Oxford Studies in Political Philosophy Volume 5 [5] 2019931471, 9780198841425, 9780198841432, 9780192578433

Citation preview

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

OXFORD STUDIES IN POLITICAL PHILOSOPHY

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved. Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Oxford Studies in Political Philosophy Volume 5

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Edited by D AV I D S O B E L, PE T E R VA L L E N T Y N E, A N D S T EV E N WA L L

1

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

1 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 © the several contributors 2019 Te moral rights of the authors have been asserted First Edition published in 2019 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

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Library of Congress Control Number: 2019931471 ISBN  978–0–19–884142–5 (hbk.) ISBN  978–0–19–884143–2 (pbk.) Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY 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.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Acknowledgments

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Tis is the ffth volume of Oxford Studies in Political Philosophy. Te chapters assembled here were frst presented as papers at a workshop in Tucson, Arizona in October 2017. We would like to thank all those who attended this event, with special thanks to Rosie Johnson, who oversaw most of the organization. All of the chapters in this volume were reviewed by referees, most of whom serve on the editorial board of Oxford Studies in Political Philosophy (see http://www.oxfordstudiespoliticalphilosophy.org). We very much thank these referees for their eforts in helping to make this ffth volume a success. Tanks also to the Center for the Philosophy of Freedom at the University of Arizona for providing funding for the workshop. Finally, we would like to express our gratitude to Peter Momtchilof for supporting this series and for his expert guidance.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved. Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Contents List of Contributors

ix

Part I:  Power and Legitimacy 1. Power and Equality Daniel Viehof 2. Rescuing Public Justifcation from Public Reason Liberalism Fabian Wendt 3. Injustice, Reparation, and Legitimacy Stephen Galoob and Stephen Winter 4. Justifying Uncivil Disobedience Ten-Herng Lai

3 39 65 90

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Part II:  Political, Legal, and Moral Relations 5. Discrimination and Subordination Sophia Moreau 6. Protecting Vulnerable Languages: Te Public Good Argument Alan Patten 7. Localized Restricted Aggregation Victor Tadros

117

Index

205

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

147 171

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved. Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

List of Contributors

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Stephen Galoob is Associate Professor at the University of Tulsa College of Law. Ten-Herng Lai is a PhD student in the School of Philosophy at the Australian National University. Sophia Moreau is Associate Professor in the Faculty of Law and the Department of Philosophy at the University of Toronto. Alan Patten is Howard Harrison and Gabrielle Snyder Beck Professor of Politics at Princeton University. Victor Tadros is Professor of Criminal Law and Legal Teory at the University of Warwick. Daniel Viehof is Assistant Professor of Philosophy at New York University. Fabian Wendt is Research Associate at the Smith Institute for Political Economy and Philosophy at Chapman University in Orange, California. Stephen Winter is Senior Lecturer in Politics and International Relations at the University of Auckland.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved. Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

PA RT I

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

P OW E R A N D L E G I T I M A C Y

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved. Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

1

Power and Equality

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Daniel Viehof

A number of democratic theorists have recently sought to vindicate the ideal of political equality (that is, the ideal of an equal distribution of political power) by tying it to the intrinsic value of egalitarian relationships. According to these “social” or (as I will usually say) “relational egalitarian” arguments for distributing political power equally, such a distribution is an essential component of certain intrinsically valuable relationships, and required for ours to be a “society of equals.”1 Te motivation for adopting such a relational egalitarian account of political equality is twofold. Te frst is a matter of “ft.” Many citizens of democratic societies accept that there is distinctive value in democratic decisionmaking. Similarly, many citizens accept that there is distinctive authority associated with democratic decisions. Neither this value nor this authority seems to be fully accounted for by appeal to procedure-independent outcome considerations. Instead they appear to depend on the egalitarian character of democratic procedures: making decisions as equals is intuitively of independent moral signifcance. Yet articulating what the signifcance of egalitarian procedures consists in, in a way that accommodates its (at least partial) independence from non-procedural considerations, has been difcult. Relational egalitarian arguments, many of their proponents think, provide a relatively straightforward explanation of why procedurally egalitarian decision-making matters.

1  (Schefer 2015), p. 21. Relational (or “social”) egalitarian arguments for democracy or political equality are suggested in, e.g., (Anderson 1999, 2010, 2012; Kolodny 2014a, 2014b; Viehof 2014; Schefer 2015). Tough Tomas Christiano’s argument for democracy, in (Christiano 2008), shares some features with relational egalitarian accounts, it is sufciently diferent not to be easily subsumed under this header, and so I will set it aside here.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

4

Daniel Viehof

But relational egalitarian accounts do not merely ft existing intuitions about the importance of political equality. Tey also (and this is the second reason for adopting them) promise to provide independent support for our commitment to this ideal. One of the main challenges in defending procedural egalitarian commitments is to escape the worry that one has simply restated, in slightly diferent terms, the very democratic intuition one is trying to justify. Relational egalitarian arguments avoid this concern by highlighting these commitments’ continuity with other values we care about outside of politics narrowly conceived. Even those who are not already wedded to democratic procedures, or who are uncertain of their democratic commitments, may recognize that equality is an ideal central to many of our relationships. If that ideal carries over—directly or indirectly— from these relationships to our political arrangements, and if it requires an egalitarian distribution of decision-making power, then this could provide independent support for democratic procedures and the demands they make on us. I am sympathetic to the relational egalitarian approach. And yet I have come to think that vindicating the ideal of political equality on its basis is more challenging than has often been recognized. To explain what the challenge consists in is the purpose of this chapter. I begin, in Section 1, by explaining what the project of vindicating the ideal of political equality amounts to. Section 2 outlines the basic structure of the relational egalitarian argument for political equality, and highlights a signifcant ambiguity in it. Two diferent paradigmatic examples of egalitarian relationships commonly underpin these arguments for democracy: that of an egalitarian society, a society in which everyone has equal social status (rather than the kind of unequal status we associate with hierarchical societies governed by, e.g., caste or class structures); and that of egalitarian relationships, such as friendships or marriages among equals. Tese two examples, though plausibly related, are not neatly aligned. And, I argue in Sections 3 to 6, they have diferent implications for the distribution of power, and the applicability of relational egalitarian intuitions to our political community. While egalitarian rela­ tionships like friendship do include a positive requirement of equal power, the ideal of equal status does not. It merely demands that unequal power be socially justifed in some ways (ways that are compatible with our basic moral equality) and not others (ways that are not). And while the ideal of equal status straightforwardly applies to large political communities, it is open to doubt whether the ideals associated with friendship do; and even if these doubts can be overcome (or at least kept in check), the resulting picture makes the value and authority of democratic institutions much more conditional on the actual attitudes

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Power and Equality

5

of citizens (historic and contemporary) than defenders of the ideal of political equality may have hoped for. 1.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Political equality is a matter of how political power is distributed among the members of a particular group. Political power is constituted by the opportunity to infuence political decisions, which usually take the form of laws and other directives that are regularly coercively enforced against, or widely considered binding for, the group’s members.2 So to have equal political power is to have an equal opportunity to infuence political decisions that apply to one’s group.3 What does it take to vindicate the ideal of political equality, by which I mean, vindicate that political equality is an ideal or value in its own right? It is not enough to show that egalitarian political institutions (institutions which distribute political power equally) are in fact valuable, as their value could derive from considerations that are quite independent of political equality. As Steven Wall has pointed out, “For the ideal of political equality to be vindicated, it must be shown to be more than a mere by-product of a sound justifcation.”4 Tis means, for instance, that a vindication of political equality cannot rest on purely instrumental defenses of democracy: even if these defenses could establish that some egalitarian distribution of decisionmaking power would best bring about good outcomes (suitably specifed), the value of the egalitarian distribution would be derived from the value of the outcomes, which is specifable without reference to political equality.

2  So not all power is political power, and a commitment to equal political power need not go hand in hand with a commitment to equal power more generally. But our concern with equal political power is plausibly not unrelated to a broader concern with equal power, and an account of political equality and its value should elucidate that relation. 3  Two points are worth fagging. First, the opportunity to infuence, rather than actual infuence, is what matters here because someone may have equal power yet fail to exercise it. Second, an opportunity to infuence must be distinguished from an opportunity to acquire an opportunity to infuence. If I can only vote at time t2 if I register at time t1, then I have an opportunity at t1 to acquire the opportunity to infuence the decisions at t2. But this doesn’t mean that I have the power at t1 to infuence the decision. And if I fail to register at t1, I lack the opportunity to infuence the decision at t2, and thus lack the relevant power. Tis is a conceptual point about power, separable from the normative question whether my having, but not using, an (equal) opportunity to register at t1 bears on whether I can complain that I lack (equal) political power at t2. 4  (Wall 2007), p. 417.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

6

Daniel Viehof

But even among theories that treat political equality as more than a mere by-product, it is worth drawing a distinction between those that treat political equality as an ideal in its own right, and those that do not. What would it be to treat political equality as more than a mere by-product and yet not as an ideal in its own right? On some views, equality simply sets a moral baseline from which distributions of political power must start. If there is no (adequate) reason for distributing power diferently—to move away from the baseline—then there is reason to distribute it equally. (In Isaiah Berlin’s words, “equality needs no reasons, only inequality does so . . .”5) But though equality is (on such views) special because it sets the baseline, and any move away from it requires justifcation, it is also nothing but a baseline. If there is a good reason to move away from the baseline—a good reason for an unequal distribution—then equality does not provide a countervailing reason to stick (or remain close) to an equal distribution. Putting the point slightly technically: On the baseline view, the presence of reasons for an unequal distribution does not simply outweigh the reasons we have to distribute power equally. Rather, insofar as equality is nothing but a baseline, the presence of suitable considerations favoring inequality cancels the reason we would otherwise have had to distribute power equally. Equality, in such cases, can make a non-instrumental contribution to the realization of some non-derivatively valuable good; but it is not itself an essential component of that good, insofar as that good can in principle be realized even under conditions of inequality. To make this quite abstract point more concrete, consider an infuential position in democratic theory with such a “baseline” structure: David Estlund’s argument for democracy by appeal to a “reasonable acceptability requirement,” and in particular his proposal that democracy is distinctly acceptable because its justifcation can avoid making “invidious comparisons” among citizens.6 As some critics have pointed out, Estlund builds into his account of political justifcation a basic asymmetry between unequal and equal relations of rule.7 Tus, when Estlund concludes that a democratic—egalitarian— distribution of political power is acceptable where a non-egalitarian is not, the endorsement of political equality is not a mere by-product of a justifcation that is otherwise unconcerned with an equal distribution of power. Nonetheless, what Estlund is ultimately concerned with is not whether power is distributed equally, but whether its distribution can be justifed 5  (Berlin 1999 [1956]), p. 84. 6  (Estlund 2008), p. 37: “[i]nvidious comparisons purport to establish the authority and legitimate power of some over others in ways that universal sufrage does not, and so invidious comparisons must meet a burden of justifcation that universal sufrage does not.” 7  See, e.g., (Arneson 2009) and (Kolodny 2014a).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

7

to all qualifed points of view. So if an unequal distribution can be justifed without invidious comparison, and is acceptable to all qualifed points of view, the fact that the distribution deviates from standards of equality is not regrettable, because an equal distribution of power is not a value in its own right. By contrast, on other views, an equal distribution of political power is not simply a baseline, nor a mere by-product, but instead an ideal in its own right. On such views, there are non-instrumental reasons in favor of distributing power equally; and these reasons survive the presence of reasons against doing so. Many democratic theorists believe that these reasons in favor of political equality prevail against most competing reasons in favor of an unequal distribution of power. For the purposes of clarifying the conceptual point at issue, however, this is less important than another observation: even if the reasons for distributing political power unequally prevail, they do not cancel the reasons favoring political equality. Tey merely outweigh them. And so there is something to regret where we cannot realize simultaneously the value that speaks in favor of political equality and the value that speaks in favor of political inequality. On such a view, equality is either itself a nonderivatively valuable good, or (more plausibly) an essential component of such a good. In either case we can sensibly think of it as being an ideal in its own right, insofar as whatever gives us reason to realize equality can itself not be understood without it.8 Tis distinction, between views that treat equality as a mere by-product, a mere baseline, or an ideal in its own right, seems to me of general theoretical interest for thinking about political equality (and indeed equality more generally). But, more importantly for the purposes of this chapter, the distinction is relevant because, as I understand them, relational egalitarian arguments for political equality generally aspire to vindicating it as an ideal in its own right.9 Indeed, it may plausibly be among the main motivations for relational egalitarian views that they promise to establish something more than a mere by-product or baseline justifcation of equality (political and other). I do not purport to show here that this aspiration is worth sharing. I merely mean to point out that it sets a standard against which to assess the success of relational egalitarian arguments. 8  So to say that political equality is an ideal in its own right is not to say that it may not be in some sense derivative of some other good, as long as it is also the case that a complete specifcation of that other good makes essential reference to political equality. See (Viehof 2017). 9  I take this aspiration to be present, for instance, in both (Kolodny  2014a) and (Viehof  2014). More generally, insofar as relational egalitarians are (at least in part) concerned with establishing democracy’s authority, a mere baseline view will generally be inadequate, for reasons briefy discussed at the end of Section 4.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

8

Daniel Viehof 2.

Te relational egalitarian account of political equality rests on the following line of thought:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(1) Relational Equality: Certain kinds of egalitarian relationships have non-derivative value. (2) Equal Power: A (roughly) equal distribution of (some forms of ) power among the parties is an essential component of such relationships. (3) Political Relationships: Our political community should instantiate relationships of this sort. (4) Political Equality: So (some forms of ) power should be distributed equally among the citizens. Where it is, the institution has special value (Democracy’s Value) and special authority (Democracy’s Authority). As it stands, this is evidently incomplete. In particular, even if (1), (2), and (3) are true, it does not yet follow that we should distribute political power equally because an equal distribution of power, though necessary, may not be sufcient for the instantiation of non-derivatively valuable egalitarian relationships. Under what conditions Political Equality does follow will depend on a more detailed account of egalitarian relationships and their instantiation conditions. I will briefy return to this toward the end of this chapter. But before I can get there, I need to discuss in more detail (1), (2), and (3). Let me begin with Relational Equality. Te starting point of the relational egalitarian approach is the observation that certain egalitarian relationships have non-derivative value. Tus Elizabeth Anderson has argued that egalitarians are fundamentally committed “to creat[ing] a community in which people stand in relations of equality to others.”10 According to Samuel Schefer, “equality is an ideal governing certain kinds of interpersonal relationships,” and egalitarians should care about “the establishment of a society of equals, a society whose members relate to one another on a footing of equality.”11 And the editors of a recent volume on relational (or, as they say, “social”) equality ofer the following characterization of the position their book elucidates: “[E]quality is foremost about relationships between people . . . When we appeal to the value of equality, we mean the value primarily of egalitarian and nonhierarchical relationships.”12 I am sympathetic to the thought that equality is a constitutive component of certain non-derivatively valuable relationships, and that societies in which the relevant form of equality is instantiated realize an ideal of which other societies, which do not instantiate it, fall short. But these claims, even if 10  (Anderson 1999), p. 289.    11  (Schefer 2015), p. 21. 12  (Fourie, Schuppert et al. 2015), p. 1.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

9

true, are open to signifcantly diferent interpretations. To see this, consider the two quite diferent sets of examples from which discussions of relational equality commonly start. One case to which relational egalitarians regularly appeal to illustrate the ideal of relational equality is that of a society not governed by social hierarchies assigning positions of inferiority or superiority to diferent people. Tus David Miller invokes the ideal of a society “that is not marked by status divisions such that one can place diferent people in hierarchically ranked categories, in diferent classes for instance.”13 Niko Kolodny, when introducing the idea that “in virtue of how a society is structured, some people can be . . . ‘above’ and others ‘below’,” ofers some paradigm cases of problematic social hierarchy: “Te servant is ‘subordinate’ to the lord of the manor, the slave ‘subordinate’ to the master . . . Te plebian is ‘lower than’ the patrician, the untouchable ‘lower than’ the Brahmin, and so on.”14 At their most extreme, such caste societies (as I will, for ease of reference, call societies that paradigmatically violate the ideal of equality Miller, Kolodny, and others are concerned with) assign a place in the hierarchy based on parentage or similar features beyond a person’s control.15 But caste societies, in the sense at issue here, may exist even where someone had control over the fate that led them to be assigned a lower rank on the social ladder. (Consider societies permitting peonage, in which people essentially discharge their debts by selling themselves into temporary slavery, and are viewed as equivalent to slaves while the peonage relation lasts.) Te contrast to such a caste society is then a society that assigns equal social status to all citizens, and disallows inequalities that would be incompatible with it. Another case often invoked by proponents of relational equality is a wellfunctioning friendship or similar relationship.16 Friendship and (at least more recently, and in some societies) marriage are commonly seen as quintessentially egalitarian relationships.17 We have a reasonably straightforward grasp of the ideal that friends should be one another’s equals, and we can think of a 13  (Miller 1997), p. 224. 14  (Kolodny 2014b), p. 292. See (Anderson 2012), p. 40, for a more detailed list of historically signifcant forms of social inequality. 15  Elizabeth Anderson refers to a specifc prohibition on consigning people “to inferior ofce on the basis of identities or statuses imputed at birth” as “the anticaste principle.” (Anderson 2012), p. 106. I use the notion of a caste society in a more general fashion. 16  Friendship, marriage, etc. are discussed in some detail by (Schefer 2015), Sect. 1.2, (Viehof 2014), Part IV. Even those who do not discuss them in detail recognize these relationships as examples that fall within the general purview of relational equality. See, e.g., (Kolodny 2014b), p. 304. 17  For a thoughtful discussion of friendship’s egalitarian character (that does, however, overemphasize the signifcance of consensus among friends), see (Mansbridge 1980), pp. 8–14.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

10

Daniel Viehof

variety of ways in which a friendship may fall short of this ideal. Imagine, for instance, that one friend considers herself entitled to special treatment that her friend has no claim to (the friend owes it to her to be attentive, or grateful for her friendship, but she has no reciprocal duty to him), or asserts power over her friend that her friend lacks or that she denies to him (as when she insists that she gets to decide where they go on holiday together if she pays, or that she should pick their destination because she has better taste). Such a friendship, in which one friend efectively deems herself the other’s superior (or inferior), would intuitively be defcient because it falls short of an ideal of how friends should relate to each other—specifcally, as equals. I think that relational egalitarian arguments for political equality must pay attention to diferences between these two examples, and the associated intuitions underpinning Relational Equality, because they have quite diferent implications for Equal Power and Political Relationships. In a nutshell: If we start from the anti-caste intuition to defend relational egalitarianism, we have an easy time explaining why our fndings apply to political relations in society at large. After all, caste is an essentially societal phenomenon. But we have a hard time explaining why relational equality requires equal power: unequal distributions of political power need not amount to objectionable social hierarchy of the sort we associate with caste or class structures. On the other hand, if we start from the example of friendship, we have a relatively easy time explaining the need for equal power. But we have a hard time establishing that the relevant norms apply to political society. Let me conclude this section by contrasting the relational egalitarian arguments for political equality that are the focus of this chapter with other arguments with which they may easily be confused. On the relational egalitarian arguments I discuss, equal power is itself an essential component of a non-derivatively valuable relationship. By contrast, there are other arguments that also appeal to the non-derivative value of certain relationships (including, perhaps, relationships we tend to associate with equality), but grant at best indirect signifcance to equal power. Tus one might, with Rousseau’s Second Discourse, greatly care about the relational (dis)value of dependence, and favor political equality because it inhibits dependence relations.18 Or one might, in line with neo-republican views, take nondomination to be the central value governing relationships among co-citizens, and argue that democracy contributes to its realization.19 It would be unsurprising if someone attracted to the ideal of relational equality also felt the pull of some of these other relational ideals. Indeed it is natural to think 18  Cf. (Neuhouser 2014).

19  (Pettit 2012).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Power and Equality

11

that an ideal egalitarian relationship will instantiate not only the ideal of equality, but also other ideals of roughly the sort just gestured at. Yet the support for political equality that these other relational ideals provide is structurally sufciently diferent, and subject to sufciently distinct worries and objections, that this chapter will limit itself to discussing the more direct arguments for political equality that ft the schema outlined at the beginning of this section.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

3. Tis section discusses the anti-caste paradigm of relational equality. Behind this conception of relational equality lies the following thought: Caste societies, in which some people are socially “above” and others “below,” are intuitively morally problematic. Tere is something objectionable about a society that distinguishes between peasants and lords, plebeians and patricians, untouchables and Brahmins. And, relational egalitarians propose more specifcally, what is objectionable about such arrangements are not merely their instrumental consequences, or the fact that those deemed “below” are treated in ways that are anyway problematic quite apart from the fact that others are “above,” or even that those who are below act in obsequious ways we fnd demeaning. Instead the social hierarchy is inherently problematic. Someone can say: “Te social arrangements under which we live treat me as another’s social inferior, and him as my superior,” and that is meant to be an objection in its own right to these arrangements. Finally, for those who appeal to this conception of relational equality to defend political equality, inequality in power is (unless qualifed in certain quite specifc ways) itself constitutive of social hierarchy, rather than being merely a causal antecedent of certain hierarchical social relations. To assess the plausibility of this position, this section discusses what precisely social hierarchy of the sort we associate with caste or class amounts to, and why such “social status hierarchy” (as I will call it) may be deemed distinctly problematic. Section 4 considers whether the absence of social status hierarchy requires an equal distribution of political power. To determine what is morally problematic about social status hierarchies, we need to understand what they are. Tis is not, in the frst instance, a moral inquiry but a conceptual one: an attempt to identify, and properly characterize, core features of a particular social phenomenon. Still, part of what seems to unify diferent instances of the phenomenon is that we view them as morally problematic; and we would expect this to matter for our analysis of the phenomenon’s central features. I treat as paradigmatic

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

12

Daniel Viehof

instances of the phenomenon the kinds of caste or class20 societies mentioned earlier: societies in which some are peasants and others lords, some untouchables and others Brahmins, some plebeians and others patricians. I focus on three characteristics of such societies: they involve status inequality; the inequality is not a matter of mere diference, but instead establishes a hierarchy; and the hierarchy structures society as a whole. Clarifying these characteristics should in turn help us identify what is distinctly morally problematic about paradigmatic instances of social status inequality.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

i. Society as a Whole Let me consider the last point frst. Te existence of a caste structure (like the existence of a class hierarchy, a patriarchal structure, etc.) is a feature of a society as a whole, rather than of a particular relationship. When we think, for instance, of the sense in which the servant is “below” the lord of the manor, we do not just mean that, within their particular relationship, the servant is subordinate. We also mean that their positions as master and servant generalize, and shape all other social relationships that they have. Te servant, we may say, it not just his master’s servant. Even if he currently has no master, he remains a servant, and others will relate to him as such. Similarly, the master is not just his servant’s master. He will be a master even if he currently has no servants, and others will relate to him in what they think is a manner appropriate to his status. A social hierarchy is properly attributable to society as a whole if it structures relationships among members of the society in general. Te relevant notion of generality bears on both the content of social norms and the norms’ existence conditions. First, if you know that I am an untouchable in a caste society, you know not only how you should relate to me (in this regard), you also know the relation in which I stand to all other members of society, since that relation is itself determined by caste. Social status is, in Hohfeldian language, a “multital” relation (like property), not a “paucital” relation (like contract).21 (And like property, the social status associated with caste or class is insulated from certain forms of detailed attention to individual peculiarities. I will return to this point below.) 20  So class, as it fgures here, is centrally about social status. Tere are infuential alternative notions of class, indebted to Marx or Weber, which focus instead on a person’s relation to the means of production, or capacity to generate income in the market. Class understood in these latter ways is evidently important in its own right. But the moral questions it raises are (at least in the frst instance) distinct from relational-egalitarian concerns about inequality. For discussion, see (Turner 1988). 21  (Hohfeld 2001).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

13

Second, for our society to be structured by a particular hierarchy, the norms governing relations among people with diferent status must have social reality: they must be “systematically sustained by laws, norms, or habits” that are sufciently widespread to properly count as representative of society as a whole.22 We may call these “societal norms” for short. A full-blown account of social status hierarchy (which is beyond the scope of this chapter) would need to explain under what conditions norms are properly attributed to society as a whole, rather than refecting the view of just a single person or a small sub-group. It would, in particular, have to explain how disagreement among members of a society about which norms properly govern it will afect the existence of societal norms, norms representative of society as a whole. Often the legal system will function as a mouthpiece for society’s view of norms. But not all social norms will be embodied in legal norms. And sometimes legal norms are in fact in tension with social norms; and it cannot be taken for granted that in such cases, the former prevail. (Tink of the long struggle about caste in India after the ofcial legal rejection of caste structures.) Let me add three clarifcatory observations. First, we need not assume that a society is governed by a single social status hierarchy. Instead societies are usually structured by various intersecting social status hierarchies: gender, race, class, and so on. To say that a social status relation governs society as a whole is thus not to say that it governs it exclusively. Second, the features just highlighted are not unique to status hierarchies, but apply more generally to social diferentiation that is attributable to society as a whole. Tus in a society that distinguishes between the status of child and the status of adult yet does not treat one as superior to the other, the fact that I am an adult structures all of my relations to everyone else qua child or fellow adult, and the norms involved are sustained by society. (Te distinction between status diferentiation and status hierarchy is discussed further below.) Tird, a society in the relevant sense is not limited to a group the size of a modern political community. For instance, a high school may be a “society” in the relevant sense, governed by internal norms that structure relations among all students and are sustained by the students’ attitudes and actions.23 (Tis matters mostly because it expands the range of examples with which we can work to get a grip on the phenomenon in question.) 22  (Anderson 2012), p. 42. 23  Perhaps a friendship too counts as a “society” so understood, and the demands of social status equality also apply to it qua small-scale society. Tis would, I think, be a feature rather than a bug. More importantly, it would not prevent us from also insisting that additional norms apply among friends qua friends (rather than qua fellow members of a small-scale society).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

14

Daniel Viehof

Tat caste or class is a feature of society as a whole in turn explains why not all instances of inequality amount to status hierarchy of the sort we associate with these phenomena. For example, that some people think of themselves as superior to others (and perhaps even that those particular others happen to think of themselves as inferior) is compatible with the absence of castes and classes if the claim to superiority is not sustained by societal norms. And even if it is recognized that one person has a special claim on another, and that claim is supported by societal norms, the asymmetry in claims need not amount to a hierarchy that mars society as a whole if the socially recognized relation is limited to the two parties, and does not structure their relations to many other people.24

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

ii. Status But even inequalities that are socially recognized, and structure relations among all members of society, need not create social hierarchies of the sort we associate with caste or class. To see this, consider the somewhat mundane, but also relatively tractable, example of a high school. Te school could be structured by caste hierarchies: the jocks reign supreme, the geeks are somewhere near the bottom, and so on. But it need not be. And it need not be even if there are socially recognized inequalities that structure relations among all students. Imagine, for instance, that each term the school publishes a complete ranking of all students’ academic performance. So everyone knows where they are vis-à-vis anyone else when it comes to academic standing. And imagine too that there is a social norm in the school that students are expected to care about, and admire, academic success, and express that admiration toward those who do well. Te social life of this high school, though it sustains inequality, need nonetheless not instantiate status hierarchies. Just imagine the relation between two students, one ranked close to the top of the class, 24  Consider peonage. Tere is evidently something intrinsically bad about it: the person who is indebted must work for the other, without (at that moment) adequate compensation, and without signifcant control about whether to do such work. Tat alone likely sufces to make peonage objectionable, and deserving of abolition. It may also follow that the relation between debtor and creditor is one that is importantly unequal, unequal in a way that undermines certain relations between them. (Friends, for instance, would have to forgive another’s debt for the friendship to be sustainable.) But as long as what has changed is only the debtor’s relation to the creditor, and not the debtor’s relation to others in society, peonage does not introduce the kind of status hierarchy with which we are currently concerned. Te fact that historically, peonage was associated with social hierarchy refects in part the fact that peonage existed in societies where those working for others in various positions were generally deemed to be of lower status. It is this further association that explains why peonage creates a distinctive problem of social hierarchy, of the sort we associate with caste or class.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

15

the other close to the bottom. Tat one has performed better academically, and is thus worthy of admiration, and that such admiration ought to be expressed—the more successful student ought to be congratulated, say— does not, I think, justify the judgment that the higher-ranked student has superior social status in the school.25 What distinguishes positive judgments, or even rankings, in general, and judgments of social hierarchy of the sort associated with superior or inferior status in particular? It is a central feature of status that it attributes to us a range of rights and duties that are one step removed from the characteristics on which the attribution of that status seems to rest. Tink of the legal status of “minor”: It attributes to someone a whole range of legal incidents that are at least partly mediated by the very idea of “minor,” rather than directly justifable by appeal to the characteristic that make us one (viz., being below the age of majority). And this is not a feature of legal status alone. Sociologists concerned with social status also emphasize in their studies “the prestige accorded to individuals because of the abstract positions they occupy rather than because of immediately observable behavior.”26 Even moral status may plausibly be thought to have this character.27 Generalizing from these observations, I propose that status involves a gap between what triggers the attribution of a particular status to someone (their quality) and what response to the bearers of superior status is thought to be appropriate given that status (their claim). Status, in other words, is a non-eliminable intermediate step in the justifcation of its bearer’s claim, a step that makes the claim about something other than simply the underlying quality (age, behavior, performance).28 Tis explains why we need not think of the high school as instantiating status inequality: while social norms require responding in certain ways to other students’ academic performance, 25  Tis is not to say that the judgment that is being made is normally inert or irrelevant. A lower-ranked student may envy the higher-ranked student, or resent her for her success, and yet not take the other to be her social superior. 26  (Gould 2002), p. 1147. See also, e.g., (Chan and Goldthorpe 2004), p. 383: Status order is “a set of hierarchical relations that express perceived and typically accepted social superiority, equality or inferiority of a quite generalised kind, attaching not to qualities of particular individuals but rather to social positions . . . or to certain . . . ascribed attitudes.” Note that some sociologists discussing status are ultimately interested in the microprocesses that determine how individuals evaluate others, and how various evaluations interact in establishing mutual (but not necessarily societal) rankings. See, e.g., (Jasso 2001). See (Turner 1988) for a general treatment of status in sociology and social theory. 27  See, e.g., the discussion of “range properties” central to moral status in (Waldron 2002), and of “evaluative abstinence” and “opacity respect” in (Carter 2011). 28  Cf. Kolodny’s discussion of “consideration,” or “those responses that social superiors, as social superiors, characteristically attract.” (Kolodny 2014b), p. 297. As Kolodny explains, “although their basis may be some narrow and accidental attribute of the person, the responses constitutive of consideration are focused on the person and his or her interests, claims, or imperatives as a whole.” (Kolodny 2014b), p. 298.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

16

Daniel Viehof

the link between that performance and the appropriate response is sufciently close that we don’t think of it as involving a more general judgment about the person that exceeds the specifc quality at issue. (Matters would have been diferent if, for instance, the higher-ranked students had been entitled not to receive warm words, but to be obeyed, or to have their belongings carried around by their fellow students.)

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

iii. Hierarchy Tat status comes with a whole bundle of rights and duties in turn explains why it is worth distinguishing clearly between status diferences and status hierarchy. Adults and children do not have the same legal status. Nor do married people and single people. And yet we would not ordinarily think that with regard to these examples, one group’s legal status is superior to the other’s. Teir status diferences—the diferent rights and duties they have qua minors or adults, or qua married or single people—do not involve claims that we associate with one party’s superiority over the other. Tere is a status diference here, but no status hierarchy. Or, to use terminology sometimes adopted by sociologists, there is “diferentiation” but no “stratifcation” of status. And it is status hierarchy or stratifcation—or, as I will usually continue to call it, “status inequality”—that really concerns us. How do we distinguish between social status inequality and a mere diference in social status? It is tempting to adopt what I will call the simple approach here: A is B’s social superior, and their relation is thus one of status inequality, if the relevant societal norms specifcally assign A greater benefts than they assign to B, or grant her greater rights, or give her greater power. Let me say, for short, that the norms assign “advantages” to A over B.29 On this view, I can identify someone as my social superior by identifying how our society’s norms distribute advantages between us. But the simple approach, though tempting, is ultimately inadequate. For part of our aspiration in developing an account of social status hierarchy is to make sense of the complaint someone has when he says, “Te social arrangements under which we live treat me as another’s social inferior, and him as my superior,” where this is an objection in its own right to these arrangements. Te sense in which society treats another as my superior (or inferior) must, in other words, be inherently morally problematic. And 29  How do “advantages” relate to Kolodny’s “consideration”? If “consideration” is meant to pick out responses to superiors that are not inherently problematic, then “consideration” and “advantage” may come to the same thing; but then Kolodny still needs to explain which form of consideration is morally objectionable. If “consideration” is meant to pick out responses that are inherently morally problematic, then some of the phenomena that Kolodny is interested in do not amount to “consideration.”

