked his version of contractualism on the moral imperative of the ex post POV. To quote an early version of the argument for an ex post POV: ³ Rahul Kumar, “Defending the Moral Moderate: Contractualism and Common Sense,” Philosophy and Public Affairs 28, no. 4 (Oct. 1999): p. 295 (emphasis added).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

          

61

Suppose that A is a principle which it would be rational for a self-interested chooser with an equal chance of being in anyone’s position to select. Does it follow that no one could reasonably reject A? It seems evident that this does not follow. Suppose that the situation of those who would fare worst under A, call them the Losers, is extremely bad, and that there is an alternative to A, call it E, under which no one’s situation would be nearly as bad as this. Prima facie, the losers would seem to have a reasonable ground for complaint against A.⁴

In What We Owe To Each Other, Scanlon restates the case for an ex post POV as follows: “In assessing the rejectability of [a] principle” based on burdensomeness, “we can begin . . . by taking the maximum level of burdensomeness and asking whether that would give a potential agent reason to reject the principle.” Where the activity in question has benefits for many people but “involve[s] risk of serious harm to some . . . [i]t is obvious what the generic reasons would be for rejecting such a principle from the standpoint of someone who is seriously injured.”⁵ When combined with the maximin function built into Scanlon’s Greater Burden Principle (GBP), an ex post POV produces the following result: whichever representative hypothetical person, peeking ahead, discovers she will be the one most seriously harmed by adopting a given principle may reasonably reject it on that basis, provided that the harm to her will be serious as an absolute matter and some alternative course of action with a less-bad worst outcome is available. The result, writ large, is that virtually every action can reasonably be vetoed by someone. Indeed, avoiding just this form of moral gridlock was one of Rawls’s motivations for insisting on his much more stringent version of an ex ante POV built into the Original Position.⁶

⁴ Scanlon, “Contractualism and Utilitarianism,” pp. 122–3 (emphasis added). It has been suggested that Scanlon’s substitution of representative types and generic reasons signals a shift to an ex ante POV. Jeffrey Brand-Ballard, “Contractualism and Deontic Restrictions,” Ethics 114, no. 2 (Jan. 2004): pp. 269–300. There is some textual support for that reading, but on balance I don’t think it can be what Scanlon intended. As Elizabeth Ashford has suggested, to read Scanlon that way is to assume that he has jettisoned sub rosa the central commitment of his version of contractualism: that every candidate principle must be judged based on its actual effects on particular individuals (that is, from an ex post POV). Elizabeth Ashford, “The Demandingness of Scanlon’s Contractualism,” Ethics 113, no. 2 (Jan. 2003): p. 277. The better reading, I think, is that the substitution is meant to avoid the impossible informational burdens of having to consider what every potentially affected individual actually would prefer, as well as to screen out idiosyncratic or objectionable reasons that are not plausible candidates for universal agreement. ⁵ Scanlon, What We Owe to Each Other, p. 207 (emphasis added). ⁶ John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press 1971), p. 140. On the other hand, as I discuss in Chapter 8, the assumption of extreme risk aversion built into Rawls’s Difference Principle threatens moral gridlock by a different route. For more recent discussions of the gridlock problem, see Sophia Reibetanz, “Contractualism and Aggregation,” Ethics 108, no. 2 (Jan. 1998): pp. 302–4; Ashford, “The Demandingness of Scanlon’s Contractualism,” pp. 298–9; Derek Parfit, On What Matters, vol. 1 (Oxford: Oxford University Press, 2011), pp. 348–9.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

62

      ?

Scanlonian contractualists, following Scanlon’s lead, initially adopted an ex post POV as well,⁷ believing that it was logically entailed by their dual commitments (i) to adopt Scanlon’s “individual reasons” restriction, which limits the reasons that count for or against principles to reasons held by a single individual, without regard to the number of other individuals who share those reasons; and (contra Rawls) (ii) to endow individuals with knowledge of their differentiated, situated selves when they choose the general principles they wish everyone to live by.⁸ As I suggested in Chapter 3, an ex post POV does not follow from either of those commitments. Contractualists’ conclusion that it does conflates the question of whether people should be stripped of knowledge they possess at the time agreement is sought (no, if one is committed to thick selves) with the question of when agreement should be sought. Having committed to an ex post POV for weighting complaints against a candidate principle, the challenge facing contractualists is to find a way to confine its role so as to avoid moral gridlock, without going all the way to an ex ante POV. In the next section, I consider the principal strategies contractualists have explored.

2. Finding a Viable Middle Way In Section 2.1, I start with the qualification that has implicitly carried the laboring oar in most contractualist arguments: limit the ex post POV to principles that, as of the epistemic moment that hypothetical agreement is sought, are “actually” known to be certain to cause harm to an identified victim. All other cases, in which outcomes are uncertain at the moment of agreement, are handled by some version of an ex ante POV. I then take up a variety of other strategies. The first group (Sections 2.2 and 2.3) adheres to a pure ex ante POV, but gives a veto to those who face disproportionately greater expected harms if a given principle is adopted. I return to this strategy at much greater length in Chapter 6, when I consider the revisionist contractualist position, which has renounced an ex post POV in favor of an ex ante one. Where relevant, distributionally sensitive ex ante POVs can provide a meaningful and (I believe) desirable alternative to conventional aggregation. But their relevance in ⁷ Michael Otsuka, “Risking Life and Limb,” in Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glen Cohen, Norman Daniels, and Nir Eyal (Oxford: Oxford University Press, 2015); Reibetanz, “Contractualism and Aggregation,” pp. 296–311; James Lenman, “Contractualism and Risk Imposition,” Politics, Philosophy and Economics 7, no. 1 (2008): pp. 99–122; Kumar, “Defending the Moral Moderate,” pp. 275–309. F. M. Kamm and Arthur Ripstein reach the same conclusion from a noncontractualist perspective. See F. M. Kamm, Intricate Ethics (Cambridge, Mass.: Harvard University Press, 2007); Arthur Ripstein, Force and Freedom (Cambridge, Mass.: Harvard University Press, 2009). ⁸ Reibetanz, “Contractualism and Aggregation,” p. 303.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

63

the real world is limited. Even where relevant, contractualists are likely to balk at applying them when the social costs of doing so become substantial. The second group starts with a pure ex post POV (Sections 2.4, 2.5, 2.6) or a pure ex ante one (Section 2.7), but then qualifies it in a way that pushes it in the other direction. The open question with all of these approaches is whether, when all the dust settles, the hybrid principles that result are meaningfully distinct from aggregation. The third group (Sections 2.8 and 2.9) tries to combine ex ante and ex post POVs in some fashion—in the first case by judging some cases from an ex ante POV and some from an ex post, in the second by requiring that all cases be considered from both POVs, in a kind of reflective equilibrium. For different reasons, I believe neither approach can generate a viable alternative to aggregation.

2.1 Certain v. Uncertain Harms As discussed in earlier chapters, one of the most powerful intuitions running through the nonconsequentialist literature is that conduct that is deemed certain to result in harm, typically to identifiable others, is factually and morally distinct from conduct that poses a mere risk of harm. Contractualists typically cash out this intuition by weighting complaints about harms deemed certain to result at the full value of the ex post harm to the hypothetical Losers, and complaints about harms uncertain to result at something less than that (sometimes expected value, sometimes a more complicated calculus).⁹ The importance contractualists and other nonconsequentialists attach to the distinction between absolutely certain consequences and possible, probable or (in some cases) even statistically certain consequences is neatly illustrated by Frances

⁹ Different authors have drawn the line between “certain” and “uncertain” harms in different places. Reibetanz, Scanlon, and Kumar, for example, require that the identity of the victim (or, in Scanlon’s case, a hypothetical stand-in for the victim) be known with certainty in order for the would-be victim’s complaint to be weighted at the full ex post value of the harm; any complaints about possible harms to potential victims must be discounted by the probability of their befalling a particular potential victim. See Reibetanz, “Contractualism and Aggregation,” pp. 301, 308; Scanlon, What We Owe To Each Other; Kumar, “Defending the Moral Moderate,” p. 295. Otsuka requires only that it be certain that someone will be harmed in order for the harm to any potential victim to be weighted at the full ex post value of the harm—a difference that, in most large-number samples, will put almost all risky conduct on the “certain” side of the line. He would also weight harms that fall on the “uncertain” side of the line differently, discounting a given individual’s risk not by the probability that he in particular will be harmed, but by the probability that someone will be harmed. Both differences push the scheme strongly towards an ex post valuation across the board, with resulting moral gridlock. Otsuka, foreseeing this difficulty with the second difference, suggests that in valuing uncertain harms, we could revert to individualized expected value, at least for individuals who are expected to benefit over their lifetimes if society permits reasonable risks. See Otsuka, “Risking Life and Limb.” As discussed in Section 2.2, this exception is likely to swallow his alternative rule.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

64

      ?

Kamm’s Ambulance Cases, referred to briefly in Chapter 3. Ambulance I requires us to decide whether the town may authorize its ambulances to speed on the way to the hospital, whenever it is the case that doing so will save five sick passengers for every one pedestrian killed as a result of the speeding. Kamm first poses the case from a straightforward ex ante POV: may the town acquire an ambulance and permit it to speed? Yes, says Kamm: “[P]reventing a greater number of equally great losses (deaths) seems sufficient reason” to justify it. In support of that conclusion, she observes (significantly) that “[p]eople take risks of death all the time so as to increase their probability of survival, as well as for lesser goods.”¹⁰ In Ambulance II, Kamm asks whether, in addition to authorizing the driver to speed, the town may precommit (through technological constraints or by verbal agreement) not to allow the driver to brake to save the pedestrian’s life, whenever it is the case that the time lost by braking will cause five patients inside the ambulance to die. Clearly such an agreement would be impermissible, Kamm concludes, because even if everyone consented to that decision rule ex ante, “the death would be imposed deliberately at a time when we know that it can no longer be in the interest of the [pedestrian] to risk it. It was just in her interest ex ante to run the risk of being put in this position later.”¹¹ Not surprisingly, many share Kamm’s intuition that the driver in Ambulance II must brake to save the pedestrian, notwithstanding the pedestrian’s earlier consent to be sacrificed under such circumstances. In defense of that intuition, James Lenman argues that a policy that “is not acceptable at every time [to every person] is plausibly acceptable at none.” Since, in Ambulance II, there is a moment in time in which one person—the now identified about-to-be victim of the speeding ambulance—can complain that the driver is not taking “reasonable precaution against [her] coming to harm,” any prior consent the victim might have given in Ambulance I to resolve this by aggregation is retroactively rescinded.¹² If we eliminate the original agreement in Ambulance I (which in Kamm’s view is superseded in any event in the case of Ambulance II), the same intuition explains contractualists’ widespread agreement about how to handle cases in which the outcomes of available actions are stipulated to be known with certainty ab initio. These include cases in which we must decide whether to inflict certain harm on identified victims to save more numerous others (trolley-type problems), as well as cases in which we must decide whether to rescue one identified person from certain harm or death at the cost to others of qualitatively lesser goods (lesser harm, money, lost pleasure). Like the rescission cases, in both the “certain” harm and “certain” rescue cases there is a moment in time at which it is unambiguously not in the interest of the about-to-die person to consent to be sacrificed for the

¹⁰ Kamm, Intricate Ethics, pp. 273, 275. ¹¹ Kamm, Intricate Ethics, p. 273. ¹² Lenman, “Contractualism and Risk Imposition,” p. 117.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

65

greater good. The only difference from the rescission cases is that that moment occurs the first and only time we seek hypothetical consent. In such situations, as in the rescission cases, contractualists would permit the person who knows her life is at stake to withhold consent based on that knowledge. Indeed, Kamm defends the duty of easy rescue in almost identical terms to the right not to be bound by one’s prior consent: [S]uppose that we did argue even for the permissibility of investing in cures for truly minor problems affecting many, such as headaches, rather than in a cure of a rare fatal disease, on the ground that it is reasonable for each person to take a small risk of being the one who will die in order to have headache cures at hand for his many, certain-to-occur headaches. This does not imply that here and now we should not save someone from dying from a rare fatal disease, if we could, rather than cure millions of headaches . . . . It could be wrong to leave him to die on the grounds that it was reasonable ex ante, in order to produce the aspirin for headaches, for each person to take a small risk of dying because no help for him would be available when he fell fatally ill. It is here and now that the irrelevant utilities of headache cures do not aggregate to override saving the life.¹³

Although Scanlon does not explicitly invoke the distinction between certain and uncertain consequences, I believe it similarly drives his conclusion that our duty to rescue someone from imminent harm is much stronger than our duty not to put her in harm’s way to begin with. Suppose, says Scanlon, Jones is trapped under live transmission wires, which are inflicting severe pain on him (although no permanent physical injury). In order to rescue him immediately, we would have to cut off transmission of the World Cup soccer game for twenty minutes, depriving hundreds of millions of viewers of the pleasure they would have received from watching those twenty minutes live. Alternatively, we can leave Jones to suffer for an hour until the game is over. Scanlon concludes that of course we must save Jones now, because it is not permissible to leave someone to suffer an hour of severe pain, merely to give recreational pleasure to others, no matter how numerous those others are.¹⁴ But suppose we are deciding instead whether to build a new system of transmission towers that will improve the quality of reception for millions of television viewers. We know that in the course of a project like this, some number of workers are likely to suffer harms as great as Jones’s. Scanlon concludes it is permissible to proceed with the project, notwithstanding that it “involve[s] risk of serious harm to others,” provided the builders use “adequate precautions” to reduce that risk. (I take up the meaning of “adequate precautions” in Section 2.6. For now, it suffices ¹³ Kamm, Intricate Ethics, pp. 36–7 (emphasis added). ¹⁴ Scanlon, What We Owe to Each Other, p. 235.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

66

      ?

to say that it is clearly intended to permit us to impose some risks of serious harm on the would-be Losers that we would not be permitted to impose if the consequences and/or the victim were known with certainty ex ante.) The reason we are permitted to sacrifice the construction workers even as we must save Jones, says Scanlon, is that the construction case involves “failing to prevent accidental injuries rather than either intentionally inflicting serious harm on a few people or [as in Jones’s case] withholding aid from people who need it.”¹⁵ While Scanlon does not explicitly invoke the uncertain/certain harm distinction here, I think this is what he must mean in contrasting accidental and intentional harms. That is to say, “accidental” refers to situations in which we act without full knowledge of the consequences (“accident” in the sense of “Gee, I didn’t see that one coming”) and “intentional” refers to the opposite.¹⁶ The resulting hybrid POV seems difficult to justify. It presupposes distinctions in kind, both factual and moral, between certain and uncertain harms that are doubtful at best. And it holds the reasonableness of a given principle hostage to the adventitious factual posture of the hypothetical that must be resolved pursuant to it, producing general principles that are both morally arbitrary and unstable. I take up each of these concerns in turn. 1. Are acts that cause harm and those that impose a risk of harm factually distinct forms of conduct? As discussed in previous chapters, the philosophical literature on harm to others has been organized around a set of canonical hypothetical choices that, like Kamm’s Ambulance II, involve identified victims and consequences that are stipulated to be known with certainty ex ante. But in the real world, no conduct is absolutely certain to harm others, judged ex ante. This is true even of harms that are intended (in the strong sense of desired or the weak sense of foreseeable); a fortiori, it is true of accidental harms as well. In predicting the unintended consequences of conduct, we are therefore always dealing with greater and lesser degrees of uncertainty. Thus, from a temporally ex ante perspective, the problem of harm is the problem of risk. Furthermore, the probability that harm will result from a given act is itself an artifact of how the question is framed—in particular, the universe of acts over ¹⁵ Scanlon, What We Owe to Each Other, p. 236 (emphasis added). Consent also plays a role in this and other hypotheticals. Scanlon explicitly requires that, for the construction project to be permissible, the construction workers must consent to the risk they are assuming. Consent is surely morally relevant, but its relevance is complicated, and cannot ultimately be separated from our judgment about the reasonableness of the risk to which someone has implicitly or explicitly consented. I discuss this point further in Section 2.6. ¹⁶ It is not clear what else Scanlon could have in mind here, since, in the two scenarios he describes as intentional, he clearly does not mean intent to refer to motive. What the two examples have in common is that the actor knows with certainty the consequences of acting or not acting.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

67

which we predict consequences and the epistemic POV from which we assess them. The failure to distinguish these framing decisions from the underlying causal relationship between act and consequences has led to considerable confusion in the literature. Three decisions are critical here: whether to take the smallnumber or large-number perspective; if we take the large-number perspective, whether to treat statistical certainty as certainty; and, most importantly, at what epistemic moment to assess probabilities. Small-number versus large-number perspective. Suppose we estimate that if we distribute a new flu vaccine, roughly one out of every one million people inoculated will have an adverse reaction, resulting in death. We plan to inoculate ten million people. As I discussed in Chapter 2, there are two ways to describe the likelihood that death will result from the inoculations, sometimes differentiated as “probability” versus “frequency.” Probability describes the odds that any given person who receives the vaccine will die (one in a million). Frequency describes the total number of expected deaths if we inoculate ten million people (ten). Both are describing the same set of acts, predicted to cause the same consequences with the same likelihood; they are simply describing the likelihood from different perspectives. The nonconsequentialist literature, however, frequently treats the two formulations as if they were describing different acts that entail different degrees of risk, with different normative implications. James Lenman, for example, argues that if the question on the table is, may an individual go for a drive, knowing she is thereby putting others at some minuscule risk of harm, the answer is yes, in part because “the risks imposed are very small.” But if the question is, what speed limit should the government adopt, we should apply a much stricter standard, in part “because the small risks governments impose are typically imposed over very large populations . . . so that it becomes an actuarial certainty some people will be harmed.”¹⁷ Read literally, the contrast Lenman sets up here is between prohibiting a given individual to drive and permitting the government to regulate the manner in which all drivers drive. Surely Lenman will conclude that individuals are permitted to drive and that they are required to drive prudently, whether the question is posed from the perspective of what individuals owe to each other or from the perspective of how government should regulate driving. The real contrast he means to pose, I assume, concerns the level of care required in each case. To make this concrete, suppose the question is, how fast is it permissible to drive through a city street? Lenman’s analysis suggests that he would specify a higher speed limit (say, 40 mph) if the question is framed in terms of the duty of care I owe individually to others, and a lower speed limit (say, 25 mph) if it is ¹⁷ Lenman, “Contractualism and Risk Imposition,” p. 109 (emphasis added).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

68

      ?

framed in terms of the level of care the government should legally require of all of us. The reason for the difference is that the odds are very low that I in particular will kill anyone if I drive at 40 mph rather than 25 mph (on a particular occasion? over a lifetime?), but it is a statistical certainty that, over a sufficiently large number of car trips governed by the speed limit, at least one additional person will be killed if the speed limit is set at 40 mph rather than 25 mph. Surely it cannot make sense to have anything turn on whether we frame identical conduct imposing identical risks in probabilistic or frequentist terms. Is statistical certainty certain enough? Virtually all acts, if measured over a large enough number of occurrences, are statistically certain to produce at least one death or serious injury. If statistical certainty counts as certainty for these purposes, then virtually all conduct is certain to result in someone’s death or serious injury if we assess certainty over a large enough class of events. A few nonconsequentialists, including Lenman, have taken that view.¹⁸ Most have not, requiring some form of “objectively determined certainty” (about the harm that will result, about the identity of the victim) in order to treat harm as certain to result.¹⁹ For reasons suggested above, I don’t think the distinction between “objective certainty” and “statistical certainty” can withstanding scrutiny. But even if it could, either way we resolve the issue—treat statistical certainty as certainty or treat it as uncertainty—virtually all conduct will end up on the same side of the divide, pushing us to either a pure ex ante POV or a pure ex post POV across the board. Knowledge is a dependent variable. Probabilities are epistemic facts. They describe what is known or knowable about future consequences as of the particular moment at which we assess them. In the real world, the appropriate epistemic POV from which to assess consequences is determined by practical needs, and will generally be whatever we know at the moment we must commit ourselves to some course of action. If we are trying to decide whether a fail-safe valve on an oil rig is safe enough, we have compelling practical reasons to act on the basis of the best information available as of the moment we must choose in reliance on that information. In contractualist thought experiments, in contrast, the appropriate POV from which to assess the acceptability of general principles to guide our actions is not determined by exogenous circumstances. It is internal to the contractualist argument and needs to be settled by appeal to normative criteria.

¹⁸ For a particularly strong version, which not only treats statistical certainty as certainty but also treats statistical foresight as providing the mens rea necessary for murder, see Lisa Heinzerling, “Knowing Killing and Environmental Law,” N.Y.U. Environmental Law Journal 14, no. 3 (2006): pp. 521–34. ¹⁹ Otsuka, for example, would require “objective certainty” that a predetermined individual will die, but does not require that decision-makers themselves know who that individual is. See Otsuka, “Risking Life and Limb.”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

69

De facto, most contractualists have equated the moment of agreement on general principles with the moment at which a particular group of people must decide how to act in a particular situation pursuant to those principles—whether to flip the switch on the trolley, whether to rescue Jones, etc. When combined with the Scanlonian commitment to let individuals act on the basis of whatever knowledge they are imagined actually to possess at the time of agreement, the result is an epistemic POV that is contingent on the happenstance of the factual posture in which a real or hypothetical dilemma arises. If, as in Kamm’s Ambulance I, we imagine that each person must decide whether to allow the ambulance to speed at a time when she has no reason to think that her expected outcome deviates from anyone else’s in the town, we conclude that she could reasonably assent to it. If she happens to have the opportunity to revisit the issue in Ambulance II, when an ambulance is about to mow her down, we conclude it is no longer reasonable to hold her to her assent in Ambulance I, and she must be allowed to change her mind, based on the updated information about her selfinterest. That is the burden of Lenman’s argument that a policy that “is not acceptable at every time [to every person] is plausibly acceptable at none.” It is hard to overstate the argumentative consequences of contractualists’ having de facto equated the moment people must agree on general principles with the moment that some subset of them must choose how to act in a particular situation pursuant to those principles (the “here and now,” in Kamm’s words). As I suggested in Chapter 3 with respect to trolleyology, if contractualists were to decide on further thought that the only morally appropriate POV from which to choose general principles governing the risks we may impose on each other is before any of us knows exactly what risks we will face and how they will play out for us—to conclude, that is, that agreement must be reached from some version of a truly ex ante POV—then by contractualists’ own acknowledgement, virtually all socially useful but potentially harmful conduct may be (must be?) regulated by some form of interpersonal aggregation, in which the number of potentially affected individuals and the probability of harm or benefit to each of them counts. A few questions will remain in nonaggregationists’ domain: Are there some future interests that individuals should not be allowed to put at risk? What do we do about individuals who are in a position to know ex ante that the expected outcomes for them are worse than for others?²⁰ But all the remaining moral quandaries are swept away. No more trolley problems, no more Ambulance Cases, no more technicians trapped under transmission wires, because we are no longer interested in what general principles someone would agree to once she knows for a fact that her own life is at stake in the choice. The only morally relevant question is, what would she have said before she knew that?

²⁰ For discussion of the latter issue, see Section 2.2.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

70

      ?

If, on the other hand, contractualists really mean to insist on the moral imperative of an ex post POV—insist, that is, that before deciding whether to adopt a given principle individuals have a right to know how things will turn out for them in their actual lives if it is adopted—the result, inescapably, will be moral gridlock. But instead, contractualists have, de facto, split the difference. They have equated the moment people should agree on a general principle with the moment some subset must resolve an exigent crisis pursuant to that principle, and then endowed the members of that subset with whatever knowledge each of them happens to possess at that moment about how things will turn out for them, should the exigent crisis be resolved in accordance with the proposed principle. The POV that results is an odd hybrid of ex ante and ex post. If, at the moment of exigent choice, someone happens to know that adopting Principle X will result in her certain death, her complaint about Principle X will carry the full disvalue of death (Ambulance II). If she knows only that she, like everyone else, faces a one in a million chance of death, her complaint will be discounted by that probability (Ambulance I). If, at the moment of exigent choice, everyone is certain about their fates and every principle will have the same worst outcome (e.g., someone dies), some other supplementary rule is needed to break the tie (e.g., Scanlon’s tiebreaker rule which would let the numbers count, the doing/allowing distinction which gives the victory to the pedestrian in Ambulance II). If at that moment everyone is uncertain about her fate and everyone faces identical risks, we would expect representative individuals to choose the principle that yields the best expected aggregate outcome (the “right” result in Ambulance I). But if the parties have asymmetric information about their fates, the hybrid rule produces outcomes that even its proponents recognize to be unacceptable. The problem has been much discussed in the context of Sophia Reibetanz’s “Unexploded Mines” hypothetical, the facts of which are this: One hundred workers are working in a field, in which an unexploded mine is known to be located. A nearby person, Y, is the only one who can disarm the mine. If Y disarms it, Y is certain to get pneumonia; if Y doesn’t disarm it, one of the 100 workers is certain to be seriously injured, but it is unknown which one. The injury to that one worker will be 10 times worse than pneumonia will be for Y.²¹

Under the hybrid rule, Scanlon’s GBP produces the “unacceptable” answer that Y doesn’t have to disarm the mine. Because Y knows for certain what will happen to him ex post if he disarms the mine, Y’s complaint about pneumonia gets

²¹ Reibetanz, “Contractualism and Aggregation,” pp. 302–3.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

71

weighted at the full disvalue of pneumonia. On the other side, since “it is not known, in advance, which of the hundred laborers will encounter the mine,” each laborer’s complaint about the potential severe injury to himself is discounted by the odds he will incur it, yielding a complaint of 1/100 x 10x, which is 1/10th of Y’s. Contractualists have suggested a number of ways to fix the problem.²² But the problem itself is an artifact of the way contractualists construct the choice situation from which a hypothetical agreement on general principles is sought: (i) locate the moment of agreement on a general principle to govern choice X at the moment some arbitrarily selected group of people must make an exigent X-type choice in accordance with it; (ii) limit the representative persons whose complaints are to be counted to the members of that group; and then (iii) weight each of their complaints about candidate principles based on whatever information they just happen to possess at that moment about how things will turn out for them if a given principle is adopted. What is the right epistemic moment from which to seek hypothetical agreement on general principles? I don’t think there is a simple, or single, answer to the question. It depends in part on one’s normative commitments, in part on the nature and scope of the principles in question. But the answer that much of the contractualist literature has, de facto, adopted seems to me singularly hard to defend. At a minimum, it needs to be defended. I return to this question in Chapter 6, when I consider more recent developments in contractualism. 2.

Right result, wrong reason?

None of this is to deny that most people respond very differently to acts that are virtually certain to harm others than they do to acts that expose others to a “mere” risk of harm. If that difference cannot be explained by moral or factual distinctions between “certain” harm and risk, what explains it? One obvious answer lies in the emotional and psychological valence of immediacy and certainty. Contemplating imminent death to an identified victim tends to concentrate the mind wonderfully on the horror of causing or permitting it to happen. For the person facing imminent death, the horror of knowing that others could have saved her but are choosing not to is likely swamped by the prospect of death itself. But she is not the only (or indeed maybe even chief) party whose sensibilities are at risk in such cases. There is the agent who, in order to follow through on a prior agreement, must now knowingly kill (or fail to rescue) an identified victim. There are also the bystanders who must witness an “intentional” killing or avoidable death up close and personal. And when that death or serious harm can be prevented with mere money or the trivial loss of pleasure—Baby ²² See, for example, Otsuka, “Risking Life and Limb”; Reibetanz, “Contractualism and Aggregation,” p. 303 n.12.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

72

      ?

Jessica in the well, Scanlon’s Jones trapped under transmission wires—it seems more than a little cold, maybe downright inhuman, to refuse to incur those costs simply because, in the aggregate, the costs are large. Perhaps, as some have argued, our heightened emotional response to imminent, identifiable harms itself has a moral basis.²³ Perhaps it is explicable only in psychological or emotional terms. In a practical sense, it may not matter. If, for whatever reason, it is the case that none of us will want to stand by and do nothing while the driver in Ambulance II mows down one pedestrian to save five passengers, as we all had assented to in Ambulance I, and if those feelings are going to be strong enough, in the event, that neutral agents, interested parties, and observers will revolt against the prior agreement not to interfere and will condemn anyone who doesn’t repudiate it too, then such feelings have to be taken into account in some fashion, if we are to cobble together a stable policy response to the problem. But how exactly we should take them into account depends (inter alia) on whether we think we are dealing with two independently valid responses to different questions or simply temporally inconsistent responses to the same one. 3.

A Pyrrhic victory?

Resolving risk from an ex ante POV and certain harms from an ex post POV solves the moral gridlock problem for most real-world problems, for the simple reason that in real life, people generally must commit to a course of action while the outcome is still uncertain, and rarely have a chance to withdraw their consent once their own fate becomes clear. It is easy to miss this truth, coming at the problem through the lens of the contemporary philosophical literature on harm to others, because of the prominent role given to hypotheticals involving “certain” harms (trolley problems, etc.) and the marginalization of risk. One of the unfortunate consequences of this inverted picture of reality has been to encourage contractualists to think that if they cede the problem of risk to aggregation but require “certain” harms to be resolved by nonaggregative principles, they will have carved out a meaningful middle way—one that retains a substantial role for nonaggregation while avoiding moral gridlock. In fact, as I discuss in Chapter 2, under the “we get certainty and aggregationists get uncertainty” compromise, almost all of the work of regulating potentially harmful conduct is ceded to aggregation. The odd assortment of one-off easy rescue and “certain” harm cases that populate the philosophical literature—trolley problems, Jim and the Indians, Ambulance II—in which we must choose between two bad outcomes certain to occur depending on the choice we make are, in effect, just an indulgence of our nonaggregationist intuitions. ²³ See, for example, Sarah Miller, “Need, Care and Obligation,” Royal Institute of Philosophy Supplement 57 (2005): pp. 141–3; Lenman, “Contractualism and Risk Imposition,” p. 116; Parfit, On What Matters, vol. 2, p. 211.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

73

I do not think it could be otherwise, because the implications of refusing to aggregate in determining how to regulate harm to others will be palatable only as long as the occasions for deploying it are rare. As I suggested in Chapter 3, whether or not one agrees that it is morally wrong to kill one identified person to save five, the fact is, social life will go on pretty much as before if every time we encountered that dilemma in its various postures, we let the five (or 500 or 5,000) die.²⁴ But if acts that read as “certain” to harm others were as common as ones that read as “mere” risks, the social costs of refusing to let the numbers count would be morally unacceptable to all. Scanlon’s response when the going gets tough under the GBP leaves little doubt where he would come out. Scanlon recognizes that if the maximin decision rule built into the GBP were applied to all activities that posed a risk of serious harm to someone, we could not (for starters) build sports stadiums or skyscrapers, drive cars, or fly airplanes, because the risk of death to bystanders (should it materialize) would clearly trump the inconvenience or foregone pleasure if we ban such activities. But, faced with that implication, instead of biting the bullet, Scanlon abandons the GBP as “too confining.”²⁵ In its place, Scanlon proposes that we reach a reasonable compromise among competing interests, one that takes into account both our interest in “avoid[ing] bodily injury” and also the “cost[s] from the point of view of potential agents” of “a general prohibition on . . . acting” in a fashion that would put others in harm’s way.²⁶ In short, we should accept the very principle that the “individual reasons” restriction was meant to rule out: that relatively trivial ex post benefits to one group can offset serious costs to another, provided the former group is large enough and the latter group small enough.

2.2 Allow Ex Ante Disadvantaged Parties to Reject an Unequal Distribution of Ex Ante Risks As I noted above, individual agreement sought from an ex ante POV will tend to approximate the aggregative solution, provided that everyone faces roughly the same expected costs and benefits from a proposed activity. What if they don’t?

²⁴ Indeed, contractualists are often unwilling to live with the implications of the ex post POV even in cases of “certain” harms. Rahul Kumar, for example, qualifies his commitment to an ex post POV for certain harms as follows: you may use the information you possess about your own fate only to the extent doing so is “impartially justifiable,” which he equates with the probability that an average person will benefit (suffer) from the principle. Rahul Kumar, “Defending the Moral Moderate,” p. 298. I’m not entirely certain how Kumar intends this to operate, but it appears to weight all benefits and harms on an ex ante basis, producing straight aggregation. ²⁵ Scanlon, What We Owe to Each Other, p. 209. For analysis of Scanlon’s inconstancy to the GBP in the context of the duty to rescue, see Ashford, “The Demandingness of Scanlon’s Contractualism,” pp. 298–302. ²⁶ Scanlon, What We Owe to Each Other, pp. 204, 205.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

74

      ?

Discussion of this issue in the contractualist literature has focused on a hypothetical involving a community of Amish farmers who, by virtue of living under a flight path, face a greater than average risk from falling planes while deriving few if any benefits from air travel.²⁷ But the same issue arises in the real world in less fanciful forms. Land use regulation is a common site for such intergroup conflicts, because the negative consequences of a particular land use are often geographically concentrated while the benefits are more widely diffused. A standard case here would be the siting of a toxic waste dump. Placing the dump near a poor community (Poorville) may well be optimal from an aggregative (welfarist) perspective, because the dump will typically depress land values in Poorville less than in nearby richer communities. But because they will bear a disproportionate share of the risks from the toxic dump and at best enjoy only a pro rata share of the benefits from whatever activities generate the toxic waste, the residents of Poorville could reasonably object that the optimal solution is unfair to them. A Scanlonian contractualist is committed to taking those objections seriously. But what exactly does that entail? We could (and sometimes do) require those disproportionately benefitted by siting the dump in Poorville (the company generating the waste, the larger polity that gets at least their fair share of the benefits without bearing any of the costs) to compensate Poorville for expected or actual harm. But from a contractualist perspective, this only partially addresses the complaint of the residents of Poorville, who (by hypothesis) don’t not want the toxic dump there at all. When the potential harm to them is severe and not fully compensable by money, they might well refuse to agree to the siting for any amount of money (or at least any amount anyone is likely to offer them). Now what? Broadly speaking, there are three possible responses. The first is to force Poorville to accept the toxic dump along with some amount of cash compensation that is presumably less than the real cost to them. Writ large, this amounts to a welfarist regime without a strict Paretian constraint. My guess is that most contractualists will resist this solution, at least for serious risks to health. If not, then they have ceded the problem of risk to aggregation full stop, leaving to nonaggregative principles only the question of compensation. Second, we can judge the fairness of the ex ante distribution of risks over some larger class of activities. If, over some suitably broad class, the (expected) losers on one risky project are likely to be the (expected) winners on another, we can treat those two events as offsetting. I take up this alternative in Section 2.3. For now, I note only that where death is one of the risks, many contractualists will balk at that solution.

²⁷ See Lenman, “Contractualism and Risk Imposition,” p. 121, n. 40; Ashford, “The Demandingness of Scanlon’s Contractualism,” pp. 298–9. The example is a variant of a hypothetical originally posed by Scanlon, in which the group facing a disproportionate share of the downside risks was the poor rather than the Amish. See Scanlon, What We Owe to Each Other, pp. 208–9.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

75

Third, we can insist that Poorville be given a disproportionate say in the initial siting decision, at the extreme a veto over it. This solution seems more in the spirit of Scanlonian contractualism. It may well be both appropriate and doable in discrete cases, where the social costs of giving the losers a veto are manageable and the expected costs to them if the project goes forward are severe (health hazards, destruction of a community, etc.). Again, land use disputes are a prime example here. That the losers in such cases are often the poor strengthens the argument, in my view and the view of many others, for giving them a veto on fairness grounds. But contractualists’ willingness to bite the nonaggregationist bullet in these cases is likely to run out pretty quickly. Many large-scale construction or public works projects concentrate expected costs on one group, and many of the expected costs are not fully compensable with money. If every disproportionately burdened group were given a veto over the project in question, we would be back to gridlock pretty quickly. Giving the group a disproportionate say over the principle but something short of a veto would lessen the risk of moral gridlock, but at the cost of leaving the group still vulnerable to demands that it sacrifice its own welfare for the good of the whole. And of course, giving anything less than a veto raises the question of degree: Should the disproportionately burdened minority get two votes for every one of the majority? Ten votes? How should we go about deciding this? It is notable that Scanlon’s willingness to give the expected losers a veto in such situations runs out almost immediately. Having posed a variant of the Amish problem, Scanlon concludes that of course air travel cannot be banned, because to do so would be “too confining,” notwithstanding that his own GBP unambiguously requires that result.²⁸ Finally, it is critical to keep in mind the limited reach of this exception to aggregation, even in theory. It is triggered only when the expected net benefits of an activity are distributed unequally. It does not apply when expected costs and expected benefits are distributed roughly equally, but ex post there will big losers. Many if not most accidental harms fall in the latter camp.

2.3 Aggregate risks intrapersonally Because the Amish pose such an extreme case, their complaint that they are bearing the risks of social progress without expecting to reap any of the rewards might well hold, whether we assess air travel in isolation or bundle it in with other risks the Amish are likely to face over their lifetime. But in the typical case, the choice between isolated and bundled risks will dramatically change the results

²⁸ Scanlon, What We Owe to Each Other, p. 209.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

76

      ?

under the GBP. Arguments for bundling (henceforth, intrapersonal aggregation) have been around for a long time with respect to compensation for harm.²⁹ While the technique has gotten less attention in setting acceptable levels of risk, those who have considered it generally support it.³⁰ In my view, it is difficult to defend not bundling in some fashion. Aggregating risks intrapersonally would undoubtedly change contractualists’ judgment about the reasonableness of risky conduct in some number of cases. But it won’t solve the basic problem facing nonaggregationists: an ex ante POV proves too little and an ex post POV too much. It simply replicates it across bundled groups of activities.

2.4 Take into account all the consequences of a principle What contractualists are seeking agreement on is not rules governing individual actions but rather “general principles of action.” In Scanlon’s words, this necessitates our taking into account not only the consequences of particular actions, but also the costs imposed on potential agents and society at large from adopting “a general prohibition on . . . acting” in a fashion that would put others in harm’s way.³¹ One of Scanlon’s motivations for adopting this “all things considered” consequentialism is to avoid moral gridlock.³² Whether it can achieve that result without landing us in aggregation full stop is less clear. Scanlon doesn’t say enough about how other costs and benefits are to be taken into account to allow us to answer that question. But the general description Scanlon provides sounds sufficiently like conventional interpersonal summing of costs and benefits that I think it is fair to put the burden on Scanlonians to explain how it differs.

2.5 Tiebreakers Go to the Ex Ante POV Since Taurek’s famous 1977 article, contractualists, like other nonconsequentialists, have sought a way to justify letting the numbers count, at least as a tiebreaker when we are forced to choose between harms of equal severity. As discussed in ²⁹ Daryl Levinson, “Making Government Pay: Markets, Politics, and the Allocation of Constitutional Costs,” University of Chicago Law Review 67, no. 2 (2000): pp. 345–432; George P. Fletcher, “Fairness and Utility in Tort Theory,” Harvard Law Review 85, no. 3 (Jan. 1972): pp. 537–73. ³⁰ Lenman, “Contractualism and Risk Imposition,” p. 108; Reibetanz, “Contractualism and Aggregation,” pp. 299–300; Scanlon, What We Owe to Each Other, p. 237; Otsuka, “Risking Life and Limb,” p. 16. ³¹ Scanlon, What We Owe to Each Other, pp. 171, 203–5, 264, 273. ³² Scanlon, What We Owe to Each Other, p. 170.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

77

Chapter 3, Scanlon has explicitly adopted this “tiebreaker” rule, and it has figured prominently in other contractualist accounts.³³ Whether it offers a viable middle way is unclear. First, there is the question of scope. Almost all conduct carries with it some risk, however remote, of death to someone; outside of the world of philosophical hypotheticals, none carries with it a non-statistical certainty of death. If we ignore probabilities (as we are required to do under a true ex post POV), every case is a tie, because every case could result, in the worst-case scenario, in death. At that point, we have aggregation full stop, in two steps rather than one. We can avoid that conclusion by letting probabilities count in some fashion, but if we do that we no longer need a tiebreaker rule to let the numbers count. We have done it directly, thereby giving up on contractualists’ central commitment to “individual reasons.” Assuming one could come up with a principled way to keep some but not all cases outside the jurisdiction of the tiebreaker rule (I am doubtful), any cases that are outside revert to the ex post POV, in which the numbers do not count. This compromise limits the unpalatable consequences of an ex post POV by limiting the cases to which it applies. But unless the cases that fall outside the tiebreaker solution are few in number and trivial in social importance, I doubt whether contractualists can live with the implications of refusing to let the numbers count in such cases. And if the cases that fall outside the tiebreaker rule are few in number and trivial in social importance, we face the same question raised by the certain/uncertain harms compromise: Does limiting the operation of nonaggregative principles to such cases supply a meaningful middle way, or is it instead just a moral sideshow in a basically aggregative regime?

2.6 Avoidability of Risk Waives Individual Complaints Another middle way that contractualists have explored is to distinguish between permissible and impermissible conduct based on whether the victim could have avoided the risk at a reasonable cost to herself.³⁴ Under Scanlon’s version (which he terms the “Value of Choice” view), “[i]f a person has been placed in a sufficiently good position, this can make it the case that he or she has no valid complaint about what results, whether or not it is produced by his or her active choice.” A person will be placed in a sufficiently good position if the party responsible for the hazard has “done enough to . . . reduce the likelihood that anyone will be injured.”³⁵ One of the things the responsible party can do is to ³³ Scanlon, What We Owe to Each Other; Kamm, Intricate Ethics; Reibetanz, “Contractualism and Aggregation”; Parfit, On What Matters. ³⁴ It is not always clear whether this argument is addressed to the permissibility of the risky conduct or the obligation to compensate victims. My comments should be read to apply only to the former. ³⁵ Scanlon, What We Owe to Each Other, p. 258.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

78

      ?

inform would-be victims of the danger so they can take reasonable precautions to avoid it. Another is to take reasonable precautions herself to minimize the risk. Whether Scanlon’s Value of Choice model offers a viable middle way once again depends on exactly what it means to “do enough” or take “reasonable” precautions. Welfarists would answer the question by finding the optimal tradeoff between the costs and benefits of safety precautions. As discussed in more detail in Section 2.7, it is unclear whether contractualists would answer the question differently, and if so, how.

2.7 Require Greater Safety Precautions than Under Straight Aggregation In the typical case of risky but socially productive conduct, the choice we face is not whether to permit risky act X or forbid it, but rather what level of precaution to require as a precondition of doing X. (To put it another way, forbearing from doing X at all is the limit case of the available levels of precaution.) A number of contractualists have suggested that the middle way lies in requiring a higher level of precaution when undertaking potentially harmful conduct than would be required by a straight summing of costs and benefits. In What We Owe to Each Other, Scanlon suggests one way we might do that: rather than permitting less precaution the greater the social benefits of the activity (as he takes aggregation to do), we should demand the same level of precaution irrespective of the level of such benefits.³⁶ The contrast Scanlon sets up here is based on a misunderstanding of how aggregationists would think about the problem. In standard cost/benefit analyses, three numbers are relevant in determining whether activity X should go forward at all, and, if so, the optimal level of safety precautions to employ: the social benefits of X; the expected harms if X is done with a given level of precaution; and the costs of undertaking that level of precaution. Once it is established that there is some level of safety precautions at which going forward with X will increase social welfare notwithstanding the risks, the social benefits of the project are irrelevant to calculating the optimal level of safety precautions. For a welfarist, the optimal level depends solely on the marginal tradeoff between money spent on safety precautions and increased safety. The optimal level of safety precautions is reached when the incremental cost of additional precautions will exceed the incremental benefits (in increased safety) they will produce.

³⁶ In Scanlon’s words: “We do not think that a higher level of safety must be provided for workers on a building that will benefit only one family as opposed to an apartment house or a public bridge.” Scanlon, What We Owe to Each Other, p. 236.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

79

It is not clear that contractualists would make either of these determinations differently. Notwithstanding the unambiguous implication of the GBP that one may never trade off lives for lesser goods, Scanlon suggests such tradeoffs are permissible: “In each case, in order to defend the practice in question we need to argue that the importance of the social goal justifies creating the risk.”³⁷ Once contractualists conclude that at some level of safety precautions it is reasonable to permit activity X to go forward, it is unclear whether they would determine the required level of safety precautions differently from aggregationists. By way of illustration, consider the following. A city is deciding whether or not to put up a new sports stadium. The city estimates that the stadium will generate subjective entertainment value for spectators equal to $100 per person per year. An expected 100,000 persons will attend games per year, for a total expected benefit of $10 million per year. If $1 million is invested in safety precautions during construction in order to protect passersby, four pedestrians are expected to die. If the amount is raised to $10 million, expected deaths are reduced to two. If it is raised to $50 billion, expected deaths are reduced to one. No amount of money spent on safety precautions will reduce expected deaths to zero. Evaluating this project from an ex post POV, a Scanlonian would have to conclude that the city may not build the stadium, whatever level of safety precautions it takes, because the complaint of the one passerby who (peeking ahead) knows she will be killed will outweigh whatever amount of pleasure spectators would derive in the aggregate over the life of the stadium. As discussed in Section 2.1, however, Scanlonians reject a pure ex post POV if the potential harms to others are uncertain at the time the decision is made whether to build the stadium, as in the real world they will always be. Instead, they permit it to be built notwithstanding the non-zero risk of death to passersby, provided the social benefits of doing so are great enough and “reasonable” precautions are taken. What level of precaution is “reasonable”? As I discuss in Chapter 2, it is often suggested in the contractualist and other nonconsequentialist literatures that if we could do anything more to reduce expected deaths, it would be unreasonable for us not to do it. In the stadium project, that would mean it is unreasonable for the builder to spend any less than $50 billion on safety precautions. But in most arenas of life, as undoubtedly in this project, the point at which further investments in safety will cease to produce any incremental gain in safety is far beyond the projected benefits of the project. The result of interpreting “reasonable” in this fashion will thus be moral gridlock by a different route. No building project will show a net expected benefit given the enormous amount the builder is morally required to spend on safety precautions, and hence no building project may go forward.

³⁷ Scanlon, What We Owe to Each Other, p. 264.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

80

      ?

So, presumably, there is a point at which even contractualists will conclude that an additional investment in safety will produce so little incremental benefit that it is unreasonable to require it to be made, and hence the project may go forward without it. As noted above, for aggregationists that point is reached when the incremental cost of a further increase in safety precautions is greater than the incremental benefit (in increased safety) it will produce. Would contractualists make that calculus differently? As I suggest in Chapter 5, discussions of this issue in the nonconsequentialist literature strongly suggest that the answer is no.

2.8 Limit the Ex Post POV to Serious Harms Another middle way that has been suggested is to handle serious and non-serious harms differently. One approach, adopted by Scanlon, Keating, and others, is a form of threshold deontology: when the potential harm passes some threshold of seriousness, those who, peeking ahead, learn they will be the big losers are given either a veto or the right to demand that their interests be protected by especially stringent precautionary measures. Non-serious harms, on the other hand, are handled by some form of aggregation.³⁸ Like all disjunctive criteria, dividing the universe of potentially harmful conduct in this fashion leaves contractualists with two unattractive alternatives. If a line is drawn and adhered to strictly, it produces results that are morally arbitrary: If one hour of pain is serious enough to be a serious harm, why not fifty-nine minutes? On the other hand, if some flexibility is introduced to counter that arbitrariness, the distinction threatens to unravel because of transitivity issues. More significantly, the worst possible outcome of most risky activity is very bad—bad enough to be considered serious by anyone’s definition of serious. Thus, we are back at the impasse we face under the tiebreaker rule: If we measure the seriousness of potential harm by the worst possible outcome undiscounted by the probability of its occurrence, as is required by an ex post POV, then virtually all conduct threatens serious harms, and we are back to moral gridlock. Alternatively, we could adopt some form of continuous function that weights expected harms more heavily, the more serious they are. While this approach would produce different answers from conventional cost/benefit analysis and other forms of unweighted aggregation, it is still a form of aggregation in the sense that harms to me can be offset by benefits to you (albeit at a different ratio than under unweighted aggregation). It is also unclear what justifies giving more weight to serious harms than such harms would be given in unweighted ³⁸ See, for example, Gregory Keating. “Rawlsian Fairness and Regime Choice in the Law of Accidents,” Fordham Law Review 72, no. 5 (Apr. 2004): p. 1876.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

81

aggregation. If the argument is that serious harms are, well, a lot more serious, that self-evident truth should be reflected in the subjective disutilities that individuals assign to serious and trivial harms respectively. That is to say, a properly done cost/benefit analysis based on subjective preferences would normally assume steeply increasing marginal disutility as one moves up the scale of harms. If contractualists are evaluating expected outcomes based on subjective preferences, what is the argument for putting even more weight on serious harms than the would-be victims themselves do? If they are evaluating them based on something else, what is that something else? A paternalistic calculus of what we think people would prefer in the long run? A perfectionist account of what they ought to prefer? A desire to protect society from ex post revolt against ex ante agreements that turn out very badly? And how do we decide what weight serious harms should be given under any of these criteria?

2.9 Consider Every Case from Both Points of View Concluding that a pure ex ante POV and a pure ex post POV are both dead-ends for contractualists, James Lenman suggests the solution is to meld the two perspectives, in a kind of reflective equilibrium: If we think then of ex ante commitment and rejection as morally constrained and sensitive to the strains of commitment, and if we think of ex post agreement as reasonably sensitive to the earlier epistemic perspective in which principles have been applied in conditions of risk, it perhaps become easier to think of these approaches as complementary rather than necessarily competing.³⁹

Applying the method to Bernard Williams’s famous hypothetical of the nonnegligent lorry driver who, through no fault of his own, kills a young child who has darted into the street, Lenman argues we should accept both that the lorry driver’s conduct must be justifiable to the parents from their ex post (heartsick) perspective and also that “[e]ven ex post there may be constraints on what [the parents] can reasonably reject.” “If we take this point on board,” Lenman concludes, we can go some way along with Scanlon in privileging the perspective of the most burdened affected person ex post provided we also require that person to make due allowance for the epistemic perspective of the agent ex ante and it is here that probability information plausibly, properly, and indeed necessarily kicks in.⁴⁰ ³⁹ Lenman, “Contractualism and Risk Imposition,” p. 116. ⁴⁰ Lenman, “Contractualism and Risk Imposition,” pp. 115–16.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

82

      ?

It is not possible to have it both ways. If by “due allowance,” Lenman means that the parents can object ex post only if the driver’s actions were unreasonable and that unreasonableness is to be judged from an ex ante POV, then the parents’ ex post POV does no work. We have a purely ex ante system in which the numbers are all that count. This seems to be Lenman’s own conclusion in the lorry driver case. Considering the driver’s conduct from the driver’s own (ex ante) POV, Lenman asserts that the parents “should plausibly reasonably conclude the driver’s choice not to be morally at fault.”⁴¹ If, on the other hand, by “due allowance” Lenman means that the parents’ ex post objections will sometimes convert an action that was reasonable ex ante into one that in retrospect we judge to be unreasonable, then Lenman has recreated the paradox of moral luck that he rightly argues dooms a pure ex post POV as a theory of action. On what basis do we decide whether or not the parents’ ex post objections are reasonable, if not on the basis of the ex ante reasonableness of the driver’s conduct? If we decide the parents’ objections are reasonable, do we therefore conclude, in retrospect, that the lorry driver should have acted differently? And what follows from that? That the parents may blame the lorry driver even though he acted just as we would have had him act at the time he had to choose? That they are entitled to compensation? In the end, the only restriction Lenman himself appears to put on the ex ante POV concerns the rescission cases discussed in Section 2.1.

3. Facing the Inevitability of Tradeoffs In my view, the distributionally sensitive versions of an ex ante POV discussed in Sections 2.2 and 2.3 are the only alternatives to conventional aggregation contractualists have proposed to date that are clearly viable outside of the rarified world of trolleyology. It is possible that some of the other suggested approaches, once fleshed out, will generate a viable middle way. I am doubtful, however, for two reasons. First, even in their very sketchy form, the hybrid approaches offered to date suggest exactly the sorts of interpersonal tradeoffs that are at the heart of standard aggregation. Second, and more importantly, the problem on the table— defining the permissible limits of socially useful but risky conduct—is one that cannot be solved without such tradeoffs. The reason for this is straightforward. Scanlonian contractualism, like most Kantian-inflected versions of nonconsequentialism, is committed to two foundational principles: that each of us has a legitimate interest in being free to pursue our own projects in life (freedom of conduct); and that, in doing so, each of us has

⁴¹ Lenman, “Contractualism and Risk Imposition,” p. 116.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

83

a right not to be interfered with by others, most especially in the form of serious bodily harm (freedom from harm). But most conduct in pursuit of one’s own projects carries with it some risk (however remote) of harming others. As a result, we cannot protect one of those interests without compromising the other. Where a given act is very, very unlikely to harm others (changing channels on the TV, making the bed), we can effectively ignore the risk associated with it. But such conduct, precisely because it poses no meaningful threat to others, does not need to be regulated under any normative criteria. Harm that results from criminal conduct is an easy case in the opposite direction: under any normative criteria, we place no value on leaving the criminal free to pursue his own (criminal) projects, and hence we need not compromise the would-be victims’ right to be free from harm in order to protect the criminal’s freedom. In between those two extremes lies the vast universe of conduct that is prima facie socially productive and hence prima facie permissible, but that poses some nontrivial risk of harm to others. The inherent conflict between freedom of conduct and freedom from harm is posed in stark and unavoidable terms in such garden-variety cases of risk. The implicit compromise struck is expressed in lots of different ways (you can’t harm someone else “negligently” or “unreasonably,” the right to be free from harm must give way when the social costs of deferring to it are just too great). These different formulations may or may not strike that compromise differently from each other or from unweighted aggregation. But in the end, they all (necessarily) reduce to an interpersonal tradeoff between these two interests. However unattractive one might find aggregation in principle and however difficult to administer in practice, it at least faces that necessity squarely. Contractualists have yet to do so, or explain how it can be avoided. I return to this matter in Chapter 6, when I take up the second wave of contractualist arguments.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

5 Tortious Harms Tort law adjudicates claims between private parties, in which one side (the plaintiff ) alleges that the other side (the defendant) has harmed, or is threatening to harm, her legally protected interests. The potential universe of claims that can be pursued under a tort theory is very large. But the vast majority of tort claims actually pursued involve accidental harms, in both senses in which I have used the term in previous chapters: they are harms that are not intended in the sense of not desired, and that are not specifically foreseen to be a consequence of the act (“Gee, I never saw that one coming”). Car accidents, slip and fall cases, and malpractice suits are the bread and butter of the torts docket in state courts. In such cases, the wrong generally alleged is that the defendant acted negligently.¹ As in the philosophical literature on harms to others, the word “negligence” has many synonyms (without due care, unreasonable, etc.), none of which has a clear operational meaning. But courts have to decide actual cases, and from their decisions one can infer the sorts of conduct that are likely to be deemed negligent: driving 40 mph through a school zone, failing to put up scaffolding to protect passersby from falling debris from construction work, leaving a loaded gun in an unlocked cabinet, etc. Over the past fifty years, academic writing in tort theory has been slowly colonized by welfarists, who equate negligence with conduct that flunks a cost/ benefit analysis (CBA) because the private benefits to the actor are less than the social costs imposed on everyone else. While courts generally don’t define what they mean by negligence, the determinations themselves are generally consistent with the conclusions one would likely reach under a CBA-type analysis. But nonconsequentialist arguments have long played a significant role in tort theory and tort doctrine. While their role is greatly diminished, they have not been vanquished entirely. The arguments offered are parallel those in moral theory, but present somewhat different problems when confronted with the reality of legal institutions. This chapter examines the arguments and how they get cashed out in practice.

¹ For the balance of the discussion, “action” or “conduct” should be understood to refer to a chosen activity undertaken with a specified level of precaution against harm to others. Thus, “driving at 50 mph” is one act; “driving at 65 mph” is a different one. When I speak of the state prohibiting or permitting an act, I therefore mean prohibiting or permitting a given activity when undertaken with the specified level of precaution. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

 

85

In lieu of CBA or some other form of aggregation, nonconsequentialists endorse a two-pronged test for establishing that conduct is tortious (legally wrong), a test that they argue is already incorporated in tort law. The plaintiff must show that (i) she was actually harmed by defendant’s conduct, and that (ii) the conduct violated the defendant’s duty of care towards her.² The first requirement—harm to the interests of others—is generally taken to exclude the type of harm that motivated Mill to articulate the harm principle: private conduct that does not compromise the interests of others (other than their nosy preferences about how their fellow citizens live their lives). But within the remaining universe of conduct that is (potentially) harmful to others, the requirement of harm to others cannot help to distinguish good risk imposition from bad. In contrast, the second requirement—that the conduct in question must violate a duty of care toward others—is capable of differentiating among different forms of potentially harmful conduct. But to do so, it must specify what sorts of conduct that duty requires, permits, or prohibits. Most nonconsequentialist tort theorists have left that question for another day. Where they have suggested an answer, it is not clearly distinct from some version of CBA. At the end of the day, then, the “harm plus fault” test either provides no answer to the question of what dangers it is permissible to expose others to, or provides an answer that reduces to aggregation manqué. It has, however, answered a different question: when are we required to compensate the victims of our risky conduct for their losses? While the two questions are related, they are nonetheless distinct, and the failure to distinguish clearly between them is one of the central analytic problems with the literature. To the extent nonconsequentialists have distinguished the two questions and are addressing themselves only to the latter— correcting wrongs—their arguments are orthogonal to my main concern here, which is how we ought to decide what sorts of risky conduct are permissible in the first place. Rather than offering an exhaustive account of the nonconsequentialist literature on torts, I want to put on the table what seem to me the central, recurring problems that have led it down a dead end in the quest for a systematic alternative to aggregation to regulate risky conduct. It may be possible to come up with some other alternative that works, although I am highly doubtful for reasons I touch on briefly at the end.

² It is frequently argued that libertarians, in contrast to all other nonconsequentialists, regard harm to others as sufficient to create a prima facie case that the actor’s conduct was wrongful, without requiring any separate showing that the actor is at fault (point (ii)). See, for example, Jules Coleman, “Tort Law and Tort Theory,” in Philosophy and the Law of Torts edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), p. 201. Libertarians often characterize their own positions in such a fashion, as in Nozick’s famous metaphor of a boundary that cannot be crossed. But, in fact, libertarian arguments always depend (and must depend) on some notion of fault to go through; the fault criterion is simply implied rather than stated. I return to this issue in Chapter 8.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

86

 

1. Eight Problems with the Nonconsequentialist Analysis of Tort Law 1.1 The Dangers of an Immanent Critique The last thirty years have seen the rise of a new (or arguably renewed) scholarly approach to evaluating the common law, which might be termed the immanent critique.³ Rather than starting with an external metric of value—optimizing welfare, protecting specified rights, giving parties what they have (or would have) chosen for themselves—and assessing the goodness or badness of an existing legal regime in terms of it, proponents argue we should start with the existing legal regime and work out. That is to say, we should start by giving the best interpretation we can of existing practice and identifying the norms that are immanent in that practice, and then ask whether those norms yield principles we wish to endorse. During roughly the same period, moral philosophy has witnessed a similar development, with the rise of a kind of common-law method of moral reasoning about our duties to others. In place of the norms of legal practice, our intuitions about the “right” answers to hypothetical dilemmas function as social facts that unconsciously embody moral wisdom, a proper understanding of which will lead us, by process of generalization and reconciliation, to a set of moral principles for handling harm to others.⁴ One can get to the same place whether one works from the inside out or the outside in (or, in Jules Coleman’s terms, “bottom up” or “top down”⁵). Proponents of an immanent critique have offered a number of plausible arguments for starting from the inside, including the impossibility of evaluating moral principles divorced from their applications, and the fact that widely shared moral intuitions are entitled to some presumption of moral correctness. Thus, starting with what is, understood in the most sympathetic light, may be the most reliable and expeditious route to what ought to be. But there are obvious dangers to the immanent critique that both the legal and philosophical literature on harm to others have (to my mind) fallen prey to. First, it is easy to misunderstand what you are looking at when you view it only from the inside. In seeking to give the “best” explanation of tort law as a freestanding ³ Richard Craswell, discussing the same development, refers to the literature as “interpretive theories.” “In That Case, What is the Question? Economics and the Demands of Contract Theory,” Yale Law Journal 112, no. 4 (Jan. 2004): p. 918. Recent expositors of an immanent critique of private law include Jules Coleman, Ernest Weinrib, Dennis Patterson, Peter Benson, Stephen A. Smith, Daniel Markovits, John Goldberg, and Benjamin Zipursky. ⁴ See, for example, Frances Kamm, Intricate Ethics (Cambridge, Mass.: Harvard University Press, 2007); John Mikhail, Moral Cognition: Rawls’s Linguistic Analogy and the Cognitive Science of Moral and Legal Judgment (Cambridge: Cambridge University Press, 2011). ⁵ Coleman, “Tort Law and Tort Theory,” p. 184.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

87

institution, legal philosophers have committed themselves to the view that it can best be understood by staying within its own borders. But tort law is just one arm of the state’s regulatory apparatus for controlling harmful conduct. The decision to analyze tort law in isolation from the rest of that regulatory apparatus has led to a number of misconceptions about tort law itself, some of which I take up below.⁶ Second, it is easy to lose sight of the fact that what is is entitled at most to a presumption that it is right. It might be wrong. At some point, the presumption has to be defended in light of articulated norms. Most of the nonconsequentialist literature on tort law has not gotten to that last step. As a result, it often confuses convention for some external notion of morality, and institutional designs adopted for pragmatic reasons for ones that reflect a deeper normative commitment. (Again, I take up examples below.)

1.2 Conflating Prohibition and Compensation The state faces two decisions in regulating harmful conduct: (i) what risks should we allow individuals to impose on each other; and (ii) who should bear the costs of risky conduct? These are very different questions. In economic terms, the first is an allocative question: what conduct is optimal from a societal point of view? The second is a distributive one: how do we distribute the costs that result from permitting or prohibiting a given form of conduct? To put the difference in operational terms, the first asks, would we want to prohibit (risky) act X from ever occurring if we could? The second asks, assuming act X has occurred, with or without legal permission, and it has resulted in actual harm to one or more persons, should the victims be compensated, either by the actor or by the state? How the state resolves the distributive question will indirectly affect the allocative decision by affecting the incentives actors have to engage in act X in the first place. But these are nonetheless separable questions, both analytically and operationally. The distinction between the permissibility of conduct and the distribution of costs that arise from permitting or prohibiting it has been familiar to legal audiences at least since Calabresi and Melamed’s famous 1972 article spelling

⁶ While I am setting the compensation question to the side, it is worth noting that the immanent critique of tort law has worked mischief in that area as well. In focusing on tort law in isolation from other parts of the regulatory system, proponents of a corrective justice approach to compensation implicitly assume that the only available source of compensation for the injured party is the injurer—an assumption that simultaneously justifies the corrective justice approach that helped itself to it, and insulates that approach from obvious criticisms that arise when one considers alternative compensation schemes (e.g., private or social insurance). For a critique of the corrective justice approach for that parochial vision, see Jeremy Waldron, “Moments of Carelessness and Massive Loss,” in Philosophical Foundations of Tort Law, edited by David G. Owen (New York, Oxford University Press, 1995), pp. 387–408.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

88

 

out the four possible resolutions of the prohibition/compensation question.⁷ In the philosophical literature on harm to others, however, the two questions are frequently conflated. Within the legal academy, that confusion can be traced in large part to the fact that the literature comes at the problem through the lens of the tort system, which (in the case of torts that are not considered “strict liability”) conflates the two questions by design. That is to say, defendants who are found to have acted negligently—meaning to have violated the standard of care we require of actors in their position—are also required to compensate the victims for the injuries they cause. But this is not a natural feature of the world. It reflects a prior policy determination to link the two in the design of the negligence portion of the tort system. We can (and in other contexts routinely do) decouple judgments about whether conduct is wrong in the sense that we would have prohibited it if we could have, from judgments about whether for sundry policy reasons (corrective justice, individual welfare, cost internalization, risk spreading, etc.) the victim should be compensated, and if so, whether the compensation should come from the person who caused the harm, from a private insurance scheme, or from the state. The failure to keep these questions separate is, I believe, the first casualty of the immanent critique of law. This point needs to be underscored. In the past, debates about the propriety of analyzing tort law in isolation from the rest of the state’s risk-regulation apparatus have centered on whether tort law should be assumed to (or required to) cohere with the principles underlying the rest of the regulatory state. Why should it, nonconsequentialists have argued? Why shouldn’t it follow its own logic of corrective justice?⁸ The problems with interpreting tort law as a freestanding institution that I press here are different. I am not worried that such a practice has encouraged people to ignore the differences between tort law and the regulatory state. I am worried that it has encouraged them to ignore the similarities, and thereby misdescribe the tort system itself. Our tort system is not simply engaged in ex post corrective justice, via compensation; it is engaged in ex ante risk regulation as well, via the standards of due care it generates to determine liability for negligence in the first place.⁹ Sometimes nonconsequentialists explicitly undertake to answer both the prohibition and compensation questions, but by conflating the two, answer neither.¹⁰ In the legal context, however, most nonconsequentialists have been preoccupied ⁷ Calabresi, Guido, and A. Douglas Melamed. “Property Rules, Liability Rules, and Inalienability: One View of the Cathedral,” Harvard Law Review 85, no. 6 (Apr. 1972): p. 1085. ⁸ For a forceful argument to this effect, see Jules Coleman, The Practice of Principle: In Defense of a Pragmatist Approach to Legal Theory (Oxford: Oxford University Press, 2001), pp. 36–8. ⁹ Once we focus attention on the latter function of tort law, I do think it becomes easier to explain why we might think tort law ought to cohere with the other institutions for risk regulation, at least with respect to standards of conduct. But that is not my immediate concern here. ¹⁰ Nozick’s treatment of the problem in the first part of Anarchy, State, and Utopia is a particularly spectacular train wreck along these lines. I discuss this example in Chapter 7.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

89

with the compensation question, but—perhaps because they do not focus on the prohibition question at all—assume it is resolved pari passu with the compensation question. Others have acknowledged the distinction between the prohibition and compensation questions, but have either postponed or explicitly disavowed the obligation to supply a nonconsequentialist answer to the former. That tack avoids confusion, but at the cost of acknowledging the limited policy relevance of the nonconsequentialist literature on torts. To the extent the literature is concerned solely with the obligation to compensate others for harm one has as a factual matter done to them and has nothing to say about prohibition, it has nothing to say about the question on the table here: whether there is a coherent alternative to aggregation for regulating risky conduct.

1.3 Treating Imposition of Risk and Imposition of Harm as Different Forms of Conduct Closely related to the foregoing problem, a substantial strain in the nonconsequentialist literature has approached the problem of harmful conduct by splitting it into two categories: conduct that has already harmed another and conduct that imposes a risk of future harm. But as I argued in previous chapters, imposition of risk and imposition of harm are not distinct forms of conduct. They are identical conduct viewed from ex ante and ex post perspectives, respectively. Confusion on this point has produced many of the most serious confusions in the nonconsequentialist tort literature, including (a) the belief that actual harm to others is both a necessary condition for culpability and a sufficient condition to create at least a presumption of culpability; and (b) the belief that we can coherently judge completed harms under nonconsequentialist criteria, while ceding risky conduct to some form of aggregation. Finally, confusion on this point has spurred an entire literature trying to solve a nonexistent problem: how risky conduct that has not yet resulted in harm can be deemed wrongful. I take up each of these in turn.

1.4 Completed Harm Is Necessary for Culpability Those nonconsequentialists who start with the intuition that the proper way to think about the wrong of harm is from an expected utility framework have no difficulty recognizing that the problem of risk is the problem of harm, viewed from an ex ante rather than ex post perspective. But many start instead with the presumption that until we have actually harmed others, we have done nothing wrong, leading to the following dilemma: “On what grounds can a rights-based

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

90

 

political theory justify prohibiting risky actions” that have not yet ripened into material harm?¹¹ Within the legal academy, Ernest Weinrib has been the strongest defender of the view that actual injury is necessary for culpability.¹² But he is hardly alone in that view. Consider this from John Goldberg and Benjamin Zipursky: The wronging of one person by another is the very essence of the enterprise, and until such an event happens, there is no occasion to inquire whether an actor can or should be held to have acted wrongfully by violating a moral or legal obligation of conduct.¹³

Or this, from Judith Thomson: Suppose I play Russian roulette on you. (Gun with six chambers, one bullet.) And suppose that nothing happens: the bullet was not under the firing pin when I fired. Suppose I did this without your knowledge, so that you were caused no fear. Did I infringe a right of yours? It does not seem obvious that I did.¹⁴

The same concern has been expressed by numerous others, and has driven them to search for alternative principles to explain the wrong of risk.¹⁵ The apparently widespread conviction that we cannot condemn conduct until we know its consequences is one of the more perplexing features of the nonconsequentialist tort literature. The view that results should count for a lot is obviously tangled up with the epistemic difficulties of guessing exactly what risks a given course of conduct poses ex ante. But nonconsequentialists’ attachment to the ex post perspective clearly goes beyond anything that can be explained by such difficulties. Within the legal literature, I think it is explained in large part by the fact that nonconsequentialist legal theorists have generally approached the problem of harm to others through the lens of the private law tort system. Here, then, is the second casualty of the immanent critique: only by viewing the tort system in isolation rather than as an integral part of the state’s regulatory apparatus would ¹¹ Kenneth F. Rogerson, “Rights at Risk,” Southwest Philosophy Review 1 (1984): p. 119. ¹² See, for example, Ernest J. Weinrib, The Idea of Private Law (Oxford: Oxford University Press, 2012), p. 153; Ernest J. Weinrib, “The Disintegration of Duty,” in Exploring Tort Law, edited by M. Stuart Madden (New York: Cambridge University Press, 2005), p. 143. ¹³ John C. P. Goldberg and Benjamin C. Zipursky, “Tort Law and Moral Luck,” Cornell Law Review 92, no. 6 (Sept. 2007): p. 1138. See also John C. P. Goldberg and Benjamin C. Zipursky, “Unrealized Torts,” Virginia Law Review 88, no. 8 (Dec. 2002): p. 1689: “A claim for heightened risk—even if that risk is intelligible as a harm—does not invoke the sort of harm that defendants have a duty to take care to avoid causing.” ¹⁴ Judith Jarvis Thomson, “Imposing Risk,” in Rights, Restitution, and Risk: Essays in Moral Theory, edited by William Parent (Cambridge, Mass.: Harvard University Press, 1986), p. 173. ¹⁵ See (3) in Section 1.5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

91

one be led to conclude that no conduct can be judged impermissible until it has resulted in harm to others. The US, like most western countries, has divided the task of regulating harmful conduct between the private law tort system (which adjudicates only those cases in which there is an identifiable plaintiff with standing to sue) and the public regulatory system (which can intervene at any point, with or without identifiable victims). It is true that the tort system is generally called upon to judge conduct only after it has resulted in harm to identifiable others. But this feature does not reflect deep moral principle; it reflects administrative expediency. The only parties with standing to bring a tort suit are those who have been harmed by the conduct, or who are put at a heightened risk of harm. In theory, one can seek an (ex ante) injunction to prohibit conduct that will put one at heightened risk of serious harm. But to get an injunction, a plaintiff must show that she is likely to be seriously harmed and that compensatory damages will not provide an adequate remedy.¹⁶ In addition, she would have to know about the risk in time to get into court before the act in question has been completed. Only rarely are all these criteria met. As a result, it is rare for a plaintiff to seek, let alone get, an injunction.¹⁷ Thus, almost all tort suits come up in an ex post posture, in which the remedy the plaintiff seeks is not to prevent the harm (which by hypothesis it is too late to do) but to be compensated for it. Hence, actual harm is a necessary element of most tort suits. But that is because of the remedy sought, not the scope of the right. While the tort system operates largely ex post, much of the rest of the enormous state apparatus for regulating harmful conduct does not. Criminal codes routinely punish conduct that puts others at extreme risk of injury (drunk driving, criminally negligent violation of safety standards, etc.) without requiring that the conduct actually have resulted in injury. In the civil sphere, environmental regulations, health and safety regulations (building codes, product standards, workplace regulations), licensing requirements for high-risk jobs, securities regulation, etc., all operate ex ante. Sometimes ex ante enforcement takes the form of requiring pre-clearance to demonstrate compliance with regulatory standards (e.g., you can’t market a new

¹⁶ Dan B. Dobbs, Law of Remedies: Damages-Equity-Restitution 2, 2nd ed. (Eagan, Minn.: West Publishing Co., 1993), } 8.10, p. 535. ¹⁷ Goldberg and Zipursky, otherwise strong defenders of the notion that an already-completed harm is essential to the very notion of tortious conduct, provide a nice counterexample: cases in which the defendant’s negligent conduct results in a heightened risk of contracting a disease in the future. In such cases, they argue, the plaintiff should be entitled to an injunction allowing her to recoup the medical costs of monitoring the progress of the disease going forward—not as compensation for a “completed wrong,” but as a prophylactic measure to avoid any wrong (tortious harm) in the first place. “Unrealized Torts,” pp. 1709, 1711. Their example presents one of the few instances in which ex ante prevention is a plausible remedy for a plaintiff put at risk of harm by defendant’s conduct— plausible only because of the odd happenstance that there is a long lag time between when the party put at risk of injury is first identified (and hence has standing and motive to sue) and when harm to her will first materialize, if it does.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

92

 

drug without FDA approval). Sometimes it takes the form of penalties for violating those standards, whether or not actual harm ensues.¹⁸ Outside of the US, these and other regulatory regimes swamp the tort system in social and economic importance. Even in the US, which is alone among industrialized democratic countries in its attachment to the “adversarial legalism” of the tort system, regulatory regimes are an increasingly important part of harm regulation, and have subsumed many functions that used to be performed by the tort system.¹⁹ They not only operate ex ante; they are designed to operate ex ante, in order to prevent harm from occurring in the first place. Whatever its source, the belief that actual harm is necessary for conduct to be judged culpable has a number of curious implications. First, it requires us to suspend moral judgment of an act until it is too late to avoid it, thereby relegating rights-holders to (at best) ex post compensation for the involuntary loss of their rights. That result is surely ironic, given that the conception of rights that most nonconsequentialists embrace implies that cash compensation for injury resulting from a rights violation is always a second-best solution. The first-best is to prevent the violation to begin with, through deterrence, injunctive relief, etc. How can it be that the just state is prohibited from taking the only sort of action that would fully prevent injustice? Second, as I suggested in previous chapters, a moral theory that cannot judge whether actions are wrongful until they are completed is not a theory of action. If we conclude, ex ante, that it is permissible for someone to drive down the street at 25 mph, because on balance the activity is socially useful even if it poses an irreducible risk of harm to others, but conclude ex post, after someone is hurt, that it was wrong after all for reasons not fully captured by a utilitarian calculus, we have created a paradox. That paradox is economically encapsulated in the common practice of describing the wrongfulness of conduct by reference to its actual consequences. Thus, Judith Thomson describes the job of moral theory to be to answer the question “what harms we may or may not cause in what circumstance.”²⁰ Tony Honoré describes the “conduct” for which one may incur tort liability to include “causing untoward consequences [by] conduct of a potentially

¹⁸ A nonconsequentialist might be tempted to respond that the latter case does involve completed harm—the harm of violating mandated safety standards. But that violation is not a moral wrong in itself. There is nothing intrinsically immoral about driving 70 mph, or putting a handrail at the height of 3 rather than 2 feet off the ground. Such acts are prohibited only because they are thought to create a risk of harm to others (in legal terminology, they are malum prohibitum rather than malum in se). The ultimate harm to others, not the violation of safety standards promulgated to prevent it, is the only harm we care about here. ¹⁹ Robert A. Kagan, Adversarial Legalism: The American Way of Law (Cambridge, Mass.: Harvard University Press, 2003). One recent study puts the current cost of the US tort system at 2.3 percent of gross domestic product. “U.S. Tort Costs: 2004 Update: Trends and Findings on the Cost of the U.S. Tort System,” p. 13, http://www.legalreforminthenews.com/Reports/Tort_Costs_2004_Update.pdf. ²⁰ Thomson, “Imposing Risk,” p. 185 (emphasis added).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

93

dangerous sort.”²¹ Goldberg and Zipursky characterize tort law as prohibiting “[t]he doing of realized wrongs,” not “the doing of unrealized . . . . wrongs,” and as imposing “a duty . . . to not injure, rather than a duty to . . . not engage in injurious conduct.”²² The paradox created by insisting that we suspend final judgment of conduct until its consequences are known is, of course, the paradox of moral luck I discussed in Chapter 2: we decide what course of action is morally required of us ex ante, on the basis of expected outcomes, but conclude ex post that we made the wrong choice after all, when things turn out badly through no fault of our own. These time-inconsistent judgments are logically problematic, however, only if they are purporting to answer the same question. Often, they aren’t. Consider, for example, a variant on Bernard Williams’ famous hypothetical of Gauguin, who must choose between his family and his art. When Gauguin states, ex ante, that all things considered it is right for him to choose his art over his family, he is proposing a theory of rational (moral) action. When he states, ex post, after he has failed as an artist, that he regrets his earlier choice, he is expressing an emotional response to a choice that turned out badly. Our propensity to feel regret or remorse for the bad consequences of (all things considered) prudent choices may be unfortunate, in the sense that it guarantees much unhappiness in life to no good purpose. But it is not logically inconsistent. It simply reflects the fact that our rational and emotional lives are operating on different planes, and responding to different stimuli. The same cautionary point holds here. If what nonconsequentialists really mean by a duty not to injure is that we have a duty to compensate for any injury that results from our conduct—if, that is, they are really proposing a standard for compensation and not conduct—then the duty they propose does not create a paradox. It simply amounts to a version of strict liability (you broke it, you fix it) for conduct that is judged culpable on other, unstated grounds. It may be harder than proponents think to come up with an unmoralized notion of “broke” or to defend the moral basis of a strict liability compensation scheme, but the proposed regime is not paradoxical. If, on the other hand, the duty not to injure is intended to prescribe a standard of conduct with respect to the action in question—you have a duty to act in a fashion that will not cause X consequences, by which we mean that we will prohibit any actions that will turn out to have caused X—it is a paradox, and has to be resolved in practice by choosing one temporal perspective or the other. Given that we go through life in only one temporal direction, if that duty is meant to provide a standard for conduct, it must be based on information available to the ²¹ Tony Honoré, “The Morality of Tort Law—Questions and Answers,” in Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), p. 75. ²² Goldberg and Zipursky, “Unrealized Torts,” pp. 1645, 1652.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

94

 

actor ex ante, at the moment of choice. That information, by necessity, must be about the expected, not actual, consequences of the act under consideration. Of course, even hardcore adherents to the view that harm is necessary for culpability agree that something is wrong with pointing a loaded gun at someone’s head and firing what turns out to be (through no virtue of the shooter’s) a blank, or brandishing a loaded gun carelessly in a crowded downtown area whether or not it accidentally goes off. The question is, what exactly is wrong with it, and does it sound in nonconsequentialist principles? As I discuss in Sections 1.5 and 1.6, nonconsequentialists have split on this question.

1.5 The Wrong of Risk Can Be Explained by Nonconsequentialist Principles A number of possibilities have been floated for explaining the wrong of risk in deontological terms. 1.

Harm includes expected harm.

The most straightforward one has been to redefine harm to include having one’s prospects statistically worsened. Thus, David McCarthy supplants the traditional liberal harm theory with what he calls “risk liability theory,” which holds that X is liable to Y “if X performs an action that she knows or ought to know will impose a risk of harm on Y,” provided there is no excuse for it (“excuses aside”).²³ The shift to an ex ante (expected harm) perspective provides an answer to the question posed by Judith Thomson: if harm is necessary for culpability, how can the mere act of imposing a risk of future harm be wrong under the harm principle? But it does so at the cost of highlighting a second and more serious problem. Judged ex ante, almost all conduct poses a risk of harm to others. If that is sufficient to make conduct culpable, then we are back to the problem of moral gridlock discussed in previous chapters: almost all action is rendered culpable. If it is not sufficient, then something else is doing the work in differentiating permissible from impermissible conduct. I take up this problem in Section 1.7. 2.

Risk-creation is a completed harm.

Others have argued that imposing risk on others is a completed harm. Richard Epstein gestures in that direction, in arguing that in creating dangerous situations, we create a “store of energy” that is released into a harm-causing force—a view ²³ David McCarthy, “Liability and Risk,” Philosophy and Public Affairs 25, no. 3 (Summer 1996): p. 251.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

95

that invites us to view risk creation as just the first step in a continuous, multi-part act of harming others, like shooting an arrow from a bow aimed right at someone’s heart.²⁴ Others have gone further, and argued that imposition of risk is not the first step towards a completed harm but the last. In the case of risks that have run their course and resulted in no injury, the completed harm generated by risk imposition could be psychological trauma from a near-miss. In the case of ongoing risks (e.g., exposure to toxic substances that impose a risk of later disease), it could be long-term anxiety about whether the harm will materialize, or out-ofpocket monitoring costs to detect it as soon as it does materialize.²⁵ This solution doesn’t seem to have much to recommend it, beyond expediency. The mere imposition of risk may well constitute a completed harm, whether or not the principal threatened harm comes to pass. Knowing that a loaded gun is pointed at your head is surely an injury in itself, with lasting psychological consequences, even if the gun never goes off. But our widespread—indeed, universal—intuition that there is something wrong with imposing certain kinds of risk on others does not seem to depend on the presence of such completed harm or to track its severity. The victim of an attempted murder need not know the attempt has been made for us to conclude it was wrongful. We need to shed those intuitions, or come up with a better explanation for them. 3.

Risk-creation violates a different right from the right to be free from harm.

Others concede that risk imposition falls outside the reach of the anti-harm principle, but argue that it violates other interests of the would-be victims that sound broadly in deontological principles. A number of people have identified that supplemental right as a right not to be subjected to conduct that expresses contempt for (or otherwise impermissibly devalues) its potential victims. Thus, Jean Hampton argues that the reason that attempted murder is wrong, even if the attempt fails, is because of what the attempt “conveys about the intended victim’s worth.”²⁶ Rahul Kumar describes the wrong as “[t]he denial of the value of [another’s] humanity” by “culpably failing to comply with [the others’] legitimate

²⁴ Richard A. Epstein, “A Theory of Strict Liability,” Journal of Legal Studies 2, no. 1 (Jan. 1973): pp. 151–204, 179. The metaphor leaves unclear whether the force released must actually cause harm for the act to be tortious. If the answer is yes, Epstein hasn’t succeeded in making risk-creation itself tortious. ²⁵ See Goldberg and Zipursky, “Unrealized Torts”; Christopher H. Schroeder, “Corrective Justice and Liability for Increasing Risks,” UCLA Law Review 37, no. 3 (Feb. 1990): pp. 439–78; Kenneth W. Simons, “Corrective Justice and Liability for Risk-Creation: A Comment,” UCLA Law Review 38, no. 1 (Oct. 1990): pp. 113–42; Schroeder, “Corrective Justice, Liability for Risks, and Tort Law,” pp. 143–62; Claire Finkelstein, “Is Risk a Harm?,” University of Pennsylvania Law Review 151, no. 3 (Jan. 2003): pp. 963–1002. ²⁶ Jean Hampton, “Correcting Harms versus Righting Wrongs: The Goal of Retribution,” UCLA Law Review 39, no. 6 (Aug. 1992): pp. 1681.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

96

 

expectations.”²⁷ The same principle would presumably apply to Russian roulette, criminal negligence, etc. Stephen Perry identifies that supplemental right as a “second order interest recursively derived from [one’s] first order interest in not being physically injured.” So, “I can have a right that you not try to physically injure me, even though a violation of that right which did not cause me physical injury would not itself be a harm.”²⁸ Others have located it in an independent, countervailing “right to self-defense,” which permits one to take preventive action to forestall injury even when no wrong has yet been committed.²⁹ All of these efforts to handle risk under supplementary deontological principles are unsatisfying. Hampton’s supplementary right to be free of expressive harms, like the argument that the imposition of risk is an ex post harm, is wildly underinclusive. The contempt for others expressed by certain acts may give an additional reason to condemn the most egregious forms of risky behavior. But it doesn’t explain why we might want to prohibit or otherwise regulate gardenvariety behavior (e.g., driving 65 mph on ice-slicked highways) that imposes an excessive risk of loss on others but hardly rises to the level of malum in se conduct. Moreover, even as to the most egregious forms of behavior, the right to be free from expressive harms, like the recharacterization of risk as an ex post harm, seems to miss the point of what is really wrong with such behavior. Surely most people would think that the central wrong entailed in someone’s pointing a loaded gun at your head and attempting to fire it is not the contempt the would-be killer thereby expresses for your worth as a person or the transitory anxiety he causes you, but the fact that he actually could have killed you, and failed to do so through no virtue of his own. It is hard to see how nonconsequentialists can avoid the conclusion that what is wrong with risky conduct, finally, is the risk of ex post harm it poses. At least it is hard to see how, if (as I believe) most would conclude that what is wrong with actually killing someone, finally, is not the contempt it expresses for the victim or the transitory anxiety it imposes on the about-to-be-killed victim but the killing itself. Indeed, it is hard to see how ostensibly freestanding principles

²⁷ Rahul Kumar, “Who Can Be Wronged?,” Philosophy and Public Affairs 31, no. 2 (2003): p. 109. ²⁸ Stephen R. Perry, “Harm, History, and Counterfactuals,” San Diego Law Review 40 (2003): p. 1307 (emphasis added). ²⁹ See, for example, Dennis McKerlie, “Rights and Risk,” Canadian Journal of Philosophy 16, no. 2 (Jun. 1986): p. 241; Samuel C. Wheeler III, “Self-Defense: Rights and Coerced Risk-Acceptance,” Public Affairs Quarterly 11, no. 4 (Oct. 1997), arguing that the right to carry (and use) handguns for selfdefense purposes is consistent with the classic libertarian account of rights. Judith Jarvis Thomson, Rights, Restitution, and Risk: Essays in Moral Theory, edited by William Parent (Cambridge, Mass.: Harvard University Press, 1986), pp. 160–1, enlists the right of self-defense to fashion an intermediate position that seems, if anything, harder to defend: the individual right of self-defense permits the state to levy a penalty on top of compensatory damages should harm result, as a means of “indirect prevention,” but apparently would not permit it to levy penalties on risky conduct for the same purpose, in the absence of any as-yet materialized harm. Why should we encourage the state to do indirectly (and imperfectly) what we will not permit it to do directly?

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

97

like “a right to self-defense” or a “right to be free from expressive harms” can be other than parasitic on the judgment that the conduct in question poses, ex ante, too great a risk of ex post harm to be tolerated. Why else is the expression of contempt in this particular form—brandishing a loaded gun with intent to shoot— prohibited, when its expression in most other forms is not? Why else do we have a supervening right of self-defense that allows us to prevent certain forms of conduct from occurring, even in the absence of any demonstrated harm? If that analysis is correct—and I have a hard time seeing how it could not be— then such “supplementary” principles do not supplement the anti-harm principle as conventionally interpreted (i.e., to require ex post harm). They quietly eviscerate it, by disguising as an unrelated right what is in fact an ex ante judgment on the permissibility of imposing certain risks. Perry’s strategy—characterizing the right to be free from risk of harm as a second-order right derived recursively from the first-order right to be free from harm—forthrightly acknowledges that parasitic relationship, raising a different puzzle: what, operationally, turns on bifurcating these interests into primary and secondary ones? Finally, it is hard to escape the sense that all of these efforts to handle risk under the anti-harm principle notwithstanding the absence of (ex post) harm, like the decision to kick it out and resolve it under other principles, are ad hoc. What is the meta-theory that tells us when we trigger a right to self-defense or a right to be treated with respect, and how we are to trade off those supplementary rights of the would-be victim against the starting assumption that the would-be injurer has a right to do what she wishes with herself and her property unless and until she harms another? In most cases, that question is unanswered, again leaving us in doubt what the real criteria for judging the permissibility of conduct are, and whether in the end they amount to something clearly distinct from some form of CBA.

1.6 Risk Should Be Ceded to Welfarism, Keeping (Completed) Harm in the Deontological Fold Many nonconsequentialists, however, have reluctantly concluded that the problem of risk cannot be resolved by nonconsequentialist principles. Consider again this lament from Nozick: It is difficult to imagine a principled way in which the natural-rights tradition can draw the line to fix which probabilities imposed unacceptably great risks upon others. This means that it is difficult to see how, in these cases, the natural-rights tradition draws the boundaries it focuses upon.³⁰

³⁰ Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 75.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

98

 

Or this from Judith Thomson: if risky conduct is morally problematic notwithstanding the absence of completed harm, then risk imposition does generate an independent problem for moral theory. For there is a further question which then arises, beyond the question what harms we may or may not cause in what circumstances, namely, the question what risks of what harms we may or may not impose in what circumstances.³¹

But risk does not pose an “independent problem for moral theory.” It is the problem of harm to others, judged from an ex ante rather than ex post perspective. Once we recognize that we are talking about the same conduct judged from different temporal perspectives, the problem with judging the imposition of risk under different criteria from the imposition of harm becomes obvious. It requires the state to judge the permissibility of the same conduct twice, and under different standards. Take the hypothetical case of a new vaccine with the high probability of preventing AIDS, but carrying some low-level risk of death from adverse reactions to the vaccine. Suppose the FDA, the arm of the administrative/regulatory state charged with regulating drug safety, reviews the vaccine under some version of CBA and approves it for distribution. The vaccine is widely distributed, resulting in one death and an estimated thousands of lives saved by the end of the first year. The surviving family of the one victim files a tort claim, arguing that the drug company violated its duty of care in selling a defective product. The court must determine as a threshold matter whether the drug company did in fact violate a duty of care in distributing the vaccine. If—as it routinely does—the court adopts the same standard of due care as the regulatory agency used, either in determining the negligence question de novo or in concluding that regulatory safety standards, where they exist, define the level of care required, then the two stages of review present no conflict.³² But then the court is simply importing CBA into the common law, not adopting a different standard. But suppose the court decides to impose a different and tougher standard—say, a “duty not to injure”— which duty, the court concludes, the drug company violated as of the moment that someone died from the vaccine. Now what? Once again, if what the court really means by a “duty not to injure” is not that the company should not have marketed the drug but that it has a duty to ³¹ Thomson, “Imposing Risk,” p. 185. ³² Consider, for example, a tort suit based on a claim of negligent driving. Negligence will in fact be determined by looking at whether the driver has complied with the rules (e.g. speed limits, right of way) and standards of safe driving (e.g., exercising caution in a crowded intersection) devised by the administrative state. This is generally true of negligence standards: they mirror regulatory standards, to the extent such standards speak to the relevant conduct.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

99

compensate victims for any injuries that result, there is no contradiction. The two stages of review are simply answering different questions: The FDA decides the question of prohibition under a CBA, and leaves it to the courts to decide the question of compensation under whatever standard they wish. But if the “duty not to injure” in fact means what it purports to mean—that the company should not have distributed the vaccine in the first place, given that its distribution (as it turned out) resulted in injury—the company (along with the rest of us) has a real problem. Recognizing that continued distribution is likely to result in further deaths, should the company continue to distribute the vaccine or should it not? Do we want it to or do we not? Assuming someone has the standing to request the court to enjoin future distribution of the vaccine, should the court grant the injunction?

1.7 Harm Is Sufficient to Create at Least a Presumption of Wrongful Conduct Few scholars have argued that harm to others is sufficient to condemn the conduct that caused it.³³ Many, however, regard the fact that harm resulted from a given act as sufficient to create a strong presumption that the act was wrongful, acknowledging that in most cases an additional factor—generally, some version of fault—may be required to establish culpability conclusively. As noted above, the presence of harm is a necessary precondition to establish a right to compensation under tort law, since without harm of some sort, there is no loss to be compensated. But the presence of harm cannot help determine whether the conduct that caused it is impermissible (by which I mean that we would have prohibited it ex ante if we could have), for the reasons discussed at length in previous chapters. From an ex ante perspective—the only perspective from which we can judge the permissibility of conduct without creating the paradox of moral luck—virtually all conduct poses some risk of harm to others, but (by definition) none of those risks has yet eventuated in actual harm. If harm to others includes imposing a risk of harm on others, then virtually all conduct is harmful ex ante. If it includes only completed harms, then no conduct is harmful ex ante. Either way, the requirement of harm cannot help differentiate conduct we wish to prohibit from conduct we wish to permit. Something else must be doing all the work.

³³ For an argument verging on that position, see Lisa Heinzerling, “Knowing Killing and Environmental Law,” N.Y.U. Environmental Law Journal 14, no. 3 (2006): pp. 521–34. Again, if we read Heinzerling’s condemnation as a demand for compensation not prohibition (and set aside the Coasian problem of determining which of the joint causes of a given social cost “caused” it and which was the victim of it), it is a perfectly coherent (if not practical or morally attractive) requirement. It just amounts to a strict liability standard. But if it is meant as a theory of action—that is, meant to prohibit any conduct that might lead to serious harm—it appears to condemn virtually all action.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

100

 

Once again, I believe that rights theorists have reached the contrary (erroneous) conclusion because they have approached the problem through the lens of the tort system—the third casualty of the immanent critique. The reason for this is simple. Whatever risks an act may pose ex ante, few of those risks ripen into ex post harm. As a consequence, a rule that treats harm to others as prima facie culpable appears not to force us to condemn all action. Goldberg and Zipursky seem to offer exactly this justification for the ex post harm requirement in the following passage, suggesting that nonconsequentialists’ tolerance of moral luck may at root not be a tolerance of luck at all, but instead a defense of liberty. It is worth quoting at length: [A] system that recognized as actionable not only duties not to cause physical injury, but also duties not to cause various increments of risk of physical injury would potentially create undue de facto burdens on citizens’ freedom of action. . . . It may be that, from certain moral perspectives, identical acts of careless driving, one of which ripens into a tort, the other of which does not, are equally deserving of sanction. Hence from those perspectives, there is an element of arbitrariness as to when tort liability attaches. But viewed from the standpoint of liberal political theory, the insistence on realization is not arbitrary. Rather, it harnesses chance to create a kind of buffer zone for free action: Unless and until injurious conduct actually causes an ultimate harm, it is not subject to sanction through a privately commenced lawsuit. In this regard, the employment of duties of non-injury in tort is akin to the rule against prior restraint of speech: It serves as a prophylactic by permitting a certain amount of undesirable conduct in order to ensure that liberty is preserved. Experientially, this appears intuitive: Most people take advantage of the buffer zone created in part by the requirement of ultimate harm at one time or another, for example, by occasionally driving unreasonably.³⁴

The authors back away from the more radical implications of this justification in the paragraph that follows, suggesting that perhaps all they are policing against is potential abuse of process by private plaintiffs, along with the administrative burden from the multiplicity of lawsuits that would ensue if all risky behavior were subject to private causes of action by potentially harmed parties. But taking them at their word, the quoted passage (inadvertently) puts its finger on the moral luck paradox buried in the ex post view. Suspending judgment on the permissibility of conduct until we know its consequences does not, contra the authors, “permit[] a certain amount of undesirable conduct.” It permits it all, relegating the victims to ex post compensation under a (roughly) strict liability standard.³⁵ ³⁴ Goldberg and Zipursky, “Unrealized Torts,” p. 1654. ³⁵ Alternatively, if we read this standard to mean that anything that could turn out, ex post, to harm others is impermissible, it prohibits virtually all conduct.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

101

As the authors note, this is precisely the effect of our presumption against prior restraints on speech in the context of the First Amendment. We generally will not suppress speech ex ante; potential victims must wait until the speech in question has been shown to have wronged them (because it is slanderous, libelous, etc.). But by that point, the injured party’s only recourse typically is monetary compensation. We live with the resulting higher level of harm in the context of the First Amendment, because we attach a uniquely high value to freedom of speech. We could follow the same procedure in the context of risky conduct, permitting it all and relegating victims to ex post compensation. Thus, for example, we could refuse to judge whether a scaffolding has been put up safely until we see whether it collapses and kills someone.³⁶ I seriously doubt this is the result that the authors want. But it is the result they will get—a fact that is easy to miss looking at the problem from the ex post perspective of the tort system. It is easy to miss because courts, judging conduct ex post, are not called upon to decide whether they would have prohibited the conduct if they could have. That horse is out of the barn. The only thing they are called upon to decide is whether compensation is required. And since they don’t have to answer the prohibition question, it is easy to think that in answering the compensation question they have somehow answered the prohibition question as well.

1.8 “Fault” plus Harm Makes Conduct Wrongful As noted above, the fact that a given act resulted in harm is generally taken to create only a presumption that the act was culpable. Nonconsequentialists acknowledge that in most cases, something more is required to establish culpability conclusively. That something more, most often, reduces to a requirement that the conduct in question be “wrongful.”³⁷ Sometimes the requirement of wrongfulness is unconsciously smuggled into the definition of harm-in-fact, like the “benevolent definition of a sou as a small coin to be given to the poor.”³⁸ Thus, Thomson, after starting with an apparently very broad, unmoralized definition of harm as “unwanted outcomes,” silently accommodates her intuition that many unwanted outcomes should not be ³⁶ As I suggested at the start, the actor’s (speaker’s) knowledge that he will have to compensate for any (ex post) harm will indirectly chill risky conduct (speech) by increasing its expected cost. But that consequence is unintended, presumably undesired (if we take the authors at their word about the value of freedom), and unlikely to deter the same conduct we would choose to prohibit ex ante on the basis of expected harm. ³⁷ On the centrality of fault to a moral conception of tort, see David G. Owen, “The Fault Pit,” Georgia Law Review 26, no. 3 (Spring 1992): p. 710: “[T]he dominance of fault in the law of torts is a moral inevitability. Fault lies at the very heart of tort law and provides it with a meaningful, moral definition.” ³⁸ Felix Cohen, “The Ethical Basis of Legal Criticism,” Yale Law Journal 41, no. 2 (Dec. 1931): p. 203.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

102

 

actionable by limiting unwanted outcomes to those outcomes that “infringe a right of [the victim’s].”³⁹ Goldberg and Zipursky, likewise starting with a very broad duty of care under negligence law—“to take reasonable care not to cause [a physical] injury through affirmative conduct”—restate it as a duty not to cause “physical injuries [through one’s] misfeasance.”⁴⁰ More often, the requirement of wrongfulness is explicitly layered on top of the requirement of harm-in-fact. Thus, to be culpable, conduct must not only cause harm; the harm must also be “wrongful.” The requirement of wrongfulness is phrased in a variety of ways. The conduct must violate [a] right”⁴¹ or a “legitimate” or “protected interest,”⁴² fail to give people what they are “due,”⁴³ violate a “duty of care,”⁴⁴ constitute a “wrong”⁴⁵ or “improper treatment of the victim,”⁴⁶ be “not excused,”⁴⁷ be unreasonable,⁴⁸ be a harm with respect to which the actor is “at fault,”⁴⁹ or be inflicted “negligently,” “unjustifiably,” or “impermissibly.”⁵⁰ For legal purposes, some of these verbal differences may imply a difference in the types ³⁹ Thomson, “Imposing Risk,” p. 173. Thomson’s examples here include actions that inflict physical pain but no lasting injury (pinching someone’s nose) and actions that amount to aesthetic nuisances (a neighbor puts unsightly plastic geese on her lawn). These injuries may be relatively trivial, transitory, or easily avoided. For these and other reasons, we may decide, as a normative matter, that the actions should be privileged. But if the question is whether, as a matter of fact, they inflict an unwanted outcome on the victim, the answer is yes. ⁴⁰ Goldberg and Zipursky, “Unrealized Torts,” p. 1697 (emphasis added). ⁴¹ Jules Coleman, Risks and Wrongs (New York: Oxford University Press, 2002), p. 335. ⁴² Peter Birks, “The Concept of a Civil Wrong,” in Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), p. 40; Thomson, “Imposing Risk,” pp. 174–5; Kumar, “Who Can Be Wronged?,” p. 107. ⁴³ Joseph Raz, “Autonomy, Toleration, and the Harm Principle,” in Issues in Contemporary Legal Philosophy: The Influence of H.L.A. Hart, edited by Ruth Gavison (New York: Oxford University Press, 1987), p. 329. ⁴⁴ Goldberg and Zipursky, “Unrealized Torts,” p. 1659. ⁴⁵ Joel Feinberg, Harm to Others (Oxford: Oxford University Press, 1987), p. 36. ⁴⁶ Owen, “The Fault Pit,” p. 717. ⁴⁷ McCarthy, “Liability and Risk,” p. 250: X is liable “if X performs an action that she knows or ought to know will impose . . . harm on Y,” provided there is no excuse for it (“excuses aside”); Julie Tannenbaum, “Emotional Expressions of Moral Value,” Philosophical Studies 132, no. 1 (Jan. 2007): p. 47: you must never kill another unless “the other’s right to life is overridden, forfeited, or waived.” ⁴⁸ Gregory Keating, “A Social Contract Conception of the Tort Law of Accidents,” in Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), pp. 54–60. ⁴⁹ Feinberg, Harm to Others, p. 36; Owen, “The Fault Pit”; Stephen Perry, “On the Relationship Between Corrective v. Distributive Justice,” in Oxford Essays in Jurisprudence, edited by Jeremy Horder, 4th series (Oxford: Oxford University Press, 2000), pp. 237–8. ⁵⁰ Ernest J. Weinrib, “Causation and Wrongdoing,” Chicago-Kent Law Review 63, no. 3 (1987): pp. 448–9; Judith Jarvis Thomson, “Some Questions about Government Regulation of Behavior,” in Rights, Restitution, and Risk, edited by William Parent (Cambridge, Mass.: Harvard University Press, 1986), p. 167; Coleman, Risks and Wrongs, p. 332; Stephen R. Perry, “Responsibility for Outcomes, Risk, and the Law of Tort,” in Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), pp. 112–15; Martin Stone, “The Significance of Doing and Suffering,” in Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), pp. 161–2; Arthur Ripstein and Benjamin C. Zipursky, “Corrective Justice in an Age of Mass Torts,” in Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2 Anarchy, State, and Utopia 1), pp. 219–20; Richard W. Wright, “Right, Justice and Tort Law,” in Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), pp. 169–70; David G. Owen, “Philosophical

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

     

103

and levels of proof required. But for current purposes, the differences are irrelevant. Each formulation, like Thomson’s silently moralized definition of harm itself, boils down to an assertion that culpable harms are all those harms-in-fact that we have a right to be free from. As a result, we are left with a scheme that essentially says: ex post harm to others’ interests gives rise to liability, except when it doesn’t.⁵¹ Many rights theorists have acknowledged that fault is just a placeholder for a standard of permissible conduct yet to be articulated. But until it is articulated, it is impossible to say whether nonconsequentialist principles can supply a coherent alternative to aggregation in distinguishing permissible from impermissible risk imposition. Where rights theorists have tried to flesh out a standard for permissible conduct, the criteria they suggest often seem vanishingly close to CBA. Consider, for example, this formulation of our duty of care, which the author acknowledges reduces to “a calculus of costs and benefits . . . to the actor and other persons”: By definition, an accident diminishes the quality (and perhaps quantity) of a victim’s life and other goods, which produces suffering for the victim. We may assume that human suffering is undesirable and so should be avoided, ex ante, or remedied, ex post. [Fn: Unless the cost of such avoidance or remedy is, by some fair measure, excessive.] Moreover, because accidents consume human and other social resources, society suffers harm to its aggregate stock of goods by accidents. [Fn: Unless the conduct that caused the accident generated more goods than it consumed.]⁵²

Or this: A rough account of why [a] drunk driver has wronged [a] pedestrian [he has put at risk] ought to appeal to the failure to comply with the pedestrian’s legitimate Foundations of Fault,” in Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), pp. 219–20, 224–5. ⁵¹ Scholarship in criminal law has (not surprisingly) fallen into the same tautology in defining the class of harmful conduct that is criminally wrongful. The Model Penal Code, for example, defines it as conduct that “unjustifiably and inexcusably inflicts or threatens substantial harm.” Model Penal Code } 1.02. See also Feinberg, Harm to Others, pp. 31–6. While beyond the scope of this chapter, in my view many other principles offered as decision rules in the nonconsequentialist literature (respect peoples’ autonomy, sovereignty, self-ownership, treat others as ends and not means) similarly are labels attached to decisions reached on other, unstated grounds. ⁵² Owen, “The Fault Pit,” pp. 716, 722–3. Owen, who has sought to defend a moralistic view of fault, tries to climb back from the limb he has put himself out on here by arguing that his motives for embracing the utilitarian solution are moralistic, in particular, a Dworkinian “equality of concern and respect” for the interests of other persons. But he goes on to define acts that are “morally justifiable in terms of equality” as those that are “likely to achieve a good for the actor and others that is greater than any harm foreseeably risked to the victim and others”—that is to say, those that would be recommended by CBA. Owen, “The Fault Pit,” p. 721.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

104

 

expectations of the driver that she operate her vehicle in a manner conducive to keeping the risk at which others are put as a result of her activity within certain acceptable limits.⁵³

In a similar vein, Joel Feinberg argues that the culpability of an actor should turn on three factors: the probability of expected harm, the magnitude of expected harm, and the independent value of the risky act to the actor himself and to society at large.⁵⁴ Charles Fried argues that “it is wrong to expose the person or property of another to undue risk of harm, but what risk is undue is a function of the good to be attained and the likelihood and magnitude of harm.”⁵⁵ While conceding its similarity to CBA, Fried gamely tries to keep his standard in the deontological fold by enshrining its aggregative solution as something each side has a categorical right to demand: Once “the weighing has been done, then it is also absolutely wrong to go against the conclusion of that process.”⁵⁶ In other words, in return for being willing to rename their welfarist policy recommendation a “right,” welfarists get to specify the content of that right. This is a version of rights theory one imagines any welfarist could happily live with. Either way, rights theorists have not offered an alternative to aggregation for regulating risks—in the first case, because they have ceded the problem to others (including welfarists), and in the second, because they appear to have implicitly adopted the welfarist solution as their own.

2. Can We Resolve the Question of What Risky Conduct to Prohibit and What to Allow Under Other Nonconsequentialist Principles? Up until this point, I have limited consideration to whether the criteria of “harm” plus “fault” can yield a coherent decision rule distinct from CBA for determining what risky conduct to prohibit. I don’t think the limitation rules out a lot, as most of the nonconsequentialist literature on torts revolves around these two criteria. But a full treatment of the problem would have to consider other criteria invoked to distinguish permissible from impermissible conduct. I’ve touched on many of them in previous chapters: whether harm will result from an act or an omission; ⁵³ Kumar, “Who Can Be Wronged?,” p. 107. See also the discussion in Chapter 4 of Scanlon’s requirement that we adopt “reasonable precautions,” with reasonableness to be determined by weighing the risk of serious harms to some individuals against the benefits that will be realized by others if the activity goes forward. My point here is not to argue that all these formulations of a duty of care reduce to CBA. It is to argue that they are sufficiently close that it falls to nonconsequentialists to explain how they differ. ⁵⁴ Feinberg, Harm to Others, pp. 190–3. ⁵⁵ Charles Fried, Right and Wrong (Cambridge, Mass.: Harvard University Press, 1978), p. 12. ⁵⁶ Fried, Right and Wrong, p. 12; see also pp. 156–7.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

        

105

whether the harm is intended, in the strong sense; whether the harm is foreseeable even if not intended; whether the actor has specific or only statistical foresight of the possible harms and possible victims; and whether harm is deemed “certain” to result, or merely possible or probable. Whatever moral traction each of these criteria may have in the trolley-type, one-off, individual choice scenarios that dominate the philosophical literature on harm to others, there are strong reasons to think they are irrelevant to wide-scale risk regulation.⁵⁷ The state is a collective entity, not a real person. Unlike private parties, it has an affirmative duty to aid its citizens. In dispatching that obligation in the regulatory context, it typically articulates rules to govern private conduct rather than acting directly itself. All of these factors may moot many, if not all, of the agent-centered prerogatives that are taken to support the act/omission distinction. In addition, the acts I am concerned with here—socially useful conduct that carries some risk of harm to others—are for the most part identically situated ex ante with respect to intentionality, foreseeability of harm, certainty of harm, and identifiability of the type of harm and the likely victim(s). Harm is never intentional, in the strong sense of desired; it is always an unwanted byproduct of conduct pursued for other reasons. Typically, harm is foreseeable as a probabilistic matter; in some cases the frequency and extent of harm is foreseeable to something approaching a statistical certainty. But exactly when and to whom it will occur is not foreseeable, and since no individual instance of harm is certain to occur, it is theoretically possible (although in many cases exceedingly improbable) that no harm will result. Because most socially useful but risky conduct falls out the same way with respect to each of these factors, such factors are incapable of differentiating good risks from bad.

⁵⁷ Michael Otsuka makes the same observation. See Michael Otsuka, “Risking Life and Limb,” in Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

6 Can Contractualism Be Saved? In his 1998 book, What We Owe to Each Other, T. M. Scanlon offered an ambitious contractualist theory for determining the moral obligations of individuals to each other.¹ As I discussed in Chapter 4, Scanlonian contractualism rests on two central tenets. The first is that moral obligations are based on the “individual reasons” of representative, situated, persons. The second is that, presented with competing individual reasons, the moral agent must choose among proposed alternative actions by reference to the Greater Burden Principle (GBP): greater harms trump lesser harms. The first tenet, common to most forms of nonconsequentialism, rules out aggregation of similar complaints in assessing the relative burdens imposed by proposed alternatives. The second, a distinctive feature of Rawlsian and Scanlonian contractualism, adopts a maximin decision rule: the individual with the most to lose in a given choice is allowed to veto it, provided there is an alternative that will lead to a less-bad worst outcome. As I discussed in Chapter 4, in his original version of contractualism (henceforth “ex post contractualism”), Scanlon proposed that individual complaints be weighted in accordance with the actual harm that will, in the event, befall a representative individual as a result of a given choice. Over the past decade, many contractualists (Scanlon included) have concluded that ex post contractualism is unworkable, because it cannot handle the garden-variety problem of decisionmaking under uncertainty (risk). Attention among contractualists has now shifted to developing an alternative version (henceforth “ex ante contractualism”) that can solve that problem while still generating outcomes that are meaningfully distinct from utilitarianism.² The version that has emerged preserves the second tenet of ex post contractualism (the maximin decision rule). But instead of measuring individual burdens by the actual harms representative ex post losers will suffer, ex ante contractualism ¹ T. M. Scanlon, What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998). ² Those who have embraced some version of ex ante contractualism include Johann Frick, “Contractualism and Social Risk,” Philosophy and Public Affairs 43, no. 3 (Summer 2015): pp. 175–223; Aaron James, “Contractualism’s (Not So) Slippery Slope,” Legal Theory 18, no. 3 (Sept. 2012): pp. 263–92; Rahul Kumar, “Risking and Wronging,” Philosophy and Public Affairs 43, no. 1 (Winter 2015): pp. 27–51; and T. M. Scanlon, “Reply to Zofia Stemplowska,” Journal of Moral Philosophy 10, no. 4 (2013): 508–14. Not all contractualists have shifted their attention to uncertain outcomes. See, for example, Alex Voorhoeve, “How Should We Aggregate Competing Claims?” Ethics 125, no. 1 (Oct. 2014): pp. 64–87. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

   ? 107 measures them by the expected harms to representative potential victims, calculated as of the epistemic moment of hypothetical agreement. The move from actual to expected harms eliminates one of the more bewildering implications of ex post contractualism: that a gamble that was in everyone’s interest to take ex ante is nonetheless impermissible if we know it will fall out differently for (as yet unidentified) people ex post. As Johann Frick argues, ex ante contractualism automatically produces the sensible answer in such cases of noncompetitive claims. Since everyone faces the same expected outcomes, absent idiosyncratic risk preferences or tastes everyone will share the same preferences over available choices, eliminating all disagreement. That answer will also coincide with the aggregative solution, since individuals who “naturally” face the same risks ex ante with the same preferences are in the same relationship to each other as those put behind a thick veil of ignorance. The problematic cases that remain for contractualists are ones in which individual prospects diverge at the moment of choice (“competitive claims”). In keeping with the individual reasons restriction (the first tenet), ex ante contractualists have proposed weighting each individual’s complaint by her expected outcome (the sum of harms and benefits discounted by the probability of occurrence) under a given alternative. In keeping with the GBP (the second tenet), agents must choose among available alternatives in accordance with a maximin function. Thus, if Jones faces a 10 percent chance of losing a leg under choice A and Rodriguez faces a 20 percent chance of losing a leg under choice B, Rodriguez gets to reject choice B in favor of choice A, since the 20 percent chance of losing a leg that Rodriguez faces is worse than the 10 percent chance that Jones faces. That result will not change if, in addition to Jones, 999 other people also face a 10 percent chance of losing a leg under choice A. Since we evaluate each complaint individually, Jones’s complaint against A is not strengthened by the fact that 999 other people share it. In short, in ex ante contractualism as in ex post, the numbers of affected individuals are irrelevant to each individual’s complaint, and hence irrelevant to the operation of the GBP. Proponents of ex ante contractualism have been careful not to claim too much on its behalf, suggesting that it “captur[es] an important class of pro tanto moral reasons that contribute to making actions right or wrong, but that do not by themselves determine an action’s rightness all things considered.”³ To quote Aaron James, Taking the theory on its own terms, the proper test of adequacy is not whether it could have determinacy of a kind it does not strive for but rather whether its general structure helps us reason through various specific cases in perspicuous ways.⁴ ³ Frick, “Contractualism and Social Risk,” p. 181 (emphasis omitted). ⁴ James, “Contractualism’s (Not So) Slippery Slope,” p. 291.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

108

   ?

In this chapter, I question whether ex ante contractualism can meet even that more modest test of adequacy. The basic moral intuitions on which contractualism rests are that distribution matters, and that sometimes the appropriate way to address distributional concerns is to tilt individual policy decisions in the direction of those who have the most to lose.⁵ That strikes me as a morally plausible view. But it is shared by many forms of nonconsequentialism, including prioritarianism and egalitarianism, as well as by forms of weighted aggregation. Contractualism’s claim to supply a distinctive pro tanto moral reason must stand or fall on its distinctive ethical contribution, which is the GBP. Doubtless, some number of cases can be resolved under the GBP, whether it is applied ex post or ex ante. (Whether they should be resolved by the GBP is a separate question, which I take up briefly at the end.) But the number is small, and confined to a highly stylized set of tragic choices in which the potential claimant class is closed, the interests of each member are quantifiable, and the epistemic perspective of each member is accurately captured in the epistemic perspective of the “neutral” agents deciding on their behalf. The tragic choices we face daily in the real world do not conform to those strictures. As applied to them, the GBP generates outcomes that misrepresent, often dramatically, the interests of persons who actually stand to gain or lose under a given choice. As a result, ex ante contractualism eliminates one serious problem (ex post contractualism’s inability to cope with risk) but replaces it with another: in many if not most cases, the individual-level probabilities fed into the GBP do not to correspond to the interests of actual persons who could be affected by the choice, now or in the future. For a theory that is grounded in respect and concern for embodied individuals, that problem would seem to be fatal. I take up that problem in Sections 1 and 2. In Sections 3‒6, I turn to other serious challenges facing ex ante contractualism that are not necessarily fatal, but

⁵ Assuming the declining marginal utility of money or goods, a utilitarian welfare function will automatically assign more weight to a poor person’s reasons than a rich person’s. By “tilt in the direction,” I mean assign even greater weight than would a standard welfare function, at the extreme (as in the GBP), giving the worst-off person a veto over the decision, conditioned on there being a lessbad worst outcome. “Individual policy decisions” refers to decisions about the distribution of specific in-kind harms or benefits, in contrast to broad-based tax and transfer policies that seek to redistribute overall wealth. There is a lively debate in the legal literature about the relative merits of using legal rules v. broad-based tax and transfer policies for redistributive purposes. The general consensus among economists is that, setting aside political feasibility, from a welfarist perspective the latter generally dominates the former. For a classic statement of that view, see Louis Kaplow and Steven Shavell, “Why the Legal System Is Less Efficient than the Income Tax in Redistributing Income,” Journal of Legal Studies 23, no. 2 (Jun. 1994): pp. 667–81. As I discuss below, the “individual reasons” commitment of contractualism focuses principally on the former.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

   

109

that have yet to be adequately addressed. These include deciding what categories of information are admissible under contractualism’s natural veil of ignorance; justifying contractualists’ de facto choice of the right epistemic moment at which to assess the complaints of parties to a hypothetical agreement; justifying the use of a maximin function; and making a plausible case that the GBP would apply broadly enough to justify treating it as a pro tanto moral reason of general application, as opposed to an ad hoc solution in a handful of cases. The last two of these problems carry over from ex post contractualism; the first two are introduced by the shift to ex ante contractualism.

1. Identifying the Relevant Claimants The GBP is a relative function: the answer it produces depends on the relative positions of different claimants. That fact puts enormous pressure on how we identify the set of relevant claimants. The pressure is greatly magnified by its maximin decision rule, in which outcomes are driven solely by the claims of the worst-off member(s) of a group. If the set over which a maximin function is applied is (Jane = −1, Bill = −4, Clara= −5), Jane’s and Bill’s complaints are assigned zero weight and Clara’s complaint prevails. If we expand the set to include (David = −100), Clara’s complaint is assigned zero weight as well, and David’s complaint prevails. In contrast, in an additive function like aggregation, the contribution that an individual’s expected harm makes to the rankings of different choices is invariant with the group over which the function is evaluated. If Jane is expected to lose 3 units of utility if society chooses X over Y, her contribution to a function comparing the total welfare generated by X and Y will be 3 units, irrespective of the composition of the group. As I discussed in Chapter 4, contractualists have taken the position, implicitly or explicitly, that the relevant class of competing claimants is determined by the stipulated facts of the hypothetical under consideration. If we are told that we must choose between developing a vaccine that will save ten children from death and developing a vaccine that will save 1,000 children from paralysis, the competing claimants are those 1,010 children. If we are told that a passerby has to choose between saving five people each with a 50 percent chance of success or one person with a 100 percent chance of success, the competing complainants are those six people. If the scarce resource in question is not transferable beyond a clearly bounded set and there are no knock-on effects from the choice, the justification for limiting the set of claimants to the members of that set is straightforward. This is the situation posited in trolley-type hypotheticals. Take Taurek’s famous example of five people stranded on one rock and one person on a nearby other. The innocent

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

110

   ?

bystander can, by virtue of her accidental presence at the scene, save one, five, or none.⁶ Those are her only socially available choices. But when the resource in question is money or any other transferable and fungible good, the number of potentially competing claimants is vastly larger than the on-stage group. This is the situation we face in most if not all collective resource allocation decisions in the real world. Consider the following numerical variant on one of the examples (Mass Vaccination (Known Victims)) that Frick uses to illustrate the operation of the GBP in the face of ex ante competitive claims.⁷ Assume we have to choose between two vaccines to treat a terrible virus that, if left untreated, will leave 10,000 identified children blind. Each of the vaccines would cost $10 million to produce, and we can afford to produce only one of them. Vaccine One will save every one of the 10,000 children from blindness but paralyze one of their legs. Vaccine Two will allow 9,900 children to survive the virus completely unharmed. However, because of a known particularity in their genotype, Vaccine Two is certain to be completely ineffective for a hundred identified children, with the consequence that they will go blind. Assuming that blindness is worse than paralysis in one leg, the GBP appears to dictate that we go with Vaccine One. But why is the relevant group of complainants limited to these 10,000 children? Money is scarce but fungible, which means that all potential claims on that $10 million are ultimately competitive. Suppose that we could save 5,000 people (or for that matter, just one) from death if we redirected the $10 million to purchase mosquito netting for the developing world, or run an ad campaign to increase organ donations, or repair our crumbling infrastructure. Since death is worse than either blindness or paralysis in one leg, under the GBP shouldn’t any one of those alternative uses of the money win out over Vaccine One or Two? If so, what is the justification for limiting consideration to the claims of the 10,000 on-stage children? There may be pragmatic reasons to do so in a particular case. The relevant legislative body appropriating the funds may restrict their use to domestic purposes, or medical purposes, or developing vaccines to treat this particular virus. The governmental agency that receives the funds will almost certainly have a limited subject-matter mandate that rules out most uses of the funds. In any given case, it may be infeasible for decision-makers to rank-order more than two or three alternatives. But unless some practical or legal constraint limits the use of the funds to Vaccine One or Vaccine Two, all that “Mass Vaccination (Known Victims)” tells us is how to rank-order two claims on the $10 million that are

⁶ John C. Taurek, “Should the Numbers Count?,” Philosophy and Public Affairs 6, no.4 (Summer 1977): pp. 293–316. ⁷ Frick, “Contractualism and Social Risk,” p. 183.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 - 

111

(from the moral perspective of contractualism itself) arbitrarily plucked out of a potentially infinite set of competing claims. Contractualism (ex ante or ex post) is not committed in principle to such myopia, but it has a large stake in perpetuating it. First, the larger the number of claimants, the less feasible any individual-level inquiry into expected consequences to them. Second, the larger the number of claimants, the larger the number of claimants whose interests will count for nothing under a maximin rule. For most correctly specified problems, contractualists are likely to find the resulting divergence between contractualist and utilitarian solutions unacceptable—a fact acknowledged by contractualists themselves. Finally, the larger the number, the less likely that any arbitrarily selected group of claimants will contain the worst-off claimant. This is the lesson of Peter Singer’s and Elizabeth Ashford’s challenge to Scanlon’s classic “easy rescue” hypothetical. A child is drowning in a shallow pond. You happen to walk by and see the child in distress. You could wade in and save her at negligible cost to yourself (ruining a pair of shoes, say). Of course you must save her, Scanlon concludes, because the burden on you of doing so is trivial compared to the benefit to the child.⁸ But as Singer and Ashford observed, the same calculus holds if, instead of saving a drowning child you happen upon, you have the opportunity to save the life of an anonymous starving child halfway around the world at the cost of the same pair of shoes.⁹ That leaves contractualists with two choices, both of which they are likely to find unpalatable: deny that the starving child has an equal claim to be rescued, or accept that she does and thereby lose any purchase on the problem. Taking a less myopic view of competing claims on scarce resources inevitably pushes contractualism away from its focus on individual pair-wise comparisons and toward a collectivist solution that threatens to swallow contractualism whole. The duty of easy rescue can no longer plausibly be described as arising from the fact that the interests of the would-be rescuer and rescued are linked in a particular transaction. Like utilitarianism and Rawlsianism, both of which it threatens to bleed into, that broader conception of duty arises from an impersonal obligation to help all those considerably less fortunate than one’s self.

2. Calculating Individual-Level Probabilities Any function (including utilitarianism) that compares the (dis)value of different expected outcomes requires that we be able to assign commensurable values to

⁸ Peter Singer, “Famine, Affluence, and Morality,” Philosophy and Public Affairs 1, no. 3 (Spring 1972): pp. 229–43, 231; Scanlon, What We Owe to Each Other, pp. 223–9. ⁹ Singer, “Famine, Affluence and Morality”; Elizabeth Ashford, “The Demandingness of Scanlon’s Contractualism,” Ethics 113, no. 2 (Jan. 2003): pp. 273–302.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

112

   ?

them. That it is often difficult to come up with values that are better than noise counts against both contractualism and utilitarianism, and there is no reason to think that it counts more heavily against contractualism.¹⁰ For the balance of the discussion, I set this shared difficulty aside to focus on the additional problems contractualism faces as a result of its requirement that expected outcomes be calculated at the individual level. For a consequentialist, concerned only with the frequency and severity of outcomes in the aggregate population, individual-level probabilities are irrelevant.¹¹ Offered by a consequentialist, the statement that each member of a population of 100,000 has a .0005 (1/50,000) chance of dying from risk X is not a statement about individual prospects. It is a statement about the expected number of deaths in a given population, framed in probabilistic rather than frequentist terms. In contrast, individual-level probabilities of harm are critical to ex ante contractualism. Along with the severity of expected harm, they determine the weight given to each individual’s complaint. Ex ante contractualists’ reliance on individual-level probabilities is problematic for a number of reasons, which I take up in the balance of this section.

2.1 Being Exposed to a Risk of Harm Is Not the Same as Being Harmed As discussed in previous chapters, in the fantasy world of complete foresight—the world of trolleyology—the ex post distribution of harm and the ex ante distribution of risks of harm converge. Once we move from that fantasy world to the real one, where epistemic uncertainty is ubiquitous, ex ante contractualists have (rightly) concluded that ex post outcomes cannot guide the ex ante allocation of risks. But as others have noted, in abandoning a focus on ex post outcomes,

¹⁰ Indeed, it arguably counts less heavily against ex ante contractualism, because a maximin function requires only ordinal rankings of outcomes, while utilitarianism requires cardinal values to be assigned to every outcome. ¹¹ I set to the side possible demoralization costs from choosing the apparently optimal consequentialist solution. As an empirically contingent matter, such costs might lead consequentialists to take the distribution of probabilities (or, indeed, the personal identifiability or proximity of potential victims) into account on consequentialist grounds. I also set to the side nonconsequentialist concerns like the right of agents, moved by the seeming fairness of equalizing probabilities, to act on their agent-relative preferences, or the duty we have to respect reasonable, divergent views on the significance of epistemic risks. For discussion of some of these factors, see Norman Daniels, “Can There Be Moral Force to Favoring an Identified over a Statistical Life?,” in Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015); Marcel Verweij, “How (Not) to Argue for the Rule of Rescue: Claims of Individuals versus Group Solidarity,” in Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015), p. 137.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 - 

113

nonconsequentialists necessarily distance themselves from the fundamental truth that motivated Scanlonian ex post contractualism in the first place: what matters most to real, embodied individuals is what actually happens to them, not what might have happened but didn’t. When ex post remediation of harm is possible, the much greater significance of what actually happens may supply a strong argument for remediation on (distributional) fairness grounds, however fair or voluntary the ex ante distribution of risk. But when the outcome is death or some other grave harm that cannot be remediated, the fair allocation of ex ante risks, indirect and imperfect as it is, may be the only tool available to address distributional concerns. It is worth keeping in mind, however, that distribution of risks is already one very large remove from what embodied individuals have the strongest reason to care about. At the moment, contractualists and fellow travelers are divided on whether the fair distribution of risks is normatively significant enough to ground a moral theory built on individual reasons. Most of the philosophical articles in the recently published collection Identified versus Statistical Lives are devoted to this debate.¹² Most ex ante contractualists have assumed the answer is yes, and have turned their attention to how to calculate and incorporate individual-level risks into the GBP. Others have expressed skepticism that, in Nir Eyal’s words, we have “inherent duties to prioritize those at concentrated risk as a matter of distributive fairness.”¹³ That skepticism may be warranted. But if ex post outcomes and ex ante risks are both rejected as the appropriate measure of individual burdens to be fed into the GBP, what measure will substitute?¹⁴

2.2 How Do We Assess Individual-Level Risks? Assuming that the equitable distribution of risks is sufficiently weighty to undergird a moral philosophy grounded in individual reasons, a further challenge faces ex ante contractualism: does the method used to quantify individual-level risks produce results that correspond to what individuals themselves would recognize as expressing their interests, to what contractualists themselves mean by individual reasons, or to other plausible intuitions of fairness?

¹² See, in particular, the contributions of Brock, Otsuka, Hare, and Verweij, Identified v. Statistical Lives. ¹³ Nir Eyal. “Concentrated Risk, the Coventry Blitz, Chamberlain’s Cancer,” in Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015), p. 94. ¹⁴ For thoughtful discussions of some of the issues raised in this section, see the contributions of Dan Brock, Michael Otsuka, Casper Hare, and Marcel Verweij in Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

114

   ?

The answer is sometimes but usually not. An adequate explication of the issue requires more space than I have here. But here in brief are the key problems: 1. Objective (meaning observable) probabilities have meaning only in relation to a class of events (a “collective” or a “reference class”). In Richard von Mises’s words, they express the probability (frequency) of encountering “a certain attribute in a given collective,”¹⁵ and are a property of that collective, not of any individual member or a single draw from the collective. 2. Statements about individual-level risks are statements about degrees of belief; hence the common descriptor, “epistemic risks.” A belief about a risk that a particular individual faces may be more or less justified, given what we know about the world. But it is not “objective,” in the sense of being an observable property of the individual. 3. In some cases, we may have a strong basis for a justified belief about a particular individual’s risk. Take the classic trolley problem. Suppose we are confident that if the trolley continues down track 2 unimpeded, it will kill all ten people, but that even if we do not divert the trolley, there is an 80 percent chance that some other mechanism will stop it before it hits the ten. In that case, we have a high level of confidence that all ten are in the identical position with respect to death, because we understand how the relative causal mechanisms would operate. In such a case, it doesn’t seem a stretch to call the .2 increase in the chance of death each person’s individual reason. 4. Suppose, instead, that a vaccine to prevent a deadly illness has been administered to one million people, ten of whom have died from an adverse reaction. Scientists don’t know what caused those ten and no others to die.¹⁶ ¹⁵ Richard von Mises, Probability, Statistics and Truth, 2nd ed. (New York: The MacMillan Co., 1956), p. 12. ¹⁶ The contrast between the types of uncertainty posed in points (3) and (4) is sometimes described as the difference between uncertainty that is “a property of our incomplete knowledge” and uncertainty that is “an inherent part of the phenomenon being predicted.” Steven Goodman illustrates the difference as follows, in the context of medical risks: “[C]linical outcomes are produced by a causal chain of events, albeit a chain we may not fully understand. If we understand it, we would be certain about the outcome. If this is so, then what does probability mean when applied to an individual case? Consider what would happen to 100 identical copies of Mr. Smith if each underwent [an operation with a 40 percent mortality rate]. . . . Would 40 Mr. Smiths die? Or would all 100 Mr Smiths either live or die, with 40 percent representing our uncertainty about which it would be? The first of these scenarios represents the stochastic interpretation of risk, implying that risk is a physical characteristic of Mr. Smith. In the second scenario, Mr. Smith’s fate is determined, and potentially explainable. What we are calling his risk is actually a measure of our incomplete knowledge: the degree-of-belief interpretation.” Steven N. Goodman, “Probability at the Bedside: The Knowing of Chances or the Chances of Knowing?” Annals of Internal Medicine 130, no. 7 (1999): pp. 604, 605. James Robins put the point this way: “In a large ideal epidemiologic study, one can accurately determine the total number of [years of life lost] due to exposure because the total depend[s] only on the data . . . . On the other hand neither the probability of causation nor the correct apportionment of the total number of [years of life lost] among the exposed subjects can be empirically determined from the data, since both depend on the unknown and empirically unverifiable . . . biological mechanisms (at the sub-cellular and genetic level) by which exposure causes disease.” James Robins, “Should Compensation Schemes Be Based on

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 - 

115

Saul is considering whether or not to be vaccinated. Is the 10/1,000,000 frequency of death relevant to Saul’s individual position? It doesn’t reflect his objective probability of death: that is either 0 or 1. But if the class is wellconstructed and observed outcomes numerous enough, in the absence of information concerning causal factors that would allow us to narrow the class down, most of us would regard the information as relevant to our individual situation, and rationally take it into account in deciding whether or not to get vaccinated. 5. But what if there is no objective basis on which to establish the relevant class and/or no reliable data about the frequency of bad outcomes? This is the case in most of the tragic tradeoffs under uncertainty that we face in the real world, as well as those that are posed in the nonconsequentialist literature. Contractualists have sidestepped the problems entailed in identifying the relevant class and the frequency of bad outcomes by stipulating both numbers, and then equating each individual’s interest with the resulting ratio. In constructing the class, they have also confounded two very different forms of uncertainty: objective uncertainty about outcomes, and their own subjective uncertainty about the identity of the winners and losers. The resulting risks of harm assigned to different individuals is at best noise, and sometimes quite a bit worse, systematically favoring some groups over others without any plausible normative justification. Consider the following problem, which is the subject of an ongoing debate in biomedical ethics.¹⁷ There are roughly 1.1 million HIV-infected people in the US. The average age of that group is 35. If left untreated, the average member will die at age 45. The current treatment regimen for AIDS, if followed for the rest of their lives, will extend the life of an average member to 70 years (an additional twenty-five years), at a lifetime cost of roughly $350,000 in present value terms. On those assumptions, treatment will add 27.5 million life-years (1.1 million × 25 additional years) to the currently infected population, at an aggregate cost of $385 billion (1.1 million × $350,000). the Probability of Causation or Expected Years of Life Lost?,” Journal of Law and Policy 12, no. 2 (2004): pp. 543–4. ¹⁷ The numbers are chosen for illustrative purposes, but on the treatment side, are roughly in line with current estimates. For opposing views on the choice of treatment v. prevention, see Dan W. Brock and Daniel Wikler, “Ethical Challenges in Long-Term Funding for HIV/AIDS,” Health Affairs 28, no. 6 (Nov. 2009): pp. 1666–76; and Johann Frick, “Treatment versus Prevention in the Fight against HIV/ AIDS and the Problem of Identified versus Statistical Lives,” in Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015), pp. 182–202. Hypotheticals in the nonconsequentialist literature on risk that present structurally similar problems that are deflected by the author’s stipulating class size include Frick, “Contractualism and Social Risk,” p. 212; Sophia Reibetanz, “Contractualism and Aggregation,” Ethics 108, no. 1 (1998): p. 302; Kumar, “Risking and Wronging,” pp. 34–5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

116

   ?