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Power and Equality

17

yet the features highlighted up to now—that society as a whole assigns certain unequal advantages to A over B, in a way that seems justifcatorily detached from underlying considerations—are not, jointly or alone, inherently problematic. Consider the following example: Medical Services:  A society grants certain people (medical doctors on duty) a right to park their car in spots where others are not permitted to park. It also gives them fashlights that they can attach to their cars, and when they turn them on, others are expected to scramble out of the way and let the doctor pass. In some ways—and, crucially, with regard to those features our analysis of social status inequality has focused on up to now—this case is difcult to distinguish from another.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Lord’s Carriage:  A society grants certain people (Lords) the right to park their carriage in places where others are not permitted to park. It also gives Lords certain insignia, and if those are attached to the Lord’s carriage, others (commoners) are expected to scramble out of the way and let the Lord’s carriage pass. On the simple view, the special advantages that doctors have in Medical Services would establish them as our social superiors, just like the Lord in Lord’s Carriage. But intuitively it is quite clear that, though the doctor’s advantages could be that, they need not amount to superior social status (which, remember, is meant to be inherently objectionable). For these advantages, despite their unequal distribution, can also intuitively be compatible with mere social diferentiation. Whether the doctors’ advantages amount to diferentiation or hierarchy depends, I propose, on how they are justifed. And since what matters are the norms attributable to society, it depends, more specifcally, on how society takes the advantages to be justifed. (For the sake of simplicity, I will generally continue to speak of justifcation simpliciter. It is worth keeping in mind that the issue is the justifcation as viewed by society, or social justifcation.) If the societal norms granting doctors such advantages are justifed by appeal to the interests of everyone around here, where all of these interests are treated as equally signifcant, then possession of these advantages does not translate into social superiority. I would not, in that case, look at a doctor who races past me in her car with her fashlight on and think “Society treats her as my social superior,” the way that a peasant may have looked at the lord of the manor as he passes by in his carriage. And when I see the doctor later at a bar, I wouldn’t normally fear that she would take herself to be my social superior and decline to talk to me. (She might still do so. But if

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

18

Daniel Viehof

she does, this refects her personal views rather than society’s judgment of our respective status.) If, by contrast, doctors are given such advantages, not because society believes them to be suitably instrumentally justifed in light of everyone’s equally relevant interests, but because doctors are deemed to have more important interests or claims—to be ultimately more important than we are—then their advantages do amount to social hierarchy rather than mere diferentiation. (Similarly, if society takes these advantages to be justifed instrumentally, but the instrumental justifcation itself rests on assumptions about the diferential moral importance of diferent persons, then the advantages mark, though they may not constitute, social status hierarchy.) With this conceptual analysis of social status hierarchy in place, we can turn to the normative question why such hierarchy is inherently morally problematic. Te distinction just drawn, between status hierarchy and status diferentiation, suggests an initial answer: If we are all moral equals, matter equally, etc., then social status hierarchy is objectionable because it treats us as if we were not. Te distribution of advantages associated with social status hierarchy lacks adequate social justifcation. Te emphasis on social in the previous sentence is important if the analysis of social status hierarchy is to capture the distinctiveness of the relational egalitarian complaint. After all, if the issue were simply that some people are given objectively unjustifed advantages to the detriment of others, then this complaint could easily be accommodated by conceptions of equality in contradistinction to which relational egalitarian positions have usually been developed.30 What makes the complaint at issue here distinctive is its concern with social status hierarchy as a social fact: at issue is not simply whether an unequal distribution is objectively justifed, but whether it can be justifed from within the normative commitments of society at large without presupposing that some people (some people’s interests or claims) are of greater ultimate moral signifcance than others (their interests or claims).31 Te attribution of social status hierarchy to a society is thus an interpretive exercise that requires judgments about the normative basis on which society endorses particular social norms, most obviously norms that distribute unequally certain advantages. Where, on the best interpretation 30  See, e.g., (Anderson 1999, 2012). 31  It is compatible with this account that social status inequality exists even though an objective egalitarian justifcation for the distribution of advantages is in principle available, if that justifcation is not recognized, or indeed recognizable, by the citizens. So a concern with social status inequality, as a phenomenon that depends on people’s views of how inequalities are justifed, may provide support for theories that care about whether justifcations of social or political arrangements are accessible to, or endorsed by, those they govern.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Power and Equality

19

available to those living under these norms, society’s endorsement of these norms cannot rest on normative and factual premises that treat everyone’s interests or claims as of fundamental equal importance, these norms embody society’s implicit (and sometimes explicit) judgment that some people matter more than others. Social status hierarchies, we may say, embody society’s judgment that some people are fundamentally more important than others; and they exist—as a social fact—where those living in a society cannot reasonably see how the unequal distribution of advantages could be given a social justifcation compatible with everyone’s equal fundamental moral signifcance.32 Tis may have various detrimental efects on our capacity to engage in egalitarian relationships across class- or caste-lines, or on our selfrespect. But it is, crucially, also inherently objectionable: it is a morally deplorable feature of a society that its norms embody mistaken judgments of fundamental inequality even if this has no further efect on people’s attitudes and relationships.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

4. Section 3 ofered a reconstruction of social status hierarchy and its moral signifcance. In this section I want to explore what social status hierarchy, so understood, entails for our assessment of political equality. Specifcally, I argue that, once social status hierarchy is properly understood, it becomes difcult to defend the ideal of political equality by appeal to the anti-caste intuition. If the previous account of social status hierarchy is correct, there need be no complaint based on status hierarchies just because some people have certain advantages or superior entitlements, including greater power. For as long as society justifes these inequalities in a way that does not treat one person (or her fundamental interests and claims) as more important than another, the inequalities are compatible with our status as social equals. And because 32  Tere is thus an expressive dimension to social inequality, if by this we mean that such inequality matters centrally because it is reasonably taken to refect a certain view of people’s fundamental moral signifcance. Te expressive dimension in turn afects— constitutively—the possibility of certain kinds of relationships, relationships in which people see each other as equals. For views that emphasize the expressive dimension of status inequality, see (Fourie 2012) and (Scanlon 2003). But unlike Scanlon (and perhaps Fourie), I think that what is required for problematic status inequalities is neither that certain inequalities “could only be understood as intended to express the view that they were inferior” ((Scanlon  2003), p. 213, my emphasis) nor that certain inequalities, though lacking “the aim of expressing inferiority, nonetheless had the efect of giving rise to feelings of inferiority on the part of most reasonable citizens” (p. 213, second emphasis added). It sufces that the inequality, though not intended to express any view, in fact is reasonably taken to express such a view; and when it does, this constitutively undermines certain valuable relationships, even if no one in fact feels inferior as a result.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

20

Daniel Viehof

the distribution of political power has signifcant instrumental efects on many people other than the power-holder, it is often possible to explain, quite straightforwardly, the beneft of an unequal distribution of power without appealing to the superior importance of one person’s interests or claims. To make this suggestion both more concrete and more plausible, consider an example of unequal political power that, it seems to me, fts this description.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Necessary Representation:  An egalitarian tribe, one in which all adults are generally assumed to have equal status, comes into confict with another tribe about shared hunting grounds. Some agreement needs to be negotiated. Te difculty is, however, that both tribes are nomadic, which makes it difcult to send an emissary. So when one person (call her R) happens to come upon a member of the other tribe, they take the opportunity to negotiate a wide-ranging set of rules for hunting that will minimize future confict. Ten each of them returns home to their own tribe and presents them with the agreement reached. R delivers the rules to her tribe, and the tribe expects all of its members to abide by them—not because R was authorized in advance to make the decision, or because a majority of the tribe’s members agree with the rules, but because, given the importance of having rules that coordinate interactions with the other tribe, and the difculty of negotiating with the other tribe, following R’s rules is the best way to solve the urgent moral problem posed by the inter-tribal confict. I think there is no doubt that R has greater political power here than any other member of the tribe: she decided what rules would bind all of them with regard to hunting in a certain area. And yet I also think that she need not therefore be deemed their social superior. In other words: whatever complaint R’s fellow citizens may have about this arrangement empowering R (and I do not deny that they could have justifed complaints), their complaint cannot reasonably be that, if R is so empowered, then R is granted superior social status. Tis remains true, it is worth adding, even if the agreement will be binding for many years into the future, and so R’s decision will afect how the tribe will live for a long time to come (because, say, a suitable encounter with a member of the other tribe is sufciently uncommon). And I think it is also true even if, as the example assumes, the negotiations cover a wide range of issues, touching on many features of tribal life. Tis example provides intuitive support for the claim that not all inequalities in power amount to social status hierarchy. Furthermore, it fts with the explanation I ofered for why inequalities in advantages (including inequalities in power) need not undermine equality of social status. Whether they do depends precisely on why society grants someone special advantages, including greater power. And given what I said in setting up the example, we here have an explanation of R’s superior political power

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

21

that does not depend on any assumption that she is, or is thought by the other members of the tribe to be, their social superior—as someone who somehow matters more than they do. Let me emphasize here the particular dialectic of the argument: my claim is not that there is nothing problematic about R’s unequal power. My claim is, rather, that whatever we think is problematic about it (if anything), it cannot be that R, by dint of her greater power, has become her fellow tribemembers’ social superior, since that claim is false. So if we are not willing to give up the thought that R’s greater power is objectionable, or at least regrettable, then we need to look elsewhere for a justifcation of that judgment. In light of these observations, let me discuss in more detail Niko Kolodny’s defense of political equality based on relational egalitarian concerns.33 Kolodny’s paradigmatic examples of relational inequality include, as I  mentioned earlier, servant/lord of the manor, slave/master, plebeian/ patrician, and untouchable/Brahmin. In other words, he is centrally (though perhaps not exclusively) concerned with what I have called social status inequality.34 Kolodny also argues that such inequality is instantiated, in a fairly obvious way, where society gives some people greater power or de facto authority than others. Even if the society otherwise shows equal concern for people’s interests, and for their claims to means that enable the pursuit of their personal life plans, it is nonetheless a presumptively unequal society if (i) some have “greater relative power (whether formal or legal, or otherwise) over others, while not being resolutely disposed to refrain from exercising that greater power as something to which those others are entitled”; or (ii) some have “greater relative de facto authority (whether formal or legal, or otherwise) over others, in the sense that their commands or requests are generally, if not exceptionlessly, complied with (though not necessarily for any moral reasons)” and they lack (once again) the right disposition to refrain 33  Elizabeth Anderson, the other prominent relational egalitarian proponent of democracy, is not open to the worry I raise here, at least on one reading of her argument. On that reading, Anderson’s relational argument for democracy is quite indirect: Democracy is not required by relational equality as such. Rather, relational equality requires that public ofcials act for public ends, public ends are determined by the public interest, and people should be given a democratic say in determining what the public interest requires if we are to make sure that everyone’s interests are to count equally. See (Anderson 2010), p. 107. (On another interpretation, Anderson relies on a story about delegation not dissimilar to Kolodny’s, and assumes that the public ends whose pursuit is compatible with relational equality must—as a conceptual matter—be set by the people themselves. Tis account would be subject to worries similar to the ones discussed here.) 34  Niko Kolodny has suggested to me that he may have had in mind something closer to a view on which relations of inequality—including asymmetries of power and authority—are problematic in general, independently of specifc relationships and their value. But for reasons I briefy discuss in Section 7, I think this is a rather less plausible position than the one I discuss here.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

22

Daniel Viehof

from exercising that authority.35 So for Kolodny, inequality of power or de facto authority as such poses a (presumptive) problem for our social status equality. Kolodny is clear that not all social relations that assign diferential power to people give rise to worries about status inequality. He recognizes, for instance, that many private associations—churches, employment relations, families—involve unequal power and may yet avoid worries about relational inequality. But he thinks he has a straightforward explanation for the special objection we have to an unequal distribution of political power: Private relations usually include exit options, or other opportunities to avoid standing in the unequal power relation.36 As a result, it is within our power to determine whether others have unequal power over us; and that itself reduces the impact that the inequality has on our relationships. Political power, by contrast, usually arises in relations that lack signifcant exit options or other opportunities to avoid being under another’s power. I agree that the presence of exit options does indeed explain why we are often (though not always) much less concerned with inequalities of power within certain private relations. But though correct, the appeal to exit options is insufcient to deal with Necessary Representation: just as in any other political community, membership in the tribe is not easily given up, and so the unequal power of the tribal emissary R cannot be compensated for by other members’ opportunity to avoid being bound by the outcome of her negotiations. Te distinction between private and political decisions is not the only resource Kolodny deploys to explain why sometimes we are relatively unconcerned with unequal power. He also suggests that unequal power is unproblematic where the person who has greater power is merely the agent of those over whom the power is exercised.37 Tis explains, Kolodny suggests, why political representation need not pose a threat to our equal social status: our representatives (and, by extension, someone like R in my example) may have more power or de facto authority than we do. But they are nonetheless not our superiors because they have and exercise this power qua agents of the people, who have merely delegated decision-making power to the ofce holder. 35  (Kolodny 2014b), p. 295. Kolodny also mentions, as a third possibility, someone’s having “attributes (for example, race, lineage, wealth, perceived divine favor) that generally attract greater consideration than the corresponding attributes of others” (p. 296). I think consideration is indeed more closely tied to issues of caste inequality. But as we just saw, unequal consideration amounts to caste inequality only if it lacks a suitable social justifcation. In light of this, and because Kolodny himself is content to forgo appeals to consideration, and reach democratic conclusions via appeals to the signifcance of unequal power or de facto authority (p. 298), I set aside this third possibility. 36  (Kolodny 2014), p. 304. 37  (Kolodny 2014b), p. 317.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

23

I agree that there need be no problem of social status inequality between citizens and their representatives. Speaking purely anecdotally, many years ago I used to know my local MP reasonably well. And though I knew that he had power that I lacked, I never thought that he was my social superior.38 But is this best explained by the fact that the MP was my (or, rather, my community’s) agent? It depends on what the agency relation amounts to. On one understanding of what delegation amounts to, it may require that the principal has substantive control over the agent. I doubt, however, that this is strictly necessary to explain why my MP’s greater power did not make him my social superior. For it is highly doubtful that citizens do have substantive control over their MPs. Clearly individual citizens lack such control. And I in fact doubt that even the community as a whole possesses it. (Just consider the signifcant divergence between a representative’s voting patterns and her constituents’ preferences that is common in many democracies.) One response to this would simply be that our democracies fall short of the ideal of delegation, and thus also of realizing social equality between MPs and ordinary citizens. But I don’t think this is adequate. For I accept (non-idiosyncratically, I believe) both that my local MP was subject to neither my control nor the control of the community, and that he was nonetheless not my social superior. On another view, the central feature of delegated power is precisely that it is not justifed by, and exercised for the sake of, the interests of the powerholding agent, but by, and for, the interests of the principal. Power may thus count as delegated even if the principal has no control over the agent. (In practice it may, however, often make sense to introduce such control precisely to ensure that the agent acts for the principal’s beneft.) Tis, I think, ofers a more plausible account of why my MP is not my social superior. But it also entails that what creates conditions of social inferiority and superiority is not possession of unequal power or de facto authority as such. Rather, whether unequal power constitutes relations of social inferiority or superiority depends on what justifes this inequality. I thus suggest that relational egalitarian arguments that start from a concern with caste or class hierarchies do not provide reasons for valuing political equality as such. Tis follows from the fact that the distribution of power or de facto authority as such is not an independent constituent of unequal status relations of the sort we associate with caste and class. Instead political equality is, on the anti-caste view, nothing more than a baseline: If society’s 38  I leave it again open whether there may be other complaints about our MPs’ greater power. My sole point here is that, whatever complaints we have about representative institutions, we cannot plausibly complain that the correlative inequalities in power constitute relations of social status hierarchy.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

24

Daniel Viehof

justifcation for distributing political power unequally does not rest on an appeal to the equal interests or claims of the community’s members, then the unequal distribution gives rise to complaints about social status inequality. But if society’s justifcation for the inegalitarian distribution of power takes every member to be of equal signifcance, then political inequality is not even regrettable insofar as our concern is solely with social status hierarchy. Let me conclude this section with two brief observations about the implications of this argument for Democracy’s Value (the claim that democratic institutions have special value) and Democracy’s Authority (the claim that they have special authority) respectively. When it comes to Democracy’s Value, the fact that the anti-caste argument only establishes equal power as a baseline (rather than vindicate it as an ideal) may, on some views, be of greater theoretical than practical signifcance. For whether an inegalitarian distribution of power is compatible with social status equality depends on whether an adequate social justifcation of such inequality is available. And whether it is available depends on the conditions under which particular justifcations can be attributed to society as a whole, a matter about which I have said very little. Tus someone may respond to the argument ofered here by suggesting that a justifcation can be attributed to society only if there is a high degree of consensus among citizens (or reasonable citizens, or . . .) regarding its content; and that there is no such consensus when it comes to the purported egalitarian benefts of an inegalitarian distribution of power.39 But then it might turn out that the anti-caste argument is sufcient, in practice, to establish democracy’s distinctive egalitarian value. I in fact believe that the conditions under which we can plausibly attribute a particular justifcation to society are (in some ways) less demanding, so that it is rather easier for a society to satisfy the requirements of social status equality while distributing political power unequally. But since I cannot solve this matter here, I simply fag its importance. And in any case, even on the most generous interpretation of the conditions under which we attribute a justifcation to society, the anti-caste argument cannot establish Democracy’s Authority. Kolodny suggests that, “If I were to disregard the democratic decision, then I would be depriving others of equal opportunity to infuence this very decision. For infuence over the decision, in the sense relevant in this context, is not simply infuence over what gets engraved on tablets or printed in registers; it is infuence over what is actually done. Insofar as relations of social equality are partly constituted by precisely that equal opportunity for infuence, I would be, by depriving

39  Tis suggestion is evidently modelled on Estlund’s argument briefy mentioned in Section 2.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Power and Equality

25

others of that infuence, relating to them as a social superior, at least in that instance.”40 But if equal power (or “equal opportunity for infuence”) is not a constituent component of equal social status, then this argument runs into trouble.41 Unequal power may be compatible with social equality if it is suitably justifed. If one thinks that acting contrary to unjust legal demands is an adequate justifcation for claiming special power, then justifed resistance or disobedience need not give rise to a complaint about social status inequality. If I thought I could disobey because I was special, superior to my fellow citizens, then there would indeed be a problem. But if I thought instead that anyone who found himself in my situation—anyone confronted with this unjust law, and able to disobey—would have reason, and permission, to act as I do, then I would not be taking myself to be anyone’s social superior, and my disobedience would not have to be incompatible with our equal social status.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

5. Let me turn next to the friendship conception of relational equality, which takes as its starting point paradigmatically egalitarian relationships like friendship or marriage. Tough perhaps not wholly independent of the anti-caste version of relational equality, it is clearly not neatly aligned with it. Tere is no neat alignment, because in a society that is deeply structured by social hierarchies like caste, people are capable of having egalitarian friendships or marriages, if not across caste- or class-lines, then at least with people who share their status. (Two servants can realize an ideal of egalitarian friendship among themselves even though they are both “below” their master.) But neither is there complete independence, because in a society governed by class hierarchies it is difcult for a master and a servant to have a friendship among equals. Even if they both try as hard as they can to ignore the inegalitarian norms, the fact that these norms are socially enforced will make it difcult to escape the societally imposed inequalities, and avoid having them foisted upon their own interpersonal relationship.42 Te following discussion of the friendship conception of relational equality will focus on two points in particular. First, equal power is, I think, a constituent 40  (Kolodny 2014b), p. 315. 41  I set aside here the further problem that my disobedience need not be authorized by any norm attributable to society, which would seem a precondition for social status hierarchy. 42  Indeed, some sociologists use density of friendship relations as an indicator of class structure (understood in the sense discussed in Section 3): (Chan and Goldthorpe 2004).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

26

Daniel Viehof

component of egalitarian friendship. Tus an appeal to this conception of relational equality will avoid many of the problems we encountered in the previous sections. But, second, we must ask whether the ideal of friendship can plausibly be thought to govern our political relations—or, rather, which features of friendship are essential for triggering the demand for equal power, and whether these features plausibly have a counterpart in the political domain. Consider a friendship, marriage, or similar relationship. I assume that participants in such relationships have special concern for one another, and thus take the other person’s interests to make demands on them that are greater than those made by the interests of outsiders. But special concern is not enough for friendship. Tere must also be a commitment to equal concern. Friends take the demands made by their friends’ interests to be symmetrical to those that their own interests make on their friends.43 More specifcally, they each accept in principle that “the other person’s equally important interests . . . should play an equally signifcant role in infuencing decisions made within the context of the relationship” and they each have “a normally efective disposition to treat the other’s interests accordingly” in their deliberation, “constraining [their] decisions and infuencing what [they] will do.”44 But even special concern and equal concern together do not exhaust our ideal of friendship. Tere is also a requirement of equal power over the relationship. And this requirement is not a mere by-product, but a constituent component of our egalitarian ideal of friendship. Friends should have equal power—understood as equal opportunity for infuence—over the character of their relationship and the norms governing it; and failure to distribute power over the relationship equally means that the relationship falls short of its egalitarian ideal. Consider an example: Imagine spouses who each accept that the other’s interests are as important as their own in determining how they should relate to each other, and who each have the disposition to act accordingly. Nonetheless they may end up disagreeing about the character and norms of their relationship, or how they should interact or act together. Tey may disagree because equal concern underdetermines what they should do; or because they difer as to what equal concern exactly requires, whether because they diverge on what interests properly count as part of their marriage, or because they disagree about how weighty various interests are. To make the 43  I don’t want to exclude the possibility that there may be other relationships that give rise to special obligations and yet lack that symmetrical character. But these would not be relationships of friendship, and would lack the distinctive value that friendships have. 44  (Schefer 2015), p. 25.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

27

example more concrete: Spouses in an egalitarian marriage may disagree about whether they owe it to their neighbor to invite her to a party they are holding, even though they both would be happier if the neighbor didn’t come. (So their interests are aligned, but their judgments about what to do in light of these interests are not.) If one of them unilaterally goes ahead and invites the neighbor even though he knows that his spouse thinks they ought not to, then this is, I think, a presumptive problem for their relationship. As a one-of event, it may be relatively minor: what ultimately matters is equal power over the relationship as a whole, rather than any one-of decision.45 So if there will be future opportunities for the other spouse to decide how they proceed in the face of disagreement, the current decision to issue an invitation unilaterally may not seem especially problematic. But if something like this happens frequently, and isn’t balanced across the parties to the relationship, then it would, it seems to me, threaten their egalitarian relationship, simply because the person extending the invitation exercises (and, in recognizing that he does, implicitly asserts a right to) unequal power over the relationship. (Similarly, if the decision at issue, though one-of, is sufciently important to seriously change the shape of the relationship, and if the other party foreseeably won’t have an opportunity to equally shape the relationship in the future, then there is a problem.) Someone might accept the example but insist that it does not show that equal power plays the particular role in our ideal of friendship that I have proposed. I will consider two versions of this response. First, someone might argue that friendship requires consensus among the friends about the character of their relationship and the norms governing it.46 Te problem with the example mentioned is not that one partner exercises unequal power by issuing the invitation; it is, rather, that the spouses do not agree on how to proceed as a couple. (So in a sense, equal power over the relationship is important. But this is only because friends must agree, and so may be thought to have—equal—veto power over actions undertaken qua friends. What ultimately matters is that the parties reach a consensus, not that they have equal power.) 45  Not all forms of power are equally problematic from the perspective of friendship. Persuading me of the wisdom of a course of action, though it involves a form of power, does not pose the same problem as authoritative directives, threats, or even ofers. Why? For friends to relate to each other as friends, they must see each other and themselves as possessing certain agential capacities. Among these is the capacity to appropriately respond to reasons central to the friendship, including reasons about how best to understand its character and norms. So a friend (qua friend) should properly treat her rational convictions regarding the proper character of the friendship as her own, rather than attribute them to another, even if that other played a role in bringing the conviction about by rational argument. 46  See, e.g., (Mansbridge 1980), pp. 9–10.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

28

Daniel Viehof

But this seems to me to be a mistaken view of friendship. As friends, we do have reason to coordinate, or settle, on a common understanding of the norms governing our relationship. Tis partly refects the instrumental value of shared norms, and partly the importance of reciprocity (and perhaps especially reciprocity visible to the parties) in relationships among friends. But such coordination can be achieved not just by consensus, but also by taking turns in deciding contested decisions, deferring to an impartial third party, or adopting some other egalitarian decision procedure that we accept as binding. If my spouse and I disagree about certain important matters— how to treat our neighbor, raise our kids, etc.—then this might put a strain on us because we might fnd ourselves torn between the demands of the relationship and the duties we have to others (the neighbor, our children, etc.). But this does not make the relationship any less successful qua friendship or marriage than it would be if we had simply been in agreement about these matters.47 Second, someone might agree that the distribution and exercise of power matters in a friendship (and not just because friendship is committed to consensus), but suggest that the real problem with, say, unilaterally issuing an invitation to the neighbor is that it amounts to making use of an arbitrary power advantage to settle how the couple will proceed in the face of disagreement. What makes these power advantages arbitrary is that they are unjustifed: there is no good reason why the partner who prefers extending the invitation should be able to settle the matter the way he did. More generally, one may think that friendship is incompatible with unjustifed power advantages, but not with justifed ones—and so equal power is, even in relations among friends, a baseline but not an ideal in its own right. But then the friendship conception of relational equality would fare no better than the social status conception when it comes to vindicating the ideal of political equality. Yet this view too seems to me mistaken, because even (otherwise) justifed power advantages may be problematic from the point of view of egalitarian friendship. One way to see this is to recognize that, in the examples mentioned, the party’s use of power need not be unjustifed—except insofar as there is a distinct requirement of equal power. For if it were indeed morally wrong not to invite the neighbor, and the spouse extends the invitation because he recognizes this, then it would seem that he has a justifcation for doing what he did. Now perhaps the thought is that, though his use of the power was 47  Tis is compatible with recognizing that certain kinds of disagreements may make our relationship impossible: if we disagree so deeply that we cannot even see each other’s actions and attitudes as governed by a commitment to equal concern, say, then this will make it difcult for us to sustain our relationship over time.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Power and Equality

29

justifed, the fact that he had the power was not. But what could explain that his possession of the power was unjustifed except that, in possessing the power, he was able to bypass his partner’s disagreement and thus exercise unequal power over the relationship? I don’t see any plausible answer, and thus conclude that our assessment of the situation does presuppose a genuine commitment to equal power among the spouses, not just as a baseline, but as a requirement in its own right. Another way of making this point is to highlight cases where an unequal distribution of power is perhaps even more obviously justifed. If one partner is much more reliable in judging what course of action would be best, but usually cannot persuade the other within the time frame in which a decision has to be made, then an instrumental concern with outcomes would reasonably justify empowering the more reliable partner to make decisions when the conditions just sketched are met. And even if this includes pretty much all of the relevant decisions that have to be made together, on a mere baseline view this would not be regrettable. Yet I think a friendship that would have this shape would be decidedly lopsided, and worse as a friendship. Tis is so even if, all things considered, the instrumental benefts of the unequal distribution would make up for the resulting loss in the value of the relationship. In other words, even if the reasons for political equality are defeated by the reasons against, the inegalitarian distribution of power is regrettable—and so equal power is an ideal in its own right, rather than a mere baseline.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

6. I take away from Section 5 that friendship and similar relationships involve a genuine commitment to an ideal of equal power: friends ideally have (roughly) equal opportunity to infuence the character of their relationship and the norms governing it. Tus relational egalitarian arguments that start from the intrinsic good of friendship, rather than the paradigmatic evil of caste or class societies, can relatively straightforwardly vindicate an ideal of equal power in certain relationships. Tey face, however, a distinctive challenge: they must explain how the ideal of friendship can plausibly be extended beyond the relatively small, face-to-face relationships in which it is usually at home, to cover a much larger political community of the sort governed by modern democratic institutions. How do we justify applying the demands of friendship to an entire polity? One strategy would be to accept that the requirement of equal power applies, in the frst instance, to small-scale interpersonal relationships; but to then argue that the laws that govern our community at large themselves

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

30

Daniel Viehof

shape how we may relate to friends and spouses.48 Yet this strategy runs into trouble. Consider, for instance, an arrangement that denies voting rights to everyone who fails to pass a political knowledge test. Even if half the population were disenfranchised as a result, the actual power each remaining voter has over the laws governing the community would be minuscule. So if our concern were with how much power two friends have over their relationship, the fact that one of them is enfranchised and the other is not would be of limited importance: there are many sources of diferential power among friends, and diferential enfranchisement would have much less impact on the overall power balance than many other social inequalities that friends can regularly tolerate. So the friendship conception of relational equality can plausibly vindicate the ideal of political equality only if political relations are themselves governed by (something very much like) the norms we ordinarily associate with friendship. And this may be doubted. Tere are a number of features that may seem to clearly set apart friendship (and similar relationships) from relations we have (and indeed could have) with our fellow citizens in a political community. To assess the force of this concern, the following discussion asks whether any of the features that most plausibly set apart friendship from political relations centrally bear on whether requirements of equal power apply among friends. What may most obviously distinguish paradigmatic cases of friendship from political relations are the size of the group and the kind of interaction the members engage in. Friendship commonly involves face-to-face interactions (or their mediated counterparts: phone conversations, letters . . .) and is (partly for that reason) limited to groups of a manageable size. By contrast, what we think of as political relations arise commonly among groups the membership of which is many magnitudes larger than even large-ish groups of friends; and consequently there couldn’t be face-to-face interaction among all, or even a signifcant portion, of the polity’s members. But although size and face-to-face interactions are important for under­ standing central aspects of friendship, these features do not seem crucial for understanding the applicability of egalitarian demands of equal power. If groups grow too large to allow for regular face-to-face interaction, this changes the character of the relationship in important ways: the idea that a certain form of emotional intimacy, or certain kinds of interactions that presuppose face-to-face encounters, are central to the group’s character must be abandoned. But other important features could survive: members of the group may continue to take themselves to have special obligations to each other, to be specially committed to each other in particular ways, and so on. And, crucially, 48  Cf. (Viehof 2014), p. 363.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

31

I see no reason for thinking that the mere loss of face-to-face interaction, and the mere increase in size beyond what we associate with standard cases of friendship, would undermine the members’ sense that they should have an equal opportunity to shape the character of the relationship and the norms governing it. Another crucial feature of friendship is its distinctively non-instrumental value, and the fact that those party to it must value it—and each other—in a correspondingly non-instrumental fashion.49 Tis is not to deny that friendship also has instrumental value. (When I am in trouble I may fnd it instrumentally benefcial to have friends rather than be friendless.) Still, there is an important sense in which the instrumental value cannot be too central to it: On a plausible view, the special obligations we have to friends depend on the special value of our relationship, and the special value of our relationship depends on our valuing the relationship, and each other, appropriately—which means, crucially, not just (or even primarily) as instrumentally benefcial. By contrast, it may be suggested, political relations are centrally instrumental in orientation: we make decisions as part of a political community because we need to solve certain problems together. Political relations may, for instance, be necessary for doing justice: without forms of collective action that are made possible by large-scale authoritative decision-making, we couldn’t discharge moral obligations that we owe to one another. And the instrumental value of political relations is largely independent of the attitudes we take toward them. Yet this attempt to distinguish friendship from political relations overstates their diferences. Most importantly, even if political relations have instrumental value because they enable us to realize justice among us, and we would have political obligations on purely instrumental grounds, it may also be true that political relations have additional non-instrumental value (and citizens corresponding obligations) under the right conditions. And among these conditions may be that the citizens suitably value one another, and their relationship, non-instrumentally. A more plausible objection to the friendship conception emphasizes not the diference between instrumental and non-instrumental value, but rather that friendship is optional in an important sense, while political relationships are mandatory, because required for the realization of justice. Correlatively, one might think, what our political relationships should look like is signifcantly constrained by considerations of justice; by contrast, the character of a friendship, though subject to some external moral norms, is importantly 49  For an infuential articulation of this line of thought, see (Schefer 1997).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

32

Daniel Viehof

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

underdetermined by such norms, and it is up to those who are party to the friendship to fll in the picture. More specifcally, obligations among friends depend in part on the actual reciprocal expectations of the parties, in part on what their past interactions have made reasonable to expect of each other; and so these obligations can be shaped, intentionally and unintentionally, by the parties. Tis malleability, one might think, in turn explains the importance of equal power over our friendship: Part of what it is to be friends is to create the friendship together, by shaping its character and the norms governing it. If this is central to friendship, then a commitment to relating to one another as equals requires giving parties an equal opportunity to shape the relationship together as equals— and so we can explain the requirement of equal power among friends.50 But does this fundamentally distinguish friendship from political relations? Even if our political relations are mandated, and signifcantly constrained, by antecedent moral duties (including, centrally, duties of justice), these moral duties underdetermine what our political and social arrangements ought to look like. So the norms governing our political life are also malleable in important ways: diferent political relationships may difer in character, and in the norms that govern relations among co-citizens. And that there are limits to such malleability, set by considerations of justice, does not fundamentally distinguish political relations from friendships, which are similarly constrained by moral requirements. Given the discussion up to now, I do not see why it should in principle be impossible to extend the relational egalitarian ideal with which we are familiar 50  We must distinguish this account of equal power from another view for which it may be mistaken: the view that we each have a personal autonomy interest in shaping our own lives, and thus also in shaping our relationships; and that, in light of our commitment to equal concern, we would also try to advance those interests equally within the relationship, by giving people equal power over it. Elsewhere I have expressed misgivings about such an argument (Viehof 2017). But my concern here is merely to distinguish it from the relational egalitarian account that I sketched. Crucially, the interest that the relational account focuses on is not a general interest in giving shape to our lives, but a specifc interest in shaping this relationship. If the concern were with a general interest in giving shape to our lives, then it would be possible that one person’s interest would be advanced by having control over the relationship, and the other’s by having control over other features of her life. But then we would lack the specifc focus on equal power over the relationship that is, I think, central to our understanding of friendship. Neither does the account just sketched assume that each friend has an interest in shaping the relationship in particular (rather than, as on the view distinguished in the previous paragraph, their life in general), which must then be weighed against similar interests other friends have. Instead it assumes that each friend has an interest in shaping the relationship as an equal together with others; and that the value of shaping and creating the relationship that is internal to the relationship (rather than derived from the more general concern with personal autonomy) is conditional on the shaping and creating being undertaken as equals.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

33

from friendship to the political domain. Tat doing so is possible must, however, not blind us to some limits of this argument when it comes to vindicating political equality as an ideal for an actual political community. Two points in particular deserve highlighting. Te frst follows from the discussion of malleability as it exists in friendships and in political relations: though each is in principle constrained in some ways and malleable in others, political relations may in practice be much more severely constrained, leaving much less room for the distinctive importance of creating the relationship together as equals. Tis may impose important limits on the value of egalitarian political arrangements, and the authority of democratic procedures. In fact, there is a genuine worry that any argument that appeals to the value of egalitarian relationships will be confronted with the fact that, however valuable the relationship is, political outcomes are also of enormous instrumental signifcance—for one, they also afect many other egalitarian relationships, like marriages—so that ensuring that the outcomes are as good as they can be might in practice often take precedence over any concern with the intrinsically valuable relationship in which we might stand to our fellow citizens.51 Te second, and to my mind more important, point is that even if the demands of equal power familiar from the case of friendship may in principle be extended to political relationships, the conditions under which they so extend are much more restrictive than they would have been on the anti-caste model. Te egalitarian requirements associated with the anti-caste model ultimately depend on not much more than our general commitment to people’s equal moral status, and the thought that societal norms should not deny that status. Tis explains why the anti-caste model is in principle compatible with unequal power: such inequality need not cast doubt on society’s commitment to viewing us all as fundamentally equal. It also entails, however, that the egalitarian demands associated with the anti-caste model apply to societies in general, independently of specifc local conditions. Matters are quite diferent when it comes to the demands associated with friendship and analogous political relationships. Tese demands—including, specifcally, the demand of equal power—rest on the existence of intrinsically valuable interpersonal relationships. And these relationships exist only if the parties are in some way or another committed to them: they must value their relationship, and grant a suitable role in their deliberation to its norms. To be clear, the relationship can exist even if the parties do not live up to its ideals: Up to some point, we remain friends even if we are both bad friends and regularly neglect the special obligations we owe to each other. And in a group of people who relate to each other in a certain way, there may be 51  For worries along these lines see, e.g., (Stemplowska and Swift 2018).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

34

Daniel Viehof

signifcant disagreement about the precise character of the relationship, and yet the relationship plausibly exists and makes normative demands. Still, at some point—most obviously if people lack all disposition to treat each other as equals, but plausibly already well before then—the relationship begins to disintegrate, and its value and normative force to disappear. So on the argument that starts from the friendship conception of relational equality, the demands of political equality, and the reason we have to obey democratic decisions, will be conditional on local circumstances in ways that they would not have been on the alternative anti-caste model.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

7. Tis chapter has sought to address a particular problem for recently prominent relational egalitarian accounts of political equality: Some infuential relational egalitarian arguments take as their starting point the problem of social status hierarchy, and are concerned with the ideal of a society not structured by castes or classes. Tis ideal straightforwardly applies to political relations. But it does not in fact vindicate the ideal of equal power, and so not the ideal of political equality either. Others take as their starting point the ideal of egalitarian friendship. Tis ideal does seem to impose a requirement of equal power. But it is much less obvious that this ideal applies to our political community. Let me conclude by returning to the question, briefy touched upon earlier in this chapter, why I have focused on the particular examples of friendship and caste or class structures to make sense of the relational egalitarian argument for political equality. In Section 2 I emphasized that these two cases fgure most prominently in recent discussions of relational equality and democracy. One response to the argument I ofer in this chapter is to suggest that the special attention given to these cases in recent discussions has been misguided; or, at least, that recent discussions (and, as a result, also this chapter) have paid insufcient attention to alternative examples of egalitarian relationships that would in fact simultaneously satisfy the twin conditions that this chapter put center-stage: that equal power be an ideal in its own right (Equal Power), and that it be an ideal applicable to largescale political communities (Political Relationships). What about (to mention just the examples put to me by various audiences) a philosophy department running its afairs collegially, people on a camping trip planning their weekend together, and members of a kibbutz collectively deciding how to organize their common economic life? Such alternative examples could indeed be useful for making sense of the relational egalitarian argument for political equality. But I suspect that their

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Power and Equality

35

usefulness will consist in sharpening our understanding of the relational egalitarian commitments that were implicated by the caste and friendships examples, rather than in putting on the table a diferent set of foundational egalitarian ideals. To make this thought more concrete: I think that egalitarian friendship, and social status equality, are ideals in their own right. By contrast, when I think about various other examples—like philosophy departments, camping trips, or kibbutzim—and their non-derivative value, I am inclined to think that they have such value when they instantiate ideals we paradigmatically associate with friendship or the absence of social status hierarchies. Tese examples may nonetheless be helpful in bringing out what is truly central to the relevant ideal, and what is just a contingent feature of the paradigmatic relationship with which the ideal is often associated. For instance, refecting on the camping trip may help us recognize that relations among friends need not involve emotional intimacy, but can instead just consist in certain joint pursuits. Still, if someone were to point out that, ideally, fellow campers should have equal decision-making power over their endeavor, I would agree, not because there is an independent ideal of equal power that is associated with camping trips, but because there is an ideal of a camping trip among friends that egalitarian decision-making would help instantiate. Tis response—assimilating new examples to the two I have focused on in this chapter—might not always succeed. In particular, it would be inappropriate if the alternative egalitarian relationship cited were truly committed to an ideal of equal power, but the instantiation conditions of that ideal, as exemplifed in that relationship, were quite diferent from those that govern the ideal when it comes to friendship or a society without caste or class inequalities. Yet none of the alternatives I have encountered seem to me to satisfy this requirement. Finally, I have assumed throughout that a relational egalitarian argument for political equality would start from examples of particular relationships, and preferably ones not too closely tied to politics narrowly conceived. In doing so, I have implicitly set aside two other positions. First, one might think that power inequalities between persons are presumptively problematic as such, independently of any particular kind of relationship (other than the relation of unequal power) in which they stand. Tere are special conditions that may make such inequalities unproblematic (or at least less problematic)— exit options, voluntariness, etc.—but if these are absent, the sheer fact that power is distributed unequally is objectionable. On such a view, our objection to unequal power does not rest on a prior account of friendship or equal social status with which unequal power proves incompatible; and so vindicating the ideal of political equality would not require identifying the conditions under which such relationships exist among us, in our political community.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