Suppose we estimate that if we invest just a fraction of that amount in developing a vaccine, we can prevent 50,000 people a year from contracting HIV in the US alone. The average lifespan of an uninfected person is 80 years. Over the same thirty-five-year period that HIV treatment would add an aggregate of 17.5 million life-years for already infected persons, we could thereby prevent 1.75 million people (50,000 × 35) from contracting AIDS, saving 61 million life-years (1.75 million × 35 additional years) at a fraction of the cost. The marginal cost of distributing the vaccine once it is developed is effectively zero, so every year thereafter, at no additional cost, we could save an additional 1.75 million life years. If the vaccine were distributed world-wide, including to countries that currently cannot afford HIV treatment, we could increase life-years saved many-fold, still at no additional cost. If we use whatever amount is left over from that original $385 billion on other cost-effective opportunities to increase lifespan, we can increase life-years many-fold more. If the goal is to maximize total life-years (the consequentialist solution), investing in prevention is the obvious choice. If instead we follow nonconsequentialists’ “fair distribution of risks” principle using a maximin decision rule, prevention loses out to treatment resoundingly. The identities of persons currently infected with HIV are known or easily discoverable, limiting the reference class to 1.1 million people on the treatment side. On the prevention side, we know little or nothing about the identity of the people who would be saved by prevention, almost all of whom are not yet in existence. Lacking that information, we have no basis on which to narrow the reference class of future claimants to less than the entire future population of the US (or the world, if we assume our obligations of aid are global). However effective prevention might have turned out to be and however many more life-years it would have saved, prevention will always lose out to treatment under the nonconsequentialist solution, for two reasons. It intentionally ignores the numbers of persons who will be similarly affected by a given choice (nonconsequentialists’ refusal to “let the numbers count”), and then massively discounts the individual-level burden of a given choice in proportion to our ignorance about the identities of the ultimate victims. The result is to systematically favor current over future persons, the proximate over the distant, the familiar over the unfamiliar. The preference for those whose identity we happen to know (or imagine we know) is antithetical to the most basic commitment of nonconsequentialists: “For a moral principle to be valid is for it to be justifiable to each person from her own point of view.”¹⁸ There is no actual person alive now or who will be alive at any ¹⁸ Frick, “Contractualism and Social Risk,” p. 190 (emphasis added); see also Rahul Kumar, “Defending the Moral Moderate: Contractualism and Common Sense,” Philosophy & Public Affairs 28, no. 4 (Autumn 1999): p. 295.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 “ ” 

117

time in the future whose epistemic viewpoint is reflected in the statistic that every American born in the next one hundred years has a 50,000/300 million (or every human a 50,000/7 billion) chance of being saved by prevention. All that statistic reflects is the vast ignorance of those of us currently alive about the identity of future persons who would have been saved had we opted for prevention over treatment.¹⁹ If our ignorance of their identity determines the weight of their claims with respect to the choice we are about to make between treatment and prevention, future persons will lose out every time. Agents alive now, faced with making a decision that will bind future generations, will almost always have more information about the defining characteristics (in some cases including actual identity) of present claimants than they will about future ones. As a result, the complaints of future claimants will be systematically diluted—often massively so—by current agents’ ignorance of who they will turn out to be. Future claimants will generally not have an opportunity to revise past agents’ choices when they can finally speak for themselves. That train will have left the station. Once they come into existence and discover that they are the ones whose lives would have been extended had we chosen prevention instead, why should they conclude that we took their individual reasons seriously? After all, the cost to them individually (premature death) as a result of our opting for treatment over prevention is the same as the cost would have been to those alive now and infected with AIDS had we chosen prevention instead (and, an aggregationist would add, there are many more of them).²⁰ Why should they not conclude that they have written off by our partial moral vision?

3. The “Full Information” Requirement As I discussed in Chapter 4, Scanlon’s original (ex post) version of contractualism endows representative individuals (or their agents) with two kinds of information about a proposed course of conduct. The first is all individual-level information about its expected consequences that would “naturally” be available to individuals at the moment of hypothetical agreement (what Scanlonians call a natural veil of ignorance).²¹ That includes general risks associated with such conduct, aspects of the individual’s own situation that put her at greater or lesser risk than others, and ¹⁹ Indeed, that ignorance will persist into the lifetimes of those future persons. As Marcel Verweij notes, successful prevention is the dog that doesn’t bark in the night: neither we nor the ultimate beneficiaries will ever know for sure whether they were the ones who would have died but for our investment in prevention. “How (Not) to Argue for the Rule of Rescue.” ²⁰ See Verweij, “How (Not) to Argue for the Rule of Rescue,” pp. 141–3. ²¹ Frick, “Contractualism and Social Risk,” p. 190. For Scanlon’s version, see T. M. Scanlon, “Contractualism and Utilitarianism,” in Utilitarianism and Beyond, edited by Amartya Sen and Bernard Williams (Cambridge: Cambridge University Press, 1982), pp. 120–8.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

118

   ?

her own subjective preferences (over possible outcomes, about risk-taking, etc.). The second is the actual ex post consequences to representative individuals under each available option. In the revised (ex ante) version of contractualism, Scanlonians have removed the second kind of information, but have retained the first. I touched on the problems with the first kind (Scanlon’s natural veil of ignorance) in Chapter 4. But in the context of ex post contractualism, those problems were overshadowed by the second (actual ex post consequences). Who cares what she “naturally” would have known about her odds of being harmed when she also knows (peeking ahead) whether she will in fact be harmed? With the knowledge of ex post consequences removed, the problems with Scanlon’s natural veil of ignorance come to the fore. Operationalizing the natural veil entails two decisions: determining what counts as information “naturally” available to individuals or their agents at the moment of hypothetical agreement, and determining the appropriate (epistemic) moment for securing that agreement. I start with the first, and take up the second in Section 4. As I discussed in Chapter 4, the natural veil is intended to mimic the actual knowledge an individual would possess at the moment of hypothetical agreement. In Rahul Kumar’s words, “Valid moral principles are principles that must be justifiable to each person (provided she is appropriately motivated), from her own point of view, with no informational restrictions.”²² Contractualists have cashed out the “no informational restrictions” requirement as follows. In addition to whatever information is already known to them, agents are required to seek out further information that could differentiate the expected burdens that individual claimants face, provided they can do so relatively costlessly.²³ The example Frick provides of a “simple and costless” task is requiring agents to administer to each child an already developed genetic test that can identify with 100 percent accuracy those children who will be killed by a proposed vaccine.²⁴ The example is meant to suggest that agents can easily ascertain what additional information is easily obtainable, and that stretching the definition of naturally available information to include the obligation to obtain it will not significantly compromise the “naturalness” of the information made available to claimants’ agents. ²² Rahul Kumar, “Defending the Moral Moderate,” pp. 275–309. ²³ Frick, “Contractualism and Social Risk,” p. 193; Kumar, “Risking and Wronging.” More precisely, if agents fail to seek out easily knowable information that would have identified (or at least narrowed the class of) individuals facing a higher risk, agents must treat those individuals as if they had been identified, thereby converting a case of noncompetitive claims into a case of competitive claims and giving the victory to those hypothetical individuals who could have been identified as the big losers but weren’t. The upshot is to push ex ante contractualism strongly in the direction of ex post contractualism, with all of its insoluble problems. ²⁴ Frick, “Contractualism and Social Risk,” p. 193.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 “ ” 

119

Perhaps Frick and Kumar are right, but I’m doubtful. There is a lot of daylight between simple and costless on the one hand and extremely costly on the other, neither of which is defined.²⁵ In addition, it is not clear that cost should be, or in most cases will be, the major determinant of whether an agent has a moral obligation to seek out additional information. As a result, when all the dust settles, extending the “natural information” requirement to include information that is unknown but knowable is likely to produce knowledge conditions that are no less artificial (in the sense of constructed) than under the Rawlsian veil, and in many cases a good deal less impartial. Consider the following variant on Frick’s Mass Vaccination (Known Victims) hypothetical. A policy-maker (Fred) must decide whether to invest limited societal resources in mass-producing Vaccine A or Vaccine B to prevent a potentially deadly disease. Clinical trials show the expected adverse reactions to Vaccine A to be lower than those to Vaccine B, but they provide no information that would allow Fred to differentiate among individual claimants. Unlike in Frick’s original hypothetical, there is no genetic test already developed that could cheaply and easily identify the children who will actually suffer each of these harms. Given the information available, the GBP dictates that Fred, acting as a neutral agent on behalf of each of the children, select Vaccine A. But before Fred pulls the trigger on Vaccine A, a representative of one of the girls in the claimant group points out that the subjects in the clinical trials on which Fred is relying were all boys, and that there is a nontrivial possibility that girls will respond differently to the two vaccines. To test that hypothesis, Fred would have to order a second clinical trial, this time on girls, at considerable additional cost and delay. Is Fred obliged to order the second clinical trial before he makes the choice? Until very recently, the answer routinely given in the medical world was no: researchers and practicing physicians (the overwhelming majority of whom were men) assumed that what was good for men was good for women, and policymakers and funders never required them to test that assumption.²⁶ I would guess that ex ante contractualists, along with most of the rest of us, would answer the ²⁵ Rahul Kumar takes up the same example of genetic tests that can determine whether any given individual will have an adverse reaction to the vaccine, but with one difference: The determination will require “a series of complex tests,” not just one, and because the test results can be inconclusive, each test may require “several repetitions over a span of time in order to produce reliable evidence”; Kumar, “Risking and Wronging,” pp. 46–7. That added cost and delay is enough for Kumar to conclude that it is permissible for an agent not to administer the tests. But, of course, others could easily conclude that the point at which further investigation becomes “too costly” is far in excess of the costs Kumar rejects as too high. After all, we’re talking about avoidable death here. ²⁶ It wasn’t until 1994, under pressure from women researchers who had entered the medical profession in meaningful numbers over the previous decade, that the federal government required federally-funded studies to include roughly equal numbers of men and women and test for gender disparities in the data collected, absent some compelling reason not to. Many studies still do not comply with that requirement. Carolyn M. Mazure and Daniel P. Jones, “Twenty Years and Still Counting: Including Women as Participants and Studying Sex and Gender in Biomedical Research,” BMC Women’s Health 15 (2015): available online at http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4624369/.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

120

   ?

question differently now. But that answer, like the answer to Frick’s genetic test example, is too easy. In Frick’s original example, given the proven reliability of the test and the cheapness of administering it, any doctor who failed to administer it before vaccinating her patient would be committing malpractice. In the gender case, the answer is too easy for a different reason. Given the raft of medical studies over the past twenty years that have found significant gender differences in the efficacy of preventive measures, diagnostic techniques, and treatment across a wide range of health problems, most impartial agents (including all contractualists, I am guessing) would conclude that they cannot responsibly ignore the possibility that gender is an important variable in predicting outcomes in this case. Here is a more difficult question: If we are morally required to investigate the relationship of gender to vaccination outcomes, are we also required to do the same for race, country of origin, income class, early childhood environment, diet, and a host of other individual characteristics that might affect outcomes? Doesn’t each member of these demographic subgroups have an equal moral right to know, to the extent it is knowable, what her individual prospects are? It is tempting to say yes, study them all, but we can’t and we won’t. On what basis then would an ex ante contractualist have us make the choice? It can’t be based on the social significance of the characteristic (e.g., race v. diet). That difference is of no moment to any individual at risk of death, who wants to know only what his prospects of dying are. It also cannot be based on the contributions that each answer will make to aggregate well-being: that is an aggregationist’s concern. Consistent with contractualist premises, it could be based on the likelihood that further investigation would establish that a given characteristic has explanatory power. But as the history of medical research attests, we often can’t know the likelihood without undertaking the investigation. Inevitably, the choices we make—indeed, the facts that it even occurs to us we don’t know and might be able to find out upon further investigation—will be shaped by a host of subjective factors unrelated to the costs of undertaking further investigation, the likelihood of finding a statistically significant correlation, or any other neutral criteria. Those factors include the social power of different subgroups in the population, social and cultural norms, and whom an agent feels sympathy for. The same factors will inevitably influence which of the many facts already known to someone the agent will seek to learn. Each of these subjective factors pulls agents strongly towards partiality of one form or another.²⁷ In this connection, Alex Voorhoeve’s recent attempt to justify deference to the sympathies of agents on the ground that they reflect the exercise of a

²⁷ Verweij expresses similar concerns about the argument that we ought to favor rescue over prevention because rescue enhances community solidarity. “How (Not) to Argue for the Rule of Rescue,” pp. 146–8.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    ()   ? 121 Smithian “sympathetic identification” seems hopelessly optimistic.²⁸ Our “sympathetic identification” could, in theory, lead us to a more impartial and ecumenical view of the world. But it is much more likely to confirm our existing biases—not just in the moral realm, but in the cognitive one as well.²⁹

4. When Is the Appropriate (Epistemic) Moment for Agreement? However contractualists cash out the “available information” condition, the information thereby admitted will depend on what contractualists deem to be the right epistemic moment for assessing it. For ex post contractualists, the answer is clear: it is the moment when we know the actual consequences to different individuals as a result of the choice at hand. That knowledge is imputed to agents in their temporally ex ante posture, whether or not they “naturally” would have access to it at that moment (yes if the choice is made under epistemic certainty, no if it is not). Most ex ante contractualists have not addressed the appropriate epistemic moment explicitly.³⁰ But their answer is implicit in the near-universal practice of assessing individual-level probabilities at the epistemic moment that the “facts” posed in a hypothetical require an exigent choice to be made. As suggested in Chapter 4, I believe it is impossible to defend the view that the right epistemic moment at which to assess individual-level interests is a mere question of “fact” exogenous to the normative dimensions of contractualism. Take the choice posed by Frick in the Mass Vaccination (Known Victims) hypothetical above. We must choose between two vaccines to treat a virus which, if left untreated, will leave 10,000 children blind. Vaccine One will save each of the 10,000 from blindness but leave each with a paralyzed leg. Vaccine Two will allow 9,900 children to survive the virus completely unharmed, but will be ineffective for a hundred children, who will therefore go blind. In assessing the choice from the epistemic perspective of a given child (Elena), which epistemic moment is the right one? Before Elena “naturally” knows whether she will be one of the 10,000 children threatened with blindness (the Rawlsian veil)? After she “naturally” learns that she is one of the 10,000 children but before she learns whether she is one of the hundred or one of the 9,900 (the posture of Frick’s “noncompetitive claims”)? After she “naturally” learns that she is one of the hundred (the posture of Frick’s “competitive claims”)? The more local and ²⁸ Voorhoeve, “How Should We Aggregate Competing Claims?,” p. 69. ²⁹ On the cognitive front, as John Halstead observes, the difficulty we have in grasping the import of large numbers creates a strong bias in favor of nonaggregation. John Halstead, “The Numbers Always Count,” Ethics 126, no. 3 (Apr. 2016): p. 789–802. ³⁰ Frick’s argument for “stage-wise ex ante contractualism,” discussed below, is an exception.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

122

   ?

granular the “natural” epistemic position chosen, the closer ex ante contractualism comes to the complete knowledge condition of ex post contractualism. Conversely, the more general and less granular the “natural” epistemic perspective chosen, the closer ex ante contractualism comes to the “no particularized knowledge” condition of the complete veil of ignorance, and with it, to the aggregative solution. In practice, contractualists have opted for the local and granular in constructing hypotheticals and assessing the strength of an individual’s complaints under the available choices. In general, readers have not questioned that choice. To the extent contractualists have defended it explicitly, it is by reference to contractualism’s “individual reasons” restriction: the only complaints that count are those held by embodied individuals. But as I argued in Chapter 3, the individual reasons restriction cannot resolve the choice of epistemic moment. Elena is Elena whichever of the three moments we pick, and she knows whatever she naturally knows at that moment. The choice has to be made on other, normative, grounds. The “stage-wise ex ante contractualism” adopted by Frick, Frances Kamm, and others extends the privilege given to the local and granular beyond the epistemic moment of the original agreement (t1), to include information that “naturally” becomes available after the original hypothetical agreement is struck (t2).³¹ To put this in operational terms, agents will generally respect an individual’s choice to assume a risk if it is in her interest to do so at the moment she agrees to assume it. If the payoffs from that gamble are instantaneous or inexorably set in motion at the time of the initial agreement (“single-stage contractualism,” in Frick’s terms), the parties are stuck with those payoffs. If, however, as a matter of contingent social fact, the parties have the opportunity to revisit and revise their decision before performance becomes irrevocable, agents must reevaluate the gamble on behalf of each party in light of each new piece of information acquired. If the original gamble is still justifiable to each person given the updated information available, then it stands. If it isn’t, then (in Frick’s words) the justifiability of the initial gamble at t1 “unravels from the end, as if by backward induction.”³² ³¹ Frick, “Contractualism and Social Risk,” pp. 205–6; Frances Kamm, Intricate Ethics (Oxford: Oxford University Press 2007), pp. 274–5. See also James Lenman, “Contractualism and Risk Imposition,” Politics, Philosophy and Economics 7, no. 1 (Feb. 2008): p. 117. Kamm and Frick both resolve the conflict between earlier and later epistemic viewpoints in favor of the later. Lenman suggests we resolve it by weighing both viewpoints in some kind of reflective equilibrium; “Contractualism and Risk Imposition,” p. 116. As I discuss in Chapter 4, that solution is not viable, and in practice will cash out in favor of the later viewpoint. ³² Frick, “Contractualism and Social Risk,” pp. 205–6. Contra Frick, this is not the same as saying individuals may not alienate the right in question ex ante but may only waive it contemporaneously with performance. If the original contract is never revisited, under the “unraveling” view it remains in force and governs the disposition of the case. In contrast, under the “no prior alienation” view, there would be no valid agreement in place to resolve the dispute, leaving the parties to hammer out some resolution at the moment before irrevocable performance, if the opportunity arises, and if not, ex post. This is not a technical distinction. Under the former view, most ex ante contracts would end up enforceable. Under the latter, none would.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    ()   ? 123 In addition to extending the as-yet unjustified preference for the local and granular, stage-wise ex ante contractualism eliminates the option to create a binding precommitment. In defense of that result, Frick, Kamm, and others have argued that stage-wise contractualism follows from “[t]he moral ideal underpinning contractualism [which] is that we regulate our behavior in such a way that our actions are, at all times, justifiable to each other.”³³ But this argument conflates two questions: whether we must be able to justify our actions to each other at both t1 and t2 (yes, per ex ante contractualists), and what constitutes a good justification at either time. Advocates of stage-wise contractualism assume that X’s hypothetical agreement to a given action at t1 cannot justify enforcing it against X at t2 when newly acquired information reveals it is no longer in X’s interest. But they haven’t offered an argument for that conclusion. The obvious reason to enforce the original agreement at t2, notwithstanding newly acquired information, is that that’s what the parties agreed should happen when they entered into the agreement at t1. An analogy to the institution of private contracting may be illuminating here. Normally, if a person enters into a binding contract at t1 committing him to sell his house to a stranger for $500,000 at t2, we would say that at t1 he alienated the right not to sell his house at that price to the buyer at t2. That’s what a contract does. If we allow parties to a contract entered into at t1 to walk away from it if, at any point prior to delivery, they conclude it is no longer in their interest (the “unraveling” view), we relegate them and their contracting partners to a barter system: all parties are stuck with whatever deal they can extract for themselves at the latest moment before performance becomes irrevocable, given whatever “natural” bargaining leverage they possess at that moment. This is precisely the position that stage-wise ex ante contractualism forces parties into, albeit with an individual’s advantage coming not from market power but from the fact that she faces the worst expected outcome at that epistemic moment. The only difference from an arms’ length barter is that under the GBP, the leverage that parties gain from being the expected big loser is severely constrained by the other pro tanto reasons contractualists would have us weigh in coming to an “all things considered” answer. That is a hugely important difference in practice, but it is external to the operation of the distinct pro tanto reason (the GBP) offered by contractualism. Ex ante contractualists are, of course, under no obligation to conform their version of contractualism to the institution of real-world contracts just because they have adopted the same descriptor. But the project of contractualism, which is to come up with principles that suitably motivated people can agree to live by, seems to entail some version of precommitment. It is not clear why the version adopted for contracts in the real world—if you agree to do X, you are bound to do

³³ Frick, “Contractualism and Social Risk,” p. 205 (emphasis added).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

124

   ?

X absent compelling circumstances—isn’t the appropriate one here. At a minimum, the alternative view that stage-wise contractualism has adopted (agreements are always renegotiable in light of newly acquired information) has to be argued for. It doesn’t follow in any way from the “individual reasons” restriction at the heart of contractualism. Frick backs away somewhat from stage-wise ex ante contractualism in his discussion of Human Experiment, the facts of which are this: At t1, a doctor organizes a lottery amongst a group of 100 paraplegic children, which selects two of them by a random process. At t2, he conducts medical experiments on these two children, foreseeing (but not intending) that they will die in the course of the experiments. He knows for certain that the knowledge gained in this way will allow him to cure the remaining 98 children of their paraplegia.

Frick, concluding that it would be impermissible for the doctor to proceed with the experiment at t2, suggests that if the children (through their agents) “had alienated their right to life through some ex ante agreement,” he might be willing to enforce that agreement in the face of a later attempt to back out of it at t2.³⁴ But, he argues, that is not the case in Human Experiment. Why not? By stipulation it is in every child’s interest at t1 to choose Human Experiment, which choice entails a two-stage process: participation in a lottery at t1 followed by experiments on the losers at t2. Let us assume that every child’s legally empowered agent explicitly consents to that two-stage process at t1. If that is not enough to infer that each child “had alienated their right to life through some ex ante agreement,” what more is required? In the end, I don’t think contractualists’ reluctance to enforce the agreement in Human Experiment derives from the substance of the gamble the parties took at t1 or the absence of meaningful consent at t1 to enforcement of the lottery results at t2.³⁵ It derives, I would guess, from the anticipated horror associated with enforcing Human Experiment at t2. That horror does not come from the fact that (by stipulation) two children will die as a result of the experiment; the same is true as a matter of statistical certainty for many if not most collective risks voluntarily ³⁴ Frick, “Contractualism and Social Risk,” p. 206. ³⁵ Frick appears to concede as much when he states that he would likely refuse to enforce Human Experiment at t2 even if those who lost the lottery at t2 had clearly consented at t1 to follow through, and maybe even if they consented again at t2; Frick, “Contractualism and Social Risk,” pp. 206–7. Frick suggests a third explanation for what he intuits to be the right answer: we believe that persons lack the moral power to alienate their claims against certain kinds of treatment by advance consent. (Frick’s example here is advance consent to sexual intercourse). Frick, “Contractualism and Social Risk,” p. 207 n. 34. But this too would not differentiate Human Experiment from the typical voluntary assumption of risk case. The only difference between Human Experiment and the typical case is that in the latter there is usually no socially-available opportunity to reconsider the wisdom of one’s choice before it becomes irrevocable, while in the former there is.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 .     

125

assumed in real life. It comes rather from the fact that the deaths will be the result of direct human agency up close and personal, and both children will go to their deaths knowing they still could have been saved if the doctor had chosen to disregard the agreement reached at t1.³⁶ Frick’s Automatic Experiment hypothetical inadvertently pushes this alternative explanation to the fore. Automatic Experiment is identical to Human Experiment, except that at t1, the parties agree to a precommitment device that will automatically circumvent what would otherwise have been a “naturally” occurring opportunity to revisit and repudiate their agreement at t2. If the “natural” causal chain of events had made it impossible to revisit the agreement at t2—that is, if this were a single-stage agreement at t1—Frick concludes that the agreement at t1 would be morally binding. But because the parties have actively chosen at t1 to “set in motion an unstoppable device that will automatically perform the lethal experiments at t2,” he concludes it is not.³⁷ Why not? Frick argues that “[t]he only rationale for setting up such a device . . . is precisely to avoid the need to perform an action at t2 that we know could not be justified to each person at that time.”³⁸ Perhaps, but the more likely motivation, I would guess, is that the parties wish to avoid the psychological and emotional strain of having to execute their agreement at t2, once the identity of the losers is revealed. If at t1 everyone agrees to tie themselves to the mast for that reason, why shouldn’t their agreement be morally binding? Are people morally required to maximize their future opportunities to reconsider prior choices?

5. Individual v. Impersonal Conceptions of Moral Responsibility Up until now, I have focused on the problematic features of the GBP in action. Many of these problems, I believe, arise from tensions between two conflicting moral intuitions that animate the contractualist project: a commitment to individual reasons, and an impersonal commitment to aid the worst-off in society. Reflecting the animating idea that morality must take the difference between individuals seriously—or at least more seriously than utilitarianism or Rawlsianism does—contractualism builds in a number of individualistic features, many of which I touched on above. First, claims under the GBP attach to particular individuals, as do obligations to satisfy them. Second, the strength of an individual’s complaint is based on her ex ante prospects under a candidate principle, assessed in light of the information “naturally” available to her or her agent at the ³⁶ That horror may well have a moral dimension, but if it does, it is orthogonal to the concerns addressed by the GBP. ³⁷ Frick, “Contractualism and Social Risk,” p. 211. ³⁸ Frick, “Contractualism and Social Risk,” p. 211 (emphasis added).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

126

   ?

moment of hypothetical agreement. Third, claims and obligations arise between two individuals, X and Y, as a consequence of their being entangled in a particular choice situation; contractualism does not impose a free-floating obligation on X or Y to their fellow citizens or the world at large. Fourth, X’s and Y’s obligations to each other are limited to the stake each has in the choice in question: X may be morally obliged to forgo all (expected) gains from that choice so as to minimize Y’s (expected) losses, but she is not obliged to do more. Finally, the GBP as generally applied compares changes in well-being as a consequence of an exigent, isolated choice put before us. In all five respects, contractualism bears a structural kinship to a market model of distributive fairness, in the sense that individual obligations are created and bounded by an isolated transaction between two or more persons.³⁹ But the moral obligation instantiated in the GBP—to protect the worst-off in society—is at root an impersonal one. In contrast to a “just price” theory of distributive justice in real-world contracts or a corrective justice approach to tortious injuries, the obligation does not arise from individual malfeasance. As Frick states, contractualism is concerned with conduct that is “intuitively permissible.”⁴⁰ It arises instead from the unavoidably interdependent nature of our lives. Whenever we engage in intuitively permissible activities like driving a car or constructing a bridge, we unavoidably impose (risks of) harm on others who have a prima facie right to be free from such harms. When we spend $10 million of our collective resources to save one child, we necessarily deprive others of whatever benefits that money could have generated for them. Second, the maximin function is a comparative measure, rooted not in the absolute benefits and burdens imposed on a given individual but in the relative distribution of (incremental) benefits and burdens among individuals. Third, in weighing individual complaints, most contractualists would take into account an individual’s level of overall well-being, in addition to changes in her well-being as a result of the choice at hand.⁴¹ It is not clear how contractualists

³⁹ To be clear, the claim is not that the GBP in any way embraces a free-market solution to the scope of our duties to each other. Quite the contrary. The claim is rather that insofar as contractualism mimics some of the structural features of the market, those features constrain and distort the impersonal, welfarist, objectives that I take also to be at the core of contractualism. I omit here a sixth feature that Scanlon and others cite to as evidence of the individualistic nature of contractualism: that it proceeds by identifying, “by a series of pairwise comparisons, the action or policy that satisfies the strongest individual claim or—its flipside—generates the weakest individual complaint.” Frick, “Contractualism and Social Risk,” p. 176; see also Scanlon, What We Owe to Each Other, ch. 5. I omit it because the method of pairwise comparison has no effect on the ultimate answer generated by the GBP, which can be reached straightforwardly by applying a maximin function to the group as a whole without stopping to consider whose claim dominates in each of the possible pairwise comparisons within the group. ⁴⁰ Frick, “Contractualism and Social Risk,” p. 179. ⁴¹ Scanlon, “Contractualism and Utilitarianism,” pp. 122–3; Frick, “Contractualism and Social Risk,” p. 215; Reibetanz, “Contractualism and Aggregation,” p. 299; Voorhoeve, “How Should We Aggregate Competing Claims?,” pp. 66, 69 n. 9.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi



127

envision blending the two measures. More importantly, it is not clear why, if they care about levels of overall well-being, they would care at all about expected changes in well-being from an isolated transaction. Finally, while on rare occasion—rare in life, not in the philosophical literature—one might find oneself the only person in a position to rescue another, the individual obligation one thereby acquires to provide that aid is a secondary one created by happenstance. The primary obligation, contractualists would concur, runs to all of us in virtue of our membership in the relevant moral community. The effort to fit what is at root an impersonal obligation into an individualistic, transactional framework has produced a moral theory that sits uneasily in both camps. If my obligation to another arises not from my having treated her unfairly but rather from her greater need, the adventitious fact that we have competing stakes in a particular choice would seem morally irrelevant to the scope of that obligation. That is the essence of the Singer/Ashford challenge to Scanlon. Once one accepts that our individual duty to the less fortunate extends equally to far-flung strangers we have no connection to as it does to the drowning child we happen upon, the individualistic, transactional framework of contractualism loses all moral traction. What we are left with is some version of collectivism, be it welfarism, Rawlsianism or other forms of prioritarianism, or egalitarianism.

6. Conclusion Two other problems posed by the GBP deserve mention. The first, which some might consider the elephant in the middle of the room, is the moral implausibility of a maximin function. The normative problems with it have been widely discussed in the decades since Rawls made it a centerpiece of A Theory of Justice in his Difference Principle, and I will not rehearse them here. As I discuss in Chapter 8, Rawls quietly distanced himself from the Difference Principle over the years. But it lives on in the GBP, as well as in the writings of the many contemporary Rawlsians who continue to treat the Difference Principle as the starting—and often stopping—point for answering any distributive problem. Contractualists have limited the implausible implications of the maximin function built into the GBP by the other pro tanto considerations that constrain it. But the question remains, why preserve it at all? What advantage does it offer over prioritarianism or other measures of distributive equity that make no such extreme and implausible demands to begin with? Second, once we take into account the many other pro tanto moral reasons that contractualists would have us weigh in the balance, it is not clear what role remains for the GBP. Those other reasons include whether the would-be claimant assumed the risk of harm; whether the activity in question is pursued with due

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

128

   ?

care;⁴² whether it would be “too confining” not to permit the risky activity in question;⁴³ whether the outcome dictated by the GBP deviates too much from the optimal welfarist outcome;⁴⁴ the absolute severity of the harms in question;⁴⁵ whether expected losses in one transaction may be offset by gains in others;⁴⁶ whether “the policy alternatives at issue are all morally permissible” (if they are, Aaron James suggests, “an aggregative decision is not an issue of right or wrong, justice or injustice”); and whether “a legitimate public institution [has been] entrusted with scarce resources for a specified social purpose, and an aggregative decision would advance this aim.”⁴⁷ As James acknowledges, many of these other reasons push contractualism strongly in the direction of aggregate welfarism. The last two appear to cede public policy determinations to aggregation tout court.⁴⁸ I agree with Frick that contractualists need not produce a precise algorithm to show how the GBP should be weighed against these other considerations in order to defend its relevance.⁴⁹ But at a certain point, one is entitled to ask—as James himself does—whether it will play any significant role, once it is transported out of the carefully controlled conditions of the philosophical laboratory and into the real world. If not, then the one distinctive substantive contribution of contractualism—the GBP—reduces to an ad hoc decision rule with a vanishingly small domain. If there is a case to be made on the other side, it cannot be made by reference to carefully constructed thought experiments that stay within the confines of that domain. It has to be made by reference to the sorts of trade-offs we must make daily in our personal and collective lives, almost all of which require that the serious problems with ex ante contractualism be faced and answered.

⁴² For a discussion of how the requirement of due care smuggles aggregation in by the back door, see Chapters 4 and 5. ⁴³ Scanlon, What We Owe to Each Other, p. 209; James, “Contractualism’s (Not So) Slippery Slope,” p. 272. See also Kumar, “Risking and Wronging,” p. 34 n. 12. ⁴⁴ Frick, “Contractualism and Social Risk,” pp. 222–3; James, “Contractualism’s (Not So) Slippery Slope,” p. 282. ⁴⁵ Frick, “Contractualism and Social Risk,” pp. 223 n. 48. ⁴⁶ Reibetanz, “Contractualism and Aggregation,” pp. 299–300; James, “Contractualism’s (Not So) Slippery Slope,” p. 276. ⁴⁷ James, “Contractualism’s (Not So) Slippery Slope,” pp. 286–7. ⁴⁸ James comes close to conceding as much when he states that under his approach, aggregation would likely end up as the “policy default.” But he defends the continuing relevance of contractualism on the ground that “the default . . . will always stand open to scrutiny from a fundamentally nonaggregative moral perspective”; James, “Contractualism’s (Not So) Slippery Slope,” p. 289. One always has the option to reassess one’s moral perspective. The question is: when and why should we exercise it on behalf of contractualism? And do those exceptions add up to a distinct moral theory with a nontrivial domain? ⁴⁹ Frick, “Contractualism and Social Risk,” p. 223.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

7 Is Nozick a Libertarian? With this chapter, I turn from moral to political philosophy, starting with Nozickean libertarianism. The distinction between the two, however, is less clear than that division suggests. A theory of legitimate state power is simultaneously a theory about the scope of individual prerogatives and duties. Which side of the line we put a given theory on has less to do with any analytic distinction between the two than with what the theory foregrounds. Libertarianism, which I take up in this chapter and return to in Chapters 11, 12, and 13, pushes the complementarity of personal and political morality into high relief. Libertarianism is first and foremost a theory of individual rights. To quote the famous manifesto with which Anarchy, State, and Utopia (hereinafter ASU) begins: Individuals have rights, and there are things no person or group may do to them (without violating their rights). So strong and far-reaching are these rights that they raise the question of what, if anything, the state and its officials may do.¹

But as the final phrase suggests, it is also a theory of legitimate state power, which is reduced to almost nothing by the robust theory of individual rights presented in Part II of ASU. The same complementary relationship between individual rights and state power structures the two other political theories that I take up in the second half of the book: Rawlsianism and left-libertarianism. They differ from (right) libertarianism only in that the relative scope of individual rights and collective prerogatives is flipped. Rawls (the subject of Chapters 8 and 9) starts (contra utilitarians) with a strong commitment to rational choice built into the heuristic of a social contract, but immediately drains it of any power by eliminating all distinctions among individuals. That move leaves the just state unconstrained in its ability to pursue a collective vision of the (political) good. Left-libertarianism, the subject of Chapter 10, reaches essentially the same result within a Lockean framework. Like conventional (right) Lockeanism, it starts with the complementary Lockean principles of self-ownership (embodying individual rights) and the Lockean proviso (embodying collective rights), but flips their importance, yielding

¹ Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. ix. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

130

   ?

a distributive outcome that differs little from Rawlsianism and other forms of egalitarianism. Part II of ASU (“The State”) presents the core libertarian argument for an expansive view of individual rights and a concomitantly limited view of state power. Elaborating on the manifesto that opens the book, Nozick states: No one has a right to something whose realization requires certain uses of things and activities that other people have rights and entitlements over. Other peoples’ rights and entitlements to particular things (that pencil, their body, and so on) and how they choose to exercise these rights and entitlements fix the external environment of any given individual and the means that are available to him . . . The particular rights over things fill up the space of rights, leaving no room for general rights to be in a certain material condition.²

From these and other statements throughout Part II, one can glean the essential features of Nozickean rights: (1) They are absolute: If you are the owner of X (your body, your land), you control it absolutely. (2) One of the key rights of control you possess over X by virtue of owning it is the right not to have your interests in X interfered with or altered in any way without your consent. As a corollary, whatever price you negotiate for agreeing to an alteration is just: “[A]n entitlement theorist would find acceptable whatever distribution resulted from the party’s voluntary exchanges.”³ Even self-styled libertarians generally accept the need to limit an owner’s rights of enjoyment in some respects. Some of those limitations are innocently packaged in the classic common-law adage, sic utere tuo ut alienum non laedas (use your property so as not to injure others), the legal version of the Kantian principle that you may pursue your own projects in life, provided you do not interfere with others’ prerogatives to do the same. Other limitations arise from a recognition that the state must deploy some coercive powers to solve collective action problems for the good of all. Among libertarians, Nozick stakes out a position pretty far out on the “no interference” end in Part II: In general, people have a right not to be forced to do things they don’t want to do with themselves or their property, subject only to the right of the night-watchman state to use force, if necessary, to protect the like rights of others. This prohibition on interference holds, even if interference would leave others or even the person interfered with better off. Thus, for example, compulsory redistributive taxation is impermissible in a Nozickean world, even if everyone ² Nozick, Anarchy, State, and Utopia, p. 238 (footnotes omitted). ³ Nozick, Anarchy, State, and Utopia, p. 188.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   ? 131 supports helping the needy and concludes that because of collective action problems, coordinated action is a more efficient means to that end than uncoordinated individual action.⁴ If everyone would really prefer coordinated action, says Nozick, then get their consent. If you can’t get everyone’s consent because of dissenters, hold-outs, or the prohibitively high transactions costs of getting consent—tough. That’s what it means to have a historical rather than end-state theory. “[I]t would violate moral constraints to compel people who are entitled to their holdings to contribute against their will.”⁵ (3) Consent means explicit consent. As Nozick famously put it, “tacit consent isn’t worth the paper it’s not written on.”⁶ You may not imply a rights holder’s consent to a rights waiver from the fact that she has continued to reside physically in the jurisdiction that stripped her of that right. (To put it another way, individuals have a right of internal exit: they can stay put in a jurisdiction but refuse to abide by individual laws if they deem those laws unjust from a libertarian perspective.)⁷ And (pace H. L. A. Hart), you also may not imply consent to pay for benefits received from the state.⁸ (4) Natural rights are straightforwardly derivable from libertarian premises: “A line (or hyperplane) circumscribes an area in moral space around an individual. Locke holds that this line is determined by an individual’s natural rights, which limit the action of others.”⁹ Moreover, the rights derived from libertarian premises are not just abstractions like self-ownership. They are “particular rights over particular things.”¹⁰ (5) The boundaries around individual rights are non-overlapping. In Nozick’s words, “[i]ndividual rights are co-possible.”¹¹ (6) Rights occupy most of the space of social interaction, drastically limiting the domain of collective choice. “The exercise of . . . rights fixes some features of the world. Within the constraints of these fixed features, a choice may be made by a social choice mechanism based upon a social ordering.”¹² But, warns Nozick: After we exclude from consideration the decisions which others have a right to make, and the actions which would aggress against me, steal from me, and so on, and hence violate my (Lockean) rights, it is not clear that there are any decisions remaining about which even to raise the question of whether I have a right to a say in those that importantly affect me. Certainly, if there are any left to speak

⁴ Nozick, Anarchy, State, and Utopia, p. 268. ⁵ Nozick, Anarchy, State, and Utopia, p. 268. ⁶ Nozick, Anarchy, State, and Utopia, p. 278. ⁷ For Nozick’s insistence that only such a right of internal exit could legitimize the more-thanminimal state, see Anarchy, State, and Utopia, pp. 173–4, 292–3. ⁸ Nozick, Anarchy, State, and Utopia, pp. 90–3. ⁹ Nozick, Anarchy, State, and Utopia, p. 57. ¹⁰ Nozick, Anarchy, State, and Utopia, p. 238. ¹¹ Nozick, Anarchy, State, and Utopia, p. 166. ¹² Nozick, Anarchy, State, and Utopia, p. 166.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

132

   ?

about, they are not significant enough a portion to provide a case for a different sort of [that is, a more extensive] state.¹³

Over the past forty years, this conception of rights has been criticized on a number of grounds, in particular its analytic indeterminacy and dubious normative underpinnings. I agree with many of these criticisms, and explore some of them in later chapters. But with the exception of criticisms going to points (4) and (5) above, I don’t press them here. Instead, I want to suggest that when one reads the three parts of the book together, one begins to doubt that Nozick has any theory of individual rights. More precisely, he appears to have at least three mutually inconsistent theories: (i) utilitarianism; (ii) Lockean libertarianism; and (iii) anything goes, provided that citizens have some unspecified level of choice among legal regimes. Insofar as one can detect any predominant theory, it is not libertarian, or even liberal, but utilitarian. Some of the inconsistencies in Nozick’s treatment of property rights are doubtless explained by the book’s origins. As others have pointed out, the three parts of ASU started out as three unconnected essays, and in many ways that’s where they ended up. Other inconsistencies can be written off to the analytically casual nature of Nozick’s enterprise. If that were the end of the story, Nozick’s inconstancy to his professed libertarian commitments would be of limited philosophical interest. But many of Nozick’s lapses are symptomatic of the problems nonconsequentialists of all stripes encounter in translating abstract articulations of rights theory into concrete rules. His de facto solution is the same as that adopted by other nonconsequentialists when confronted with the inability of nonaggregative solutions to resolve the problem of harm to others: when the going gets tough, turn utilitarian. I start by considering Nozick’s treatment of individual rights in Parts I and III of ASU, then return to reconsider his defense of expansive individual rights in Part II. I conclude with some thoughts about how best to understand Nozick’s inconstancy to libertarian principles.

1. Anarchy The task Nozick sets for himself in Part I of ASU is to show that the minimal state could have come about from the state of nature (SON) through a series of uncoordinated, private transactions, none of which violates any individual’s rights.¹⁴ If some of the hypotheticals he uses in Part II to test libertarian property

¹³ Nozick, Anarchy, State, and Utopia, p. 270 (emphasis omitted). ¹⁴ Nozick, Anarchy, State, and Utopia, pp. xi, 113.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



133

rights aren’t exactly taxing (e.g., ruling out forcible eyeball transfer from the twoeyed to the blind), he has gone to the other extreme here. Nozick starts out with an imaginary SON populated with multiple independent protective associations (IPAs) in close proximity to each other. Each of the IPAs is pursuing the same prima facie legitimate end (protecting its own members from wrongful transgressions by nonmembers). In the course of doing so, each imposes identical risks on nonmembers by virtue of its unavoidably error-prone enforcement procedures. Being able to show that individual rights are co-possible—point (5) above—is essential to any theory of rights. If there is an absolute right to do X, but A’s doing X entails B’s not doing X, or not doing it to the same extent as A, then some metaprinciple is required to adjudicate the conflict between A’s and B’s asserted rights to do X. The consequence of that meta-principle will be to leave A, B, or both with a less than absolute right to do X. Stepping back for the moment from the details of Nozick’s baroque thought experiment, there would seem to be only three places a good (Nozickean) libertarian could go from that starting point. The first is to get the unanimous, explicit agreement of the members of all IPAs to give up their rights to their own IPA in return for the benefits of a minimal state. Nozick rejects this solution as implausible, given that some people may prefer to remain independent, while others may hold out in negotiations for strategic reasons. In addition, even if all are willing to cooperate, such agreements are often prohibitively costly to reach. But these are just the common obstacles that make unanimous consent to collective action unattainable in fact in any large group. The way around the problem that most social contractarians adopt is some version of implied consent. Nozick, however, rejects that solution as not consent at all.¹⁵ It is tantamount, says Nozick, to giving someone a book, and grabbing money from him to pay for it on the grounds he has “nothing better to spend the money on” (and by implication also would have valued having the book at least at its cost). The fact that we are “social products,” says Nozick, “does not create in us a . . . floating debt which the current society can collect and use as it will.”¹⁶ The second alternative is to decide that individuals don’t, after all, have a right to their own IPAs in the SON—that in the SON, might makes right, and moralized Lockean rights arise only after a state, by whatever means, comes into being. This strategy, however, renders Part I of the book irrelevant, and Nozick ostensibly rejects it out of hand.¹⁷ I return to this in Section 1.5 below.

¹⁵ See point (3) above. ¹⁶ Nozick, Anarchy, State, and Utopia, p. 95. ¹⁷ See Nozick, Anarchy, State, and Utopia, p. 118: “Our explanation does not assume or claim that might makes right.”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

134

   ?

The third alternative is to give up on the idea of a minimal state and make one’s peace with anarchy. It is true that doing so would require one to forego the social gains that a minimal state makes possible. But that’s the price you pay for being a strict libertarian with a strong notion of consent. It can (and in collective action situations often will) lead to suboptimal results for everyone. If that bothers you enough to dispense with consent, then you are a consequentialist, not a libertarian—the characterization Nozick inevitably invites in rejecting this third alternative as well. Instead, Nozick opts for a fourth alternative, which is a complicated amalgam of alternative (2) and the implied consent version of alternative (1). His solution takes on board all the objectionable features of both of those alternatives, from the perspective of a strict libertarian, and adds a few of its own. Here are the essential moves.

1.1 Doing Away with Consent In Nozick’s “invisible hand” tale, the minimal state emerges from the anarchic SON when the de facto dominant protective association (DPA) unilaterally forces all nonmembers to give up their IPAs and join it instead. Thus, in place of actual or implied consent, Nozick does away with consent entirely. The door to this solution is opened with surprising insouciance. After asserting, famously, that “[a] line (or hyperplane) circumscribes an area in moral space around an individual,” determined for Lockeans by an individual’s natural rights, Nozick opines that that principle raises the following question: “Are others forbidden to perform actions that transgress the boundary or encroach upon the circumscribed area, or are they permitted to perform such actions provided that they compensate the person whose boundary has been crossed?” Unraveling this question, says Nozick, “will occupy us for much of this chapter.”¹⁸ But no libertarian should need thirty pages to answer that question. Three words suffice: they are forbidden. That’s what consent means. Presumably the IPAs are free to leave the territory now ruled by the DPAturned-ultraminimal-state if they wish. But if they stay put, they will be deemed to have relinquished any independent right of self-protection. To put it another way, there is no right of internal exit for dissenting IPA members. That result is arguably even more objectionable from a libertarian perspective than would be implying consent to pay for benefits, pursuant to something like H. L. A. Hart’s principle of fairness. Hart’s principle covers free-riders: holdouts who are happy to enjoy the benefits of membership but refuse to pay for them. In Nozick’s

¹⁸ Nozick, Anarchy, State, and Utopia, p. 57 (emphasis in original).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



135

ultraminimal state, dissenters don’t want to free-ride on others; they want to remain independent but have been denied that option. In the chapter on “Prohibition, Compensation and Risk,” Nozick tries to reconcile the coercion he licenses here with a strong consent version of libertarianism by arguing that what was taken from IPAs—an absolute right to selfdefense—was never theirs to begin with, because in exercising that right, the IPAs unavoidably jeopardized the rights of others through their unreliable enforcement procedures.¹⁹ This strategy might seem promising at first, but it is hopeless, because it pits identical rights that are not co-possible against each other. I turn to that problem now.

1.2 What Is the Rights Violation Here? The face-off between identical interests that are not co-possible puts any rightsbased argument on an obvious collision course between incompatible intuitions about who is the wronger and who the wronged. When the DPA forbids other IPAs from protecting their own members, who is the (would-be) victim here? The DPA, which is protecting its members from the risky conduct of all the other IPAs (albeit at the same time imposing an identical risk on those other IPAs)? Or the other IPAs, which have lost the right to protect themselves, merely because doing so imposes some risks on others—a loss that arguably is incompatible with a free society?²⁰ Rather than trying to reconcile these warring intuitions in some fashion, Nozick simply embraces them in seriatim. The result is an analytic train wreck, in which Nozick flips between three of the four possible answers to the two questions that have to be resolved in designing legal rules: who gets the right, and how is the right protected? Two principal forms of protection are in play in Nozick’s argument— in legal terminology, a property rule and a liability rule. A property rule gives the rights-holder the entitlement to enjoin others from interfering with the exercise of that right, and the lesser-included right to charge whatever price he can exact for waiving his entitlement. A liability rule, in contrast, allows the non-rights holder

¹⁹ Nozick, Anarchy, State, and Utopia, pp. 108, 114. Nozick is ambiguous as to whether the “unreliability” is procedural (“fair” substantive rules applied in an imperfect fashion) or substantive (rules that are substantively unfair to the members of the other IPAs). He seems to lean toward the former, raising a further question: why should we assume that the procedural imperfections would systematically disadvantage the members of the other IPAs, compared to an IPA’s own members, such that the former is entitled to some sort of special protection? ²⁰ Nozick, Anarchy, State, and Utopia, p. 78: “Since an enormous number of actions do increase risks to others, a society which prohibited such . . . actions would ill fit a picture of a free society as one embodying a presumption in favor of liberty, under which people permissibly could perform actions so long as they didn’t harm others in specified ways.”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

136

   ?

to interfere unilaterally with the exercise of a right, provided that she pays compensation equal to the minimum amount the loser IPAs would have settled for to give up their rights voluntarily (in normal economic parlance, their minimum reservation price).²¹ Nozick initially assigns the entitlement to the IPAs when he assumes that they have a right to defend their members, and protects that entitlement by a property rule (the IPAs may enjoin the DPA from interfering with their risky enforcement activities). He then switches the entitlement to the DPA (to be free from IPAs’ risky enforcement activities), protecting that entitlement by a property rule (the DPA may enjoin the IPAs from interfering with its enforcement activities). Finally, he abruptly shifts the entitlement back to the IPAs, but protects it by a liability rule rather than a property rule: The DPA may enjoin such activities, provided it pays compensation to the IPAs equal to the diminution in IPA members’ welfare as a result of the injunction.²² Again, it is easy to write this mess off to the casually speculative nature of Nozick’s enterprise. But there is a deeper cautionary tale here for all nonconsequentialists. As I argued in previous chapters, the nonconsequentialist literature on harm to others typically assumes that the anti-harm principle is more or less self-executing: it is presumptively violated whenever one party harms another (qualifications to follow). But since the popularization of the Coase Theorem in the 1970s, legal academics have understood that the presence of harm-in-fact resolves nothing. As I discuss in Chapter 1, all disputes between two or more parties involve conflicting desires about how to deploy scarce resources. However we resolve the dispute, one side will be harmed as a matter of fact, in the sense that it will no longer be able to do as it wishes with impunity. As a result, the question of who is wronging whom cannot be resolved by an unmoralized (factual) determination of who is harming whom. It can be resolved only on the basis of normative commitments (implicit or explicit) that lead us to favor one side’s interests over the other’s.²³ In the typical hypotheticals that populate the nonconsequentialist literature, this analytic truth is easy to miss, because the hypotheticals focus on visible and violent forms of harm-in-fact (in Nozick’s words, acts involving “primarily, ²¹ Nozick, Anarchy, State, and Utopia, p. 63. ²² Nozick refers to this solution as the “Principle of Compensation”; Anarchy, State and Utopia, pp. 82–3. As I discuss in Chapter 11, Nozick’s final resting place here is inconsistent with his strong view of property rights once in the minimal state. There, with the exception of extortionate monopoly pricing for necessities, he concludes that owners have the right to retain the scarcity rents built into the market price of their goods or services. For a similar observation, see Eric Mack, “Nozick on Unproductivity: The Unintended Consequences,” in Reading Nozick: Essays on Anarchy, State, and Utopia, edited by Jeffery Paul (New Jersey: Rowman & Littlefield, 1981), pp. 160–90. ²³ On the tendency among contemporary libertarians to tie legal or moral culpability to factual causation of harm, see Mark Kelman, “The Necessary Myth of Objective Causation,” Chicago-Kent Law Review 63, no. 3 (Oct. 1987); Mark Kelman, Review, “Taking ‘Takings’ Seriously: An Essay for Centrists,” California Law Review 74, no. 5 (Oct. 1986).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



137

physical aggressing” against others), which allows the author to traffic in clear and widely shared intuitions about who is wronging whom.²⁴ We all know who will win the dispute between the two-eyed and the blind over forcible eyeball transfers, or the able-bodied and the society that wishes to harvest her organs to save ten lives,²⁵ just as we all know who will win as between A, who wishes to stick her knife into B’s back just for the hell of it, and B, who would prefer that A not do so.²⁶ Because the conclusion is never in doubt in these “slightly hysterical” examples, to borrow Nozick’s apt characterization,²⁷ the reasons adduced in support of it tend to go unscrutinized, in particular the assumption that the mere presence of harm-in-fact to the “victim” resolves the question. In contrast, the conflict that the Nozickean SON sets in motion forces the Coasian problem front and center, because all of the IPAs are engaged in identical conduct (self-defense) that imposes an identical risk of harm to others (inadvertent boundary crossings), and that conduct is of the sort we regard as prima facie socially productive. The noncompossibility of the rights at stake in Part I of ASU follows directly from the fact that all the IPAs are identically situated at the start. Given that symmetry, the only way to get from multiple IPAs to a minimal state without favoring one IPA over another is to endow all IPAs with equal prerogatives, and let them work it out among themselves—by force, in a Hobbesian SON, by explicit agreement in a Lockean one.²⁸ As noted above, Nozick rejects both of these alternatives. Understandably lacking any clear moral intuition about how to resolve this stand-off—whether to regard the imposition of risk as a right or as an incursion on others’ rights, and how to break the tie between identically positioned parties—Nozick just cycles among the various possibilities.

1.3 Measuring Compensation for the Losers In the chapter on “Prohibition, Compensation and Risk,” Nozick considers three possible measures of compensation for the loser IPAs. The first, “market compensation,” mimics the result of a property rule of liability: it is “that price that would have been arrived at had a prior negotiation for permission [to extinguish the loser IPAs’ right of self-defense] taken place.”²⁹ The second, confusingly named “full compensation,” mimics the result of a liability rule: it is the loser ²⁴ Nozick, Anarchy, State, and Utopia, p. 32. ²⁵ Nozick, Anarchy, State, and Utopia, p. 206. ²⁶ Nozick, Anarchy, State, and Utopia, pp. 170, 282. ²⁷ Nozick, Anarchy, State, and Utopia, p. 206. ²⁸ Nozick himself raises this objection but never answers it: “It might be thought that moral considerations require allowing another to do whatever you do; since the situation is symmetrical some symmetrical solution must be found . . . . What moral right does [one party] have to impose this asymmetry, to force others not to behave as he does?” Anarchy, State, and Utopia, p. 125 (emphasis omitted). ²⁹ Nozick, Anarchy, State, and Utopia, p. 65.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

138

   ?

IPAs’ reservation price for giving up their rights to self-defense voluntarily.³⁰ The meaning of the third, “compensation for disadvantages,” is a bit elusive. At one point, Nozick defines it as the diminution in preference-satisfaction suffered by the prohibited individual “as compared to the normal situation.”³¹ “Normal” appears to refer to the situation of others, but it is unclear what others and how we are to calculate the value of their “normal situation.” At various points, Nozick states unambiguously that “compensation for disadvantages” provides less compensation than would full compensation.³² At other points, however, he defines it in a manner that appears indistinguishable from full compensation.³³ I will assume for present purposes that it entails something less than full compensation. For a libertarian, the choice among these three options should be clear. The second-best alternative to requiring the DPA to get the loser IPAs’ consent to extinguish their rights to self-protection would be to make the DPA pay them whatever price the loser IPAs could have extracted in arms’ length negotiations to give up that right—in Nozick’s parlance, market compensation. Indeed, in another context Nozick goes even further, requiring supercompensatory damages (a penalty levied on top of market compensation) for unconsented-to boundary crossings.³⁴ Instead, Nozick relegates the loser IPAs to either full compensation or compensation for disadvantages. As Nozick notes, the effect of full compensation is to give to the DPA all of the joint gains from having one state with a monopoly on power, a result that he himself describes as “clearly unfair.”³⁵ A fortiori, compensation for disadvantages, which leaves the loser IPAs with even less, is unfair as well. Why then choose either? Nozick’s defense is that something less than market compensation is fair when the exercise of one person’s rights imposes a risk of harm on others, because in such cases both sides have some legitimate claim of right: the DPA, to protect itself from the loser IPAs’ risky procedures, the loser IPAs, to protect themselves via procedures that might never result in harm to the DPA.³⁶ But the DPA and the loser IPAs don’t just have offsetting claims of right; they have identical claims of right. Given that, Nozick’s solution seems to add insult to the loser IPAs’ injury. They not only lose the right to the protective agency of their choice; they also get no share in the surplus value generated by the forcible ³⁰ Nozick, Anarchy, State, and Utopia, p. 63. ³¹ Nozick, Anarchy, State, and Utopia, pp. 82–3. ³² See, for example, Nozick, Anarchy, State, and Utopia, pp. 87, 145–6. ³³ Nozick states that compensation should be set at “one extremity of the contract curve,” which he (correctly) equates with what he terms full compensation. Nozick, Anarchy, State, and Utopia, p. 84. ³⁴ Nozick, Anarchy, State, and Utopia, p. 57. ³⁵ Nozick, Anarchy, State, and Utopia, pp. 63–5. ³⁶ Nozick, Anarchy, State, and Utopia, pp. 82, 84, 145–6. What Nozick actually argues is that “compensation for disadvantages” is a fair compromise between full compensation and no compensation. Id Nozick, Anarchy, State, and Utopia, p. 146. But the central challenge to libertarian principles comes not from the decision to give less than full compensation but the decision to give less than market compensation.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



139

creation of a monopoly state, and indeed, under the “compensation for disadvantages” measure, are likely to find themselves in a worse position than they were in the SON. Worse yet, in conceding that both sides have legitimate claims, Nozick threatens to win the battle but lose the war for libertarianism, as that concession exposes the wholesale inability of libertarian principles to delineate “particular rights over particular things.” I return to this point in Section 1.5 below.

1.4 Might Makes Right While Nozick goes to much trouble to present the minimal state as an outcome of a non-rights-violating invisible-hand procedure, in the end, might makes right in Nozick’s SON.³⁷ The hand that produces the outcome is invisible only in the very limited sense that it does not operate through an explicit social choice procedure. Beyond that, it lacks the two key attributes of the Smithian invisible hand: voluntary, private transactions that produce an optimal outcome (here, the minimal state) without any one’s trying to produce it. Instead, it replaces the visible hand of extralegal collective action by each of the IPAs with the visible hand of one dictatorial DPA that conquers and absorbs all the others by force. The fact that might makes right in Nozick’s tale would seem to doom this ostensibly libertarian justification for the minimal state from the start, rendering the rest of the argument in Part I superfluous. At the end of the day, then, the hypothetical invisible-hand process that Nozick invokes to ferry us from anarchy to the minimal state violates virtually every libertarian right. The DPA gains a monopoly on coercive force by forcefully extinguishing every other IPA’s natural right to self-defense. The loser IPAs are relegated to compensation (under the Principle of Compensation) equal at best to their minimal reservation price for ceding their rights to self-defense. If the losers are unhappy with that outcome, they are left with exactly the choice that Nozick denounces in Part II as no choice at all: If you don’t like it, leave the territory; if you stay put, you will be deemed to have consented to the legitimacy of the

³⁷ Nozick, Anarchy, State, and Utopia, pp. 108–9: Although all IPAs may act to protect their members against other IPAs’ unfair procedures by seeking to prohibit the use of those unfair procedures, only the DPA “will be able to do so with impunity,” because the DPA “occup[ies] a unique position by virtue of its power,” and as a consequence “alone is in a position to act solely by its own lights.” Nozick denies that this argument amounts to “might makes right.” Nozick, Anarchy, State, and Utopia, p. 118. But it is hard to see what else it could mean, as the disavowal itself inadvertently goes on to concede: “Our explanation does not assume or claim that might makes right. But might does make enforced prohibitions, even if no one thinks the mighty have a special entitlement to have realized in the world their own view of which prohibitions are correctly enforced”; Nozick, Anarchy, State, and Utopia, p. 118. What could the two sentences possibly mean, other than “I’m not saying might makes right. I’m simply saying might makes might.”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

140

   ?

resulting state. Indeed, the losers in Part I are arguably worse off than the losers in the more-than-minimal redistributive state imagined in Part II. The latter can at least console themselves that they lost pursuant to a majority (or supermajority) vote, in which their vote counted as much as everyone else’s. In contrast, the losers in Part I were vanquished solely by force. And how does Nozick justify this rout of libertarian rights? At first, he asserts that it is the only way to get us to “civil government,” which alone can remedy (in Locke’s words) the “inconveniences of the state of nature”—that is, its suboptimality from a consequentialist perspective.³⁸ He then extends that argument to any case in which it would cost more to obtain consent than it would be worth, and still further to any case in which doing away with consent would be the “efficient” solution because “the transactions costs of reaching a prior agreement are greater, even by a bit, than the costs of the posterior compensation system,” provided that the benefits of the act are “great enough.”³⁹ That position seems vanishingly close to straight utilitarianism.⁴⁰ Indeed, Nozick reveals his true utilitarian colors from the start of Part I, when he declares that one could justify the state by showing it “would arise by a process involving no morally impermissible steps, or would be an improvement if it arose,” and then conditions the first route on being able to show that it too would lead to improvements rather than deterioration.⁴¹ If we are to take Nozick at his word here—and the rest of Part I certainly suggests we should— then the entire historical invisible-hand justification can be dispensed with, in favor of a comparison of the goodness or badness of the end-states of anarchy and the minimal state.⁴² But of course this is exactly the instrumental justification for coercive collective action that Nozick repudiates in Part II, arguing that one may never “compel people who are entitled to their holdings to contribute against their will.”⁴³ ³⁸ Nozick, Anarchy, State, and Utopia, p. 10. ³⁹ Nozick, Anarchy, State, and Utopia, p. 73. ⁴⁰ It is not clear that requiring significant benefits from overriding rights meaningfully differentiates Nozick’s proposal from straight utilitarianism, as circumventing the market for gains that are anticipated to be small is unlikely to be optimal from a welfarist perspective in any event. For another statement casually embracing a utilitarian cost/benefit calculus, see Nozick’s discussion at p. 79 of what forms of pollution to allow: “Presumably [society] should permit those polluting activities whose benefits are greater than their costs,” which should be determined via the Kaldor/ Hicks criterion. ⁴¹ Nozick, Anarchy, State, and Utopia, p. 5 n.* (emphasis added). ⁴² Indeed, one could argue it must be dispensed with, as—to quote the Nozick of Part II—“[c]ontract arguments embody the assumption that anything that emerges from a certain process is just. Upon the force of this fundamental assumption rests the force of a contract argument”; Nozick, Anarchy, State, and Utopia, p. 208. ⁴³ Nozick, Anarchy, State, and Utopia, pp. 238, 263–4, 269. Nozick seems to suggest at the very end of Part II that he might waive consent even for the more-than-minimal state, judging the justness of resulting societies instead by the extent to which they embody and protect the moral side constraints of individual rights. Indeed, Nozick goes even further down the consequentialist road, to suggest that consent is neither necessary nor sufficient to validate the society it produces: A society that protects individual rights but arose unjustly is much preferable, says Nozick, to one that arose

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



141

1.5 What Is Left of Libertarian Rights? Nozick intermittently recognizes that dispensing with consent when there are significant enough welfare gains from doing so leaves libertarian rights in a shambles. But he tries to contain the damage by suggesting that we may jettison rights for utility by paying compensation only when we “are acting in selfprotection in order to increase [our] own security” from potential harm resulting from the actions of others, and those actions “might actually have turned out to be harmless.”⁴⁴ In short, the conduct must “risk crossing another’s boundary” with a probability of less than 1.0.⁴⁵ The unstated assumption here is that deontological principles can handle cases of certainty in either direction. If a given act is certain to result in harm to others, it is wrongful and may be prohibited without paying compensation.⁴⁶ If it is certain not to result in harm to others, it is permissible, and hence any waiver of that right must be negotiated. Only where an action might or might not turn out to be harmful do deontological principles run out. In such cases, says Nozick, “[i]t is difficult to imagine a principled way in which the naturalrights tradition can draw a line” between acceptable and unacceptable risks imposed on others. “This means that it is difficult to see how, in these cases, the natural-rights tradition draws the boundaries it focuses upon.”⁴⁷ This brings us back to the central argument of Part I. The category of conduct that “merely” risks harm to others is not small. It encompasses virtually everything we do, including conduct that we routinely prohibit without requiring compensation to the would-be transgressor (e.g., driving 60 mph through crowded city streets) and conduct that we permit people to engage in with impunity (e.g., driving prudently down a crowded city street, knowing that doing so imposes some irreducible risk of harm to pedestrians), as well as conduct that we permit people to engage in provided that they compensate victims for any harm that results (e.g., so-called strict liability torts). If, as Nozick suggests, he would have us assign risky conduct to one of these three categories based not on the natural rights of the actor but on the social utility of the conduct, he, like the Kantian nonconsequentialists discussed in Part I, has effectively ceded the regulation of most human conduct to utilitarianism.

justly but fails to protect individual rights. Nozick, Anarchy, State, and Utopia, p. 294. In short, the end-states justify the means. ⁴⁴ Nozick, Anarchy, State, and Utopia, p. 114. ⁴⁵ Nozick, Anarchy, State, and Utopia, p. 74. ⁴⁶ For reasons suggested in Section 1.2, articulated libertarian premises cannot resolve hypothetical cases of certain harm either without an assist from other unstated normative premises that generally turn out to be welfarist manqué. ⁴⁷ Nozick, Anarchy, State, and Utopia, p. 75.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

142

   ?