36

Daniel Viehof

But it seems to me we have important reasons to avoid this position. First, treating unequal power as a general problem, independent of particular relationships, would be most plausible if unequal power gave rise to the same complaint in a wide range of intuitively central examples. Yet one of the upshots of this chapter has been that the deep structure of our objections to unequal power in fact varies signifcantly across diferent paradigmatic cases. If our intuitions about inequality of power in the context of caste or class structures treat equal power as a baseline, whereas our intuitions about, say, friendship and marriage treat equal power as an ideal in its own right, then this casts signifcant doubt on the thought that there is indeed a general problem of unequal power. Second, and perhaps more seriously still, a view that treats unequal power as a general problem needs to account for the many relationships in which power inequalities do not seem even presumptively objectionable as such. For instance, I think the unequal power I have over my seven-year-old daughter is not morally problematic as such, qua unequal power. (It may, however, be problematic on other grounds that have nothing to do with equality.52) And yet my daughter did not enter into the relationship voluntarily, she lacks exit options, etc. Examples like this strongly suggest that the particular kind of relationship in which we stand to another plays a crucial role in determining whether it matters that we have equal power. An alternative position concedes that the ideal of equal power is relationshipspecifc, but suggests that political relationships are a sui generis source of egalitarian demands, independently of any appeal to friendship or social status equality. Tis avoids the worries faced by the view that requirements of equal power are unmediated by particular relationships. But it comes at the cost of giving up on what, in the introduction to this chapter, I suggested is one of the promises of the relational egalitarian defense of political equality: to provide independent argumentative support for our democratic intuitions, rather than simply restating them.53 52  I discuss this issue further in (Viehof 2017). 53  Versions of this chapter have been presented at the 5th OSPP Workshop in Tucson (AZ); at the Colloquium in Political and Legal Teory at Queen’s University, Kingston (ON); at a workshop on Political Equality at NYU; at a “New Work in Legal Philosophy” Workshop, organized by Hrafn Asgeirsson and supported by the University of Surrey School of Law; at Princeton’s University Center for Human Values; at the Kadish Workshop in Law, Philosophy, and Political Teory at Boalt Law School, UC Berkeley; at the Centro de Investigación y Docencia Económicas (CIDE) in Mexico City; and the Eastern APA. I also presented an early sketch in Amanda Greene and Han van Wietmarschen’s graduate seminar in political philosophy at UCL. For their questions and comments I am grateful to the participants at these events, and especially to Daniel Baker (my discussant at Berkeley), Chuck Beitz, Kristen Bell, Colin Bird, Tom Christiano (my discussant in Tucson), Josh  Cohen, Dave Estlund, Marc Fleurbaey, Carina Fourie (my discussant at NYU),

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Power and Equality

37

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Bibliography Anderson, E. (1999). “What Is the Point of Equality?” Ethics 109(2): 287–337. Anderson, E. (2010). Te Imperative of Integration. Princeton, NJ: Princeton University Press. Anderson, E. (2012). Equality. Te Oxford Handbook of Political Philosophy. D. Estlund. Oxford: Oxford University Press. Arneson, R. (2009). Te Supposed Right to a Democratic Say. Contemporary Debates in Political Philosophy. T. Christiano and J. P. Christman. Chichester, UK; Malden, MA: Wiley-Blackwell: 197–212. Berlin, I. (1999 [1956]). Equality. Concepts and Categories. Princeton, NJ: Princeton University Press: 81–102. Carter, I. (2011). “Respect and the Basis of Equality.” Ethics 121(3): 538–71. Chan, T. W. and J. H. Goldthorpe (2004). “Is Tere a Status Order in Contemporary British Society? Evidence from the Occupational Structure of Friendship.” European Sociological Review 20(5): 383–401. Christiano, T. (2008). Te Constitution of Equality: Democratic Authority and Its Limits. Oxford: Oxford University Press. Estlund, D. (2008). Democratic Authority: A Philosophical Framework. Princeton, NJ: Princeton University Press. Fourie, C. (2012). “What Is Social Equality? An Analysis of Status Equality as a Strongly Egalitarian Ideal.” Res Publica 18: 107–26. Fourie, C., F.  Schuppert and I.  Wallimann-Helmer (2015). Te Nature and Distinctiveness of Social Equality: An Introduction. Social Equality: On What It Means to Be Equals. C. Fourie, F. Schuppert, and I. Wallimann-Helmer. Oxford: Oxford University Press: 1–18. Gould, R. V. (2002). “Te Origins of Status Hierarchies: A Formal Teory and Empirical Test.” American Journal of Sociology 107(5): 1143–78. Hohfeld, W.  N. (2001). Fundamental Legal Conceptions as Applied in Judicial Reasoning. Aldershot, UK; Burlington, VT: Ashgate/Dartmouth. Jasso, G. (2001). “Studying Status: An Integrated Framework.” American Sociological Review 66(1): 96–124. Kolodny, N. (2014a). “Rule Over None I: What Justifes Democracy.” Philosophy and Public Afairs 42(3): 195–229. Kolodny, N. (2014b). “Rule Over None II: Social Equality and the Justifcation of Democracy.” Philosophy and Public Afairs 42(4): 287–336. Johann Frick, Stephen Galoob, Amanda Greene, Alex Guerrero, Athmeya Jayaram, Dimitri Landa, Melissa Lane, Claudio Lopez-Guerra, Steve Macedo, Andrei Marmor, Liam Murphy, Kristi Olson, Alan Patten, Ryan Pevnick, Jon Quong, Melissa Schwartzberg, Annie Stilz, Victor Tadros, Jean Tomas, Justin Tosi (my discussant at the APA), Moises Vaca (my discussant at CIDE), Laura Valentini, Steven Wall, Jay Wallace, Gregoire Webber, Han van Wietmarschen, and Ekow Yankah. For written comments or detailed discussions of the argument I owe thanks to Stefen Ganghof, Takuto Kobayashi, Felix Koch, Sam Schefer, Stefan Sciarafa, Kevin Vallier, Juri Viehof, Jake Zuehl, and especially Niko Kolodny and Sophia Moreau.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

38

Daniel Viehof

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Mansbridge, J. J. (1980). Beyond Adversary Democracy. Chicago, IL: Te University of Chicago Press. Miller, D. (1997). “Equality and Justice.” Ratio X(3): 222–37. Neuhouser, F. (2014). Rousseau’s Critique of Inequality: Reconstructing the Second Discourse. Cambridge: Cambridge University Press. Pettit, P. (2012). On the People’s Terms. Cambridge: Cambridge University Press. Scanlon, T. M. (2003). Te Diversity of Objections to Inequality. Te Difculty of Tolerance. Cambridge: Cambridge University Press: 202–18. Schefer, S. (1997). “Relationships and Responsibilities.” Philosophy and Public Afairs 26(3): 189–209. Schefer, S. (2015). Te Practice of Equality. Social Equality: On What It Means to Be Equal. C.  Fourie, F.  Schuppert, and I.  Wallimann-Helmer. Oxford: Oxford University Press. Stemplowska, Z. and A. Swift (2018). Dethroning Democratic Legitimacy. Oxford Studies in Political Philosophy 4. D. Sobel,P. Vallentyne, and S. Wall. Oxford: Oxford University Press. Turner, B. (1988). Status. Minneapolis (MN): University of Minnesota Press. Viehof, D. (2014). “Democratic Equality and Political Authority.” Philosophy and Public Afairs 42(4). Viehof, D. (2017). “Te Truth in Political Instrumentalism.” Proceedings of the Aristotelian Society 117(3): 273–95. Waldron, J. (2002). God, Locke, and Equality: Christian Foundations of John Locke’s Political Tought. Cambridge: Cambridge University Press. Wall, S. (2007). “Democracy and Equality.” Te Philosophical Quarterly 57(228): 416–38.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

2

Rescuing Public Justifcation from Public Reason Liberalism

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Fabian Wendt

Public reason liberals from John Rawls to Gerald Gaus uphold a principle of public justifcation as a core commitment of their theories. Critics of public reason liberalism have sometimes conceded that there is something compelling about the idea of public justifcation. But so far there have not been many attempts to elaborate and defend a comprehensive liberalism that incorporates a principle of public justifcation.1 In this chapter, I will spell out how a principle of public justifcation could be integrated into a comprehensive liberalism and I will argue that everything worthwhile about public justifcation can be extracted from public reason liberalism. I fnd it interesting that most political philosophers seem to either reject the idea of public justifcation completely, or else make it the one-and-only center of their political thinking. To me it seems obvious that the truth must lie in the middle. When laws are publicly justifable, then they are acceptable to everyone, and it is hard not to see this as something worth caring about; on the other hand, it is surely not all that matters. Te picture I advance in the chapter thus tries to give the idea of public justifcation its due without infating its signifcance. It is the picture of a comprehensive liberalism with a place for public justifcation. I hope to be able to show that this middleground position is highly plausible. I proceed in three steps. First, I explain what I take to be at stake between public reason liberals and comprehensive liberals (section 1). Second, I explain what role public justifcation could play in a comprehensive liberalism, and I give two examples (section 2). Tird, I present three objections against the possibility of freeing public justifcation from public reason liberalism without losing what is worthwhile about public justifcation: that it disables public 1  Eberle 2002, Lott 2006, Ebels-Duggan 2010, and Wall 2016 may count as exceptions.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

40

Fabian Wendt

reason (section 3), that the point of public justifcation is lost outside public reason liberalism (section 4), and that public reason liberalism’s distinctively meta-level perspective is lost in comprehensive liberalism (section 5). I try to rebut all three objections. As a caveat, I should point out what I will not (be able to) do in this chapter. First, I will not try to defend a particular rationale for public justifcation. I presuppose that there is something attractive about public justifcation and I argue that we can have all that is attractive about public justifcation without endorsing public reason liberalism. Tis is the main point of the chapter. However, when dealing with the objection that the point of public justifcation is lost outside public reason liberalism, I will at least sketch some arguments for why we should care about public justifcation. Second, I will not defend and spell out a particular version of a comprehensive liberalism that incorporates public justifcation. Te chapter moves on a more abstract level, arguing that public justifcation could be incorporated in all kinds of comprehensive liberalism.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

1. PUBLIC REASON LIBERALISM Public reason liberalism is a family of views that form one of the dominant schools of thought in contemporary political philosophy. It is a family of views, and so there are many disagreements among proponents of public reason liberalism (and not all of them might even endorse the label “public reason liberalism”). Nonetheless I think it is fair to say that the core commitment of public reason liberalism is a commitment to a principle of public justifcation. Tis principle sometimes comes in disguise; in John Rawls’s work, for example, it is called a “liberal principle of legitimacy.” Te principle of public justifcation requires the public justifability of something, and public reason liberals difer in their views about the proper subject of public justifcation. It could be laws, constitutional essentials, political decisions, coercion, claims to authority, moral rules, or something else. For the sake of simplicity, I will here work with a principle of public justifcation that applies to laws, but my arguments should also apply to principles of public justifcation that apply to diferent things. Readers who prefer to apply the principle of public justifcation to something else may replace “laws” with whatever they regard as the proper subject of the principle of public justifcation. As a general formula, laws are publicly justifable when all members of the relevant “public” have sufcient reason to accept them, whereby “having sufcient reason” is taken to be relative to the individuals’ values and beliefs, not relative to some external standard. Steven Wall thus contrasts public

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Rescuing Public Justifcation

41

justifcation with correctness-based justifcation.2 While the former aims to provide reasons that actually speak to the relevant parties, the latter aims to provide reasons that are simply based on the facts and sound moral principles. Of course this formula still allows for many diferent interpretations of public justifcation. Proponents of public justifcation also have diferent views about who is to constitute the relevant public, about whether public justifcation requires public reasons (i.e. reasons that are accessible to all members of the public) or allows for a convergence of non-public reasons, and about other details. I am going to skip over all these issues. Let me provide two paradigmatic examples for principles of public justifcation as they have been advocated by public reason liberals. Rawls’s “liberal principle of legitimacy” requires that “the basic structure and its public policies are to be justifable to all citizens.”3 And Gerald Gaus’s “public justifcation principle” says, in one version, that a “coercive law L is wrongful unless each and every member of the public P has conclusive reason(s) R to accept L.”4 Later, in his book Te Order of Public Reason, Gaus uses the term “not authoritative” instead of “wrongful,” and applies the principle not only to laws, but also to moral rules.5 What is special about public reason liberalism is not just that it advances a principle of public justifcation. Tat principle also has a very special status. For Rawls, it is a principle of “legitimacy,” and for Gaus it determines whether a rule or law is “authoritative.” So the principle of public justifcation gives laws an important moral status. Without satisfying the principle of public justifcation, they are not “legitimate” or “authoritative.” In what follows, I will use the terminology of “legitimacy,” not “authoritativeness.”6 So let us say that the core commitment of public reason liberalism is the following principle: (P1)  Laws are legitimate only if they are publicly justifable. 2  Wall 2002: 386,  2010: 126–7, 133, 136–7. In a similar vein, Eberle distinguishes between “rational” and “public” justifcation (2002: 61–6). 3  Rawls 1993/1996: 224. Te more ofcial statement of the “liberal principle of legitimacy” says that “our exercise of political power is fully proper only when it is exercised in accordance with a constitution the essentials of which all citizens as free and equal may reasonably be expected to endorse in the light of principles and ideals acceptable to their common human reason” (1993/1996: 137; see 1993/1996: 217). 4  Gaus 2010a: 21. 5  Te “basic principle of public justifcation” says: “A moral imperative ‘φ!’ in context C, based on rule L, is an authoritative requirement of social morality only if each normal moral agent has sufcient reasons to (a) internalize rule L, (b) hold that L requires Φ-type acts in circumstances C and (c) moral agents generally conform to L” (2011a: 263). 6  One could think that for laws to be legitimate just means that they are “authoritative” or “binding” for their subjects, since this comes closest to what legitimacy means in talk about state legitimacy. I would like to stay agnostic on this. 

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

42

Fabian Wendt

To be illegitimate or not authoritative is certainly not a minor moral worry about a law. It arguably implies that the law ought not to be enacted and enforced. A public reason liberal who said that it can sometimes be permissible to enact and enforce illegitimate laws in light of important values beyond public justifability would concede my point right from the start and become a comprehensive liberal with a heart for public justifcation. As a principle of legitimacy, (P1) thus cannot be outweighed by competing other moral considerations (it is not a mere “pro tanto” principle), and it does not allow for exceptions (it is not a mere “prima facie” principle). It specifes a strictly necessary condition for legitimacy. But this is not yet sufcient to characterize public reason liberalism. Let me take Rawls’s and Gaus’s versions as paradigmatic examples again. For Rawls, public reason liberalism (he speaks of “political liberalism”) advances not only a principle of legitimacy, but also a theory of justice. And, as is well known, a conception of justice is to be compatible with a plurality of reasonable “comprehensive doctrines” and hence has to avoid taking a stance on moral, philosophical, and religious issues that are disputed among reasonable comprehensive doctrines. Rawls speaks of a “method of avoidance.”7 A conception of justice has to be “freestanding” and ft like a module into diferent comprehensive doctrines.8 Although Rawls’s own conception of “justice as fairness” is his favorite one, he emphasizes that there are other political conceptions of justice, and that “in any actual political society a number of difering liberal political conceptions compete with another.”9 How do justice and the principle of legitimacy relate in Rawls’s account? Because the principle of legitimacy requires the public justifcation of consti­tutional essentials and matters of basic justice, it requires that constitutional essentials and matters of basic justice conform to a political conception of justice, but not to a particular one.10 (All political conceptions of justice are publicly justifable.) In that sense, “legitimacy” (the principle of public justifcation) is “weaker” than justice, yet “related” to justice, as Rawls says.11 And both theories of justice and legitimacy are to abstain from 7  Rawls 1985: 231, 240 n. 22, 1987: 12. 8  Rawls 1985: 230–1, 1987: 3–4, 7–8, 1993/1996: 9–10, 12, 40, 2001: 182–3. 9  Rawls 1993/1996: xlviii; see 1993/1996: xlix, 223, 226, 1997: 770, 774–5. 10  Rawls also says that the liberal principle of legitimacy would be chosen in the original position of “justice as fairness” (1996, 137 n. 5; see also 1996, 225–26; 2001, 89). Yet because he acknowledges that the liberal principle of legitimacy allows coercion in the name of other political conceptions of justice besides “justice as fairness,” the principle arguably also needs a justifcation that is independent from “justice as fairness.” See Neufeld 2010, Freeman (2004, 2027–28).  11  Rawls 1993/1996: 427–8; see also Freeman 2007: 377–9. Similarly, according to Quong, laws are to be publicly justifable because there is reasonable disagreement about justice (2011: 131–5, 137, 219).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Rescuing Public Justifcation

43

controversial moral truths and hence remain in the sphere of what is publicly justifable. A law can be unjust (or not fully just) and at the same time legitimate, but a law cannot be just and at the same time illegitimate. Just like Rawls wants to free his political theory from controversial philosophical, religious, and moral issues, Gaus wants to theorize independently from controversial moral truths and from the question of what reasons “there are” (in contrast to what reasons “people have”).12 One of the main diferences between Rawls’s and Gaus’s public reason liberalism is that Gaus gives the principle of public justifcation a much wider scope. It applies not only to constitutional essentials, but to all laws and also moral rules. Moreover, he assumes an “order of justifcation” and applies the principle of public justifcation to abstract rights, frst of all.13 Tese rights then determine what  laws and rules are publicly justifable. Gaus allows a pluralism of moral standards including standards of justice among the “members of the public” to whom moral rules and laws must be justifable.14 Some of the members of the public are classical liberals, others are egalitarians, and so forth. All members of the public will have to accept less than what they regard as optimal or perfectly just, when the principle of public justifcation does its work: “[W]hen engaging in collective justifcation about a common framework for living, we have reason to endorse common rules even when they do not align with our convictions about what is optimal.”15 It is not perfectly clear, though, whether Gaus would be willing to say that one of the members of the public could be right about justice,16 since he sometimes suggests that justice is whatever set of common laws and rules are publicly justifable.17 So public reason liberalism not only advances a principle of public justifcation, it also excludes all other moral considerations from political theorizing and in particular from theorizing about legitimacy. So there is a second core claim of public reason liberalism: (Q)  Moral considerations beyond public justifability are irrelevant for the legitimacy of laws. A clarifcation about (Q) is important, though: Of course moral con­ siderations beyond public justifability are relevant within public reason liberalism insofar as the question whether a law is publicly justifable 12  Gaus 2011a: 229–35. 13  Gaus 2011a: 275, 387, 510–11. Elsewhere, I argue that treating rights as the subject of public justifcation lacks a coherent rationale in Gaus’s theory (Wendt 2016a, 176–7). 14  Gaus 2011a: 2, 277–8, 445, 548. 15  Gaus 2011a: 502–3; see 2016: 215. 16  See Gaus 2011a: 429, 445–6, 2016: 15, 183, 208, 216, 249–50. 17  Gaus 1996: 121, 2010b: 237, 239. In his recent book, Te Tyranny of the Ideal, he argues that an “Open Society” that provides a framework for a plurality of “perspectives on justice” is needed to even understand one’s own ideal of justice (2016: 243, ch. 2).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

44

Fabian Wendt

depends on what considerations of justice, fairness, liberty, equality, peace, stability, autonomy, democracy, prosperity, public health, culture, art, the environment, and so on are accessible to people, given what they believe and value. Te principle of public justifcation obviously has to work on something. But the point is that once we know which laws are publicly justifable and which laws are not, public reason liberalism claims that this is all that matters for assessing their legitimacy. When there is a confict between what is publicly justifable and what is justifable based on a correctness-based justifcation, only the former is relevant for legitimacy.18 It should be noted that one can be a public reason liberal and thus endorse (P1) and (Q) and yet advance a substantive, correctness-based justifcation for (P1) and (Q). Tis possibility is not formally excluded by the content of (P1) and (Q). I will get back to this in section 5. (Q) does not imply that public reasons liberals have to declare public justifability a necessary and sufcient condition for the legitimacy of laws, as in the following principle:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(P2)  Laws are legitimate if and only if they are publicly justifable. Tis is so because there can be other non-moral necessary conditions for a law’s legitimacy or authoritativeness, for example that the law is actually in place and recognized as such. Tus I will work here with (P1), not (P2). Of course one can imagine political theories that endorse (P1) while rejecting (Q). Public justifability would then be regarded as a necessary condition for legitimacy, but some other moral qualities would be taken as necessary as well. But public reason liberalism, as I want to understand it here, is committed to both (P1) and (Q). I will call liberal political theories that reject both (P1) and (Q) “comprehensive” forms of liberalism, for lack of a better term. Liberal political theories that endorse (P1), but reject (Q), are hybrids and will be ignored here. Critics of public reason liberalism have often targeted specifc versions of public reason liberalism, like, most prominently, John Rawls’s “political liberalism,” but there have also been attempts to refute public reason liberalism as a whole. Tus it has been argued that the principle of public justifcation is self-defeating.19 And it has been argued that there is no coherent reason 18  Some may doubt that there could be such conficts; Rawlsians may say that Rawlsian justice is publicly justifable, since it fts like a module into all reasonable comprehensive doctrines, and Gausians may say that abstract moral rights are themselves a subject of public justifcation. But public reason liberals usually concede that there could be moral truths, including truths about justice and moral rights, and so conficts between what is publicly justifable and what is justifable on a correctness-based justifcation that refers to justice, moral rights, or something else are certainly possible. 19  Wall 2002, 2013b; Enoch 2013: 170–3.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Rescuing Public Justifcation

45

to idealize the constituency within public reason liberalism and that therefore public reason liberalism leads to anarchism (contrary to public reason liberals’ intentions).20 My main goal in this chapter is not to fnd a defeating argument against public reason liberalism. My main goal is to show that everything that looks plausible and worthwhile about the idea of public justifcation can be extracted from public reason liberalism. I want to make a positive case for a  comprehensive liberalism that gives public justifcation its due. But I also think that my case for a public justifcation-friendly comprehensive liberalism is a real challenge for public reason liberalism. If everything that is plausible and worthwhile about public justifcation can be extracted from  public reason liberalism, then it is hard to see why anyone should embrace public reason liberalism. Public reason liberalism suddenly looks baseless and ill-motivated, when its core idea can be taken away without loss.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

2. HOW TO UPHOLD PUBLIC JUSTIFICATION OUTSIDE PUBLIC REASON LIBERALISM Critics of public reason liberalism have sometimes remarked that there is something compelling about the principle of public justifcation, even though public reason liberalism has to be rejected as a whole. Tere indeed is something worrisome about non-publicly justifable laws.21 But how could we integrate public justifcation into a comprehensive liberalism? Comprehensive forms of liberalism, like all normative political theories, contain normative and evaluative claims about the realm of politics. Such claims can be formulated as “principles” (and sometimes it is convenient to do so). Some principles may be absolute principles, other principles may be pro tanto principles, i.e. sometimes outweighed by other principles, and other principles may be prima facie principles, i.e. holding merely as rules of thumb, but knowing exceptions.22 Principles may invoke a great variety of categories. 20  Enoch 2013: 164–70, 2015. 21  See Raz 1998: 51; Wall 2010: 137,  2013a: 488; Enoch 2015: 138–40. What they want to rescue may in the end not be public justifability, though, but “non-subjugation” (but see also Wall 2016). Non-subjugation is the ideal of not forcing people to act against conscience, not subjecting them to one’s directives. In contrast to public justifcation, it does not work with any idealization of persons. I think that there is something valuable about public justifcation and about non-subjugation, but I concentrate on the former in this chapter. On non-subjugation see Wendt 2017. 22  Particularists like Dancy (2004) hold that there are no “principles” in ethics, but they need not reject loose prima facie principles that allow for exceptions. Teir point is that every property that usually makes a positive contribution can in certain contexts make a negative contribution to one’s overall evaluation of something.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

46

Fabian Wendt

Tey may use the vocabulary of good and bad, right and wrong, legitimate and illegitimate, or refer to justice and many other specifc moral categories. It is easy to see, then, that a comprehensive liberalism, while rejecting (P1) and (Q), could endorse some principle of public justifcation that is weaker than (P1). Tere are several possibilities. Here are four: 

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(P3)  (P4)  (P5)  (P6) 

Laws are legitimate if they are publicly justifable. Te public justifability of laws contributes to their legitimacy. It is a desideratum of legitimacy that laws are publicly justifable. Public justifability is a good-making quality of laws.

(P3) is a principle of legitimacy that takes public justifability as a sufcient condition for legitimacy, but allows for the existence of legitimate, but not publicly justifable laws. It does not regard public justifability as a necessary condition for legitimacy. Because laws may be legitimate for other reasons than their public justifability, accepting (P3) is compatible with a rejection of (Q).23 (P4) still is a principle of legitimacy, but public justifability is presented as neither a necessary nor sufcient condition for legitimacy, but as one of several factors that can contribute to the legitimacy of laws. (P5) is somewhat similar, but holds that, no matter what other factors contribute to legitimacy, it is always a desideratum of legitimacy to have publicly justifable laws.24 (P4), in contrast, allows for perfectly legitimate laws that are not publicly justifable. To make sense of the idea of a “desideratum of legitimacy” in (P5), legitimacy has to be conceived as a gradual notion that allows for degrees of legitimacy. Tere may be a threshold of legitimacy that laws have to pass if they are to be permissibly enacted and enforced, but even laws above that threshold can fail to meet all desiderata of legitimacy and hence difer in their degrees of legitimacy. Principles (P3) and (P4), in contrast, are compatible with a non-gradual view of legitimacy, where legitimacy is a pure all-or-nothing matter. (P6), fnally, is not a principle of legitimacy at all. It just states that public justifability is a good-making quality of laws, neither necessary, nor sufcient, nor contributory for legitimacy, nor a desideratum of legitimacy. It just says that whether laws are publicly justifable or not matters morally, among other things. Obviously, (P6) is a weaker principle than for example (P3), and so a comprehensive liberalism 23  As I said, even accepting (P1) is compatible with rejecting (Q), but it results in a “hybrid” position, not in a public justifcation-sensitive comprehensive liberalism. Accepting (P1) is to depart from comprehensive liberalism, since the mark of comprehensive liberalism is the claim that there can be legitimate laws that are not publicly justifable. 24  Wall points out that one need not understand Rawls’s principle as a strict necessary condition for legitimacy; one could also understand it as a mere desideratum (Wall 2014: 417). Tis would make Rawls’s theory a hybrid.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Rescuing Public Justifcation

47

that incorporates (P3) is in a certain sense closer to public reason liberalism than a comprehensive liberalism that incorporates (P6). But what distinguishes all forms of comprehensive liberalism from public reason liberalism is their rejection of (P1) and (Q). Let me now provide two more concrete examples for how a principle of public justifcation like (P3), (P4), (P5), or (P6) could fgure in a comprehensive liberalism. Take natural rights libertarianism. For a natural rights libertarian, central principles of justice are principles about people’s natural rights. Tese natural rights are self-ownership rights as well as rights to appropriate external resources, maybe subject to some sort of Lockean proviso. Tese principles of justice, cashed out in terms of natural rights, may be based on deeper principles that may have to do with respect for persons, or with an ideal of persons as project pursuers, or with human rationality or happiness. Further, they will imply less deep principles about what state institutions ought to do and what they ought not to do, etc. How could public justifability be integrated into natural rights libertarianism? I do not think it would make much sense to add it to the libertarian theory of justice. But libertarians may not only endorse a theory of justice, but also care about other values, like peace, friendship, and, indeed, public justifability, and so they could endorse a principle of the form of (P3), (P4), (P5), or (P6). Tat principle may rest on other foundations than the libertarian principles of justice, but it may also rest on the same foundations. And, together with libertarian justice, it may co-determine the less deep principles about what the state ought and ought not to do, and it might also co-determine verdicts about the legitimacy of laws, if it is (P3), (P4), or (P5). So understood, public justifability will not be part of libertarianism understood as a theory of justice, but it will be part of the canon of values a libertarian can and should endorse. As a side-note: I often speak of “values” in a rather loose sense. Libertarian justice can be taken as a “value,” even though it is formulated in terms of people’s rights, not values. I think it is convenient to speak of “values” as sets of considerations that apply to evaluative and normative reasoning and have some common core or common rationale. Now take, as another example, a liberal perfectionist theory. Tat theory does not start with people’s rights, but places the notion of the good center stage. Although there is a great variety of perfectionist theories, let us here assume that it is a theory that formulates a list of what is objectively good, like autonomy, pleasure, knowledge, love, appreciation of art, development of excellences, etc. All these claims about what is good can be formulated as principles again (if convenient). Liberal perfectionism will also uphold some principles about how promoting the good is (pro tanto or prima facie) right. And it will uphold a principle saying that political institutions should (pro tanto or prima facie) be designed such that they promote the good for all

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

48

Fabian Wendt

citizens (although some liberal perfectionists may also hold that active political promotion of some goods is self-defeating). Justice may be conceived as the set of distributional principles that apply to the political promotion of the good. Tus a perfectionist may uphold a prioritarian constraint on how the good may be politically promoted, saying that the worst-of have a certain priority. How could public justifability be added to such a perfectionist theory? As principle (P6), it could function as another item on the list of objectively good things. As (P3), (P4), or (P5), it could work as a principle that constrains how the good may be politically promoted, a principle that allows public justifability and perfectionist values to co-determine the legitimacy of laws. Tese are just two examples for comprehensive forms of liberalism. Whatever your moral and political values are, you can (and probably should) add public justifability in one way or another. So why should one side with public reason liberalism and exclude all moral considerations beyond public justifability from assessments of the legitimacy of laws, as (Q) requires? I will now discuss three objections against the possibility of extracting public justifcation from public reason liberalism. Tese objections all come from the perspective of public reason liberals and basically state that what is worthwhile about public justifcation is lost outside public reason liberalism. I will argue that the objections fail and that hence we can and should reject public reason liberalism with its claims (P1) and (Q) and instead endorse a more modest principle of public justifcation within a comprehensive liberalism.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

3. FIRST OBJECTION: NO PUBLIC REASON Te frst objection is that public reason liberalism is obviously not only about public justifcation, but also about public reason, and that one loses the idea of public reason when public justifcation is embedded in a comprehensive liberalism. Public reason is to be distinguished from public justifcation; it is a concept that is important to formulate norms for public reasoning and public debate (“duties of civility”). In Rawls’s version of public reason liberalism (“political liberalism”), public reason basically is a set of ideas; its content is given by the family of political conceptions of justice.25 Because political conceptions of justice ft into all reasonable comprehensive doctrines, the contents of public reason are public reasons, reasons that are common ground for all reasonable doctrines: “Public reason . . . specifes the public 25  Rawls 1993/1996: lii–liii, 217, 226, 241, 1997: 773–4, 2001: 92. Maybe the content of public reason is not only constituted by the family of political conceptions of justice, but also by other “political values” besides justice. See Rawls 1997: 776; Freeman 2007: 388–90.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Rescuing Public Justifcation

49

reasons in terms of which [particular] questions are to be politically decided.”26 So in the end, public reason simply is the set of public reasons. Te Rawlsian duty of civility, then, is a moral duty that applies to judges, legislators, state ofcials, and, under certain circumstances, to citizens, and requires them to present and act upon public reasons.27 Duties of civility are central to Rawls’s public reason liberalism and, with modifcations, to many other public reason liberals. Andrew Lister, for example, argues that they are essential if we are to have some sort of community in pluralist societies.28 Now why should we lose the idea of public reason, and relatedly, the idea of duties of civility, when public justifcation is extracted from public reason liberalism? Once we reject (Q) and so allow all kinds of moral considerations in our thinking about the legitimacy of laws, there is no realm of political inquiry where only public reasons ought to prevail. Citizens will often regard moral considerations that are not translatable into public reasons as relevant for their assessment of laws, and comprehensive liberalism allows this. Because this is so, one cannot plausibly ask citizens to only rely on public reasons when arguing with others about political matters. Hence public reason and the duty of civility presuppose (P1) and (Q), and we cannot save them when we give up public reason liberalism in favor of a public justifcation-friendly form of comprehensive liberalism. A little bit more formally, the argument is:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(1) Public reason presupposes some realm where only public reasons ought to prevail. (2) In comprehensive liberalism, there is no such realm. (3) Terefore, there is no place for public reason in comprehensive liberalism. To answer this objection, one should emphasize, frst of all, that not all public reason liberals endorse an idea of public reason. Some public reason liberals allow for a convergence of non-public reasons in public justifcation.29 When public justifcation does not require public reasons, while public reasons are what constitutes public reason, then there is no space for public reason within a public reason liberalism that works with a conception of public justifcation that allows a convergence of non-public reasons. Of course, in such forms of public reason liberalism there is no space for a duty of civility to present public reasons in public, too. Tere could be other duties of civility, for example duties not to vote for laws one believes not to be publicly 26  Rawls 1993/1996: liii. 27  Te duty does not apply when talking privately (in the “background culture”), though. According to what Rawls calls the “wide view of public reason,” citizens are always allowed to present private reasons as long as they present public reasons in due course. Rawls 1993/1996: li–lii, 1997: 783–5, 2001: 90. 28  Lister 2013a: ch. 5. 29  Gaus 2011a: 283–7; Vallier 2014: ch. 4.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

50

Fabian Wendt

justifable, but public reason liberals of the convergence brand are usually rather skeptical about such duties as well, at least with regard to citizens.30 But this is not yet an answer to the objection. If one cares about the idea of public reason, then it would be a worry if a public justifcation-sensitive comprehensive liberalism could not preserve the idea of public reason. But, as far as I can see, one can quite easily make room for public reason outside public reason liberalism, just like one can make room for public justifcation outside public reason liberalism. Once you have a moderate principle of public justifcation like (P3), (P4), (P5), or (P6), one has all one needs to generate a set of ideas to call “public reason.” Assuming that public justifcation is interpreted as requiring public reasons, in this moderate principle of public justifcation, one can simply refer to public reason as the set of public reasons. And once one has the idea of public reason, one can of course also formulate a duty of civility to present public reasons when publicly debating issues of public concern. In other words, the frst premise of the above argument is false: public reason does not presuppose some realm where only public reasons ought to prevail. It does not presuppose (Q). It just presupposes that there is some principle of public justifcation that requires public reasons. Now this may seem a little too easy. Te objection was that the idea of public reason presupposes a realm of normative inquiry where non-public moral considerations are to be excluded. It may be technically possible to construe some sort of public reason outside public reason liberalism, but the objector could insist that it lacks a rationale when there is no realm where only public reasons ought to prevail. I think this is wrong. Once one acknowledges the value in public justifability, one can also acknowledge the importance of presenting public reasons in public. What is this value in public justifability? Some will argue that public justifability serves stability and helps to build mutual trust; others will argue that it is an expression of respect; still others will say that it constitutes some sort of community (more on these reasons to care about public justifcation in section 4). Te same considerations may ground the ideas of public reason and duties of civility. Tis does not imply that duties of civility have to be conceived as “absolute.” In fact it is much more plausible that other moral considerations matter as well, and that hence the duty of civility can sometimes (maybe rarely) be outweighed by other moral considerations. Tis is acknowledged by some public reason liberals like James Boettcher and Andrew Lister.31 Some may still worry that something is lost when we abandon public reason liberalism and allow all kinds of considerations in our thinking about legitimacy: Tere no longer is a proper “public point of view.” Of course, a 30  Gaus 2010a; Vallier 2014: ch. 6. 31  Boettcher 2007: 233, 2012: 168–70; Lister 2013a: 109–10, 128–9.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Rescuing Public Justifcation

51

lot depends on what exactly is meant by the “public point of view.” But if the public point of view is constituted by public reason, then there still is a public point of view in a comprehensive liberalism that makes room for public justifcation (at least if what I argued above is correct). What is new is only that taking the public point of view is no longer regarded as all that matters for the legitimacy of laws.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

4. SECOND OBJECTION: PUBLIC JUSTIFICATION LOSES ITS POINT Te second objection is that public justifcation loses its point when it is extracted from public reason liberalism. I cannot discuss all the points public justifcation could be said to have, of course. I will focus on three I fnd most convincing. I will try to show that public justifcation can be integrated into a comprehensive liberalism without any loss. In fact, it is more plausibly integrated into a comprehensive liberalism. A frst rationale for public justifcation is stability. In modern pluralist societies, where deep disagreements and conficts are normal and pervasive, publicly justifable laws have a stabilizing efect. Tis is so for two reasons. First, when laws are publicly justifable, they are acceptable for all citizens. Acceptability is not by itself stabilizing, because laws can be acceptable without actually being accepted, when people fail to see that they are justifable to them. But at least it seems more likely that publicly justifable laws will sooner or later actually be accepted because, after all, everyone has sufcient reason to accept them. And actually accepted laws certainly serve stability and peace, since they do not give rise to resentment and anger.32 Second, public justifability could also serve stability because it allows people to rely on public reason in public debates (at least when public justifcation is conceived as requiring public reasons).33 Everyone wants to have laws that are justifable to him or her, and when others publicly signal that they care about public justifability, this can help to foster trust and cooperation, which can obviously have a stabilizing efect. Now why should public justifcation lose its stabilizing efects when integrated into a comprehensive liberalism? Because allowing non-public moral reasons in people’s deliberations about the legitimacy of laws is a threat to peace already. It is a threat to peace because morality “does not fax its 32  See also Wendt 2016a: 139–44. 33  For discussion of the stabilizing efect of public reason see Weithman 2011: 327–35; Gaus  2011b; Hadfeld and Macedo  2012; Trasher and Vallier  2015; Klosko  2015; Weithman 2015.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

52

Fabian Wendt

demands down from above,” as Gaus puts it,34 and so allowing non-public moral reasons in people’s deliberations means that diferent people’s views about moral truth will clash. People will try to push each other around in the name of what they regard as right and good. For that reason, we have to strictly bracket controversial convictions about moral truth in order to achieve stability. So the argument goes something like this:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(1) According to comprehensive liberalism, one has to consider all kinds of moral reasons when assessing the legitimacy of laws. (2) Because people often deeply disagree about moral issues, it leads to conficts and instability when people consider all kinds of moral reasons in their assessment of the legitimacy of laws. (3) To avoid these conficts and instability, we have to exclude non-public moral reasons from the assessment of the legitimacy of laws, i.e. endorse public reason liberalism with its principles (P1) and (Q). Tus formulated, it is at frst an argument about how people should assess the legitimacy of laws, not about how people may debate the legitimacy of laws. It is about people’s thinking, not their talking. But one could easily formulate a derivative argument that makes the same statements about the latter. It should also be noted that the conclusion presupposes a principle of public justifcation that requires public reasons and does not allow for a convergence of non-public reasons. Tere are three replies to the argument. First, even if we concede that premise (2) is (more or less) correct, (3) does not follow. We do not have to maximize stability and do everything we can to avoid any conficts and instability. Endorsing a weaker principle of public justifcation, like (P3), (P4), (P5), or (P6), together with related duties of civility, may be sufcient to achieve sufcient stability, even if not maximal stability. Second, one should not overestimate the impact of people’s thinking about legitimacy for the stability of a society. Premise (2) may not be straightforwardly false, but it should be put in perspective. Tere are so many other things that contribute to a society’s stability. A political culture of civility, respect, and toleration can also be maintained among citizens who do not give much weight to public justifability and who do not refrain from assessing laws’ legitimacy in light of all kinds of moral considerations. Citizens may simply believe that fostering attitudes of civility, respect, and toleration is the morally right way to deal with disagreements and conficts. Tere may well be a correctness-based justifcation for civility, respect, and toleration. It is a mistake to believe that a commitment to comprehensive liberalism leads to people trying to impose their views on others by force. Moreover, besides political 34  Gaus 2011a: 11.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Rescuing Public Justifcation

53

culture, there are many other factors that contribute to stability. Sophisticated modus vivendi arrangements and checks and balances are important for stability.35 Community ties are relevant for stability, as well as economic development and economic interdependencies. Claudia Mills argues that a common history of living together is more important than shared principles, Bernard Dauenhauer points out how a shared religion as well as linguistic and cultural heritages can increase stability, and Joseph Raz suggests that “afective and symbolic elements may well be the crucial cement of society.”36 How people think about legitimacy may indeed have some efect on stability, but it would be exaggerated to conclude that therefore public justifability has to be all that matters for the legitimacy of laws. Too many other factors contribute to stability as well. A public reason liberal may reply that only people’s commitment to public justifcation can provide stability “for the right reasons.”37 But one may wonder why a commitment to peace or cultural and afective ties should not count as the right kind of reasons as well. Moreover, even a modus vivendi that is sustained by pure self-interest based reasons can be highly stable, and one may well ask why these self-interest based reasons should not count as the “right kind,” if our concern is stability.38 Tird, a simple point: what academic philosophers think about the legitimacy of laws and about how people ought to think about the legitimacy of laws will unfortunately have little efect on how people actually think about it. So in contrast to what (3) says, adopting public reason liberalism in academic philosophy will be of little help with the practical problem of stability. What matters for stability is that publicly justifable laws are acceptable to all and thus make actual acceptance more likely. Tere is no reason to assume that this efect cannot be had if we adopt a comprehensive liberalism. For all three reasons—because stability is not the only thing that matters, because other things beyond people’s way of thinking about legitimacy contribute to stability, and because of the little efect of academic philosophy— the stability rationale for public justifcation can be maintained in compre­ hensive liberalism. Indeed it is more plausibly maintained in comprehensive liberalism, because other values beyond stability are given their due, and because the impact of public justifability (and public reason) on stability is acknowledged, but not exaggerated. 35  Hershovitz 2000: 224–5. 36  Mills 2000: 192, 194, 197–203; Dauenhauer 2000: 212; Raz 1990: 30–1. 37  See Rawls (1987, 11; 1996, xxxix–xliii, 142–3, 145, 147–8; 2001, 185–6, 195).  38  See also Wendt (2016a, 141–2; 2016b, 359–60). Some reasons may not be of the “right kind” because they do not allow us to live in community (Rawls, 1996, 202, 146–7; 2001, 199–200). But the community concern is distinct from the stability concern.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

54

Fabian Wendt

A second reason to care about public justifcation may be that it is an expression of respect for persons.39 Both the public justifability of laws and people’s reliance on public reason or caring about public justifcation can be regarded as expressions of respect for persons. I will here work with the former version: the fact of a law being publicly justifable is taken as an expression of respect for all persons subject to it. One may well ask whose respect could be expressed by publicly justifable laws, but I cannot go deeper into this issue here. I will also not try to answer the question of what is respected in persons. It could be their reason, their integrity, or something similar. Why should publicly justifable laws not be able to express respect for persons in comprehensive liberalism? After all, we envisage a comprehensive liberalism that does include a principle of public justifcation—a principle like (P3), (P4), (P5), or (P6). So why should we adopt public reason liberalism? Te answer probably is that respect for persons is so important that it should be made an “absolute principle,” not possibly outweighed by other moral considerations. Te argument could go roughly like this:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(1) Publicly justifable laws are an expression of respect for persons. (2) If publicly justifable laws are an expression of respect for persons, then not publicly justifable laws are disrespectful. (3) Respecting persons is extremely important and should not possibly be outweighed by other moral considerations in assessing the legitimacy of laws. (4) Terefore, laws can only count as legitimate if they are publicly justifable, as expressed in public reason liberalism’s core principle (P1). Te problem with comprehensive liberalism and its rejection of (P1) is that it allows for disrespectful but at the same time legitimate laws, and so the respect rationale for public justifcation cannot be preserved in comprehensive liberalism. One line of response to this objection is to reject premise (2). To have publicly justifable laws may be an expression of respect for persons, but this does not imply that not publicly justifable laws are disrespectful. Tere may be other ways to express respect for persons.40 Following Christopher Eberle, one could argue that it is an expression of respect for persons to care about whether laws are publicly justifable, and to actively pursue public justifcation in public, but that it is not disrespectful to have laws that are 39  Macedo 1990: 47; Nagel 1991: 159; Bird 1996: 75–6; Larmore 1999: 607–8, 610; Rawls 2001: 91; Eberle 2002: 87–8, 94; Neufeld 2005: 284–7; Boettcher 2007: 230–3; Freeman 2007: 330, 343–4, 411; Gaus 2011a: 17, 19; Nussbaum 2011: 18–20; Vallier 2014: 31–3; Wall 2016; Wendt 2016a: 152–4. 40  For a related argument see also Lott 2006: 87–9 and the replies in Boettcher 2012: 170 and Wendt 2016a: 155–7.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Rescuing Public Justifcation

55

not publicly justifable, when there is a correctness-based justifcation for them.41 If that is right, it would not only rebut the objection, but also speak in favor of integrating public justifcation into a comprehensive liberalism. Another line of response is to reject premise (3). Maybe respect is not so important that it should not possibly be outweighed by other moral considerations in our assessment of the legitimacy of laws. First of all, one should note that not all morally required treatment of persons is a matter of respect. Not to murder an innocent person is not an expression of respect, just like it sounds odd to say that murder is “disrespectful.” Murder is wrong not because it is disrespectful, but because it is a serious violation of a basic moral right. What is “disrespectful” is failing to do things that one owes persons, but that one owes them not as a matter of their basic moral rights.42 Because this is so, respect for persons should not be regarded as the most important moral imperative. Justice and people’s rights may well be regarded as more important. Values like peace and stability may also be regarded as more important. If that is so, respect for persons is a moral consideration that can indeed sometimes be outweighed by other moral considerations in the assessment of the legitimacy of laws.43 But then the same holds for public justifcation as an expression of respect for persons: it can also sometimes be outweighed by other moral considerations, be it considerations of justice and people’s moral rights, considerations about peace, or something else. Second, sometimes a law may not be publicly justifable and hence disrespectful even though its purpose is to fght disrespectful treatment of persons, and even though it achieves its purpose. It is not evident that such a law could never be morally justifed, all things considered.44 Tus premise (3) is false. To put it the other way around: it is a worry about public reason liberalism that it does not allow for trade-ofs between public justifability and other values, since rights, justice, peace, etc. can sometimes be more important than respect for persons as it is expressed in publicly justifable laws.

41  Eberle 2002: chs. 4–5; see also Lott 2006 and Ebels-Duggan 2010. One could go one step further and argue that having publicly justifable laws is not an expression of respect for persons at all. It is just the caring about public justifcation and pursuing of public justifcation that is an expression of respect for persons. 42  Tings are more complicated if you think that persons have a basic moral right against being treated disrespectfully. If people have such a right, then we should say that what is “disrespectful” is not to do things that we owe persons, but that we owe them not as a matter of other basic moral rights besides the right against being treated disrespectfully. 43  Of course it is possible to advance a broader notion of respect according to which anything we owe persons is a matter of respect (including not to murder them, etc.). But this notion of respect would merely hide the fact that the respect that is expressed in publicly justifable laws can sometimes be outweighed. 44  Wall 1998: 86–7; Lister 2013a: 72, 2013b: 324–5.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

56

Fabian Wendt

I conclude, then, that the respect rationale for public justifcation—if sound—can well be maintained outside public reason liberalism. In fact it is more plausibly maintained in comprehensive liberalism, because respect can indeed be outweighed by other moral considerations.45 A third possible rationale for public justifcation is that publicly justifable laws are needed if we are to have a community of mutual moral authority. Tis is what Gaus argues. In a somewhat Wittgensteinian manner, he wants to provide an “internal” foundation for the principle of public justifcation, showing that, on refection, a commitment to public justifcation is internal to our everyday moral practice.46 One central aspect of our moral practice is that we hold each other responsible for what we do, and that we blame others when they do wrong, and accordingly feel moral emotions like indignation. Yet when a person cannot understand why her action was wrong, then we do not think that blame and indignation are appropriate.47 In that sense, we do not claim moral authority in such cases. Tis is also why, for example, we do not blame little children, animals, or psychopaths for what they do.48 Gaus concludes that we presuppose the following principle in our moral practice: “A moral prescription is appropriately addressed to Betty only if she is capable of caring for a moral rule even when it does not promote her wants, ends or goals and she has sufcient reasons to endorse the relevant rule.”49 It is easy to see that the second part of this principle is a principle of public justifcation. When a moral rule (or, in our context, a law) is not justifable to a person, then we cannot blame that person for not acting in accordance with that rule, and it would be inappropriate to feel indignation. Te point of the principle of public justifcation is to constitute a moral community, a community of mutual moral authority, where it makes sense to hold each other responsible and to blame those who violate rules or laws. Tus to adhere to the principle of public justifcation is highly important because it means to sustain moral relations with other people. As with stability and respect, I here grant that this indeed is a rationale for public justifcation. Te objection is that this rationale cannot be preserved in a public justifcation-sensitive comprehensive liberalism. Te reason is that comprehensive liberalism allows all kinds of moral considerations to 45  One objection to my line of argument could be that my focus on laws allows me to say that it is not so bad if some laws are not respectful; it allows me to easily accept tradeofs. Tings look diferent when we apply the principle of public justifcation not to laws, but to the “basic structure” of society (its constitution etc.). Here, respect and public justifcation become all-or-nothing matters. In reply, I think one should resist thinking about respect and public justifcation in all-or-nothing terms. Even a society’s constitution can be publicly justifable in parts, and not publicly justifable in other parts. For that reason, my arguments can work in exactly the same way when we apply the principle of public justifcation not to laws, but to the basic structure of society. 46  Gaus 2011a: 226. 47  Gaus 2011a: 184, 258. 48  Gaus 2011a: 210. 49  Gaus 2011a: 222. He calls it the “principle of moral autonomy.”

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Rescuing Public Justifcation

57

determine the legitimacy of laws, and so we can end up with legitimate laws that are not publicly justifable, and hence upset our moral community by destroying relations of mutual moral authority. If public justifcation is the cement of our moral community, then a little bit of public justifcation will not do. We have to make it a strictly necessary condition of legitimacy, as public reason liberalism’s principle (P1) does. So the argument goes as follows:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(1) It is highly important to have a community with relations of mutual moral authority. (2) Not publicly justifable laws upset a community with relations of mutual moral authority. (3) Terefore, laws can only count as legitimate if they are publicly justifable, as expressed in public reason liberalism’s core principle (P1). In response, one can doubt premise (2). One can doubt that a moral community with relations of mutual moral authority requires that all laws are publicly justifable. Arguably a person can stay a member of the moral community even when some laws are not justifable to him or her. It might be the case that one cannot claim moral authority over that person with regard to some moral rules or laws, but membership in the community need not be defned so narrowly. Te person can stay a member as long as most laws are justifable to him or her. Tis more fexible view of community membership will probably also be supported by stability concerns. Hence a community of mutual moral authority can be maintained without (P1). Second, having a moral community of mutual authority is, like stability and respect, not all that matters, even if it is highly important. In particular, while having a community of mutual moral authority is highly important, it is much less important to have everyone included in that community. Terefore, one can deny that (3) follows from (1) and (2). Sometimes other values, like peace, justice, and rights, indeed are more important than having everyone included in a community of mutual accountability. Tis is (implicitly) conceded by Gaus, when he says that some people cannot be members of the moral community.50 He does not explicitly present moral arguments for this claim, but there are moral arguments for it. Because other values besides moral community matter for whether laws are legitimate or not, the moral community rationale is better maintained in a public justifcation-sensitive comprehensive liberalism that rejects (P1) and (Q). Tis has a further welcome side-efect. Rejecting (P1) and (Q) helps to explain why there is no “blameless liberty to act as we see ft” with regards 50  Gaus 2011a: 282–3, 2014: 566, 2016: 222. See also Van Schoelandt 2015.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

58

Fabian Wendt

to psychopaths and others outside the moral community.51 We have laws with regard to psychopaths, animals, and others outside our community of mutual moral authority, and their legitimacy depends on moral values beyond public justifability. I conclude that the moral community rationale for public justifcation can be maintained outside public reason liberalism, and that it should be maintained outside public reason liberalism, since we cannot escape the fact that moral considerations beyond having a moral community of mutual relations of authority matter for the legitimacy of laws.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

5. THIRD OBJECTION: PUBLIC REASON LIBERALISM WORKS ON A DIFFERENT THEORETICAL LEVEL Another objection to the viability of public justifcation outside public reason liberalism is that public reason liberalism works on a diferent theoretical level than comprehensive liberalism, and that hence public justifcation cannot fulfll the same function when it is transferred from public reason liberalism to comprehensive liberalism. Tis is the picture: While traditional political theories inquire into what the objectively best laws would be, public reason liberalism asks how we can get along given that we (reasonably) disagree about what the best laws would be. Because fghting about what the best laws would be leads to all kinds of troubles, public reason liberalism moves to a meta-level and refuses to be another party on the battlefeld. Instead, it tries to fnd ways to contain the frst-level fghting in ways that are acceptable to all. Hence public reason liberalism wants to play a diferent role: instead of being a sectarian party, to use Rawls’s expression,52 it wants to be something like a mediator.53 As Gaus puts it, public reason liberalism is not concerned with uncovering “moral truth,” but with fnding a self-sustaining “moral constitution,” i.e. a “shared moral framework all can live with in a social world where our understandings of moral truth clash.”54 Public justifcation is the core idea within this meta-level project, and we lose this meta-level when we move back to a comprehensive liberalism, since comprehensive forms of liberalism are players on the frst-level battlefeld. Tis is why we cannot 51  Tat there is such a blameless liberty is afrmed in Gaus 2011a: 463. Enoch takes this as a reductio ad absurdum of Gaus’s theory (2013: 170). 52  Rawls 1985: 246, 1987: 20, 1993/1996: 129, 180. 53  Tis is perceptively described in Enoch 2015: 134–7; also 2013: 174–6. 54  Gaus 2010b: 244, 2014: 564; see 2016: 178, 181, 183. Te term is from Rawls 1980: 539.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Rescuing Public Justifcation

59

preserve what is worthwhile about public justifcation when we extract it from public reason liberalism. So the core of the objection is roughly this:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

(1) We should engage in second-level moral reasoning and search for a shared framework all can live with. (2) Te principle of public justifcation is a principle in second-level moral reasoning. (3) Comprehensive liberalism engages in frst-level moral reasoning. (4) Terefore, the principle of public justifcation can no longer work as a principle in second-level reasoning when it is integrated in comprehensive liberalism. Te answer to this objection is that comprehensive liberalism can (and should) make a distinction between two diferent levels in the moral evaluation of laws.55 On the frst level, we try to determine what the best laws would be; on the second level, we determine what laws would be acceptable given that other people disagree with us about what the best laws would be. A comprehensive liberal need not endorse the principle that it is permissible to simply impose one’s view on others. A comprehensive liberal can happily endorse premise (1) and be highly sensitive to the moral dimension in dealing with conficts and disagreement. A comprehensive liberal can also see the importance and value in a “moral constitution.” So premise (3) is wrong (or at least incomplete): comprehensive liberalism can engage in both frst- and second-level moral reasoning. What is more, on the second level values like public justifability come into play, and so the principle of public justifcation can stay on the second level of moral thinking when it is embedded in a comprehensive liberalism. Tis means that conclusion (4) is wrong, too. Public justifcation can stay on the same level of moral thinking in both public reason liberalism and comprehensive liberalism. Tis distinction between the two levels of moral evaluation also sheds new light on the above distinction between public justifcation and correctnessbased justifcation. Te dichotomy stresses that there are two diferent forms of justifcation with diferent goals, but this obscures the fact that public justifability functions as one consideration among many other considerations within correctness-based justifcation, when it is integrated into a compre­ hensive liberalism. On the frst level of the evaluation of laws, we engage in correctness-based justifcation and evaluate laws in terms of their justice, efciency etc., and we determine what the best law would be in light of these values. On the second level we take into account that others disagree about what the best law would be, and thus we consider moral values that become 55  For an elaboration and defense of the distinction between two levels see Wendt 2016a: ch. 3.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

60

Fabian Wendt

relevant under such circumstances of disagreement.56 Here, public justifability co-determines what the all things considered best laws and the all things considered acceptable laws are within a correctness-based justifcation. What public reason liberalism needs, then, is an argument as to why political philosophy should only engage in second-level moral reasoning when assessing the legitimacy of laws.57 (Of course, public reason liberals are interested in frst-level reasoning when they apply the principle of public justifcation. But applying the principle of public justifcation is itself an exercise in second-level reasoning as it is defned here, and, according to public reason liberals, it is all that counts for the legitimacy of laws.) I already mentioned one such argument: that we should not become one sectarian party in the moral and political conficts we face. But one may continue asking why it is so important to avoid sectarianism. Tis is basically equivalent to asking for a rationale for public reason liberalism with its core claims (P1) and (Q). Now when public reason liberals try to provide such a rationale for public reason liberalism, they have to point to the importance of stability, respect, moral community, or something else. As soon as they do this, they seem to provide a correctness-based justifcation for public reason liberalism (as is conceded by some public reason liberals58). But then it seems that public reason liberalism cannot be completely “unsectarian,” and hence cannot live up to its own standards. And when non-sectarianism is not an option anyway, one may well continue making frst-level evaluations of laws and stick to a comprehensive liberalism. In reply, Charles Larmore argues that the respect rationale for public reason liberalism is merely “minimally moral” and as such compatible with all reasonable comprehensive doctrines; it does not make public reason liberalism another sectarian doctrine.59 But, in reply to Larmore, one may well argue that the respect rationale cannot be so minimal and unsectarian if it is to ground the quite radical claims (P1) and (Q).60 Some public reason liberals refuse to provide a rationale for public reason liberalism, following the mentioned “method of avoidance,” and so they may 56  Some foundational values for public justifcation are relevant on the frst level of moral evaluation: stability and respect matter not only under circumstances of disagreement, of course, but determine what the best laws would be. Public justifability, on the other hand, is a pure second-level value; it does not matter when there is no disagreement about what the best laws would be. 57  More precisely, it needs an argument as to why we should only engage in that part of second-level reasoning that employs the principle of public justifcation. 58  Larmore  1999: 608–11; Neufeld 2005: 287; Estlund  2008: 23, 51–2, 57–60; Quong 2011: 2, 56, 159, 313; see also Bajaj 2017: 3148–50. Tis would mean that Raz is right that public reason liberalism cannot be epistemically abstinent (1990: 14–15). It is also related to Wall’s insistence on the primacy of the frst-person standpoint in moral reasoning (2010: 136–40). Gaus concedes that primacy (2011a: 225–6, 228–9). 59  Larmore 1999: 623–4; see also 1990: 353–4, 1994: 61, 1999: 600, 605, 608. 60  Wall 2002: 390–1; Quong 2013: 272.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Rescuing Public Justifcation

61

indeed be able to avoid “sectarianism.” But the question why we should endorse public reason liberalism is a pressing one for public reason liberals. (P1) and (Q) are far from self-evident, after all. Some Rawlsians delegate the answer to comprehensive doctrines (and hence out of political philosophy). Rawls writes that a “reasonable judgment of the political conception must still be confrmed as true, or right, by the comprehensive doctrine.”61 In a similar vein, Jonathan Quong argues that public reason liberals should see it as the task of comprehensive doctrines to ultimately fnd the right reasons to accept the public reason project. Public reason liberalism “passes the buck” to reasonable citizens.62 But this just means that we have no argument to engage with when asking for a reason why we should endorse public reason liberalism. A second strategy for fnding a rationale for public reason liberalism could be to defend skepticism about frst-level moral reasoning.63 Te main difculty for this strategy would be to explain why sound second-level moral reasoning about public justifability and legitimacy should nonetheless be possible. But anyway, as mentioned earlier, most public reason liberals do not want to make this move and stay agnostic about the existence of moral facts and moral truths and the possibility of sound frst-level moral reasoning.64 Tus, for the time being, we may well endorse a public justifcation-sensitive comprehensive liberalism, and conclude that everything worthwhile about public justifcation can be enjoyed outside public reason liberalism.65

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

References Bajaj, Sameer 2017: Self-Defeat and the Foundations of Public Reason. Philosophical Studies 174: 3133–51. Barry, Brian 1995a: Justice as Impartiality. Oxford: Clarendon Press. Barry, Brian 1995b: John Rawls and the Search for Stability. Ethics 105: 874–915. Bird, Colin 1996: Mutual Respect and Neutral Justifcation. Ethics 107: 62–96. 61  Rawls 1997: 801; see also 1993/1996: 128–9. 62  Quong 2011: 226–42, 2013: 274–5. 63  Tis is not exactly what Barry does, since he defends skepticism about the good, but not justice, and justice is a frst-level value (1995a: 168–73, 1995b: 902–3). 64  See Nagel 1987: 227–31; Rawls 1987: 12–13, 1993/1996: 62–3, 150; Larmore 1994: 79; Bird 1996: 78–80; Quong 2011: 243–54; also Gaus 2011a: 229, 233, 2016: 183. 65  For helpful comments and discussion, I would like to thank Colin Bird, Alistair Isaac, Benjamin Kiesewetter, Frodo Podschwadek, Kevin Vallier, Steve Wall, an anonymous reviewer for Oxford University Press, and the audiences at the 5th Oxford Studies in Political Philosophy workshop in Tucson in October 2017, the political philosophy workshop in Flensburg in July 2017, the workshop on “Liberty and Well-Being” in Oxford in June 2017, and a research seminar in Hamburg in November 2016.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

62

Fabian Wendt

Boettcher, James W. 2007: Respect, Recognition, and Public Reason. Social Teory and Practice 33: 223–49. Boettcher, James W. 2012: Te Moral Status of Public Reason. Journal of Political Philosophy 20: 156–77. Dancy, Jonathan 2004: Ethics without Principles. Oxford: Clarendon Press. Dauenhauer, Bernard P. 2000: A Good Word for a Modus Vivendi. In V. Davion and C. Wolf (eds.): Te Idea of a Political Liberalism: Essays on Rawls (pp. 204–20). Lanham, MD: Rowman & Littlefeld. Ebels-Duggan, Kyla 2010: Te Beginning of Community: Politics in the Face of Disagreement. Te Philosophical Quarterly 60: 50–71. Eberle, Christopher J. 2002: Religious Conviction in Liberal Politics. Cambridge: Cambridge University Press. Enoch, David 2013: Te Disorder of Public Reason. Ethics 124: 141–76. Enoch, David 2015: Against Public Reason. In D. Sobel, P. Vallentyne, and Steven Wall (eds.): Oxford Studies in Political Philosophy 1 (pp. 112–42). Oxford: Oxford University Press. Estlund, David 2008: Democratic Authority: A Philosophical Framework. Princeton, NJ: Princeton University Press. Freeman, Samuel 2007: Rawls. London: Routledge. Gaus, Gerald 1996: Justifcatory Liberalism: An Essay on Epistemology and Political Teory. Oxford: Oxford University Press. Gaus, Gerald 2010a: Te Place of Religious Belief in Public Reason Liberalism. In M. Dimova-Cookson and P. Stirk (eds.): Multiculturalism and Moral Confict (pp. 19–37). London: Routledge. Gaus, Gerald 2010b: Coercion, Ownership, and the Redistributive State: Justifcatory Liberalism’s Classical Tilt. Social Philosophy and Policy 27: 233–75. Gaus, Gerald 2011a: Te Order of Public Reason: A Teory of Freedom and Morality in a Diverse and Bounded World. Cambridge: Cambridge University Press. Gaus, Gerald 2011b: A Tale of Two Sets: Public Reason in Equilibrium. Public Afairs Quarterly 25: 305–25. Gaus, Gerald 2014: Te Good, the Bad, and the Ugly: Tree Agent-Type Challenges to “Te Order of Public Reason.” Philosophical Studies 170: 563–77. Gaus, Gerald 2016: Te Tyranny of the Ideal: Justice in a Diverse Society. Princeton, NJ: Princeton University Press. Hadfeld, Gillian K. and Stephen Macedo 2012: Rational Reasonableness: Toward a Positive Teory of Public Reason. Te Law & Ethics of Human Rights 6: 7–46. Hershovitz, Scott 2000: A Mere Modus Vivendi? In V. Davion and C. Wolf (eds.): Te Idea of a Political Liberalism: Essays on Rawls (pp. 221–30). Lanham, MD: Rowman & Littlefeld. Klosko, George 2015: Rawls, Weithman, and the Stability of Liberal Democracy. Res Publica 21: 235–49. Larmore, Charles 1990: Political Liberalism. Political Teory 18: 339–60. Larmore, Charles 1994: Pluralism and Reasonable Disagreement. Social Philosophy and Policy 11: 61–79. Larmore, Charles 1999: Te Moral Basis of Political Liberalism. Te Journal of Philosophy 96: 599–625. Lister, Andrew 2013a: Public Reason and Political Community. London: Bloomsbury.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Rescuing Public Justifcation

63

Lister, Andrew 2013b: Te Classical Tilt of Justifcatory Liberalism. European Journal of Political Teory 12: 316–26. Lott, Micah 2006: Restraint on Reasons and Reasons for Restraint: A Problem for Rawls’s Ideal of Public Reason. Pacifc Philosophical Quarterly 87: 75–95. Macedo, Stephen 1990: Liberal Virtues: Citizenship, Virtue, and Community in Liberal Constitutionalism. Oxford: Clarendon Press. Mills, Claudia 2000: “Not a Mere Modus Vivendi”: Te Bases for Allegiance to the Just State. In V. Davion and C. Wolf (eds.): Te Idea of a Political Liberalism: Essays on Rawls (pp. 190–203). Lanham, MD: Rowman & Littlefeld. Nagel, Tomas 1987: Moral Confict and Political Legitimacy. Philosophy & Public Afairs 16: 215–40. Nagel, Tomas 1991: Equality and Partiality. Oxford: Oxford University Press. Neufeld, Blain 2005: Civic Respect, Political Liberalism, and Non-Liberal Societies. Politics, Philosophy & Economics 4: 275–99. Neufeld, Blain2010: Reciprocity and Liberal Legitimacy. Journal of Ethics & Social Philosophy: 1–7. Nussbaum, Martha C. 2011: Perfectionist Liberalism and Political Liberalism. Philosophy & Public Afairs 39: 3–45. Quong, Jonathan 2011: Liberalism without Perfection. Oxford: Oxford University Press. Quong, Jonathan 2013: On the Idea of Public Reason. In J. Mandle and D. Reidy (eds.): A Companion to Rawls (pp. 265–80). Oxford: Blackwell. Rawls, John 1980: Kantian Constructivism in Moral Teory. Te Journal of Philosophy 77: 515–72. Rawls, John 1985: Justice as Fairness: Political not Metaphysical. Philosophy & Public Afairs 14: 223–51. Rawls, John 1987: Te Idea of an Overlapping Consensus. Oxford Journal of Legal Studies 7: 1–25. Rawls, John 1993/1996: Political Liberalism. New York: Columbia University Press. Rawls, John 1997: Te Idea of Public Reason Revisited. Te University of Chicago Law Review 64: 765–807. Rawls, John 2001: Justice as Fairness: A Restatement. Cambridge, MA: Harvard University Press. Raz, Joseph 1990: Facing Diversity: Te Case of Epistemic Abstinence. Philosophy & Public Afairs 19: 3–46. Raz, Joseph 1998: Disagreement in Politics. American Journal of Jurisprudence 43: 25–52. Trasher, John and Kevin Vallier 2015: Te Fragility of Consensus: Public Reason, Diversity and Stability. European Journal of Philosophy 23: 933–54. Vallier, Kevin 2014: Liberal Politics and Public Faith: Beyond Separation. London: Routledge. Van Schoelandt, Chad 2015: Justifcation, Coercion, and the Place of Public Reason. Philosophical Studies 172: 1031–50. Wall, Steven 1998: Liberalism, Perfectionism and Restraint. Cambridge: Cambridge University Press. Wall, Steven 2002: Is Public Justifcation Self-Defeating? American Philosophical Quarterly 39: 385–99.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

64

Fabian Wendt

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Wall, Steven 2010: On Justifcatory Liberalism. Politics, Philosophy & Economics 9: 123–49. Wall, Steven 2013a: Political Morality and Constitutional Settlements. Critical Review of International Social and Political Philosophy 16: 481–99. Wall, Steven 2013b: Public Reason and Moral Authoritarianism. Philosophical Quarterly 63: 160–9. Wall, Steven 2014: Perfectionist Justice and Rawlsian Legitimacy. In J. Mandle and D. Reidy (eds.): A Companion to Rawls (pp. 413–29). Oxford: Blackwell. Wall, Steven 2016: Te Pure Teory of Public Justifcation. Social Philosophy and Policy 32: 204–26. Weithman, Paul 2011: Why Political Liberalism? On John Rawls’s Political Turn. Oxford: Oxford University Press. Weithman, Paul 2015: Reply to Professor Klosko. Res Publica 21: 251–64. Wendt, Fabian 2016a: Compromise, Peace and Public Justifcation: Political Morality Beyond Justice. London: Palgrave Macmillan. Wendt, Fabian 2016b: Te Moral Standing of Modus Vivendi Arrangements. Public Afairs Quarterly 30: 351–70. Wendt, Fabian 2017: Compromise and the Value of Widely Accepted Laws. In C. Rostbøll and T. Scavenius (eds.): Compromise and Disagreement in Contemporary Political Teory (pp. 50–62). London: Routledge.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

3

Injustice, Reparation, and Legitimacy

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Stephen Galoob and Stephen Winter

Legitimate states have the right to rule over specifc people and territory. At a minimum, the right to rule is a power to make law and, in so doing, to change the normative situation of those who are ruled. Having the right to rule places a legitimate state in a special relationship with the people it governs and in regard to other states. Te right to rule is not immutable, however. Some injustices can undermine a state’s legitimacy. Te efects of such injustices can be narrow or broad. An injustice might compromise the legitimacy of a state narrowly in relation to those who are victims of state wrongdoing, or more broadly in relation to those who have not been victimized, or both. Moreover, by responding to  wrongdoing that compromises its legitimacy, a state’s acts of redress might bolster its right to rule. Here too the efect can be narrow or broad. Reparation could (in principle) enhance a state’s legitimacy over those who have been the victims of injustice, as well as over those who have not. Every account of political legitimacy should be able to explain how injustice and reparation afect a state’s right to rule. Tis chapter makes two metatheoretical points regarding political legitimacy. Te frst (in sections I–II) is that injustice and reparation can directly afect a state’s legitimacy, so any adequate theory of legitimacy must be able to account for these phenomena. We develop this point through several claims about the specifc efects of injustice and reparation on political legitimacy. Tese efects are intuitively plausible and widely recognized, although they have not previously been presented or analyzed systemically. Te second metatheoretical point (in sections III–V) is that explaining these efects has general implications for theorizing legitimacy. In light of these implications, many prominent theories of legitimacy cannot straightforwardly account for the signifcance of injustice and reparation. We demonstrate this point by analyzing A. John Simmons’s voluntarist theory of political legitimacy. Although Simmons accepts many of the efects we identify, his theory

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

66

Stephen Galoob and Stephen Winter

ultimately fails to explain how injustice can compromise and reparation can bolster a state’s legitimacy. We then describe four strategies for explaining the signifcance of injustice and reparation, two of which seem particularly promising. A concluding section explores some broader implications of our insights for theorizing legitimacy.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

I. Political legitimacy is a complex of rights, including an entity’s right to make law for its subjects and over its territory, its right to enforce these requirements, and its right that other entities not interfere with its lawmaking and enforcement.1 For the sake of exposition, we assume that legitimacy is a property of states and that a state’s political legitimacy correlates with an obligation of those subject to authority to obey the state’s legal dictates.2 A theory of political legitimacy establishes, at a minimum, the conditions under which a state could have this complex right to rule and the contours of that right.3 Teories of legitimacy difer along many dimensions, including the proper grounds of legitimacy—that is, in their conception of the necessary and sufcient conditions for a state to have the right to rule.4 Most theories construe the grounds of legitimacy in terms of history and/or structure. For historical approaches, legitimacy is based on interactions between individuals and states. A state has the right to rule over (only) those with whom it has the relevant interactions. Diferent theories identify diferent historical interactions as relevant, including procedural fairness, consent, and the acceptance of benefts.5 By contrast, structural approaches ground legitimacy on features of a state’s existing political institutions. Structural approaches also difer about which features of states ground the

1  David Copp, “Te Idea of a Legitimate State,” Philosophy and Public Afairs 28 (1999): 3–45, 18. 2  Te correlativity thesis is inessential to our key metatheoretical claims about the efects of injustice and reparation on legitimacy. We assume correlativity in order to mirror the account of A.  John Simmons analyzed in sections  III and  IV. See A.  John Simmons, Justifcation and Legitimacy: Essays on Rights and Obligations (New York: Cambridge University Press, 2001), 137; A. John Simmons, Moral Principles and Political Obligations (Princeton, NJ: Princeton University Press, 1979), 14–15. 3  Bernard Williams, In the Beginning Was the Deed: Realism and Moralism in Political Argument (Princeton, NJ: Princeton University Press, 2005), 5. 4  Arthur Applbaum, “Legitimacy without the Duty to Obey,” Philosophy & Public Afairs 38 (2010): 215–39, 216. 5  H.L.A. Hart, “Are Tere Any Natural Rights?” Te Philosophical Review 64 (1955): 175–91.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Injustice, Reparation, and Legitimacy