2. Utopia Nozick’s Utopia is a Tieboutian paradise, in which every imaginable sort of community is on offer: ones to satisfy the preferences of “visionaries and crackpots, maniacs and saints, . . . capitalists and communists and participatory democrats,” Nozick and Rawls, you name it.⁴⁸ In this utopia, individual consent operates at the level of choice of which community to join rather than the collective rules governing that community. “[I]n a free society people may contract into various restrictions which the government may not legitimately impose upon them. Though the framework is libertarian and laissez-faire, individual communities within it need not be.”⁴⁹ If Swedenville wants to run itself on social-democratic principles, god bless it. If you don’t like it, don’t move to Swedenville. So, the best of all possible worlds that Nozick can imagine is one in which we are given a choice among a reasonably diverse menu of communities and then told, with respect to each of them, take it or leave it. But to the naïve observer, that arrangement sounds a lot like the formal position of citizens in most parts of the world. If American citizens don’t like the laws adopted by their fellow countrymen by majority rule (indeed, by any means), they are free to leave. If they would prefer a more socialist-democratic alternative, there’s Canada and Western Europe. If they’d prefer something closer to the libertarian ideal of unregulated capitalism, there’s the Cayman Islands.⁵⁰ Nozick of Part II, however, explicitly rejects the Tieboutian justification for permitting a hundred flowers to bloom at the national level: “The minimal state is the most extensive state that can be justified. Any state more extensive violates people’s rights.”⁵¹ Nozick of Part II would allow a very narrow exception when every citizen explicitly consents to a given deviation from the minimal state.⁵² But this exception—if it is even an exception—is a null set, since once a polity exceeds a trivial size, one could never get unanimous consent to any term of governance. Is it possible to reconcile Nozick’s vision of Utopia at the community level in Part III with his vision of the just (minimal) state in Part II? To a libertarian sensibility, the problems with the “let a hundred flowers bloom” solution in Part III ⁴⁸ Nozick, Anarchy, State, and Utopia, p. 316. ⁴⁹ Nozick, Anarchy, State, and Utopia, p. 320 (emphasis omitted). ⁵⁰ One might legitimately question how much that right of exit is worth in a world in which rights of entry are severely limited. But, as suggested below, libertarian principles don’t lend themselves well to a functional interpretation of rights. More importantly, if closed borders are the problem, that suggests a very different solution from the one proposed by the Nozick of Part II: not that we pressure any given country to reconfigure its political arrangements along libertarian (or any other) lines but that we pressure all countries of the world to open their borders. ⁵¹ Nozick, Anarchy, State, and Utopia, p. 149. ⁵² See Nozick, Anarchy, State, and Utopia, p. 293.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



143

are clear, and well-rehearsed in the literature on Tieboutian sorting through jurisdictional exit.⁵³ Even in (ideal) theory, there will never be enough communities on offer for any person to realize all of her preferences about political arrangements. In order to assemble a critical mass of co-citizens, everyone will have to make compromises. In the real world, the choice among political communities is further constrained by a host of non-political considerations (job opportunities, financial constraints, proximity to friends and family) that for many people are determinative. And once they are located in a given community (by choice or by birth), the costs of exit for many people become prohibitively high.⁵⁴ Thus, inevitably, some libertarians will voluntarily opt into (or stay put in) a particular community—say, Swedenville—because the total package of social, economic, and political benefits on offer is better than in any other community (or at least not so much worse that it is worth the costs of exit), but want to opt out of particular Swedenville laws that in their view violate their rights (its high property tax rates, its compulsory education laws, etc.). As noted above, the Nozick of Part II concludes that the just state is morally required to give its citizens the right to opt out of rules with which they disagree at the national level. But faced with the same hypothetical demand at the local level, the Nozick of Part III concludes otherwise. It is enough, he argues, that you are given a diverse range of choices—which choices, incidentally, need not include a libertarian option.⁵⁵ Once you have chosen Swedenville, you are bound by all of its laws, whatever they are. You may leave anytime you wish, but you have no right to stay put and opt out of particular laws you disagree with.⁵⁶ Nor does it matter whether you disagree with the law as a matter of mere preference or because it violates (libertarian) rights: “A nation or protective agency may not compel redistribution between one community and another, yet a community such as a kibbutz may redistribute within itself (or give to another community or outside individuals).”⁵⁷ What explains this radical disjunction in the legitimate reach of collective action at the local and national levels? The most obvious and plausible explanation is that it is easier to move from Massachusetts to New Hampshire to improve one’s tax-and-expenditure package than to move from the US to the Cayman

⁵³ See William A. Fischel, ed., The Tiebout Model at Fifty: Essays in Public Economics in Honor of Wallace Oates (Cambridge, Mass.: Lincoln Institute of Land Policy, 2006). For further discussion of Tiebout’s model, see Chapter 12. ⁵⁴ I return to these matters in Chapters 12 and 13. ⁵⁵ Nozick, Anarchy, State, and Utopia, p. 320. ⁵⁶ See Nozick, Anarchy, State, and Utopia, p. 321: “[F]ounders and members of a small community may, quite properly, refuse to allow anyone to opt out of equal sharing, even though it would be possible to arrange this.” ⁵⁷ Nozick, Anarchy, State, and Utopia, p. 321.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

144

   ?

Islands.⁵⁸ Given that relative ease, we need not be as solicitous of individuals’ preferences about local governance if they stay choose to put. Having so much turn on the practical feasibility of exit is not without its dangers for libertarians. It opens them up to demands more generally to evaluate freedom on substantive rather than formal grounds, demands they rightly perceive would be their undoing. In addition, as I suggest in Chapter 12, even within the more limited sphere of jurisdictional choice, having so much turn on the robustness of exit options plunges libertarians into a normative and empirical morass it is not easy to climb out of. On the other hand, concern with the meaningfulness of choice at least sounds in values (liberty, self-determination) that are congenial to libertarianism. Nozick, however, refuses to distinguish the local and national cases on this basis: Even if almost everyone wished to live in a communist community, so that there weren’t any viable noncommunist communities, no particular community need also . . . allow a resident individually to opt out of their sharing arrangement. The recalcitrant individual has no alternative but to conform. Still, the others do not force him to conform, and his rights are not violated. He has no right that the others cooperate in making his nonconformity feasible.⁵⁹

Instead, he suggests the cases should be distinguished on the following ground. Many people will be offended by the knowledge that their fellow community members have refused to comply with some of the laws of the community. Such offense constitutes a harm, which is itself a rights violation, but only if one has one’s face rubbed in it daily. That will be the case with respect to dissenters with whom one must interact face-to-face on a daily basis, but not dissenters in farflung regions of the country.⁶⁰ The distinction Nozick urges here is an empirical one, and it may well be wrong. In the world of mass media, we often have much more intimate knowledge of events in far-flung places than we could glean from casual contact with our neighbors. But even granting the accuracy of Nozick’s empirical assertion, it is stunning that any libertarian would think that thwarting others’ nosy preferences about how you live your life might constitute a cognizable harm that trumps your right to live as you wish.⁶¹ And should it turn out that Nozick’s ⁵⁸ Nozick also considers and rejects as empirically unproved another possible defense: that nations can more easily accommodate internal exit than communities. Anarchy, State, and Utopia, p. 322. To a true libertarian, however, the ease of accommodating rights should be irrelevant to our obligation to accommodate them. ⁵⁹ Nozick, Anarchy, State, and Utopia, p. 322. ⁶⁰ Nozick, Anarchy, State, and Utopia, p. 322. ⁶¹ Just to remind Nozick what a libertarian would say about this argument: “Others have no right to a say in those decisions which importantly affect them that someone else . . . has a right to make”; Anarchy, State, and Utopia, p. 270.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

145

empirical hunch is wrong, and the offense we take at how others live knows no geographical bounds within a given country, does that mean that our conception of Lockean rights in the minimal state must be reconfigured to accommodate this newly cognizable harm?

3. The Minimal State Reconsidered Up until now, I have implicitly treated the libertarian credentials of Nozick of Part II as impeccable. But in fact, many of the same problems that dog Nozick of Parts I and III are present in Part II as well. The only reason they aren’t more prominent there is that, with a very few exceptions, Nozick never spells out the “particular rights over particular things” implied by libertarian principles, relying instead on (in his words) “placeholders for [those] conventional details.”⁶² As I suggest above, the exceptions tend to be the easy cases: acts of physical aggression in which our universally shared intuitions (and most policy arguments) clearly favor the person physically harmed. Because they are easy cases from any normative perspective, they do not put libertarian (or any other) principles to a serious test. Since ASU was published, numerous commentators have observed that, depending on how the details are filled in, Nozick’s three principles of justice (Justice in Acquisition, Justice in Transfer, and Rectification) could house virtually any distributive scheme of “particular rights over particular things,” from requiring us to give the United States back to the Native Americans, to Rawlsianism, to something approximating Nozickean libertarianism. I want to make a different point here: that whatever scheme ultimately emerges, it cannot be derived straightforwardly from libertarian principles. What Nozick dismisses as merely “conventional details” emerge from thousands of micro decisions about how to balance conflicting but prima facie legitimate interests, and libertarian principles—which do not admit of the need to compromise—cannot tell us how to make them. Nozick acknowledges as much with respect to the Principle of Rectification, which requires us to “balanc[e] the conflicting considerations” that argue in favor of the wrongly deprived party (or their successors in interest) on the one hand, and the settled expectations of third party beneficiaries of that unjust acquisition on the other.⁶³ And, one could argue, he implicitly acknowledges it in the countless questions he asks but never answers throughout Part II. But once one gets beyond the easy cases, equally difficult choices have to be made at almost every juncture in fleshing out Nozick’s other two principles:

⁶² Nozick, Anarchy, State, and Utopia, p. 150.

⁶³ Nozick, Anarchy, State, and Utopia, p. 173.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

146

   ?

Justice in Acquisition and Justice in Transfer. Consider for example income taxation, the one genuinely hard case that Nozick tackles in Part II. Famously equating it with forced labor, Nozick concludes income taxation is clearly impermissible under Justice in Acquisition.⁶⁴ Nozick’s target in making this argument is taxation used to finance the redistribution of wealth. But the argument applies with equal force to taxes used to finance the operations of the minimal state. Nozick of Part I avoids confronting the impermissibility of the latter form of taxation by ignoring the public goods nature (that is, the nonexcludibility) of many of the “protective services” the minimal state would provide. Only if individual residents can be excluded from enjoyment of those services if they refuse to pay for them can such services be financed through voluntary (market) transactions.⁶⁵ And Nozick of Part II ignores how we are to finance the minimal state entirely. If the operations of the minimal state cannot be financed through a voluntary market model, at some point coercive taxation will be required. If coercive taxation violates libertarian rights when the proceeds are used for redistributive purposes, why does it not equally violate them when used to finance the minimal state? There are a number of possible answers (including H. L. A. Hart’s “benefit” principle, which, as noted above, Nozick rejects). But they all require adopting a much more qualified notion of consent than Nozick’s Justice in Transfer gestures towards. Those qualifications, whatever they are, are not going to come from Nozick’s three principles of justice. Finally, the one “particular right over a particular thing” that Nozick deals with in detail in Part II—our right to appropriate things out of the commons—he resolves the same way he derives the minimal state in Part I: by doing away with consent, and requiring only that we compensate those whom we have deprived of “enough and as good” without their consent. As in Part I, the compensation required under this weak reading of the Lockean proviso is Nozick’s ungenerous measure of “full compensation”: they are entitled only to as much as will return them to same place on the indifference curve that they would have occupied in the absence of private property.⁶⁶ As Nozick plausibly concludes, this is functionally equivalent to giving an unfettered right of appropriation without compensation that can be overridden only in extreme cases, since “the baseline for comparison is so low as compared to the productiveness of a society with private appropriation that the question of the Lockean proviso being violated arises only in the case of catastrophe (or a desert-island situation).”⁶⁷ As with the DPA’s unilateral destruction of the IPAs in Part I, all of the surplus value generated by expropriators’ unilaterally converting the commons to private property goes to the expropriators ⁶⁴ Nozick, Anarchy, State, and Utopia, p. 172. ⁶⁵ Nozick, Anarchy, State, and Utopia, pp. 110–15. ⁶⁶ Nozick, Anarchy, State, and Utopia, p. 178. ⁶⁷ Nozick, Anarchy, State, and Utopia, p. 181.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



147

themselves. Nozick’s justification for doing away with consent in both cases is the immense utilitarian benefits to be derived from allowing private appropriation out of the commons.⁶⁸ Again, the point is not to criticize that resolution; it is to insist that it cannot be derived from libertarian premises.

4. Conclusion What are we to make of Nozick’s inconstancy to libertarian principles? There are a number of possibilities here. 1. Nozick is a fair-weather libertarian. When the going gets tough—meaning when it doesn’t give him the answer he wants—he deserts. 2. Nozick isn’t a libertarian. The only distinctly libertarian thing he cares about is blocking income redistribution. All the rest of the trappings are there to justify that result in terms that seem to have broader philosophical interest. 3. Nozick is a libertarian; the concessions he has made in Parts I and III are either minor or unnecessary. This may be true of some of the concessions, but not most. The problems that stymie Nozick are ones that any libertarian has difficulty handling: What background options must someone have to make her consent voluntary? What should we do about collective action problems that cannot be solved by unanimous consent? And—to anticipate the subject of Chapter 10—can we really derive “particular rights over particular things” from broad and vague libertarian principles like self-ownership, the rights to the fruits of one’s labor, the rights to transfer what one owns, etc.? In Nozick’s optimistic view, once we exclude the clear cases where I have an unfettered right to do X and the equally clear cases where I have no right to do X, almost no details remain to be filled in.⁶⁹ But in reality almost everything remains to be filled in. The details fill hundreds of thousands of volumes of civil-law statutes and court decisions in contracts, torts, real property, intellectual property, trusts and estates, civil procedure, corporations, debtor/creditor law, environmental law—the list goes on and on. By and large, the details resolve problems that would arise under a minimal or a maximal state, and that the broad principles of libertarianism cannot resolve. That’s why they are not clear cases. And what is true of law is (as I suggested at the start) true of morality as well, which is faced with resolving the same array of facially plausible but incompatible assertions of rights. Many rights theorists have been more conscientious than Nozick in filling in the details of their proposed schemes. But in the end, none has shown that broad principles such as “we own ourselves and the products of our labor” can generate answers to the everyday problems we actually face, without an illicit assist from ad ⁶⁸ Nozick, Anarchy, State, and Utopia, p. 171.

⁶⁹ Nozick, Anarchy, State, and Utopia, p. 270.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

148

   ?

hoc intuitionism, naked self-interest, or (as I argued in early chapters) welfarism manqué, snuck in through vague concepts like “due care,” “negligence,” or “boundary crossing.” Which brings us to the last possibility, the same one I pressed with respect to other forms of nonconsequentialism in the first Part of the book: 4. Libertarianism is in significant trouble once it gets beyond a relatively limited set of clear-cut cases that everyone agrees on. In my view, this is the overarching lesson to be learned from ASU at almost fifty years’ distance.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

8 Rawls, Risk, and the Maximin Principle How would a rational (self-interested) person choose, lacking almost all information bearing on the probable outcomes of available choices? This is the challenge Rawls created for himself in A Theory of Justice in imposing his thick veil of ignorance on the representative person who is (hypothetically) asked to choose the basic structure of society.¹ She has a general sense of the range of possible consequences of a given choice, but no information concerning the probability that a particular consequence will actually come to pass. Rawls famously argued that choosing from that position of radical uncertainty, a rational person would use a maximin (more precisely, a leximin) decision rule to resolve how to distribute goods among different members of society,² resulting in his famous Difference Principle: inequalities in the distribution of primary goods are permissible only to the extent that the position of the worst-off in society will be improved by allowing such inequalities. In Rawls’s and most readers’ minds, the Difference Principle, along with the strict lexical priority Rawls assigned to basic liberties over other primary goods, is what distinguishes Justice as Fairness most clearly from what Rawls took to be its chief rival—average utilitarianism— and what I would argue is in fact its chief rival: sufficientarianism. If Rawls’s reliance on a leximin rule cannot be justified, then Justice as Fairness loses its most distinctive feature as a theory of justice. Rawls’s use of a leximin strategy has attracted more attention and criticism than any other aspect of A Theory of Justice. He offers two principal justifications for adopting it: risk aversion and uncertainty aversion.³ I take them up in turn.

¹ John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971). ² A maximin rule dictates that we choose whatever alternative will maximize the minimum (worst) outcome of a particular choice. A leximin rule applies a maximin function recursively over a series of lexically ordered choices. As applied to the distribution of resources in society, for example, a maximin rule would dictate that we choose whatever will most improve the situation of the worst off. A leximin rule says, after we have done that, we then maximize the outcome for the second worse-off person, and so on up to the best-off person. I use the terms interchangeably here. ³ He offers a third as well: the demands of reciprocity, publicity, stability, and fraternity. See David Reidy, “John Rawls,” in IVR Encyclopedia of Philosophy of Jurisprudence, Legal Theory and Philosophy of Law (London: Springer, 2008). I take up this justification in Chapter 9. While this might well be more persuasive than risk aversion or uncertainty aversion, it does not rest on a theory of rational choice. It is also subject to the same objection I raise below with respect to risk aversion and uncertainty aversion. At most, it establishes the superiority of a maximin rule over average utilitarianism. It does not establish its superiority over sufficientarianism, a more plausible response to the problems raised. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

150

, ,    

1. Maximin Is the Rational Choice in Expected Value Terms, Under the Unique Circumstances of the Original Position The basic challenge to Rawls’s rational choice justification for a maximin rule was laid out within a few years of its publication.⁴ In brief, if (as Rawls postulates) his representative persons would “assume that they normally prefer more primary social goods rather than less” and would therefore “attempt[] to win for themselves the highest index of primary social goods” possible,⁵ a maximin strategy is rational only on the assumption that people display infinite risk aversion. Rawls repeatedly denied he was resting the case for the Difference Principle on the assumption of infinite risk aversion.⁶ But it is difficult to see how else to interpret comments like the following: “[T]he two principles are those a person would choose for the design of a society in which his enemy is to assign him his place”;⁷ the Difference Principle would be the rational choice for someone who “has a conception of the good such that he cares very little, if anything, for what he might gain above the minimum stipend that he can, in fact, be sure of by following the maximin rule”;⁸ “each must suppose that the marginal utility of these fundamental interests is infinite; this requires anyone in the original position to give them priority and to adopt the two principles of justice”;⁹ and finally, “[t]he standard of utility approaches maximin as risk aversion increases without limit.”¹⁰ Over the years, many critics have argued that Rawls’s assumption of infinite risk aversion is empirically implausible, and that the maximin decision rule it produces yields outcomes almost everyone would regard as unacceptable. Rawls did not dispute either of these claims in general,¹¹ but argued that they did not apply to the unique circumstances posed by the original position, in which (i) the representative person must choose under conditions of radical uncertainty, and (ii) the choice—of the basic structure of a just society—is extraordinarily high stakes. ⁴ David Lyons, “Rawls versus Utilitarianism,” Journal of Philosophy 69, no. 18 (Oct. 1972): pp. 535–45; Kenneth J. Arrow, “Some Ordinalist-Utilitarian Notes on Rawls’s Theory of Justice,” Journal of Philosophy 70, no. 9 (May 1973): pp. 245–63; John C. Harsanyi, “Can the Maximin Principle Serve as a Basis for Morality? A Critique of John Rawls’s Theory,” American Political Science Review 69, no. 2 (Jun. 1975): pp. 594–606. ⁵ John Rawls, A Theory of Justice, Rev. ed. (Cambridge, Mass.: Harvard University Press, 1999), pp. 123, 125. ⁶ See Rawls, A Theory of Justice, Rev. ed., pp. 149, 512–13. ⁷ Rawls, A Theory of Justice, Rev. ed., pp. 132–3. ⁸ Rawls, A Theory of Justice, Rev. ed., p. 134. ⁹ John Rawls, “Some Reasons for the Maximin Criterion,” American Economic Review 64, no. 2 (May 1974): p. 143. ¹⁰ Rawls, “Some Reasons for the Maximin Criterion,” p. 143. Rawls qualified some of these statements in later work, but the Difference Principle itself remains irrational except on the assumption of infinite risk aversion. ¹¹ Rawls, A Theory of Justice, Rev. ed., p. 133.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

        

151

In making this argument, Rawls, following conventional usage, distinguished between risk (taking a gamble with known odds) and uncertainty (taking a gamble with unknown odds). In the case of risk, Rawls accepted that for most people the rational decision rule is some version of expected utility maximization: select the alternative with the highest expected value, calculated by discounting the utility of each possible outcome by its probability and summing the results.¹² The result is average utilitarianism (choose so as to maximize average utility in society). In the original position, however, he argued that “[t]he parties have no basis for determining the probable nature of society or their place in it,” and hence no basis for calculating the expected value to them (or, indeed, anyone) of any particular political arrangement.¹³ John Harsanyi, fleshing out the implications of his essentially identical veil of ignorance in 1955, argued that in the absence of any information about probabilities, the rational strategy would be to treat all possible outcomes as equiprobable (the principle of insufficient reason).¹⁴ Rawls, rejecting Harsanyi’s conclusion, argued that not just the principle of insufficient reason but any form of probabilistic reasoning is irrational when “knowledge of likelihoods is impossible, or at best extremely insecure.”¹⁵ But if all forms of probabilistic reasoning are irrational in the absence of any information about probabilities, how should a rational person choose, and why would he opt for a maximin rule? Rawls offered three arguments for a maximin function as the product of rational choice under conditions of radical uncertainty. First, he argued that the only way to produce a definitive decision rule based on expected values is to assume that people are infinitely risk averse. This is so, because the implications of that preference—choose whatever will minimize the likelihood of the worst possible outcome— are invariant with the distribution of probabilities.¹⁶ In contrast, if the representative person’s degree of risk aversion is anything less than infinite, his rational strategy will depend at least to some extent on the probabilities attached to different outcomes—information that by stipulation he does not have. Rawls’s observation is correct, and explains why it would be computationally convenient to the Rawlsian thought experiment if people turned out to display infinite risk aversion. But it does not establish that they do. ¹² Rawls, A Theory of Justice, Rev. ed., p. 134. ¹³ Rawls, A Theory of Justice, Rev. ed., p. 134. ¹⁴ John C. Harsanyi, “Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility,” Journal of Political Economy 63, no. 4 (Aug. 1955): pp. 309–21. See also Harsanyi, “Can the Maximin Principle Serve as a Basis for Morality?” ¹⁵ Rawls, A Theory of Justice, Rev. ed., p. 134. Rawls is in good company here. The basic critique of the principle of insufficient reason is that it purports to derive knowledge from ignorance, yielding a circular result: If we know nothing about probabilities, what evidence could possibly support the assumption of equiprobability? And if there is evidence to support it, we are not in fact ignorant of probabilities, and Bayesians at least can straightforwardly justify a prior belief in equiprobability based on that evidence. See Alan Hájek. “Interpretations of Probability,” The Stanford Encyclopedia of Philosophy, December 19, 2011, https://plato.stanford.edu/entries/probability-interpret/. ¹⁶ Rawls, A Theory of Justice, Rev. ed., p. 134.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

152

, ,    

His second argument is that “the person choosing has a conception of the good such that he cares very little, if anything, about the minimum stipend that he can, in fact, be sure of by following the maximin rule.”¹⁷ This restates what it means to say that people are infinitely risk averse, but once again does not establish that they are. The third argument is the only one that seeks to defend the empirical truth of the claim that people are infinitely risk averse. Rawls acknowledges the implausibility of that claim in the general case, but argues it is plausible given the “grave risks” posed by the high-stakes choice at issue in the original position.¹⁸ That argument was made easier for Rawls by the fact that in A Theory of Justice, he sought to establish only that his two principles dominated average utilitarianism, which offers no guarantee against truly horrifying outcomes (starvation, enslavement, etc.) for individuals. But as I discuss at greater length in Chapter 9, beginning as early as 1974, Rawls acknowledged that the superiority of a maximin rule would be much harder to defend if it were compared instead to some form of sufficientarianism, in which we guarantee minimal freedoms and a basic income for all but after that revert to some form of average utilitarianism. By the time the revised version of A Theory of Justice was published in 1999, Rawls distanced himself further from the Difference Principle, tilting strongly toward a less draconian form of sufficientarianism.¹⁹ Given the enormous influence that the Difference Principle has had on moral and political philosophy over the past forty-five years and the influence that Rawls’s imprimatur continues to exert on the reception of an argument, it is to be regretted that Rawls’s change of heart and its implications for Justice as Fairness are not more widely known. But there is a deeper puzzle about why and when Rawls’s two principles of justice would imply policy conclusions that diverge from average utilitarianism. As both sides acknowledge, the two approaches will often converge on the same solution.²⁰ The question is, why would they ever diverge? In A Theory of Justice, Rawls recognized that a significant degree of risk aversion is built into the von Neumann/Morgenstern utility function on which Harsanyi relied in calculating expected value.²¹ The incorporation of risk aversion into that function means that, all other things held constant, average utilitarianism, like the Difference Principle, will place more weight on avoiding bad outcomes than achieving good ones. Rawls, ironically in view of his own reliance on the Ellsbergian “principle of uncertainty aversion” (see Section 2), sought to distinguish his thought experiment from Harsanyi’s on the ground that the risk aversion built into the von ¹⁷ Rawls, A Theory of Justice, Rev. ed., p. 134. ¹⁸ Rawls, A Theory of Justice, Rev. ed., p. 134. ¹⁹ Rawls, A Theory of Justice, Rev. ed., p. xiv. ²⁰ Rawls, A Theory of Justice, Rev. ed., p.142; Arrow, “Some Ordinalist-Utilitarian Notes”; Harsanyi, “Can the Maximin Principle Serve as a Basis for Morality?” ²¹ Rawls, A Theory of Justice, Rev. ed., p. 282; Harsanyi, “Can the Maximin Principle Serve as a Basis for Morality?”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

 

153

Neumann/Morgenstern utility function reflects peoples’ psychological attitudes toward risk-taking, which Rawls deemed arbitrary from a moral point of view. Rawls was simply wrong about that. It does not reflect psychological attitudes toward risk-taking; it reflects the weights that individuals assign to different outcomes.²² Since Rawls (at least in his rational choice mode), like preferencebased utilitarians, sought to model the actual degree of risk aversion the representative person in the original position would display, any differences in the conclusions reached by the Rawlsian thought experiment and average utilitarianism should reflect different empirical assumptions about preferences rather than any disagreements in principle. The same is true of the strict lexical priority that Rawls assigns to liberty over fair equality of opportunity in his first principle of justice, and the lexical priority he assigns to the first principle of justice over the second. (As Rawls notes, that lexical ordering is itself an instantiation of the maximin principle, as it forbids any tradeoffs between a given priority and all lesser-ranked priorities.) Any comprehensive utility function would include individuals’ preferences not just over material goods, but over all other conditions that affect their own conception of a good life. That would include the value people attach to religious freedoms, political equality, and the host of other intangible goods contained in Rawls’s conception of liberty and fair equality of opportunity. If, as Rawls posits, people in the original position would choose to give strict lexical priority to protecting liberties over fair equality of opportunity, and both of the foregoing over the Difference Principle, those same preferences should be reflected in a welfarist function. Once again, then, any differences in the weights attached to liberty or fair equality of opportunity in the Rawlsian thought experiment and in average utilitarianism should reflect different empirical assumptions about what people would actually choose, rather than any disagreements in principle.

2. Uncertainty Aversion “Uncertainty aversion” refers to people’s psychological propensity to avoid uncertainty, even at a potentially high cost in expected utility. Uncertainty aversion was first brought to public attention in Daniel Ellsberg’s famous 1961 article, which Rawls relies on in A Theory of Justice to justify the rationality of preferences that would clearly be paradoxical under expected utility theory.²³ Experimental evidence to date about the prevalence of uncertainty ²² Harsanyi, “Can the Maximin Principle Serve as a Basis for Morality?,” p. 600. ²³ Daniel Ellsberg, “Risk, Ambiguity, and the Savage Axioms,” Quarterly Journal of Economics 75, no. 4 (Nov. 1961).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

154

, ,    

aversion has been mixed.²⁴ Susan Hurley has argued that even a modest degree of uncertainty aversion coupled with a Pareto constraint could lead individuals to choose a maximin rule over average utilitarianism in the original position.²⁵ Once again, the harder and more meaningful comparison would be not to average utilitarianism but to a form of sufficientarianism that is less draconian than a maximin rule. Whatever effect uncertainty aversion would have on individuals’ choices in the original position as a matter of fact, it is unclear why it is entitled to normative weight. Unlike risk aversion, which describes the relative utilities individuals attach to different possible outcomes, uncertainty aversion describes a psychological propensity to avoid gambles that is unrelated to the expected outcomes of those gambles. By definition, the psychological discomfort triggered by uncertainty will end the moment the outcome of the gamble is known. From a welfarist perspective, even transitory disutility of this sort counts in calculating expected utilities. Except in extreme cases, however, it won’t count for very much, compared to the utilities attached to the potential outcomes themselves. But it seems doubtful that a transitory psychological state has any normative role to play in deriving principles of justice, let alone the central role that Rawls gives it in defending the rationality of a maximin solution.

²⁴ See Johanna Etner, Meglena Jeleva, and Jean-Marc Tallon, “Decision Theory Under Ambiguity,” Journal of Economic Surveys 26, no. 2 (Apr. 2012); Peter P. Wakker, “Uncertainty Aversion: A Discussion of Critical Issues in Health Economics,” Health Economics 9, no. 3 (Apr. 2000): pp. 261–3; Craig R. Fox and Amos Tversky, “Ambiguity Aversion and Comparative Ignorance,” Quarterly Journal of Economics 110, no. 3 (Aug. 1995): pp. 585–603. ²⁵ S. L. Hurley, “Cognitivism in Political Philosophy,” in Well-Being and Morality: Essays in Honour of James Griffin, edited by Roger Crisp and Brad Hooker (Oxford: Oxford University Press, 2000).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

9 The Unwritten Theory of Justice Rawlsian Liberalism versus Libertarianism

As noted in previous chapters, Rawls originally conceived of A Theory of Justice (TJ) as an alternative to utilitarianism. While other theories of justice—perfectionism, intuitionism, and rational egoism—make cameo appearances in the book, Rawls’s case for Justice as Fairness rests in significant part on its claimed superiority to utilitarianism, in virtue of its refusal to “justif[y] institutions on the grounds that the hardships of some are offset by the greater good of others.”¹ The word “libertarianism” does not appear in the book (although Rawls likely had a Smithian version of libertarianism in mind in his brief discussion of “a system of natural liberty”).² Three years later, Nozick published Anarchy, State, and Utopia³ (hereinafter ASU). Part II of ASU, which lays out Nozick’s Lockean theory of individual rights and the minimal state, is explicitly framed as a rejoinder to Rawls’s TJ, and fully a third of its argument is given over to a close reading and critique of the book, in particular the Difference Principle. The publication of ASU singlehandedly made libertarianism a political philosophy to be reckoned with in academic circles, at precisely the moment it was on the ascendency in political circles. For the general readership of TJ, it also singlehandedly enshrined libertarianism rather than utilitarianism as the chief rival to “justice as fairness,” and put the Difference Principle at the center of Rawlsianism. Almost fifty years later, to most nonspecialists, Rawlsianism is the Difference Principle, and the most durable part of Nozick’s argument has proved to be his critique of that principle. As mismatched as they are in aspirations and method, ASU and TJ have become the towering bookends of political theory, each the other’s chief foil. By the time Rawls was working on the revised edition to TJ, libertarianism was well entrenched in the political scene. Rawls surely knew that to his general readership, Justice as Fairness and the many variants it spawned (equality of resources, equal opportunity for welfare, equal access to advantage, democratic equality) were generally viewed as egalitarians’ most cogent answers to libertarianism. Rawls chose not to reformulate his argument to take cognizance of the

¹ John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971), p. 15. ² Rawls, A Theory of Justice, pp. 72–5. ³ Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974). Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

156

    

changed political and philosophical landscape. As a result, the revised version of TJ, like the original one, is framed as a rejoinder to utilitarianism, making no mention of libertarianism or libertarian principles, beyond the same brief discussion of “systems of natural liberty.”⁴ Thus it comes to be that Rawls’s TJ has had two parallel lives in political theory. The first—the version Rawls wrote—is framed as an alternative to utilitarianism, and in particular utilitarianism’s failure to take seriously the separateness of persons and each individual’s right to pursue her own projects in life. The second—the version “received” by its general audience—is framed as an alternative to libertarianism, and in particular libertarianism’s failure to take seriously our moral obligations to the well-being of our fellow citizens. Notwithstanding the received view of TJ and ASU as the opposing poles of contemporary political philosophy, over the past thirty-five years a small cottage industry has emerged, dedicated to showing how little is required, logically speaking, to turn Rawls into Nozick and Nozick into Rawls.⁵ From one perspective, this development is hardly surprising, given the shared, foundational commitment of Rawlsianism and libertarianism to liberal individualism. As Loren Lomasky put it, “Is it possible to deny the fundamentally libertarian flavor of a theory in which [maximal liberty for all] enjoys lexical priority?”⁶ My aim here is to explore where and why, starting from that common commitment, Rawls’s Justice as Fairness and libertarianism come apart. I focus on their respective treatments of the just distribution of wealth, as this is the point at which the two theories most clearly diverge. Among traditional (right) libertarians, I consider (in addition to Nozick) the work of Loren Lomasky, Jan Narveson, Eric Mack, James Buchanan, Israel Kirzner, Richard Epstein, Gerald Gaus, Randy Barnett, John Simmons, David Schmidtz, John Hasnas, and Murray Rothbard.⁷ I also consider the arguments for “left-libertarianism” offered by Michael Otsuka, Peter Vallentyne, Hillel Steiner, and others.⁸ (Unless otherwise indicated, “libertarianism” refers to right-libertarianism alone.) ⁴ See John Rawls, A Theory of Justice, Rev. ed. (Cambridge, Mass.: Harvard University Press, 1999), pp. 58–8, 62–4. ⁵ See, for example, Loren E. Lomasky, “Libertarianism at Twin Harvard,” in Natural Rights Liberalism from Locke to Nozick, Part I, edited by Ellen Frankel Paul, Fred D. Miller, Jr., and Jeffrey Paul (Cambridge: Cambridge University Press, 2005); John Stick, “Turning Rawls Into Nozick and Back Again,” Northwestern University Law Review 81, no. 3 (Spring 1987): pp. 363–416; James M. Buchanan and Loren E. Lomasky, “The Matrix of Contractarian Justice,” Social Philosophy and Policy 2, no. 1 (Autumn 1984): pp. 12–32; John E. Roemer, “A Challenge to Neo-Lockeanism,” Canadian Journal of Philosophy 18, no. 4 (Dec. 1988): pp. 697–710; Allan Gibbard, “Natural Property Rights,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), pp. 23–30. ⁶ Lomasky, “Libertarianism at Twin Harvard,” p. 180. ⁷ I omit consideration of the virtue-based (or teleological) arguments for libertarianism most associated with Ayn Rand and (among contemporary writers) Douglas Rasmussen and Douglas Den Uyl. ⁸ I take a closer look at left-libertarianism in Chapter 10.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

    

157

In defending their respective theories of justice, both Rawls and libertarians (including ostensibly take-no-prisoners libertarians like Nozick) move back and forth between instrumental arguments—what will make peoples’ lives go better overall / produce a stable society—and arguments from rights-based principles. I take up the former briefly at the end, but for the most part focus on the latter. The reason for this is simple. Disagreements about the instrumental value of a given arrangement of property rights or a given distribution of wealth rest principally on disagreements about the facts. That peoples’ empirical assumptions tend to track their ideological predispositions is a significant (if slightly depressing) fact about human nature. But for my purposes, if that is all that separates Rawls and Nozick, there is no mystery about how to bring them together: change one or both of their empirical assumptions. In contrast, turning Rawls into a principled libertarian or Nozick into a principled Rawlsian egalitarian requires us to understand where, starting from a shared commitment to liberal individualism, they faced the same choices and chose differently. For the most part, we don’t have to guess how libertarians would describe their points of divergence from Rawls. Much of the academic libertarian revival sparked by ASU, like Part II of ASU itself, is framed as a rejoinder to Rawls. From the perspective of libertarians, Rawls has assumed the rhetoric of liberal individualism, but at every juncture has proved himself (in Lomasky’s words) “more committed to his egalitarian redistributionist conclusions than he is to the [liberal individualist] premises that generate those results.”⁹ The upshot, in libertarians’ estimation, is that Rawls is a faux liberal individualist. While Rawls did not address modern-day libertarianism in TJ, he did touch on the subject in other writings. But his comments are relatively brief, and focused exclusively on the Nozickean version, in particular Nozick’s derivation of the minimal state.¹⁰ As a result, for the most part, we have to infer how Rawls would have responded to the substantial libertarian critique of TJ that has amassed over the past forty-five years. Finally, there is the question of method. In comparing Rawls to his libertarian counterparts, should we regard TJ as a procedural (constructivist) theory of justice or a substantive one? Rawls’s answer was, both. The two principles are to be tested by “find[ing] the arguments in their favor that are decisive from the standpoint of the original position,” from which vantage point there are “no given antecedent principles external to [the representative person’s] point of view to which they are bound.” But they are also to be tested “by comparison with our considered judgments of justice,” meaning our substantive ethical commitments.¹¹

⁹ Lomasky, “Libertarianism at Twin Harvard,” p. 199. ¹⁰ See John Rawls, Political Liberalism (New York: Columbia University Press, 1993), pp. 262–5; John Rawls, The Law of Peoples (Cambridge, Mass.: Harvard University Press, 1999): pp. 49–50. ¹¹ Rawls, A Theory of Justice, Rev. ed., p. 132.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

158

    

At this point, I would venture that most people agree with Thomas Nagel’s view, first expressed when TJ was published, that the device of the original position offers no independent support for Rawls’s substantive theory of justice. In Nagel’s words, “The egalitarian liberalism which [Rawls] develops and the conception of the good on which it depends are extremely persuasive, but the original position serves to model rather than justify them.”¹² At various points Rawls himself invites or offers the same assessment.¹³ On the other hand, given the central rhetorical role that the heuristic of rational choice plays both in unfolding Justice as Fairness and establishing Rawls’s bona fides as a liberal individualist, one cannot ignore it. I will therefore consider both the procedural and substantive defenses Rawls offers for his theory of justice. Libertarians have pursued both procedural and substantive arguments for the just state as well, although (with the notable exception of Nozick) generally not in the same work. Most, however, have focused on the latter. As John Simmons observes, contemporary libertarian political philosophers have said relatively little about how a state gains political legitimacy, and with it the right to coerce its citizens.¹⁴ The omission is surprising, because libertarian principles hardly imply indifference to the process by which individuals come under the coercive authority of the state, and most versions of libertarianism (whether they start from self-ownership, autonomy/liberty, or concepts of mutual advantage) would seem to make individual consent a necessary precondition to legitimacy. For anarcho-libertarians like Simmons,¹⁵ Murray Rothbard,¹⁶ and John Hasnas,¹⁷ the explanation is simple. Believing that only explicit consent can justify subjecting an individual to the power of the state and that such consent cannot be obtained, they have concluded that the state cannot be justified. The majority of libertarians, perhaps sharing anarcho-libertarians’ skepticism about consent but unwilling to accept their conclusion, have, sub silentio, given up on consent. Instead, they have argued that the state is legitimized insofar as its political arrangements, however arrived at, are just. As noted above, on one view of TJ—probably the predominant view at this point—this is precisely what Rawls did as well.

¹² Thomas Nagel, “Rawls on Justice,” The Philosophical Review 82, no. 2 (Apr. 1973): p. 233. ¹³ See Rawls, A Theory of Justice, p. 141, Political Liberalism, pp. 25–6, A Theory of Justice, Rev. ed., p. 16. ¹⁴ See A. John Simmons, “Consent Theory for Libertarians,” Social Philosophy and Policy 22, no. 1 (Jan. 2005): pp. 330–56. ¹⁵ Simmons, “Consent Theory for Libertarians,” pp. 330–56. ¹⁶ Murray N. Rothbard, “Robert Nozick and the Immaculate Conception of the State,” Journal of Libertarian Studies 1, no. 1 (1977): pp. 45–57. ¹⁷ John Hasnas, “The Obviousness of Anarchy,” in Anarchism/Minarchism: Is Government Part of a Free Country, edited by Roderick T. Long and Tibor R. Machan (Burlington: Ashgate Publishing Company, 2008), pp. 111–33.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

159

But a small band of libertarians have pursued the classical liberal social contractarian tradition, deriving the legitimacy of the state from hypothetical individual consent to its arrangements pursuant to a properly moralized bargain. Notable among these are David Gauthier,¹⁸ James Buchanan,¹⁹ Jan Narveson,²⁰ and Loren Lomasky.²¹ Nozick is, perhaps ironically, the most famous expositor of both the second and third approach. As I suggested in Chapter 7, the Nozick of Part I of ASU pursues a procedural justification for the state that is essentially contractarian, although not labeled as such. The Nozick of Part II, abandoning procedural justifications for substantive ones, argues that the state is legitimate insofar as it conforms to just (Lockean) rights. In this respect at least, Rawls’s and Nozick’s projects parallel each other. I start by comparing procedural derivations of the just state in Rawlsian and libertarian accounts, and then turn to their respective substantive principles of justice.

1. Constructing the Choice Position Rawls explains the motivation for the original position as follows: On the contract interpretation treating men as ends in themselves implies at the very least treating them in accordance with the principles to which they would consent in an original position of equality. For in this situation men have equal representation as moral persons who regard themselves as ends and the principles they accept will be rationally designed to protect the claims of their person.²²

Most social contractarians of a libertarian stripe would find little to quarrel with in this formulation. The disagreements all come from how Rawls on the one hand and libertarians on the other construct the situation from which such persons must choose, which in turn determines what they perceive to be in their rational self-interest. By “persons,” Rawls insists he means “determinate-persons,” each with her own “place in society,” “class position or social status, . . . fortune in the distribution of natural assets and abilities, . . . intelligence, strength and the like[,] . . . psychological ¹⁸ David Gauthier, Morals by Agreement (Oxford: Oxford University Press, 1986). ¹⁹ James M. Buchanan, The Limits of Liberty (Chicago: Chicago University Press, 1975), “The Ethical Limits of Taxation,” The Scandinavian Journal of Economics 86, no. 2 (Jun. 1984): pp. 102–14. ²⁰ Jan Narveson, The Libertarian Idea (Philadelphia: Temple University Press, 1988). ²¹ Loren E. Lomasky, Persons, Rights, and the Moral Community (New York: Oxford University Press, 1987). ²² Rawls, A Theory of Justice, Rev. ed., pp. 156–7.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

160

    

propensities,” and “conception of the good.”²³ This, more than any other feature of his contractualist thought experiment, is (in Rawls’s view) what differentiates Justice as Fairness from Harsanyi-like contractualist arguments for average utilitarianism, and establishes its superiority. To quote Rawls, “the utilitarian argument assumes that the parties have no definite character or will, that they are not persons with determinate final interests, or a particular conception of their good, that they are concerned to protect.”²⁴ In short, “[t]hey are, we might say, ‘bare-persons,’ ” each with the same chance of “being any one of a number of persons complete with each individual’s system of ends, abilities, and social position” expressed by the “same deep utility function.”²⁵ But having populated the original position with determinate persons motivated by rational self-interest, Rawls famously puts them behind a thick veil of ignorance, which prevents them from knowing what those determinate traits are. Rawls argues that his determinate person behind the veil differs from utilitarians’ bare-person, because the former knows he is different from other individuals, although he doesn’t know how. As a consequence, while (contra utilitarianism) ethical considerations do not operate “in the characterization of the parties” in the original position, they operate in the characterization of the choice situation, by suppressing the information required for parties to favor their own interests over others.”²⁶ But to libertarians (and many others), this is a distinction without a difference. As Nozick famously responded in ASU, turning Rawls on himself, the choosing self behind the veil satisfies Rawls’s ambition (contra utilitarianism) to “‘take seriously the distinction between persons’ . . . only if one presses very hard on the distinction between men and their talents, assets, abilities, and special traits.”²⁷ The resulting “bare deontological ego”²⁸ populating Rawls’s original position may suffice to establish Rawls’s credentials as a moral individualist in the limited sense that, contra communitarian or other group-based accounts of justice, in evaluating institutional arrangements he counts only individuals’ interests. But from the perspective of libertarians, the gulf between that “ bare deontological ego” and the flesh-and-blood differentiated self at the heart of liberal individualism is vast—indeed, unbridgeable. As Lomasky puts it: Everything pertaining to persons except their personhood, whatever that could be when abstracted from ends, character, abilities, and relations to material possessions, is thoroughly socialized . . . . The conclusion is thoroughly illiberal.²⁹

²³ ²⁴ ²⁵ ²⁶ ²⁷ ²⁸ ²⁹

Rawls, A Theory of Justice, Rev. ed., pp. 11, 152, 164. Rawls, A Theory of Justice, Rev. ed., p. 152. Rawls, A Theory of Justice, Rev. ed., pp. 150, 152. Rawls, A Theory of Justice, Rev. ed., pp. 160, 166, 512. Nozick, Anarchy, State, and Utopia, p. 228. This is Lomasky’s term; see Persons, Rights, and the Moral Community, p. 137. Lomasky, Persons, Rights, and the Moral Community, p. 140.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

161

Rawls’s motivations for imposing the veil are, in the view of most libertarians, thoroughly illiberal as well. Rawls defends the veil in (at least) two different ways in TJ. The first is that it prevents distributive shares from being “improperly influenced by . . . factors [that are] arbitrary from a moral point of view,” including natural and social endowments and an individual’s choice of ends.³⁰ At this point, most libertarians conclude that Justice as Fairness is simply not a liberal individualist theory. Nozick, Lomasky, and others have argued that Rawls here conflates two questions: whether people deserve their natural and social endowments; and whether they deserve to keep them and decide for themselves how to exercise them. Even if the answer to the first question is no, Lomasky argues, because such endowments are what make individuals determinate people rather than “barepersons,” “[a] robust liberalism must . . . maintain that contingencies become imbued with moral weight once they are intimately attached to the lives that persons actually live.”³¹ Rawls’s second justification is that without some sort of veil of ignorance, we cannot hope to reach agreement, because “we cannot reasonably expect our views to fall into line when they are affected by the contingencies of our different circumstances.”³² The original position solves that problem by eliminating all such differences, thereby making “the deliberations of any one person . . . typical of all.”³³ Rawls is clear that eliminating the potential for disagreement about moral principles is not just a happy byproduct of his having adopted the veil of ignorance for other reasons. It is, rather, one of the chief motives for adopting it: “Indeed, other things being equal, the preferred description of the initial situation is that which introduces the greatest convergence of opinion.”³⁴ At this point, libertarians (and many others) have concluded that Justice as Fairness is not a contractualist theory either. Starting with the desideratum of unanimous agreement, Rawls, by his own acknowledgement, has reverseengineered the original position to guarantee it. But surely, this puts the cart before the horse in any consent-driven theory of legitimation. If determinate persons, each calculating her rational self-interest with full knowledge of her actual circumstances in life, cannot reach agreement on a just state, then—as the anarcho-libertarians have concluded—so much the worse for agreement. Finally, the consequences of imposing the Rawlsian veil are, from the libertarian vantage point, thoroughly illiberal as well. By making “the deliberations of any one person . . . typical of all,” Rawls turns the problem of social choice into a

³⁰ Rawls, A Theory of Justice, pp. 59–60, 72. ³¹ Lomasky, Persons, Rights, and the Moral Community, p. 140. ³² Rawls, A Theory of Justice, Rev. ed., p. 453. ³³ Rawls, A Theory of Justice, p. 263. ³⁴ Rawls, A Theory of Justice, p. 517. Rawls’s argument here is reminiscent of one of his arguments for the Difference Principle, discussed in Chapter 8: that the only way to produce a definitive decision rule based on expected values is to assume that people are infinitely risk averse.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

162

    

problem of individual choice under conditions of uncertainty. As a result, what is rational for the representative individual to choose coincides with what an impartial observer, weighing each individual’s (by stipulation identical) preferences equally, would choose. What we are left with, in Eric Mack’s words, is a vision of society that reflects collectivist, not individualist, values: “An abiding feature of liberal individualism—and more particularly moral individualism—is the deep-seated rejection of the idea of a shared substantive social end or hierarchy of ends to which all members of society are to be devoted.”³⁵ For their part, by insisting on determinate, situated persons who know who they are, libertarian contractarians face a host of difficult decisions that the Rawlsian veil allows Rawls to sidestep: What material goods, talents, and social opportunities (including opportunities to exit the bargaining table) do we endow the hypothetical bargainers with? If we start from where people happen to be as a matter of social fact, without regard to the justice of how they got there, why should their resulting hypothetical agreement carry any moral weight? And if instead we construct a fictional set of goods, talents, and social opportunities, where do they come from? How do we deal with holdouts (strategic or otherwise)? Which real-life constraints on bargaining should we build into the imagined bargain and which suppress? As I discuss in detail in Chapters 12 and 13, depending on how libertarians resolve each of these issues, it is a relatively simple matter, if not to turn libertarians into Rawlsians, at least to move them quite far in that direction.

2. The Content of Liberty Rawls grounds the substantive argument for his two principles of justice in the Kantian imperative to treat all individuals as ends in themselves, in keeping with their moral status as free and equal beings.³⁶ Once again, most libertarians would find little to quarrel with in that formulation. From that shared sentiment, however, the two camps part company almost immediately over the content Rawls gives to liberty in his first principle and the content he gives to equality in his second. Consider the following quintessentially Rawlsian defense of the lexical priority of liberty: The moral adequacy of a society of project pursuers can be appraised by reference to the protection it affords to individualism: this is almost a universal ³⁵ Eric Mack, “Individualism and Libertarian Rights,” in Contemporary Debates in Political Philosophy, edited by Thomas Christiano and John Christman (London: Wiley-Blackwell, 2009), p. 133. ³⁶ Rawls, A Theory of Justice, Rev. ed., p. 157.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

163

truth. It is difficult to imagine a civil order in which individualism thrives but liberty rights are not a prominent component of that society’s moral grid. Therefore it will not oversimplify much to cast the fundamental moral imperative of a well-ordered society as: Each person is to have an equal right to the most extensive basic liberty compatible with a similar liberty for others.

The author of this passage is not Rawls but Loren Lomasky who, as if to underscore the common intellectual provenance of libertarianism and Rawlsianism (and, one suspects, to tweak Rawlsians), helps himself in conclusion to Rawls’s first principle of justice.³⁷ In fleshing out the first principle, Rawls does achieve some common ground with libertarians, and also most clearly differentiates Justice as Fairness from utilitarianism. In the end, however, that common ground is (in the view of libertarians) severely limited by Rawls’s construction of the basic liberties.

2.1 The Meaning of Liberty The most expeditious way for Rawls to package a social welfarist agenda in liberal individualist terms would have been to construe liberty as “positive liberty,” in Isaiah Berlin’s famous taxonomy. With the notable exception of political liberties, to which I return below, Rawls explicitly declines to do so.³⁸ But having largely rejected substantive equality in favor of formal equality for purposes of the first principle, Rawls builds it into the second principle via “equality of fair opportunity” and the Difference Principle. Thus, concludes Rawls, does “the two-part basic structure allow[] a reconciliation of liberty and equality.”³⁹ It is a nice question whether Rawls’s division of labor between a largely negative conception of liberty and a positive conception of equality gets him to a significantly different conclusion than if he had simply given liberty a positive interpretation to begin with. But, as I discuss below, from the libertarian perspective the net result with respect to economic rights is the same. His positive conception of equality cedes whatever common ground Rawls might have been thought to establish with libertarians by adopting a negative conception of liberty in the first place.

³⁷ Lomasky, Persons, Rights, and the Moral Community, p. 100. ³⁸ See Rawls, A Theory of Justice, Rev. ed., pp. 179, 198, 291; Rawls, The Law of Peoples, p. 49; Rawls, Political Liberalism, pp. lvii, 327; Harry Brighouse, “Political Equality in Justice as Fairness,” Philosophical Studies 86, no. 2 (May 1997): pp. 155–84. ³⁹ Rawls, A Theory of Justice, Rev. ed., p. 179.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

164

    

2.2 What Are the Basic Liberties? Rawls’s list of the “basic liberties,” amended slightly in the revised version of TJ to account for criticisms from H. L. A. Hart and others, includes: [P]olitical liberty (the right to vote and be eligible for public office) and freedom of speech and assembly; liberty of conscience and freedom of thought; freedom of the person, which includes freedom from psychological oppression and physical assault and dismemberment (integrity of the person); the right to hold personal property and freedom from arbitrary arrest and seizure.⁴⁰

As Lomasky notes, “conspicuously absent from this catalogue are economic liberties, including freedom of contract to buy and sell, to employ and be employed, or to accumulate and invest.”⁴¹ Throughout his writings, Rawls remained agnostic about the choice between private and collective ownership of the means of production, arguing that the question must be settled at the (later) legislative stage, when the veil is partially lifted to allow for knowledge of the “historical conditions and the traditions, institutions, and social forces of each country.”⁴² But he expressly rejected the possibility that private property (other than personal property) was one of the basic liberties. Whatever their other disagreements, all libertarians are unified in support of strong private property rights, although they get there by different routes. Lockean libertarians view private property rights (in one’s self, in the external things one acquires through one’s labor or by other just means) as an extension of selfownership and at least coequal in importance.⁴³ Kantian and instrumental libertarians, in contrast, support them because and only insofar as they are valuable in securing other desired ends (the ability to pursue one’s projects in life, mutual advantage defined in more utilitarian terms). Rawls addresses his very brief remarks on libertarianism to the Lockean strain.⁴⁴ But instrumental libertarianism has always been a significant part of libertarian thought. Among contemporary libertarians, it is arguably much more influential than the Lockean tradition represented by Nozick. In an earlier generation, one would have included among the defenders of private property rights on instrumental grounds von Mises, Hayek, Milton Friedman, and sometimes Ayn Rand. More recent proponents include James Buchanan, Israel Kirzner, Richard ⁴⁰ Rawls, A Theory of Justice, Rev. ed., p. 53. ⁴¹ Lomasky, “Libertarianism at Twin Harvard,” p. 180 (footnote omitted). ⁴² Rawls, A Theory of Justice, Rev. ed., pp. xv–xvi; see also Rawls, Political Liberalism, p. 298. ⁴³ See Nozick, Anarchy, State, and Utopia; John Hospers, “The Libertarian Manifesto,” in Morality in Practice, 7th ed., edited by James Sterba (Wadsworth Publishing Company, 2003), pp. 21–7; Rothbard, The Ethics of Liberty, pp. 113–19. ⁴⁴ See Rawls, Political Liberalism, p. 298; Rawls, The Law of Peoples, pp. 49–50.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

165

Epstein, David Gauthier, Gerald Gaus, Chandran Kukathas, Jan Narveson, Randy Barnett, Eric Mack, and Loren Lomasky. The instrumentalist camp also stands in a more interesting relationship to Rawls, who purports to evaluate property rights—along with all other candidate primary goods—by the same instrumental criterion as they: Is it a good without which “persons cannot . . . achieve their essential aims”?⁴⁵ That raises the possibility that all that divides Rawls from most libertarians are their differing empirical assumptions about the instrumental value of property to liberty. If Rawls had been persuaded that a strong regime of private property rights was in fact essential to enable people to pursue their own projects in life, or libertarians were persuaded that it is not, would their disagreements vanish? One assumes not, but why, exactly? The likely answer on Rawls’s side is that, in the end, the priority he puts on political and civil liberties over economic liberties does not depend on the ends that persons in the original position would actually choose for themselves and the instrumental value of different liberties in pursuit of those ends. Rather, it reflects Rawls’s own conception of the good. Rawls equivocated between the two justifications in the original edition of TJ, but in the revised edition unambiguously endorses the latter.⁴⁶ From the libertarian perspective, of course, this gives away the store. As Lomasky puts it, Rawls’s willingness to give priority to his own “higher order” conception of the good cannot be squared with his ostensible commitment to respect equally the “different and indeed incommensurable and irreconcilable conceptions of the good” that individuals hold.⁴⁷ The latter would seem to require that we regard wealth accumulation, just as much as political participation, as a legitimate project in life if that is in fact what individuals want for themselves. The opposite perfectionist notions are no doubt at work sub rosa in libertarians’ ostensibly instrumental justifications for laissez-faire capitalism and strong private property rights. It is possible that a redistributive, social welfare state leads to worse outcomes than laissez-faire capitalism, especially for the least well-off, as Lomasky and other libertarians have argued.⁴⁸ But it is possible it does not, and I think it is a fair surmise that libertarians’ empirical hunches here follow their political predilections and not the other way around.

⁴⁵ John Rawls, “Social Unity and the Primary Goods,” in Utilitarianism and Beyond, edited by Amartya Sen and Bernard Williams (Cambridge: Cambridge University Press, 1982): pp. 172–3; Rawls, Political Liberalism, p. 298. ⁴⁶ Rawls, A Theory of Justice, Rev. ed., p. xiii; see also John Rawls, “Kantian Constructivism in Moral Theory,” Journal of Philosophy 77, no. 9 (Sept. 1980): pp. 525, 527, 547. ⁴⁷ Lomasky, “Libertarianism at Twin Harvard,” pp 183–4, quoting Rawls, Political Liberalism, p. 303. ⁴⁸ See Lomasky, “Libertarianism at Twin Harvard,” p. 191.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

166

    

2.3 The Lexical Priority of Liberty In his famous response to Rawls’s first principle, H. L. A. Hart argued that Rawls’s commitment to the “most extensive total system of liberty compatible with a similar liberty for others” ignores the reality that one person’s exercise of her basic liberties may impinge on another’s basic liberties. This is, of course, the Coasian problem discussed in Chapter 7, posed in the language of liberty rather than freedom from harm. Rawls, responding to this objection in “The Basic Liberties and Their Priority,”⁴⁹ revised the desideratum to a rhetorically more modest “fully adequate scheme” of liberties,⁵⁰ and explicitly acknowledged that liberties must “be limited when they clash with one another,” language that made its way into revised version of TJ.⁵¹ One could argue that both of these qualifications were implicit in Rawls’s original formulation of the “most extensive total system of liberty compatible with a similar liberty for others.” The challenge facing Rawls was to show that such clashes could be resolved by principles that sound in liberal individualism rather than utilitarianism. This problem is hardly unique to Rawls. The burden of the first half of this book is that it inheres in all nonconsequentialist schemes (most definitely including libertarianism). Rawls explicitly deferred the resolution of the problem to the constitutional stage,⁵² and most libertarians have de facto deferred it as well. The omission matters for present purposes because, were both sides forced to explain how we should choose among plausible but incompatible liberty interests, both would likely be driven to some form of welfarist calculus, thereby bringing the two closer not only to utilitarianism but to each other as well. The little that Rawls does say about reconciling clashing liberty interests supports that prediction. Rawls minimizes the problem by arguing that basic liberties are not “infringed when they are merely regulated, as they must be, in order to be combined into one scheme as well as adapted to certain social conditions necessary for their enduring exercise.”⁵³ But the distinction between “infringing” liberties and “regulating” them is a verbal one only: All regulation infringes liberty. And Rawls’s commonsense reconciliation of conflicting liberty interests does not clearly differ from what most people would intuit to be the optimal aggregative solution.⁵⁴

⁴⁹ John Rawls, “The Basic Liberties and Their Priority,” in The Tanner Lectures on Human Values, vol. III, edited by Sterling McMurrin (Salt Lake City, Ut.: University of Utah Press, 1982): pp. 3–87. ⁵⁰ Rawls, “The Basic Liberties and Their Priority,” p. 5. ⁵¹ Rawls, A Theory of Justice, Rev. ed., p. 54. ⁵² Rawls, Political Liberalism, p. 336. ⁵³ Rawls, Political Liberalism, p. 295. ⁵⁴ Rawls’s analysis of speech regulation is a case in point. See Rawls, Political Liberalism, pp. 335–6, 340–56.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

   

167

In the libertarian camp, Nozick’s confused attempt in Part I of ASU to adjudicate between identical warring rights to self-protection in the state of nature provides its own cautionary tale. As I discussed in Chapter 7, in the end Nozick extricates himself by recourse to the decidedly non-libertarian principle that might makes right, itself justified by the decidedly non-libertarian desire to maximize social good by forcing warring factions into one minimal state.

3. The Meaning of Equality The obligation of a just society to lessen economic and social inequalities arises under three different requirements of Justice as Fairness: the obligation under the first principle to equalize the fair value of political liberties, by requiring that all citizens have the same opportunities to hold office, influence elections, etc., irrespective of wealth or social class; the obligation under the second principle to assure fair equality of opportunity, meaning that citizens should have equal chances to acquire the skills on the basis of which merit is assessed; and the Difference Principle, which guarantees the “sufficient all-purpose means to enable all citizens to make intelligent and effective use of their freedoms.”⁵⁵ As many have noted, guaranteeing the fair value of political liberties and fair equality of opportunity potentially entails redistribution as extensive as what is required by the Difference Principle. Given the lexical priority of the first two principles over the third, there may be little work left for the Difference Principle to do in the just Rawlsian state. Notwithstanding that possibility, disagreements among Rawlsians, utilitarians, and libertarians about the just state’s obligations to remedy social and economic inequalities have focused almost exclusively on the Difference Principle. Hence, I will do the same. Rawls’s Difference Principle has provoked more criticism than any other aspect of Justice as Fairness. As I discussed in Chapter 8, the maximin solution built into the Difference Principle is rational only if we assume extreme (indeed, infinite) risk aversion—a highly implausible assumption even in the peculiar circumstances of the original position. The maximin solution, however, is a rational choice from the perspective of those who actually turn out to be worst off ex post. Rawls’s official position, of course, is that the choice of principles is to be made from an ex ante and not an ex post perspective.⁵⁶ But he frequently elides the two, and on occasion unambiguously adopts an ex post perspective, including in a passage that provoked a strong rejoinder from Nozick in ASU. Positing two individuals, A (the more favored) and B (the least favored), Rawls answers the hypothetical complaint of A that “he

⁵⁵ Rawls, The Law of Peoples, p. 49.

⁵⁶ Rawls, A Theory of Justice, Rev. ed., p. 143.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

168

    

is required to have less than he might since his having more would result in some loss to B” by noting that A is still better off than he would have been under a system of noncooperation.⁵⁷ Nozick responds that the same defense could be given for any scheme of cooperation.⁵⁸ The question left unanswered by Rawls, he argues, is why A isn’t justified in holding out for a scheme that gives him a larger share of the benefits than he would receive under the Difference Principle.⁵⁹ But as Nozick also notes in a lengthy footnote, there is a deeper, structural problem with Rawls’s argument: how to make sense of it from the “perspective of the original position.”⁶⁰ In the original position, there would be no A and B; there would be only the representative person with some uncertain chance of being A and some uncertain chance of being B. If that representative person chose the Difference Principle as the arrangement that he judged to be the most advantageous to his future self, given the possibilities of turning out to be A or B, to whom is he now complaining? And when is now? Nozick concludes (rightly in my view) that the only sensible way to construe A’s and B’s complaints, as Rawls presents them, is to imagine that we have stepped out of the original position and are viewing matters ex post through the eyes of A (the big winner) and B (the big loser). From that perspective, B, knowing he is the loser and A the winner, will understandably be unmoved by the statement that he should have less so that A might have more.⁶¹ But if B’s argument from the ex post perspective is decisive for Rawls, his contractualist thought experiment is not merely superfluous; it is misleading. For in that case, the notion of impartiality that motivates the Difference Principle is not that we should stand as equals in choosing the basic structure of society, but rather that we should stand as equals (at least as regards primary goods) at every moment of our lives. If it is a strain to fit the first notion of impartiality into the framework of mutually disinterested rational choice, it is impossible to fit the second. The moral force of B’s argument from the ex post position is that we must share what we have with the least among us, not because it is in our actual selfinterest to do so, and not because, from behind the veil, we would have calculated it to be in our self-interest as a form of insurance against the worst outcomes we might face, but because it is what we owe to others who are unambiguously not ourselves but whom we should nonetheless care about because they are less fortunate than we are. Whether that duty is best realized through a maximin

⁵⁷ Rawls, A Theory of Justice, p. 103. ⁵⁸ Indeed, he helps himself to the argument in Part I of ASU to defend the dominant protection agency’s right to expunge all independent protective agencies, and in Part II to defend an extremely weak version of the Lockean proviso. See Chapters 7 and 10 for further discussion. ⁵⁹ Nozick, Anarchy, State, and Utopia, p. 196. ⁶⁰ Nozick, Anarchy, State, and Utopia, p. 196–7, n.*. ⁶¹ Rawls’s slide between ex ante and ex post perspectives is structurally identical to Scanlon’s slide in his original version of contractualism (see Chapter 4).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  . . .