67

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

right to rule.6 Hybrid accounts identify various structural and/or historical grounds for legitimacy.7 A state’s wrongdoing can afect its right to rule. Te following example illustrates this point. State Injustice:  Politia is a legitimate state and has been since time immemorial. Te constitution of Politia guarantees each citizen’s right to vote and many other fundamental rights. Te citizens of Politia are wealthy and expect future material improvement. Politia is characterized by a fair distribution of social resources and political processes that are open to participation by all citizens. All Politians are treated as equals under law. Each Politian is, at the age of majority, allowed to choose whether to remain in Politia or (with considerable publicly-provided resources and with rights of reentry into Politia) to emigrate to another state. Politia is not Utopia, but its right to rule can be appreciated on both structural and historical grounds. Politia contains a discrete minority called the Targetarians, who difer in both ethnicity and religion from their fellow citizens. One day, the  political leaders of Politia eschew normal legislative procedures to enact a series of new rules called “Te Policy.” Te Policy disenfranchises Targeterians and expropriates their holdings as part of an efort to forcibly resettle the group. Te Policy also deprives Targetarians of the right to practice their religion and assigns them an inferior legal status, according to which they lack the political and due process rights that other citizens of Politia have. Te Policy treats the Targetarians in ways that will signifcantly diminish the life prospects of group members. However, Te Policy does not restrict the rights of any other citizens of Politia. 6  For example, justice-based theories hold that a state’s legitimacy depends on the justness of its institutions. E.g., John Rawls, “Justice as Fairness: Political Not Metaphysical,” Philosophy & Public Afairs 14 (1985): 223–51. Democratic theories see the state’s right to rule as conditional on its institutions satisfying certain democratic criteria. E.g., Joshua Cohen, “Procedure and Substance in Deliberative Democracy,” in Deliberative Democracy, eds. James Bohman and William Rehg, 407–37 (Cambridge, MA: MIT Press, 1997); Niko Kolodny, “Rule Over None II: Social Equality and the Justifcation of Democracy,” Philosophy and Public Afairs 42 (2014): 287–336. Instrumental theories ground legitimacy on a state’s capacity to achieve certain political goods, like the protection of rights or the promotion of welfare. E.g., Steven Wall, “Democracy, Authority and Publicity,” Journal of Political Philosophy 14 (2006): 85–100. 7  Examples of hybrid theories of legitimacy include Anna Stilz, “Nations, States, and Territory,” Ethics 121 (2011): 572–601, 578–82; Allen Buchanan, “Political Legitimacy and Democracy,” Ethics 112 (2002): 689–719, 703; Henry S. Richardson, Democratic Autonomy: Public Reasoning about the Ends of Policy (New York: Oxford University Press, 2002), chapter 2. Likewise some theories invoke several grounds of legitimacy, but see one ground as more fundamental than the other(s). See, e.g., David Estlund, Democratic Authority: A Philosophical Framework (Princeton, NJ: Princeton University Press, 2009).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

68

Stephen Galoob and Stephen Winter

Indeed, proceeds from the sale of expropriated Targetarian holdings are used to fnance public projects that beneft non-Targetarian citizens of Politia. Te Policy negatively afects Politia’s legitimacy vis-à-vis the Targetarians, no matter what theory of legitimacy one adopts.8 To the extent that Targetarians had a (pro tanto, content-independent) duty to obey legal directives of Politian ofcials before Te Policy, this duty is meaningfully diminished (and perhaps nonexistent) afterwards. Tese observations suggest the following claim. C1:  Certain injustices compromise the political legitimacy of a state vis-à-vis the persons or groups who are the target(s) of injustice.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

C1 is the personal delegitimating efect of injustice. Te case for C1 is more than intuitive. C1 is echoed in the Declaration of Independence’s observation that a “long train of abuses and usurpations” by the British gave American colonists a “right . . . to throw of such Government and to provide new Guards for their future security.”9 C1 also refects arguments made by many other targets of political injustices.10 Several philosophers argue for or assume C1.11 Moreover, the idea that injustice compromises a state’s right to rule over the victims of injustice has a long pedigree in political theory.12 Te personal delegitimating efect of injustice is automatic and direct. In State Injustice, Te Policy is sufcient to explain why Politia’s right to rule 8  Te State Injustice example (like the State Reparation example below) is overdetermined in order to demonstrate that injustice (and reparation) can afect a state’s right to rule on any theory of political legitimacy. We thank an anonymous reviewer for suggesting the need to clarify this point. 9  U.S. Declaration of Independence (1776). Readers who deny our intuition in the State Injustice case are invited to modify the example in order to increase the seriousness of the injustice. (Suppose, for example, that Politia commits genocide against the Targetarians.) So long as some injustice would compromise Politia’s legitimacy, then C1 and C2 are correct. 10  Examples include Simón Bolívar’s “Jamaican Letter” of September 6, 1815, available http://faculty.smu.edu/bakewell/bakewell/texts/jamaica-letter.html, and the Zapatista Declaration of War; see Declaration from General Command of Ejercito Zapatista Liberacion Nacionala, First Declaration of the Lacandon Jungle (January 1, 1994), available https://library.brown.edu/create/modernlatinamerica/chapters/chapter-3mexico/primary-documents-with-accompanying-discussion-questions/document-9frst-declaration-from-the-lacandon-jungle-today-we-say-enough-is-enough-ya-bastaezln-command-1993. 11  E.g., Jef Spinner-Halev, Enduring Injustice (New York: Cambridge University Press, 2012), 134; Tommie Shelby “Justice, Deviance, and the Dark Ghetto,” Philosophy & Public Afairs 35 (2007): 126–60, at 151; Samuel Schefer, “Immigration and the Signifcance of Culture,” Philosophy & Public Afairs (2007): 93–125, 114, n. 20. 12  See the examples in Michael Baylor, “Political Tought in the Age of the Reformation” in George Klosko, Te Oxford Handbook of the History of Political Philosophy (2011).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Injustice, Reparation, and Legitimacy

69

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

over the Targetarians is changed. No other consideration need be invoked. Likewise, there is no intermediate step needed to explain why the Targetarians’ duty to obey might be weaker, and perhaps canceled, in the wake of Te Policy. Teories of legitimacy difer about how, exactly, the personal delegitimating efect occurs. Some theories conclude that all injustices have this personal delegitimating efect.13 Others see only a subset of injustices as having this efect.14 Our analysis is agnostic about which of these conclusions is correct. C1 posits only that some injustices could have the personal delegitimating efect. Further, C1 does not necessitate any particular conclusion about the efect of injustice. For example, a theory might suggest that Politia’s right to rule over the Targetarians “dissolves”15 or that it merely degrades in quality.16 Te essential point is that an injustice can afect the state’s right to rule over those against whom the injustice is perpetrated. C1 does not exhaust the ways that injustice can afect legitimacy. Injustices can compromise a state’s right to rule over citizens who are not themselves victims of injustice. To be sure, some (like Hobbes) deny this proposition.17 But Hobbes’s position is both an outlier and counterintuitive. Appreciating the broader efects of injustice motivated Toreau’s contention that injustices against Mexicans, indigenous peoples, and African Americans compromised the right of the United States to rule over him.18 It also rings true to the experiences of “white” activists during America’s Civil Rights movement and “liberal” opponents of the Apartheid regime in South Africa. Moreover, a variety of philosophers contend that a state’s unjust actions can undermine 13  For example, Robert Nozick’s historical account of legitimacy implies that all injustices (regardless of whether they are perpetrated by the state) compromise the state’s legitimacy in relation to the target of the injustice. Lawrence Davis, “Comments on Nozick’s Entitlement Teory,” Te Journal of Philosophy 73 (1976): 836–44, 839. 14  An example of this position is Rawls’s contention that legitimacy does not require that laws “be just by a strict standard of justice,” so long as they are not “too gravely unjust. At some point, the injustice of the outcomes of a legitimate democratic procedure corrupts its legitimacy, and so will the injustice of the political constitution itself. But before this point is reached, the outcomes of a legitimate procedure are legitimate whatever they are.” John Rawls, Political Liberalism (New York: Columbia University Press, 1991), 428. See also Ronald Dworkin, Justice for Hedgehogs (Cambridge, MA: Harvard University Press, 2011), 321–2; Christopher Kutz, “Justice in Reparations: Te Cost of Memory and the Value of Talk,” Philosophy & Public Afairs 32 (2004): 277–312, 304; Robert Post, “Democracy and Equality,” Te ANNALS of the American Academy of Political and Social Science 603 (2006): 24–36, 33. 15  John Locke, Two Treatises of Government, ed. Peter Laslett (Cambridge: Cambridge University Press, 1963), §221. 16  Phillip Pettit, “Legitimacy and Justice in Republican Perspective,” Current Legal Problems 65 (2012): 59–82, 62, 64. 17  Tomas Hobbes, Leviathan, ed. C.  Macpherson (Harmondsworth: Penguin, 1975), ch. 21. 18  Henry David Toreau, Walden and Other Writings (New York: Modern Library, 2000).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

70

Stephen Galoob and Stephen Winter

its legitimacy vis-à-vis an entire citizenry.19 Finally, fundamental notions of international human rights law presuppose that a state’s unjust policies toward some citizens can undermine its right to rule over other citizens who are not victims.20 Tis more general efect of injustice also operates at the inter-state level. Normally, other states would commit a serious wrong by interfering with the exercise of a legitimate state’s authority. Yet at least some injustices perpetrated or allowed by a state seem to undermine its right against interference in its internal afairs by foreign actors.21 Tis efect is implicit in the (admittedly controversial) international law doctrines of “humanitarian intervention” and “responsibility to protect.” State Injustice provides intuitive support for both of these general efects. Politia’s perpetration of Te Policy undermines its right to rule over all its citizens, not merely the Targetarians. Other states seem to have less powerful reasons (and, perhaps, no reason at all) to avoid interfering with Politia’s exercising its political authority. Te following claims capture these insights.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

C2:  Certain injustices compromise the political legitimacy of a state more generally, vis-à-vis (C2A) citizens who are not the target of the injustice; and/or (C2B) other states. C2 describes the general delegitimating efects of injustice. Like the personal delegitimating efect, these general delegitimating efects are direct and automatic. In the State Injustice case, for example, Politia’s enacting Te Policy compromises its right to rule. No further actions by the Targetarians or others are needed to realize this efect (although, perhaps, further actions might be necessary to authorize other states to intervene in Politia’s internal afairs).

19  For example, Williams, In Te Beginning Was the Deed, at 71–3; Bas van der Vossen, “Te Asymmetry of Legitimacy,” Law and Philosophy 31 (2012): 565–92, 566; Christopher Heath Wellman, “Rights and State Punishment,” Te Journal of Philosophy 106 (2009): 419–39, 436; Gary Bass, “Jus Post Bellum,” Philosophy and Public Afairs 32 (2004): 384–412, 398–9. 20  Evan Fox-Decent and Evan Criddle, “Te Fiduciary Constitution of Human Rights,” Legal Teory 15 (2009): 301–36, 320; John Tasioulas, “Human Rights, Legitimacy, and International Law,” Te American Journal of Jurisprudence 58 (2013): 1–25, 11. 21  Andrew Altman and Christopher Heath Wellman, A Liberal Teory of International Justice (New York: Oxford University Press, 2009), 148; Joseph Raz “Human Rights without Foundations,” in Te Philosophy of International Law, eds. Samantha Besson and John Tasiolas, 321–38, 328 (Oxford: Oxford University Press, 2010); Fox-Decent and Criddle, “Te Fiduciary Constitution of Human Rights,” at 347–8; Allen Buchanan, “Reciprocal Legitimation: Reframing the Problem of International Legitimacy,” Politics, Philosophy & Economics 10 (2011): 5–19, 13; Van der Vossen, “Te Asymmetry of Legitimacy,” at 568.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Injustice, Reparation, and Legitimacy

71

Of course, just as with C1, conceptions of legitimacy vary in explaining C2 and in identifying which injustices trigger these efects. In addition, theories might difer about whether C1 operates in the same way as C2, as well as whether C2A and C2B operate identically. Te basic insight refected in C1 and C2, however, is that a state’s perpetration of injustice can undermine its right to rule over both those who are the victims of injustice and those who are not. II.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

C1 and C2 indicate that political legitimacy is a dynamic, rather than static, property. A state’s legitimacy depends, at least in part, on what it does. Tis dynamism works in the opposite direction as well. Just as legitimate states can lose their legitimacy, it seems possible for states to acquire or bolster legitimacy.22 One way that a state can do so is by rectifying injustices or providing reparation. We defne reparation as an act or policy that responds to a wrong or injustice and ameliorates the signifcance of that wrong or injustice. Tere are diferent conceptions of reparation’s ameliorative efect.23 Opinions also difer about the goals of reparation—for example, efectuating a return to the status quo ante,24 returning the victim(s) to the point of counterfactual indiference,25 or providing what is needed for the victim(s) to stand as equals with others.26 22  Copp, “Te Idea of a Legitimate State,” at 31. See also Jeremy Waldron, “Superseding Historic Injustice,” Ethics 103 (1992): 4–28, 14; Stephen Winter, Transitional Justice in Established Democracies: A Political Teory (London: Palgrave Macmillan, 2014), ch. 5; Christopher F. Zurn, “Te Logic of Legitimacy: Bootstrapping Paradoxes of Constitutional Democracy,” Legal Teory 16 (2010): 191–227, 224. 23  For example, corrective justice approaches portray reparation as satisfying the wrongdoer’s remedial duties, which arise out of the wrongdoing. Jules Coleman, “Te Practice of Corrective Justice,” Arizona Law Review 37 (1995): 15–31, 30. Restorative justice approaches describe reparation as enabling survivors and perpetrators to share community. Margaret Urban Walker, Moral Repair (New York: Cambridge University Press, 2006), 107–8. Retributive approaches see reparation as a way to punish wrongdoers. David Hershenov, “Restitution and Revenge,” Te Journal of Philosophy 96 (1999): 79–94. Equitable approaches emphasize the importance of disgorging unjustly derived benefts. Hanoch Dagan, “Restitution and Slavery: On Incomplete Commodifcation, Intergenerational Justice, and Legal Transitions,” Boston University Law Review 84 (2004): 1139–76. 24  Robert E. Goodin, “Teories of Compensation,” Oxford Journal of Legal Studies 9 (1989): 56–75. 25  Christopher Morris, “Existential Limits to the Rectifcation of Past Wrongs,” American Philosophical Quarterly 21 (1984): 175–82, 176–8. 26  Tomas Pogge, “Historical Wrongs: Te Two Other Domains,” in Justice in Time: Responding to Historical Injustice, ed. Lucas Meyer, 117–34 (Baden-Baden: Nomos Verlagsgesellschaft, 2004).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

72

Stephen Galoob and Stephen Winter

We do not here argue for any particular understanding of reparation. Instead, our contention is that every account of political legitimacy requires a theory of reparation. To be sure, a theory of reparation need not follow from any particular conception of legitimacy. Two theories of legitimacy might converge about the grounds of a state’s right to rule but diverge in explaining how reparation could bolster this right. To see how reparation might bolster political legitimacy, consider the following sequel to the State Injustice scenario. State Reparation:  A few years after Te Policy, the leaders of Politia come to realize it as an injustice. Based on this realization, Politia rescinds Te Policy and enacts a series of measures called the “Reparation Policy.” Reparation Policy compensates Targetarians for the harms they have sufered and provides a public apology for (and a memorialization of ) Te Policy’s unjustness through a truth commission, museum exhibits, and other public memorials. Reparation Policy also modifes Politia’s political institutions to safeguard human rights, establish legal institutions that more credibly guarantee the rule of law, and provide the Targetarians with greater self-determination. We think that Reparation Policy would have the efect of bolstering Politia’s right to rule over the Targetarians.27 Tis intuition supports the following claim.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

C3:  Certain acts of reparation bolster the political legitimacy of a state vis-à-vis the persons or groups who were the target of injustice. C3 is the personal legitimating efect of reparation. Tis efect has been posited by a number of scholars28 and is implicit in the arguments of redress participants.29 As with the delegitimating efects of injustices, C3 is 27  Some readers might resist this intuition. Because the wrongs perpetuated in Te Policy were so egregious, the thinking might go, the gestures comprising Reparation Policy are inadequate to bolster Politia’s right to rule. We encourage readers who have this intuition to amend the scenario in order to make the wrongfulness of Te Policy less severe and the gestures in Reparation Policy even more rigorous. If Politia’s right to rule would be bolstered in this amended case, then C3 and C4 are true. 28  See, e.g., Pablo De Greif, “Justice and Reparations,” in Te Handbook of Reparations, ed. Pablo De Greif, 451–77 (New York: Oxford University Press, 2006); Mihaela Mihai, “Transitional Justice and the Quest for Democracy: A Contribution to a Political Teory of Democratic Transformations,” Ratio Juris 23 (2010): 183–204, 198; Winter, Transitional Justice in Established Democracies, at 226; Lea Ypi, “What’s Wrong with Colonialism,” Philosophy & Public Afairs 41 (2013): 158–91, 187. 29  For example, C3 is a plausible interpretation of New Zealand’s Waitangi Tribunal’s statement that reparation can end the Crown–Māori relationship of grievance and inaugurate a new partnership in lawmaking. Waitangi Tribunal, Ko Aotearoa Tēnei: Report on the Wai 262 Claim (2011). https://www.waitangitribunal.govt.nz/news/ko-aotearoatenei-report-on-the-wai-262-claim-released.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Injustice, Reparation, and Legitimacy

73

both  automatic and direct. In the State Reparation case, the enactment of  Reparation Policy improves Politia’s legitimacy. No further act or acknowledgment is needed for this change to occur (although, to be sure, reparation might also have indirect efects on Politia’s legitimacy). If injustices can have personal and general delegitimating efects and reparation can have a personal legitimating efect, then reparation might also bolster legitimacy more generally.30 In the State Reparation case, Politia’s right to rule over people other than the Targetarians seems to be bolstered by the enactment of the Reparation Policy. Te same conclusion also applies to Politia’s claim against interference by other states. Tese efects are represented in the following claims, which we call the general legitimating efects of reparation.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

C4:  Certain acts of reparation can bolster the political legitimacy of a state more generally, vis-à-vis (C4A) citizens who were not the target of injustice; and/or (C4B) other states. Tese efects have been postulated by other writers,31 although explanations for them difer across theories of legitimacy. Such theories also likely diverge about which reparative acts realize C3 and C4. Nevertheless, if any redressive gestures could have the efects described in C3 and C4, then an adequate theory of political legitimacy should be able to explain this link between reparation and the right to rule. Call C1–C4 collectively “the efects.” Each efect seems to be part of our normative world. Each is intuitively plausible and embraced from diverse perspectives, including the perspectives of those who have been targets of injustice and those who agitate for redress. However, our case for the efects has been indicative, not defnitive. A theory of political legitimacy should either be able to describe and explain each of the efects of injustice and reparation, or else ofer a principled argument why (despite appearances to the contrary) one or more of these efects does not exist.

30  It is not strictly paradoxical to accept C1–C3 but deny C4. A theory of legitimacy could coherently deny C4 if diferent considerations explain C1 and C2 or C1 and C3. However, if the same considerations explain C1, C2, and C3, then C4 follows from the conjunction of C1–C3. 31  De Greif, “Justice and Reparations,” 464–6; Lucy Allais, “Restorative Justice, Retributive Justice, and the South African Truth and Reconciliation Commission,” Philosophy & Public Afairs 39 (2011): 331–63, 359–60; Martha Minow, Between Vengeance and Forgiveness (Boston, MA: Beacon Press, 1999), 88; Richard Black and Saskia Gent, “Sustainable Return in Post-Confict Contexts,” International Migration 44 (2006): 15–38, 17; Seana Shifrin, “Reparations for U.S. Slavery and Justice Over Time,” in Harming Future Persons, eds. David Wasserman and Melinda Roberts, 333–9 (Dordrecht: Springer, 2009).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

74

Stephen Galoob and Stephen Winter

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

III. Yet a satisfactory explanation of the efects is elusive. In the next two sections, we demonstrate this point by examining A.  John Simmons’s voluntarist theory of legitimacy. Simmons is an appropriate foil not only because he is a preeminent and infuential contemporary theorist of legitimacy, but also because he recognizes the efects and criticizes other theories for failing to account for them. For example, Simmons echoes C1 in contending that profound injustices must “at the very least limit the obligations that wrongdoing states can unilaterally impose on those they have wronged.”32 Simmons notes, consistent with C2A, that certain “unjust policies and laws” can “deprive [the state] of the moral authority to impose on its citizens any obligation of obedience or support.”33 Simmons also sees reparation as integral to political legitimacy, arguing that the “rectifcation[] of past wrongs” is an ingredient of a state’s “ma[king] itself fully legitimate.”34 Tis statement is consistent with C3, that the reparation of such injustices can bolster the state’s legitimacy vis-à-vis victims of injustice. Te logic of C4A is implicit in Simmons’s contention that “real states are always morally required to be  honestly working toward full rectifcation” of the injustices they have perpetrated, regardless of whether the victims of injustice are still citizens.35 Finally, Simmons embraces C4B in contending that all contemporary states must “make sincere eforts to do their parts in moving us toward a fully rightful world,” including “rectifying the wrongs done by our states in their bloody-handed transformations into their contemporary forms.”36 It is fair, then, to expect Simmons’s theory of legitimacy to account for the efects. However, central and infuential elements of Simmons’s theory are incompatible with them. Before arguing for this conclusion in section IV, we describe three core tenets of Simmons’s theory of legitimacy that are relevant to our analysis.

32  A. John Simmons, “Democratic Authority and the Boundary Problem,” Ratio Juris 26 (2013): 326–57, 341–2. For Simmons, so-called “Kantian” theories of legitimacy (such as those advanced by John Rawls, Tomas Nagel, and Ronald Dworkin), which base a state’s legitimacy on the justness of its political structures, are faulty because they cannot explain this efect of profound injustices on legitimacy. Ibid., at 342. 33  A. John Simmons, “Disobedience and Its Objects,” Boston University Law Review 90 (2010): 1805–31, 1811. See also Simmons, “Democratic Authority and the Boundary Problem,” 342; “Voluntarism and Political Associations,” Virginia Law Review 67 (1981): 19–37, 31. 34  A.  John Simmons, Boundaries of Authority (New York: Oxford University Press, 2016), 247. Tis link between reparation and legitimacy is a feature of what Simmons calls “Lockean nonideal theory.” Ibid., 210. 35  Simmons, Boundaries of Authority, 246. 36  Ibid., 249.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Injustice, Reparation, and Legitimacy

75

First, Simmons sees a state’s legitimacy as independent of its justifcation. For Simmons, a state’s justifcation turns on considerations about whether there are reasons for citizens to accept its authority. For example, states might facilitate wealth production, respect human rights, enable political participation, or promote equality. Each of these characteristics is relevant to assessing whether, or to what degree, a state is justifed—that is, whether its rule is “on balance morally permissible (or ideal) and . . . rationally preferable to all feasible nonstate alternatives.”37 Yet, Simmons argues, a state’s legitimacy does not depend on whether it exhibits these characteristics. A state’s “legitimacy (or illegitimacy) with respect to me and its possession of other good (or bad) characteristics are . . . simply independent variables, in the same way that a business’s general good (or bad) qualities are independent of its right to have me as a client.”38 For Simmons, the view that justifcation is independent of legitimacy stands in contrast to “Kantian” approaches, which base conclusions about a state’s legitimacy directly on considerations related to its justifcation.39 A second tenet of Simmons’s theory is its solution to what he terms the “particularity requirement.”40 For Simmons, an adequate account of political legitimacy must show why a specifc state has the right to rule over identifable subjects and within a defnite territory. To fulfll the particularity requirement, a theory must be able to explain why a given citizen is subject to the authority of a specifc state “above all others.”41 In virtue of the particularity requirement, Simmons construes legitimacy as both relational and bipolar. Legitimacy is relational in that it describes a “specifc relationship between individual and institution.”42 It is bipolar in that those who are not part of the citizen–state dyad have no bearing on whether the state has legitimate authority over a specifc citizen.43 A state’s unjust actions toward others “need not afect the voluntariness of one’s consent to it. Supposing only that the unjust institution does not happen to be doing violence to me, I can freely consent to its authority.”44 37  A. John Simmons, Justifcation and Legitimacy, at 126. 38  A.  John Simmons, “Consent Teory for Libertarians,” Social Philosophy and Policy 22 (2005): 330–56, 354 (emphasis added). Simmons has recently described his position as “pluralist,” in that it sees both historical and structural considerations as relevant to determining political legitimacy. Boundaries of Authority, 73. Such a “pluralist” interpretation of Simmons’s account is in tension with seeing justifcation and legitimacy as independent variables. 39  Simmons, Justifcation and Legitimacy, at 145–55. 40  Simmons, Moral Principles and Political Obligations, at 31. 41  A. John Simmons, “Te Particularity Problem,” APA Newsletter on Philosophy and Law 7 (2007): 18–27, 19. 42  Simmons, Justifcation and Legitimacy, 155. 43  Simmons, “Consent Teory for Libertarians,” 356. 44  Simmons, Moral Principles and Political Obligations, 78. See also “Consent Teory for Libertarians,” 351.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

76

Stephen Galoob and Stephen Winter

According to Simmons, only voluntarist theories can capture the independence of justifcation from legitimacy while also satisfying the particularity requirement.45 Terefore, a third core tenet of Simmons’s theory is a voluntarist conception of legitimacy on which “genuine and nondefective consent is both necessary and sufcient for . . . legitimacy.”46 For Simmons, each person has a natural right to rule herself, and a person can transfer this right to a political entity (paradigmatically) through consenting to its rule.47 To be sure, many other theories posit consent as necessary to ground a state’s political legitimacy.48 Simmons’s version goes further in construing a citizen’s consent as both necessary and sufcient to establish a state’s right to rule. As such, Simmons contends that it is possible for a profoundly unjust state to wield legitimate authority over a person, so long as that person has genuinely consented to the state’s authority.49 IV.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Tese three core tenets (that a state’s justifcation is independent of its legitimacy; the political relationship is bipolar; and a citizen’s genuine consent is both necessary and sufcient to ground a state’s right to rule over her) render Simmons’s voluntarist theory unable to account for the efects of 45  E.g., Simmons, “Te Particularity Problem,” 24. 46  Simmons, “Consent Teory for Libertarians,” at 351; see also A. John Simmons, “Consent, Free Choice, and Democratic Government,” Georgia Law Review 18 (1983): 791–819, 792. To expand, Simmons sees consent as genuine only when it is intelligent and voluntary: the person must intend to consent to the state’s rule and know what she is consenting to. Simmons, “Consent, Free Choice, and Democratic Government,” 797–9. For Simmons, satisfying these genuineness conditions is necessary for consent to constitute “freedom freely surrendered.” Simmons, Boundaries of Authority, 21. 47  Simmons, Moral Principles and Political Obligations, 62–3. 48  E.g., Amanda Greene, “Consent and Political Legitimacy,” in Oxford Studies in Political Philosophy 2, eds. David Sobel, Peter Vallentyne, and Steven Wall, 71–97, 81–2 (Oxford: Oxford University Press, 2015). 49  Simmons recognizes that a “genuinely illiberal” state (for example, one that provided for “an absolute monarch or dictator (who is free to prey upon the masses), [or allowed] for political participation for only a privileged few, for draconian censorship and invasion of privacy, or for a police state”; one that “require[d] regular gladiatorial games, provide[d] that members of disliked racial or ethnic groups be denied citizenship or even entry into the society . . . and guarantee[d] that no societal resources whatsoever be used to improve circumstances outside of the society (except where such neglect would violate outsiders’ rights, as opposed to their needs or wants)”; or one that “provide[d] for human slavery, allowing slaves and masters to be selected by any means one can imagine”) could have the right to rule over any subject who has “freely chosen the coercion to which [he is] subject.” Simmons, “Consent Teory for Libertarians,” 350–2, 355. Te only limitation Simmons envisions is that a legitimate state cannot in principle perpetrate “violations of the rights of nonconsenting outsiders.” Ibid., at 351.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Injustice, Reparation, and Legitimacy

77

injustice and reparation on political legitimacy. Tis inadequacy applies even if one grants, for the sake of argument, that all three tenets are true. In this section, we lay out the main reasons why Simmons-type voluntarism cannot explain the efects, while in section V we explore some ways that alternative theories of legitimacy might explain them. Simmons’s theory clearly captures C1, the personal delegitimating efect of injustice. Te State Injustice case shows how. Suppose that, prior to enactment of Te Policy, the Targetarians had genuinely consented to being ruled under the Politian constitution. If the enactment of Te Policy were outside the ambit of what the Targetarians consented to, then Te Policy would violate the terms of their consent and thereby undermine Politia’s right to rule over the Targetarians. For Simmons, a state’s ultra vires actions are by defnition outside of “the terms of consent given”50 and therefore undermine its legitimacy. However, Simmons’s theory cannot capture the general delegitimating efects described by C2A and C2B. In the State Injustice case, Te Policy afects citizens who are not Targetarians by automatically impairing Politia’s right to rule over them. Yet Simmons-type voluntarism cannot straightforwardly explain this efect. Perhaps Te Policy could afect non-Targetarians in virtue of its being ultra vires—if the terms of consent for non-Targetarians included a prohibition covering events like Te Policy. (We consider this possibility further in section V.) Suppose, for the sake of argument, that no non-Targetarian insisted on protections against the abuse of Targetarians as a condition of Politia’s right to rule. In other words, construe the terms of the political relationship as wholly bipolar, such that no citizen’s consent specifcally concerns anyone other than himself or herself. Given this supposition, Simmons’s voluntarist theory faces two options. Te frst is to deny that Te Policy automatically undermines Politia’s legitimacy vis-à-vis non-Targetarians, contravening C2A. Te second option is to accept that something other than a citizen’s volitional exercise (namely, a state’s enactment of an unjust policy that victimizes others) is relevant to whether the state has the right to rule over her. Yet the latter option would deny that political legitimacy is determined solely by factors arising from the bipolar relationship between a citizen and the state. In other words, accounting for C2A requires construing political relationships as multipolar. Similarly, Te Policy compromises Politia’s external legitimacy in a way that (consistent with C2B) might license a form of intervention by other states. Yet Simmons’s theory cannot capture this efect either. Doing so would imply that considerations other than the volitional commitments of 50  Simmons, Justifcation and Legitimacy, 129.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

78

Stephen Galoob and Stephen Winter

a citizen determine, at least in part, which state (if any) has the right to rule over him or her. Simmons’s theory of legitimacy therefore cannot explain how Te Policy directly changes the normative situation of non-Targetarians in the ways that C2B contemplates. Nor can Simmons’s theory capture C3, the personal legitimating efect of reparation. Tis efect is inconsistent with the notion that a citizen’s volitional commitments are the sole determinant of a state’s legitimacy. Accounting for C3 requires allowing that reparative actions by a wrongdoer (e.g., the state of Politia in the State Reparation case) can directly change the normative status of citizens (e.g., the Targetarians). If so, then considerations other than a citizen’s consent determine whether the state has the right to rule over him or her. Yet to accept this possibility is to deny that a citizen’s genuine consent is both necessary and sufcient to determine a state’s political legitimacy over him or her. For similar reasons, Simmons’s theory is also incompatible with the general legitimating efects of reparation described in C4A and C4B. Both of these efects imply that reparative actions can change a state’s right to rule over those who are not the targets of reparation. Yet on the bipolar construal of the political relationship that Simmons ofers, a state’s treatment of others can never automatically determine whether that state has the right to rule over someone. Tus, both clauses of C4 confict with Simmons-type voluntarism for the same reason that C3 does: if a state’s reparative gesture can directly bolster its legitimacy, then someone’s consent is insufcient to determine the state’s right to rule over him or her. To summarize, Simmons’s theory of legitimacy can account for C1, the personal delegitimating efect of injustice. However, his theory cannot account for the other efects of injustice and reparation. Yet an adequate theory of political legitimacy should be able to account for C2–C4. Terefore, core tenets of Simmons’s voluntarist view render it descriptively inadequate. More broadly, any theory that adopts one or more of these core tenets risks descriptive inadequacy. V. We now consider four ways that legitimacy theories might account for the efects of injustice and reparation. We formulate these possibilities as gambits, or moves that a theory might deploy. None of these gambits is compatible with Simmons-type voluntarism, since each would require abandoning some or all of the core tenets described above. However, other theories of legitimacy (voluntarist or not) might utilize these gambits. We frst describe each gambit, then analyze some of its broader implications for theorizing legitimacy.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Injustice, Reparation, and Legitimacy

79

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Te “Other-Regarding” Gambit As suggested in section IV, one move is to construe any volitional exercise sufcient to ground political legitimacy as automatically (even if implicitly) referencing the treatment of others. On this gambit, legitimizing consent to a state’s authority must include standards for how the state may treat others. If a state violates these terms by mistreating others, then it undermines its right to rule over those who are not subject to mistreatment. Tis “otherregarding” gambit would capture C2, the general delegitimating efect of injustice. To illustrate this gambit in the context of State Injustice, suppose that the terms of any citizen’s consent to Politian rule implicitly included a condition requiring the rightful treatment of all fellow citizens, including the Targetarians. If so, then implementing Te Policy not only violates the terms of consent by Targetarians, but also violates the terms of consent for non-Targetarians. Tis gambit ofers a unifed explanation of C1 and C2: both types of delegitimation efect are explained by Politia’s ultra vires action against the Targetarians. Simmons endorses something like this gambit in arguing that consent is nullifed if it is “in important ways defective—for example, if it is nonvoluntary, unintentional, uninformed, or given to arrangements that violate others’ rights.”51 In violating the rights of Targetarians, the argument goes, Te Policy runs afoul of the consent by other Politians. Despite Simmons’s endorsement, the “other-regarding” gambit is broadly incompatible with core tenets of his theory. As noted above, capturing C2 requires attributing mandatory terms to a citizen’s consent. Tis attribution would violate Simmons’s contention that consent be “freedom freely surrendered”52 because it would deny that each citizen has the power to set the terms on which he or she consents to a state’s rule. Capturing C3 and C4 would require attributing even more baroque mandatory terms, for example provisos pertaining to rights violations and the reestablishment of consent in response to reparation. Requiring such terms as part of genuine consent would make political legitimacy subject to a series of conditionals (e.g., if the state acts wrongly, then citizen consent is violated; if the state’s wrongdoing is followed by reparation, then the terms of consent are not violated). In efect, the “other-regarding” gambit would construe both structural factors (such as the justness of a state’s institutions) and historical factors (such as the treatment of citizens and remedying of injustices) as relevant to a state’s legitimacy. Tis blending of structural and historical factors conficts with Simmons’s contention that a state’s legitimacy can be 51  Simmons, “Consent Teory for Libertarians,” 341. 52  Simmons, Boundaries of Authority, 21.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