169

rule or by some other redistributive metric, it expresses a powerful moral perspective. But the perspective it expresses is not that dictated by liberal individualism, conventionally construed. It is the perspective of those who (in Rawls’s words) “agree to share one another’s fate,” as would members of family who “commonly do not wish to gain unless they can do so in ways that further the interests of the rest.”⁶² Libertarians, not surprisingly, have pounced on that statement and similar ones in TJ as evidence that Rawls is a liberal individualist in self-description only. Judged by the commitments that actually motivate him, he reveals himself to be a communitarian. He conceives of society not (as libertarians would insist) as the mere aggregation of whatever individuals comprise it, with no identity or interests of its own, but instead as the highest organism, the “just social union of social unions,” without which we are all “mere fragments.”⁶³ Lomasky, assessing the resulting divide between libertarians and Rawls, concludes that the reason that Rawls does not end up a libertarian is “not because Rawls’s theoretical underpinnings are fundamentally hostile to libertarian perspectives . . . . [T]hey are in fact hardly more than a hair’s breadth away from yielding a recognizably libertarian position.”⁶⁴ It is instead because Rawls is more committed to his egalitarian redistributionist conclusions than he is to the premises that generate those results. Whenever he enters into wide reflective equilibrium, opposition to libertarianism is one of those relatively fixed points unlikely to be dislodged . . . . [H]is continued inability to come to terms successfully with libertarianism is due to an internal tension between his methodology and his convictions. One or the other has to give; invariably, it is former.⁶⁵

But as John Stick notes, the same might be said of the other side, suggesting that the disagreements of Rawls and Nozick are founded upon the political content of their views, and not upon differences in method.⁶⁶

4. And Yet . . . At this point, it is easy to see where Rawls and libertarians come apart and hard to see how to get them to agree on much of anything except by persuading one or the other to change their political commitments. But a number of other considerations put that conclusion in doubt.

⁶² ⁶⁴ ⁶⁵ ⁶⁶

Rawls, A Theory of Justice, p. 102, 105. ⁶³ Rawls, A Theory of Justice, Rev. ed., p. 464. Lomasky, “Libertarianism at Twin Harvard,” p. 199. Lomasky, “Libertarianism at Twin Harvard,” p. 199. Stick, “Turning Rawls into Nozick and Back Again,” p. 387.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

170

    

On Rawls’s side, TJ makes two concessions that hold the real-world implications of Justice as Fairness hostage to human motivation and psychology. The first is Rawls’s concession to the “strains of commitment.” Whatever principles we might otherwise regard as just, Rawls argues, must be tempered in light of the possibility that they will have “consequences [the parties] cannot accept” or can accept “only with great difficulty.”⁶⁷ It is possible to understand this argument as addressed to the rational person choosing from the original position—that is, as arguing it would be rational for that person, choosing principles ex ante, to take into account people’s likely ex post reactions to really bad outcomes. If so, the argument collapses into the argument for the ex ante rationality of extreme risk aversion, with all the problems that attend that argument. But I read Rawls’s strains of commitment argument differently: however rational it might be ex ante for people to take a given gamble, one cannot count on them to take their licks quietly ex post, when things in fact turn out very badly for them. Whatever the moral arguments for and against bailing people out of the adverse consequences of their informed choices, the strains of commitment argument sensibly suggests that we may want to do so as a matter of prudence, to maintain the stability of the social contract over time. Judged from this perspective, the Difference Principle dominates both average utilitarianism and libertarianism, because it protects would-be losers against the worst possible outcomes that the other two theories of justice would permit.⁶⁸ This argument for the Difference Principle surely proves too much, however, because the threat to stability that Rawls is most worried about can be avoided by guaranteeing a basic minimum in the context of utilitarianism, libertarianism, or a sufficientarian version of egalitarianism, without going all the way to the Difference Principle. But more importantly, concerns about the strains of commitment can as easily be turned on Justice as Fairness as its rivals. Nozick led the way here, noting that while Rawls worries about the potential resentment of those who turn out to be worst off towards any distributive principle less generous to them than the Difference Principle, he ignores the potential resentment of those who turn out to be the most fortunate and who feel they are now being bled dry for the sake of the worst off.⁶⁹ As Lomasky suggests, one might also worry that everyone other than the worst off would resent the draconian effects of the maximin rule built into the Difference Principle.⁷⁰ Finally, in the real world, where people are not choosing from behind a veil of ignorance, one would have to worry about the strains of commitment not just for those who want to defect from Justice as Fairness once they see how things turn

⁶⁷ ⁶⁸ ⁶⁹ ⁷⁰

Rawls, A Theory of Justice, Rev. ed., p. 153. Rawls, A Theory of Justice, pp. 176–8, Political Liberalism, p. 17, The Law of Peoples, pp. 49–50. See Nozick, Anarchy, State, and Utopia, p. 196; Rawls, A Theory of Justice, Rev. ed., pp. 470–1. See Lomasky, “Libertarianism at Twin Harvard,” pp. 185–6.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  . . .

171

out for them, but also for those who never accept its normative vision to begin with—for example, those who believe that economic liberties are as important as political liberties, or whose conception of the good life is to have the opportunity to amass great wealth. It’s hard to say what Justice as Fairness would look like once revised to accommodate the strains of commitment that it actually provokes, but one cannot rule out the possibility that it would push Rawls’s two principles substantially in the direction of libertarianism. The second concession Rawls makes to human motivation and psychology is to allow any inequalities that work to the advantage of the worst off. The consequence of this concession, as G. A. Cohen and others have argued, is to hold the distributive consequences of the Difference Principle hostage to the real-world motivations of the wealthy. Pace Cohen, the incentives necessary to induce the well-off to work forty rather than thirty hours a week or shift from painting to doctoring for the benefit of the least well-off do not account for the major inequalities we observe in contemporary American society. But the picture could conceivably change if we took a broader view of labor decisions, to include the sorts of entrepreneurial risks that arguably have contributed the most to raising everyone’s (including the least well-offs’) standard of living in the industrialized world over the last three centuries, while generating unprecedented levels of inequality. If, in the end, libertarians were proved right as an empirical matter about the universal advantages of an unregulated market economy, the Difference Principle would commit Rawls to join in their endorsement of laissez-faire capitalism, and with it a distributive scheme that most libertarians could happily live with. On libertarians’ side, a number of factors could well nudge the programmatic implications of libertarian principles much closer to Rawlsian egalitarianism than libertarians assume. The possibility that has gotten the most attention over the last two centuries is whether and how to restrict individuals from appropriating property from the commons for their private use. As I discuss in Chapter 10, libertarians of different stripes have given widely divergent answers to what falls within the commons, the extent of private rights one acquires through a legitimate act of appropriation, and what it means to leave others a fair share of the commons for their own use. Those divergent answers, in turn, produce widely disparate distributive outcomes, from a strong private property regime with minimal obligations to others at one end, to a distributive scheme at least as egalitarian as Rawls’s and in some cases quite likely more so at the other. A number of other less-discussed factors could well push libertarianism in the direction of Rawlsianism. The first, reflected in Nozick’s Principle of Rectification, is the moral imperative most Lockeans feel to undo the fruits of past injustices. As even Nozick concedes, given the historical record, complying with the Principle of Rectification in the world we actually live in could well have radical implications

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

172

    

(starting with returning extensive portions of the US to Native Americans) that swamp in importance all the other distributive implications of libertarianism. Second, as Lomasky notes, Nozick is an outlier among prominent libertarians in denying any collective obligations to the worst off.⁷¹ Most libertarians would guarantee a basic minimum to those who cannot obtain it for themselves, and many regard the refusal to do so as indefensible. In Lomasky’s words, “To the extent that successful civility requires the provision of aid, welfare rights merit a place alongside liberty rights.”⁷² This concession is enormously significant, both in closing the gap between Rawlsianism and libertarianism and in neutralizing what is, for most people, the most powerful selling point of Justice as Fairness: its guarantee that no one will be left destitute in the name of principle, be it the principle of average utilitarianism or libertarianism. Third, libertarians who reject the moral purity of anarchy for the pragmatic advantages of a coercive state have just begun to confront the compromises with principle that may be necessary to secure those advantages. Having achieved the minimal state by force in Part I of ASU, the Nozick of Part II regains religion on the sanctity of individual consent, famously asserting that compulsory taxation is a form of slavery. But most non-anarchic libertarians recognize that the same collective action problems that necessitate a state in the first place also necessitate giving the state the authority (i) to levy compulsory taxes, at a minimum to fund public goods; (ii) to solve, through compulsory regulation, the problems of externalities that cannot realistically be solved by private contract; and (iii) to operate through majoritarian or supermajoritarian rule, at the cost of minority interests. Pragmatic libertarians have spent much of the last century trying to work out principled limits to each of these necessary concessions. Without prejudging the ultimate success of these efforts, once again I think it not unlikely that at the end of the day, libertarians will find themselves moving much closer to social welfarists on pragmatic grounds. Finally, there is the question of what Rawls himself might have made of his two principles of justice, revisiting them at a half-century’s distance. Rawls left a number of clues that, taken together, suggest he might well have revised them significantly in a direction congenial to libertarians.⁷³ Two potential changes in particular are worth noting. The first concerns individual responsibility. Rawls famously equivocated in TJ about whether individuals’ distributive share should reflect their work effort and other “voluntary” choices made against the backdrop of an otherwise just world. ⁷¹ Lomasky, “Libertarianism at Twin Harvard,” pp. 195–6. ⁷² Lomasky, Persons, Rights and the Moral Community, p. 126. ⁷³ On the evolution of Rawls’s thinking after the original publication of A Theory of Justice, see James Sterba, “Rawls and a Morally Defensible Conception of Justice,” in Morality in Practice, 7th ed., edited by James Sterba (Belmont, Calif.: Wadsworth, 2003), pp. 62–70.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

  . . .

173

The Difference Principle, however, seems to resolve the question in the negative, basing individuals’ distributive shares on the positions individuals find themselves in, without regard to how they got there. (The Difference Principle would of course allow greater return to greater work effort to the extent necessary to induce efforts that redound to the benefit of the least well-off. But the inequalities that result from that concession are forward-looking and instrumental, not a backward-looking measure of desert.) Rawls’s apparent position on this issue is anathema to libertarians, who regard the right to the peaceful enjoyment of the fruits of one’s labor as “the core prescriptive postulate of libertarianism.”⁷⁴ But libertarians are hardly alone in this regard. Beginning with Ronald Dworkin’s choice-sensitive resource egalitarianism, many contemporary liberal egalitarians have distanced themselves from what they take to be Rawls’s position on this point. As G. A. Cohen said of Dworkin’s repudiation of Rawls, it “has, in effect, performed for egalitarianism the considerable service of incorporating within it the most powerful idea in the arsenal of the anti-egalitarian right: the idea of choice and responsibility.”⁷⁵ In the Preface to the revised edition, Rawls suggests that if he were writing TJ anew, he would move much closer to the luck egalitarian position, and stress equal citizenship rather than equality of material goods as the motivation for redistributing wealth.⁷⁶ As Andrew Williams notes, depending on one’s empirical assumptions, the shift to choice-sensitive egalitarianism on the left could well give libertarians most of what they want in terms of outcomes, and would certainly give them the portion of economic rights they most care about in principle.⁷⁷ Second, in writings after the original publication of TJ, Rawls qualified his support for the Difference Principle significantly. Responding to Kenneth Arrow’s⁷⁸ and John Harsanyi’s⁷⁹ critical reviews of TJ, Rawls defended his use of the maximin criterion but conceded that other conceptions of justice might in the end turn out to be more reasonable.⁸⁰ Beginning in 1987 with “The Idea of an Overlapping Consensus,”⁸¹ he suggested that the core moral intuition behind the Difference Principle was a sufficientarian version of egalitarianism that guarantees

⁷⁴ Eric Mack, “Libertarianism,” in The Oxford Handbook of the History of Political Philosophy, edited by George Klosko (Oxford: Oxford University Press, 2011), p. 673. ⁷⁵ G. A. Cohen, “On the Currency of Egalitarian Justice,” Ethics 99, no. 4 (Jul. 1989): p. 933. ⁷⁶ Rawls, A Theory of Justice, Rev. ed., pp. xiv–xv. ⁷⁷ Andrew Williams, “Liberty, Equality and Property,” in The Oxford Handbook of Political Theory, edited by John S. Dryzek, Bonnie Honig, and Anne Phillips (Oxford: Oxford University Press, 2006). ⁷⁸ Kenneth Arrow, “Some Ordinalist-Utilitarian Notes on Rawls’s Theory of Justice,” Journal of Philosophy 70, no. 9 (May 1973): pp. 245–63. ⁷⁹ John C. Harsanyi, “Can the Maximin Principle Serve as a Basis for Morality? A Critique of John Rawls’s Theory,” The American Political Science Review 69, no. 2 (Jun. 1975): pp. 594–606 ⁸⁰ John Rawls, “Some Reasons for the Maximin Criterion,” American Economic Review 64, no. 2 (May 1974): pp. 141–66. ⁸¹ John Rawls, “The Idea of an Overlapping Consensus,” Oxford Journal of Legal Studies 7, no. 1 (Spring 1987): pp. 1–25.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

174

    

“all citizens sufficient material means to make effective use of [their] basic rights.” That goal, he acknowledged, could be achieved through a wide range of schemes, of which the Difference Principle was just one example.⁸² And in the Preface to the revised version of TJ, Rawls acknowledged it might not even be the most plausible example, and gave what was at best a half-hearted defense of its comparative virtues. Indeed, it is not merely half-hearted; as I suggested in Chapter 8, it is also misleading, because Rawls omits the one comparison that most people would think puts the Difference Principle at the greatest relative disadvantage: Rawls’s two principles of justice versus the first principle plus average utilitarianism with a sufficientarian minimum.⁸³ If one imagines that Rawls circa 2020 would in fact have ditched the Difference Principle for some version of choice-sensitive sufficientarianism, it is not out of the question that his disagreements with many libertarians about the just distribution of wealth would reduce to haggling over numbers—that is, over exactly how large a basic minimum is needed to ensure a decent life.

5. Conclusion From the selfish perspective of contemporary readers, it is regrettable that the other TJ—the one framed as an alternative to libertarianism rather than utilitarianism—will remain unwritten. One can only speculate how engaging with libertarianism directly, and later the libertarian critique of TJ, might have moved Rawls to revise Justice as Fairness. But there seem to be two obvious directions in which he could have gone, both foreshadowed above. Whether either of them would have left him with what he hoped to achieve—“a viable systematic doctrine” to rival utilitarianism—is more doubtful. The first (Rawls 1.0) would have been to ditch any perfectionist notion of the good in favor of the “incommensurable and irreconcilable” ends that determinate individuals actually embrace, and to revise the list of primary goods accordingly to include those things that are essential to attain those ends, whatever (within reason) they may be. It seems likely that the resulting principles of justice would have included much broader protections for private property—at least insofar as private property results from individual effort—and would have ditched the Difference Principle entirely, in favor of some guaranteed decent minimum for all. Faced with that revised version of Justice as Fairness, one imagines that most libertarians would concede, however grudgingly, that it is a plausible interpretation of liberal individualism, even if not theirs. ⁸² Rawls, Political Liberalism, pp. 156–7; see also John Rawls, “The Idea of Public Reason Revisited,” The University of Chicago Law Review 64, no. 3 (Summer 1997): p. 774. ⁸³ Rawls, A Theory of Justice, Rev. ed., p. xiv.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



175

The second (Rawls 2.0) would have been to abandon any pretense of seeking terms of cooperation that would be to the mutual advantage of determinate-selves, and to substitute two principles, both unabashedly reflecting Rawls’s substantive notion of the good: strong protections for civil and political liberties, coupled with a broad-based, ongoing redistribution of wealth to the less fortunate in accordance with some social welfarist metric. It seems likely that this revised version of Justice as Fairness would look much like the original, with one notable exception. The strict maximin rule built into the Difference Principle would have to be jettisoned in favor of a more plausible compromise between egalitarian and welfarist ends. It is an open question how different Rawls 2.0 would look from average utilitarianism. But there are a couple of reasons to think it wouldn’t look all that different, once the implications of average utilitarianism are fleshed out. First, while we tend to focus on what utilitarianism implies about the optimal distribution of wealth, wealth is hardly the only—or, in many cases, even most important— determinant of happiness, once a subsistence level of income is reached. Insofar as people share a strong preference for protecting basic political and civil liberties, those preferences would count in any utilitarian calculus as well. Thus, while utilitarians would have different motivations for protecting basic liberties than Rawls did, the actual protections implied by the two regimes might not look all that different. Second, as economists have long argued, assuming that the marginal utility of wealth decreases as a person’s total wealth increases, aggregate utility will be maximized by equalizing wealth, all other things held constant.⁸⁴ But of course all other things are not held constant. In particular, the distributive scheme that is optimal from a utilitarian perspective must balance the welfare gains from redistributing wealth against the welfare losses from the disincentive effects of an implicit 100 percent marginal tax rate on the rich. Rawls’s Difference Principle struck its own version of that tradeoff between equality and efficiency. As with protections afforded political liberties, Rawls’s motivations for the tradeoff are different from utilitarians’: not to maximize aggregate welfare but to maximize the position of the least well-off. But once the Difference Principle is jettisoned for a more plausible compromise between equality and aggregate welfare, it is not clear that the ideal redistributive scheme implied by Rawls 2.0 would look very different from the optimal distribution implied by average utilitarianism. Whether it did or not, Rawls 2.0, unlike TJ itself, would share with utilitarianism something more telling from the libertarian perspective: a frank acknowledgement that the motivation for redistribution is other-regardingness, not mutual disinterest. Whatever libertarians might make of Rawls 2.0 on the merits, one imagines they would regard that acknowledgement as a major victory.

⁸⁴ Barbara H. Fried, The Progressive Assault on Laissez Faire: Robert Hale and the First Law and Economics Movement (Cambridge, Mass.: Harvard University Press, 1998): pp. 152–5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

10 Left-Libertarianism The left-libertarian movement, which crystallized in the early 2000s with the publication of Peter Vallentyne and Hillel Steiner’s two-volume collection of the same name, seeks to harness the normative premises of the libertarian right to the political agenda of the egalitarian left.¹ Its historical antecedents can be found in a long tradition of what one might term left-Lockeanism—a defense of a limited form of capitalism, which entitles people to keep that portion of the value of their product added by their own labor, but no more. Important historical figures in left-Lockeanism include the early agrarian radicals Thomas Paine and William Ogilvie, the liberal socialists of the mid-nineteenth century, and radical land reformers including J. S. Mill, (the early) Herbert Spencer, and, most famously, Henry George. In addition to Vallentyne and Steiner, its contemporary proponents and sympathetic fellow travelers include Philippe Van Parijs, Allan Gibbard, Michael Otsuka, Baruch Brody, and James Grunebaum. Left-libertarianism stakes out a middle ground between the two dominant strains of contemporary nonconsequentialist political philosophy: the conventional libertarianism of people like Nozick on the right and the egalitarianism of people like Rawls, Dworkin, and Sen on the left. Neither the right nor the left in this somewhat oversimplified political topography distinguishes sharply between internal endowments (for which I shall use “talents” as a placeholder) and external resources in their respective versions of distributive justice. The libertarian right permits individuals to assert strong ownership rights over both; the egalitarian left permits them to assert strong ownership rights over neither. Left-libertarians, in effect, split the difference. They side with the libertarian right in embracing a right of self-ownership, albeit a considerably less robust version. But they side with the egalitarian left in holding that individuals have no right to a disproportionate share of the external resources of the world—a position (borrowing further from the right) that they house in Locke’s famous proviso that each individual may appropriate only so much of such resources as leaves others with

¹ The first volume, Peter Vallentyne and Hillel Steiner, eds. The Origins of Left-Libertarianism: An Anthology of Historical Writings (New York: Palgrave, 2000), contains readings from historical antecedents to left-libertarianism. The second volume, Left-Libertarianism and its Critics: The Contemporary Debate (New York: Palgrave, 2000), draws together contemporary readings on leftlibertarianism. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

-

177

“enough, and as good.”² Unlike right-libertarians, however, they define external resources very broadly, and give a strictly egalitarian interpretation to “enough, and as good.” For left-libertarians, that middle way represents the road not taken by libertarianism when it veered off into radical individualism in the late nineteenth century, led (in an abrupt about-face) by Spencer himself.³ This marriage of a narrow conception of self-ownership with a broad conception of external resources subject to an egalitarian sharing rule has led many left-libertarians to quite egalitarian outcomes. Exactly how egalitarian varies dramatically, depending on what is included in the external resources that are subject to an egalitarian sharing rule; how much of the value of those external resources is treated as a commons; and how that commons is to be distributed among the polity. On the narrowest view—a Georgian land tax, with fair compensation paid to existing landowners and the net tax revenues distributed equally among all citizens—left-libertarianism probably implies little change from the existing distribution of resources in developed countries like the United States. On the broadest view, it implies a distributive scheme at least as egalitarian as Rawls’s Difference Principle, and, for those who (unlike Rawls) endorse a global obligation to share equally in external resources, likely more so. Notwithstanding (or perhaps because of) their egalitarian conclusions, leftlibertarians have taken great pains to stress that those conclusions follow not from any attachment to broad-based egalitarianism per se, but instead from their libertarian commitments as “a matter of contingent fact.”⁴ Thus we have a program for distributive justice that ostentatiously declares itself to be, in Baruch Brody’s words, “redistribution without egalitarianism” (or, more precisely, equality without egalitarianism).⁵ It is hardly surprising that redistributive proposals that have surfaced in the current political climate self-consciously fly under a libertarian banner. The last significant strain of left-Lockeanism in British and American political thought rose to prominence in the 1880s through the 1930s, a period in which economic

² John Locke, The Second Treatise of Government, in Two Treatises of Government, edited by Peter Laslett (Cambridge: Cambridge University Press, 1967) (1690), §§ 27, 33. ³ Steiner, Hillel, “Original Rights and Just Redistribution,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 74., vol. II, p. 110. For an account of Spencer’s conversion from left-libertarianism to radical individualism, see Michael W. Taylor, Men Versus the State: Herbert Spencer and Late Victorian Individualism (New York: Oxford University Press, 1992), pp. 246–52; David Wiltshire, The Social and Political Thought of Herbert Spencer (New York: Oxford University Press, 1978), pp. 119–31. ⁴ Michael Otsuka, “Self-Ownership and Equality: A Lockean Reconciliation,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, p. 150. ⁵ Baruch Brody, “Redistribution without Egalitarianism,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, pp. 31–47.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

178

-

laissez faire enjoyed widespread support in both countries.⁶ Our own political times are more hospitable to laissez faire, and less to egalitarianism, than any since that period. In this political climate, strategic motives would surely counsel rewrapping the aims of equality in libertarian garb, and it is not hard to detect such motives at work in left-libertarianism. Yet self-ownership, the “libertarian” part of left-libertarianism, clearly holds a genuine allure for many on the left. The question is, why? G. A. Cohen suggests one possible answer: The embrace of self-ownership reflects adaptive preference formation for the “politically bereaved.”⁷ Faced with a world turning increasingly to the right, many on the left may have been driven to rethink whether there might not be something they can live with in resurgent libertarian premises. For most left-libertarians, however, the allure of self-ownership is clearly heartfelt, as Cohen himself poignantly demonstrates in Self-Ownership, Freedom, and Equality, his Bunyanesque chronicle of his own ultimately victorious struggle to free himself from the grip of the self-ownership thesis. Left-libertarians, all of whom have chosen to remain in the self-ownership fold, are not indifferent to the fear that drove Cohen out of it—that affirming the right of self-ownership may legitimate substantial inequalities in wealth or welfare. But unlike Cohen, they have persuaded themselves that they can fashion a version of self-ownership that will not ineluctably lead to that outcome. The political times being what they are, it seems churlish for any left-leaning social welfarist to resist the invitation to wrap the case for equality in the mantle of individual liberty. But I think it should be resisted nonetheless. I focus on two central problems. First, self-ownership cannot do the work that left-libertarians assign it, any more than it can do the cognate work assigned by the libertarian right. Second, the robust interpretation of the Lockean proviso that left-libertarians embrace to distance themselves from the right threatens to eliminate the distinction between left-libertarianism and more conventional strains of egalitarianism. I end by suggesting that left-libertarians’ choice to justify equality by reference to liberty raises strategic concerns as well.

1. Self-Ownership The locus classicus for the libertarian concept of self-ownership is the famous passage in John Locke’s Second Treatise on Government: “Though the Earth and

⁶ See Barbara H. Fried, The Progressive Assault on Laissez Faire: Robert Hale and the First Law and Economics Movement (Cambridge, Mass.: Harvard University Press, 1998), chs. 1 and 4. ⁷ G. A. Cohen, Self-Ownership, Freedom, and Equality (Cambridge: Cambridge University Press, 1995), pp. 253–7.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

-

179

all inferior Creatures be common to all Men, yet every Man has a Property in his own Person. This no Body has any Right to but himself.”⁸ What concrete conclusions follow from this moral imperative has been subject to endless debate. The conventional right-libertarian view is concisely summarized by Vallentyne, paraphrasing G. A. Cohen: The core idea is that agents own themselves in just the same way that they can have maximal private ownership in a thing. This maximal private ownership is typically taken to include the right to fully manage (to use, and to allow or prohibit others from using); the right to the full income; the right to transfer fully any of these rights through market exchange, inter vivos gift, or bequest; and the right to recover damages if someone violates any of these rights. Redistributive taxation (e.g., of income or wealth) is incompatible with these rights of maximal private ownership.⁹

For current purposes, the critical point is the last: that from the cardinal “principle that each person is the legitimate owner of his own powers,” it follows that redistributive taxation is theft.¹⁰ In the hands of right-libertarians, this strong version of self-ownership, coupled with the erroneous belief that the concept of self-ownership can be cashed out into detailed legal arrangements through logical deduction, has led to conclusions that strike most people as absurd, and for good reason. Take, for example, Samuel Wheeler’s argument that taxation is morally akin to physical violence: No significant moral difference in kind obtains between eliminating my ability to play softball by taking my knees away and eliminating my ability to play the market by taking my money away. Crimes against property are just crimes against persons which tend not to produce immediate sensations of pain. Theft, taxation, and disembowelment are different forms of the same kind of violation of rights.¹¹

Rather than engaging all of the moral difficulties inherent in the assertion that disemboweling someone and levying an 8 percent sales tax on his luxury ⁸ Locke, Second Treatise, § 27, p. 187. ⁹ Peter Vallentyne, “Critical Notice of G.A. Cohen’s ‘Self-Ownership, Freedom and Equality,’ ” Canadian Journal of Philosophy 28, no. 4 (Dec. 1998): p. 611. ¹⁰ Cohen, Self-Ownership, Freedom, and Equality, pp. 216, 222. ¹¹ Samuel Wheeler, “Natural Property Rights as Body Rights,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, p. 242. This is, if anything, the more plausible step in Wheeler’s argument that self-ownership implies that taxation is theft. The first, even less plausible, step seeks to establish that if the right of selfownership exists, it must be absolute, because “no moral justification exists for drawing the line anywhere.” Wheeler, “Natural Property Rights as Body Rights,” vol. II, pp. 241–2.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

180

-

purchases are “the same kind of rights violation,” Wheeler simply relocates these difficulties in the concession that they are “different forms” of that same kind. In this hermetically sealed world of formal analogic reasoning, questions such as, “But might there be some reasons why we would condemn forcibly removing someone’s kidney or sticking a knife in someone’s back that don’t necessarily carry over to the state’s imposing an ad valorem property tax?” are treated as longwinded rhetorical questions to which the only possible answer is no. While left-libertarians whittle down the requirements of “universal maximal self-ownership” substantially from the robust version embraced by the right, most share with the right the conviction that self-ownership implies a “determinate, complete and consistent” set of legal rights, and that one of those rights is to the products of one’s own labor,¹² which in the view of some prohibits taxing labor income except to pay for public goods.¹³

1.1 The Legal Realist (Functional) Reformulation of Property Oliver Wendell Holmes, Jr., prefigured the Legal Realist revolution in a formative 1894 article, which decomposed abstract legal concepts like “the right to compete” and the “duty to keep a contractual promise” into the complicated functional relations they house.¹⁴ In two important articles published in 1913 and 1917, Wesley Hohfeld formalized and extended Holmes’s basic insight, devising a more systematic and precise vocabulary to describe the range of functional relations created by legal rights, especially property rights.¹⁵ By “functional relations,” Hohfeld and other Legal Realists meant the entitlement to act (or a prohibition on acting) in certain concrete ways with respect to property. What are loosely described as “rights,” Hohfeld argued, in fact comprehend a number of distinct legal entitlements conferred on the holders, the scope of which is defined by the correlative legal duties placed on all others (the universe of non-rights holders). The details of Hohfeld’s complicated typology for those reciprocal rights and duties need not detain us here. What is important for present purposes is that Hohfeld’s scheme reoriented the legal analysis of property relations in two significant and lasting respects. ¹² James Grunebaum, “Autonomous Ownership,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, p. 68. ¹³ Otsuka, “Self-Ownership and Equality,” vol. II, p. 155; Steiner, “Original Rights,” vol. II, pp. 75–7. As discussed in Chapter 7, Nozick would not allow an exception even for public goods, but he is an outlier among right-libertarians in this regard. ¹⁴ Oliver Wendell Holmes, Jr., “Privilege, Malice, and Intent,” Harvard Law Review 8, no. 1 (Apr. 1894): p. 1. ¹⁵ Wesley Hohfeld, “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning,” Yale Law Journal 23, no. 1 (Nov. 1913): pp. 16–59; Wesley Hohfeld, “Fundamental Legal Conceptions as Applied in Judicial Reasoning,” Yale Law Journal 26, no. 8 (Jun. 1917): pp. 710–70.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

-

181

First, by deconstructing the supposedly unitary concept of “ownership” into its functional parts, Hohfeld implicitly revealed how intricate and malleable it is. Thus, party A could retain legal title to Greenacre, but give away to B the right to live on the land during A’s lifetime, to C the right to farm the land, to D the right to mine it, to E the right to use it as a right of way to reach E’s property, and to F the right to inherit the estate in its entirety upon A’s death. Picking up where Hohfeld left off, his fellow Realists showed that even if A retained an undivided interest in Greenacre (in legal terminology, a fee simple absolute), her rights to use Greenacre and to prevent others from using it were legally constrained in numerous ways. She could not, for example, use her land for various purposes that might harm her neighbors, or prevent the state from taxing the value of her land to pay for public services or from restricting almost all development for environmental reasons. That one could, as a matter of customary legal practice, be the nominal owner of property and yet be stripped of many, if not all, economically valuable rights in that property doesn’t preclude the possibility that there are some essential rights that ought to come with ownership. But it strongly suggests that identifying and defending them will be no easy matter. Second, and more importantly, by redescribing rights as a network of reciprocal powers and disabilities, Hohfeld showed that when we enlarge one party’s formal powers, we necessarily diminish everyone else’s. Thus, when we say that Scarlett has the absolute right to operate a meat-rendering plant on Tara, what we are saying functionally is that her neighbors are legally powerless to stop her from doing so, even if her actions will make their own property less usable, at the extreme destroying all of its value. Fifty years after Hohfeld, Ronald Coase made essentially the same point in economic terms in his famous article, “The Problem of Social Cost.”¹⁶ All social costs generated by activities, Coase argued, are really the joint costs of conflicting desires in a world of scarce resources. Thus, Scarlett’s use of Tara as a meat-rendering plant is a social problem (that is, imposes costs on society) only because her neighbors don’t want to smell the effects of it and cannot costlessly avoid doing so. Details aside, a seditious message about rights lurks just below the surface of Coase’s and Hohfeld’s analysis. Coase’s morally neutral description of the social problem of (for example) smell pollution as a joint cost of Scarlett’s and her neighbors’ conflicting desires strips away the usual veneer of rights talk plastered over such problems, revealing two aspects usually obscured from view: wherever we put the entitlement, we will necessarily sacrifice one party’s interests for others’, and the traditional conceptual apparatus of nonconsequentialism cannot guide the choice.

¹⁶ R. H. Coase, “The Problem of Social Cost,” Journal of Law and Economics 3 (Oct. 1960): pp. 1–44.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

182

-

The recognition that an individual’s exercise of her own rights may infringe another’s was in the air long before Hohfeld and Coase. It was captured eloquently in Mill’s discussion in On Liberty, to which Holmes’s analysis of the problem was undoubtedly indebted. But, in the context of property relations, the Realists sharpened that analysis by careful attention to the myriad legal disabilities imposed on third parties through the creation of any ownership right. More significantly, they showed that such conflicts were not occasional but universal: all property rights necessarily infringe the liberties of others as all place reciprocal burdens on others, and in a world of scarcity such burdens are often substantial. The lessons of the functionalist revolution in property rights have not been entirely lost on libertarians, right or left. Both camps pay verbal obeisance to the Realist insight, sometimes with a nod to Hohfeld himself, about the divisible nature of ownership, the reciprocal nature of legal rights and legal burdens, and the clash among conflicting claims of self-ownership that it inevitably produces.¹⁷ But when it comes to cashing out “self-ownership” in concrete legal rights, most libertarians (right and left) have ignored those lessons, clinging to the optimistic view that one can derive from the abstract principle of “self-ownership” a detailed regime of unqualified rights over one’s self and one’s product. As James Grunebaum puts it, “[A]ll the rules of autonomous ownership follow logically from the principles of autonomy.”¹⁸, ¹⁹

1.2 Self-Ownership Viewed Through a Functional Lens The basic right-libertarian argument, proceeding from the premise of selfownership to the conclusion that taxation is theft, goes as follows: (1) Self-ownership implies, uncontroversially, that each of us has “full liberal ownership” of our own bodies.²⁰ Full liberal ownership means that each of us has the right “to control that body free of coercive interference.”²¹

¹⁷ For respectful invocations of Hohfeld’s analysis, see Steiner, “Original Rights,” vol. II, pp. 91–2; Hillel Steiner, “Capitalism, Justice, and Equal Starts,” Social Philosophy and Policy 5, no. 1 (Autumn 1987): p. 52; Robert Nozick, “Distributive Justice,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, pp. 178–9; Cohen, Self-Ownership, Freedom, and Equality, p. 222 n. 29; John Christman, “SelfOwnership, Equality, and the Structure of Property Rights,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, p. 361 n. 5; Brody, “Redistribution without Egalitarianism,” vol. II, pp. 32–3. ¹⁸ Grunebaum, “Autonomous Ownership,” vol. II, pp. 68, 70. ¹⁹ For an extensive discussion of the issues raised in this section, see Fried, Progressive Assault, ch. 3. ²⁰ Hugh Breakey, “Property,” Internet Encyclopedia of Philosophy, https://www.iep.utm.edu/propcon/#H3. ²¹ Murray N. Rothbard, For a New Liberty: A Libertarian Manifesto (New York: Collier, 1973): p. 34.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

-

183

(2) That right logically implies the right to control the expenditure of energy and talent housed in that body (i.e., ownership of one’s labor). (3) Ownership of one’s labor logically implies ownership of the fruits of one’s labor.²² Thus, to take Steiner’s example, if someone makes a bench with his labor (which he owns) and lumber and tools (which we presume arguendo that he also owns), he “unproblematical[ly]” must, through some logic akin to the mathematical principle of additivity, own the bench as well.²³ (4) Ownership of the fruits of one’s labor logically implies not just the right to use them or give them away, but also the right to keep whatever one gets from selling them to others. Starting from those same four propositions, left-libertarians have reached radically divergent conclusions, not just from right-libertarians but also from each other, about the concrete rights thereby implied. There is, of course, a short list of acts that all libertarians, left or right, along with pretty much everyone else, agree are prohibited by self-ownership: having one of your eyeballs forcibly extracted to transplant it in the blind; torturing, killing (without justification, presumably—that great verbal equivocation) or enslaving you; or (in Van Parijs’s example of motiveless meddling) requiring you to get the permission of the collectivity to scratch your own nose.²⁴ Once we get beyond these easy cases, there is little agreement on any core rights implied by self-ownership. Some of these disagreements derive from a more fundamental disagreement about what ends self-ownership is meant to vindicate. The literature, right and left, moves (often seamlessly) between the formalist “negative liberty” definition of self-ownership, which is secured by prohibiting others from interfering with one’s self and one’s possessions, and the functional “positive liberty” definition, which is secured by the substantive capacity to effect one’s desires. That ambivalence/ambiguity is perfectly captured by Van Parijs’s name for his own version of left-libertarianism, “real libertarianism.” This, like Grunebaum’s statement that forms of ownership should be judged by how well they promote the “well-being needed for autonomy”²⁵ and Will Kymlicka’s statement that “substantive selfownership” requires an “equal claim to the[] resources and liberties” necessary for each of us to act “according to [our] conception of [our]selves,”²⁶ tilts

²² Murray Rothbard, “Property and Exchange,” in Left-Libertarianism and its Critics, vol. II, p. 219. See also Steiner, “Original Rights,” vol. II, p. 76–7. ²³ Steiner, “Original Rights,” vol. II, p. 77. ²⁴ Cohen, Self-Ownership, Freedom, and Equality, pp. 70, 243–4; Philippe Van Parijs, Real Freedom for All: What (If Anything) Can Justify Capitalism? (Oxford: Oxford University Press, 1995), p. 8. ²⁵ Grunebaum, “Autonomous Ownership,” vol. II, pp. 53, 71. ²⁶ Will Kymlicka, “Property Rights and the Self-Ownership Argument,” in Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, p. 317.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

184

-

strongly towards positive liberty.²⁷ Others, including Steiner and Vallentyne, tilt more to the negative liberty version of self-ownership that dominates libertarianism on the right. But even libertarians who agree on the basic ends a commitment to selfownership is meant to vindicate read out of that commitment vastly different sets of concrete rights. That fact casts further doubt on the strong claim, suggested by Steiner and others, that a particular regime of property rights is logically entailed in the concept of self-ownership. It also raises the concern that leftlibertarians, like those on the right, are pulling some very thick conclusions out of some very thin premises, giving them the latitude to find in self-ownership whatever they are looking for. I want to suggest that it could hardly be otherwise, given the inherent indeterminacy and malleability of the concept of ownership as it is used to flesh out the four propositions on which libertarianism (right and left) is built. Consider first proposition (1): that self-ownership implies full liberal ownership of one’s body, which in turn implies a right “to control [one’s] body free of coercive interference.” What exactly does this mean operationally? Suppose I stand two inches from you and blow smoke in your face. Or suppose I imitate your voice in a commercial, passing myself off as you. Have I coercively interfered with your right to control your body? Examples like these highlight the central Coasean problem embedded in the concept of full liberal ownership. Enlarging your rights over your body in any of the foregoing situations necessarily constricts my rights, often including rights over the use of my own body. Rather than confronting that difficulty head on, libertarians and fellow travelers typically deflect it with examples of intrusions that trade on uncontroversial resolutions of the Coasean problem. As Exhibit A, consider Cohen’s assertion that the principle of “universal maximal self-ownership ensures that my right to use my fist as I please stops at the tip of your nose, because of your rights, under universal maximal selfownership, over your nose.”²⁸ Well, maybe. But even this apparently uncontroversial statement skirts a host of difficulties lurking in the war of fist against nose. Suppose my fist grazes your nose by accident on a crowded street. Or suppose I use my fist to hit you, because you have used yours first to hit me, or because you ²⁷ I read Van Parijs as (deliberately) ambiguous rather than ambivalent. His true allegiance is clearly to the positive liberty version of self-ownership. See, for example, Real Freedom for All, pp. 21–4. ²⁸ Cohen, Self-Ownership, Freedom, and Equality, pp. 214–15. Cohen thinks that the two rights in conflict—autonomy over nose versus autonomy over fist—can be distinguished on the ground that the right over one’s own nose is a right of self-ownership of one’s own body, while the right to move one’s fist in such a way that it lands on someone else’s nose is a right of a different sort. But it is hard to see how Cohen could get that distinction out of self-ownership, other than by reading it in. For Nozick’s and Eric Mack’s versions of the argument, using a knife instead of a fist, see Robert Nozick, Anarchy, State and Utopia (New York: Basic Books, 1974): p. 171; Eric Mack, “Self-Ownership, Marxism, and Egalitarianism (Part II): Challenges to the Self-Ownership Thesis,” Philosophy, Politics and Economics 1, no. 2 (Jun. 2002): pp. 237–76.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

-

185

have verbally threatened to attack me? It may be intuitively obvious to some readers which right of self-ownership should win in these cases. But the concept of self-ownership will not help get them to that intuition. Self-ownership doesn’t fare any better as a generative principle when we turn to the other three propositions: (2) the right to one’s labor, (3) the right to the fruits of one’s labor, including (4) the right to its exchange value. Without any operational content given to full liberal ownership in proposition (1), the argument that it implies propositions (2), (3), and (4) seems driven solely by an empty pun on the word “ownership.” That is, because (1) we “own” our body, (2) we must therefore “own” our labor; and (3) because we “own” our labor, we must therefore “own” the fruits of our labor, (4) which necessarily include whatever we can get for selling them. Even a cursory survey of the practical issues that arise in operationalizing the concepts that we “own our own labor” and “own the fruits of that labor” suggests that propositions (2), (3), and (4) raise substantially different concerns, and that none is logically entailed by the proposition(s) from which it is purportedly deduced. In arguing that people “own their own labor” (proposition 2), libertarians are principally concerned with preserving each person’s right to control his labor choices (that is, whether, how much, and at what job to work). Almost all libertarians (right and left) would agree that this right rules out the state’s “[a]uthoritatively assigning people to jobs.”²⁹ But, aside from the singular cases of military conscription and prison labor, no one is seriously proposing such a policy. The real dispute between left and right is over taxation of labor income—that is, propositions (3) and (4). Libertarians, right and left, face equally serious challenges in fleshing out the concrete rights implied by proposition (3): We own the fruits of our labor.³⁰ But the challenges, as well as the underlying autonomy interests at stake, are different enough to call into question whether propositions (2) and (3) have anything in common. Herman Melville may “own” his own body and mind, in the sense that no one can lock him in a room to force him to write Moby Dick or lock him out to force him not to, and he may own the paper on which he writes it down, in the sense that someone who takes it from him without permission has committed theft. But neither form of ownership sheds light on the myriad questions that arise in defining Melville’s intellectual property rights in the novel he does write. To name just a few: May others, without the author’s permission, publish the text and sell it for profit? The year after it was written? Fifty years after? Two hundred years ²⁹ See, for example, Grunebaum, “Autonomous Ownership,” vol. II, pp. 50–1, suggesting that this is the only form of collective action that his version of self-ownership precludes. ³⁰ In what follows, I focus on the use and exchange value of labor and property, setting to the side disagreements among the left about other rights that right-libertarians generally assume to be entailed by full liberal ownership, including the right to make inter vivos gifts or bequests and the right to recover damages if someone violates any of these rights.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

186

-

after? Lend a copy to a friend? To a hundred friends via the Internet? Write a parody of it, or a sequel to it? Sell T-shirts with a picture of a white whale and the name Moby Dick emblazoned above it? Name their rock band “Moby Dick”? When confronted with these and a host of other questions that must be resolved in assigning intellectual property rights, most people quickly recognize how morally arbitrary many of the answers seem, and how little guidance can be gleaned from the principle that people own the fruits of their labors. But the difficulties are no less in the case of tangible property. They are just obscured by two factors. The first is the existence of a clear physical boundary around the perimeter of such property, which invites observers to think that all use rights can somehow be derived by analogy to physical trespass.³¹ The second is the existence of longstanding conventions about how to resolve conflicting claims with respect to real property, which lead many to confuse the customary with the logically or normatively entailed. This is the point at which left-libertarians have departed most sharply from the right. They have done so not by denying that individuals own the “fruits of their labor,” but by construing “fruits of one’s labor” so narrowly as to render it close to a null set in a complex economy. Vallentyne would include only goods produced without use of any external (scarce) resources.³² Otsuka would narrow the set further, to include only those goods produced autarkicly (that is, without any assistance from others) and for personal consumption (thus not requiring external markets to realize their value). So, in Otsuka’s fanciful example, if you weave your own hair into clothing for your own use, the state may not tax its use value to pay for clothing for bald folks devoid of any talent for weaving.³³ Which brings us to proposition (4): People own the exchange value of their labor, and by extension, any property purchased with that exchange value. This is where the rubber finally meets the road for most libertarians, right and left. Here, right-libertarians are pretty much of one mind: self-ownership means you get to keep it all, subject at most to a modest “benefits” tax to finance the minimal state. They don’t, however agree on why. Most take the exchange value of labor or property to be part of the fruits of one’s labor, and hence protected under proposition (3) (the “taxation is theft” argument). Some (including Nozick when he is in the mood) jump over proposition (3), arguing that taxation of labor income amounts to a form of forced labor (in clear violation of propositions (1) and (2)), because, by commandeering the ³¹ Here are just a few of the allocation decisions that cannot be resolved by reference to the perimeter boundaries of real property: the right to emit air, noise, or light pollution; rights to the airspace over one’s property; water rights; the right to deny others access over it to the oceanfront; commercial zoning restrictions; whether you may convert the property into rental property; safety standards for construction; and of course ad valorem property taxes. ³² Vallentyne, “Critical Notice,” p. 622. ³³ Otsuka, “Self-Ownership and Equality,” vol. II, p. 155; see also Christman, “Self-Ownership,” vol. II, p. 346.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

-

187

fruits of our labor, the state indirectly commandeers the labor itself (the “taxation is slavery” argument).³⁴ As I suggested in Chapter 7, the latter argument seizes on one formal likeness between taxation and slavery, disregarding all the differences in the degree and kind of constraints imposed by the two. Those differences do not merely weaken the analogy; in most peoples’ view, they defeat it. That is, they make taxation not merely a lesser form of slavery, but in fact not “the sort of thing” that slavery is at all.³⁵ All left-libertarians and fellow travelers construe proposition (4) more narrowly, but don’t agree on why, or where to draw the line. Within the “taxation is theft” set of arguments, for those like Otsuka who have already whittled down Proposition (3) (owning the fruits of one’s labor) to almost nothing, nothing is at stake in deciding whether or not it extends to the exchange value of such fruits. Among those who construe proposition (3) more broadly, views about the permissibility of taxation vary widely. Brody, for example, while including in the “fruits of one’s labor” products that result from multiple factor inputs, would protect from taxation only that portion of market value that is generated by the owner’s labor.³⁶ Van Parijs would protect only that portion of labor income that is not attributable to scarcity rents on labor.³⁷ Within the “taxation is slavery” set of arguments, views also vary widely. As noted above, at one extreme Cohen swallows whole the right-libertarian view that taxation of labor income amounts to part-ownership of others, complete with an epigraph from Nozick to that effect, a conclusion that forces him, in his view, to the hard choice of renouncing self-ownership to preserve equality.³⁸ Left-libertarians, having stuck with the self-ownership thesis, not surprisingly ³⁴ Nozick, Robert. Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 172. As I discuss in Chapter 11, right-libertarians have offered two other arguments for the nontaxation of exchange value that are not based on the concept of the creator’s self-ownership. The first is that people are entitled to keep whatever they are paid for their marginal product because it reflects the value they generated for society. The second, espoused by Nozick in his principle of Justice in Transfer, is that the creator’s right to keep the exchange value of whatever she created is based not on her self-ownership, but on the buyer’s. ³⁵ Cohen, Self-Ownership, Freedom, and Equality, p. 231; T. M. Scanlon, “Liberty, Contract, and Contribution,” in Markets and Morals, edited by Gerald Dworkin, Gordon Bermant, and Peter Brown (Washington D.C.: Hemisphere, 1977), p. 66 n. 8. The argument also curiously ignores other distortions that taxation introduces in individuals’ labor decisions and that could with equal force be said to compromise individual autonomy over such decisions. I have in mind the fact that virtually any form of taxation distorts taxpayer choices about how much to work (at the extreme, whether to work at all). A taxpayer who, facing a 70 percent tax on income, opts out of the labor market entirely in favor of untaxed leisure will thereby avoid being forced to work directly or indirectly for the government, and hence avoid Nozick’s particular concern about forced labor. But she will nonetheless have had her life plans altered (to her detriment) by government action in a fashion one would imagine most libertarians would regard as troubling on self-ownership grounds. My point, obviously, is not to argue that rightlibertarians haven’t gone far enough in their hostility to taxation. It is rather to suggest that the ostensible basis for their hostility, reflected in the “taxation is slavery” argument, is ill-thought-out and unlikely to survive scrutiny. ³⁶ Brody, “Redistribution without Egalitarianism,” vol. II, p. 35. ³⁷ Van Parijs, Real Freedom for All. I discuss Van Parijs’s argument in more detail below. ³⁸ Cohen, Self-Ownership, Freedom and Equality, pp. 229, 214, 216.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

188

-

have clustered near the other end, construing the analogy of taxation to slavery much more narrowly. Otsuka, for example, concludes that his version of selfownership permits any tax on the talented, provided only that they, like everyone else, are not “forced, on pain of starvation, to work for the badly off”³⁹—a qualification that, redolent as it is of Marxist exploitation theory, will cause (right) libertarians to suspect that self-ownership has been hijacked by the enemy. That suspicion will not be allayed by the arguments of Grunebaum, Van Parijs, Kymlicka, and others, to the effect that self-ownership doesn’t merely permit progressive taxation; it affirmatively requires it.⁴⁰ Others have taken the analogy of taxation to slavery more seriously, but have limited its application to an ex ante tax on earnings capacity (an endowments tax). Such a tax, they argue, pushes taxpayers into higher-wage work or longer working hours in order to pay their tax bill, thereby raising concerns (in Dworkin’s words) about the “slavery of the talented.”⁴¹ In contrast, an ex post tax on income or wealth, even at confiscatory rates, doesn’t force the talented to produce anything or prohibit them from doing so. “All it does is redistribute what is produced.” Thus, concludes Christman, agents’ self-ownership is not violated by an income or wealth tax, even at confiscatory rates, because they are still “fully able to plan strategies, construct projects, and pursue goals.”⁴² From a functional perspective, however, the distinction between an income or wealth tax, on the one hand, and an endowments tax on the other is much less stark than left-libertarians imagine. Under an income or wealth tax, decisions about what to produce to begin with—that is, how much to work and at what job—are made in light of the (ex post) tax liability that will attach to earnings, and are affected in various ways by that impending liability. For many people—full-time workers who work at jobs that maximize their earning potential—an income or wealth tax will have the identical effect on labor choices as an endowments tax. In other cases, the effects differ significantly. Libertarians may find some of these effects more disturbing than others. But it is attention to precisely these material differences that gets derailed by Christman’s and others’ misleading assertion that an ex post tax on earnings does not “force” anyone to do anything.

³⁹ Otsuka, “Self-Ownership and Equality,” vol. II, p. 161. ⁴⁰ Grunebaum, “Autonomous Ownership,” vol. II, p. 72; Van Parijs, Real Freedom for All, pp. 25, 37. ⁴¹ Ronald Dworkin, “What is Equality? Part 2: Equality of Resources,” Philosophy and Public Affairs 10, no. 4 (Autumn 1981): p. 312. ⁴² Christman, “Self-Ownership,” vol. II, p. 357. See also Van Parijs, Real Freedom for All, pp. 9, 63–4; Eric Rakowski, Equal Justice (Oxford: Oxford University Press, 1991); Peter Vallentyne, “Introduction: Left-Libertarianism—A Primer.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, pp. 9–10. Vallentyne adds the condition that the producer must make at least minimal use of external (scarce) resources—a condition that will virtually always be met in a market economy. “Critical Notice,” p. 622. For further discussion of this argument, see Barbara H. Fried, “As Compared to What? Taxing Brute Luck and Other Second-Best Problems,” Tax Law Review 53, no. 3 (Spring 2000): pp. 377–96.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 

189

More generally, all government interventions in the economy alter the relative attractiveness of different work options and of work versus leisure. To that extent, they distort (coercively interfere with, if you will) labor choices. Which of those distortions one might find unacceptable depends on the autonomy interests one deems most important (e.g., maximizing choice about jobs or work effort, minimizing involuntary contributions to the federal fisc), and how one ranks different constraints on autonomy (e.g., by what people subjectively value, by what seems most critical to self-determination in some perfectionist sense). “Self-ownership” or “ownership of one’s labor” cannot produce answers to these questions, since their operative meaning is ultimately supplied by those answers. All of these efforts to limit the portion of one’s marginal product attributed to self-ownership start from the broad intuition that without society (which generates demand for our product, a mechanism for exchange, co-workers, natural resources and other inputs, education, and the like), almost all of our marginal product would disappear. That intuition seems indisputable as a matter of fact. But it has no logical stopping point, leaving the libertarian half of the leftlibertarian project hostage to how broad an interpretation is given to the left half, housed in Locke’s proviso.

2. The Proviso As I stated at the outset, left-libertarianism has two distinct commitments: to selfownership, and to an egalitarian distribution of natural resources. In Section 1, I focused on the former. I turn now to the latter. Left-libertarians anchor the case for a more egalitarian distribution of natural resources in Locke’s famous proviso: each person may appropriate external resources by mixing his labor with them, provided that “there is enough, and as good [of such resources] left in common for others.”⁴³ Like self-ownership, Locke’s proviso gestures in the direction of an important moral intuition. Like self-ownership, however, it lacks any determinate operational implications, and has been construed so variously as to cast doubt on whether left-libertarianism is really a coherent or distinct philosophical approach. The precise relationship envisioned between the two halves of the left-libertarian project is not always clear. In practice, most left-libertarians appear to interpret the two principles as virtual complements, so that whatever is not an incident of self-ownership becomes a collective asset subject to the strongly egalitarian sharing rule left-libertarians read into the proviso. Given that interpretation and the narrow reading most left-libertarians have given to self-ownership, it is

⁴³ Locke, Second Treatise, § 27.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

190

-

hardly surprising that most read the proviso quite broadly, at the extreme obliterating the libertarian half of the project housed in self-ownership. Hence the doubt expressed earlier: whether the robust interpretation leftlibertarians give to the proviso can be distinguished functionally from more familiar strains of liberal egalitarianism. Producing a reading of the proviso that is distinct requires, at a minimum, that proponents (a) identify a discrete set of scarce natural resources subject to the proviso that stops somewhere short of the more comprehensive list of assets that traditional egalitarians think must be equalized; and (b) devise a redistributive scheme correcting only inequalities in the distribution of those scarce resources, not inequality more generally. On both scores, the record is unlikely to allay suspicions that left-libertarianism is just liberal egalitarianism in drag.

2.1 What “Natural” Resources are Subject to the Egalitarian Strictures of Locke’s Proviso? What the proviso covers depends upon the answer to two questions: what resources count as “natural,” and how do we define “enough, and as good”? The latter is not an issue for left-libertarians, who (correctly in my view) treat any scarce resource as definitionally flunking that requirement, a matter I return to in Chapter 11. It is an issue for right-libertarians, but this too I defer until Chapter 11. Locke himself was preoccupied with land and the natural bounty it holds (wild animals, plant life, and the like). Well into the nineteenth century, that preoccupation was shared among reformers on the left, including J. S. Mill, (the early) Herbert Spencer, and, most notably, Henry George, whose proposal for a “single tax” on land rents is the most famous example of left-Lockeanism in action.⁴⁴ Contemporary right-libertarians have continued this single-minded focus on land and other tangible natural resources, with some (like Israel Kirzner) reducing the scope of Locke’s proviso even further. But most left-libertarians have signed on to a much broader view of nature, and with it a much broader view of the state’s obligations to equalize the fortunes of its citizens pursuant to the proviso. At the minimalist end, Hillel Steiner (in the early versions of his argument) and Nicolas Tideman adopt the traditional Georgian view that the proviso covers only the physical resources of the natural world, along with the traditional Georgian solution narrowly tailored to remedy the perceived wrong: a natural

⁴⁴ For further discussion of the land rents version of left-Lockeanism in the nineteenth century and its gradual metamorphosis into a broader rent-theory Lockeanism by the end of the century, see Fried, Progressive Assault, ch. 4.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 

191

resources tax that expropriates the market value of land and other (unimproved) natural resources.⁴⁵ At the other extreme is Van Parijs’s suggestion that the external resources subject to equitable division include “the whole set of external means that affect people’s capacity to pursue their conceptions of the good life, irrespective of whether they are natural or produced.”⁴⁶ Among the external means proposed for inclusion are intergenerational gifts and bequests; all traditional public goods (laws, police force, public works); the community’s cultural heritage (language, art, science); the country’s physical productive capacity; and well-functioning markets.⁴⁷ Hillel Steiner goes even further in his later work, suggesting that external assets subject to the proviso should include the “germ-line genetic information” contained in all living things, a view that, at a minimum, seems to turn all talents into a collective asset.⁴⁸ Van Parijs suggests the same with respect to effort, arguing that a culturally transmitted work ethic should be included among the “external means that affect people’s capacity to pursue their conceptions of the good life.”⁴⁹ At that point, when talents and propensity for hard work are both transported out of the self-ownership camp and into collective resources subject to Locke’s proviso, it is hard to see what remains for the self-ownership half of left-libertarianism to operate on. Most left-libertarians stop short of that point, explicitly excepting talents and effort from the proviso. But as noted in the first part of the chapter, many take away with the left hand what they have given (the better-endowed) with the right, arguing that if individuals use any resources subject to the proviso to exploit their talents, the state is justified in taxing all or a portion of the return to those talents and effort. Christman, for example, picking up on an argument with a venerable history on the left, suggests that because society is the but-for cause of all market incomes, it may treat them as a collective asset subject to the proviso—the complement of his view that self-ownership protects only use value.⁵⁰ Vallentyne,

⁴⁵ Steiner, “Original Rights,” vol. II, p. 108; Nicolas Tideman, “Commons and Commonwealths: A New Framework for the Justification of Territorial Claims” and “The Shape of a World Inspired by Henry George,” cited in Peter Vallentyne and Hillel Steiner, eds. “Introduction,” The Origins of LeftLibertarianism: An Anthology of Historical Writings (New York: Palgrave, 2000), vol. I, p. 8. Unlike Henry George, Steiner read into the proviso global obligations, leading him to argue that the revenue generated should be contributed to a fund for global equality. ⁴⁶ Van Parijs, Real Freedom for All, p. 101. ⁴⁷ Bruce Ackerman and Anne Alstott, The Stakeholder Society (New Haven: Yale University Press, 1999), pp. 13–14, 32; Van Parijs, Real Freedom for All, p. 101; Rolf Sartorius, “Persons and Property,” in Utility and Rights, edited by R. G. Frey (Minneapolis: University of Minnesota Press, 1984), p. 211. ⁴⁸ Steiner, “Original Rights,” vol. II, p. 105. ⁴⁹ Van Parijs, Real Freedom for All, p. 101. ⁵⁰ Christman, “Self-Ownership,” vol. II, pp. 358. See also Ackerman and Alstott, The Stakeholder Society, pp. 13–14, 32; Kenneth Arrow, “Nozick’s Entitlement Theory of Justice,” Philosophia 7, no. 2 (June 1978): pp. 265–80. For similar arguments from an earlier generation of political theorists on the left, see Fried, Progressive Assault, chs 1 and 4.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

192

-

Gibbard, and others reach the same conclusion by the somewhat different route of bargaining theory.⁵¹ Others have stopped somewhere short of collectivizing all market incomes, seeking a more exacting separation of individual and social contributions to exchange value. One suggestion is that we should collectivize only that portion of market price that reflects factor rents—that is, the return to labor or capital attributable to the scarcity of such factors rather than the costs of supplying them. The rationale here, also with a venerable history on the left, is that factor rents are a purely social product, because they result from the existence of a market in which demand exceeds available supply at constant costs.⁵² In the contemporary left-libertarian literature, Van Parijs develops this line of argument in a particularly interesting direction. Notwithstanding the very broad definition of assets he suggests might be included in external assets subject to the proviso, he suggests taxing only one component: the employment rents that result in “a non-Walrasian economy”—that is, an economy in which, for a variety of reasons, labor markets don’t clear, resulting in some equally skilled people having jobs while others are unemployed, or some holding more desirable jobs at a nonmarket-clearing wage.⁵³ Instead of directly taxing away such rents, Van Parijs proposes approximating the distributive result of such a tax scheme with the “grosser, . . . . unoriginal[]” solution of an optimal progressive income tax, pursuant to which we would tax “wages . . . up to the point at which the tax yield, and hence the basic income financed by it, is maximized.”⁵⁴ That solution, effectively operationalizing the Rawlsian Difference Principle on the tax side, is the same one

⁵¹ Vallentyne, “Introduction,” vol. II, p. 9; Allan Gibbard, “Natural Property Rights,” in LeftLibertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. II, pp. 24–8. In brief, both authors suggest that society may appropriate through taxation an amount equal to what it could have charged private parties for the use of natural resources, were those resources collectively owned. Given the state’s monopoly position in such a hypothetical market, it could charge up to the full benefit private parties derive from the use of such resources. For the vast majority of talents, which are valueless without such resources, that benefit is equal to 100 percent of the market value of those talents. The resulting hypothetical bargain allows the state to price discriminate against the talented, charging them more than the poor for access to the same quantity of natural resources simply because, by virtue of their talents, they will derive greater value from its use. Such price discrimination seems a strained interpretation of the “equal shares” notion of equality expressed in the proviso. The motivation for it, like the motivation for Christman’s argument, seems to lie instead in some more sweeping egalitarianism. ⁵² Fried, Progressive Assault, ch. 4. I return to this argument in Chapter 11. ⁵³ Van Parijs, Real Freedom for All, p. 141. ⁵⁴ Van Parijs, Real Freedom for All, pp. 115–16. The precise measurement of employment rents, Van Parijs suggests, would equal the difference between actual wages and the (lower) amount each worker would receive were the market for their jobs to clear. That amount, in turn, should equal the difference between current earnings and expected replacement income during the year following termination of current employment. Van Parijs, Real Freedom for All, pp. 108, 264, n. 35. For similar proposals, using a similar measure of employment rents, see Juliet B. Schor and Samuel Bowles, “Employment Rents and the Incidence of Strikes,” Review of Economics and Statistics 69, no. 4 (Nov. 1987): pp. 584–92; Bert Hamminga, “Opstaan voor iemand misstaat niemand,” Maandschrift Economie 47 (1983): pp. 395–410 (both cited in Van Parijs, Real Freedom for All, p. 264 nn. 34, 35).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 

193

that Rawls and many contemporary egalitarians gravitate to on straightforward egalitarian grounds. Of course, convergence on a shared solution hardly proves that left-libertarians are simply liberal egalitarians manqué. However, when one turns to the transfer side of the left-libertarian program, that is, the uses made of whatever tax revenues the state collects pursuant to the proviso, that conclusion seems unavoidable.

2.2 Redistribution of the Tax Base As Steiner notes, left-libertarianism logically implies an equal distribution of any taxes collected from resources subject to the proviso. (More precisely, it implies a combined tax and transfer scheme that leaves all individuals at the end of the day with equal shares of such resources, in kind or in cash equivalents.) Van Parijs and others argue that in prosperous societies, such a tax and transfer scheme could generate a sizable basic income for everyone, thereby satisfying basic-needs social welfarism. There may be room to move even further towards equality without abandoning Lockeanism entirely by interpreting “equal share” to permit the state to equalize the value individuals derive from mixing those assets with their talents. But the one thing Lockean libertarianism clearly seems to rule out is a combined tax and transfer scheme designed to compensate individuals for inequalities in personal endowments. At least it would seem to do so, if self-ownership is to be given any content beyond minimal, universally accepted prohibitions on forced labor, forced organ transplants, and similar forcible incursions on bodily integrity. Yet, when it comes to redistributing the proceeds of a Lockean tax, this is precisely what many left-libertarian schemes propose, including many that scrupulously adhere to a Lockean regime on the tax side. Brody defends that result on left-libertarian grounds, arguing that the ones who have been cheated out of their fair share of external resources are the poor, with the result that an egalitarian transfer scheme effects the Lockean proviso.⁵⁵ Most left-libertarians, however, have not tried to bring their redistributive programs into the left-libertarian fold. Instead, they typically concede that the proposed redistributive transfer scheme is intentionally designed to compensate for unchosen inequalities in personal endowments, but argue that result is consistent with left-libertarianism (even if not motivated by it) because some relatively thin notion of self-ownership is preserved on the tax side.⁵⁶ ⁵⁵ Brody, “Redistribution without Egalitarianism,” vol. II, p. 43. ⁵⁶ Variants of this argument can be found in Van Parijs, Real Freedom for All, which couples a Lockean tax on employment rents with a guaranteed basic income that, among other things, compensates for “dominated diversity” in internal endowments; Ackerman and Alstott, The Stakeholder Society; Sartorius, “Persons and Property”; Peter G. Brown, “Food as National Property,” in Food Policy: The Responsibility of the United States in the Life and Death Choices, edited by Peter G. Brown

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

194

-

All of these arguments depend on a sharp division of moral labor between the tax and spending side of fiscal policy, with taxes levied in accordance with Georgist left-libertarianism but the proceeds of the tax redistributed in whole or in part on egalitarian criteria. Left-libertarians are hardly alone in maintaining that moral division of labor. Rawls, for example, argues that his highly egalitarian Difference Principle should govern redistributive transfers, but that fairness to individual taxpayers requires that the revenues needed to finance such redistributive transfers should be raised through a proportionate (flat-rate) consumption tax.⁵⁷ One can see the same schizophrenic impulse at work among political progressives in the late nineteenth century, who split the difference in the opposite direction: they supported (on welfarist grounds) a progressive rate structure to raise the revenue necessary to run the minimal state, but rejected as socialistic the state’s use of such tax revenues for redistribution to advance the same welfarist goals.⁵⁸ Many contemporary right-libertarians continue to adhere to the same schizophrenic approach, albeit with a more modest commitment to welfarism on the tax side.⁵⁹ There may well be pragmatic reasons to distinguish between the tax and transfer sides of fiscal policy. Vallentyne hints at one, arguing that by limiting redistributive transfers to the size of the social fund derived from a tax on natural resources, we “place[] clear—and arguably [politically??—BF] plausible—limits on our duties to others.”⁶⁰ But beyond such prudential considerations, it is hard to see what recommends the division. The resulting schemes, which judge the tax and transfer sides of fiscal policy by wholly different distributive criteria, seem morally incoherent. If the just state may not take more from the talented in virtue of their greater talents—the premise of left-libertarianism—why may it give more to the untalented in virtue of their lesser talents? Otsuka’s answer is that distribution in inverse relation to talents is justified because it is “fair.”⁶¹ In a similar vein, Van Parijs defends differential transfers to compensate for extreme differences in internal endowments on the ground that without such transfers, the and Henry Shue (New York: The Free Press, 1977); Otsuka, “Self-Ownership and Equality,” vol. II, p. 162; Vallentyne, “Critical Notice,” pp. 621–3. ⁵⁷ John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971): pp. 278–9. ⁵⁸ Fried, Progressive Assault, at pp. 152–5, 302 n. 269. ⁵⁹ I have in mind various proposals for a so-called flat tax issuing from libertarian quarters, which are in reality progressive tax schemes administered through a so-called “degressive” tax schedule. All income below a certain level (approximately $25,000 for a married couple with two children) is exempt from taxation entirely; income above that level is taxed at a flat rate. As proponents concede, welfarist concerns are the only compelling justification for exempting a basic minimum from taxation. Indeed, it may well be that welfarism is the only compelling argument for proportionate (rather than steeply regressive) tax rates above that basic minimum. See Barbara H. Fried, “Why Proportionate Taxation?” in Tax Justice, edited by Joseph J. Thorndike and Denis J. Ventry (Washington, D.C.: Urban Institute Press, 2002), pp.149–92. ⁶⁰ Vallentyne, “Critical Notice,” p. 621. See also Brown, “Food as National Property.” ⁶¹ Otsuka, “Self-Ownership and Equality,” vol. II, p. 171, n.37.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi



195

severely handicapped and other seriously disadvantaged persons would unfairly lose the opportunity for “real freedom.”⁶² But the intuitions of fairness on which Otsuka, Van Parijs, and others draw here are the intuitions of liberal egalitarianism. If left-libertarians are free to help themselves to those intuitions on the redistributive side, why bother with left-libertarianism at all, and in particular with its very different and much narrower claim of equality on the tax side: that the earth belongs in equal shares to all of us? When all the dust settles, egalitarian commitments on the transfer side are likely to swamp any of the liberal/libertarian constraints placed on the tax side, leaving no visible trace of the latter. This is surely the result in Rawls’s proposed division of moral labor between taxation and transfers, as I have argued elsewhere.⁶³ It is likely the result in most left-libertarian schemes as well, at least if the value of external resources is large (and under the more expansive definitions of external resources, it will be), thereby generating a substantial tax base for redistribution.