80

Stephen Galoob and Stephen Winter

determined based solely on historical considerations and the core tenet of his theory that a citizen’s consent is both necessary and sufcient to determine a state’s legitimacy.53 Furthermore, the “other-regarding” gambit denies the bipolar construal of the political relationship. On the “other-regarding” gambit, a state must respect the rights of some in order to have the right to rule over others. To adopt this requirement is to see the political relationship as multipolar, with every political relationship necessarily referencing and implicating a myriad of other agents,54 including non-citizens.55 On the “other-regarding” gambit, then, the state’s right to rule over you is conditioned on how it treats others. More broadly, this multipolar construal of the political relationship places the “other-regarding” gambit in tension with any voluntarist theory of legitimacy. Te imposition of these mandatory, other-regarding terms blurs the line between voluntarist and non-voluntarist approaches, including “Kantian” views in which a state’s legitimacy is a function of its justifcation. Moreover, if the political relationship is multipolar, then the citizen’s consent to reasonable other-regarding conditions is implicit. Yet if assent to these terms can be inferred, then it is an open question why assent to other reasonable terms (including the just state’s right to rule in the frst place) cannot also be inferred. While “Kantian” theories of legitimacy allow for such inferences,56 voluntarist theories characteristically rule them out. Further, invoking such mandatory, other-regarding terms invites Joseph Raz’s charge that consent is unnecessary as a ground for legitimacy—that it “cannot be used as a way of endowing anyone [or entity] with authority where that person [or entity] had none.”57 In sum, the “other-regarding” gambit provides a prima facie explanation of the efects. However, it entails a multipolar construal of the political relationship, one that weakens the distinction between justifcation and legitimacy. Te move is difcult to reconcile with any theory of legitimacy 53  E.g., Moral Principles and Political Obligations, 15; “Consent Teory for Libertarians,” 340. 54  Tis multipolar construal of the political relationship is captured by Michael Tompson’s discussion of the “manifold of persons” implicated under a particular normative order. See Michael Tompson, “What Is It to Wrong Someone? A Puzzle about Justice,” in Reason and Value: Temes from the Moral Philosophy of Joseph Raz, eds. R. Jay Wallace, Philip Pettit, Michael Smith, and Samuel Schefer, 333–84 (Oxford: Oxford University Press, 2004), 354. 55  Aaron Maltais, “Political Obligations in a Sea of Tyranny and Crushing Poverty,” Legal Teory 20 (2014): 186–209, 205. 56  E.g., Tomas Nagel, Equality and Partiality (New York: Oxford University Press, 1991), 36; Arthur Ripstein, Force and Freedom: Kant’s Legal and Political Philosophy (Cambridge, MA: Harvard University Press, 2009), 184–5. 57  Joseph Raz, Te Morality of Freedom (Oxford: Oxford University Press, 1986), 90.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Injustice, Reparation, and Legitimacy

81

that draws a conceptual distinction between a state’s justifcation and its legitimacy.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Te “Evidential” Gambit A second move focuses on the indirect ways that an injustice can compromise a state’s legitimacy. Above, we described the efects as direct and automatic: a state’s right to rule is undermined or strengthened solely in virtue of its wrongdoing or reparation. Yet the efects might instead be modeled in a more circuitous way. A state that perpetrates or permits injustice demonstrates that it is likely to commit or permit further injustices.58 On the “evidential” gambit, an injustice would compromise a state’s legitimacy based on what it reveals about what the state is likely to do in the future, rather than in virtue of the wrongdoing itself.59 On this gambit, the State Injustice case might be analyzed as follows. Te Policy wrongs the Targetarians directly, thereby compromising Politia’s right to rule over that group. In addition, Te Policy demonstrates to other citizens and states that Politia is disposed to act ultra vires, and is therefore likely to commit future ultra vires actions. Tat disposition would, in turn, explain why Te Policy compromises Politia’s legitimacy in relation to non-Targetarian citizens (under C2A) and other states (under C2B), especially if a prohibition on Politia being likely to act ultra vires is among the terms of Politians’ consent. In addition to explaining the delegitimating efects of C1 and C2, this “evidential” gambit also could explain C3 and C4. Just as the commission of an injustice provides evidence relevant to assessing whether a state has the right to rule, the provision of reparation demonstrates the probability that a state will act justly in the future. For example, in the State Reparation scenario, Reparation Policy could bolster Politia’s legitimacy by demonstrating that Politia will ameliorate the negative efects of its prior injustices and, therefore, will likely avoid similar injustices going forward. Tis “evidential” gambit is incompatible with Simmons-type voluntarism. First, this move denies that a citizen’s volitional commitments fully determine a state’s political legitimacy. Rather, factors other than a citizen’s consent and a state’s intra vires actions—namely, the likelihood of the state’s protecting rights and promoting justice in the future—drive conclusions about the state’s right to rule. Second, the “evidential” gambit violates the independence

58  E.g., Walker, Moral Repair, 30–4. 59  E.g. Pablo De Greif, “Teorizing Transitional Justice,” in Transitional Justice, eds. Jon Elster, Rosemary Nagy, and Melissa Williams, 31–77 (New York: New York University Press, 2012), 47.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

82

Stephen Galoob and Stephen Winter

of justifcation and legitimacy by grounding legitimacy on a non-historical feature—namely, the state’s prospective likelihood of protecting rights. Te “evidential” gambit faces independent objections. One problem is that the “evidential” gambit provides a revisionist explanation of the efects. At its core, the “evidential” gambit denies that injustice and reparation directly or necessarily bear on a state’s political legitimacy. Rather, injustices and reparation only matter derivatively, in terms of a state’s future actions. Views that attribute solely derivative signifcance to a state’s injustice and reparation face independent objections.60 Admittedly, such a forward-looking perspective might be relevant to fully appraising the impact of injustice and reparation.61 Yet to contend that the signifcance of an injustice should be appraised only in light of what it reveals about the probability of future injustices is, in efect, to concede that those future injustices have nonderivative signifcance. It would be strange if future injustices had direct signifcance for legitimacy, but present injustices had only indirect signifcance. In any event, the presuppositions of the “evidentiary” gambit stand in need of a principled defense.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Te “Permeability” Gambit A third move relaxes the distinction between a state’s justifcation and its legitimacy. On this “permeability” gambit, a state’s lack of justifcation can afect its legitimacy. In other words, a state’s unjust policies might be sufcient to support the conclusion that it lacks the right to rule, even if its justifed policies are insufcient to establish that it has the right to rule. On this gambit, justifcation and legitimacy are connected, but the connection between them is epiphenomenal. Simmons attributes something like this move to Locke, who stipulates that a state must be minimally justifed in order to be capable of being legitimated through consent. For Locke, someone’s consent to a state’s rule that is not at least as good (on prudential grounds) as the conditions of the state of nature cannot bind him.62 Te “permeability” gambit therefore sees a state’s justifcation as conditioning its legitimacy: only states that meet a sufcient level of justifcation can be the receptors of genuine consent. Te “permeability” gambit also resembles Rawls’s contention that legitimacy and justice operate along the same dimension of assessment, such that a state cannot be legitimate if its procedures and laws are not “too gravely unjust.”63 60  Walker, Moral Repair, 66–70. 61  Shifrin, “Reparations for U.S. Slavery and Justice Over Time,” at 335. 62  Locke, Two Treatises of Government, §§ 221 and 225; see also Simmons, Justifcation and Legitimacy, 129 n. 18. 63  Rawls, Political Liberalism, 428.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Injustice, Reparation, and Legitimacy

83

Te “permeability” gambit provides a straightforward explanation of C1 and C2. In the State Injustice case, Te Policy undermines Politia’s right to rule over the Targetarians in virtue of its being ultra vires (C1). Te Policy might undermine Politia’s legitimacy more generally by placing the state below a minimal justice threshold (C2). Once Politia crosses the threshold, the consent of non-Targetarians no longer binds them to Politia’s rule. Te “permeability” gambit might also explain the legitimating efects of reparation (C3 and C4). If the Reparation Policy moves Politia above the minimal justifcation threshold, then it removes potential defeaters to Politia’s legitimacy (although these changes need not, by themselves, restore Politia’s right to rule over anyone). Te “permeability” gambit is incompatible with the core tenets of Simmons’s theory. If injustice compromises legitimacy, then a state’s perpetration of injustice against third parties can directly afect its right to rule over any citizen. Terefore, the “permeability” gambit posits the generic political relationship as multipolar, abandoning Simmons’s bipolar construal. Further, the “permeability” gambit also denies that a subject’s volitional exercise is sufcient to establish a state’s political legitimacy regarding her. C3 and C4 envision that reparation can, by itself, bolster a state’s right to rule over citizens who have been the victims of injustice and those who have not themselves been victimized. If these efects are real, then at least one factor unrelated to a subject’s volitional exercise is relevant to determining whether the state has the right to rule over that subject. Finally, the gambit requires political legitimacy to depend upon a state’s justifcation. A broader consequence of the “permeability” gambit concerns its use of a threshold of justness to evaluate political legitimacy. As we envision this gambit, only injustices that push a state below the threshold trigger C2, and only reparation that pulls a state above the threshold triggers C4. Yet these contentions raise signifcant, unanswered questions. To employ the “permeability” gambit, one must explain why injustices that move a state’s level of justifcation below the threshold bear on its legitimacy, but injustices that do not push the state below the threshold (or injustices committed by a state already below the threshold) do not. Similar points would apply to reparative acts that do not cross the threshold of justifcation. In other words, defending the “permeability” gambit requires an independent explanation of those thresholds. Tis explanation will be complicated if, for example, diferent thresholds apply to legitimacy over a specifc person (C1), general domestic legitimacy (C2A), and international legitimacy (C2B). Additionally, the “permeability” gambit construes legitimacy as a discontinuous variable.64 Above the threshold, a state has the right to rule; 64  E.g., Frank Michelman, “Legitimacy and Autonomy: Values of the Speaking States,” Brooklyn Law Review 79 (2014): 985–1004, 994; Pettit, “Legitimacy and Justice in Republican Perspective,” 64.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

84

Stephen Galoob and Stephen Winter

below the threshold, it does not. For Locke, this construal is perhaps explained by the ambition to identify the grounds of allegiance to a single, unitary government. Teorists who have a diferent ambition (such as explaining the phenomenon of political authority in complex and pluralistic states) might favor construing legitimacy as a scalar variable. Yet the “permeability” gambit is inconsistent with construing legitimacy as scalar—that is, with seeing structural and/or historical considerations as strengthening or weakening a state’s right to rule.65 We do not mean to resolve the conceptual debate about whether legitimacy is binary or scalar. Rather, the point is that any view that construes legitimacy as scalar will have difculty adopting the “permeability” gambit.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Te “Right of Resistance” Gambit A fourth move invokes a Lockean “right of resistance.”66 Recall that, on Simmons’s view, each person has a natural right to rule herself and govern­ ment operates as a trust for the protection and promotion of this natural right. Some injustices violate this trust and activate a “right of resistance” among citizens.67 For Locke, the exercise of the right of resistance dissolves the state’s right to rule.68 Te right of resistance is held individually by each person in a society, as well as in common by society as a whole. When an individual invokes the right of resistance, the state’s legitimacy is dissolved vis-à-vis her. Similarly, when a citizenry invokes its collective right of resistance, it dissolves the state’s right to rule over them. Te “right of resistance” gambit has two conditions. First, the state must systematically violate rights. Second, society (perhaps represented by the majority) must invoke its collective right to resistance. Te state’s breach of trust is not sufcient to dissolve its legitimacy. Rather, this breach licenses the invocation of the collective right of resistance. To use legal terminology, an available but unexercised collective right of resistance is an encumbrance that renders the state’s right to rule liable to revocation. As such, any specifc injustice does not dissolve the state’s right to rule, since the power to efectuate this change is held by the citizenry. Something like the “right of resistance” gambit is invoked in arguments for a remedial right of secession in the wake of injustice and oppression.69 65  E.g., Zurn, “Te Logic of Legitimacy,” 201–2. 66  A.  John Simmons, On the Edge of Anarchy (Princeton, NJ: Princeton University Press, 1993), 149. John Locke, Two Treatises of Government, §222. 67  Simmons, On the Edge of Anarchy, 158. 68  Ibid., 163. 69  Peter Radan, “Secessionist Referenda in International and Domestic Law,” Nationalism and Ethnic Politics 18 (2012): 8–21; Allen Buchanan, Justice, Legitimacy, and Self-Determination: Moral Foundations for International Law (New York: Oxford University Press, 2003), 331. Tanks to Ten-Herng Lai for this suggestion.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Injustice, Reparation, and Legitimacy

85

Te “right of resistance” gambit can explain both the personal and general delegitimating efects of injustice. In the State Injustice case, Te Policy licenses individual Targetarians to invoke their respective rights of resistance. Te activation of these rights explains the personal delegitimating efect of C1. Te Policy also licenses invocation of the collective right of resistance held by Politia’s non-Targetarian citizens. Te licensing of this collective right would not, by itself, dissolve Politia’s government. Rather, it would empower the majority of Politians to dissolve the government. Tis power would render Politia’s right to rule “voidable,” rather than “void.”70 Such an encumbrance is a general delegitimating efect of the kind that C2A describes. Te “right of resistance” gambit provides a similar explanation for C2B. Te Policy would not, by itself, make it permissible for foreign agents to interfere in the internal afairs of Politia. However, Te Policy would license further steps (such as authorization from international political institutions) that, if carried out, would enable foreign agents to void or disregard Politia’s right to rule. Regarding C3 and C4, the “right of resistance” gambit sees reparation as removing potential defeaters to a state’s right to rule. In our example, the Reparation Policy both targets the objectionable efects of Te Policy and guards against their recurrence. Terefore, after the Reparation Policy, Te Policy would no longer ground a right of resistance for either Targetarians or Politians at large. Nor would Te Policy license foreign parties to intervene in Politia’s internal afairs. Te “right of resistance” gambit is incompatible with Simmons’s theory. First, this gambit sees a close connection between a state’s justifcation and its legitimacy. Te same factor (namely, whether the government has acted in accordance with the purposes for which it was entrusted) would determine both whether the state is justifed and whether it has the right to rule. Tis implication conficts with construing justifcation and legitimacy as independent variables. Second, the “right of resistance” gambit adopts a multipolar understanding of the political relationship: a state that commits an injustice against any citizen compromises its right to rule over victims and non-victims alike. Tis understanding conficts with the bipolar construal of the political relationship. Tird, the “right of resistance” gambit denies that a citizen’s volitional exercises are sufcient to determine a state’s political legitimacy, since the state’s right to rule depends on factors unrelated to a citizen’s specifc volitional exercises. More broadly, the “right of resistance” gambit, like the “evidentiary” gambit, sees injustice and reparation as afecting political legitimacy indirectly. According to the “right of resistance” approach, a state’s legitimacy in the 70  Corbin on Contracts §1.6 (2015).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

86

Stephen Galoob and Stephen Winter

wake of an injustice is a function of whether the people have exercised their right of resistance. Injustice and reparation matter only insofar as they trigger or invalidate the exercise of this right. Yet C1–C4 seem to bear directly and automatically on political legitimacy. Either the “right of resistance” gambit cannot capture the directness of these efects, or else it requires supplementation by an argument showing why injustice and reparation do not directly afect a state’s right to rule. Furthermore, the “right of resistance” gambit has conceptual implications that make it incompatible with a scalar notion of legitimacy. A state’s right to rule is either turned of (if citizens exercise their justifed right to revolt) or on (if they have not). Although we take no position on the conceptual question of whether legitimacy is a discontinuous or scalar variable, any view that construes legitimacy as scalar will have difculty incorporating the “right of resistance” gambit.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

* * * In this section, we identifed several possible ways that a theory of legitimacy might account for the efects of injustice and reparation. None of these gambits can improve the descriptive adequacy of Simmons’s version of voluntarism, since each of them jettisons core tenets of Simmons’s theory. Teorists of legitimacy who do not embrace all of the tenets of Simmons-type voluntarism might be attracted to one or more of these gambits. Te “otherregarding” and “permeability” gambits strike us as most promising, although both strategies contain important gaps that require further development. Both of these strategies have more general ramifcations for theorizing legitimacy, as they imply that political relationships are multipolar and that a state’s legitimacy depends, at least in part, on its justifcation. By contrast, the “evidentiary” and “right of resistance” gambits strike us as less promising, since both deny that injustice and reparation directly afect a state’s right to rule. CONCLUSION Injustice and reparation are central to understanding political legitimacy. We identifed four efects of injustice and reparation on legitimacy and argued that an adequate theory of political legitimacy should be able to explain these efects. Tis is our frst metatheoretical point. Our second metatheoretical point is that explaining these efects has general implications for theorizing legitimacy. To support this point, we showed that the core tenets of Simmons’s voluntarist theory preclude an explanation of all of the efects. To the extent that other voluntarist theories adopt one or more of these tenets, they also risk explanatory inadequacy.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Injustice, Reparation, and Legitimacy

87

Tese explanatory difculties are not unique to voluntarist theories of legitimacy. Non-voluntarist theories of legitimacy cannot easily account for the signifcance of injustice and reparation. For example, theories that derive political legitimacy solely from the justness of a state’s political institutions face challenges accounting for the delegitimating efects of injustices. Tis is less a problem of explanatory adequacy than a problem of internal consistency. Teories like Rawls’s “liberal principle of legitimacy” make a structuralist move by assessing a state’s legitimacy based solely on an appraisal of the structure of its political institutions and the outcomes they produce. On Rawls’s liberal principle, for example, these assessments are questions of justice.71 As Simmons points out, this structuralist move leads to an asymmetry in the treatment of past and future injustices: the former are irrelevant to assessing a state’s justness and its right to rule, while the latter are highly relevant. Yet it seems “self-evidently false . . . that wrongs that occurred yesterday are in principle morally diferent from the same kinds of wrongs occurring tomorrow.”72 In other words, a version of the liberal principle of legitimacy that incorporates the structuralist move can explain C1 and C2 only by accepting that present injustices are relevant to legitimacy. However, this identifcation conficts with the structuralist move, which classifes all historical interactions (including both past injustices and past acts of reparation) as irrelevant to legitimacy. Likewise, some democratic theories of legitimacy (in particular, “pure” accounts that see democracy as a sufcient ground for political legitimacy)73 face difculty accounting for C3 and C4, the legitimating efects of reparation. Te challenge here mirrors our critique of Simmons’s view in section IV: if reparation can directly bolster a state’s right to rule, then the democratic character of a state’s political institutions is insufcient to appraise that state’s political legitimacy. Terefore, it seems unlikely that an adequate theory of political legitimacy could be derived solely from a theory of democratic practice. Furthermore, hybrid theories of legitimacy (which utilize a variety of structural and/or historical factors to explain a state’s right to rule) might face internal difculties explaining the legitimating efects of both injustice and reparation. Explaining these efects seems to require some way of making tradeofs across diferent inputs to a theory. By way of illustration, consider Amanda Greene’s “sovereignty” theory, on which a state’s legitimacy is based on its achieving “actual quality consent to rule based on positive 71  Political Liberalism, 428. 72  Simmons, Boundaries of Authority, at 55–6. 73  See, e.g., Jeremy Waldron, Law and Disagreement (Oxford: Oxford University Press, 1999), ch. 13; Fabienne Peter, “Democratic Legitimacy and Proceduralist Social Epistemology,” Politics, Philosophy, & Economics 6 (2007): 329–53.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

88

Stephen Galoob and Stephen Winter

governance assessments,” which are “judgment[s] by a particular subject of some sort of governance success” by the state.74 In the State Reparation case, Greene’s theory might explain the legitimating efects of Reparation Policy in terms of the evidence it provides for a positive governance assessment of Politia. On this logic, Reparation Policy bears directly on “quality consent,” but it does not afect “actual consent” (for the reasons described in section IV in connection with Simmons’s view). Yet seeing these two aspects of Greene’s sovereignty theory as having independent value conficts with Greene’s contention that they have only dependent value: for Greene, consent to rule makes no contribution to legitimacy unless it is based on a positive governance assessment, and (presumably) vice versa.75 In order to account for C3 and C4, then, a hybrid view would need to clarify whether the relationship between its component elements is dependent or independent. If a hybrid view does not allow for tradeofs across components,76 then it would fail to attribute signifcance to injustices and/or reparations that afect some components, but not others. Our argument therefore suggests that a hybrid account of legitimacy would need to enable tradeofs across components in order to explain a case like State Reparation.77 Our preliminary discussion here does not establish that any of these theories is explanatorily inadequate. Nor do we mean to convey pessimism about the possibility of explaining the efects.78 Rather, these brief observations provide further support for our second metatheoretical point: certain commitments might restrict a legitimacy theory’s power to explain the signifcance of injustice and reparation. Our analysis has several broader implications for understanding political legitimacy. First, because the efects invoke moral phenomena like wrongdoing and reparation as part of explaining political legitimacy, any theory that draws a sharp distinction between a state’s justifcation and its legitimacy is likely to be explanatorily inadequate. If the efects are real, then justifcation and legitimacy are not independent properties. At a minimum, they are mutually dependent, and perhaps they are even more intimately intertwined. A second implication is that the best account of political legitimacy will probably have both historical and structural aspects. Te efects imply that 74  Consent and Political Legitimacy, at 81, 83–4. 75  Ibid., at 86–7. 76  Examples of this position include a hybrid theory that assigns lexical priority to one component, or one that deems its components to be incommensurable. 77  See, for example, the “balancing” view of legitimacy proposed by Zofa Stemplowska and Adam Swift: “Dethroning Democratic Legitimacy,” in Oxford Studies in Political Philosophy 4, eds. David Sobel, Peter Vallentyne, and Steven Wall, 3–27 (Oxford: Oxford University Press, 2018). 78  Indeed, one of us has articulated an account of political legitimacy that aims to capture each of the efects. See Winter, Transitional Justice in Established Democracies, at chapters 3 and 5.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Injustice, Reparation, and Legitimacy

89

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

legitimacy is a dynamic property. If so, then it is unlikely that any set of political arrangements could fully determine a state’s legitimacy. Facts about wrongdoing and reparation must be capable of informing conclusions about legitimacy. On the other hand, purely historical accounts also seem inadequate. Assessments of political legitimacy require appraising the structural features of the state, even if only to account for the general delegitimating efects of injustice and the general legitimating efects of reparation. Tird, our analysis suggests the need for a broader understanding of “particularity”—that is, the delimitation of a state’s right to rule to specifc subjects and over defnite territory. Simmons construes the political relationship as a dyad between the citizen and the state. However, fully appreciating the efects of injustice and reparation on legitimacy requires a broader focus, an acknowledgment that a state’s right to rule over someone or someplace can be directly afected by its treatment of others. Compared to a bipolar understanding of political legitimacy, a multifaceted understanding is better suited to capture the efects of injustice and legitimacy. A fourth implication is that any complete theory of legitimacy must include some account of how the rectifcation of injustice could afect a state’s political legitimacy—that is, a theory of reparation. Such a theory of reparation is necessary to explain how a state’s providing redress to some people can afect its right to rule over them and others. Yet most theorizing about reparation is isolated from core debates in political philosophy. Likewise, few accounts of political legitimacy devote serious attention to theorizing reparation, although all extant states are founded on histories of  injustice. Understanding how reparation works is therefore vital to appreciating how states might be legitimate, and vice versa.79

79  Tanks to Evan Fox-Decent, Bill English, Jon Gartof, Keith Hankins, John Hasnas, Geof Kemp, Ten-Herng Lai, Ethan Leib, Loren Lomasky, Arie Rosen, Alex Sarch, Katherine Smits, Kevin Vallier, Daniel Viehof, Kevin Walton, Martin Wilkinson, and audiences at the Georgetown Institute for Study of Markets and Ethics, Oklahoma Junior Scholars Conference, the Workshop for Oxford Studies in Political Philosophy, the Legal Philosophy Roundtable in Wellington, NZ, the New Zealand Political Studies Association, and the European Consortium for Political Research for feedback. Te editorial team of Steve Wall, Peter Vallentyne, and David Sobel has been remarkable.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

4

Justifying Uncivil Disobedience

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Ten-Herng Lai

A prominent way of justifying civil disobedience is to start from the idea that there is a pro tanto duty to obey the law and to argue that the considerations that ground this duty sometimes justify forms of civil disobedience.1 Te basic idea is that we have a duty to obey the law that derives from certain important substantive or procedural values that are secured by compliance. However, even societies like ours, which seem reasonably just such that this pro tanto duty to obey the law holds, are likely to fall short of the substantive or procedural values that they aim to deliver. And while it will often be feasible to address these shortcomings within the bounds of the legal system, sometimes the only efective way of doing so will involve engaging in certain types of illegal political activities. Disobeying the law may be the best way of realizing the substantive or procedural values that underpin the duty to obey the law. Under these circumstances, disobedience may be justifed.2 Proponents of this view take for granted that, in societies that are reasonably just, only civil forms of disobedience are capable of being justifed in this way. According to the Bedau (1961) defnition that gained signifcant infuence through John Rawls’s A Teory of Justice (1999), civil disobedience consists of “a public, nonviolent, conscientious yet political act contrary to law usually done with the aim of bringing about a change in the law or policies of the government” (p. 320). While certain aspects of this defnition can be (and have been) challenged (e.g. Morreall, 1976; Moraro, 2007; Brownlee, 2012; Celikates, 2014, just to name a few), the key idea is that disobedience is 1  Some argue that there is a moral right to engage in civil disobedience (e.g. Brownlee, 2012; Dworkin, 1978; Lefkowitz, 2007). Tese accounts are beyond the scope of this chapter. However, I am inclined to believe that my argument can be extended to those accounts: the grounds of the right to civil disobedience would most likely extend to support certain (though perhaps a more limited range of ) acts of uncivil disobedience. 2  I avoid taking a stand on whether diferent forms of disobedience are justifed as merely permissible or obligatory. Tis is an important issue, but beyond of the scope of this chapter.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Justifying Uncivil Disobedience

91

justifed only when it meets some requisite standard of civility, however construed. Activities that fail to meet this type of threshold can never be justifed. Let us call this the Orthodox View.3 Te Orthodox View comprises two theses:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Positive Tesis:  It is justifed to engage in civil disobedience within a reasonably just society insofar as and because doing so constitutes responding correctly to the considerations that ground the pro tanto duty to obey the law. Negative Tesis:  It is never justifed to engage in uncivil disobedience within a reasonably just society. Te aim of this chapter is to argue that there is something unstable about the Orthodox View. Engaging in certain kinds of uncivil disobedience— disobedience that fails to meet one or more of Rawls’s criteria or civility, however broadly construed by critics of Rawls—within a reasonably just society also sometimes constitutes responding correctly to the considerations that ground the pro tanto duty to obey the law.4 So, if it is sometimes justifed to engage in civil disobedience within a reasonably just society because doing so constitutes responding correctly to the considerations that ground the pro tanto duty to obey the law, then it is also sometimes justifed to engage in uncivil disobedience within a reasonably just society, all else being equal. If the Positive Tesis is true, then the Negative Tesis is false. In short, we should conclude either a) that civil disobedience is never justifed within a reasonably just society or b) that uncivil disobedience is sometimes justifed within a reasonably just society. Tus, civil disobedience doesn’t pick out a category that enjoys a special normative status over uncivil activities. 3  Proponents of the orthodox view include, just to name a few, Hugo A. Bedau, Martin Luther King Jr., John Rawls, Andrew Sabl, Daniel Markovits, and William Smith. In addition, activists often endorse the orthodox view. Te Occupy Central movement in Hong Kong, for example, adhered strictly to nonviolence, and as soon as the movement ended, the leaders submitted themselves to arrest. Ordinary citizens also tend to condemn activities that fall short of being civil. For example, some dismiss the possibility that the Ferguson Unrest in the U.S. can be justifed because violence was involved. 4  Te term “uncivil disobedience” has been used by Jennet Kirkpatrick (2008) to refer to violent political activities such as the interracial abolitionists mob in 1854 that attempted to rescue the fugitive slave and Baptist preacher Anthony Burns from being returned to slavery under the Fugitive Slave Act of 1850, where one guard was killed during the incident. She holds that this type of violent activity shouldn’t be dismissed prematurely as “antidemocratic.” Tat being said, she does warn against how commendable motivations can lead to condemnable and indiscriminate violence. My use of the term “uncivil disobedience” is broader: any illegal political act that fails the standards of being civil, however expansively civility may be defned. Accordingly, political activities that involve little or no violence may also be instances of uncivil disobedience.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

92

Ten-Herng Lai

Tis chapter is in four sections. In section 1, I examine a number of diferent versions of the Positive Tesis, starting from Rawls as a key reference point, but also incorporating a wide range of non-Rawlsian accounts. In section 2, I argue that, if the Positive Tesis is true, then this implies that uncivil disobedience is also sometimes justifed and, hence, that the Negative Tesis is false. In section 3, I consider and respond to the objection that my argument ignores the fact that civil disobedience enjoys a special normative status over other illegal dissents on account of instantiating certain special features: nonviolence, acceptance of legal consequences, publicity, and conscientious­ ness, and thus avoid involving actions that are wrong in themselves. In section 4, I argue that my view is distinct from and superior to two rivals: the view that we should expand the notion of civility such that civil disobedience, expansively construed, is uniquely appropriate; and the view that uncivil disobedience is appropriate in but only in unfavorable conditions. It should be emphasized that throughout the chapter, I do not take a stand on whether or why there is a general duty to obey the law. Instead, I simply work with the assumptions of diferent accounts of civil disobedience, and show that their justifcation of civil disobedience vis-à-vis the duty to obey naturally extends to justify uncivil disobedience.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

1.  JUSTIFYING CIVIL DISOBEDIENCE Te Orthodox View begins from the presumption that, in reasonably just societies, there is a pro tanto duty to obey the law (e.g. Rawls, 1999; Smith, 2011). Tis duty to obey is supposed to extend to laws and policies that are nontrivially fawed. (Otherwise there would be no point in engaging in civil disobedience to fx anything, or there would be no need to justify civil disobedience against a duty to obey that doesn’t exist.) But why think that there is any duty to obey laws that are far from perfect—say, laws that impose systematic disadvantages on certain minorities? Consider, frst, the duty to obey perfectly just laws. Tere are, broadly, two ways of grounding this duty: what I shall call substantive and procedural accounts.5 Substantive accounts appeal to certain substantive values that are realized through obedience: e.g. responding better to reasons (Raz, 1986),6 promoting justice through supporting just institutions (Rawls, 1999), or 5  Tis dichotomy is employed in order not to over-complicate the discussion. It should, however, be pointed out that those who endorse a procedural account still give a certain weight to substantive considerations, and vice versa. 6  Te basic idea of Raz’s account of legitimate authority is that if, by following the directives of some agent or agency, we would better conform to the reasons which apply to us in a given domain than through reliance upon our own judgment, that agent or agency has practical authority over us.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Justifying Uncivil Disobedience

93

fulflling our “duty to rescue” in a fair way (Wellman, 2005).7 Procedural accounts, in contrast, appeal to the non-instrumental quality of the (usually democratic) procedures in which political decisions are made. For example, the binding force of democratic decisions might be grounded on the fact that laws are constraints that we impose upon ourselves (Post, 1993; Markovits, 2005), or decisions made when all relevant points of views are properly considered (Habermas, 1996; Smith, 2011).8 How might we extend these accounts to less than perfect laws?9 Take substantive accounts frst. Joseph Raz (1986) holds that one ought to obey if but only if submitting one’s judgment to imperfect law responds to reason better than relying on one’s own judgment. Rawls (1999) holds that the natural duty toward justice demands that we support less than perfectly just institutions “as long as they do not exceed certain limits of injustice” (p. 311). Tat’s because there is no guarantee that we can establish perfectly just institutions, while establishing and supporting institutions represents our best chance of approximating a just system. As for procedural accounts, it is often argued that we ought to respect others through respecting democratically made decisions, even when the outcomes fall short of being perfect. At the same time, it seems clear that when the quality of the law deviates from the ideal too much, obeying no longer contributes to realizing the substantive or procedural values that ground the duty to obey. Obeying in such cases is pointless if not detrimental. Tis brings us to civil disobedience. In such circumstances, the duty to obey is undermined. As a last resort to address serious faws within the system, civil disobedience might very well represent the most appropriate way of responding to the substantive or procedural considerations that ground the duty to obey the law. Wellman (2005) for instance, states that if a law is unjust and if disobeying it in a particular way is efective in a certain circumstance to promote justice, one 7  Wellman (2005) grounds the duty to obey the law in what he labels as “the Samaritan duty”: we ought to rescue others from great peril provided that such rescue is not too costly for us. He holds that the state of nature threatens everyone, and can only be prevented by establishing a state to maintain order. Te burden of supporting the state, if fairly distributed through each obeying a fair law, is comparatively small. Given that the cost for each individual is small, whereas the beneft of rescuing everyone is great, each individual has a duty to obey the law. 8  It is important to point out that this is clearly not an exhaustive enumeration of the substantive and procedural accounts of the duty to obey. However, I am confdent that my analysis can be easily applied to other versions of the substantive or procedural accounts. 9  Some simply admit that once the law falls short of being perfect, the duty to obey no longer applies. Wellman (2005), for example, points out that when a law is unjust, it either fails to play any positive role in fulflling the duty to rescue and might even play a part in perpetuating injustice, or it places unfair burdens on particular individuals. Either way, the duty to rescue is not fairly distributed. Since cashing out that duty fairly is exactly what grounds the duty to obey, there is no duty to obey.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

94

Ten-Herng Lai

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

is “morally at liberty to break the particular laws . . . [one] disobeyed simply because they were unjust” (p. 86). For example, in a system that embodies racial discrimination, some might disobey the law publicly to draw attention to the oppression certain groups face and thus enable change. Because such illegal activities help to rescue others from immediate and persisting harm that arises from unjust systems, in contrast to obeying unjust laws and perpetuating injustice, they would actually be fulflling the duty to rescue. Rawls holds that civil disobedience is permissible when dealing with clear and blatant injustice such as racial discrimination, and employed as a device to promote justice. Tose who endorse procedural accounts appeal to the quality of the relevant procedures, and hold that democratic decisions are not binding if certain groups or individuals were excluded from the decisionmaking procedures, and that civil disobedience is justifed when disobeying actually enables or promotes inclusion (Markovits, 2005; Smith, 2011): say, when the voices of certain groups are continuously ignored by the majority, and can only be brought to the table through illegal and disruptive activities; or when there are bureaucratic obstacles that can only be overcome by imposing costs on the system, e.g. inconvenience and public embarrassment, and forcing negotiation to take place (Fung, 2005).10 In each case, then, the idea is that civil disobedience is justifed vis-à-vis the substantive or procedural considerations that ground the pro tanto duty to obey the law. Obeying imperfect laws or policies may be pointless if not detrimental with regard to realizing these considerations; and civil disobedience may better realize these considerations. In such cases, there is no duty to obey the law. Rather, civil disobedience is justifed.11 2.  UNCIVIL DISOBEDIENCE I shall now argue that if this way of justifying civil disobedience succeeds, then it also extends to uncivil disobedience—activities that are somewhat 10  Note that civil disobedience doesn’t necessarily breach the laws it purports to change. Sometimes breaching otherwise perfectly unproblematic laws represents the best way to respond to the underlying values that normally demand obedience. Civil disobedience that breaches the corresponding laws is labeled “direct” civil disobedience, and that which breaches other laws “indirect civil disobedience.” Te justifcation of direct and indirect civil disobedience is pretty much identical, so I do not emphasize the distinction in this chapter. 11  I have focused on unjust laws or policies. However, the justifcation of civil disobedience extends naturally to disobeying otherwise just laws or policies: sometimes disobeying them, e.g. for the sake of signaling the injustice, also serves the underlying values better than obeying those laws. Te basic idea is the same: if obeying otherwise just laws doesn’t realize the underlying political values, but disobeying them does, those values would demand disobedience.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Justifying Uncivil Disobedience

95

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

similar to civil disobedience, but that fall short of the standards of civility by failing to be public, refusing to accept punishment, or showing no respect or loyalty to the system, etc., and most notably involving violence. Paradigm examples of uncivil disobedience include whistleblowing,12 hacktivism, ecotage, and the use or threat of violence in protests.13 Whistleblowing, or more precisely governmental whistleblowing, defned by Candice Delmas (2014b), “involves the unauthorized acquisition . . . and disclosure . . . of classifed information about the state or government” (p. 78) regarding “suspected illegal or unethical conduct . . . [such disclosure] amounts to an indictment of the wrongdoing” (p. 80). Hacktivism is activism using hacking techniques, “with the intent of disrupting normal operations but not causing serious damage” (Denning, 2001, p. 241). Ecotage is “sabotage of inanimate objects (machinery, buildings, fences) that contribute to ecological destruction” (Vanderheiden, 2005, p. 427). Te use or threat of violence in protests includes politically motivated vandalism, resisting arrest, threats to escalating to more radical measures, and the actual damaging of property or harming of persons.14 Note that the discussion of civil disobedience in section 1 was confned to cases where civil disobedience was a last resort to address serious faws within the system. A more refned reading of this is that, to be justifed, civil disobedience must be necessary to address the targeted issue efectively, and the costs or harm imposed by civil disobedience must be proportionate to the severity of injustice. When it comes to uncivil disobedience, it might be 12  I follow Candice Delmas (2014b) in holding that whistleblowing falls short of civility, as it refuses to accept punishment and involves covert planning or even anonymity; but I acknowledge that some might fnd whistleblowing, at least highly idealized versions of it, compatible with the norms of civil disobedience. However, the main argument of my chapter doesn’t rest on whether whistleblowing is civil, and I am confdent that those who insist that whistleblowing is civil would agree that other activities, especially those that involve violence, are uncivil, and this would sufce for the argument in this chapter to work. 13  Tis list is clearly non-exhaustive, but should serve the purpose of discussion. I will omit cases such as secret tax evasion to dissociate with unjust governments or draft dodging when facing unjust wars, which are traditionally labeled as “conscientious refusals” or “conscientious objections.” Tere seem to be well-developed and widely accepted accounts regarding such activities. 14  Here I use “violence” in a broader sense, as Bedau (1961) did when he discussed the nonviolent condition of civil disobedience, which includes “deliberately destroying property, endangering life and limb, inciting to riot (e.g., sabotage, assassination, street fghting)” (p. 656). I am well aware that there is a fair amount of equivocation that occurs when this term is used, and it is an inexcusable (and arguably often a malicious) mistake to equivocate these diferent senses when engaging in normative arguments. In order to avoid this mistake, I will single out violence in a narrower sense—against persons—and have in-depth discussion thereon immediately after discussing violence against property in section 2.1. My argument, therefore, can also appeal to those who view only violence in a narrower sense as genuine violence.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

96

Ten-Herng Lai

thought that uncivil activities are likely to be much costlier than civil activities. Tus, uncivil disobedience, even if necessary to realize signifcant value(s) efectively, may never be proportionate, and thereby never permissible. However, the claim that uncivil disobedience is likely to be much costlier than civil disobedience is simply false. Certain forms of uncivil disobedience impose much lower costs on persons or society in general in comparison to civil disobedience. Whistleblowing, for example, often causes at most public embarrassment to the government, but that can hardly be counted as being harmful in any meaningful sense. Politically motivated vandalism indeed destroys public property, but the costs it imposes are far outmatched by instances of civil disobedience that aim at causing large-scale inconvenience, e.g. occupying and thus paralyzing transportation hubs or even airport runways (where the latter can cause worldwide chaos in the airspace). In addition, even if certain forms of uncivil disobedience impose more costs on persons or the society, this doesn’t prevent uncivil disobedience from being proportionate. First of all, whether an act is proportionate depends, at least in part, on how signifcant the values realized are. In cases where, say, ecotage helps to prevent massive environmental damages that threaten the ecosystem or people’s health, or, say, political violence deters severe racialized policing or racial violence, the values realized or protected are substantial, and thus speak in favor of those acts being proportionate. Second, and perhaps more importantly, it matters to proportionality exactly who bears the costs of particular actions. Regarding the ethics of self-defense and war, it is widely accepted that when it comes to the distribution of costs, it is better for those who are more responsible for the wrong being addressed to bear the costs in addressing that wrong; moreover, proportionality allows more costs to be distributed to the culpable—those who freely and knowingly engage in the wrongful activities that necessitate defensive actions—in comparison to their innocent counterparts (Bazargan, 2014; Draper, 2016; Montague, 2010; Tadros, 2011, 2012; Vallentyne, 2011, 2016). Applying this to uncivil disobedience, even if such activities sometimes cause much more overall harm, insofar as the harm is directed at those who are culpable, acts of uncivil disobedience can still be proportionate. Tis can be true of a variety of instances of uncivil disobedience, ranging from ecotage targeting industries that poses severe threats to the environment, to political violence that responds to racial violence, or a wide range of possible disruptive actions that can be taken against those who proft from selling frearms and/or from opposing gun control. In contrast, acts of uncivil disobedience that impose indiscriminate costs on others are more likely to be ruled out by proportionality. (Taking this into consideration, one might even suspect that acts of uncivil disobedience that target the culpable are much more preferable in comparison to acts of civil

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Justifying Uncivil Disobedience

97

disobedience that impose costs on people indiscriminately, as proportionality would more likely rule that the costs be imposed on the culpable.)15 Having established the possibility that uncivil activities might be propor­ tionate even if they might be thought to be much costlier, I will now spell out the implications of the substantive and procedural versions of the Orthodox View for the justifability of uncivil disobedience.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

2.1.  Substantive Justifcations First consider substantive accounts. Tey hold that under special circumstances engaging in civil disobedience, rather than obeying the law, constitutes the best way of realizing the substantive values that ground the pro tanto duty to obey the law. In the following, I shall show that this can and should be extended to uncivil disobedience. Tat is, under other special circumstances, it is the case that engaging in uncivil disobedience, rather than obeying the law or engaging in civil disobedience, constitutes the best way of realizing the substantive values that ground the pro tanto duty to obey the law, namely the value of responding to reasons better, fulflling our duty to rescue, or acting upon our general duty to promote justice. Where this is so, it seems clear that we should conclude that uncivil disobedience is indeed justifed. As we saw above, paradigm types of uncivil disobedience include whistle­ blowing, hacktivism, ecotage, and the use of violence in protests. Tese types of activities involve breaking laws prohibiting the leakage of classifed information, laws governing cyberspace, property laws or laws specifcally introduced to target “eco-terrorism,” or laws prohibiting the use of violence against property, persons, or the police. Under many “normal” circumstances, obeying these laws will presumably contribute to the realization of the desirable substantive ends. But there is no reason to think that this will always be so.