3. Conclusion When one combines left-libertarians’ (i) typically thin reading of self-ownership and their broad reading of proviso on the tax side, with (ii) their willingness to distribute the proceeds of a Lockean tax base so as to achieve equality, it is hardly surprising that most schemes cashing out left-libertarian premises look a lot like what issues from egalitarian quarters. Often, all that seems distinctly libertarian is the rhetoric used to justify equality: we, the fortunate, are obliged to help the less fortunate because society has helped us in myriad ways that fall within the spirit of Locke’s proviso. Given the political temper of the times, signing on to the rightlibertarian program of self-ownership subject to the proviso, and then demonstrating why those twin commitments inexorably lead (contra right-libertarianism) to substantial equality, may well seem a shrewd strategy. It is worth considering, however, that there may be some real costs to it, at least when pursued in apparent sincerity, and not (as Roemer, Brown, and others do) simply as a demonstration arguendo of how far left the premises of right-libertarianism might take one. Most versions of left-libertarianism start with the assumption that the interpretation of self-ownership that explains why slavery and forced eyeball transplants are bad also explains, prima facie, why redistributive taxation is bad. G. A. Cohen (astonishingly, in my view) took that argument seriously enough to feel he had to renounce self-ownership entirely. There are so many obvious, ⁶² Van Parijs, Real Freedom for All, p. 101. ⁶³ See Fried, “Why Proportionate Taxation?” p. 173.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

196

-

morally salient, differences between slavery and redistributive taxation that it is hard to understand why anyone with egalitarian/social welfarist instincts would sign on, sincerely or strategically, to a program that treats them as homologous. Maybe something is to be gained in disarming a handful of right-leaning skeptics by conceding the moral power of self-ownership before proceeding to jettison all the practical implications the same skeptics ever thought the concept had. But there may be much more to lose with the far greater number firmly entrenched on the right, who will correctly read left-libertarians’ willingness to take seriously the proposition that taxation is morally akin to slavery as a huge propagandistic triumph. The latter group includes not only respectable academics, but also their seamier fellow travelers in a much more disturbing sub-culture of government haters—tax protestors, off-the-grid survivalists, and so on—who, starting from the same dubious premises, come to rest in a place much harder to digest. There is a long tradition of the left’s coopting natural rights talk to its own political ends.⁶⁴ In the same spirit, left-libertarians may hope that, by harnessing the concept of self-ownership for egalitarian ends, they can reclaim the moral high ground from right-libertarians. But in conceding that self-ownership is one of the chief principles on which a just state is built, they may well give up more than they bargain for in the public relations battle for the hearts and minds of those in the murky center of American politics, who harbor instincts of both liberty and equality (of the decent social minimum sort) that could be played to. At the very least, left-libertarians do well to keep in mind the old adage: If you eat with the devil, bring a long spoon.

⁶⁴ See Fried, Progressive Assault; Daniel T. Rodgers, Contested Truths: Keywords in American Politics Since Independence (New York: Basic Books, 1987), pp. 72–5, 122–30.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

11 Wilt Chamberlain Revisited Nozick’s Justice in Transfer and the Problem of Market-Based Distribution

In his 1973 review of Anarchy, State, and Utopia (ASU), Kenneth Arrow remarked that the view that “an individual is entitled to what he creates” is perhaps the most widely held intuition about distributive justice—and, he added, one that he had considerable difficulty persuading introductory students “[is] not completely self-evident.”¹ In Chapter 10, I explored disagreements among Lockean libertarians about whether that entitlement extends to the right to retain the exchange value of the thing one creates, and in particular to the portion of that value (“scarcity rents”) attributable to the scarcity of a factor of production (land, financial capital, or natural talents). I return to that problem now to examine Nozick’s treatment of scarcity rents in ASU. Within the structure of Nozick’s Lockean-inflected argument, the natural place for him to take up the problem of scarcity rents would have been under his principle of Justice in Acquisition, which describes how people justly acquire rights in something that has never been previously owned.² Instead, he circumvents that debate by justifying the owner’s right to scarcity rents under a different principle: Justice in Transfer. That move, I will suggest, rests on a fundamental misconception about what is at stake in a market exchange, a misconception that results from Nozick’s conflating two morally unrelated issues: whether we own ¹ Kenneth Arrow, “Some Ordinalist-Utilitarian Notes on Rawls’s Theory of Justice,” Journal of Philosophy 70, no. 9 (1973): p. 248. See also Elliot M. Abramson, “Philosophization Against Taxation: Why Nozick’s Challenge Fails,” Arizona Law Review 23, no. 2 (1981): p. 759, noting that “Nozick’s ‘whoever makes something is entitled to it’ notion is . . . intuitively appealing only because it is simply a tautological reference to culture-bound axioms of thought.” ² Nozick explicitly declines to specify the “details” of the principle of Justice in Acquisition, including the fundamental question whether it rests on Lockean or other premises. But in a curious (coy?) move he proceeds to insert the details of Locke’s principle of Justice in Acquisition, ostensibly as a placeholder for his own. Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1973), p. 153. But Part II of ASU read as a whole supports the conclusion that Locke is more than a placeholder for Nozick’s view—that Nozick is a Lockean, in the broad sense of believing that our claims on property, including its exchange value, trace back to our rights of self-ownership. While I will assume for the rest of the discussion that Nozick is a Lockean, it is not critical to my main argument, which (stated more broadly) is that whatever principle establishes the justice of original acquisition, the principle of Justice in Transfer does nothing to resolve the extent of property rights in the thing acquired, in particular the right to its market value. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

198

  

our holdings free from any claims of society; and whether we have a right to give to others what we own. By smuggling the problem of scarcity rents out of Justice in Acquisition, where it rightly belongs, and into Justice in Transfer, where it does not, he avoids having to defend scarcity rents as a logical entailment of a Lockean labor theory of ownership.

1. The Problem of Scarcity Rents As I discussed in Chapter 10, Locke famously argued that in a state of nature, if we mix our labor with the land, we thereby acquire a property right not just in the value added by our labor, but in the total value of the parcel, including the value of the land itself.³ But he immediately undercut that conclusion with his famous proviso: we may appropriate land out of the commons only if we leave “enough, and as good” for others—a requirement that, read literally, would preclude appropriation of any land or other natural resources the supply of which is not limitless. Thus, putting together the two halves of Locke’s theory and giving a literal reading to the proviso leads to the paradoxical conclusion that we may appropriate land out of the commons only when it would be of no value to do so because other equally good land exists in superabundance for all comers as far as the mind can imagine. Locke missed the paradox built into his argument, because he imagined an England with scarce land (making appropriation valuable) and a fictive America with land in superabundance (leaving “enough, and as good” for all others deprived of the opportunity to appropriate land in England). But those two conditions (scarcity in England and abundance in America) could coexist only because, due to its locational disadvantages, land in America was not an economic substitute for land in England. Thus, a herder in the late seventeenth-century Cotswolds might rightly feel that the offer to relocate him to the wilds of Massachusetts after dispossessing him of the commons in his home town does not satisfy the true spirit of the requirement that he be given “enough, and as good.” If it really is “enough, and as good,” why shouldn’t the appropriator be told to leave the commons alone and move to America himself? As I discussed in Chapter 10, Locke’s libertarian heirs have disagreed strongly about how far entitlement theory extends. At one extreme, many have argued that it extends to all the fruits of our labor, including whatever we can get for them on the free market. As J. S. Mill described that view almost two centuries ago in language that could have been written by Nozick:

³ John Locke, The Second Treatise of Government, in Two Treatises of Government, edited by Peter Laslett (Cambridge: Cambridge University Press, 1967) (1690), § 27.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

    

199

Nothing is implied in property but the right of each to his (or her) own faculties, to what he can produce by them, and to whatever he can get for them in a fair market; together with his right to give this to any other person if he chooses, and the right of that other to receive and enjoy it.⁴

But many Lockeans have struggled with the justification for giving labor that portion of market value that reflects scarcity rents accruing to other factors of production (financial capital, market opportunities, and natural talents). As Mill explained it, the moral appeal of a Lockean labor theory of ownership lies in its promise of “proportion between remuneration and exertion”—something that private appropriation of scarcity rents would upset: To judge of the final destination of the institution of property, we must suppose everything rectified which causes the institution to work in a manner opposed to that equitable principle, of proportion between remuneration and exertion, on which in every vindication of it that will bear the light it is assumed to be grounded.⁵

In short, if one-tenth of the market value of Blackacre reflects the value of the owner’s labor in constructing a house on it and the other nine-tenths reflects the appreciation in the value of the land due to its natural scarcity, why should his labor alone entitle the owner to 100 percent of its value under a Lockean theory of ownership, such that the government’s taxing it amounts, in Nozick’s words, to “appropriating the actions of other persons.”⁶ Doubts about the justification for private appropriation of scarcity rents have played a significant role in debates about the distributive justice of the market over the past two centuries. In the classical economic tradition, the attack on scarcity rents focused on economic rents that capital could command from labor, which in the Marxian view were the putative source of all value. Defenders of capitalism took the attack seriously enough at the time to parry it by reconceiving profit as the “wages of abstinence” (that is, as the remuneration justly due capitalists’ exertion in abstaining from current consumption of their wealth). Henry George and his followers, taking their cue from David Ricardo, assumed that only one form of capital—land—was the residual claimant that would take all the surplus value after labor and mobile capital were paid their fixed charges, an assumption that fueled the wildly popular land tax movement in England and the US in the late nineteenth and early twentieth centuries. No less a figure than (the early)

⁴ John Stuart Mill, Principles of Political Economy (London: Parker, Son, and Bourn, 1865), Book II, ch. 2, § 3, p. 267 (emphasis added). ⁵ Mill, Principles of Political Economy, Book II, ch. 1, § 3, p. 255. ⁶ Nozick, Anarchy, State, and Utopia, p. 172.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

200

  

Herbert Spencer argued that working a piece of land gives one at most a lien against the land, equal to the added value of one’s labor.⁷ In the neo-classical tradition, the Fabian Socialists and British New Liberals generalized the Ricardian attacks on land rents to all factors of production—land, labor, and capital. They argued that any of the three factors that was in short supply could command a scarcity rent, a view succinctly captured in their “law of three rents.” Finally, cutting across economic theories, political eras, and political factions, the hostility to scarcity rents found a ready target in the rents that accrued to sole suppliers of a factor of production—that is, monopolists.⁸ Doubts about the private appropriation of scarcity rents persist in contemporary libertarian thought.⁹ Nozick’s unease is evident throughout ASU. He acknowledges it explicitly in discussing extortionate monopoly pricing.¹⁰ As discussed in Chapter 7, it is implicit at other junctures where, instead of trying to justify an owner’s right to scarcity rents in Lockean terms, for avowedly utilitarian reasons he offers the singularly un-Lockean defense that might makes right.¹¹ At the same time, defending an owner’s right to retain the full exchange value of property is central to Nozick’s project in Part II of ASU, and the reader is left with little doubt that, one way or another, Nozick is going to declare victory. As I suggested at the outset, the logical place for Nozick to have taken up scarcity rents is in fleshing out his principle of Justice in Acquisition, given that the two central questions addressed by the principle are “the things that may come to be held [by a particular process of just acquisition],” and “the extent of what comes to be held” by that process.¹² But Nozick never resolves it in that context. ⁷ Herbert Spencer, Social Statics, or, the Conditions Essential to Human Happiness Specified, and the First of Them Developed (London: John Chapman, 1851), ch. 9, §§ 2, 8–10, pp. 132, 141–4. For the problematic status of private land ownership in nineteenth and early twentieth century political theory, see Barbara H. Fried, The Progressive Assault on Laissez Faire: Robert Hale and the First Law and Economics Movement (Cambridge, Mass.: Harvard University Press, 1998), ch. 3. ⁸ For a fuller history of rent-based attacks on the market, see Fried, Progressive Assault, ch. 4. ⁹ See, for example, Richard A. Epstein, “The Utilitarian Foundations of Natural Law,” Harvard Journal of Law and Public Policy 12, no. 3 (Summer 1989): pp. 713, 730–2; David Miller, “Justice and Property,” Ratio 21 (June 1980): pp. 11–14; Eric Mack, “Distributive Justice and the Tensions of Lockeanism,” Social Philosophy and Policy 1, no. 1 (Autumn 1983): pp. 132–50. As Richard Epstein notes, it is theoretically unsatisfying and empirically unjustified to dismiss the problem as de minimis. Epstein, “Utilitarian Foundations,” p. 731. ¹⁰ The best-known example is the passage in which Nozick concedes that the Lockean proviso may prevent an otherwise just Lockean appropriator of the only water hole in town from “charg[ing] what he will” for water. Nozick, Anarchy, State, and Utopia, pp. 179, 180. In his well-known 1969 article, “Coercion,” Nozick expresses the same moral qualms about the right of a passerby who rescues a drowning person to extract an extortionate price for his efforts. “Coercion,” in Philosophy, Science and Method: Essays in Honor of Ernest Nagel, edited by Sidney Morgenbesser, Patrick Suppes, and Morton White (New York: Saint Martin’s Press, 1969), pp. 440–72, n. 29. ¹¹ The principal examples discussed in Chapter 7 were the right of the dominant protective agency to extinguish all other protective agencies and appropriate for itself the surplus value inherent in holding a monopoly on political power (Part I of ASU), and Nozick’s self-described “weak” version of the Lockean Proviso, which allows first appropriators out of the commons to claim for themselves the scarcity rents inherent in natural resources and other common assets (Part II of ASU). ¹² Nozick, Anarchy, State, and Utopia, p. 150.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

’     

201

He deliberately distances himself from one popular libertarian defense of private appropriation of scarcity rents: people are entitled to what the market pays for their goods and services, because it measures the value such goods or services generated for society (that is, their marginal product).¹³ And he offers no alternative ground on which to conclude that it is an entailment of Justice in Acquisition. Instead, he purports to justify it under his second principle, Justice in Transfer. I turn to that now.

2. Nozick’s Principle of Justice in Transfer Under Nozick’s “historical” theory of distributive justice, one rightfully owns something if one has justly acquired it under one of three principles: by acquiring something that has never previously been owned, in accordance with the principle of Justice in Acquisition; by acquiring it in transfer from someone else, in accordance with the principle of Justice in Transfer; or by acquiring it as compensation for past violations of either of the first two principles (the principle of Rectification).¹⁴ Justice in Transfer holds that we justly own something if we acquire it in a voluntary transfer from someone who rightly owns it herself. As Nozick famously summarized it, “[T]o each according to . . . what others choose to do for him and choose to give him of what they’ve been given previously (under this maxim) and haven’t yet expended or transferred.”¹⁵

¹³ Nozick attributes the argument to Friedrich Hayek; Anarchy, State, and Utopia, pp. 158–9. But Hayek himself eschews any attempt to defend as “just” whatever distribution results from the unimpeded operation of the market. He defends it partly on incentive grounds, and (more importantly) as a necessary byproduct of the personal freedom of action that comes only with the spontaneous ordering of the market. F. A. Hayek, The Constitution of Liberty (Chicago: University of Chicago Press, 1959), pp. 95–l00, Law, Legislation and Liberty, vol. 2 (Chicago: University of Chicago Press, 1976), pp. 67–78. A more apt cite (among others) would be John Bates Clark, whose The Distribution of Wealth (New York: The Macmillan Company, 1899) argued unambiguously for a moral entitlement to the market value of what one produces. Nozick’s position on the moral significance of marginal productivity is slippery at best, but seems to boil down to this. He invokes the theory of marginal productivity to argue (contra Rawls) that when things are jointly produced, we can measure each person’s contribution for purposes of implementing the theory “to each according to what he makes”: his contribution is his marginal product, as determined automatically by market prices; Anarchy, State, and Utopia, pp. 187–8. He also takes some comfort from the fact that most transfers in a market economy will be to people in accordance with their marginal product, a motive that is “reasonable and intelligible.” (Why?) Anarchy, State, and Utopia, p. 159. But he rejects the argument that we are entitled to what the market pays us simply because it reflects the value of our marginal product. See, e.g., Anarchy, State, and Utopia, pp. 158, 188, arguing that while distribution according to benefits to others (marginal productivity) “is a major patterned strand in a free capitalist society,” it is not the only one, nor is it “a standard that one should insist a society fit.” ¹⁴ Nozick, Anarchy, State, and Utopia, pp. 150ff. ¹⁵ Nozick, Anarchy, State, and Utopia, p. 160; see also pp. 186–7.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

202

  

Nozick makes much of the fact that voluntary transfers could be motivated by a number of factors, thus making the resulting distribution of goods “unpatterned,” in the sense intended by his historical theory of entitlement.¹⁶ But as he elsewhere acknowledges, the numerous examples he provides fall into one of two classes: gratuitous transfers (inter vivos gifts and inheritance, charitable contributions, grants, etc.) and market exchanges (sales of goods and services and returns on investment, including interest, dividends, and gambling proceeds).¹⁷ Nozick treats the two classes as identical for all purposes under Justice in Transfer. But as I argue below, the two pose quite different justificatory problems. Nozick suggests that Justice in Transfer plays a role in his theory of distributive justice that is theoretically modest, but important practically. It is theoretically modest, because it purports to explain only the conditions under which one can justly transfer an entitlement from one person to another, leaving the content of the entitlement to be determined by Justice in Acquisition or the principle of Rectification. It is important practically, because the actual distribution of holdings at any given moment is the product of many such transfers.¹⁸ Professing complete befuddlement as to why market prices should be rejected as “an appropriate theory of fair or just shares” in favor of some other measure of “just price,” Nozick writes: It is difficult to see why these issues should even arise here. People are choosing to make exchanges with other people and to transfer entitlements, with no restrictions on their freedom to trade with any other party at any mutually acceptable ratio. Why does such sequential social cooperation, linked together by people’s voluntary exchanges, raise any special problems about how things are to be distributed? Why isn’t the appropriate (a not inappropriate) set of holdings just the one which actually occurs via this process of mutually-agreed-to exchanges whereby people choose to give to others what they are entitled to give or hold?¹⁹

In a famous passage, Nozick argues that if (i) one million people who want to see Wilt Chamberlain play basketball choose to transfer 25 cents to him as the price of admission; (ii) all one million justly own their 25 cents to begin with; and (iii) ownership includes (among other things) the right to dispose of your property in any manner you see fit (“what was it for if not to do something with?”), then Chamberlain must be entitled to the $250,000 he has thus acquired—entitled, in

¹⁶ ¹⁷ ¹⁸ ¹⁹

Nozick, Anarchy, State, and Utopia, pp.156–7. Nozick, Anarchy, State, and Utopia, p. 149. Nozick, Anarchy, State, and Utopia, pp.151–2. Nozick, Anarchy, State, and Utopia, pp. 186–7 (emphasis and footnotes omitted).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

’     

203

particular, as against a government that seeks to redistribute a portion of it through taxation and wealth-transfer programs.²⁰ By what process could such a transfer among two persons give rise to a legitimate claim of distributive justice on a portion of what was transferred, by a third party who had no claim of justice on any holding of the others before the transfer?

asks Nozick, rhetorically.²¹ The example has been subject to numerous criticisms, including the unsupported assumption that each fan’s initial right to her 25 cents must necessarily carry with it the absolute right of transfer (proposition (iii)).²² One could, for example, defang Nozick’s argument simply by defining the property right ab initio as not including the right to transfer the property in a fashion that will upset unduly some end-state egalitarian distributive pattern. In a similar vein, proponents of a confiscatory estate tax have long argued that property rights are bestowed for reasons that evaporate upon the death of the owner. Thus, if the state declares that property held at death escheats to it, it is not abrogating property rights; it is merely enforcing them in accordance with their initial, intended configuration. One would have to defend that initial configuration as a normative matter, but then again, Nozick has to defend his as well. But the example, and the theory of Justice in Transfer that it is meant to illustrate, has a deeper problem. As I stated at the outset, it confuses two questions: whether each party has a right to the full market value of the asset he holds the moment before exchange; and if so, whether he has a right to transfer it to another person as a gift or in exchange for another asset of equal value. Nozick’s example

²⁰ Nozick, Anarchy, State, and Utopia, pp. 160–4. ²¹ Nozick, Anarchy, State, and Utopia, pp. 161–2. It is perhaps worth noting the structural similarities between Nozick’s defense of market-based distribution and Hayek’s. Both avoid the need to defend the substantive fairness of distributive outcomes in a market economy by arguing that those outcomes are justified as a necessary byproduct of an independently valuable freedom. In Nozick’s case, the relevant freedom is the freedom of both parties to a contract to give away their holdings to whomever they wish (including the party on the other side of the contract). In Hayek’s case, it is the freedom of the seller of goods or services to “use his knowledge for his own purposes”; Hayek, Law, Legislation and Liberty, vol. 2, p. 69. ²² See Thomas Nagel, “Libertarianism without Foundations,” Yale Law Journal 85, no. 1 (Nov. 1975): p. 136, reprinted in Reading Nozick: Essays on Anarchy, State, and Utopia, edited by Jeffery Paul (New Jersey: Rowman & Littlefield, 1981), pp. 201–2; Onora O’Neill, “Nozick’s Entitlements,” Inquiry 19 (1976): pp. 468–81, reprinted in Reading Nozick, pp. 308–10; Cheyney Ryan, “Yours, Mine, and Ours: Property Rights and Individual Liberty,” Ethics 87, no. 2 (Jan. 1977): pp. 126–41, reprinted in Reading Nozick, pp. 325–35. For the observation more generally that historical theories of justice, like end-state theories, may limit transfers by requiring that the transfers preserve any provisos attached to initial acquisition, see Lawrence Becker, “Against the Supposed Difference Between Historical and EndState Theories,” Philosophical Studies 41, no. 2 (Mar. 1982): pp. 267–72. As Becker notes, Nozick’s own theory requires such “tinkering” with people’s property rights to preserve others’ rights under the Lockean proviso even after original just acquisition. Nozick, Anarchy, State, Utopia, pp. 179–80.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

204

  

goes only to the latter question, but it is the former one in which the problem of scarcity rents is housed. The problem with Nozick’s argument may be easier to see by starting with a different example. Imagine that in the 1950s WC bought a parcel of vacant land in a sparsely populated county adjacent to New York City for $5,000 in cash, which cash he had saved from his earnings as a day-laborer. Over the ensuing twenty years, economic, demographic, and other societal changes spurred large numbers of people who worked in New York City to immigrate to the suburbs, driving real estate prices up a hundred-fold or more.²³ By the early 1970s, WC’s land is worth $250,000. What are WC’s Lockean rights to the market value of the land? First, assume WC continues to hold on to the land rather than sell it. Although WC has chosen not to cash out his investment, he holds an asset that, if he wished, he could cash out at any time for $250,000. By any conventional notion of wealth, it is clear that WC is wealthier now than he was twenty years earlier, by $245,000. (Suppose, for example, WC had just won the land in a contest he paid $5,000 to enter.) In support of WC’s claim to the full $250,000, right-Lockeans would argue that because WC bought the land with the fruits of his labor (the $5,000 in savings being his stored labor), he owns it absolutely, as against any claims by the state. There are a number of things to say in defense of that position, some arising from a theory of rights and some from utilitarian considerations. But, as suggested above, the position is hardly unassailable. Left-Lockeans, including the Ricardian socialists, Henry George, the early (unreconstructed) Spencer, and others would agree with the right-Lockeans that by virtue of his labor, WC is entitled to a portion of the value of that land. But they would argue that the portion is limited to WC’s actual cost, or sacrifice, in acquiring it ($5,000), plus perhaps a fair return on that cost.²⁴ Any appreciation in value above that amount is pure luck as far as WC is concerned, resulting from the confluence of a naturally constrained supply of land in commuting distance from New York City and increasing societal demand for such land. A left-Lockean would argue that if anyone has a right to that appreciation, it is not WC, who has already received remuneration in proportion to his exertion, but society at large, out of whose commons the land was taken in the first place, and whose changing ²³ For purposes of this argument, I am ignoring inflation. In the real world, inflation changes the relevant numbers for measuring real increases in wealth, but not the underlying argument. ²⁴ WC’s right under some Lockean theory of entitlement to a fair return, reflecting the going (risk-adjusted?) interest rate, raises a host of philosophical puzzles of its own, many with a rich history in political and economic thought. I have dealt with them in detail elsewhere. See Fried, Progressive Assault, ch. 4. They are hardly irrelevant to the example, as WC’s entitlement to the entire $250,000 might best be defended by arguing that it represents (taking WC’s portfolio of assets as a whole) a fair return to risk. That argument, however, is beyond the scope of the claim here, which requires only that we agree that a supernormal return on investment on its face presents moral difficulties for a Lockean theory of entitlement, which ought to be dealt with somewhere in a Lockean theory of distributive justice.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

’     

205

tastes and material wealth put a financial premium on the land’s scarce locational attributes. As Edmund James put the argument a century ago, To test the relative productivity of the state and the individual, compare the fortune accumulated by Cornelius Vanderbilt in America with what he might have accumulated had he been adopted when an infant by a family of Hottentots.”²⁵

Under the left-Lockean view, society could justly exercise its rights as a silent partner in WC’s investment by taxing WC on the increase in the value of the land as it occurs. For administrative and other prudential reasons, it might decline to exercise its right until such time as WC sells the property. Among those reasons are the difficulty of valuing property that hasn’t recently changed hands in an arm’s length market exchange, and the so-called “liquidity” problem, meaning that WC may have trouble coming up with cash to pay a tax on the annual increase in value if he hasn’t actually sold the property for cash. In our current tax system, those concerns have led to a general “realization” requirement: we generally don’t tax appreciation in the value of financial assets until the assets are sold for cash or exchanged for other non-like assets. But the fact that a society chooses not to tax the appreciation in the value of assets as it accrues does not change the fact that (under the left-Lockean view) it has the moral right to do so if it wishes. Suppose instead that when the land was worth $250,000, WC chose to give it to his children as a gift. The principle of Justice in Transfer says in effect that the right to give one’s property to someone else is an entailment of ownership. Thus, to the extent WC justly owns the land, he may transfer ownership to his children whenever he wishes. As suggested above, Nozick’s conclusion that the right of unrestricted transfer is entailed in ownership, although widely shared, is not selfevident. But the relevant point for present purposes is not the conclusion, but the premise: WC may give to his children only what he owned to begin with the moment before transfer. What did he own? Right-Lockeans would say WC owned all rights of exploitation in the property (including the right to retain its full market value upon sale). Left-Lockeans would say he owned only a portion of its value, equal to his investment of labor or capital in the property (plus a fair return), and that society has a claim on the rest. Society could exercise that claim by taxing WC on the increase in value as it occurs. Again, for a variety of administrative reasons, it might choose to defer taxation until WC transfers the property to his children, or even later, until the children sell the property at a gain (the actual

²⁵ Edmund James, “The State as an Economic Factor,” in Science Economic Discussion (New York: The Science Company, 1886), p. 32.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

206

  

solution under our income tax system). But again, the choice to defer taxation is merely pragmatic. Suppose now that when the land was worth $250,000, WC sold it to a willing buyer, X, for $250,000 in cash. Does the government have a claim on any portion of the $250,000? Nozick’s principle of Justice in Transfer says that because X voluntarily gave WC the $250,000, which (we will stipulate) X rightfully owned, WC must now rightfully own the $250,000. But I hope it is now clear that the transfer is irrelevant to resolving society’s right to a portion of the $250,000. When X gives WC $250,000, WC is not $250,000 richer. He is no richer at all. He has merely traded land worth $250,000 for $250,000 in cash, thereby changing the form but not amount of his wealth.²⁶ What made him wealthier was not the $250,000 in cash from X, but the $245,000 appreciation in value of the land over the preceding twenty years. Under the left-Lockean view, society could have taxed WC on the $245,000 appreciation in the land even if he had never sold or otherwise disposed of the property. It cannot be that society has fewer rights when WC does cash out his gain. As suggested in the two preceding examples, while society might have chosen for a variety of administrative and other reasons to wait to claim its portion of the appreciation until WC cashed out his gain, it is nonetheless the gain, and not the cash, that society claims a moral right to. With respect to that claim, X’s transfer is irrelevant. (Contrast, for example, the situation where WC buys the land for $250,000 in cash, and sells it the next day to Z for $250,000. Now the government claims nothing from WC, not because (as Nozick would have it) Z has voluntarily transferred the $250,000 to WC, but because the land never appreciated over its original cost, and hence WC gained nothing by acquiring and holding it.) Right-Lockeans would disagree with left-Lockeans’ conclusion, but only because they would disagree with the premise that WC did not own the full value of the property the moment before transfer. Both should agree, however, that the exchange itself cannot enlarge whatever preexisting rights WC had in the property. Nozick has erroneously concluded that X’s transfer is relevant to resolving WC’s rights to the appreciation in the value of the property because he misconstrues what is at stake in the transfer. Even if one accepts the core claim of Nozick’s Justice in Transfer that whatever X owns she can give to WC, X did

²⁶ This need not always be the case. If X values WC’s land more highly than the market and outbids what would have been the market price without X in the market (say, $240,000), then X’s willingness to buy does increase WC’s wealth by $10,000 (the difference between the market value of the property before X entered the market and the market value after). But even as to that last $10,000 in value, it is X’s offer to purchase at $250,000 that created the $10,000 extra value, not the actual transfer of cash. (Even if WC refuses the offer, he now possesses an asset worth $250,000 on the market, and at least that much to him or he would have sold at that price.)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

’     

207

not give the $250,000 to WC. She gave it to the economic owner of the land to which WC holds title, in order to obtain ownership herself. If, as Left-Lockeans would maintain, WC owns only $5,000 of the value of the land (arguably plus a fair return on the investment), and holds title to the remaining $245,000 in effect as a trustee of society, WC has no more right to retain the entire $250,000 than he would were the land encumbered by a private mortgage. Imagine, for example, that WC had purchased the land for $250,000, using $5,000 of his own capital and a $245,000 loan from the bank secured by the land, and that WC subsequently resold the land to X for $250,000. Notwithstanding that WC held sole title to the land, the bank in effect co-invested in the property with him, to the tune of $245,000, and has a legal right to demand that WC turn over $245,000 of the sales proceeds to it. The state, in the left-Lockean view, is a passive co-investor just like the bank, and retains a lien on any appreciation in the scarcity value of the land itself. Thus, transfer by gift and by exchange, which Nozick casually treats as posing identical problems in entitlement theory, raise completely different issues.²⁷ If X owns $250,000 in cash free of any claims by the state and transfers it to WC as a gift, the sole question is whether, as Nozick asserts, owning something for some purposes necessarily carries with it the right to transfer it to whomever we wish— or to put it in Nozick’s terms, whether “persons who have a right to hold also have a right to choose that others hold in their place.”²⁸ But if X transfers that same cash to WC in exchange for property, X’s wish to benefit WC, should it even exist (and of course it typically won’t in a market exchange), cannot determine WC’s right to that cash. Such right depends in the first instance on whether WC owns what X gave him the cash to get: the right to the full value of the property. How does all of this apply to Nozick’s example of the hypothetical exchange between Wilt and a representative fan X? It applies exactly, once one makes the necessary translation from financial capital (WC’s land) to human capital (Wilt’s basketball talent). Wilt has taken his natural, scarce, basketball talent, and added to it considerable hard work, to transform it into “human capital” (in economists’ lingo) in the form of superstar-level basketball skill. That human capital has a high value on the market, because society (due to its peculiar consumption tastes) places a high value on superstar-level basketball playing, and Chamberlain is in effect a monopoly supplier of basketball skills at that level, as a consequence of the natural scarcity of his talent. When X and the million other fans who share society’s peculiar consumption tastes pay Wilt a total of $250,000 to watch him play, Wilt is not wealthier in the broad sense. He has merely changed the

²⁷ For Nozick’s treatment of them as identical, see Anarchy, State, and Utopia, p. 161: “If the people were entitled to dispose of the resources to which they were entitled . . . , didn’t this include their being entitled to give it to, or exchange it with, Wilt Chamberlain?” ²⁸ Nozick, Anarchy, State, and Utopia, p. 168.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

208

  

form of his wealth from an ability to command a gate of $250,000 to the actual gate receipts. What are Wilt’s Lockean rights to that $250,000? A number of new wrinkles are introduced as we move from financial capital to human capital, which lead most left-Lockeans (no doubt along with most readers) to answer the question differently than in the land context. Those wrinkles are worth noting. But the core normative question is structurally identical: whether, by virtue of having been born with great talent, to which he added his labor, Wilt becomes entitled to whatever the market will pay him for exploiting that talent. Nozick, famously, replies, of course he does, because if X owned the 25 cents that she voluntarily transferred to Wilt in payment for a ticket to watch him play, then Wilt must own that 25 cents now. But as I hope the above excursion into WC’s land investment makes clear, X’s willingness to transfer the 25 cents is irrelevant to resolving the normative question in dispute: whether Wilt owns, unencumbered by any obligations to society, the thing for which he is being paid by his fans, which is the market value of his human capital. At the outset, I outlined the right-Lockean argument that he does: When Wilt mixes his labor with his natural talents, both of which he owns, he acquires all rights to the resulting asset (the ability to play spectacular basketball), including its exchange value. Left-Lockeans, on the other hand, will argue that Wilt’s natural talent is the result of brute luck, requiring no sacrifice on his part. It is true that he mixed considerable labor with his talents to produce his high level of skill, and he is entitled to fair compensation for that labor. But a willingness to work hard is not in short supply; he works no harder than the fiftieth ranked basketball player who earns 1/100th of his salary, or for that matter the typical day laborer who earns 1/1,000th. As a result, very little of the money that he commands for one night’s gate (let us say only $5,000 out of the $250,000) represents a return on hard work. The remaining $245,000 reflects scarcity rents he can extract because of the demand for spectacular basketball playing and the natural scarcity of talent able to supply it. As between Wilt, who did nothing to earn that good luck, and society, which supplied the demand for his talents as well as the material conditions that allowed him to develop them, society has the stronger claim on the scarcity rents.²⁹ It is not my purpose here to defend either of these views in preference to the other, or to some third alternative that finds a midpoint between them or departs ²⁹ As I discussed in Chapter 9, the most famous contemporary version of the argument that individual endowments are part of a common pool available for redistribution as society sees fit is given in Rawls’s A Theory of Justice. Rawls’s argument differs from the left-Lockean argument, not merely in the hypothetical contractualist procedure he employs to reach that conclusion, but also in its substantive scope. Rawls, unlike most left-Lockeans, would throw the returns to hard labor—Wilt’s $5,000 share of the gate—into the common pool as well, on the ground that the “superior character” that enables people to cultivate their talents through hard work is equally the product of (undeserved) good fortune in family and social circumstance. John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1974), pp. 103–4.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

’     

209

from different premises entirely. It is merely to point out that the choice between them has nothing to do with whether Wilt’s fans voluntarily forked out a portion of those scarcity rents, in the form of a 25 cents admission fee, to see Wilt play. To drive the point home, it should be noted that as with WC’s land, society could lay claim to a portion of those rents even if Wilt never played. Wilt is wealthier in an economic sense the moment he is born with enormous basketball talent than is an otherwise identical child without that talent. (If you doubt it, ask yourself whether you would pay something—indeed quite a lot—to be born Wilt instead of that other fellow.) Exactly how much wealthier fluctuates year by year, as the taste for professional basketball, disposable income of fans, bargaining power of players, and assessments of Wilt’s talent change, and as Wilt gets closer to the age when he can cash out that talent. But at least in theory, we could tax him on the value of that income-earning potential at the moment of birth, with appropriate adjustments each year to reflect changes in its value. Such a tax, which economists and tax theorists call an endowments tax, would be exactly analogous to a tax levied annually on appreciation in the value of WC’s land. There are many reasons why most people oppose an endowments tax on human capital. Some of those reasons—in particular problems of valuation and liquidity—also arise with respect to financial capital like WC’s land, but are more severe in this context.³⁰ Other reasons are unique to human capital. Perhaps the most serious is the one I touched on in Chapter 10: When we tax people on the full market value of their assets if put to their highest-paying use, we indirectly pressure them to put those assets to such use. That fact may not bother most people in the case of financial assets.³¹ But it disturbs almost everyone when the tax is levied on human capital. To take an example from real life and sticking with the basketball theme, if we tax Michael Jordan on what he could have earned from

³⁰ As to valuation, Wilt’s talent won’t be revealed for many years; once it is revealed, it may be hard to value until shortly before he is in a position to exploit it professionally. For people with less clearly specialized talents, even identifying the right market(s) to value those talents might be difficult. Finally, people are born with all sorts of personal endowments (loving or well-connected parents, good health, good looks, intelligence, the “right” race or sex, high socioeconomic status) that give them ex ante a much higher earning capacity than an otherwise identical person who lacks them. How do we decide which of these endowments to count for tax purposes, and how do we value them? As to liquidity, while many people might pay a lot to have Wilt’s talent, because of severe agency problems not present in most forms of financial capital far fewer would pay a lot to have a future interest in Wilt’s own development of his talent, for example, by lending him money at age 15 against his future expectancy of high earnings. ³¹ One notable exception is personal residences. Many states have adopted measures like California’s Proposition 13, which restrict the power of the state to levy a property (wealth) tax on the appreciated value of people’s personal residences. One of the motivations for such measures is to protect people’s ability to continue living in homes they have grown emotionally attached to but could no longer afford to acquire at their current market value, rather than being forced to rent or sell them to others who (given their income level) value the property highly enough to pay the tax on its full market value. Another way of describing that decision is that we let people consume the value of their personal residence rather than forcing them to sell it, even when the decision is economically suboptimal, because we think there are important interests of personhood at stake.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

210

  

the Chicago Bulls when he throws up that life to be a third-rate, unemployed baseball player, we indirectly coerce him to continue playing for the Chicago Bulls in order to pay his tax bill. Most people would find that result an unacceptable infringement on personal autonomy. In light of these considerations, virtually all left-Lockeans would hold off taking what they see to be society’s share of Wilt’s earning capacity until Wilt cashes it out by playing basketball for money. But the critical point for present purposes is that, as with WC’s land, when we ultimately do tax Wilt on the $250,000 he received from his fans, it is not because (as Nozick would have it) we think those fans have no right to give him the money in exchange for the pleasure of watching him play. It is because we believe that Wilt has no right to keep it all, because it was paid in significant part for talents he did not earn, in a strict Lockean sense, and which are exploitable only because of the society in which he lives. I’ve focused only on the portion of Wilt’s earnings reflecting scarcity rents, because that is the aspect of market exchange that has given Nozick and other right-Lockeans the most trouble, and hence the one for which Nozick’s principle of Justice in Transfer is doing the most illicit work. But the same logic that makes X’s transfer irrelevant to establishing Wilt’s Lockean right to the $245,000 that represents monopoly rents also makes it irrelevant to his right to the remaining $5,000 that represents the return to hard work. If we conclude that Wilt has a right to that $5,000, unencumbered by any societal lien, it is because we believe that twelve hours a day practicing on the basketball court represents the sort of sacrifice that entitles him to the value it creates, not because his fans paid for that value willingly. There may be better arguments for letting people keep the scarcity rents that accrue to their holdings. But within the structure of Nozick’s historical theory of distributive justice, those arguments should be made, and their difficulties faced, in elaborating and defending the content of Justice in Acquisition, in particular “the extent of what comes to be held by a particular process” of just acquisition.³² Nozick has failed to do that. Invoking Justice in Transfer in its stead obscures that failure, but it does nothing to redress it.

³² Nozick, Anarchy, State, and Utopia, p. 150.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

12 “If You Don’t Like It, Leave It” The Problem of Exit in Social Contractarian Arguments

In Chapter 9, I considered Rawls’s effort to derive the just terms of political cooperation from a hypothetical bargain struck from the original position. In this chapter and Chapter 13, I look at parallel efforts by classic liberals and libertarians that have led them to conclusions that diverge widely not only from Rawls’s, but from each other’s as well. That divergence, I suggest, is largely due to the different exit options with which they have endowed the hypothetical bargainers. In A Theory of Justice, Rawls suggests that the talented would find it in their self-interest to sign on to his highly redistributive Difference Principle, because the resulting arrangement would still leave them better off than they would be with no social cooperation at all.¹ To make the argument persuasive, Nozick observed in rejoinder, Rawls needs to explain why the talented would not demand the benefit of the even better deal they could strike by seceding with the other talented members of the social contractarian (SC) bargaining unit and heading off for some utopian Lucky Island.² Nozick’s rejoinder is telling, if not for Rawls, whose egalitarian precommitments embedded in his original position preclude such a deal from the start, then for other left-leaning contractarians hoping to derive egalitarian conclusions from rational self-interest. But it needs some explaining itself. Suppose that in any realistically constructed state of nature (SON), the talented would never actually make it to Lucky Island. Suppose, for example, that collective action problems would make it impossible to negotiate secession. Or suppose that the talented have high sunk costs (in affectional and economic ties) in their existing SON societies

¹ John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971), p. 103. As in previous chapters, I use “talented” as a placeholder for all personal characteristics, whatever their origin, that (in Nozick’s words) give individuals the ability to “accomplish something of great economic advantage to others” and hence command a higher market return for their labor. Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 193. ² Nozick, Anarchy, State, and Utopia, pp. 193–7. For G. A. Cohen’s criticism of Rawls along the same lines, see Self-Ownership, Freedom, and Equality (Cambridge: Cambridge University Press, 1995), pp. 223–6, and “The Limits of Contractual Equality: A Reply to Jacques Bidet,” Ratio Juris 8, no. 1 (Mar. 1995): p. 87. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

212

“  ’  ,  ”

that make secession from those societies unattractive; or that there is no Lucky Island left to go to by the time the SC bargain is imagined to take place; or that the costs of creating a rival social organization from scratch on Lucky Island are prohibitive. Suppose, as a result, that Nozick’s band of the talented cannot credibly threaten to secede from whatever bargaining unit they find themselves in, in order to form a rival society on Lucky Island. Are we still required to abide by the distributive bargain the talented could have obtained for themselves had they been able to get to Lucky Island costlessly, even though they cannot? To put it another way, why don’t we treat the actual social unavailability of exit as a limitation on endowments, akin to lesser talents? If—as Nozick would have it—poorly endowed members of any society are stuck with their lesser endowments because they are a given social fact, why are the talented not stuck with the impossibility of exit, because it is likewise a given social fact? The narrow question I pose in this chapter is: why is the statement “If you don’t like it, leave it” not a morally adequate response to the complaints of, say, the wealthy in America circa 2020 that their society is overtaxing them to aid the poor? The broader question I explore is how exit options are constructed in SC arguments that seek to justify a particular division of the social surplus generated by political cooperation. By exit options, I mean all alternative arrangements available to negotiators in the event they fail to reach agreement with their co-bargainers. I will focus on the social constraints on relocating from one society to a hypothetically more attractive one. In the real world, such constraints typically include sunk social capital, the transactions costs of negotiating secession, and the social unavailability of attractive alternative SCs. The exit options that parties to a hypothetical SC are endowed with are critical to the imagined outcome of their negotiations. In standard game theoretic analyses, the possible solutions (the “core”) to the problem of dividing social surplus are taken to include all allocations that give each player in the cooperative game a payoff at least as much as the greater of (1) what she could have secured through a non-cooperative strategy, and (2) what she could have achieved as a member of the most profitable coalition of players that could secede costlessly from the group to pursue their own cooperative strategy.³ Condition (1) requires that each person be at least as well off as she is in the SON. That is to say, the choice to exit the SON must be individually rational. ³ See Lewis Kornhauser, “Fair Division of Settlements: A Comment on Silver and Baker,” Virginia Law Review 84, no. 8 (Nov. 1998): pp. 1568–72; Hervé Moulin, Axioms of Cooperative Decision Making (Cambridge: Cambridge University Press, 1988), pp. 89–95. The “core” is the set of solutions in a coalitional game, such that no subset of players in the game can break away from the solution and pursue a joint strategy that leaves all members of the subset better off. The concept is thus analogous to the Nash equilibrium for non-cooperative games: a solution is stable, and hence in the core, if no deviation (by an individual or subgroup) is profitable. I discuss the application of the concept of the core to the SC bargain in more depth in Barbara H. Fried, “The Puzzling Case for Proportionate Taxation,” Chapman Law Review 2 (Spring 1999): pp. 177–9.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

“  ’  ,  ”

213

Condition (2) adds the further requirement of group rationality, restated in a slightly different form Nozick’s secession thought experiment described above: the best endowed will demand, at a minimum, that portion of the gains from social cooperation that they could have achieved through cooperation limited to themselves. But suppose the talented cannot arrange to secede costlessly. As a purely positive matter, to the extent that it is costly for the best endowed members of society (and by extension any other subgroup) to secede from the larger bargaining unit to form their more perfect union, the gains they could have captured for themselves in that more perfect union are expropriable by the less well-endowed majority.⁴ For positive political scientists modeling the likely outcome of the realworld tax game, such exit costs are obviously relevant, since they alter the credible threat point of any subgroup contemplating secession. But what is their normative relevance in SC arguments? Are they among the background conditions that ought to be treated realistically (and if so, with what consequences), or are they among those that ought to be idealized in some fashion (and if so, how)? In the first part of the chapter, I look at how exit options have been treated by different SC theorists. The survey is not meant to be (and most certainly is not) exhaustive. It should, however, suffice to demonstrate the extraordinarily wide range of solutions adopted by SC theorists across the political spectrum. Those differences, predictably, have led to radically different conclusions about the leverage the talented would possess to extract the full market value of their talents in the hypothetical SC bargain. At one extreme, SC theorists have deliberately modeled the SON on whatever (usually postindustrial democratic) society the SC thought experiment is invoked to justify or critique, and have suppressed almost none of the many constraints on exit that one would realistically find in such a society. Not surprisingly, the imagined outcome of that hypothetical bargain often looks not very different from what one would see by looking out the window. Some have gone even further, and imposed greater constraints on exit than would realistically arise—a move that locks the talented who are born into a heterogeneous society into an even worse bargain than they can actually secure for themselves in the real world. At the other extreme, many SC theorists have suppressed almost all external constraints on individuals’ ability to optimize their expected return in the SC bargain, given the natural endowments they bring to the table. Some have done so by constructing a bargaining scenario that is presumed “realistically” to imply no constraints on optimization. This approach is epitomized by the Crusoeian SON

⁴ Given that no administrable or politically acceptable tax base (income, consumption, wealth, etc.) directly taxes leisure, the rich have a less costly means of exit available to them, if tax rates get high enough: they can trade off productive labor for untaxed leisure. As a practical matter, that form of exit is likely to constrain tax rates on the wealthy far more than the more extreme sorts of exit I explore here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

214

“  ’  ,  ”

of the libertarian imagination—a world in which the potential parties to the SC are imagined to be free of any social entanglements, and hence free to join forces costlessly with the social cooperators of their dreams. Others have achieved it by constructing an SON that is taken “realistically” to imply significant constraints on exit, but bleaching out virtually all of those constraints on moral grounds. In either case, the predictable consequence of suppressing constraints on exit is to endow the talented with much greater bargaining power than they have in any existing society by giving them a more credible threat of self-segregation than they in fact possess. Notwithstanding the importance of exit options to the outcome of the hypothetical SC bargain, the choice of which exit options to endow bargainers with has received little scrutiny either from SC theorists or their critics. Although solutions to the problem of exit options vary widely in SC arguments, usual practice has been to suppress many realistically occurring constraints on exit. In the second part of the chapter, I ask whether that choice can be justified by the philosophical commitments of classic liberalism/libertarianism. I tentatively conclude the answer is no. The obvious justification for suppressing constraints on exit is that requiring individuals to internalize the costs of exit from existing social institutions is nothing more than a taunt: “If you don’t like it, leave it, if you can.” After all, the whole point of ideal bargaining theories is to imagine the deal that people would strike if they were starting afresh, fitted out only with endowments justly acquired. To require them to bear the toll charge, as it were, out of existing unjust social arrangements as a condition of any new deal seems so contrary to the spirit of the exercise as to be rejected out of hand.⁵ But on closer consideration, I will suggest, that obvious explanation faces some obvious difficulties. If one starts from the libertarian presupposition that the just state has no role to play in eliminating differences in fortune that it has not directly brought about, it seems more plausible to conclude that the SC bargain should take existing social arrangements and the costs of rearranging them as it finds them. In other words, from the libertarian perspective, such constraints are morally akin to the weather. Of course, if we require libertarian-minded theorists to treat exit options more realistically, we make it much more likely that the outcome of their SC thought experiments will simply legitimate what is. That result—turning libertarians into apologists for the status quo—may seem ironic, or at least counterintuitive. But, I suggest at the end, it is consistent with the parsimonious view of our collective responsibility to fix the exigencies of fate that animates classic liberalism and libertarianism.

⁵ For a statement to that effect, see David Gauthier, “David Hume, Contractarian,” The Philosophical Review 88, no. 1 (Jan. 1979): p. 12.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

      

215

1. Treatment of Exit Costs in SC Theories For current purposes, it is useful to distinguish between two broad groups of SC arguments: retrospective and prospective. Retrospective arguments start with an existing state, populated with individuals who look more or less like “us” (in most contemporary SC arguments, people who are citizens of a democratic, postindustrial state), and then ask whether subgroups of those citizens could, under some morally acceptable set of conditions, successfully demand to renegotiate the terms of their existing deal with the state. Prospective arguments, in contrast, start with some version of an SON, populated with individuals whose identity, personal attributes, and social position are in theory determined independently of any organized state, and ask what terms of cooperation those individuals would have agreed to in a morally acceptable bargaining process. The problem of constructing exit options comes up in both retrospective and prospective forms of SC arguments, but in a different posture. Prospective arguments, in theory, require that the author construct from scratch a morally acceptable set of exit options (including the existing non-agreement point, the SON, and all alternative rearrangements available starting from the SON). Retrospective arguments, in contrast, pose a more limited, or at least more focused, task: to assess whether the particular social constraints on exit built into existing arrangements are morally acceptable. The distinction I’ve drawn between prospective and retrospective arguments is crude at best, and does not map cleanly onto any existing SC arguments. The near impossibility of imagining a hypothetical SON that does not simply read back into it many of the features of the existing world means that all prospective SC arguments have built into them significant retrospective elements. Conversely, most retrospective SC arguments necessarily have built into them some significant prospective elements. Once the decision is made to replace the real features of the existing world with idealized ones, the task of fashioning the idealized ones in effect embeds an open-ended (i.e., prospective) constructivist project in the larger retrospective argument. At the same time, the distinction between prospective and retrospective SC arguments captures some important differences in the treatment of exit options. In particular, the retrospective approach has a built-in bias towards ratification of the features of an existing society, including exit options. This is so, because the challenge to existing arrangements posed by retrospective arguments is less demanding than that posed by prospective arguments. The former require defenders of existing arrangements to show the moral acceptability of such arrangements, rather than their moral superiority to all conceivable rival arrangements that might have arisen out of an SON constructed independently of such arrangements. In theory, that bias could be counteracted by judging the moral acceptability of

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

216

“  ’  ,  ”

existing arrangements stringently. In practice, however, retrospective SC theories tend to be much more deferential to existing arrangements (including constraints on exit) than prospective ones. I start with the treatment of exit options in retrospective arguments, and then turn to prospective ones.

1.1 Retrospective SC Theories The basic question posed by retrospective SC arguments is this: would people agree to stay put in the arrangements they find themselves in? All of the difficulties of SC arguments are, of course, housed in “would.” If the question translates to “have people in fact stayed put,” by definition the answer is yes with respect to those who have stayed put. With respect to those who have not, they have cured their problem through self-help. In either case, there is no work for SC arguments to do: every existing regime automatically vindicates itself by its survival. At the other extreme, if any constraint on exit embedded in existing social arrangements unfairly biases choice in favor of the status quo, one can never infer meaningful assent from the brute fact that people have stayed put.⁶ In either (extreme) view, then, retrospective social contractarianism is useless as a justificatory device— in the first case because its moral demands will always be met, in the second because they will never be met. Those who have put forth retrospective SC arguments have sought to avoid both extremes by arguing that some but not all real-world constraints on exit vitiate consent to the status quo. Within that spectrum, solutions have ranged widely, from versions that treat realistically almost all existing constraints on exit that would (for example) lead the rich in America circa 2020 to stay put, to those that suppress almost all existing constraints. Positioned at the realistic end of the spectrum is Jean Hampton’s Hobbes and the Social Contract Tradition.⁷ Hampton’s argument nicely illustrates both the bias towards ratification of existing arrangements built into retrospective SC arguments, and the generally ad hoc nature of the judgments about which constraints on exit to bleach out and which to treat realistically. While conceding that there will never be actual (historical) assent to existing arrangements, Hampton argues that citizens give their tacit consent to such arrangements by participating in the various social conventions that keep the ruler in power, however he came to be in power in the first place. Hampton tries to differentiate the moral underpinnings of this “social convention” version of tacit contractarianism ⁶ For a criticism of all tacit consent arguments on those grounds, see Gauthier, “David Hume,” p. 12. ⁷ Jean Hampton, Hobbes and the Social Contract Tradition (Cambridge: Cambridge University Press, 1986).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

      

217

from the version that rests consent on the mere brute fact that people have stayed put and accepted the benefits of the state.⁸ But the two seem to differ little in practice. In either case, tacit consent can be inferred, prima facie, from the fact that members have not successfully rebelled, and hence, happily or not, are accepting the benefits of and participating in the social conventions that uphold the existing state.⁹ Given the enormous practical costs of rebellion, if actual rebellion is necessary to defeat tacit consent to the ruler’s power, Hampton’s tacit consent legitimates virtually all states, however odious their practices. Hampton understandably is not prepared to go that far. Thus, she rules out as legitimate a state where citizens are “so technologically or psychologically mastered that they cannot mount effective resistance to the regime’s supporters.”¹⁰ While Hampton doesn’t explain this limitation, I take it to refer to intentional misinformation or coercive threats that prevent people from acting on the choices their informed selves would make. But Hampton appears to treat as legitimate all other real-life constraints on exit that would render resistance irrational in almost all cases. Thus, while acknowledging that rebellion will succeed only in extraordinary circumstances, she is nonetheless prepared to treat non-rebellion in the face of those prohibitive costs as tacit consent.¹¹ Hampton is similarly willing to treat the failure of dissidents to emigrate individually from an existing regime as tacit consent to that regime, even where the costs of emigrating are prohibitive. Thus, individuals in Hampton’s SC thought experiment find themselves in a position very close to the actual position they inhabit in the real world. Provided they are not choosing under threat of force or other compulsion, if, in the real world they inhabit, they choose to put up with the existing regime in preference to any of the alternatives actually available to them, they have legitimated the regime’s sovereignty over them. Clearly, that view seems to elevate the statement, “If you don’t like it, leave it,” from a taunt to a dispositive moral rejoinder. At the other extreme, consider the version of idealized exit options offered by James Buchanan in “The Ethical Limits of Taxation.”¹² Buchanan, harnessing the

⁸ Hampton, Hobbes and the Social Contract Tradition, pp. 276–9. ⁹ It is possible that Hampton means her test to differ from the more lenient version of tacit consent in one respect: that she would treat an attempt at rebellion, even if unsuccessful, as a withdrawal of tacit consent. If so, it is unclear how such a test would be applied in practice. What constitutes an attempted rebellion, such that we can no longer infer unanimous (tacit) consent? Is, for example, a failed attempt at tax protesting or other quasi-legal challenges to the system enough? Or is armed rebellion required? In any event, given the prohibitive costs of undertaking rebellion against an existing regime in any form, the difference between attempting rebellion (and failing) and not attempting it at all may matter little in practice. ¹⁰ Hampton, Hobbes and the Social Contract Tradition, pp. 279, 282. ¹¹ Hampton, Hobbes and the Social Contract Tradition, p. 282. ¹² James M. Buchanan, “The Ethical Limits of Taxation,” The Scandinavian Journal of Economics 86, no. 2 (Jun. 1984): pp. 102–14.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

218

“  ’  ,  ”

Rawlsian principle of “maximal equal liberty” to libertarian ends, argues that the principle limits the state’s legitimate taxing power, because one of the fundamental liberties the principle protects is the liberty to exit a polity internally. By internal exit, Buchanan means that individuals are permitted to stay put physically, but opt out of the rules of the polity within whose geographical borders they reside.¹³ Buchanan does not propose a formal opt-out option through juridical secession from one’s place of residence. Instead, he proposes requiring the state to cash out the value of that option indirectly, by guaranteeing that all citizens will retain at least as much of their income/wealth as they could have secured for themselves through formal internal secession.¹⁴ Seeking to hoist Rawls on his own petard, Buchanan concludes that because the principle of maximal (equal) liberty takes lexical priority over the Difference Principle in Rawls’s scheme, the constraint on redistributive taxation built into the former trumps the demands for a redistributive tax and transfer system built into the latter. It seems like a stretch, to put it charitably, to characterize Buchanan’s internal secessionist thought experiment as a form of Rawlsian liberty, given that Rawls himself barred exit of any sort through his stipulation of a closed society. (I discuss this aspect of Rawls’s theory below.) More importantly, it seems like a stretch for Buchanan the libertarian to embrace it. Simply ensuring that all individuals are legally free to leave if they wish seems more consistent with the negative definition of liberty at the core of libertarianism.¹⁵ The people in Buchanan’s thought experiment all have that formal right. Be that as it may, for the rest of the discussion I will cut Buchanan’s argument loose from the Rawlsian maximal liberty rubric, and treat it as a restatement of the conventional SC argument that people are entitled to an arrangement with the state that gives them, at a minimum, the value of the best alternative arrangement they could secure for themselves by seceding under specified conditions. Once that minimum is met, Buchanan argues, the state is free, at least as far as the ethical constraint of maximal liberty is concerned, to engage in whatever redistributive taxation it wishes.¹⁶ As with all SC arguments, the key question is: What conditions does Buchanan have in mind? Notwithstanding some ambiguous prose, Buchanan can’t mean the conditions that actually exist, for the same reason that Hampton can’t mean that tacit consent is given whenever individuals fail to rebel. What he has in mind is some form of “idealized internal exit option [that] places ethical limits on the

¹³ Buchanan focuses solely on “internal” exit options, ignoring external exit, the more conventional form of exit. His singular focus may simply reflect the belief that internal exit options are more valuable to SC bargainers than external ones, and hence represent the binding constraint on SC deals. “The Ethical Limits of Taxation,” p. 112. ¹⁴ Buchanan, “The Ethical Limits of Taxation,” p. 108. ¹⁵ Buchanan, “The Ethical Limits of Taxation,” p. 107. ¹⁶ Buchanan, “The Ethical Limits of Taxation,” pp. 109–10; James M. Buchanan, “The Gauthier Enterprise,” Social Philosophy and Policy 5, no. 2 (Spring 1988): p. 86.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

      

219

absolute level of taxation,” even when individuals lack the positive ability to enforce those limits on their own through “strategic threat behavior.”¹⁷ That raises two questions. What justifies requiring an idealized exit option? And how does Buchanan construct that option for the purpose of estimating its cash-out value? Buchanan says nothing on the first question, beyond asserting that maximal liberty is a good, and an (implicitly idealized) exit option is a form of liberty worth maximizing. One can, however, extract from his discussion a fairly clear answer to the second question. With one critical exception, Buchanan treats every factor influencing the value of exit options realistically, based on the real-world constraints found in the society whose practices are under scrutiny. Among those Buchanan treats realistically are: (1) the relative productivity of workers in the existing society and in different subcoalitions within it; (2) economies of scale, if any, in running a minimal state that produces the “good” of legal protection; (3) the size of the group contemplating exit; and (4) economies of scale for economic productivity “in the large,” due to factors such as specialization and division of labor.¹⁸ As Buchanan correctly notes, the answers to each of these questions in many cases will make secession of the talented economically infeasible.¹⁹ Now to the exception, the fifth factor that Buchanan considers: (5) “the potential barriers to political secession that would be imposed by the costs of organizing coalitions.”²⁰ Unlike the first four, Buchanan concludes these costs should not be treated realistically. Acknowledging that “[i]n any actual assessment of potential secession, these costs must clearly be reckoned,” Buchanan argues that for purposes of invoking the secession option “as an ethical principle for assessing levels of taxation, the organizational costs barrier is not necessarily relevant.”²¹ Buchanan doesn’t explain what “not necessarily relevant,” means, but I think a fair reading is that we are ethically required to disregard such costs in assessing the value of exit options. If so, Buchanan’s position amounts to saying that a person is entitled to the value she hypothetically could realize were she to wake up and find herself in an already functioning state, the citizens of which are drawn from the

¹⁷ Buchanan, “The Ethical Limits of Taxation,” pp. 108, 110 n.10. ¹⁸ Buchanan, “The Ethical Limits of Taxation,” pp. 109–12. ¹⁹ Briefly, the greater the relative productivity of workers in the existing society (due to jointness in productivity) and the greater the economies of scale in running the minimal state and for economic productivity “in the large,” the lower the value of exit options for any subcoalition, all other things being equal. ²⁰ Buchanan, “The Ethical Limits of Taxation,” p. 112. Buchanan doesn’t specify what costs he has in mind. At a minimum, they would surely include collective action costs such as costs of negotiation, hold-outs, and free-riders. It may be that the category should be read more broadly, to include all realworld transactional barriers that would prevent a rival society from coming into existence (e.g., the unavailability of land for a new society, sunk costs in the existing society, start-up costs in organizing a new complex society, etc.). ²¹ Buchanan, “The Ethical Limits of Taxation,” p. 112.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

220

“  ’  ,  ”

most talented subgroup in her own society who would have her for a member (Lucky Island). Why should the organizational costs of arranging exit, alone among all the social frictions that keep individuals put, be treated in an idealized rather than realistic fashion? And why (as I read Buchanan) are we morally required to suppress them entirely? Why, in short, should we give individuals the benefit of a bargain that is too costly for them ever to negotiate on their own behalf? I will return to these general questions in Section 2. But I note here that Buchanan’s insistence that exit costs ought to be suppressed in retrospective SC thought experiments is hard to square with his more famous, insistently real politique position about exit options in prospective SC arguments. In the latter case, Buchanan has argued forcefully that we must start with the actual status quo in the Hobbesian anarchistic SON—including, for example, slavery maintained through coercive private force—“because, quite simply, there is no other place from which to start.”²² Why not insist on the same social realism in constructing a retrospective SC argument? That is, why not treat each moment in an existing, unjust state as a Hobbesian SON, with any renegotiations going forward constrained by the real-world viability of exit from the status quo?

1.2 Prospective SC Theories As noted above, the question posed by retrospective SC theories is: would we stay put in our current arrangements, were we making that decision under existing conditions cleansed of certain morally impermissible features? In contrast, the question posed by prospective SC theories is: could we theoretically have gotten to our current arrangements by consent, starting at some point (the SON) that is determined independently of our current arrangements? As in retrospective agreements, how we construct the hypothetical alternatives available to the parties will substantially determine the outcomes. But the problem presents itself in a different and more complicated form in prospective agreements. Unlike retrospective SC arguments, which construct the initial bargaining situation starting with the social world as it exists, prospective SC arguments provide no determinate, historical situation from which to construct a realistic version of exit options. They require us to imagine from scratch the identity of the parties to the bargain, the social situation they find themselves in (and hence what their non-agreement point is), and the available alternatives to that non-agreement point. They require us, in short, to conjure an entire social world out of nothing. Are there any bounds to authorial inventiveness here? ²² Buchanan, “The Gauthier Enterprise,” p. 84. See also James M. Buchanan, The Limits of Liberty (Chicago: University of Chicago Press, 1975).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

      

221

With respect to human traits, one could at least suppose some traits (e.g., narrow rationality) to be universal enough that it would be “realistic” to import them into any bargaining situation. But it is hard to see how one could make the same claim with respect to the identity of the bargainers and the characteristics of the social world in which they find themselves. As a result, the SC theorist’s choice among the infinite array of theoretically possible bargaining scenarios is unconstrained by realism, making every aspect of the bargaining situation presented in prospective SC arguments idealized, in accordance with normative premises that are buried in the choice of hypothetical facts. This most certainly applies to the basic configuration of the SON, with whatever constraints on exit the chosen configuration is taken “realistically” to imply. Most contemporary (prospective) SC arguments draw from three basic variants on the SON: a neo-Lockean SON, in which the individuals negotiating the social contract live together in a stateless society characterized by some significant but still suboptimal degree of social cooperation; a neo-Hobbesian SON, in which the negotiators live in close physical proximity to each other, but in a dysfunctional society characterized by little or no industry or other social cooperation; and a Crusoeian SON, in which each would-be negotiator lives in his own isolated autarchy. I start by examining the treatment of exit options in the Hobbesian/ Lockean versions, and then turn to the Crusoeian version. In a nod to realism, Hobbesian and Lockean SC theorists take the bargainers to be whoever happens to be present in the (dys)functional SON from which the bargain is struck. But, of course, there is no real SON here; it is invented out of whole cloth. The threshold decision the author must make is whether the bargainers are heterogeneous in tastes and talents. If they are, all of the potential distributive conflicts that create a motive for the talented to secede are present, putting enormous pressure on how we resolve exit options. If they are not, the problem of exit simply disappears: the talented have no incentive to secede from their self-segregated SC group, and the untalented, who are not a part of the group, have no chance of persuading them to do so. Most Hobbesian and Lockean SC theorists have resolved this threshold question in favor of a significant degree of heterogeneity. In a move that obviously imports significant retrospective elements into the prospective exercise, authors generally assume that those who are present in the SON are whoever happens to be a citizen or member of “whatever actual State it is to which we wish to apply Hobbesian theory at a given time.”²³ The assumption of heterogeneous populations ²³ Gregory S. Kavka, Hobbesian Moral and Political Theory (Princeton, N.J.: Princeton University Press, 1986), p. 240. Thus, if we are testing whether prospective SC agreements would ever yield America circa 2020, we imagine a Hobbesian or Lockean SON populated with the citizens of America circa 2020. As Gauthier notes, it is difficult to defend this hybrid procedure, in which we import the retrospectively identified population into a prospective SON scenario, because “the division of persons into societies is itself a subject for assessment” under a prospective SC account. David Gauthier, “The

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

222

“  ’  ,  ”

immediately suggests the following two “realistic” features of the imagined Hobbesian/Lockean SC. First, the more talented would find it in their selfinterest to secede from the larger, preexisting bargaining group constituted by the SON, and, by credibly threatening to do so, could thereby extract a favorable deal for themselves if they are willing to stay put. Second, because the initial bargaining group throws the talented and less talented together, it builds in costs to secession that (all other things being equal) bias the outcome against it. Rather than counteracting that bias by suppressing some of the “realistic” costs of exit their bargainers would face, some prospective SC theorists have exacerbated it by prohibiting secession outright, or imposing prohibitive costs on it. Rawls, although hardly representative of self-interested bargain theorists more generally, is in this respect a case in point. Confronted with the “reality” that selfinterest would drive the more talented to secede from the larger bargaining unit, Rawls introduces two idealized features into the original bargaining position, either of which is sufficient to cut off secession. The first is the veil of ignorance, which deprives people of knowledge of their actual talents, thereby depriving the talented of any self-interested motive to secede. The second is the stipulation that the bargaining group is a closed system, from which secession during one’s lifetime is categorically barred.²⁴ Gregory Kavka, writing from a more conventional liberal perspective, reaches the same no-secession result by a different route. While stipulating that individual secession from the negotiating group must be permitted, Kavka rules out the formation of coalitions and agreements among individuals and subgroups prior to secession. Thus, subgroups are prohibited from making “binding agreements among themselves (except by adopting a social contract).”²⁵ But if they are unhappy enough about the direction that group-wide negotiations towards an SC are taking, they may secede one by one, and presumably, having done so, cast about for a more congenial alternative group of cooperators to negotiate with. What motivates Kavka’s peculiar compromise on exit options? The answer, as best as I can tell, is that it allows Kavka, contra Rawls, to give people in the original position knowledge of their own talents and a right to exit from the negotiating group, but simultaneously ensures that those entitlements will do the talented absolutely no good.²⁶ Thus, while acknowledging that once the talented know Social Contract: Individual Decision or Collective Bargain?,” in Foundations and Applications of Decision Theory, vol. 2, edited by C. A Hooker, J. J. Leach, and E. E McClennan (New York: Kluwer Academic Publishers (1978), p. 61. ²⁴ John Rawls, “Kantian Constructivism in Moral Theory: The Dewey Lectures,” Journal of Philosophy 77, no. 9 (Sept. 1980): pp. 515–72. ²⁵ Kavka, Hobbesian Theory, pp. 189–90. ²⁶ This is my reconstruction of the argument in Kavka, Hobbesian Theory, pp. 240–2. Kavka stops short of the “thick self” that the standard libertarian argument would insist on. Thus, while people are permitted to know their own talents, they are not permitted to know their social position in life. Kavka,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

      

223

their own talents, self-interest would lead them to pool their talents and march for the door, Kavka hastens to assure his readers that this will never actually come to pass, in part because of the very condition that Kavka himself imposes: participants will be forbidden from forming coalitions “from which secession groups would be most likely to arise.” Without such coalitions, Kavka argues, it will be much riskier for each individual to secede, since she must do so on a leap of faith that other well-endowed individuals will follow suit, making it much less likely she will try.²⁷ True enough. But the argument just pushes the justificatory problem one level down: why is Kavka permitted to insert this (idealized) restriction on coalitions, thereby making secession extremely unlikely? At a minimum, the choice surely reflects a lack of commitment to the “thick self” that Kavka and most other liberal- and libertarian-minded SC theorists have (in an anti-Rawlsian move) embraced. Now, suppose that, instead of assuming heterogeneity in the initial bargaining group, we assume that by historical accident or otherwise, the lucky and the unlucky have been physically and socially segregated from the start. To illustrate the radically different course that SC bargains are likely to take given this assumption, consider Gauthier’s parable of the purples and the greens.²⁸ Gauthier assumes the existence of two societies (the purples and the greens) that have developed, unbeknownst to each other, on two different islands. The purples have created an ideal Lockean/Gauthierian SON that maximizes their opportunities through market-optimal behavior and peaceable division of surplus in accordance with Gauthier’s principle of minimax relative concession, the details of which need not concern us here. The greens, in contrast, have squandered their opportunities in Hobbesian warfare and Malthusian procreation.²⁹ Thus, the difference in their situations is created by differences in chosen behavior. Nothing would fundamentally change in Gauthier’s argument, however, if the same difference was the result of differences in inborn talents rather than effort. Now suppose the purples set sail one day from their island and discover the beleaguered greens. The question that interests Gauthier is, what obligations, if any, do the purples have to help out the less fortunate greens? None, Gauthier concludes, because such obligations could arise only from a violation of the Lockean proviso. As there has been no past interaction between the purples and Hobbesian Theory, pp. 193–5. The motivation for this compromise seems to be similar to the motivation for Kavka’s odd compromise on exit options: he thereby creates the appearance of thick selves, but significantly limits the possibility that real differences among those selves will defeat unanimous agreement. See Kavka, Hobbesian Theory, p. 198. As with exit options, it is hard to discern any principled basis for this compromise. It seems rigged to produce the appearance of a real bargain among real people, without any of the unpleasant realities that would normally accompany it. ²⁷ Kavka, Hobbesian Theory, p. 241. ²⁸ David Gauthier, Morals by Agreement (Oxford: Clarendon Press, 1986). ²⁹ Gauthier, Morals by Agreement, p. 282.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

224

“  ’  ,  ”

the greens, such a violation is not possible.³⁰ Thus, Gauthier concludes, the purples may refuse to deal with the greens outright. If they do choose to deal with them, they may limit their interactions to one-off private contracts that they perceive to be to their own benefit. In particular, they need not agree to any SC bargain in which the greens would share in the social surplus generated by the purples.³¹ The question of interest here is, freed of any moral obligation to act other than in their self-interest, what would self-interested purples do? The answer seems clear. Beginning now from a position of self-segregation, the better-endowed purples can maintain the benefits of that position costlessly. In such a situation, one would expect (as Gauthier himself suggests) that the inequality in the position of the purples and the greens at the moment they discover each other “will be preserved in any rational cooperative arrangement among purples and greens.”³² Thus, however they structure their relations with the greens, the purples should be able to retain for themselves all the benefits of their greater endowments. To overcome that “realistic” outcome, we would have to impose on the purples an affirmative obligation of care that they are unwilling to assume voluntarily. No social contractarian would be willing to make that move directly; if she were, recourse to an SC argument to derive that moral obligation would be unnecessary. Few would be willing to make it indirectly either, by stripping the purples of knowledge of their superior position through some device akin to the Rawlsian veil of ignorance.³³ Note the embarrassments this outcome poses for arguments like Kavka’s. If the talented happen to find themselves thrown together with the less talented in the SON, Kavka is willing to rig the background conditions of negotiation to ensure that the talented can never self-segregate by secession. If instead the talented happen to find themselves segregated from the start on some Lucky Island populated only with their own kind, Kavka will not intervene to require the talented to expand their SC negotiating group to include the less fortunate. Why should so much turn on the imagined historical accident by which the SON takes

³⁰ Gauthier, Morals by Agreement, p. 283. This need not be the case under the conventional interpretation of the proviso. The absence of past interaction doesn’t guarantee that the purples have not violated the proviso. They could have violated it, for example, if the greens’ and purples’ differential wealth is owing to differential access to common and scarce natural resources. ³¹ Gauthier, Morals by Agreement, pp. 283–6. ³² Gauthier, Morals by Agreement, p. 283. ³³ I suppose this is as good a point as any to note the embarrassment that the parable of the greens and purples poses for someone like Rawls, who, by eschewing any first-order international obligations of justice (at least in his early work), is driven more or less to Gauthier’s conclusion in the parable of the purples and the greens. To put it another way, by adopting a closed SC system in which the bargaining units track existing nation-states and secession is barred, Rawls preserves not only whatever heterogeneity of talents and social position exists within nation-states; he also preserves whatever homogeneity exists, as compared to an SC in which all the world’s people are participants and the Rawlsian Difference Principle would operate to mandate worldwide redistribution of resources.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

      

225

shape? And given that the authors of prospective SC arguments can imagine that historical accident any way they wish, what constrains them from rigging the outcome by supposing the degree of homogeneity that will get them to the desired result? Crusoeian SON arguments have achieved precisely this result—rigging the SON to ensure the self-segregation of the talented—by slightly different means. Rather than imagining a Hobbesian or Lockean SON made up only of the talented, Crusoeian arguments imagine an SON devoid of any preexisting social arrangements. To see the structurally parallel roles the two versions of the SON play in guaranteeing self-segregation of the talented, consider the following. Assume that we have two groups of people, the Luckies (the more talented) and the Unluckies (the less talented). Assume further that a minimum of four people is required for the efficient production of private and public goods, along with other harder to quantify social advantages from cooperation. Any number less than four will leave each individual member of society no better off with cooperation than without. Now consider two variants, the first the classic Crusoeian bargaining scenario, the second the homogeneous Hobbesian/Lockean SON: (i) Assume we start with eight Crusoes on eight different islands, four Luckies and four Unluckies. The eight simultaneously discover each other one day out on the high seas. Assume that their respective levels of talent are transparent to all, and that there are no practical constraints on their ability to regroup themselves in any fashion they wish. Intuiting that economies of scale require a minimum of four people for optimal well-being, the four Luckies band together and take up residence on newly discovered Lucky Island. The four Unluckies are thereby left to fend for themselves on Unlucky Island, and strike whatever arm’s length trades they can with the Luckies. This is our classic Crusoeian SC tale. Structurally, the role the island metaphor plays in the tale is to ensure that there is no preexisting bargaining group from which the talented must extricate themselves by secession. Free to choose without impediment those with whom they will bargain, the talented will naturally choose each other. (ii) Assume now that the four Luckies start out together on Lucky Island, a happy coincidence of birth having deposited them all there. Over time, they informally band together in social cooperation. When they have just begun to bargain over an SC for their four-person society, they get wind of the fact that there is an Unlucky Island nearby on which four Unluckies, through a somewhat less happy coincidence, have started out together. Hearing that the Unluckies are also in the process of negotiating their own SC, the Luckies consider, and immediately reject as contrary to their self-interest, the possibility of merging their two bargaining units, or regrouping in some other combination of fours that will mix the Luckies and the Unluckies.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

226

“  ’  ,  ”

This, of course, is the case of the homogeneous SONs posed by Gauthier’s parable of the purples and greens. As noted above, Gauthier concludes that, as in case (i), the Luckies owe the Unluckies nothing. Note what is going on here. In both cases, the Luckies start out in an imaginary original position from which they can achieve their optimal alliance, that is, they can end up together on Lucky Island, at no incremental transactions costs over any alternative arrangement. Being rational creatures, they will naturally choose self-segregation on Lucky Island over those other alternatives. In both cases, the natural libertarian response is to conclude that it is wholly within the moral prerogative of the Luckies to do so, subject at most to the obligations imposed by the Lockean proviso to correct any unequal division of natural resources between Lucky and Unlucky Islands. If the Crusoeian version has any rhetorical advantage over the homogeneous SON society that would explain its greater appeal to the libertarian imagination, it probably lies in two things. First, the Crusoeian version underscores and reinforces the image of the individual as prior to society, making society rather than being made by it. Second, it better obscures the author’s role in producing selfsegregation of the talented as the adventitious outcome of self-interested bargaining. If one is told that the SON is composed of a number of self-contained defense groups that will each be negotiating towards its own SC and that those groups happen to be perfectly stratified by talents, one is likely to respond, says who? If, on the other hand, one is told that in an SON composed of a number of isolated Crusoes, the talented Crusoes have sought each other out as bargaining partners, one is likely to attribute that outcome to the rational agency of the fictional Crusoes, rather than to authorial fiat. But does the Crusoeian metaphor even deliver what it promises: the perfect selfsegregation of the talented? I think not, for the following reason. The ability of the Luckies to achieve perfect self-segregation depends upon two things: that there exists a moment in time in which the Luckies face no barriers to self-segregation, and that the perfect (optimal) set of co-citizens presents itself at that moment in time. The Crusoeian metaphor supplies the first of these things, by supposing that the formal state is created by the same bargain that creates social cooperation of any sort. But it does not supply the second. To see why, consider the following variant on case (i): (iii) Assume, as in case (i), that eight different Crusoes independently go out in search of compatriots. But this time, only four find each other, and the four happen to include two Luckies and two Unluckies. Recognizing that a four-person society that includes two Unluckies is better for them than any available alternative, the two Luckies happily join forces with the two Unluckies on Demos Island, and the four inhabitants strike a deal among themselves for a fairly egalitarian division of the surplus generated by their social cooperation. In time, the two stray Luckies who remained stranded on the high seas are washed up on Demos Island. The two Luckies who are already on Demos

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

      

227

immediately realize that they would be much better off if they could ditch the two Unluckies they are saddled with, team up with the two Luckies who just washed up, and head off with them for Lucky Island. But, by that time, doing so would be prohibitively costly. Now, unlike cases (i) and (ii), there was never a moment in time in which the two Luckies on Demos could feasibly have segregated themselves on Lucky Island, since by the time they discovered the two other Luckies, Lucky Island was, by hypothesis, no longer a socially available alternative for them. In short, while the Luckies started out with no social encumbrances that would have prevented them from optimizing their choice of co-citizens, they have acquired such encumbrances through voluntary choices that were, at the time made, completely rational. Now what? The thrust of Buchanan’s argument in the retrospective context is that the two Luckies stuck on Demos Island are nonetheless entitled to renegotiate with their two Unlucky compatriots and demand the share of the joint surplus they would have gotten had they somehow made it costlessly to Lucky Island. It is unclear how far this argument reaches. If taken to the extreme, it suggests that whenever one of the talented can name some hypothetical social arrangement (Lucky Island) that would have been more advantageous to her than the one she is stuck in, she gets to invoke that more advantageous arrangement in setting her threat point in the idealized bargaining game. Setting aside the mind-boggling practical problems of administering a perpetual, free option of this sort, it seems extraordinary to embrace it in principle. It is also hard to reconcile the moral case for such an option with deeply held libertarian beliefs that people must live with the consequences of their and their predecessors’ path-dependent choices. Where does this brief excursion through exit options leave us? It suggests, above all, that where SC theorists locate themselves on the continuum between pure realism and Buchanan’s (idealized) frictionless secession to Lucky Island is more the product of moral ad hocery than anything else. This is true of the threshold decisions about the social unit in which to locate the bargainers (an actual contemporary society circumscribed by existing national borders, a Hobbesian or Lockean pre-state SON, a Crusoeian SON, and so forth) and the identity of the parties to the bargain. Those decisions, as I suggested above, have enormous influence on what “realistic” constraints on exit the parties are imagined to face. It is true as well of the subsequent decisions about which of those constraints to treat realistically and which to suppress. The moral ad hocery at work in deciding which aspects of the SC bargaining scenario to treat “realistically” and which to moralize is exacerbated by the empirical ad hocery involved in stipulating what realism implies—whether, for example, economies of scale or political unrest would lead the talented to conclude that exit is not in their self-interest, or altruism would lead them to care for the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

228

“  ’  ,  ”

least fortunate of the world, notwithstanding that pure self-interest argues for exit.³⁴ Given the counterfactual nature of SC exercises and the infinite number of obstacles to secession one could imagine arising in any scenario, such ad hocery is perhaps unavoidable. Unavoidable or not, however, it casts considerable doubt on whether the hypothetical outcome of an SC bargain can ever be anything more than an elaborately tricked-out version of the inputs supplied by authorial fiat. This is hardly a new criticism of SC arguments. As I discussed in Chapter 9, the moment that Rawls invoked his veil of ignorance, thereby depriving SC bargainers of any information about their personal identities, he guaranteed that the resulting social choice would optimize aggregate welfare, in accordance with his stipulated welfare function.³⁵ It is also not a fatal criticism. But it does mean that most if not all of the action in SC arguments lies in authorial choices to enlarge or constrict the opportunities that the talented, knowing their talents, have to selfsegregate in the world in which they actually reside. Those choices cannot justify the outcomes they produce, as they require justification themselves.

2. The Case for Suppressing Social Constraints on Exit To what conclusions would a more thoughtful consideration of exit options lead? I turn to that question now, focusing on the mainstream liberal/libertarianminded SC tradition that seeks to model self-interested bargains among relatively thick selves. As suggested above, while such theorists have ranged widely in their treatment of exit options, the dominant tendency is to suppress at least some of the costs that the talented would realistically face if they tried to secede, and by suppressing them, increase the value of their threat to do so. That move raises two questions. Why should we give the talented a better deal in this respect than they could negotiate in the world in which they actually find themselves? And, if we are going to give them a better deal than they could get for themselves, how much better and in what ways? In considering the first question, it may be helpful to situate the problem of social constraints on exit in the context of two other exogenous factors that influence the threat value of exit options: (1) the constraints of the market; and (2) the constraints on exit imposed by law. Most libertarians regard the market as the paradigmatic example of constraints that should be treated realistically. Viewing the market as an impersonal, and hence morally neutral, institution, akin to the weather or other external ³⁴ For Kavka’s empirical hunch that all of these factors would push against exit of the talented, see Hobbesian Theory, p. 241. ³⁵ John C. Harsanyi, “Cardinal Utility in Welfare Economics and in the Theory of Risk-Taking,” Journal of Political Economy 61, no. 5 (Oct. 1953): p. 434; Brian Barry, Theories of Justice: A Treatise on Social Justice, vol. 1 (Berkeley: University of California Press, 1989), pp. 334–5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

       

229

circumstances that Crusoe might encounter on his desert island, they conclude that the market value of private endowments should be attributed not to society but to those endowments. This view of the market as an impersonal force majeure underlies the libertarian argument that people are entitled to the market value of their labor. As David Gauthier put it, while our marginal product depends on supply, demand, and other exogenous factors for which we are not causally responsible, those constraints are the interpersonal analogues of those experienced by Robinson Crusoe alone on her island. If her freedom is not constrained by her limited talents and particular interests, then it remains unconstrained when she leaves her island and comes to relate her talents and interests with those of others.³⁶

In the context of SC arguments, that view leads straightforwardly to Buchanan’s decision to treat “realistically” the various factors that affect the market price of private and public goods: economies of scale in private production and provision of government services, the relative productivity of workers in different subgroups, and the relative sizes of subcoalitions. At first cut, constraints on exit imposed by law are, for classic liberals and libertarians, an equally easy case on the other side (that is, for idealized treatment). As Kavka puts it, the implicit aim of any SC exercise is to “understand[] the State as an artifact created and chosen by independent individuals.”³⁷ This requires that our choice not be “essentially ‘rigged’ ” by a political society that creates in us the very reason we use to choose it and that appears to justify its existence.”³⁸ It is easy to see why SC theorists would take this to require that we suppress any unjustifiable constraints on exit imposed by the state, the justice of whose arrangements is being tested by the SC thought experiment. Without such a requirement, we would simply reenact in the SC bargain the paradox of all coercive threats: the more repressive a society, the more likely its members are to “choose” to stay put. The class of unjustifiable state-imposed constraints on exit would surely include any outright prohibition on exit. It would include as well any penalties on exit, the sole motive for which is to indirectly compel people to stay. Once we get beyond these easy cases, however, there is a lot of trouble lurking in the word “unjustifiable” for autonomy-minded SC theorists—no less trouble, perhaps, than in the context of the social (nongovernmental) constraints on exit that are the subject of

³⁶ Gauthier, Morals by Agreement, p. 91. Gauthier’s defense of market-based distribution as a byproduct of personal liberty echoes Hayek’s defense discussed in Chapter 11. ³⁷ Kavka, Hobbesian Theory, p. 200. ³⁸ Hampton, Hobbes and the Social Contract, p. 271.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

230

“  ’  ,  ”

this chapter. For present purposes, however, I want to focus on the easy cases, taking them as a paradigmatic case for moralized treatment. The laissez-faire view that the market and the state occupy distinct spheres, and that the market should be regarded as an impersonal force majeure morally equivalent to the weather, is controversial, and I don’t mean to endorse it here.³⁹ But one can at least understand the moral intuitions on which the view rests. The question is, starting from those intuitions, do the socially imposed costs of exit that the talented realistically face in arranging self-segregation look (in morally relevant ways) more like state-imposed constraints on exit or more like the constraints of the market? I want to suggest that there is a pretty good argument for the latter. Like the market, the sorts of social costs of exit I am concerned with here are not imposed by the state. They inhere in the nature of human beings, the difficulties of collective action, and the path dependence of all social arrangements. They arise whenever human beings are considering altering existing social arrangements, whatever those arrangements. Why should we not regard such costs as a natural feature of social life, and, hence, like the natural constraints experienced by Crusoe alone on his desert island, something the parties to the SC bargain have no moral claim to have suppressed? The best (or at least most frequent) explanation why not is that such social constraints on exit, like state-imposed constraints, inherently tilt the outcome of the SC thought experiment in favor of the status quo.⁴⁰ As a factual matter, the claim is obviously true. The greater the number of historical features of a particular society we replicate in the hypothetical SC bargain, the more likely we are to produce a bargaining outcome that looks a lot like that society. The question is, why is that a problem? Why are we morally required to depart from social realism in any particular case? I don’t propose to provide a comprehensive answer here. But I want to suggest why, with respect to the social costs of exit, the justification for departing from realism is far from obvious. First, unlike the legal constraints on exit imposed by an unjust state, social constraints on exit can arise by a process that is (from a libertarian perspective) morally neutral. This is of course the point of Crusoeian scenario (iii) that I posed at the end of Section 1. In that scenario, by the time two Luckies discover others like themselves, they are already voluntarily enmeshed in a heterogeneous society ³⁹ For the historical argument that the market is inextricably entwined with the state, see Barbara H. Fried, The Progressive Assault on Laissez Faire: Robert Hale and the First Law and Economics Movement (Cambridge, Mass.: Harvard University Press, 1998), ch. 3. ⁴⁰ See, for example, David Gauthier, “Bargaining and Justice,” in Ethics and Economics, edited by Ellen Frankel Paul, Jeffrey Paul, and Fred D. Miller, Jr. (Oxford: B. Blackwell for the Social Philosophy and Policy Center, 1985); Gauthier, “David Hume,” p. 12: arguing that only an autarkic SON can abstract away from existing social arrangements and thus assure the “ex ante character of the agreement.”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

       

231

(Demos Island) from which they cannot extricate themselves except at a prohibitive cost. As I suggested in the discussion above, the burden of Buchanan’s argument is that, although the Luckies stuck on Demos Island cannot actually make it to Lucky Island, they are nonetheless entitled to renegotiate their deal with their Unlucky compatriots on Demos Island to extract the benefit of the hypothetical deal they could have gotten on Lucky Island had they only been able to get there. The moral basis for this claim seems doubtful, even from a libertarian perspective. To see why, consider the following reasons why the two Luckies who have taken up residence on Demos might find themselves unable to exit and head for Lucky Island: (a) It costs too much to travel to Lucky Island, the only available place to set up their more perfect union. (b) It costs too much for the four Luckies to negotiate the terms of any new arrangement on Lucky Island and coordinate their actions. (c) The two Luckies on Demos Island have high sunk investments there that they cannot recoup if they leave for Lucky Island. Suppose, for example, that the two Luckies and the two Unluckies that founded Demos built water and sewage systems at considerable expense. (Or in place of physical infrastructure, substitute social infrastructure, in the form of friends and family, language, culture, business connections, and so forth.)⁴¹ The Luckies can’t physically move their share of the Demos infrastructure to Lucky Island, and the Unlucky residents of Demos refuse to buy out their share. As a result, if the Luckies move to Lucky Island, they will have to build a new infrastructure from scratch, at a cost that will wipe out any gains they could otherwise realize by trading up in compatriots. (d) The Luckies can’t leave Demos Island because they promised the Unluckies they wouldn’t, at a time when they didn’t know that a more attractive alternative would later present itself. In all four cases, the costs of exit result from circumstances that appear (from a libertarian point of view) morally neutral. Given that fact, it is hard to see why anyone of a vaguely libertarian stripe would think that the Luckies already on Demos Island have a right to renegotiate their agreement with their Unlucky compatriots to extract the share of the social surplus they would have enjoyed on Lucky Island had they only been able to get there costlessly, while they simultaneously insist that the Unluckies have no right to demand the larger share they could have bargained for were their talents only greater.

⁴¹ I return to the problem of social infrastructure in Chapter 13.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

232

“  ’  ,  ”

While it is ludicrous to map the development of real societies onto any of these thin, fictive models, it seems fair to suggest that the history of most contemporary societies looks a lot more like scenario (iii) in Section 1 than the spontaneous, costless self-segregation of the talented presented in scenarios (i) and (ii). Societies develop over time in a haphazard fashion. Alternatives present themselves when we are no longer in a position to capitalize on them, having already embarked on a path-dependent course that it is costly or impossible to turn back from. This describes not only the process by which any real-world society comes into existence, but also, by definition, the position that all members of subsequent (post-founding) generations find themselves in. Whatever ideal state we might imagine forming were we born citizens only of the world, with no ties to anyone or anything, it is not the one we choose, born (as we all are) in a particular time and place, and held to that place by ties of family, friends, language, work, familiarity, and so on. That fact, of course, pushes most people to want to stay put, thereby making the surplus value they enjoy from doing so expropriable by their fellow citizens. This may be regrettable from the perspective of those who could have made out like bandits had they been able to make it to Lucky Island at a reasonable cost. But why do those regrets have a greater command on our moral attention than those occasioned by bad weather, or low demand for our talents, or any other exogenous factors in our lives that we are required (in the libertarian view of things) to take as we find them? Second, even were it morally desirable to eliminate the constraints that existing social arrangements place on choice, how could we ever do so? The libertarian objection to imagining the parties to an SC to be embedded in any particular set of social arrangements is clear: whatever such arrangements are imagined to be, they will bias the outcome of SC bargaining in the direction of that status quo, and the particular direction will be the product of historical accident. But the Crusoeian solution adopted by Gauthier and others, which posits no social interaction in the SON, doesn’t solve that problem. It merely substitutes a different set of fictional arrangements, ones that are no less an expression of (imaginary) historical accident than any other. Indeed, lacking any warrant in reality, they might be thought to be somewhat more. And rather than preserving the ex ante character of the agreement that ensues, it merely biases the outcome in a different direction— in the classic Crusoeian SC scenario, towards self-segregation of the talented, rather than some modestly redistributive SC contract among a more heterogeneous set of SC bargainers whose lots were thrown together by chance. If the Unluckies who would have benefitted from that modestly redistributive contract were heard to complain that by insisting on a “no interaction” baseline, Crusoeians have simply stolen from them their (realistic) bargaining advantage, what could Crusoeians say in response? The only possible answer is that we are, for some reason or another, morally required to assume a baseline of no interaction. As the preceding comments suggest, the reason why cannot come from the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

       

233

requirement of moral neutrality among different, historically contingent, alternatives. It surely cannot come from the demands of realism. Where then does it come from? Finally, even if one could somehow come up with a persuasive moral justification for enlarging the exit options of the talented beyond those realistically available to them in any existing society, how far should we enlarge them? Buchanan’s secessionist thought experiment in effect demands that we give the talented the benefit of the cooperative arrangement that they could have reached if only they had been able to secede costlessly with the optimal subcoalition of current members of their society. But why stop there? Why not give the talented the benefit of the even better deal they could have secured by assembling an ideal set of compatriots culled from all existing societies, to band together in a utopia of the talented on Lucky Island Internationale? However we define the community that is party to the on-stage negotiations, an infinite number of outsiders wait in the wings to join with defectors in alternative coalitions. Why treat these potential, extraterritorial coalitions less generously than internal ones? Or for that matter, why stop at people who happen to be alive right now? If we are engaging in fantasy baseball here, why not let the most talented of any generation name their hypothetical alternative compatriots from among the universe of people who at one time existed, or who could be imagined to exist? The obvious reason why not—because people that don’t now exist can’t actually join them on Lucky Island—seems less obvious when one realizes that the other talented folks who do exist can’t actually join them on Lucky Island either, due to the prohibitive costs of exit. Why is one social impossibility dispositive, and the other irrelevant? Where does that leave us? First, given the determinative role that exit options play in shaping the imagined social contract that issues from SC bargaining, the task of specifying and justifying the exit options with which bargainers are endowed deserves more attention than it has gotten to date in the SC literature. Second, it is far from clear where a serious inquiry into the problem of exit options would end up. For SC arguments in the Rawlsian vein, in which bargainers are deliberately stripped of all exit options and deprived of any other means to capture the value of their superior talents, the question is irrelevant. Rawls’s deference to national borders, it is true, preserves to some uncertain extent the existing economic selfsegregation of the privileged in the first world, as many friendly critics of Rawls have noted. But that result is presumably the unwanted consequence of Rawls’s other commitments (to political pragmatism, to the virtues of stability and consensus promoted by taking existing political borders as given) that pull against the universal obligations of altruism that clearly seem to be implied by the Rawlsian veil. Whatever the appropriate way to balance the competing demands of global distributive justice and nationalism in Rawlsian-type schemes, the questions raised here shed no light on the answer.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

234

“  ’  ,  ”

The bargaining models that dominate the liberal/libertarian SC tradition are another matter. All start with the presupposition that, within certain limits, we are entitled to pursue the self-interest of our relatively thick selves in whatever SC bargain we strike. The leverage that each of us has to extract concessions from the other parties to the bargain thus goes to the heart of the enterprise, and the exit options with which each of us is endowed are one of the most important determinants of that leverage. The burden of my argument here is that SC theorists in the liberal/libertarian tradition have failed to establish what fundamental principles should be appealed to in setting exit options, let alone what social arrangements are thereby implied. If one starts from the liberal/libertarian premise that the just state has only a minimal role to play in eliminating those differences in fortune that it has not directly brought about, one is pushed toward the conclusion that a just SC bargain should take existing social arrangements and the intrinsic difficulties of rearrangement as it finds them. One is pushed toward the conclusion, that is, that such constraints are morally akin to the weather. In that view, “If you don’t like it, leave it” is a perfectly acceptable response to the privileged members of a society subject to a redistributive scheme they don’t like, but for whom exit is not a socially realistic option. As noted above, by pushing the SC thought experiment back toward the realistic end of the spectrum in constructing exit options, we make it much more likely that the outcome of that thought experiment will simply mirror existing institutions. That seems the clear moral of Jean Hampton’s version of social contractarianism. By taking social arrangements and constraints on exit largely as given, Hampton ends up implicitly legitimating what is. That result, which turns libertarian-minded SCs into apologists for the status quo, may seem ironic, or at least counterintuitive. But in fact it follows directly from the relatively thick view of selves, and parsimonious view of our collective responsibility to fix the exigencies of fortune that befall those selves, that animates classic liberalism and libertarianism. If instead we take the self-interested bargaining model at the heart of libertarian SC arguments to imply an obligation to suppress the path-dependent social arrangements in which people find themselves, I see no principled basis for choosing the hypothetical arrangements to replace them. None of the alternatives proposed seems morally more compelling than reality. And some—notably Buchanan’s perpetual free option on Lucky Island—seem considerably less. Unless liberal- and libertarian-minded SC theorists can justify the de facto configuration of exit options implicitly or explicitly driving the hypothetical bargain, the resulting social contract, to borrow Nozick’s waggish observation, is probably not worth the paper on which it is not written.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

13 The Case for a Progressive Benefits Tax In Chapter 12, I took up social contractarian arguments against redistributive taxation. In this chapter, I take up the question of what a nonredistributive tax system would look like. The version that has garnered the most support from opponents of redistribution is some form of benefits taxation, in which individual tax rates are pegged to the shadow market price of the state-provided goods and services the taxpayer consumes. The central question confronting proponents of benefits taxation is, which market? A perfectly competitive market in which goods and services are priced at their marginal cost of production? A quasi-monopolistic market in which the supplier has the ability to price-discriminate among customers based on their willingness to pay? Some third alternative? Depending upon the answer, a benefits tax will produce very different distributions of the tax burden. Benefits tax proponents have implicitly assumed that the appropriate model is a perfectly competitive market for explicit public goods, an assumption that is likely to yield the flat or regressive rate structure that benefits tax proponents generally support. I argue in this chapter that that choice, like the choice among exit options explored in Chapter 12, does not follow from libertarian principles, and is hard to square with libertarians’ laissez-faire approach to private markets.

1. The State as a Shadow Market Most libertarians, contra Nozick, accept that compulsory taxation is necessary to fund the minimal state. It is necessary because many of the goods or services the minimal state would provide fall into the technical category of public goods, one of the defining properties of which is that it is technologically infeasible to bar free-riders from enjoying them (“nonexcludibility”). As a result, such goods will always be undersupplied by the private market.¹ Only the government can solve ¹ Paul A. Samuelson, “The Pure Theory of Public Expenditure,” Review of Economics and Statistics 36, no. 4 (Nov. 1954): pp. 387–9. The standard definition of public goods adopted by Samuelson requires that they exhibit not only nonexcludibility but also nonrivalrous consumption. Samuelson, “The Pure Theory of Public Expenditure,” p. 387. Noting that the latter requirement would exclude highways, police protection, and many other traditional public expenditures, Charles Tiebout and other economists opted for the broader definition that requires only nonexcludibility. Facing Up to Scarcity: The Logic and Limits of Nonconsequentialist Thought. Barbara H. Fried, Oxford University Press (2020). © Barbara H. Fried. DOI: 10.1093/oso/9780198847878.001.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

236

      

the free-rider problem, because only the government has the authority to compel payment (via taxation) to finance provision of such goods. That argument leaves open how the tax burden should be distributed among citizens. Since the early part of the twentieth century, the prevailing view has been that it should be distributed in accordance with ability to pay, a principle that is generally taken to imply that tax rates should rise with income (a so-called progressive rate structure).² A benefits tax, in contrast, pegs individual tax rates to the benefits each taxpayer receives from the state, irrespective of the taxpayer’s income or wealth. As Richard Musgrave and others have noted, the latter view implies a quid pro quo relationship between the taxpayer and the state, in which each individual’s tax burden functions as the shadow price for the public goods she voluntarily consumes.³ While benefits tax proponents typically focus only on the tax (revenue) side of the fiscal relationship between the state and its citizens, a companion proposal (“tax hypothecation”) has been made on the expenditure side: tax revenues should be allocated among government goods and services in proportion to taxpayers’ willingness to pay for them. So, John, who sends his children to private school in California, may not be taxed on any portion of the cost of public education in California because he derives no benefit from it (the benefits tax argument). California in turn can spend only as much on public education as its individual citizens collectively would agree to pay for it (the tax hypothecation argument). That a state’s fiscal policy should be driven by the preferences of its citizens is hardly a revolutionary thought. In a well-functioning democracy, that is how fiscal policy—indeed, almost all policy—is made. The difference is that democratic procedures aggregate individual preferences through voting, and go with the plurality, majority, or sometimes supermajority choice. In contrast, under an ideal benefits tax, each citizen would get the precise package of public goods and taxes to pay for them that he prefers, and the government would in turn tie production of those goods and services to the aggregate amount that citizens are willing to pay for them. All constitutional democracies take certain categories of decisions out of the hands of the voters and their elected officials and turn them over to the unilateral control of individuals most likely to be affected by the outcome. There are a host of reasons why this might be thought a good idea, but most of them fall into one of two camps: because decentralized decision-making will lead to better outcomes for society at large (the welfarist argument), or because it will give each citizen

² See Barbara H. Fried, The Progressive Assault on Laissez Faire: Robert Hale and the First Law and Economics Movement (Cambridge, Mass.: Harvard University Press, 1998), ch. 4. ³ Richard A. Musgrave, The Theory of Public Finance (New York: McGraw-Hill, 1959).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

       - 

237

control over collective decisions that trench on her personal rights (the rights-astrumps argument). In the arena of fiscal policy, both reasons have been marshaled to support replacing a majoritarian decision rule with some mechanism that tracks preferences at the individual level (hereinafter, “devolution”). The different reasons, however, point to different uses of that information. Public finance theorists have focused on the role individual-level information can play in determining optimal output levels for public goods. In contrast, benefits tax proponents on the right have focused on its role in securing the just distribution of whatever aggregate tax burden is necessary to finance public goods. I start by summarizing the welfarist argument, as it is an important backdrop to the distributive issue. As to the distributive issue, I conclude the following: (a) Determining a just price for public goods runs into many of the same difficulties that arise in determining a just price in private exchanges, and for the same reasons. While “just price” suggests a search for a price that is substantively fair, in the context of private markets libertarians (and many others) adopt a purely procedural answer: a just price is whatever price the parties to the contract voluntarily agreed to.⁴ (b) As I suggested in Chapter 12, in evaluating just price in the context of public goods and services, most libertarians go with a robust, substantive interpretation of “just”. Voluntary consent to the taxes-for-public goods deal on offer in a particular jurisdiction is insufficient to make the resulting price just. The terms must conform to those that would be produced by an idealized, competitive market. (e) Whether that inconsistency is problematic depends upon whether private and public markets are analogous in morally relevant ways. I will suggest they are—or at least are much more alike than benefit tax proponents assume.

2. The Welfarist Case for the State as Market-Mimicking Actor From a welfarist perspective, the government’s advantage in solving the free-rider problem on the revenue side does not obviate the informational advantages of having the private market set output levels for public goods. The ideal solution from a welfarist point of view would be for the government to mimic the market in setting output levels for public goods, and then use its compulsory taxing powers to extract the funds needed to pay for them, in whatever fashion minimizes the welfare costs of that burden on individual taxpayers.⁵ ⁴ For Robert Nozick’s version of the argument, see Chapter 11. ⁵ Steven D. Schwinn, “Toward a More Expansive Welfare Devolution Debate,” Lewis and Clark Law Review 9, no. 2 (Summer 2005): pp. 311–46.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

238

      

The former task, however, requires a mechanism in lieu of market exchange to force citizens to reveal the prices they would be willing to pay for different levels of public goods on offer. Starting with the pioneering work of Richard Musgrave and Paul Samuelson in the 1950s, public finance theorists generally assumed that no such mechanism could exist because of the same free-rider problem that forced the government into the business of supplying public goods in the first place. If citizens thought that their tax rates would be pegged to their consumption preferences over public goods, they would understate their true preferences, since—however little they contributed—they could not be excluded from unlimited consumption of whatever true public goods were produced. Lacking any mechanism to force taxpayers to reveal their true preferences, governments had to rely instead on the imperfect mechanism of politics. Politics, it was conventionally assumed, would produce answers that reflect the preferences of the median voter, resulting in half the population’s ending up with more public services than it would have voluntarily paid for and the other half fewer.⁶ That conventional wisdom was upended by Charles Tiebout’s groundbreaking article, “A Pure Theory of Local Expenditures.”⁷ Largely unnoticed when it was first published in 1956, the article became a focal point of academic debate after it was resurrected by Wallace Oates in 1969.⁸ The article was, in a sense, a friendly amendment to the conventional view that governments had no way to ascertain efficient output levels for public goods. While leaving that view unchallenged at the national level, Tiebout argued that a market mechanism was available and routinely employed at the local level to elicit that information. Tiebout’s basic insight was simple—so simple, really, that in retrospect it is hard to see how it could have been overlooked. He argued, in effect, that economists ⁶ The standard median-voter model was put forth by Theodore C. Bergstrom and Robert P. Goodman, “Private Demands for Public Goods,” American Economic Review 63, no. 3 (Jun. 1973) pp. 280–96. For a literature review, see David E. Wildasin, Urban Public Finance (New York: Harwood Academic Publishers, 1986). ⁷ Charles M. Tiebout, “A Pure Theory of Local Expenditures,” Journal of Political Economy 64, no. 5 (Oct. 1956): pp. 416–24. ⁸ Wallace E. Oates, “The Effects of Property Taxes and Local Public Spending on Property Values: An Empirical Study of Tax Capitalization and the Tiebout Hypothesis,” Journal of Political Economy 77, no. 6 (Nov.–Dec. 1969): pp. 957–71. The failure of the Tieboutian argument to penetrate academic consciousness prior to Oates’s revival is evident in Albert Hirschman’s famous book, Exit, Voice, and Loyalty (Cambridge, Mass.: Harvard University Press), published in 1970. Hirschman, in framing the tradeoffs between exit and voice as mechanisms for disciplining institutions, took it for granted that exit was “very nearly” unavailable in basic social organizations such as the family, the state, and the church, leaving voice as the only option for members to register their dissatisfaction; Hirschman, Exit, Voice, and Loyalty, pp. 33, 76, 121. Hirschman was concerned primarily with disciplining lazy (slack) producers—a problem not even acknowledged as a possibility in neo-classical economic models, which presupposed (in Hirschman’s terms) a “relentlessly taut [market] economy”; Hirschman, Exit, Voice, and Loyalty, p. 9 (emphasis omitted). Public finance theorists like Tiebout, in contrast, assumed a taut economy in the private market, and sought a mechanism to produce the same result in the market for public goods. But for present purposes, that distinction is immaterial. The Tieboutian argument, if known to Hirschman, would have been an important qualification to his assumption that voice is the only effective mechanism to register disapproval of a government’s priorities.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

       - 

239

had failed to identify the market mechanism because they were looking for it in the wrong place. They were looking for something that would force the residents of a given jurisdiction to disclose their true preferences over public goods—in Albert Hirschman’s famous terms, they were looking for “voice.” If they had focused instead on residents’ initial choice of where to live, they would have realized that an alternative mechanism already existed. In choosing among different jurisdictions offering different packages of taxes-for-public-goods, each citizen votes with her feet for the package that comes closest to meeting her true preferences—in Hirschman’s terms, she votes by “exit.”⁹ In Tiebout’s world, jurisdictions thus function as a kind of public goods club, which bundles together various public services and then puts the entire bundle on offer to potential members (would-be residents). Each individual accepts a particular club’s taxesfor-public goods package by moving to the jurisdiction, and, once there, staying put. Assuming an infinite variety of taxes-for-public-goods packages on offer, zero costs of relocation, perfect information on the part of consumer/voters, and no externalities between communities, each consumer should be able to secure her optimal bundle of public goods at each governmental level, simply by choosing to establish residence in the polity in which it is on offer. Indeed, as one commentator noted, in a world of perfect Tieboutian sorting, we would have no reason to limit ourselves to the existing menu of vertical options (federal, state, and city governments, in the case of the US). We could in theory unbundle all services, allowing each to be optimized separately, by having a “customtailored government . . . [to] match[] each collective purpose.”¹⁰ In the real world, of course, none of these assumptions holds perfectly. As a consequence, almost everyone will have to settle for a package that is suboptimal from their perspective (more public art or public transit than they want, not enough invested in quality public schools or public parks, etc.).¹¹ The question is whether competition functions well enough to approximate an efficient market. In Tiebout’s model, the answer essentially boils down to whether a resident of a jurisdiction has adequate exit options, such that when she chooses to stay put, we are confident she is not just choosing the least bad of a bad set of alternative packages on offer. Thus, Tiebout’s model made two important contributions to the public finance literature. The first was to show that voting was not the only way for citizens to ⁹ Hirschman, Exit, Voice, and Loyalty. ¹⁰ John D. Donahue, “Tiebout? Or Not Tiebout? The Market Metaphor and America’s Devolution Debate,” Journal of Economic Perspectives 11, no. 4 (Autumn 1997): p. 73. The phenomenon of coupling the provision of public services with legal residency is a relatively recent historical phenomenon. See Richard T. Ford, “Law’s Territory (A History of Jurisdiction),” Michigan Law Review 97, no. 4 (Feb. 1999): p. 846. One can find occasional examples of decoupling in contemporary America (e.g., water districts, cross-jurisdictional school districts, commuter taxes). ¹¹ For a discussion of this issue in the context of Nozick’s “Utopia” of jurisdictional competition, see Chapter 7.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

240

      

register preferences over public goods. With any degree of mobility among horizontal jurisdictions, citizens’ power to exit one jurisdiction for another functions much as consumer choice does in the private market. It devolves decisionmaking to the individual level, allowing each citizen (and not just the median voter) to optimize her preferences. As one commentator put it, perfect Tieboutian sorting presents a vision of a world in which “every citizen can stage a personal revolution armed only with a moving van.”¹² Tiebout’s second contribution was to suggest that at the local level, where mobility is fairly high, we might be able to come much closer to the optimal provision of public goods than previously thought. That possibility has important practical implications in the US, where a substantial percentage of government expenditures are made at the state and local level,¹³ as well as in other countries (e.g., Canada and Switzerland) that have federal systems that devolve a significant portion of fiscal power to local governments.¹⁴ In the past four decades, the Tieboutian model has given rise to an enormous body of empirical work on the extent of interjurisdictional sorting in the US, as well as whether that sorting has triggered a “race to the bottom” or a “race to the top” in various spheres of government activity, including regulation of corporations, tax rates, welfare reform, and environmental regulation.¹⁵ Although Tiebout himself focused on competition between local governmental units, horizontal competition among jurisdictions could theoretically operate at any governmental level. But as an empirical matter, the cost of exit is likely to rise as one moves up the hierarchy of jurisdictions. Tiebout assumed that in the US, the option of exit was, practically speaking, nonexistent at the federal level. That assumption is not unreasonable for individual US citizens—the focus of Tiebout’s article—for whom immigration to another country is often legally unavailable (entry, not exit, is the problem here), and prohibitively costly because of all one leaves behind: family connections, friends, job, linguistic fluency, cultural familiarity, etc. But if one extends the Tiebout model to corporations and other business entities, the picture looks quite different. It is essentially costless for many businesses to move their legal residence from one state, or even one country, to another. As a result, we not surprisingly observe wide-scale jurisdiction-shopping

¹² Donahue, “Tiebout? Or Not Tiebout?,” p. 73. ¹³ At the time of Tiebout’s article, roughly half of non-defense-related government spending was at the state or local level. William A. Fischel, ed., The Tiebout Model at Fifty: Essays in Public Economics in Honor of Wallace Oates (Cambridge, Mass.: Lincoln Institute of Land Policy, 2006), p. xiv. Since then, the percentage has, if anything, increased. ¹⁴ Stephen Ross and John Yinger, “Sorting and Voting: A Review of the Literature on Urban Public Finance,” in Handbook of Regional and Urban Economics, vol. 3, edited by Peter Nijkamp (Holland: Elsivier Science, 1999), pp. 2001–60. ¹⁵ For an overview of the empirical literature on Tiebout sorting for individuals, see Ross and Yinger, “Sorting and Voting,” pp. 2005–6.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

       - 

241

by businesses, resulting in substantial tax and regulatory competition between states and between countries to attract businesses.¹⁶ In assessing the effects of interjurisdictional competition, Tiebout confined his attention to allocative efficiency, ignoring distributional consequences. That choice did not reflect indifference on his part to distributional issues, but rather the conventional view that efficiency and equity are analytically and institutionally separable issues, and should be dealt with separately. The sensible division of labor, Musgrave had argued, was for the government to focus solely on efficiency considerations in determining output levels, and address distributional issues on the tax side in choosing between benefits taxation and some version of “ability to pay” to distribute the burden of financing that output.¹⁷ But as many quickly recognized, jurisdictional competition at the local level is likely to push local governments not just toward efficient output but also toward some version of benefits taxation to finance it. Since the demand for public goods is positively correlated with income, if citizens distribute themselves among jurisdictions based on their desired level of public goods, they will automatically end up sorting by income class as well. In addition, wealthy residents wishing to help that sorting along can do so through the visible hand of zoning laws. By mandating minimum lot size and setbacks, maximum number of bedrooms, aesthetic requirements for buildings, etc., local governments can set a high floor on the cost of residents’ housing consumption per family member, and hence on their wealth.¹⁸ Wealthy communities have an incentive to do so, in order to solve what are, from the self-interested perspective of their wealthier residents, two different forms of market failure: poorer residents free-riding on the higher level of public goods provided in rich communities (“fiscal zoning”), and dilution of the positive externalities from resident characteristics that are believed to correlate with income level and enhance the value of local services financed by the tax base (prepared and highly motivated students, law-abiding residents, similar aesthetic tastes, etc.). At the extreme, Tiebout sorting points to a world of “high income communities walled off by zoning ordinances that effectively prevent entry by low-income households.”¹⁹

¹⁶ For a more recent look at the influence of state policy on locational choices of industry, see John D. Donahue, Disunited States: What’s at Stake as Washington Fades and the States Take the Lead (New York: Basic Books, 1997), Appendix. Corporate tax immigration at the national level has received significant attention in the past decade, due to US multinationals’ wholesale transfer of revenues to offshore tax havens through non-operating subsidiaries. For companies whose principal assets are intellectual property, such transfers face few obstacles in the existing legal regime. ¹⁷ Richard A. Musgrave, “A Multiple Theory of the Budget Determination,” FinanzArchiv/Public Finance Analysis, New Series, Bd. 17, H. 3 (1956/1957): pp. 333–43. ¹⁸ Bruce W. Hamilton, “Zoning and Property Taxation in a System of Local Government,” Urban Studies 12 (1975): pp. 205–11; William A. Fischel, Regulatory Takings: Law, Economics, and Politics (Cambridge, Mass.: Harvard University Press, 1995). ¹⁹ Wallace E. Oates, “The Many Faces of the Tiebout Model in The Tiebout Model at Fifty: Essays in Public Economics in Honor of Wallace Oates, edited by William A. Fischel (Cambridge, Mass.: Lincoln Institute of Land Policy, 2006), p. 41.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

242

      

In short, competition among local governments was predicted to yield a market for public goods that mimicked the private market not just in allocative efficiency but in distributional effects as well. With perfect sorting, residents would pay only for the local public services they actually wanted, at a price that reflected the average or marginal cost of providing them—libertarians’ ideal conception of a benefits tax. Even with imperfect sorting, however, it is widely assumed that the possibility of exit precludes significant redistribution at the local level. As a result, the conventional wisdom post-Tiebout is that in a federal system, ambitious redistributive programs have to be pushed up to the highest governmental level (in the US, the federal government), from which exit is generally very costly. Local governments, constrained by market forces from doing any extensive redistribution, should instead concentrate on what they can do best: the efficient provision of local public goods.²⁰ Most empirical tests of the Tiebout hypothesis over the past forty years have found significant Tiebout sorting of individual taxpayers where the model predicts it is most likely to occur: in major metropolitan areas comprising numerous separate jurisdictions in close proximity to each other. In such geographical areas, individual citizens can approximate their preferred public goods package at very little cost, because they can change legal residence without changing jobs, friends, etc.²¹ While there is some evidence that Tieboutian sorting increases efficiency in the production of public goods, it is highly unlikely that it has come close to optimal sorting from a welfarist perspective. There are myriad issues here, explored in detail in the massive local public finance literature that Tiebout’s article spawned.²² Some concern dimensions of social welfare left out of the Tiebout model in its single-minded focus on achieving efficient output levels of public goods. To take one much-discussed example, the social homogeneity that Tiebout sorting produces, while it may significantly increase productive efficiency, imposes other social costs, including the loss of positive peer effects that can arise in more heterogeneous communities.²³ Others concern the real-world obstacles to perfect ²⁰ George J. Stigler, “The Tenable Range of Functions of Local Government,” in Federal Expenditure Policy for Economic Growth and Stability, Joint Economic Committee (Washington, D.C.: Joint Economic Committee, Subcommittee on Fiscal Policy, 1957), pp. 213‒19, cited in William A. Fischel, “Footloose at Fifty: An Introduction to the Tiebout Anniversary Essays,” in The Tiebout Model at Fifty: Essays in Public Economics in Honor of Wallace Oates, edited by William A. Fischel (Cambridge, Mass.: Lincoln Institute of Land Policy, 2006), p. 6; Richard A. Musgrave and Peggy B. Musgrave, Public Finance in Theory and Practice, 5th ed. (McGraw Hill, 1989); Charles C. Brown and Wallace E. Oates, “Assistance to the Poor in a Federal System,” Journal of Public Economics 32, no. 3 (Apr. 1987): pp. 307–30. ²¹ Dennis C. Mueller, Public Choice III (Cambridge: Cambridge University Press, 2003), pp. 199–202; Keith Dowding, Peter John, and Stephen Biggs, “Tiebout: A Survey of Empirical Literature,” Urban Studies 31, nos. 4/5 (1994): pp. 767–97. ²² For an overview of some of these issues, see Ross and Yinger, “Sorting and Voting,” pp. 2042–9. ²³ Richard Arnott and John Rowse, “Peer Group Effects and Educational Attainments,” Journal of Public Economics 32, no. 3 (Apr. 1987): pp. 287–305; Jan K. Brueckner and Kangoh Lee, “Club Theory

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

    -   

243

Tieboutian sorting even at the local level: the existence of interjurisdictional externalities; the fact that the property tax base that localities use to fund public goods does not directly track residents’ differential consumption tastes or the differential use they make of public goods; and various constraints on mobility. Constraints on mobility are particularly important here. In thinking about the forces that would keep someone put, notwithstanding a more attractive taxes-forpublic-goods package on offer elsewhere, Tiebout principally had in mind the costs of relocating (moving, finding a new job or enduring a longer commute, etc.). But many of the factors that keep people from moving could more accurately be described as location-specific social capital: family, friends, cultural opportunities, economic ties, aesthetics, a sense of place, etc. Residents contemplating exiting (or entering) a given jurisdiction cannot unbundle its taxes-for-public-goods deal from its location-specific social capital: it is a take-it-or-leave-it package deal. This reality not only undercuts efficient output of public goods; it has important distributional implications as well, as it gives jurisdictions the market power to extract the surplus value generated by their location-specific social capital in setting tax rates (hereinafter, “locational surplus”). Since only the well-off can afford to pay significant amounts for social capital, if jurisdictions use their bargaining leverage to the hilt, a plausible outcome is a rate structure that price-discriminates on the basis of income—that is, that charges the rich more than the poor for equivalent consumption of public goods.²⁴ The only way for social contractarians on the right to guarantee the outcome they want is to prohibit jurisdictions from exercising the market power they enjoy in the real world to extract payment for that locational surplus. This returns us to a variant of the problem in Chapter 12: can such a prohibition be justified, either under libertarian commitments in general or by an analogy to the private market. I turn to that issue now.

3. The State as a Market-Mimicking Actor for Distributional Purposes Starting with the Hobbesian/Lockean premise that the just state exists solely to provide public goods that private markets undersupply, libertarians have argued with a Peer-Group Effect,” Regional Science and Urban Economics 19, no. 3 (Aug. 1989): pp. 399–420; Anita A. Summers and Barbara L. Wolfe, “Do Schools Make a Difference?” American Economic Review 67, no. 4 (Sep. 1977): pp. 639–52; Robert M. Schwab and Wallace E. Oates, “Community Composition and the Provision of Local Public Goods: A Normative Analysis,” Journal of Public Economics 44, no. 2 (Mar. 1991): pp. 217–37. ²⁴ A number of other factors potentially in play leave the outcome of the hypothetical social contractarian thought experiment analytically indeterminate. To cite just one, wealth may bring much greater mobility (the financial means to commute long distances; comparable job opportunities in other locations; less emotional attachment to immediate neighborhood, etc.), which, all other things being equal, would reduce a locality’s bargaining leverage. But it also may bring less mobility, if highpaying jobs are disproportionately clustered in high-tax metropolitan areas.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

244

      

that the state should be limited to solving that problem, and, in particular, be barred from using its taxing powers to redistribute resources from the rich to the poor. In short, the state should mimic the behavior of the private market not just in setting output, but also in allocating among its citizens the costs of producing explicit public goods. Most of the efforts to give substantive content to the notion of a just price in private markets have come from those whose political sympathies are on the left. Notable examples include the (socialist) labor theory of value, the anti-monopoly movement, and other Progressive critiques of economic rents ascendant in the late nineteenth and early twentieth centuries.²⁵ In contrast, most classic liberals/ libertarians view the search for a just price in private markets as quixotic and morally unpersuasive, arguing that—with very narrow exceptions—a just price is whatever price happens to result from a bargain fairly entered into.²⁶ AngloAmerican contract law has largely followed suit. In the US, while the Uniform Commercial Code delegates broad discretion to courts to police the substantive fairness of individual contracts, courts rarely exercise that power.²⁷ State legislatures have stepped in only sparingly to mandate “fair terms” ex ante, generally in sectors of the economy thought to face some sort of systemic market failure (e.g., urban housing markets, public utilities, consumer insurance contracts). Instead, courts and legislatures have focused on the procedural fairness of a contract—in particular, on ensuring that consent was given knowingly and voluntarily. If a court concludes it was, then the substantive terms of the contract will generally be upheld. While the kinds of ignorance and mistake that should suffice to void a contract are disputed in both the legal and the philosophical literature, the parameters of that debate, if not the appropriate resolution, are clear. Not so with voluntariness. The word suggests that what we are looking for is a showing of volition, meaning that the parties consciously chose to bind themselves to the terms of the agreement. But aside from relatively rare cases involving direct physical compulsion that deprives an individual of all choice or mental incapacity that prevents her from making a meaningful choice, volition in this sense is never in doubt. When people describe an agreement as involuntary or coerced, they generally mean something quite different: that people are driven to accept the offered terms because their alternatives are considerably worse. Consider a classic example in the common law. A gunman offers his victim the choice of “your money or your life.” As Oliver Wendell Holmes, Jr., famously argued, when the victim opts for ²⁵ See Fried, Progressive Assault, ch. 4. ²⁶ The nightmare scenarios that drive even die-hard free-market libertarians to balk at letting the free market set the price are typically contracts for essentials (e.g., rescue from imminent peril, food and water needed for survival), where a monopoly supplier seeks to extort a price for the scarce good or services far in excess of its cost. ²⁷ The relevant provision here is UCC § 2–302, which deals with unconscionability.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

    -   

245

the first alternative, he does not lack volition in the narrow sense. He had a choice, and he knowingly and deliberately chose the former. Indeed, the worse the alternative on offer (“your life”), the more eagerly he would embrace the former.²⁸ If there is a reason not to hold him to that choice, it has to do with the set of choices he had available to him. Holmes’s reformulation of the problem a century ago has become the standard point of departure for analyses of coercion in the legal and philosophical literature. While that reformulation forces us to ask the right question—is there some problem with the set of alternative choices available?—it doesn’t answer it. All choices are constrained by available alternatives, and available alternatives are always limited. In that sense, one could describe every agreement as coerced (on both sides) by the fact that all other available options are, by definition, worse or they would have been chosen instead.²⁹ The question is, when should the constraints placed on our choices as a matter of fact lead us to conclude that the agreement is coerced in a moral or legal sense, and hence unenforceable? The traditional libertarian response has been, almost never, and once again, Anglo-American contract law has followed free-market instincts here. Everyone would make an exception for bargains in which one side has constrained the other’s background options through illegal means.³⁰ This takes care of Holmes’s gunman. It is true that his victim had a choice, albeit between two bad alternatives. But by law, the victim had a right to a third and better alternative, which the gunman wrongly took from him: to keep both his money and his life. But it doesn’t touch the standard commercial contract, in which the constraints on the next best alternatives available to each party are imposed by the natural, social, and legal world (the market, property rights, personal endowments, tastes, etc.) without any help from the other side. In short, we are generally not permitted to put the other side over a barrel ourselves. But if we come upon the other side already over a barrel through no fault of our own, we are generally entitled to exploit any bargaining advantage their misfortune gives us.³¹ Thus, in the context of private agreements, the law, like the free-market libertarian, has resisted importuning from the left either to police the substantive fairness of contract terms or to inquire whether the options available to the parties are sufficiently robust that we should attribute moral agency to the parties’ assent. ²⁸ Union Pacific R.R. Co. v. Pub. Serv. Comm., 248 U.S. 67, 70 (1918). ²⁹ For the classic statement and development of this analytic truism, see Robert L. Hale, “Coercion and Distribution in a Supposedly Non-Coercive State,” Political Science Quarterly 38, no. 3 (Sep 1923): pp. 470–8. For further discussion of the issue, see Fried, Progressive Assault, ch. 2. ³⁰ The exception for terms extorted through illegal means arguably just pushes the moral difficulties down a level, to the decision what to declare illegal. But here, unlike the concept of coercion itself, we generally have a well-developed set of norms to fall back on and moral intuitions to support them. ³¹ In the legal context, the difficult cases at the border are typically ones in which, partway through performance of a contract, a seller demands more money from a buyer to complete performance, threatening to breach the contract if the buyer won’t agree. The cases are difficult because threat of breach is halfway between legally and illegally obtained bargaining power.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

246

      

With very few exceptions, both have treated each party’s available alternatives, however unappealing they might be as an absolute matter, as a morally neutral fact, and have treated any agreement reached against the backdrop of those alternatives as (in Nozick’s famous phrase) a “capitalist act[] between consenting adults.”³² When it comes to the social bargain, in contrast, benefits tax proponents have a thick conception of a just price for public goods, one that reflects a highly moralized view of the jurisdictional markets in question. In particular, a just price is the price that would result were there a competitive market for explicit public goods, and the goods could be disaggregated from each other, as well as from any locational surplus the taxpayer enjoys. That is to say, it is the price that would result in a jurisdictional market with perfect Tieboutian sorting. That idealized market does not correspond to the jurisdictional markets we observe in the real world. It also doesn’t correspond to many private markets that libertarians, under their thin, procedural conception of fair price, are committed to leaving alone. Which returns us to the question posed in Chapter 12: what justifies endowing individuals with more market power in their negotiations with their fellow citizens than they would likely enjoy in real life? In Chapter 12, the market power in question was exercised in reaching hypothetical agreement with fellow bargainers on the terms of the social contract. Here, it is exercised by each would-be citizen in reaching hypothetical agreement with an already existing state, acting as an agent for its citizens, on the price of public goods. But these are just different ways to frame the identical question: what is wrong with leaving it to ordinary market forces to allocate whatever surplus value is generated by social cooperation? In answering that question in the context of the taxes-for-public-goods deal, I want to focus on locational advantages, which, for many people, are the principal source of surplus value from residing in a given jurisdiction. Some of the things that hold people to a particular location are purely pathdependent. If Alix were moving to the US with no ties to any particular location, she might choose to live in Chicago, all things considered. But she was born and raised in Cincinnati, and her preference for Chicago over Cincinnati as a place of residence is not great enough to entice her to give up her personal connections to Cincinnati and go through the hassle and expense of moving. Others inhere in the intrinsic attractions of the place itself (cultural, social, and work opportunities, ambiance, aesthetics, weather, proximity to recreational areas, etc.). In either case, to the extent that such non-public benefits are tied to staying put in a particular polity, they are bundled in with the taxes-for-public-goods deals on offer in that polity. For states like California that have significant location-specific attractions for many people, the fact that such benefits cannot be unbundled from

³² Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 163.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

    -   

247

legal residence means that California is in the position of de facto quasi-monopolist: it has the market power to set tax rates high enough to extract the entire surplus value of living in California rather than, say, Nevada or Kansas, or whatever other state represents Californians’ next best alternative. Writ large, the same is true of the choice to live in the US rather than another country, which explains why the top 0.001 percent of the income distribution in the US has not moved en masse to the Cayman Islands or some other tax haven. In contrast, places like West Virginia or Alabama, whose location-specific attributes are unlikely to give them a comparative advantage in attracting or retaining residents, have to set tax rates close to the average or marginal cost of explicit public goods in order to be competitive. To attract residents who generate positive externalities for the community (e.g., industry and other businesses that boost employment), they may have to reduce the tax rate to zero or even lower. (To put it another way, residents who bestow positive externalities on the community can act as monopsonists, offsetting the tax price they would normally be charged for the public goods they consume by the surplus value they generate for the community.) So the question facing benefits tax proponents is this. If the state is morally required to peg individual taxes to the shadow market price of the public goods that each taxpayer gets in return, of the many markets that we observe in the real world of jurisdictional competition, which is the right market to use as a benchmark? The market in which jurisdictions have the bargaining leverage to set tax rates high enough to extract a significant portion of the locational surplus their residents derive from living there (California)? The market in which some potential residents have the bargaining leverage to get public goods for free or even be paid to move there (businesses contemplating moving to Alabama)? Or the market faced by states (say, Ohio) that possess average intrinsic attractiveness to potential citizens, and are therefore limited to charging a competitive market price for explicit public goods, equal to the average or marginal cost of producing them? As I suggested above, benefits tax proponents on the right have opted for the last of these alternatives. That is, they would have us imagine a world of perfect Tieboutian sorting, in which explicit public goods are disaggregated from all other locational attributes, limiting jurisdictions’ de facto taxing power to the price of those goods in a well-functioning, transactions-costless, competitive market,³³ and a sufficient array of taxes-for-public-goods packages are on offer that no taxpayer has to pay for public goods she doesn’t want. But what exactly is wrong with letting California charge whatever premium the market will pay for living in California? If it is generally permissible for private parties to exploit their locational and other advantages in pricing goods or services ³³ Whether benefits theorists would permit taxpayers to pay less than that (viz, Alabama) is a separate question. I assume most benefits theorists would say yes.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

248

      

(e.g., the premium charged by the only gas station for 40 miles around, or by Gucci in leveraging its snob appeal to price goods at many times the cost of producing them), why is it not permissible for the state to do the same? One obvious answer concerns the greater degree of monopoly power exercised by the state than by private producers—an issue I took up in Chapter 7, in discussing Nozick’s Tieboutian vision of Utopia. That difference, however, is easy to overstate. Many of the impediments to a perfectly competitive jurisdictional market (search costs, imperfect information, externalities, a paucity of attractive alternatives) have analogues in the private market, where they cause people to settle for a “voluntary” exchange that is suboptimal from their own point of view. And, on the other side, even when exit from a jurisdiction is very costly, the rich have another valuable exit option available to constrain tax rates: substituting leisure for work. As practical matter, the threat of exit from the labor force may be as powerful a constraint on the majority’s taxing powers as the threat of exit from the taxing jurisdiction. But this is not to deny that even in free societies, people typically have fewer feasible options when shopping for a country of residence than when choosing among goods or services in the private market. Is that difference in degree great enough to constitute a difference in kind, such that libertarians are justified in fixing the (tax) price of membership in a given polity, even though they would not seek to fix the market price in private exchanges? These are hard questions. But I think there are powerful reasons to conclude no. Benefits tax proponents’ willingness to assume away any de facto constraints on individual taxpayers’ optimizing their choice among jurisdictions, like social contractarians’ demand that we idealize exit options in the social contractarian bargain, is hard to square with the libertarian ideal of a minimal state with limited obligations to ameliorate its citizens’ social, biological, or historical circumstances. And their choice of which de facto constraints to assume away and which to take as given seems morally arbitrary. Until proponents of benefits taxation can explain how they get from general libertarian principles to this particular idealized market for public goods, a commitment to benefits taxation gives us no reason to reject progressive taxation, or indeed almost any tax scheme, including whatever happens to result at any given time from the various market forces actually in play in interjurisdictional competition. Up until this point, I have assumed that the state plays no role in generating the locational surplus it is trying to tax, and hence is just capitalizing on the adventitious fact that enjoyment of that surplus is tied to residency within its jurisdictional borders. That assumption reflects a very narrow conception of public goods. Many of the reasons people value living in a particular community have to do with its built social and physical environment: norms of civility and trust, cultural institutions, good restaurants, public parks, job opportunities, transportation and housing infrastructure, architecture, good teachers, skilled

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

    -   

249

doctors, well-functioning hospitals, and all the other parts that go into making a city, state, or country a place that works, a place where people want to live. Some of these working parts can be traced back to explicit public expenditures in the form of tax exemptions for nonprofits, tax credits, or cash expenditures. But the larger part comes from social capital built up by the private, interdependent actions of residents and local businesses over many generations. That social capital explains why the wealthy in the US, many of whom make no use of the most expensive local public goods (public schools, mass transportation, public hospitals), are nonetheless willing to pay a very high (tax) price to live in Manhattan, Chicago, San Francisco, etc. Why shouldn’t the members of a polity who have collectively generated and maintained its social capital be entitled to charge for the right to enjoy it, along with the explicit public goods that the jurisdiction supplies? And if the unique value of its social capital gives a polity quasi-monopoly power, why shouldn’t its citizens be free to use that power to price-discriminate, setting membership fees in accordance with a means-tested sliding scale that roughly tracks willingness to pay? After all, no one is forcing anyone to move to California or forbidding anyone from leaving. From a libertarian perspective, what more is required to justify California’s charging whatever the market will bear to live there?