15  One can also understand this issue in terms of the distinction between narrow and wide proportionality introduced by Jef McMahan (2009). Accordingly, wide proportionality is about defensive harm imposed on those who are not liable, while narrow proportionality concerns those who potentially are. Te weight of the harm with regard to narrow proportionality is typically discounted in virtue of the target’s liability; for the target has forfeited her rights against being harmed, and thus the harm doesn’t wrong her (in contrast to being merely permissible but nevertheless wronging the target). Tis permits more harm to be imposed on the liable in comparison to their innocent counterparts, all other things being equal. Te worry that uncivil disobedience is costlier and thus never proportionate might in part be a confusion that all the costs must be treated as rightsinfringements, weighted equally in proportionality calculations. Rather, in at least some cases the extra costs are imposed on individuals who are liable to bear them, and so do not count against the wide proportionality of the disobedience.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

98

Ten-Herng Lai

Consider governmental whistleblowing. To reiterate, governmental whistle­ blowing is “the unauthorized  acquisition . . . and disclosure . . . of classifed information about the state or government” (Delmas, 2014b, p. 78) that involves state injustice. Obeying the law and concealing such secrets might eliminate the possibility of correcting, punishing, or preventing severe wrongdoings. Te cases of Chelsea Manning16 and Edward Snowden17 fall into this category. In such cases, obeying the law would seem to be positively detrimental to the substantive values underpinning the law, while whistle­ blowing would seem to contribute most efectively to the realization of these values. Correcting and punishing seriously unjust activities seems to amount to responding to reason better and to promoting justice, while preventing severe wrongdoings would be demanded by our duty to rescue, especially when the wrongdoings endanger people’s lives. Whistleblowing, therefore, under special but realistic circumstances, would seem to be justifable vis-à-vis the duty to obey by being demanded by the substantive sources of our political obligations. Hacktivist techniques include paralyzing websites through virtual sit-ins, denial-of-service (DoS), or distributed-denial-of-service (DDoS) attacks, site defacements, site redirections, and information theft (Hampson, 2012). Tere are laws prohibiting most if not all of these techniques in many countries. Again, obeying these laws would amount to giving up a wide range of useful strategies in engaging in protests and forfeiting numerous opportunities to promoting justice. Site defacements and site redirections can raise awareness of certain issues or even provide (counter-) information against problematic organizations and their websites.18 Information theft can also be a form of whistleblowing.19 Tough hacking techniques can be used for selfsh purposes and cause severe harm to society, or sometimes represents nothing more than a simple display of the hacker’s expertise, if used properly, they can help promote the substantive values that otherwise demand obedience better 16  Manning leaked a large amount of classifed information regarding the U.S.’s military conduct, including the slaughter of non-combatants. See, for example, Rizzo, J. (2012). Bradley Manning charged. CNN. http://security.blogs.cnn.com/2012/02/23/ bradley-manning-charged. 17  Snowden revealed information regarding NSA’s mass-surveillance programs, including PRISM. See, for example, Andrews, S., Burrough, B., & Ellison, S. (2014). Te Snowden saga: A shadowland of secrets and light. Vanity Fair. http://www.vanityfair.com/news/ politics/2014/05/edward-snowden-politics-interview. 18  For example, in 1999 Anonymous redirected the “trafc intended for a KKK Web site to Hatewatch” (Himma, 2007, p. 88). 19  For example, in 2010 the hacktivist Andrew “weev” Auernheimer exposed AT&T’s security breach (Delmas, 2018). He was later sentenced to forty-one months. See https:// www.wired.com/2013/03/att-hacker-gets-3-years.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Justifying Uncivil Disobedience

99

than obeying the law. Tus, hacktivism can be justifed vis-à-vis the duty to obey the law. What about the use or threat of violence against either property or persons? Te thought that resorting to violence can be in line with the grounds of our duty to obey the law might seem counterintuitive. Te private use of violence seems to be one of the most dangerous threats to social order. It often incites retaliation, which leads to more violence, thereby undermining the secure and stable society that our way of life depends on. It might also be an indication that one has arrogated to oneself the license to behave in ways prohibited to others. It is not a surprise that many diferent accounts of our duty to obey the law insist that we do our part in establishing and maintaining a civil society that prohibits the private use of violence (e.g. Locke, 2014; Rawls, 1999; Wellman, 2005). Te use or threat of violence, some might conclude, is the exact opposite of what the grounds of our duty to obey the law demands. It is simply not true, however, that under no circumstances is violence in line with the grounds of political obligation. At least in certain extreme cases, i.e. cases where resorting to violence is necessary and proportionate to bring about certain goods, the substantive considerations might demand resorting to violence instead of obeying laws that prohibit violence. Frist, consider violence against properties and objects. Consider laws banning “ecoterrorism.” Such laws prohibit direct actions against industries that cause serious harm to the environment. Environmental hazards, however, sometimes pose signifcant risks to people’s health. In specifc circumstances, successful ecotage could help prevent such risks. Tis might be the outcome of the attention ecotage draws, which translates into public pressure, the economic costs that destroy the proft of the industries in question, or simply through direct prevention or hindrance of environmental hazards. In such cases, it is legitimate to ask what the substantive considerations demand. If one endorses the duty to rescue as the grounds of obedience, then since such laws prohibit rescue, one would have to admit that ecotage rather than obedience is demanded. Regarding the duty to promote justice through supporting just institutions, in such cases it could hardly be said that institutions are just with respect to environmental protection. Since obedience in this particular case does not support just institutions, while ecotage efectively promotes justice, the latter is demanded by such a duty. In addition, it is simply not the case that obeying laws that demand inaction with regard to severe environmental hazards that threaten people’s lives would amount to responding better to reason. In terms of responding to reason, then, there is no duty to obey such laws. Ecotage, if it will genuinely save people’s lives, constitutes the most appropriate response.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

100

Ten-Herng Lai

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Now consider politically motivated vandalism, another type of violence against properties and objects. Te British Sufragettes, for example, engaged in a certain amount of violent activity including smashing windows, blowing up mailboxes, and sabotaging telephone lines. Tey chose to do so because nothing less disruptive seemed to work. Such disruption, however, is relatively minor in comparison to their disenfranchisement and the disadvantages they sufered from disenfranchisement. Or consider the vandalism of political symbols in Taiwan. Statues and monuments of the former despot Chiang Kai-shek were established long before the democratization of the regime. Countless people were persecuted and thousands tortured and murdered during the thirty-eight years of the White Terror he initiated.20 Tis so-called “beacon of freedom” personally altered the sentences of a number of innocent people to execution.21 Te display of political symbols in honor of this man, for example praising this dictator as the “savior of mankind,” is extremely and unjustly ofensive to the victims.22 Based on such beliefs, and futile attempts to remove these symbols through legal channels, surviving victims and those who sympathize with them underwent several attempts to desecrate these state-displayed political symbols.23 In either case, the use of violence to pursue just political ends is completely in line with whatever substantive values there might be.24 Let’s now turn to the use or threat of violence against persons in political protests. Tere are surely cases where resorting to violence is the only way of minimizing unnecessary confict and casualties. Both legal and illegal protests occasionally face police brutality. In normal circumstances, it is uncontroversial 20  See Huang, T. (May 20, 2005). White Terror exhibit unveils part of the truth. Taipei Times. http://www.taipeitimes.com/News/taiwan/archives/2005/05/20/2003255840. 21  See Hsiao, A. (May 25, 2015). DPP ofcial urges nationwide Chiang Kai-shek purge. Taipei Times. http://www.taipeitimes.com/News/front/archives/2015/03/25/2003614346. 22  For the justifcatory conditions of displaying political symbols, see, for example, Tsai (2016), where it is argued that frst of all, the symbol must uphold genuine political values, and second, it must be decided through a legitimate democratic procedure. Symbols that honor Chiang Kai-shek fail to meet either of these conditions. 23  See, for example, Su, C. & Chin, J. (March 1, 2016). Bust of Chiang Kai-shek vandalized in Taichung. Taipei Times. http://www.taipeitimes.com/News/taiwan/archi ves/2016/03/01/2003640556. Similar acts of vandalism occurred constantly and regularly throughout the past several years. It wasn’t until late 2017 that the government fnally yielded to the demands of the people, and started the legislation regarding the removal of symbols honoring Chiang and his authoritarian rule as part of the project of transitional justice. See, for example, Hetherington, W. (December 7, 2017). Transitional Justice Act: Schools named after Chiang Kai-shek to be renamed: education minister. Taipei Times. http://www.taipeitimes.com/News/taiwan/archives/2017/12/07/2003683572. 24  It should be emphasized that, as I suggested at the beginning of section 2, there is a signifcant diference between violence imposing costs on the culpable and nonculpable. Acts of uncivil disobedience that impose costs on the culpable would be less likely to be ruled out by proportionality. Instances of political vandalism that target governmental ofcials who act unjustly or that target unjust political symbols more often satisfy this.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Justifying Uncivil Disobedience

101

that an innocent person has the right to engage in self-defense against unjust aggressors. However, the belief that activism and civil disobedience in particular must be committed to nonviolence might lead one to hold the view that the activists and civil disobedients ought never to resort to violence.25 Once an activist or civil disobedient engages in self-defense, she has resorted to violence, and has breached the commitment to nonviolence. Accordingly, her action is unjustifable. But this is absurd. Te right to self-defense surely can’t be compromised or nullifed just because one is engaging in political activities, or because the unjust aggressor is wearing a police uniform. Terefore, the view that violence must always be prohibited during political protests must be rejected. Note that, strictly speaking, such cases don’t fall under the category of uncivil disobedience, as legitimate self-defense is typically legal or legally excused, and thus isn’t, technically, an act of “disobedience.” But some might view a protest in question (legal or illegal) and the violence employed by the protesters in self-defense as one single act, and thus dismiss the movement as “uncivil” because it “involves violence.” My discussion here is a response to those that hold such a view: if this counts as “uncivil disobedience,” so be it, but this shows that uncivil disobedience is actually justifable. I’ve argued that the case of violent self-defense in political protests should at least not be in confict with the substantive values that otherwise demand us to obey the law. In addition to self-defense, the use or threat of violence can sometimes promote substantive values more directly. Te Ferguson Unrest might be one instance: the violence involved, e.g. violently resisting arrest or threatening to resort to retaliatory activities, demonstrates and forces the public to acknowledge the seriousness of racially biased law enforcement and abuses of power that are more often left unprosecuted (Hooker, 2016). Te angry reaction from activists, some of which manifested in the form of riots, against systematic racist police violence following the rape of Téo in France is another instance. Without the more radical movements, severe police misconduct would more likely receive less than ftting punishment, for example charges against police ofcers for rape would have been reclassifed as mere aggravated assault.26 In such cases, obeying laws prohibiting violence doesn’t fulfll the duty to rescue, doesn’t support just institutions, and doesn’t respond to reason better than direct action. Tere is, therefore, no duty to obey such laws in such circumstances; resorting to violence, instead, would be demanded by the sources of our political obligations. 25  Hugo Bedau (1961) defended exactly this position: “[the civil disobedient] does not respond with violence or violent resistance during the course of his disobedience, regardless of the provocation he may have, and thus . . . he is prepared to sufer without defense the indignities and brutalities that often greet his act” (p. 656). 26  See McQueen, F. (2017). A horrifc accusation against police reignites anger in Paris suburbs. Te Conversation. https://theconversation.com/a-horrifc-accusation-againstpolice-reignites-anger-in-paris-suburbs-73314.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

102

Ten-Herng Lai

Tere is yet another way violence can contribute to the realization of the substantive grounds of our political obligation. Violent groups often “work in concert” with their nonviolent counterparts. Tis can lead to a positive radical fank efect; that is, the bargaining position of the nonviolent groups is strengthened in virtue of the presence of violent alternatives (Haines, 1984). Accordingly, the presence of violent groups increases the awareness of the issue(s) nonviolent groups attempt to address, as both groups aim at addressing the same issue(s). In addition, the presence of genuine violence helps people recognize that nonviolent groups are genuinely nonviolent. Tis prevents nonviolent groups from being mislabeled as violent, and helps them avoid the backlash and aversion violent activities typically face. Moreover, since the government and the majority would worry that the failure of nonviolent approaches would drive people to join violent campaigns, negotiating with nonviolent groups would appear to be the better option. Tis kind of “good cop, bad cop” interaction occurred during the Civil Rights Movement when Martin Luther King Jr.’s nonviolent campaign seemingly conficted with the violent approach of Malcolm X, as well as nowadays in certain environmental movements where the radical activists “assist” the mainstream nonviolent organizations by being there and being violent (Vanderheiden, 2005). It’s easy to overlook the contribution of violence. Te Civil Rights Movement, most notably, has often been presented in a sanitized fashion, where the “success” is attributed primarily or even solely to the nonviolent campaigns (Hooker, 2016). Tis obsession with nonviolence is a mistake, to say the very least. Te strategic use of violence, or at least maintaining the availability of violent alternatives, can contribute signifcantly to the realization of substantive values.27 In sum, regarding substantive accounts of the Orthodox View, we should conclude that the grounds of our political obligation might under special circumstances cease to demand obedience to laws prohibiting certain activities or even civil disobedience, but instead speak in favor of uncivil disobedience. If civil disobedience may be justifed in this way, then uncivil disobedience may be justifed in the same way.

2.2.  Procedural Justifcations What about procedural accounts? Tey hold that when legal means fail to do the job, civil disobedience is justifed in virtue of enhancing the quality 27  According to Coretta Scott King in an interview, Malcolm X intended but was unable to visit Martin Luther King Jr. when the latter was jailed in Selma in February 1965. Malcolm X instead visited Coretta Scott King, and told her: “I want you to say to him that I didn’t come to Selma to make his job more difcult but I thought that if the White people understood what the alternative was that they would be more inclined to listen to your husband.” Accordingly, this “good cop, bad cop” strategy was intentionally employed by at least some groups during the Civil Rights Movement. http://digital. wustl.edu/e/eii/eiiweb/kin5427.0224.089corettascottking.html.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Justifying Uncivil Disobedience

103

of decision-making procedures: Ignored or marginalized but important points of views are brought to the attention of the public, and this enables more inclusive democratic decisions. Might we also extend this argument to uncivil disobedience? Certain types of uncivil disobedience, though failing to meet the standards of civility in numerous ways, are mainly communicative, and it is not hard to see how they actually do enhance the quality of democratic deliberation. Whistleblowing exposes illegal or unjust conduct, and such information is necessary for citizens to make fully informed decisions. Similarly, hacktivism, ecotage, and vandalism can also be used in a communicative way, often to raise awareness about certain issues.28 It is less clear how other types of uncivil disobedience, especially those involving coercion, threats, or violence against persons, can be consistent with a procedural account. Certain instances of hacktivism, such as Anonymous’s cyber-attack “Operation Payback” against entities they perceive as being hostile toward WikiLeaks, are more attempts to retaliate than to communicate.29 Ecotage is often performed as direct action with the primary intention “to make certain present and future acts more expensive, and hence to discourage” environmentally damaging industries (Vanderheiden, 2005, p. 438). Te use or threat of violence against persons might seem to be the exact opposite of attempting to arrive at mutually acceptable solutions through deliberation. It might thus be held that most uncivil activities are incompatible with procedural considerations. However, the confict between procedural accounts and uncivil disobedi­ ence is not as stark as it might at frst appear. It could, sometimes, be questioned whether the laws targeted by uncivil disobedience are really justifed in the frst place, according to the standards of procedural accounts. If, for example, laws against ecotage were introduced solely under the infuence of corporations whose profts would be protected by such laws, while laws in favor of environmental protection were never seriously reviewed, laws against ecotage simply lack the feature of being the outcome of an inclusive democratic decision. Another example would be governmental regulations of the cyberspace. It can be seriously doubted whether those who are regulated really had a say in the issuing of those regulations (Delmas, 2018). In such cases, while uncivil disobedience doesn’t contribute to the realization of procedural considerations, it does not confict with such considerations 28  For instance, in 1996 hacktivists hacked the homepage of the United States Department of Justice and altered the title “Department of Justice” to “Department of Injustice,” as part of the protest against the Communications Decency Act (Denning, 2015 September 8). Vandalism defacing statues and spray-painting messages like “murderer” or “BlackLivesMatter” also serve as obvious examples. 29  See, for example, BBC (2010). Anonymous Wikileaks supporters explain web attacks. http://www.bbc.co.uk/news/technology-11971259. Here I leave it an open question whether Operation Payback is justifed. I’ve employed this example just to show that cyber-attacks often aren’t aimed to talk but to coerce.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

104

Ten-Herng Lai

either; for the laws that are targeted might be incompatible with the relevant procedural considerations in the frst place. Two additional points can be made concerning how to bring certain types of uncivil disobedience into line with procedural considerations, even if they don’t primarily aim to be communicative. Te frst is about the enforcement and execution of inclusive democratic decisions. Tere is, to say the least, no guarantee that the government will carry out and stick to democratically made decisions in virtue of them being made democratically. For example, it is quite possible that laws regarding environmental protection or regulations concerning permissible policing tactics are completely ignored by governmental ofcials. Moreover, it might be the case that governmental ofcials are known to constantly get away with misconduct, and that no legal measures taken by individual citizens, such as protests to expose such misconduct, can get the government to stick to the rule of law. Compared to inaction in the face of the government’s continuous neglect of the law, direct action, such as protecting the environment through ecotage or keeping policing in check through the use or threat of violence, might actually amount to enforcing the execution of democratic decisions, and hence be more in line with procedural considerations.30 Te second point concerns the robustness and stability of deliberations. According to Locke (2014, secs. 224–6), and more recently revisited by Philip Pettit (2012), the government has disproportionate power over the people it governs. In order for the people to remain in charge of the government, instead of the government having arbitrary power over the people, the people need to be ready to rise up and overthrow the government should the government forget its place and step outside of its legitimate boundaries. In a similar vein, in order for all parties to stick to democratic procedures, instead of those in power diverging from deliberation whenever diverging is to their advantage, threats and coercions to check and balance their power might be thought to be needed. Certain types of uncivil disobedience, e.g. sabotages, hacktivism, vandalism, etc., might very well be efective means of imposing costs on those who disregard democratic procedures, and keep everyone at the table. It is worth noting that, unlike the previous point, this is not about privately enforcing democratic decisions, but about ensuring that the honoring of democratic procedures doesn’t depend on the goodwill of the government.

30  One instance is the Sea Shepherds Campaigns. Illegal commercial whaling has been often done under the guise of legal “scientifc research.” Seeing that the international laws regarding whaling are often unenforced, Sea Shepherds take the matters into their own hands and enforce the law themselves. See, for example, Dryzek (2000, p. 122), Smith (2016, p. 166), and O’Sullivan, McCausland, and Brenton (2017).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Justifying Uncivil Disobedience

105

Uncivil disobedience, therefore, even if it primarily involves threats or coercions instead of providing reasons, can still be in line with procedural accounts by making deliberation robust and stable. All this being said, it is certainly possible that procedural accounts might justify a rather diferent set of uncivil activities compared to their substantive counterparts. It is not my intention to settle which account is better or to decide exactly which types of uncivil disobediences are justifable. Te aim of this chapter is simply to show that insofar as civil disobedience is taken to be justifed with reference to certain (substantive or procedural) values that ground the duty to obey the law, then this mode of argument also extends to some instances of uncivil disobedience. In short, if the Positive Tesis is true, then the Negative Tesis is false. We should conclude either a) that civil disobedience is never justifed in a reasonably just society or b) that uncivil disobedience is sometimes justifed in a reasonably just society. 3.  IS CIVIL DISOBEDIENCE SPECIAL?

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Te above arguments naturally invite the objection that there is something special about civil disobedience that serves to distinguish it from uncivil disobedience and in virtue of which the former is uniquely justifable. What is this allegedly special feature of civil disobedience? Nonviolence. Te most straightforward possibility is that civil disobedience is necessarily nonviolent, while uncivil disobedience might involve violence. Some might believe that there is something inherently wrong with deploying or threatening violence, such that violence should always be avoided, or at least that there is always some sort of “moral stain” left behind even if violence is necessary. If this is the case, then it would follow that there is a fundamental normative diference between civil and uncivil disobedience. However, it is quite difcult to understand what this inherent wrongness of violence is supposed to be. It is true that violence normally causes signifcantly more harm than nonviolent activities. Tis makes conditions such as necessity and proportionality more difcult to satisfy. It is also true that many activities that involve violence, especially those that involve physical harm, are considered mala in se, acts that are wrong in themselves independent of the law, for example rape and murder (Ristroph, 2011). Tat being said, there seems to be no difculty in conceiving of cases where violence is completely justifed and doesn’t fall under the category of mala in se. Justifed self-defense seems to be an example, even if it involves violence; so is policing according to just law. So involving violence doesn’t ipso facto make an act unjustifable or malum in se. In the case of violent uncivil disobedience, the question is

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

106

Ten-Herng Lai

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

exactly whether they can be justifed instances of violence. To assume that violent uncivil disobedience is inherently wrong, despite the fact that it can sometimes be necessary, proportionate, and furthermore promote our shared political values, seems arbitrary. Some might want to say that the state possesses the moral right to monopolize the use of violence. Again, self-defense, at least where the state is unable to provide the defense, serves as an obvious counterexample. Moreover, in societies like ours it seems that at least sometimes resorting to violence would be necessary and proportionate. Certain issues might be so severe that it would be proportionate to damage property or (threaten to) harm people to fx it. Te system might furthermore be unresponsive to a certain degree to this particular issue and the relevant political movements for such extreme measures to be necessary. Tus, even if the state is functioning relatively well on other issues, it is highly questionable whether the state retains the power to monopolize violence on this particular issue. It might be thought violence often if not always incites violence and hatred, “while nonviolence civil disobedience leaves open the possibility of a just harmony in a scale of years rather than generations” (Sabl, 2001, p. 314). While I agree that nonviolent civil disobedience can be one way of maintaining “the possibility of a just harmony,” it is unclear why the uses of constrained violence can’t. When violence is carefully employed to promote just ends, or to deter one side from unilaterally abusing violence so that all parties would stick to democratic procedures, instances of uncivil disobedience are also making a just harmony not only possible but stable. Instances of violent uncivil disobedience that are in line with the underlying values of the duty to obey are exactly such instances. Tus, they shouldn’t be ruled out just because violence is involved. Accepting legal consequences. A second possibility is that civil disobedience expresses “respect for law.” Civil disobedients improve the law through illegal activities, but accept the legal consequences to demonstrate their loyalty to the regime (King Jr., 2002; Rawls, 1999). In contrast, people who engage in uncivil disobedience, though not necessarily, often attempt to avoid or resist punishment.31 It might be said that this distinctive attitude toward punishment is what makes civil disobedience uniquely appropriate. 31  It is typical for those who engage in whistleblowing, hacktivism, ecotage, and political vandalism to attempt to avoid punishment. Tis might be because the punishment is so severe that it is unreasonable for anyone to be willing to face such punishment. Chelsea Manning was sentenced to serve thirty-fve years in prison, and served nearly seven years (due to being commuted by then President Barack Obama). “Eco-terrorists” are treated like genuine terrorists in the U.S. It might also be because some activists just want to remain anonymous, as Anonymous’s name suggests. Tough I won’t go so far as to maintain that those who engage in uncivil disobedience necessarily avoid punishment, it should be fair to say that being non-evasive is not a typical feature of uncivil disobedience.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Justifying Uncivil Disobedience

107

Two things can be said. First of all, while accepting punishment is one way of expressing one’s loyalty to the regime, it is not the only way. Taking on great personal costs when exposing state injustice in order to make reformation possible, such as in cases of whistleblowing, seems to be just as loyal as any case of civil disobedience.32 In addition, if being loyal is nothing but bearing and expressing some sort of positive attitudes toward the regime, it is unclear why incurring any costs apart from that which is necessary to make the regime better is necessary for being loyal. By doing something to improve the moral status of the state or enhancing state legitimacy, one has benefted the regime more than most law-abiding citizens. If doing more for the sake of the regime doesn’t count as being more loyal, it’s difcult to conceive what does count. Te second thing is that the reason why respect for law is important is in need of explanation in the frst place. Te most plausible explanation is that we express mutual respect through the respect for law. We interact with other persons under fair terms of cooperation, and refuse to claim any special privileges for ourselves. Tis, however, is only the case if the system is just. If the system incorporates a number of unjust laws, then expressing positive feelings and maintaining loyalty toward it would amount to expressing disrespect toward those who are systematically exploited. If mutual respect is genuinely what grounds the respect for law, when respecting the law expresses disrespect toward others, we ought to refrain from respecting the law in order to respect persons.33 Certain types of disobedients refuse to accept legal consequences. Tis might render them uncivil, but normatively speaking this seems unimportant. One can express respect for law while refusing to accept punishment. It might also be the case that one ought not to respect the law. Either way, refusing punishment doesn’t by itself render an act of disobedience unjustifed. Publicity. Civil disobedience was traditionally defned as a public act. Some might hold that if an activity lacks publicity, it’s uncivil and therefore unjustifable. However, before jumping to conclusions, it should be questioned in what sense publicity is conceived to be important, and why. One way of understanding publicity is to focus on whether the identities of the actors were voluntarily revealed. Call this “identity publicity.” Paradigm 32  Sabl (2001) has made a similar point: accepting punishment suggests a willingness to cooperate in the future. However, it was also stated that “if the other costs of disobedience are sufciently strong, and the disobedients show a willingness to incur them in the face of great temptation, this may be sufcient to show a propensity to cooperate in the future” (p. 319). It seems that this naturally extends to cover uncivil disobedience. 33  Delmas (2014a) argues that voluntarily participating in, benefting from, and maintaining a system that is unfair amounts to freeriding and expresses disrespect toward those who are exploited. She further argues that this gives us reason to disobey the system.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

108

Ten-Herng Lai

cases of civil disobedience seem to have this kind of publicity. Tose who engage in such activities openly declare their disobedience, and furthermore submit themselves to the authorities. Tis automatically entails their identities being revealed. In contrast, uncivil disobedience might involve people breaching the laws anonymously, and furthermore putting efort into remaining anonymous. However, it can be questioned why this sort of publicity is morally relevant. One answer might be that disobedients should be legally accountable for their actions in order to express the “respect for law.” Tis is incorrect, as I have argued in the subsection for accepting legal consequences. Te more plausible answer is that people should be morally accountable for their own actions, and ought to be in a position to receive moral evaluation in order to engage in moral dialogue with the general public. If this is all there is to the moral importance of publicity, it seems that identity publicity is unnecessary. People can receive moral feedback from the general public insofar as their actions are known by the public. Tey could easily distinguish whether the public is referring to and passing moral judgment on their actions. Te importance of being morally accountable suggests another way of understanding publicity, something I would like to refer to as “anonymous publicity.” Anonymous publicity concerns actions rather than actors. It’s about an act being done and being known to have been done, so that the public can pass their moral judgment to the actors. Te revelation of the identities of the actors is not necessary for such moral dialogue to take place. Anonymous publicity is of moral relevance while identity publicity isn’t. Terefore, even if uncivil disobedience is uncivil in virtue of lacking identity publicity, it is not thereby unjustifable as it still maintains anonymous publicity. Conscientiousness. Finally, some might hold that civil disobedience is special in enshrining a morally important kind of conscientiousness: frmly upholding one’s deeply held moral convictions to the extent that one is willing to bear nontrivial costs to dissociate from and alter things one perceives as morally unacceptable. Some of the earliest and most prominent accounts of civil disobedience defne it as requiring conscientiousness (e.g. Toreau, 2016; Bedau, 1961; Rawls, 1999). Perhaps it is this feature that distinguishes civil disobedience from “ordinary ofenses” or “mere criminality” (Brownlee, 2012, p. 18). Suppose that this is true. Does it distinguish civil disobedience from other types of illegal activities that are in line with the grounds of the duty to obey the law? Te answer, I believe, is no. Consider what a conscientious person would do in a scenario where civil disobedience is futile if not detrimental, but uncivil alternatives can be efectively employed to address the issue. If the person is serious about fxing the problem, she won’t exclude

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Justifying Uncivil Disobedience

109

the possibility of engaging in justifable but uncivil activities. In special circumstances, it seems that a conscientious person would blow the whistle or engage in certain types of violent activities. It might be true that conscien­ tiousness would demand us to engage in civil disobedience under certain circumstances, but it wouldn’t limit us to do so. Indeed, we can even go a step further, and hold that conscientiousness is the mark of disobedience, both civil and uncivil. By being motivated by one’s deeply held moral convictions, rather than mere self-interest or recklessness, the breaching of law in an attempt to respond to the political values that otherwise ground the duty to obey the law is distinct from “mere criminality.” Tis distinctiveness is not solely enjoyed by justifed disobedience. Disobedi­ ence can fall short of being justifed in a number of ways; most commonly because disobedients sincerely hold misguided moral convictions or miscalculate the necessity or proportionality of their disobedient actions. Nevertheless, these shortfalls don’t prevent unjustifed disobedience from being genuine attempts of responding to the grounds of our political obligations. Tus, disobedience, even if unjustifed, should be viewed, and perhaps treated, diferently from “mere criminality.” Especially in cases where the shortfall relates to the miscalculation of necessity or proportionality, we should seriously consider what sorts of defcits in the society drove people to engage in acts of disobedience, rather than hastily dismissing them as unworthy of our attention.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

4.  OTHER APPROACHES I have argued that we should conclude either a) that civil disobedience is never justifed in a reasonably just society or b) that uncivil disobedience is sometimes justifed in a reasonably just society. Suppose we assume that civil disobedience is sufciently important that we must reject a). In that case we must accept b): that uncivil disobedience is sometimes justifed in a reasonably just society. It will be instructive to contrast this view with two alternatives: frst, the view that we should expand the notion of civility such that civil disobedience, expansively construed, is uniquely appropriate in a reasonably just society; and second, the view that uncivil disobedience is permissible under (but only under) unfavorable conditions. Expanding the notion of civility. A number of philosophers have suggested that the Rawlsian defnition of civil disobedience is unduly narrow. For example, John Morreall (1976) and Piero Moraro (2007) argue that illegal activities may potentially involve violence yet still count as civil disobedience. In addition to violence, Kimberley Brownlee (2012) argues that justifed civil disobedience may also potentially involve covert planning and the rejection of legal punishment. Robin Celikates (2014) proposed a minimalist account

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

110

Ten-Herng Lai

of civil disobedience, according to which some degree of violence, covertness, or defance can be incorporated. Tis suggests a way of saving the Orthodox View: It isn’t the case that uncivil disobedience is sometimes justifed, but that some of the activities that were dismissed by the Rawlsian defnition as uncivil instead count as instances of civil disobedience, sufciently broadly construed, and, hence, are perfectly appropriate. Perhaps the Negative Tesis is correct after all. Uncivil disobedience is indeed never permissible in a reasonably just society. It may well be right that we should adopt a broader notion of civility and interpret the Orthodox View accordingly. Tis would allow the Orthodox View to be more inclusive with respect to the kinds of activities that count as appropriate within a reasonably just society. But it is highly questionable that there is any plausible broadening that would encompass all the disruptive activities we mentioned above. Tere are clearly limits to what can count as civil. And it stretches credulity to suggest that whistleblowing, cyber-attacks, sabotages, and uses or threats of violence against persons really count as “civil” in any sense of the term that we would recognize.34 In addition, there is something unsavory about this approach. Proponents of this view tend to alter or expand the defnition of civility to make it the case that all justifed acts of disobedience just happen to be civil. If a type of action that was previously viewed as uncivil (say, because it involves violence or refuses punishment) turns out to be justifable, it becomes civil by defnition in virtue of being justifable. Tis undermines the justifcatory role of civility. A type of action is potentially justifable because it’s civil, but civil because it’s potentially justifable. Tis is simply circular. In contrast to this approach, my view doesn’t rely on the defnition of civility. Tus, my view avoids the need to defne civility to accurately capture all potentially justifable activities, and, more importantly, averts the risk of prematurely excluding potentially justifable activities due to overly narrow defnitions. Moreover, my view stays clear of the strategy of re-defnition, and thereby avoids the risk of running into a circulation. Uncivil disobedience in unfavorable conditions. Te other view that is worth mentioning holds that uncivil disobedience is sometimes justifed in unfavorable conditions. A. John Simmons (2010) has argued that the application of Rawls’s account of civil disobedience is extremely narrow: it only applies to cases that occur in what Rawls calls “near-just societies.” Tis 34  Delmas (2018), for example, argues that hacktivism can’t just be shoehorned into the traditional framework of civil disobedience: “To accommodate novel forms of digital resistance, what is called for is neither an unrefective application of an ill-ftting and too narrow concept of civil disobedience, nor an extension of the latter concept beyond recognition. Instead, we need to enrich our conceptual framework and devise additional lenses besides ECD [i.e. electronic civil disobedience] to approach these phenomena” (p. 20).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Justifying Uncivil Disobedience

111

reaction can be extended to any version of the Orthodox View. Te Orthodox View only applies to “reasonably just societies.” It might be said that, whatever is required to count as “reasonably just,” no existing society is likely to meet it. For example, no basic structure of any constitutional democracy is anything close to being designed according to the Rawlsian or any reasonable principles of justice, and rules governing any cooperative scheme are anything but fair. Perhaps the proponent of the Orthodox View is right that civil disobedience is uniquely justifed within a reasonably just society. But the interesting question is not what is appropriate within a reasonably just society but what is appropriate within the less favorable conditions that characterize the actual societies we live in. According to Simmons and others, there is no reason to think that civil disobedience is uniquely appropriate in these less favorable conditions. Rather, we should expect uncivil disobedience at least sometimes to be justifed as well. Let’s concede that uncivil disobedience is sometimes appropriate in unfavorable conditions. Te question is whether these are the only conditions in which it is justifable. In particular, is uncivil disobedience also sometimes justifed in a reasonably just society? Reasonably just societies inevitably remain fawed in a number of diferent ways, and fall short of realizing the substantive or procedural values that ground the duty to obey the law in such societies. Indeed, civil disobedience will represent the best way of remedying some of the faws, but other faws, e.g. concealed state misconduct and urgent environmental hazards, are better addressed by certain forms of uncivil disobedience. Unless we have a separate argument showing that societies that are fawed in ways that can only be appropriately responded to by some form of uncivil disobedience always fall short of being reasonably just, we have to admit that appealing to “being reasonably just” doesn’t exclude the justifability of uncivil disobedience. In contrast to this approach, my view avoids the need to provide an argument supporting why “being reasonably just” excludes the justifability of uncivil disobedience. I’m simply skeptical of whether a plausible argument can be made. In addition, this way of rescuing the Orthodox View renders the Orthodox View irrelevant: it has no application now, and I suspect that it might not ever have any application in any future society.