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

Bibliography Abramson, Elliot M. “Philosophization Against Taxation: Why Nozick’s Challenge Fails,” Arizona Law Review 23, no. 2 (1981): p. 753. Ackerman, Bruce, and Anne Alstott. The Stakeholder Society (New Haven: Yale University Press, 1999). Alexander, Larry. “Lesser Evils: A Closer Look at Paradigmatic Justification,” Law and Philosophy 24, no. 6 (Nov. 2005): p. 611. Alexander, Larry, and Michael Moore. “Deontological Ethics,” Stanford Encyclopedia of Philosophy, November 21, 2007, http://plato.stanford.edu/entries/ethics-deontological. American Law Institute. Model Penal Code: Official Draft and Explanatory Notes (Philadelphia, Pa.: The Institute, 1985). Arnott, Richard, and John Rowse. “Peer Group Effects and Educational Attainments,” Journal of Public Economics 32, no. 3 (Apr. 1987): p. 287. Arrow, Kenneth J. “Some Ordinalist-Utilitarian Notes on Rawls’s Theory of Justice,” Journal of Philosophy 70, no. 9 (May 1973): p. 245. Arrow, Kenneth. “Nozick’s Entitlement Theory of Justice,” Philosophia 7, no. 2 (Jun. 1978): p. 265. Ashford, Elizabeth. “Utilitarianism, Integrity, and Partiality,” Journal of Philosophy 97 (2000): p. 421. Ashford, Elizabeth. “The Demandingness of Scanlon’s Contractualism,” Ethics 113 (2003): p. 273. Barry, Brian. Theories of Justice: A Treatise on Social Justice, vol. 1 (Berkeley: University of California Press, 1989). Becker, Lawrence. “Against the Supposed Difference Between Historical and End-State Theories,” Philosophical Studies 41, no. 2 (Mar. 1982): p. 267. Bergstrom, Theodore C., and Robert P. Goodman. “Private Demands for Public Goods,” American Economic Review 63, no. 3 (Jun. 1973): p. 280. Birks, Peter. “The Concept of a Civil Wrong.” In Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), p. 29. Brand-Ballard, Jeffrey. “Contractualism and Deontic Restrictions,” Ethics 114, no. 2 (Jan. 2004): p. 269. Breakey, Hugh. “Property,” Internet Encyclopedia of Philosophy, https://www.iep.utm.edu/ prop-con/#H3. Brighouse, Harry. “Political Equality in Justice as Fairness,” Philosophical Studies 86, no. 2 (May 1997): p. 155. Brock, Dan W., and Daniel Wikler. “Ethical Challenges in Long-Term Funding for HIV/ AIDS,” Health Affairs 28, no. 6 (Nov. 2009): p. 1666. Brody, Baruch. “Redistribution Without Egalitarianism.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 31. Brown, Charles C., and Wallace E. Oates. “Assistance to the Poor in a Federal System,” Journal of Public Economics 32, no. 3 (Apr. 1987): p. 307.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

252



Brown, Peter G. “Food as National Property.” In Food Policy: The Responsibility of the United States in the Life and Death Choices, edited by Peter G. Brown and Henry Shue (New York: The Free Press, 1977), p. 65. Brueckner, Jan K., and Kangoh Lee. “Club Theory with a Peer-Group Effect,” Regional Science and Urban Economics 19, no. 3 (Aug. 1989): p. 399. Buchanan, James M. The Limits of Liberty (Chicago: Chicago University Press, 1975). Buchanan, James M. “The Ethical Limits of Taxation,” The Scandinavian Journal of Economics 86, no. 2 (Jun. 1984): p. 102. Buchanan, James M. “The Gauthier Enterprise,” Social Philosophy and Policy 5, no. 2 (Spring 1988): p. 75. Buchanan, James M., and Loren E. Lomasky. “The Matrix of Contractarian Justice,” Social Philosophy and Policy 2, no. 1 (Autumn 1984): p. 12. Calabresi, Guido, and A. Douglas Melamed. “Property Rules, Liability Rules, and Inalienability: One View of the Cathedral,” Harvard Law Review 85, no. 6 (Apr. 1972): p. 1085. Christman, John. “Self-Ownership, Equality, and the Structure of Property Rights.” In LeftLibertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 345. Clark, John Bates. The Distribution of Wealth (New York: The Macmillan Company, 1899). Coase, R. H. “The Problem of Social Cost,” The Journal of Law and Economics 3 (Oct. 1960): p. 1. Cohen, Felix. “The Ethical Basis of Legal Criticism,” Yale Law Journal 41, no. 2 (Dec. 1931): p. 201. Cohen, G. A. “On the Currency of Egalitarian Justice,” Ethics 99, no. 4 (Jul. 1989): p. 906. Cohen, G. A. “The Limits of Contractual Equality: A Reply to Jacques Bidet,” Ratio Juris 8, no. 1 (Mar. 1995): p. 85. Cohen, G. A. Self-Ownership, Freedom, and Equality (Cambridge: Cambridge University Press, 1995). Cohen, I. Glenn, Norman Daniels, and Nir Eyal. Identified versus Statistical Lives: An Interdisciplinary Perspective (New York: Oxford University Press, 2015). Coleman, Jules. The Practice of Principle: In Defense of a Pragmatist Approach to Legal Theory (Oxford: Oxford University Press, 2001). Coleman, Jules. “Tort Law and Tort Theory.” In Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), p. 183. Coleman, Jules. Risks and Wrongs (Oxford: Oxford University Press, 2002). Craswell, Richard. “In That Case, What is the Question? Economics and the Demands of Contract Theory,” Yale Law Journal 112, no. 4 (Jan. 2004): p. 903. Daniels, Norman. “Can There Be Moral Force to Favoring an Identified over a Statistical Life?” In Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015), p. 110. Dobbs, Dan B. Law of Remedies: Damages-Equity-Restitution 2, 2nd ed. (Eagan, Minn.: West Publishing Co., 1993). Donahue, John D. Disunited States: What’s at Stake as Washington Fades and the States Take the Lead (New York: Basic Books, 1997). Donahue, John D. “Tiebout? Or Not Tiebout? The Market Metaphor and America’s Devolution Debate,” Journal of Economic Perspectives, 11, no. 4 (Autumn 1997): p. 73. Dowding, Keith, Peter John, and Stephen Biggs. “Tiebout: A Survey of Empirical Literature,” Urban Studies 31, nos. 4/5 (1994): p. 767.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



253

Dworkin, Ronald. “What is Equality? Part 2: Equality of Resources,” Philosophy and Public Affairs 10, no. 4 (Autumn 1981): p. 283. Ellsberg, Daniel. “Risk, Ambiguity, and the Savage Axioms,” Quarterly Journal of Economics 75, no. 4 (Nov. 1961): p. 643. Epstein, Richard A. “A Theory of Strict Liability,” Journal of Legal Studies 2, no. 1 (Jan. 1973): p. 151. Epstein, Richard A. “The Utilitarian Foundations of Natural Law,” Harvard Journal of Law and Public Policy 12, no. 3 (Summer 1989): p. 713. Etner, Johanna, Meglena Jeleva, and Jean-Marc Tallon. “Decision Theory Under Ambiguity,” Journal of Economic Surveys 26, no. 2 (Apr. 2012): p. 234. Eyal, Nir. “Concentrated Risk, the Coventry Blitz, Chamberlain’s Cancer.” In Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015), p. 94. Feinberg, Joel. Harm to Others (Oxford: Oxford University Press, 1987). Finkelstein, Claire. “Is Risk a Harm?,” University of Pennsylvania Law Review 151, no. 3 (Jan. 2003): p. 963. Fischel, William A. Regulatory Takings: Law, Economics, and Politics (Cambridge, Mass: Harvard University Press, 1995). Fischel, William A., ed., The Tiebout Model at Fifty: Essays in Public Economics in Honor of Wallace Oates (Cambridge, Mass.: Lincoln Institute of Land Policy, 2006). Fletcher, George P. “Fairness and Utility in Tort Theory,” Harvard Law Review 85, no. 3 (Jan. 1972): p. 537. Foot, Philippa. “Utilitarianism and the Virtues,” Proceedings and Addresses of the American Philosophical Association 57, no. 2 (Nov. 1983): p. 273. Ford, Richard T. “Law’s Territory (A History of Jurisdiction),” Michigan Law Review 97, no. 4 (Feb. 1999): p. 843. Fox, Craig R., and Amos Tversky. “Ambiguity Aversion and Comparative Ignorance,” Quarterly Journal of Economics 110, no. 3 (Aug. 1995): p. 585. Frick, Johann. “Contractualism and Social Risk,” Philosophy and Public Affairs 43, no. 3 (Summer 2015): p. 175. Frick, Johann. “Treatment versus Prevention in the Fight against HIV/AIDS and the Problem of Identified versus Statistical Lives.” In Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015), p. 182. Fried, Barbara H. “As Compared to What? Taxing Brute Luck and Other Second-Best Problems,” Tax Law Review 53, no. 3 (Spring 2000): p. 377. Fried, Barbara H. The Progressive Assault on Laissez Faire: Robert Hale and the First Law and Economics Movement (Cambridge, Mass.: Harvard University Press, 1998). Fried, Barbara H. “The Puzzling Case for Proportionate Taxation,” Chapman Law Review 2 (Spring 1999): p. 157. Fried, Barbara H. “Why Proportionate Taxation?” In Tax Justice, edited by Joseph J. Thorndike and Denis J. Ventry (Washington, D.C.: Urban Institute Press, 2002), p.149. Fried, Barbara H. “Beyond Blame,” Boston Review, July/August 2013. Fried, Charles. Right and Wrong (Cambridge, Mass.: Harvard University Press, 1978). Gauthier, David. “The Social Contract: Individual Decision or Collective Bargain?” In Foundations and Applications of Decision Theory, vol. 2, edited by C. A Hooker, J. J. Leach, and E. E McClennan (New York: Kluwer Academic Publishers, 1978), p. 47. Gauthier, David. “David Hume, Contractarian,” The Philosophical Review 88, no. 1 (Jan. 1979): p. 3.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

254



Gauthier, David. “Bargaining and Justice.” In Ethics and Economics, edited by Ellen Frankel Paul, Jeffrey Paul, and Fred D. Miller, Jr. (Oxford: B. Blackwell for the Social Philosophy and Policy Center, 1985). Gauthier, David. Morals by Agreement (Oxford: Oxford University Press, 1986). Gibbard, Allan. “Natural Property Rights.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 23. Goldberg, John C. P., and Benjamin C. Zipursky. “Unrealized Torts,” Virginia Law Review 88, no. 8 (Dec. 2002): p. 1625. Goldberg, John C. P., and Benjamin C. Zipursky, “Tort Law and Moral Luck,” Cornell Law Review 92, no. 6 (Sept. 2007): p. 1123. Goodman, Steven N. “Probability at the Bedside: The Knowing of Chances or the Chances of Knowing?,” Annals of Internal Medicine 130, no. 7 (1999): p. 604. Grimshaw v. Ford Motor Co., 119 Cal. App. 3d 757 (1981). Grunebaum, James. “Autonomous Ownership.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 48. Hájek, Alan. “Interpretations of Probability,” The Stanford Encyclopedia of Philosophy, December 19, 2011, https://plato.stanford.edu/entries/probability-interpret/. Hale, Robert L. “Coercion and Distribution in a Supposedly Non-Coercive State,” Political Science Quarterly 38, no. 3 (Sept. 1923): p. 470. Halstead, John. “The Numbers Always Count,” Ethics 126, no. 3 (Apr. 2016): p. 789. Hamilton, Bruce W. “Zoning and Property Taxation in a System of Local Government,” Urban Studies 12 (1975): p. 205. Hamminga, Bert. “Opstaan voor iemand misstaat niemand,” Maandschrift Economie 47 (1983): p. 395. Hampton, Jean. Hobbes and the Social Contract Tradition (Cambridge: Cambridge University Press, 1986). Hampton, Jean. “Correcting Harms versus Righting Wrongs: The Goal of Retribution,” UCLA Law Review 39, no. 6 (Aug. 1992): p. 1681. Hansson, Sven Ove. “Risk and Ethics.” In Risk: Philosophical Perspectives, edited by Tim Lewens (London: Routledge, 2007), p. 21. Harsanyi, John C. “Cardinal Utility in Welfare Economics and in the Theory of RiskTaking,” Journal of Political Economy 61, no. 5 (Oct. 1953): p. 434. Harsanyi, John C. “Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility,” Journal of Political Economy 63, no. 4 (Aug. 1955): p. 309. Harsanyi, John C. “Can the Maximin Principle Serve as a Basis for Morality? A Critique of John Rawls’s Theory,” The American Political Science Review 69, no. 2 (Jun. 1975): p. 594. Hasnas, John. “The Obviousness of Anarchy.” In Anarchism/Minarchism: Is Government Part of a Free Country, edited by Roderick T. Long and Tibor R. Machan (Burlington: Ashgate Publishing Company, 2008), p. 111. Hayek, F.A. The Constitution of Liberty (Chicago: University of Chicago Press, 1959). Hayek, F.A. Law, Legislation and Liberty, vol. 2 (Chicago: University of Chicago Press, 1976). Heinzerling, Lisa. “Knowing Killing and Environmental Law,” N.Y.U. Environmental Law Journal 14, no. 3 (2006): p. 521. Hill, Thomas E. Jr. “A Kantian Perspective on Moral Rules,” Philosophical Perspectives. Vol. 6: Ethics (Chicago: University of Chicago Press, 1992), pp. 285‒304.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



255

Hirschman, Albert. Exit, Voice, and Loyalty (Cambridge, Mass.: Harvard University Press, 1970). Hohfeld, Wesley. “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning,” Yale Law Journal 23, no. 1 (Nov. 1913): p. 16. Hohfeld, Wesley. “Fundamental Legal Conceptions as Applied in Judicial Reasoning,” Yale Law Journal 26, no. 8 (Jun. 1917): p. 710. Holmes, Oliver Wendell Jr. “The Path of the Law,” Harvard Law Review 10, no. 8 (Mar. 1897): p. 457. Holmes, Oliver Wendell Jr., “Privilege, Malice and Intent,” Harvard Law Review 8, no. 1 (Apr. 1894): p. 1. Honoré, Tony. “The Morality of Tort Law—Questions and Answers.” In Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), p. 75. Hospers, John. “The Libertarian Manifesto.” In Morality in Practice, 7th ed., edited by James Sterba (Belmont, Calif.: Wadworth, 2003), ch. 2. Hurley, S. L. “Cognitivism in Political Philosophy.” In Well-Being and Morality: Essays in Honour of James Griffin, edited by Roger Crisp and Brad Hooker (Oxford: Oxford University Press, 2000), p. 177. James, Aaron. “Contractualism’s (Not So) Slippery Slope,” Legal Theory 18, no. 3 (Sept. 2012): p. 263. James, Edmund. “The State as an Economic Factor.” Science Economic Discussion (New York: The Science Company, 1886). Kagan, Robert A. Adversarial Legalism: The American Way of Law (Cambridge, Mass.: Harvard University Press, 2003). Kamm, Frances. Intricate Ethics (Cambridge, Mass.: Harvard University Press, 2007). Kaplow, Louis, and Steven Shavell. “Why the Legal System Is Less Efficient than the Income Tax in Redistributing Income,” Journal of Legal Studies 23, no. 2 (Jun. 1994): p. 667. Kavka, Gregory S. Hobbesian Moral and Political Theory (Princeton, N.J.: Princeton University Press, 1986). Keating, Gregory. “A Social Contract Conception of the Tort Law of Accidents.” In Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), p. 22. Keating, Gregory. “Rawlsian Fairness and Regime Choice in the Law of Accidents,” Fordham Law Review 72, no. 5 (Apr. 2004): p. 1857. Kelman, Mark, “Taking ‘Takings’ Seriously: An Essay for Centrists,” California Law Review 74, no. 5 (Oct. 1986): p. 1829. Kelman, Mark. “The Necessary Myth of Objective Causation,” Chicago-Kent Law Review 63, no. 3 (Oct. 1987): p. 579. Kelman, Mark. “Hedonic Psychology and the Ambiguities of ‘Welfare,’ ” Philosophy and Public Affairs 33, no. 4 (Sept. 2005): p. 391. Kornhauser, Lewis. “Fair Division of Settlements: A Comment on Silver and Baker,” Virginia Law Review 84, no. 8 (Nov. 1998): p. 1561. Kumar, Rahul. “Defending the Moral Moderate: Contractualism and Common Sense,” Philosophy and Public Affairs 28, no. 4 (Oct. 1999): p. 275. Kumar, Rahul. “Who Can Be Wronged?,” Philosophy and Public Affairs 31, no. 2 (2003): p. 109. Kumar, Rahul. “Risking and Wronging,” Philosophy and Public Affairs 43, no. 1 (Winter 2015): p. 27.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

256



Kymlicka, Will. “Property Rights and the Self-Ownership Argument.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 295. Lenman, James. “Contractualism and Risk Imposition,” Politics, Philosophy, and Economics 7, no. 1 (2008): p. 116. Levinson, Daryl. “Making Government Pay: Markets, Politics, and the Allocation of Constitutional Costs,” University of Chicago Law Review 67, no. 2 (2000): p. 345. Locke, John. The Second Treatise of Government. In Two Treatises of Government, edited by Peter Laslett (Cambridge: Cambridge University Press, 1967) (1690). Lomasky, Loren E. Persons, Rights, and the Moral Community (New York: Oxford University Press, 1987). Lomasky, Loren E. “Libertarianism at Twin Harvard.” In Natural Rights Liberalism from Locke to Nozick, Part I, edited by Ellen Frankel Paul, Fred D. Miller, Jr., and Jeffrey Paul (Cambridge: Cambridge University Press, 2005), p. 178. Luban, David. “Unthinking the Time Bomb.” In Global Basic Rights, edited by Charles R. Beitz and Robert E. Goodin (New York: Oxford University Press, 2009), p. 181. Luban, David. “Risk Taking and Force Protection.” In Reading Walzer, edited by Yitzhak Benbaji and Naomi Sussman (New York: Routledge, 2014), p. 277. Lyons, David. “Rawls Versus Utilitarianism,” Journal of Philosophy 69, no. 18 (Oct. 1972): p. 535. Mack, Eric. “Nozick on Unproductivity: The Unintended Consequences.” In Reading Nozick: Essays on Anarchy, State, and Utopia, edited by Jeffery Paul (New Jersey: Rowman & Littlefield, 1981), p. 160. Mack, Eric. “Distributive Justice and the Tensions of Lockeanism,” Social Philosophy and Policy 1, no. 1 (Autumn 1983): p. 132. Mack, Eric. “Self-Ownership, Marxism, and Egalitarianism (Part II): Challenges to the SelfOwnership Thesis,” Philosophy, Politics and Economics 1, no. 2 (Jun. 2002): p. 237. Mack, Eric. “Individualism and Libertarian Rights.” In Contemporary Debates in Political Philosophy, edited by Thomas Christiano and John Christman (London: Wiley-Blackwell, 2009), p. 121. Mack, Eric. “Libertarianism.” In The Oxford Handbook of the History of Political Philosophy, edited by George Klosko (Oxford: Oxford University Press, 2011), p. 673. Mazure, Carolyn M., and Daniel P. Jones, “Twenty Years and Still Counting: Including Women as Participants and Studying Sex and Gender in Biomedical Research,” BMC Women’s Health 15 (2015), http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4624369/. McCarthy, David. “Liability and Risk,” Philosophy and Public Affairs 25, no. 3 (Summer 1996): p. 251. McKerlie, Dennis. “Rights and Risk,” Canadian Journal of Philosophy 16, no. 2 (Jun. 1986): p. 239. Mikhail, John. Moral Cognition: Rawls’s Linguistic Analogy and the Cognitive Science of Moral and Legal Judgment (Cambridge: Cambridge University Press, 2011). Mill, John Stuart. Principles of Political Economy (London: Parker, Son, and Bourn, 1865). Miller, David. “Justice and Property,” Ratio 21 (Jun. 1980): p. 1. Miller, Richard W. “Beneficence, Duty and Distance,” Philosophy & Public Affairs 32, no. 4 (2004): p. 357. Miller, Sarah. “Need, Care and Obligation.” Royal Institute of Philosophy Supplement 80 (2006): p.141. von Mises, Richard. Probability, Statistics and Truth, 2nd ed. (New York: The MacMillan Co., 1956).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



257

Moore, Michael. Placing Blame: A Theory of the Criminal Law (Oxford: Oxford University Press, 1997), ch. 5. Moore, Michael. Causation and Responsibility: An Essay in Law, Morals, and Metaphysics (Oxford: Oxford University Press, 2009). Moulin, Hervé. Axioms of Cooperative Decision Making (Cambridge: Cambridge University Press, 1988). Mueller, Dennis C. Public Choice III (Cambridge: Cambridge University Press, 2003). Musgrave, Richard A. “A Multiple Theory of the Budget Determination,” FinanzArchiv [Public Finance Analysis], New Series, Bd. 17, H. 3 (1956/1957): p. 333. Musgrave, Richard A. The Theory of Public Finance (New York: McGraw-Hill, 1959). Musgrave, Richard A., and Peggy B. Musgrave. Public Finance in Theory and Practice, 5th ed. (McGraw Hill, 1989). Nagel, Thomas. “Rawls on Justice,” The Philosophical Review 82, no. 2 (Apr. 1973): p. 220. Nagel, Thomas. “Libertarianism Without Foundations,” Yale Law Journal 85, no. 1 (Nov. 1975): p. 140. Nagel, Thomas. “Moral Luck,” in Moral Luck, edited by Daniel Statman (Albany, N.Y.: State University of New York Press, 1979). Narveson, Jan. The Libertarian Idea (Philadelphia: Temple University Press, 1988). Norcross, Alastair. “Comparing Harms: Headaches and Human Lives,” Philosophy and Public Affairs 26, no. 2 (Spring 1997): p. 135. Norcross, Alastair. “Contractualism and Aggregation,” Social Theory and Practice 28, no. 2 (Apr. 2002): p. 303. Nozick, Robert. “Coercion.” In Philosophy, Science and Method: Essays in Honor of Ernest Nagel, edited by Sidney Morgenbesser, Patrick Suppes, and Morton White (New York: Saint Martin’s Press, 1969), p. 440. Nozick, Robert. Anarchy, State, and Utopia (New York: Basic Books, 1974). Nozick, Robert. “Distributive Justice.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 177. Nozick, Robert. Anarchy, State, and Utopia, Rev. ed. (New York: Basic Books, 2013). Oates, Wallace E. “The Effects of Property Taxes and Local Public Spending on Property Values: An Empirical Study of Tax Capitalization and the Tiebout Hypothesis,” Journal of Political Economy 77, no. 6 (Nov.–Dec. 1969): p. 957. Oates, Wallace E. “The Many Faces of the Tiebout Model.” In The Tiebout Model at Fifty: Essays in Public Economics in Honor of Wallace Oates (Cambridge, Mass.: Lincoln Institute of Land Policy, 2006), ch. 2. O’Neill, Onora. “Nozick’s Entitlements,” Inquiry 19 (1976): p. 468. Otsuka, Michael. “Self-Ownership and Equality: A Lockean Reconciliation.” In LeftLibertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 149. Otsuka, Michael. “Risking Life and Limb.” In Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glen Cohen, Norman Daniels, and Nir Eyal (Oxford University Press, 2015), p. 77. Owen, David G. “The Fault Pit,” Georgia Law Review 26, no. 3 (Spring 1992): p. 703. Owen, David G. “Philosophical Foundations of Fault.” In Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), p. 201. Parfit, Derek. Reasons and Persons (New York: Oxford University Press, 1984). Parfit, Derek. “Justifiability to Each Person,” Ratio 16, no. 4 (Nov. 2003): p. 368. Parfit, Derek. On What Matters, vol. 1 (Oxford: Oxford University Press, 2011).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

258



Parfit, Derek. On What Matters, vol. 2 (Oxford: Oxford University Press, 2011). Van Parijs, Philippe. Real Freedom for All: What (If Anything) Can Justify Capitalism? (Oxford: Oxford University Press, 1995). Perry, Stephen. “Collective v. Distributive Justice.” In Oxford Essays in Jurisprudence, edited by Jeremy Horder, 4th series (Oxford: Oxford University Press, 2000), p. 237. Perry, Stephen R. “Responsibility for Outcomes, Risks, and the Law of Torts.” In Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge, Mass.: Cambridge University Press, 2001), p. 72. Perry, Stephen R. “Harm, History, and Counterfactuals,” San Diego Law Review 40 (2003): p. 1283. Pogash, Carol, and Solomon Moore. “California Officials Fear Abduction Case May Hurt Efforts on Parole,” New York Times, 30 August 2009, http://www.nytimes.com/2009/08/ 31/us/31abduct.html. Rakowski, Eric. Equal Justice (Oxford: Oxford University Press, 1991). Rawls, John. A Theory of Justice (Cambridge, Mass.: Belknap Press, 1971). Rawls, John. “Some Reasons for the Maximin Criterion,” American Economic Review 64, no. 2 (May 1974): p. 141. Rawls, John. “Kantian Constructivism in Moral Theory: The Dewey Lectures,” Journal of Philosophy 77, no. 9 (Sept. 1980): p. 515. Rawls, John. “The Basic Liberties and Their Priority.” In The Tanner Lectures on Human Values, vol. III, edited by Sterling McMurrin (Salt Lake City: University of Utah Press, 1982), p. 1. Rawls, John. “Social Unity and the Primary Goods.” In Utilitarianism and Beyond, edited by Amartya Sen and Bernard Williams (Cambridge: Cambridge University Press, 1982), p. 159. Rawls, John. “The Idea of an Overlapping Consensus,” Oxford Journal of Legal Studies 7, no. 1 (Spring 1987): p. 1. Rawls, John. Political Liberalism (New York: Columbia University Press, 1993). Rawls, John. “The Idea of Public Reason Revisited,” The University of Chicago Law Review 64, no. 3 (Summer 1997): p. 765. Rawls, John. The Law of Peoples (Cambridge, Mass.: Harvard University Press, 1999). Rawls, John. A Theory of Justice, Rev. ed. (Cambridge, Mass.: Belknap Press, 1999). Raz, Joseph. “Autonomy, Toleration, and the Harm Principle.” In Issues in Contemporary Legal Philosophy: The Influence of H.L.A. Hart, edited by Ruth Gavison (New York: Oxford University Press, 1987), p. 313. Raz, Joseph. “Numbers, With and Without Contractualism,” Ratio 16, no. 4 (Dec. 2003): p. 346. Reibetanz, Sophia. “Contractualism and Aggregation,” Ethics 108, no. 2 (Jan. 1998): p. 296. Reidy, David. “John Rawls.” In IVR Encyclopedia of Philosophy of Jurisprudence, Legal Theory and Philosophy of Law (London: Springer, 2008). Richardson, Henry. “The Stupidity of the Cost Benefit Standard.” In Cost Benefit Analysis: Legal, Economic, and Philosophical Perspectives, edited by Matthew D. Adler and Eric A. Posner (Chicago: The University of Chicago Press, 2001), p. 135. Ridge, Michael. “How to Avoid Being Driven to Consequentialism: A Comment on Norcross,” Philosophy and Public Affairs 27, no. 1 (Jan. 1998): p. 50. Ripstein, Arthur. Force and Freedom (Cambridge, Mass.: Harvard University Press, 2009). Ripstein, Arthur, and Benjamin C. Zipursky, “Corrective Justice in an Age of Mass Torts.” In Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), p. 214.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



259

Robins, James. “Should Compensation Schemes Be Based on the Probability of Causation or Expected Years of Life Lost?,” Journal of Law and Policy 12, no. 2 (2004): p. 537. Rodgers, Daniel T. Contested Truths: Keywords in American Politics Since Independence (New York: Basic Books, 1987). Roemer, John E. “A Challenge to Neo-Lockeanism,” Canadian Journal of Philosophy 18, no. 4 (Dec. 1988): p. 697. Rogerson, Kenneth F. “Rights at Risk,” Southwest Philosophy Review 1 (1984): p. 119. Ross, Stephen, and John Yinger. “Sorting and Voting: A Review of the Literature on Urban Public Finance.” In Handbook of Regional and Urban Economics, vol. 3, edited by Peter Nijkamp (Holland: Elsevier Science, 1999), ch. 47. Rothbard, Murray N. For a New Liberty: A Libertarian Manifesto (New York: Collier, 1973). Rothbard, Murray N. “Robert Nozick and the Immaculate Conception of the State,” Journal of Libertarian Studies 1, no. 1 (1977): p. 45. Rothbard, Murray. “Property and Exchange.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 214. Ryan, Cheyney. “Yours, Mine, and Ours: Property Rights and Individual Liberty,” Ethics 87, no. 2 (Jan. 1977): p. 126. Samuelson, Paul A. “The Pure Theory of Public Expenditure,” Review of Economics and Statistics 36, no. 4 (Nov. 1954): p. 387. Sartorius, Rolf. “Persons and Property.” In Utility and Rights, edited by R. G. Frey (Minneapolis: University of Minnesota Press, 1984), p. 196. Scanlon, T. M. “Liberty, Contract, and Contribution.” In Markets and Morals, edited by Gerald Dworkin, Gordon Bermant, and Peter Brown (Washington D.C.: Hemisphere, 1977), p. 43. Scanlon, T. M. “Contractualism and Utilitarianism.” In Utilitarianism and Beyond, edited by Amartya Sen and Bernard Williams (Cambridge: Cambridge University Press, 1982), p. 103. Scanlon, T. M. What We Owe to Each Other (Cambridge, Mass.: Harvard University Press, 1998). Scanlon, T. M. “How I am Not a Kantian.” In On What Matters, vol. 2, edited by Derek Parfit (Oxford University Press, 2011), p. 116. Scanlon, T. M. “Reply to Zofia Stemplowska,” Journal of Moral Philosophy 10, no. 4 (2013): p. 508. Schor, Juliet B., and Samuel Bowles, “Employment Rents and the Incidence of Strikes,” Review of Economics and Statistics 69, no. 4 (Nov. 1987): p. 584. Schroeder, Christopher H. “Corrective Justice and Liability for Increasing Risks,” UCLA Law Review 37, no. 3 (Feb. 1990): p. 439. Schroeder, Christopher H. “Corrective Justice, Liability for Risks, and Tort Law,” UCLA Law Review 38, no. 1 (Oct. 1990): p. 143. Schwab, Robert M., and Wallace E. Oates, “Community Composition and the Provision of Local Public Goods: A Normative Analysis,” Journal of Public Economics 44, no. 2 (Mar. 1991): p. 217. Schwartz, Gary T. “The Myth of the Ford Pinto Case,” Rutgers Law Review 43, no. 4 (Winter 1991): p. 1013. Schwinn, Steven D. “Toward a More Expansive Welfare Devolution Debate,” Lewis and Clark Law Review 9, no. 2 (Summer 2005): p. 311. Simmons, A. John. “Consent Theory for Libertarians,” Social Philosophy and Policy 22, no. 1 (Jan. 2005): p. 330.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

260



Simons, Kenneth W. “Corrective Justice and Liability for Risk-Creation: A Comment,” UCLA Law Review 38, no. 1 (Oct. 1990): p. 113. Singer, Peter. “Famine, Affluence, and Morality,” Philosophy and Public Affairs 1, no. 3 (Spring 1972): p. 229. Smart, J. J. C. “An Outline of a System of Utilitarian Ethics.” In Utilitarianism, For and Against, edited by J. J. C Smart and Bernard Williams (Cambridge: Cambridge University Press, 1973), p. 3. Smith, Samuel Stanhope. Princeton Lectures on Moral and Political Philosophy, vol. II (Trenton, N.J.: 1812). Spencer, Herbert. Social Statics, or, the Conditions Essential to Human Happiness Specified, and the First of Them Developed (London: John Chapman, 1851). Steiner, Hillel. “Capitalism, Justice, and Equal Starts,” Social Philosophy and Policy 5, no. 1 (Autumn 1987): p. 49. Steiner, Hillel. “Original Rights and Just Redistribution.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 74. Sterba, James. “Rawls and a Morally Defensible Conception of Justice.” In Morality in Practice, 7th ed., edited by James Sterba (Belmont, Calif.: Wadworth, 2003), p. 62. Stick, John. “Turning Rawls Into Nozick and Back Again,” Northwestern University Law Review 81, no. 3 (Spring 1987): p. 363. Stigler, George J. “The Tenable Range of Functions of Local Government.” In Federal Expenditure Policy for Economic Growth and Stability, Joint Economic Committee (Washington, D.C.: Joint Economic Committee, Subcommittee on Fiscal Policy, 1957), pp. 213‒19. Stone, Martin. “The Significance of Doing and Suffering.” In Philosophy and the Law of Torts, edited by Gerald J. Postema (Cambridge: Cambridge University Press, 2001), p. 131. Summers, Anita A., and Barbara L. Wolfe, “Do Schools Make a Difference?,” American Economic Review 67, no. 4 (Sept. 1977): p. 639. Tannenbaum, Julie. “Emotional Expressions of Moral Value,” Philosophical Studies 132, no. 1 (Jan. 2007): p. 43. Taurek, John C. “Should the Numbers Count?,” Philosophy and Public Affairs 6, no. 4 (Summer 1977): p. 293. Taylor, Michael W. Men Versus the State: Herbert Spencer and Late Victorian Individualism (New York: Oxford University Press, 1992). Tetlock, Philip., et al., “The Psychology of the Unthinkable: Taboo Trade-Offs, Forbidden Base Rates, and Heretical Counterfactuals,” Journal of Personality and Social Psychology 78, no. 5 (2000): p. 853. Thomson, Judith Jarvis. Rights, Restitution, and Risk: Essays in Moral Theory, edited by William Parent (Cambridge, Mass.: Harvard University Press, 1986). Tiebout, Charles M. “A Pure Theory of Local Expenditures,” Journal of Political Economy 64, no. 5 (Oct. 1956): p. 416. Union Pacific Rail Road Company v. Public Service Commission, 248 U.S. 67, 70 (1918). Urbina, Ian. “Paperwork Hinders Airlifts of Ill Haitian Children,” New York Times, 10 February 2010, http://www.nytimes.com/2010/02/09/world/americas/09airlift.html? scp=1&sq=haiti%20children%20airlift&st=cse. “U.S. Tort Costs: 2004 Update: Trends and Findings on the Cost of the U.S. Tort System,” http://www.legalreforminthenews.com/Reports/Tort_Costs_2004_Update.pdf.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi



261

Vallentyne, Peter. “Critical Notice of G.A. Cohen’s ‘Self-Ownership, Freedom and Equality,’ ” Canadian Journal of Philosophy 28, no. 4 (Dec. 1998): p. 609. Vallentyne, Peter. “Introduction: Left-Libertarianism—A Primer.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), vol. I, p. 1. Vallentyne, Peter, and Hillel Steiner, eds. The Origins of Left-Libertarianism: An Anthology of Historical Writings (New York: Palgrave, 2000), vol. I. Vallentyne, Peter, and Hillel Steiner, eds. Left-Libertarianism and its Critics: The Contemporary Debate (New York: Palgrave, 2000), vol. II. Verweij, Marcel. “How (Not) to Argue for the Rule of Rescue: Claims of Individuals versus Group Solidarity.” In Identified versus Statistical Lives: An Interdisciplinary Perspective, edited by I. Glenn Cohen, Norman Daniels, and Nir Eyal (New York: Oxford University Press, 2015), p. 137. Voorhoeve, Alex. “How Should We Aggregate Competing Claims?,” Ethics 125, no. 1 (Oct. 2014): p. 64. Wakker, Peter P. “Uncertainty Aversion: A Discussion of Critical Issues in Health Economics,” Health Economics 9, no. 3 (Apr. 2000): p. 261. Waldron, Jeremy. “Moments of Carelessness and Massive Loss.” In Philosophical Foundations of Tort Law, edited by David G. Owen (Oxford: Clarendon Press, 1995), p. 387. Wallace, Jay. “Scanlon’s Contractualism,” Ethics 112, no. 3 (Apr. 2002): p. 429. Weinrib, Ernest J. “Causation and Wrongdoing,” Chicago-Kent Law Review 63, no. 3 (1987): p. 407. Weinrib, Ernest J. “The Disintegration of Duty.” In Exploring Tort Law, edited by M. Stuart Madden (New York: Cambridge University Press, 2005), p. 143. Weinrib, Ernest J. The Idea of Private Law (Oxford: Oxford University Press, 2012). Wheeler, Samuel C., III. “Self-Defense: Rights and Coerced Risk-Acceptance,” Public Affairs Quarterly 11, no. 4 (Oct. 1997): p. 431. Wheeler, Samuel. “Natural Property Rights as Body Rights.” In Left-Libertarianism and its Critics: The Contemporary Debate, edited by Peter Vallentyne and Hillel Steiner (New York: Palgrave, 2000), p. 228. Wildasin, David E. Urban Public Finance (New York: Harwood Academic Publishers, 1986). Williams, Andrew. “Liberty, Equality and Property.” In The Oxford Handbook of Political Theory, edited by John S. Dryzek, Bonnie Honig, and Anne Phillips (Oxford: Oxford University Press, 2006), p. 488. Williams, Bernard. “Moral Luck.” In Moral Luck (Cambridge: Cambridge University Press, 1981), p. 20. Wiltshire, David. The Social and Political Thought of Herbert Spencer (New York: Oxford University Press, 1978). Wolff, Jonathan. “Risk, Fear, Blame, Shame and the Regulation of Public Safety,” Economics and Philosophy 22 (2006): p. 409. Wood, Allen. “Humanity as an End in Itself.” In On What Matters, vol. 2, by Derek Parfit (Oxford University Press, 2011), p. 58. Wood, Gordon S. “Conspiracy and the Paranoid Style: Causality and Deceit in the Eighteenth Century,” The William and Mary Quarterly 39, No. 3 (Jul. 1982): p. 401. Wright, Richard W. “Right, Justice and Tort Law.” In Philosophical Foundations of Tort Law, edited by David G. Owen (New York: Oxford University Press, 1995), p. 159.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 4/2/2020, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

Index aggregation inevitability of 2–4 intrapersonal 75–6 as method of utilitarianism 1–2 nonaggregative justifications for 33 nonconsequentialists’ covert reliance on 4, 15–16, 32–3, 42, 85, 148–9 nonconsequentialists’ opposition to 1–2 weighted 80–1, 108 See also nonaggregation; utilitarianism AIDS treatment example 115–17 Ambulance I dilemma 55–7, 63–4, 68–70, 72 Ambulance II dilemma 55–7, 63–4, 66, 68–70, 72 Ammons, A. R. 40 anarcho-libertarianism 158–9, 161–2 Arrow, Kenneth 173–4, 198 Ashford, Elizabeth 47, 127 Automatic Experiment 125 average utilitarianism. See utilitarianism Barnett, Randy 156, 164–5 benefits taxation. See under taxation Berlin, Isaiah 163 blame, public policy influenced by fear of 37–9 Brody, Baruch 177–8, 188, 194 Buchanan, James 156, 159, 164–5, 218–21, 228, 230–2, 234–5 bus driver example 35–6, 81–2 Calabresi, Guido 87–8 care. See duty of care causation, Enlightenment and mechanistic approach to 26–7 CBA. See cost/benefit analysis certain v. uncertain harms, assumed moral distinction between 13, 27–9, 33, 63–73, 89–91. See also risk certainty moral psychology and 71–2 statistical 10, 14–15, 30–1, 67–8 Chamberlain, Wilt 203–4, 208–11

choices. See tragic choices choice-sensitive egalitarianism 173–4 Christman, John 189, 192–3 claimants, identification of 109–11, 116–17 Coase, Ronald 8–9, 182–3, 185 Coase Theorem 137, 166 coercion 135–6, 246–7 Cohen, G. A. 171, 173, 179–80, 185–6, 188–9, 196–7 Coleman, Jules 13, 86 commons left-libertarian conceptions of 177–8, 191–4 libertarian conceptions of 171, 177–8, 191 Lockean proviso concerning 17–18, 130–1, 147–8, 177–9, 190–4, 199 commonsense morality blame and guilt in 27, 36 and cost/benefit analysis 23, 38 failures and costs of 19–21, 38 communitarianism 169 compensation Nozick on 138–41, 147–8, 171–2 prohibition in relation to 87–9 consent 131–2, 134–6, 158–9, 217–18, 246–7 consequentialism. See utilitarianism consequentialist libertarianism 164–5 contractualism aggregation as target of 58 epistemic moment for agreement in 54–6, 62, 64, 68–71, 73, 117–18, 121–6 hybrid models of 13–15, 59, 62–83 individualistic vs. collectivist motives in 126–7 normative loading of hypothetical “facts” in 41, 49–50, 68–9, 121 precautions against risk in 77–80 premises of 58 See also ex ante contractualism; Frick, Johann; Rawls, John, and Rawlsianism; Scanlon, T. M., and Scanlonian contractualism corrective justice 13, 36–7, 88 hindsight bias built into 36–7

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

264



cost/benefit analysis (CBA) duty of care/negligence as placeholder for 16, 78–80, 84, 101–4 moral opposition to 23, 31 Scanlonian contractualism and 79–80 culpability harm as a requirement for 89–94, 99–104 libertarianism and 85n.2, 149 negligence as a component of 101–4 deontology. See nonconsequentialism Difference Principle 17, 127–8, 149–50, 152–3, 155, 163, 167–71, 173–6, 178, 193–5, 212. See also Greater Burden Principle; maximin decision rule; sufficientarianism dominant protective association (DPA) 135–7, 139–40, 147–8 DPA. See dominant protective association due care. See negligence duty of care 85, 103–5 duty of easy rescue 45–7, 65, 111, 127 Dworkin, Ronald 173, 177–8, 189 egalitarianism choice-sensitive 173 left-libertarianism compared to 17–18, 178, 194, 196 Rawls and 155–6, 163–4, 167–9, 171, 173–4, 177–8, 193–5, 212–13 Ellsberg, Daniel 153–4 Enlightenment, mechanistic theory of causation in 26–7 Epstein, Richard 36–7, 94–5, 156, 164–5 ex ante contractualism distribution of risks as individual reasons 108, 111–13 epistemic point of view 13 ex post combined with 81–2 ex post distinguished from 106–9 natural veil of ignorance 117–21 relevant claimants under 109–11 stage-wise 122–5 exit determinative role in social contract theory 212–35 effect on price of public goods 19, 236, 239–50 internal 132, 135, 218–20 Nozick and 143–5 prohibition of 223–4 Rawls and 218–19, 223, 234 social contractarianism and 18–19, 162, 212–35, 247–50 suppressing constraints on 229–35 and Tieboutian sorting 238–45

ex post contractualism. See Scanlon, T. M., and Scanlonian contractualism external resources, appropriation of 177–8 Eyal, Nir 113 Fabian Socialism 201 fair distribution of risks assessing individual-level risks 113–17 as relevant individual reasons in ex ante contractualism 113 fault. See culpability Feinberg, Joel 104 First Amendment 101 Ford Pinto case 22–3, 31 freedom of speech 101 free-rider problem 236–7, 239 frequentist perspective 14–15, 30–1, 67–8 See also numbers Frick, Johann 3, 107, 110, 118–29 Fried, Charles 104 Friedman, Milton 164–5 fruits of labor ownership of, in left-libertarianism 187–8 ownership of, in (right) libertarianism 148, 173, 184, 186–8, 199–200, 205 Gauguin, Paul 93 Gaus, Gerald 156, 164–5 Gauthier, David 159, 164–5, 224–7, 230, 233 GBP. See Greater Burden Principle George, Henry 177, 191, 200–1, 205–6 Gibbard, Allan 177, 192–3 Goldberg, John 37, 90, 92–3, 100–2 Greater Burden Principle (GBP) contractualism, central to 108, 129 defined 61 distribution of risk as the relevant burden 111–17 ex ante contractualism and 30, 106–9, 127 ex post contractualism and 12, 34, 61, 70–1, 73, 79, 106–9, 113, 126–9 implausibility of 127–8 as pro tanto moral reason 76, 107–9, 123–4, 128–9 identifying relevant claimants under 9, 31, 40, 48–9, 51, 109–11 See also leximin decision rule; maximin decision rule Grunebaum, James 177, 183–5, 188–9 Hampton, Jean 95–7, 217–18, 235 Hansson, Sven Ove 26 harms culpability in relation to 89–94, 99–104 Mill’s harm principle 85

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 one-way causation 8 reciprocal nature of 8–9 serious vs. non-serious 80–1 Harsanyi, John 60, 151–3, 160, 173–4 Hart, H. L. A. 132, 135, 147, 164, 166 Hasnas, John 156, 158 Herman, Barbara 42 Hill, T. E. 1, 24 hindsight bias defined 12, 33–4 ex ante certainty as a form of 34–5 libertarianism and 37 moral evaluation based on 12–13, 16, 33–7, 60 nonconsequentialism as a form of 33 reasons for 34–5 See also moral luck; Scanlon, T. M., and Scanlonian contractualism Hirschman, Albert 239–40 Hobbes, Thomas 138, 222–3 Hohfeld, Wesley 8–9, 181–3 Holmes, Oliver Wendell, Jr. 8–9, 181, 183, 246–7 Honoré, Tony 92–3 Horton, Willie 37–9 human capital 208–11 Human Experiment 124–5 Hurley, Susan 153–4 identified vs. unidentified victims 33, 40–2, 47–9, 64, 66, 71–2, 110, 115–17 immanent critique. See under tort law injunctions, legal 91 intellectual property rights 186–7 invisible hand 135, 140–1 IPAs. See independent protective associations James, Aaron 107, 128–9 James, Edmund 205–6 Justice as Fairness 149, 155–6, 160–2, 167, 170–1, 174–5 Justice in Transfer 18, 147, 198–9, 201–11 just price 126–7, 203, 238, 245, 247 Kamm, Frances 55, 63–6, 68–9, 122–3 Kant, Immanuel, and Kantianism ends-in-themselves principle of 2, 19–20, 162–3 and epistemic perspectives 52 on moral evaluation 36n.29 opposition to aggregation 2 perfected will 27 and property rights 164 See also ex ante contractualism; Frick, Johann; Kamm, Frances; Rawls, John, and Rawlsianism; Scanlon, T. M., and Scanlonian contractualism

265

Kavka, Gregory 223–6, 230 Keating, Gregory 80 Kirzner, Israel 156, 164–5, 191 Kukathas, Chandran 164–5 Kumar, Rahul 60, 95–6, 118–19 Kymlicka, Will 184–5, 188–9 labor. See fruits of labor lay morality. See commonsense morality left-libertarianism basic commitments of 177–8 egalitarianism compared to 17–18, 177–8 egalitarian redistribution of wealth 194–6 emergence of 177 on fruits of labor 187–8 and left-Lockeanism 177 and the Lockean Proviso 190–6 political context for 178–9, 196–7 on property rights 177–8, 181, 183, 190–6 right-libertarianism compared to 17–18, 130–1, 177–8 and self-ownership 178–81, 184, 190–2, 196–7 and the state 130–1 and taxation 188–9 left-Lockeanism and distributive justice 178–9 on fruits of labor 177 and taxation of land 191 on value of property 205–9 Legal Realism, and reconceptualization of rights 8–9, 181–3 Lenman, James 67–9, 81–2 leximin decision rule defined 149n.2 Rawls’s specification of 153 social distribution in accordance with 17, 149 See also Greater Burden Principle; maximin decision rule liability rule as measure of compensation 136–9. See also compensation libertarianism anarcho- 158, 161–2 and collective action 172 and collective obligations to the poor 172 consequentialist 164–6 critique of Rawls 157, 159–62, 165, 169, 173–4, 176 exit as issue for 144–5, 215, 229–35 and hindsight bias 37 left-libertarianism compared to 17–18, 130–1 and Lockean proviso 17–18, 130–1, 147–8, 171, 191 Rawls and 130–1, 155–7, 163–5, 169–75

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

266



libertarianism (cont.) on scarcity rents 201–2 and self-ownership 17–18, 164, 179–81, 183–5 and the state 16–17, 130, 158–9, 172, 245 and taxation 180, 183–4, 187–9, 236–7 See also left-libertarianism; Lomasky, Loren; Nozick, Robert liberty basic liberties 164–6 conflicting individual interests in 166–7 lexical priority of 163, 166–7 negative vs. positive 163–4, 184–5 Rawls and 162–7 Lifeboat I dilemma 50, 53–4, 109–10 Lifeboat II dilemma 50–4 locational surplus 244–5, 247–50. See also scarcity rents Locke, John, and Lockeanism and property 164–5, 199 proviso on natural resources (see Lockean proviso) and rectification of injustices 171–2 and rights 132, 134–5, 145–6, 155, 159 and self-ownership 17–18, 130–1, 179–80 social contractarianism and 222–3 and state of nature 138, 141 See also left-libertarianism; left-Lockeanism; Nozick, Robert Lockean Proviso “enough, and as good,” meaning of 199 left-libertarianism and 177–9, 190–4 Nozick and 147–8, 199 Lomasky, Loren 156–7, 159–61, 163–6, 169–70, 172 lorry driver example. See bus driver example Luban, David 25n.6, 47n.15 luck egalitarianism. See choice-sensitive egalitarianism Mack, Eric 156, 162, 164–5 Marx, Karl 200–1 Mass Vaccination (Known Victims) 110–11, 119–22 maximin decision rule AIDS treatment example 116 contractualism and 30, 34, 61, 73, 106–7, 109, 127 defined 106, 109, 149n.2 implausibility of 127–8 leximin compared to 149n.2 rational choice justification for 150–4, 167–8 Rawls’s use of 149–53, 167–8, 170, 175

relative nature of 109, 127 See also Difference Principle; Greater Burden Principle; leximin decision rule; sufficientarianism McCarthy, David 94 Melamed, A. Douglas 87–8 Melville, Herman 186–7 Mill, John Stuart 85, 177, 183, 191, 199–200 von Mises, Richard 114, 164–5 monopolies, and scarcity rents 201 Moore, G. E. 35n.26 Moral gridlock 10, 22, 30, 35, 45–6, 55, 61, 63n.9, 69–70, 72, 76, 79–80, 94 moral luck 12, 35–7, 82, 93, 99–100. See also hindsight bias moral psychology 41–3, 71–2 Musgrave, Richard 237, 239, 242 Nagel, Thomas 36, 41, 158 Narveson, Jan 156, 159, 164–5 natural resources. See commons natural veil of ignorance 117–19 negligence as cost-benefit analysis manqué 32, 103–4 meaning of 15–16, 84, 101–4 nonconsequentialism and 15–16 See also cost/benefit analysis von Neumann/Morgenstern utility function 152–3 New Liberals (Britain) 201 nightwatchman state. See state: minimal (nightwatchman) nonaggregation basic commitment of nonconsequentialism 1–3 defined 5 futility of, as a basic commitment 3–4 See also aggregation; nonconsequentialism; risk nonconsequentialism defined 5 ex and post ante epistemic perspectives of 11–12 future of 19–21 hegemony of 6 hybrid models of 5, 13–15, 56, 63–73, 82–3 public policy choices and 37–9 See also contractualism; left-libertarianism; libertarianism; nonaggregation; Nozick; Rawls; risk nonsacred values 23–4 Nozick, Robert 5 as closet utilitarian 17, 133, 141–2, 148 on compensation 138–41, 147–8, 171–2 on consent 131–2, 134–6, 139–43, 141n.43, 147–8, 158–9, 172

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 critique of Rawls 155, 160–1, 167–8, 170, 212 independent protective associations (IPAs) 133–40, 147–8 Justice in Acquisition 146–7, 198–9, 201–3 Justice in Transfer 18, 147, 198–9, 202–11 libertarianism of 1, 131, 133, 135–6, 140–3, 146, 148, 155, 164–5, 171–2, 177–8 Lockean Proviso, interpretation of 147–8, 199 minimal state, derivation of 133–48, 172 Rawls compared to 17, 155–8, 169 on risk as problem for nonaggregation 31–2, 37, 97, 142 and scarcity rents 198–202, 206–11 and the state 16–17, 130–1, 133–43, 159, 167, 172 on taxation 131–2, 146–7, 180, 186–9, 204, 206–7 utopia as choice among coercive polities 143–6 numbers permissible vs. impermissible risk based on 10 small vs. large 67–8 tiebreaker function of 76–7 See also aggregation; frequentist perspective; probabilistic perspective Oates, Wallace 239 Obama, Barack 20 Ogilvie, William 177 Original Position 11, 17, 61, 150–4, 157–62, 165, 168, 170, 223–4 Otsuka, Michael 13, 156, 177, 187–9, 195–6 Paine, Thomas 177 Parfit, Derek on epistemic perspectives 49–54 on trolley problems and Taurek’s dilemma 42–3, 47–8, 50–4 on uncertainty as condition for moral choices 10, 13, 28–9, 47–8 prioritarianism 108, 127–8 property rule as measure of compensation 136–9. See also compensation psychology moral viii, 41–3, 71–2 in Rawls’s theory of justice 152–3, 170–1 uncertainty aversion as matter of 153–4 public goods, provision and financing of 236–50 public policy fear of blame as influence on 37–9 nonconsequentialism’s effect on 37–9 Railton, Peter 35n.26 Rand, Ayn 164–5

267

Rational Impartialism 53 Rawls, John, and Rawlsianism conception of the good 165, 175 critiques of 6 determinate persons 157, 160–2 Difference Principle 127–8, 149–55, 163, 167–71, 173–6, 178, 193–5, 212 Difference Principle vs. sufficientarianism 152, 173–5 egalitarianism of 157, 163–4, 167–71, 173–4, 177–8, 193–5, 212–13 exit options in original position 212, 218–19, 223, 234 ex post perspective in 167–9 and individual responsibility 173 infinite risk aversion, assumption of 151–2 insulation of, from criticism 6–7 Justice as Fairness, principle of 149, 152–3, 155–6, 160–2, 167, 170–1, 174–5 libertarian critique of 157, 159–62, 165, 167–9 and libertarianism 155–7, 163–5, 169–75 and liberty 162–7 original position of 17, 61, 150–4, 157–62, 165, 168, 170, 223–4 on risk vs. radical uncertainty 150–1 and strains of commitment 170–1 and uncertainty aversion 153–4 and utilitarianism 1, 17, 21, 32, 151–3, 155–6, 160, 166–7, 175–6 and veil of ignorance (see original position of ) Raz, Joseph 1–2 reasonable precautions 15–16, 32, 65–6, 77–80. See also negligence Reibetanz, Sophia 70 Ricardo, David 200–1 Ridge, Michael 45 right-libertarianism. See libertarianism right-Lockeanism. See libertarianism rights co-possibility of 136–8 Legal Realist conception of 8–9 Nozick’s conception of 131–3 protection of 2 and rule-utilitarianism 3n.9 risk, imposition of aggregation in relation to 9–16, 24–5, 29–33, 39, 45–6, 63–73 completed harm, as a 94–7 defined 4n.11 fair distribution of 112–17 morality in relation to 27–9, 94–7 nonconsequentialism and 11–16, 29–37, 45–6, 94–9, 142

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

268



risk, imposition of (cont.) precautions against 77–80 probabilistic v. frequentist presentations of 14–15, 67–8 Rawls’s maximin rule as response to 150–4 Scanlonian contractualism and 63–73 See also hindsight bias risk aversion 150–3 Rothbard, Murray 156, 158 routine tradeoffs 23–4 sacred values 23–4 Samuelson, Paul 239 Scanlon, T. M., and Scanlonian contractualism certain vs. uncertain harms in 62–82 commentary on Parfit 42, 50–2 concept of the self in 12–13, 126–7 ex post and ex ante epistemic perspectives of 12–13, 32, 54–5, 59–82 Greater Burden Principle of (see Greater Burden Principle) and hybrid models 13, 62–82 hypothetical moment of agreement in 54–6, 62, 64, 68, 71–3, 117–18, 121–6 modified contractualism of 13, 59, 106 on Taurek’s dilemma 50–2 tiebreaker rule 50–1, 70, 76–7 unavoidability of aggregation 59, 82–3 unequal distribution of ex ante risk 73–5 See also contractualism; ex ante contractualism; trolley problems scarcity centrality of, to morality 7–8 defined 7 nonconsequentialism’s incapacity to deal with viii, 9, 19–21, 43–5 ubiquity of viii, 7, 9, 19–20, 39, 44–5, 183 scarcity rents distributive justice and 188, 199–202 in economic history 200–1 Nozick on 198–202 society’s claim on 205–7, 236 and Wilt Chamberlain 208–11 See also left-libertarianism; libertarianism; locational surplus; Nozick, Robert Schmidtz, David 156 self-ownership control of body and 185 labor and 186–7 left-libertarianism on 17–18, 178–9, 181, 183–92, 196–7 negative vs. positive definitions of 184–5

right-libertarianism on 17–18, 146–7, 164, 180, 183–90 and taxation 146–7, 172, 180–1, 183–90 Sen, Amartya 177–8 Simmons, John 156, 158 Singer, Peter 47, 127 Smart, J. C. C. 35n.26 Smith, Samuel Stanhope 26 social contractarianism exit as issue for 18–19, 213–35 individual endowments in 162 prospective 216–17, 221–9 and the redistributive state 18–19 retrospective 216–21 tenets of 18–19 SON. See state of nature Spencer, Herbert 177–8, 191, 200–1, 205–6 stage-wise ex ante contractualism 122–5 state emergence of, from state of nature 133–43 individuals’ choices among competing polities 143–6 (see also Tiebout, Charles; Tieboutian sorting) left-libertarianism and 130–1 libertarianism and 16–17, 130, 158–9, 172, 245 market-mimicking behavior of 238–50 minimal (nightwatchman) 2, 16–17, 130–1, 133–8, 140–1, 143–8, 159, 167, 172 state of nature (SON) Crusoeian 214–15, 226–8 exit options and 212–16, 222–3, 225–6 Hobbesian 138, 221–3, 226–8 Lockean 138, 222–4, 226–8 Nozick and 133–5, 138, 140–1 Steiner, Hillel 156, 177, 184–5, 191–2, 194 Stick, John 169 sufficientarianism 149, 152–4, 173–4 taboo tradeoffs 23–4 tacit consent 132, 217–18 Taurek, John 40, 50, 76–7, 109–10 taxation of appreciation in value 204–9 benefits taxation 19, 236 income vs. endowments tax 188–9 of labor income 187–90 left-libertarianism on 188–9 libertarianism on (see under libertarianism) Nozick on (see under Nozick) progressive 188–9, 237 as slavery 188–9, 196–7 as theft 180, 183–4, 187–8 transfer policy in relation to 195–6

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 8/2/2020, SPi

 tax hypothecation 237 Tetlock, Philip 23 Thomson, Judith 90, 92–4, 98, 101–3 threshold deontology 14, 80 Tideman, Nicolas 191–2 Tiebout, Charles 239–44 Tieboutian sorting 143–4, 238–45, 247 tiebreaker rule. See under Scanlon, T. M., and Scanlonian contractualism tort law 13, 84 completed harm as a prerequisite for liability 89–94 defined 84 immanent critique of 86–8, 90–1, 100 and moral luck 92–4, 99–100 negligence standard (see negligence) nonconsequentialism and 84–104 prohibition vs. compensation in 87–9 risk as a completed harm 94–7 risk regulation and 88, 91–2 risk vs. harm in 89–94, 97–9 strict liability 36n.28, 93, 100, 142 tradeoffs categories of 23–4 inevitability of 3, 7–9, 19–20, 39, 43–5, 82–3 nonconsequentialists’ failure to recognize ubiquity of 7–16 secondary factors in 15, 105 See also tragic choices tragic choices features of hypotheticals about viii, 2–4, 9–10, 24–5, 34–5, 40, 42, 108 preventability of 43–4 in real life 20, 24, 41, 43–4 scarcity and 9, 39, 45 secondary factors in 15 under- and overestimation of 43–4 See also tradeoffs; trolley problems trolley problems features of 2–3, 40, 42 influence of, in moral philosophy 2–3, 24–5, 40–1, 43–4, 72 marginality of, in real life 40–1, 43–4, 105 normative loading of hypothetical “facts” in 41, 48–9, 54–5 Parfit on 42–3 premises (set-ups) of 41–3, 54–5 psychological factors in 41–3, 71–2 relevant claimants in 109–10

269

secondary features of 15, 41, 46–8 utilitarian approach to 55–6 See also tragic choices uncertainty. See risk uncertainty aversion 153–4 undue risk. See negligence Uniform Commercial Code 245 utilitarianism average 149, 151–4, 160, 170, 172–6 government-house 14 left-liberal Kantianism, overlap with 21 nonconsequentialists’ opposition to 1–2 Nozick and 17, 133, 141–2, 148 Rawls and 1, 17, 21, 32, 151–3, 155–6, 160, 175–6 and trolley problems 55–6 See also aggregation utopia, Nozick’s version of 143–6 vaccine examples 37–8, 67, 98–9, 110–11, 119–22 Vallentyne, Peter 156, 177, 180, 184–5, 187, 192–3, 195–6 Value of Choice model 77–8 Vanderbilt, Cornelius 206 Van Parijs, Philippe 177, 184–5, 188–9, 192–6 veil of ignorance. See natural veil of ignorance; Original Position victims identifiability of 25, 42, 47–8, 63, 105, 116–17 proximity and familiarity of 47–8, 116–17, 120–1 Voorhoeve, Alex 120–1 Wallace, Jay 2 weighted aggregation 80–1, 108 Weinrib, Ernest 90 welfarism. See utilitarianism Wheeler, Samuel 180–1 will 26–7 Williams, Andrew 173 Williams, Bernard 35–6, 40, 81, 93 Wolf, Susan 42 Wood, Allen 42–4 Wood, Gordon 26 Zipursky, Benjamin 37, 90, 92–3, 100–2