CONCLUSION Te Orthodox View of civil disobedience justifes such illegal activities by appealing to the considerations that ground the duty to obey the law and suggesting that civil disobedience constitutes the most appropriate response to situations where the quality of the law deviates from the ideal. I have

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

112

Ten-Herng Lai

argued that if this approach is correct, it should be extended to uncivil disobedience. Exactly the same considerations that, in some circumstances, will support civil disobedience will, in other circumstances (that are diferent but not necessarily more unjust), support uncivil disobedience instead. I’ve considered and rejected several versions of the objection that there is something special about civil disobedience, and distinguished the view from two rivals that seek to accommodate some of the same intuitive data in diferent ways. Tus, we should be skeptical of the privileged normative place that civil disobedience is typically taken to occupy in political theory and practice. It is a mistake to overplay the importance of standards of civility, whatever they might be, when evaluating illegal dissent and considering what sorts of actions we should take in response to injustice. What is of fundamental importance isn’t whether an act is civil; rather, acts of disobedience, civil or uncivil, are justifed insofar as they are necessary, proportionate, and represent the best way of responding to the political values that purportedly ground the duty to obey the law.35,36

References

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Bazargan, S. (2014). Killing minimally responsible threats. Ethics, 125(1), 114–36. Bedau, H. A. (1961). On civil disobedience. Te Journal of Philosophy, 58(21), 653–65. 35  My view leads to two potential practical implications: Te frst relates to punishment. Te aims of punishment might be to deter undesirable actions, infict deserved sufering, get back from freeriders, and/or express condemnation. It might very well also be the case that only those who owe something due to past wrongdoing are liable to be punished (Tadros, 2011). Te punishment of justifed acts of uncivil disobedience, even the more uncivil forms, might be inconsistent with the aims and constraints of punishment. Te second is more political. According to Bernard Harcourt (2012), “the ability to brand a particular discourse as uncivil is itself a political accomplishment that refects a certain position of privilege” (p. 348). Te exact same thing might be said about those too ready to call out the incivility of certain acts of disobedience. Sometimes overplaying the importance of civility is more than a simple mistake. We should be wary of whether putting acts of disobedience under the strictest scrutiny are just deliberate attempts to silence and disempower vulnerable and desperate people. 36  Earlier versions of this chapter were presented at Disobey! Understanding the Politics and Ethics of Disobedience (2015), ANU Philsoc Seminar (2016), ANU MSPT Work in Progress Workshop (2017), Oxford Studies in Political Philosophy 5th Annual Conference (2017), and Stockholm Centre for the Ethics of War and Peace Graduate Reading Retreat (2017). I would like to thank Nicholas Southwood, Philip Pettit, Seth Lazar, Justin Bruner, Lachlan Umbers, Kimberley Brownlee, Peter Vallentyne, Sameer Bajaj, Renee Bolinger, Candice Delmas, Ned Dobos, Yun Ying Lau, Holly LawfordSmith, RJ Leland, Emily McTernan, Matthew Lindauer, Shmulik Nili, William Smith, and the participants of the above-mentioned workshops and conferences. Tis research is supported by an Australian Government Research Training Program (RTP) Scholarship.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Justifying Uncivil Disobedience

113

Brownlee, K. (2012). Conscience and conviction: Te case for civil disobedience. Oxford University Press. Celikates, R. (2014). Civil disobedience as a practice of civic freedom. On global citizenship: James Tully in dialogue (pp. 207–28). Bloomsbury Academic. Delmas, C. (2014a). Political resistance: A matter of fairness. Law and Philosophy, 33(4), 465–88. Delmas, C. (2014b). Te ethics of government whistleblowing. Social Teory and Practice, 41(1), 77–105. Delmas, C. (2018). Is hacktivism the new civil disobedience? Raisons Politiques, 1, 63–81. Denning, D. E. (2001). Activism, hacktivism, and cyberterrorism: Te Internet as a tool for infuencing foreign policy. Networks and netwars: Te future of terror, crime, and militancy, 239, 288. Rand Corporation. Denning, D. E. (2015, September 8). Te rise of hacktivism. Georgetown Journal of International Afairs. Retrieved from https://www.georgetownjournalofnternational afairs.org/online-edition/the-rise-of-hacktivism. Draper, K. (2016). Necessity and proportionality in defense. In Te ethics of self-defense (pp. 171–84). Oxford University Press. Dryzek, J. S. (2000). Deliberative democracy and beyond: Liberals, critics, contestations. Oxford University Press. Dworkin, R. (1978). Taking rights seriously (Vol. 136). Harvard University Press. Fung, A. (2005). Deliberation before the revolution toward an ethics of deliberative democracy in an unjust world. Political Teory, 33(3), 397–419. Habermas, J. (1996). Between facts and norms, trans. William Rehg (pp. 274–328). Polity. Haines, H. H. (1984). Black radicalization and the funding of civil rights: 1957–1970. Social Problems, 32(1), 31–43. Hampson, N. C. (2012). Hacktivism: A new breed of protest in a networked world. Boston College International and Comparative Law Review, 35, 511. Harcourt, B. E. (2012). Te politics of incivility. Arizona Law Review, 54, 345. Himma, K. E. (2007). Internet security: Hacking, counterhacking, and society. Jones & Bartlett Learning. Hooker, J. (2016). Black Lives Matter and the paradoxes of US Black politics: From democratic sacrifce to democratic repair. Political Teory, 44(4), 448–69. King Jr., M. L. (2002). Letter from Birmingham city jail. In Civil disobedience in focus (pp. 74–90). Routledge. Kirkpatrick, J. (2008). Uncivil disobedience: Studies in violence and democratic politics. Princeton University Press. Lefkowitz, D. (2007). On a moral right to civil disobedience. Ethics, 117(2), 202–33. Locke, J. (2014). Second treatise of government: An essay concerning the true original, extent and end of civil government. John Wiley & Sons. Markovits, D. (2005). Democratic disobedience. Yale Law Journal, 114, 1897–952. McMahan, J. (2009). Killing in war. Oxford University Press. Montague, P. (2010). Self-defense, culpability, and distributive justice. Law and Philosophy, 29(1), 75–91. Moraro, P. (2007). Violent civil disobedience and willingness to accept punishment. Essays in Philosophy, 8(2), 6.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

114

Ten-Herng Lai

Morreall, J. (1976). Te justifability of violent civil disobedience. Canadian Journal of Philosophy, 6(1), 35–47. O’Sullivan, S., McCausland, C., & Brenton, S. (2017). Animal activists, civil disobedience and global responses to transnational injustice. Res Publica, 23(3), 261–80. Pettit, P. (2012). On the people’s terms: A republican theory and model of democracy. Cambridge University Press. Post, R. C. (1993). Between democracy and community: Te legal constitution of social form. Nomos, 35, 163–90. Rawls, J. (1999). A Teory of Justice. Oxford University Press. Raz, J. (1986). Te morality of freedom. Clarendon Press. Ristroph, A. (2011). Criminal law in the shadow of violence. Alabama Law Review, 62. Sabl, A. (2001). Looking forward to justice: Rawlsian civil disobedience and its non-Rawlsian lessons. Journal of Political Philosophy, 9(3), 307–30. Simmons, A. J. (2010). Disobedience and its objects. Law Review, 90, 1805. Smith, W. (2011). Civil disobedience and the public sphere. Journal of Political Philosophy, 19(2), 145–66. Smith, W. (2016). Te boundaries of a deliberative system: Te case of disruptive protest. Critical Policy Studies, 10(2), 152–70. Tadros, V. (2011). Te ends of harm: Te moral foundations of criminal law. Oxford University Press. Tadros, V. (2012). Duty and liability. Utilitas, 24(2), 259–77. Toreau, H. D. (2016). Civil disobedience. Broadview Press. Tsai, G. (2016). Te morality of state symbolic power. Social Teory and Practice, 42(2), 318–42. Vallentyne, P. (2011). Enforcement rights against non-culpable non-just intrusion. Ratio, 24(4), 422–42. Vallentyne, P. (2016). Defense of self and others against culpable rights violators. In Te ethics of self-defense (pp. 86–109). Oxford University Press. Vanderheiden, S. (2005). Eco-terrorism or justifed resistance? Radical environmentalism and the “War on Terror.” Politics & Society, 33(3), 425–47. Wellman, C. H. (2005). Samaritanism and the duty to obey the law. In C. Wellman & J. Simmons, Is Tere a Duty to Obey the Law? (pp. 3–89). Cambridge University Press.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

PA RT I I

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

POLITICAL, LEGAL, A N D   M O R A L R E L AT I O N S

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved. Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

5

Discrimination and Subordination

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Sophia Moreau

Tere is something particularly compelling about the idea that discrimination is wrong when and because it unjustly subordinates some people to others. Te kinds of discrimination that usually give rise to the greatest moral indignation involve the creation or perpetuation of diferent classes of people, with some having a superior status and others an inferior one, in circumstances where we think that everyone ought to have an equal status. Consider, for instance, the Jim Crow laws, which turned African-Americans into second-class citizens; or dress codes for waitresses or female retail employees that mark them out as sexual objects, lacking the full and independent agency that we ascribe to men. Moreover, the idea that unjust subordination can help to explain the wrongness of certain forms of ­discrimination does not just have a hold on our moral imaginations. It is also deeply rooted in the law. Both the United States’ Fourteenth Amendment and the constitutional equality rights in Section 15 of the Canadian Charter of Rights and Freedoms have been understood—by courts, and also by academics—as prohibiting government policies that subordinate people based on certain traits.1 And of course, when private sector anti-discrimination law was frst developed in these two countries in the 1960s and 1970s, it 1  For seminal discussions of unjust subordination in the context of the United States’ Fourteenth Amendment jurisprudence, see Owen Fiss, “Groups and the Equal Protection Clause,” Philosophy & Public Afairs 5(2) (1976), pp. 107–77; Ruth Colker, “AntiSubordination Above All: Sex, Race, and Equal Protection,” New York University Law Review 61 (1986), pp. 1003–44; Mari J. Matsuda, “Voices of America: Accent, Antidiscrimination Law, and a Jurisprudence for the Last Reconstruction,” Yale Law Journal 100(5) (1991), pp. 1329–407; J.M. Balkin, “Te Constitution of Status,” Yale Law Journal 106(8) (1997), pp. 2313–74; and J.M. Balkin and Reva B. Siegel, “Te American Civil Rights Tradition: Anti-classifcation or Anti-subordination?,” 58 University of Miami Law Review 9 (2003) p. 246. Canadian courts adjudicating equality rights cases tend to speak of “stereotyping” rather than of subordination, but their discussions of stereotyping are often discussions about subordination. See, for instance, Colleen Sheppard, Inclusive Equality (New York: McGill-Queen’s University Press, 2010).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

118

Sophia Moreau

was treated as a form of quasi-criminal law that aimed to eliminate acts of prejudicial subordination, acts that deliberately denied certain privileges or benefts to members of certain social groups on the grounds that these groups were less worthy than others.2 But what exactly does unjust subordination involve? And how do discriminatory acts and policies work to subordinate certain social groups and to sustain their subordination?3 It is the aim of this chapter to answer these questions. Most of the legal scholars who analyze discrimination in terms of its contribution to subordination invoke a relatively under-specifed, intuitive idea of subordination. For instance, when Owen Fiss frst urged that the US Equal Protection Clause was best interpreted not as preventing arbitrary classifcations but as eliminating unfair subordination, he suggested that subordination was a “status harm” that involved perpetuating the lower social position of persistently disadvantaged social groups.4 But it was not his aim to develop a general account of what that status harm involved. More recently, Reva Siegel and Joel Balkin have examined the ways in which courts, in cases of discrimination, are motivated by concerns with “social stratifcation” and “the secondary social status of historically oppressed groups.”5 But they do not explain in detail what “social stratifcation” involves. Teir interest is largely in charting how concerns about social stratifcation motivate judges to adopt particular legal doctrines and to decide certain cases in certain ways. My aim in this chapter is diferent. It is to develop a philosophical account of subordination, one that can help us understand some of the ways in which diferent forms of discrimination subordinate people, and the reasons why they might be wrong, in virtue of contributing to such subordination. I will not try to specify a set of individually necessary and jointly sufcient conditions for social subordination. But I shall lay out in some detail four conditions that seem to be satisfed in most cases where the subordination of a certain group persists across diferent social contexts for some extended period of time and seems unjust. I shall do this in Section 3 of the chapter. Section 4 will consider how discrimination contributes to subordination, so understood. 2  See Denise Réaume, “Harm and Fault in Discrimination Law,” Teoretical Inquiries in Law 2 (2001), pp. 349–85. 3  Here, and often in the chapter, I shall use the term “subordination” as a shorthand for “unjust subordination.” Tis is not because my account presupposes that all forms of subordination are unjust; it is just a convenient shorthand. I do assume that whenever one social group is assigned an inferior status across a number of diferent social contexts for some lengthy period of time, this is unjust; but this assumption is quite consistent with the view that it may sometimes be just to assign a certain social group (e.g. employees) a status lower than that of another (e.g. their managers) for some specifc purpose in a certain limited context (e.g. for work-related purposes, while at their place of work). 4  Fiss, supra note 1 at 157. 5  Balkin and Siegel, supra note 1 at 256.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

119

Te scholars of discrimination law who come closest to ofering a ­ hilosophical analysis of subordination and of how discrimination suborp dinates are expressivists Deborah Hellman, Elizabeth Anderson, and Richard Pildes.6 Tey have argued that an act is wrongfully discriminatory when it subordinates a person to others in the sense that the act demeans her, or sends the message that she is of less value than others. I shall consider Hellman’s account in Section 2 of the chapter. While this account is extremely helpful in drawing our attention to the ways in which discriminatory acts send messages about the inferior status of certain groups, I shall suggest that, at least in its current form, the account is too individualistic. It focuses too much on the individual power dynamic between the discriminator and the discriminatee, when in fact we need to look at the relative amounts of power possessed by the diferent social groups to which these people belong. I shall also argue that expressivism ofers us too narrow an understanding of subordination. Subordination is not only a function of the social messages sent by particular acts or policies. Rather, it is kept in place by a variety of efects that discriminatory acts have on diferent social groups, such as perpetuating diferences in power and authority between them and rendering certain social groups or their needs invisible in certain contexts. In addition to focusing on social groups, my own account of subordination places special emphasis on one feature of subordination that is not often foregrounded. Tis is the fact that subordination often depends, for its persistence, on what I shall call “structural accommodations.” Tese are policies, practices, and physical structures that tacitly accommodate a more privileged group’s needs at the expense of the subordinate group or groups. Normally, within anti-discrimination law, we use the term “accommodation” to refer to a special measure that must be adopted in order to give the subordinated group an opportunity equal to that of the more privileged group. And we assume that the subordinate group requires an accommodation because that group has certain special needs. So, for instance, when a Muslim employee requests an altered work schedule so that he can pray at the times that his religion requires, we treat the altered schedule as an “accommodation” to which he is entitled, because of the special demands of his religion. As we have learned from feminists, critical race theorists, and disability theorists, however, at least part of the reason why these groups require an accommodation in the frst place is that our social environment has been constructed in such a way as to tacitly accommodate the needs of more 6  Elizabeth Anderson and Richard Pildes, “Expressive Teories of Law: A General Restatement,” University of Pennsylvania Law Review 148(5) (2000), pp. 1503–76, especially pp. 1533–44; Deborah Hellman, When Is Discrimination Wrong? (Cambridge, MA: Harvard University Press, 2008).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

120

Sophia Moreau

privileged groups.7 I shall be proposing that we need to think of these prior policies, practices, and structures as “accommodations”—accommodations to the more privileged social groups, which make their interests and needs seem normal and the interests of other groups seem exceptional. And I shall argue that we cannot understand the subordination of one group by another, or the real contribution of discrimination to subordination, unless we consider these “structural accommodations.” For they serve indirectly to rationalize the greater power and de facto authority that are held by these groups and the greater deference we pay to them. We can, I shall argue, see many of the policies that constitute wrongful indirect discrimination as “structural accommodations” that contribute to unjust subordination. So my account of subordination will give us a way of explaining in detail not just how direct discrimination subordinates, but how indirect discrimination, too, subordinates. Before I turn to this account of subordination, however, I need to explain what I take discrimination to involve. I shall do so in Section 1, by discussing some common examples of discrimination in the restaurant industry. I shall return to these examples at later stages in the chapter, both to evaluate the expressivist account of how discrimination subordinates and to develop my own account.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

1.  DIRECT AND INDIRECT DISCRIMINATION: UNIFORMS FOR RESTAURANT EMPLOYEES Most countries that prohibit discrimination recognize two forms of potentially wrongful discrimination: “direct discrimination” and “indirect discrimination” (or, in the United States, “disparate treatment” and “disparate impact”). In this section of the chapter, I want to lay out a rough defnition of these two forms of discrimination along with an example of both, so that we can have an idea of what it is that we think subordinates people, when we speak of “discrimination” subordinating. I shall work with standard legal ideas of discrimination, rather than frst trying to develop my own more rigorous defnitions from some independent moral standpoint, for two reasons. Te frst is that part of my aim in this chapter is a legal one. It is to see how far we can make sense of the particular phenomenon that the law carves out as “discrimination,” as wrongful because it subordinates. But the second reason why I am taking our legal concepts of discrimination as my starting point is that it seems to me that even a purely moral analysis of discrimination needs to attend at least to the broad contours of the defnitions 7  See the works cited in note 15, infra.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Discrimination and Subordination

121

of discrimination that have been carved out by the law. Tis is because our shared moral understanding of discrimination has been deeply shaped by our legal regimes for regulating it—to such an extent that if someone departed too radically from the core understanding of discrimination given to us by our laws, it might seem that they were talking about a diferent moral phenomenon entirely.8 What, then, does the law take to be potentially wrongful discrimination? Consider, for a start, the practice of some restaurants and bars of imposing a gendered dress code on their employees.9 Rather than providing a gender-­ neutral set of options and allowing each employee to choose what suits them, these restaurants require female employees to wear tighter ftting clothing designed to show the shape of their bodies, low-cut tops, shorter skirts, and high heels. Men, by contrast, are usually permitted to wear more comfortable, looser-ftting clothing that is not revealing. If you have thought about such dress codes, you have probably already refected on the messages they send about the appropriate social roles of men and women; the ways in which they mark women out as inferior; and the additional physical and health burdens they place on women, both through the tight clothes that restrict their movement and through the high heels that cause foot and back pain. Tey are standard examples of what, in most legal jurisdictions, we call “direct discrimination”:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Direct Discrimination:  A policy directly discriminates against a person, P, if the policy treats P less favorably on the basis of some trait, t, than it would treat those who lack t, either by explicitly singling out people with 8  In this respect, discrimination is arguably diferent from many other moral wrongs. We could certainly explain what murder is and why it is morally wrong without looking at the structure of criminal prohibitions on murder, just like we can explain why promises must be kept without consulting contract law. But I am not sure how an account of discrimination and its wrongness could expect to be accurate if it did not accord with at least some of the shared features of many anti-discrimination laws across diferent jurisdictions—such as the fact that most jurisdictions recognize some kind of distinction between direct and indirect discrimination, and the fact that most treat discrimination as unfair only if it occurs on the basis of a limited list of “prohibited grounds” of discrimination. For further discussion of the relevance of the law to analyses of discrimination, see my article “Equality and Discrimination,” forthcoming in Te Cambridge Companion to Philosophy of Law, ed. John Tasioulas (Cambridge University Press). For defense of the claim that there is enough common ground between the diferent anti-discrimination laws of diferent countries to talk about there being a “core understanding” of discrimination, see Tarun Khaitan, A Teory of Discrimination Law (Oxford: Oxford University Press, 2015). 9  For an overview of common practices surrounding restaurant dress codes and their impact on a variety of underprivileged groups, see Not on the Menu: Ontario Human Rights Commission Inquiry Report on Sexualized and Gender-based Dress Codes in Restaurants (March 2017), available at http://www.ohrc.on.ca/en/not-menu-ohrc-inquiryreport-sexualized-and-gender-based-dress-codes-restaurants.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

122

Sophia Moreau

t or by singling out those who have a diferent trait, u, that is in some way very closely connected to t (for instance, only those who have t can have u, or most of those who have t do have u . . .), where t is a prohibited ground of discrimination. Te dress codes for employees adopted by many restaurants also indirectly disadvantage transgendered people, pregnant women, and members of certain religions, though none of these groups is explicitly singled out by the policies. Trans employees may have to pigeon-hole themselves into the uniform for one gender or the other, even if this does not express their current gender identity. Pregnant women are often disproportionately burdened because at some point their pregnancy makes it impossible for them to ft into a tight-ftting uniform, and they then face the difculty of having to tell their employer sooner than they would have liked. Similarly, members of religious groups that require particular modes of attire may fnd it difcult or even impossible to adopt the dress codes while practicing their religion, and so might be unable to work in this industry. So, in addition to constituting direct discrimination against women as a group, the dress codes constitute what in the law we call “indirect discrimination”:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Indirect Discrimination:  A policy indirectly discriminates against a person, P, on the basis of trait t, if the policy, though it does not explicitly single P out because of t or some related trait, u, nevertheless disproportionately disadvantages those who have t relative to those who do not, where t is a prohibited ground of discrimination. Note that, as defned here, the main diference between direct and indirect discrimination is that direct discrimination explicitly singles out a certain group or person based on a certain trait or some trait that is closely connected to it, whereas policies that discriminate indirectly do not. Te latter are apparently neutral, seemingly applying the same criterion to everyone; but they nevertheless have a disproportionately disadvantageous efect on a group that shares a trait that is a prohibited ground of discrimination.10 Tere are two related practices worth examining, in connection with restaurant dress codes; for these practices constitute an even more subtle form 10  Neither this defnition of indirect discrimination nor my earlier defnition of direct discrimination presuppose that it is always wrong to engage in direct or indirect discrim­ ination. And this is both consistent with most jurisdictions’ approaches to discrimination and intuitively plausible: most recognize that certain kinds of considerations can justify discrimination, rendering it not wrongful. I shall, later in the chapter, use my account of subordination to argue that when they contribute to subordination, then direct and indirect discrimination are wrong. But I am not presupposing that they always contribute to subordination, or that they are always wrong.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Discrimination and Subordination

123

of indirect discrimination than those I have already examined. First, many restaurants do not stock any uniforms of a kind that might be easily put on and worn by people with muscular disabilities (disabilities that make it diffcult for them to put on or wear tight clothing, for instance, or to do up all of the buttons on button-up shirts). And this means that such people, when hired, are placed in the difcult position of being unable to put on their uniform without assistance, or of having to step forward and ask for a different uniform and so present themselves as “abnormal.” Second, even restaurants that have ofcially adopted gender-neutral, disability-friendly dress codes often hand new employees a training manual that has pictures only of young, svelte, conventionally attractive women dressed in the most feminine uniform options. Insofar as such manuals reinforce the image of people as useful and employable only insofar as their physical appearance is “normal,” and insofar as they suggest that it is part of a woman’s role as a waitress to use her body to gratify men, they disproportionately burden women, pregnant women, people with disabilities, and trans people. Let us now consider what expressivist theories of discrimination would say about such practices, looking in particular at Deborah Hellman’s theory of discrimination as wrongful because it demeans.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

2.  HELLMAN’S EXPRESSIVIST VIEW: SUBORDINATING BY DEMEANING Hellman has argued that discrimination is wrong when and because it puts someone down, treating them as though they are “not fully human” or “not as worthy as others.”11 She uses the term “demean” to refer to the kind of subordination that she has in mind. One might think that to demean someone is simply to act in a way that sends the social message that another person is less worthy of respect. And this is certainly how traditional expressivists have understood it. But Hellman uses the term “demean” in a specialized way, to refer to a sub-group of those acts that send the message that someone is less worthy of respect than others. For she notes, quite rightly, that not all acts that send such messages actually do afect the social status of others. For instance, she suggests, when an employee spits at her boss or a child taunts her classmate, their acts send an inferiorizing social message but do not actually lower the status of the person insulted.12 In Hellman’s view, in order to constitute objectionable subordination, a discriminatory act must both (i) send an inferiorizing message and (ii) by virtue of sending 11  Hellman, supra note 6, p. 35.

12  Ibid., 35–6.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

124

Sophia Moreau

that message, actually put the person down or lower his status. Whether it actually puts him down depends in large part, she suggests, on whether the discriminator has power over, or a higher status than, the discriminatee.13 So the employee cannot put her boss down because she lacks the power to do so; and the same is true of the child relative to her classmate. It seems right that many acts of discrimination that subordinate do so in part because of the social messages they send—messages about the inferiority of some groups or the superiority of others. And it also seems right that, in order to subordinate someone, one must afect their status in the world rather than simply expressing something about them; though, as I shall later argue when I develop my own account of subordination, it isn’t clear that subordinating a person always involves lowering their status, as opposed to simply confrming or perpetuating an already inferior status. But there are at least two respects in which the expressivist account seems problematic. First, whether one person’s expression of disrespect for another can actually put that other person down seems to depend not just on the power diferential between the two individuals, but more broadly, on the power and authority that the social groups to which they belong have, relative to each other. Consider again Hellman’s examples of spitting at one’s boss or insulting a classmate. Surely these acts could sometimes put the person down. If I am a white employee and you are my African-American boss and I spit at you in full view of all of my other fellow white employees, I can indeed lower your social status. (I can’t, of course, afect your employment status—that is, your status as my boss. But that isn’t the kind of status that is at issue here, when we are considering whether someone is put down by an act. What’s at issue is social status. And I can certainly lower your social status by spitting at you, even if I am just your employee.) I can do that precisely because, in our version of the example, I am white and you are black, and whites and blacks in our society have a certain history, relative to each other, which my act invokes. Similarly, if the child insulted by a classmate in Hellman’s example is indigenous and he is called a “drunk Indian” by a white child, the act can lower his status in the classroom, or perpetuate his already low status. Tat is because the group to which the white child belongs has historically possessed, and continues to possess, a great deal of power and authority over the ways in which indigenous children are portrayed in our society. So, in order to assess whether someone’s act puts down another, we need to look not just to the positions of these two people relative to each other, but to the social groups to which each belongs and the relationship of these groups to each other. 13  Ibid., 36.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

125

Hellman could, I believe, adopt this more group-based perspective c­ onsistently with her expressivist tenets. So this frst difculty points simply to a way in which the theory requires modifcation. Te second difculty that I shall discuss, however, is a difculty with the expressivist core of this view. As I shall now argue, it is unclear that it is only the social message sent by a particular act or policy that determines whether it unjustly subordinates. To see this, let’s turn back to my example of uniforms for restaurant employees. Consider frst the gendered dress code that requires female employees to dress in sexualized and revealing ways. It is true that this dress code sends a message about women needing to appear in a sexualized way so as to please male clients, and that, given the overall context, the message really does have the efect of confrming their already inferior status. But the dress code also seems to do many other things to women, and it does not do these things only because it sends a message about women’s inferior status. Requiring women to dress in tight clothing and heels hampers their ability to move, thereby giving them less power in the workplace than their male colleagues and male clients; so it lessens their power, quite independently of the social message it sends about them. It gives women less authority over their own self-presentation than men have, again, not because it sends a message about women’s inferiority but because it denies them a choice that it gives to their male colleagues. It presents a certain conception of how women ought to dress as preferable than others, thereby elevating the women who dress this way above those women who do not, creating a hierarchy within the class of women (a hierarchy that in turn helps to perpetuate women’s subservience to many men’s attitudes about them). And it renders invisible those women who have fgures that do not conform to this image. Although we do not yet, at this stage in the chapter, have a detailed conception of subordination to help confrm the relevance of these facts to women’s subordination, it seems implausible to suppose that these facts are irrelevant. But the expressivist account gives us no way of recognizing the relevance of such facts, except insofar as they can be reinterpreted as aspects of, or efects of, the social message of inferiority sent by the discriminatory policy. Moreover, my examples of indirect discrimination—for instance, the disproportionate burden that gendered dress codes placed on trans people, pregnant women, and members of certain religious groups—seem also to present us with cases of discrimination that subordinate people, and that do so in a similar, though not identical, way to direct discrimination. But it is unclear that the dress codes send a message about the inferiority or lack of worth of these groups: they seem rather simply to overlook their situations. So it is unclear that an expressivist could recognize this as discrimination of a sort that wrongfully subordinates. Hellman would likely reply that in her view, direct and indirect discrimination are two diferent kinds of wrongs.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

126

Sophia Moreau

Direct discrimination demeans, and so wrongfully subordinates; whereas indirect discrimination does not usually demean but is wrong for some other reason—for instance, as Hellman has recently suggested, because it compounds past injustices.14 Perhaps Hellman might even argue that “compounding a past injustice is a kind of subordination”; it is just a diferent kind of subordination. However, if both of these forms of discrimination seem to confrm or perpetuate the lower status of certain social groups, then, rather than drawing a bright line between those that demean and those that do not, perhaps we should see whether there is a single account of subordination that can allow us to explain, in rich and detailed ways, how diferent forms of discrimination work to create or confrm a person’s lower status. And, as I argued earlier, we should try to do so in a way that clearly acknowledges the role of the social groups to which that person and the discriminator belong. In the next section of the chapter I shall try to develop such an account.15 3.  TOWARD AN ACCOUNT OF SOCIAL SUBORDINATION

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

What we need is an account of subordination that considers the broader relationship between the social group (or groups) to which the discriminatees 14  Hellman, “Indirect Discrimination and the Duty to Avoid Compounding Injustice,” Chapter 5 of Foundations of Indirect Discrimination Law, ed. Hugh Collins and Tarunabh Khaitan (Oxford: Hart Publishing, 2018), pp. 105–21. 15  I don’t purport to ofer the only philosophical account of unjust subordination; though I do think it is the only one that has been systematically developed in light of concerns about how discrimination subordinates. Philosophers have discussed subordi­ nation in other contexts, and I draw on their work in what follows. For instance, feminists such as Rae Langton and Catharine MacKinnon have developed theories of subordination within discussions of pornography: see, for example, Rae Langton, Sexual Solipsism (Oxford: Oxford University Press, 2009) and Catharine MacKinnon, Only Words (Cambridge, MA: Harvard University Press, 1993). Relational egalitarians such as Elizabeth Anderson and Samuel Schefer have given analyses of the kinds of non-subordinating relationships that are required in order for us to live in a true society of equals—that is, a society in which everyone stands, to every other person, in a relationship of equality. See Elizabeth Anderson, “What Is the Point of Equality?,” Ethics 109(2) (1999), pp. 287–337 and “Te Fundamental Disagreement between Luck Egalitarians and Relational Egalitarians,” Canadian Journal of Philosophy: Supplementary Volume on Justice and Equality 36 (2010), pp. 1–23; and Samuel Schefer, “Te Practice of Equality,” in Carina Fourie, Fabian Schuppert, and Ivo Wallimann-Helmer (eds.), Social Equality: On What It Means to Be Equals (Oxford: Oxford University Press, 2015), pp. 21–44. Lastly, subordination has been recently discussed by political philosophers who argue that democracy is valuable because, and insofar as, it enables us to relate to each other as social equals, without subordination. See Niko Kolodny, “Rule Over None II: Social Equality and the Justifcation of Democracy,” Philosophy & Public Afairs 42(4) (2014), pp. 287–336.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

127

belong and the social group to which the discriminator belongs. We need an account of what I shall call social subordination—that is, the state of afairs in which one social group has a standing in society that is lower than that of another. As a preliminary to developing such an account, I should clarify what I mean by a “social group.” As I am using the term, a “social group” is an entity that has an existence apart from any particular member: one can speak about the group without reference to those who happen to be its current members. A social group shares, or is presumed to share, a certain trait. But it is not just any group of people who happen to share a certain trait, such as “people with bushy eyebrows,” that count as a social group in the sense that I am concerned with. Rather, the kinds of social groups that are subordinated either possess, or are presumed by others to possess, a particular type of trait, a trait that is socially salient, in the sense that others in society take that trait to have implications for the character and behavior of members of the group, and for the social roles that they are capable of occupying. And it is often by this socially salient trait, or a combination of it and other traits associated with a sub-group of those who possess this trait, that we identify members of that social group. So, for instance, “wearers of blue, crimson, or scarlet velvet” does not, in our current society, mark out a group of people through their shared possession of any socially salient trait. But in Tudor England, where Sumptuary Laws regulated the materials and colors that people from each social stratum could wear, it marked out the group of noblemen who stood at or above the level of a Knight of the Garter. Tis was an important social marker in those times; and it was a moral marker of sorts too, marking out people believed to be of superior moral fber. So this group of people would, at that time and in that place, have counted as a “social group” in my sense. On this understanding of a social group, all of the groups that are marked out by the sorts of traits that our laws commonly treat as prohibited grounds of discrimination constitute “social groups”—for instance, women, Jews, Haidas, and people with hearing-impairments.16 But notice both that we need not assume that members of a social group, so defned, identify closely with each other or that their well-being is in some way bound up with their group identity. Nor should we assume that social groups are homogenous, either in the aspirations of their members, or in their needs or abilities. It may be that one sub-group within a particular social group is afected quite 16  So, of course, do some groups that are not marked out by such traits—for instance, the physically unattractive, and the poor. Whether these traits should be treated as prohibited grounds of discrimination depends, on the view of discrimination that I shall later defend, on whether these groups have been subordinated on the basis of these or correlative traits.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

128

Sophia Moreau

diferently from another by a given policy. For instance, to return to our gendered dress code example, these codes impose one set of burdens on non-religious women, and an additional set of burdens on those women whose religions require them to dress in ways incompatible with the dress codes. So if we are to understand how such codes subordinate women, we may need to look at a variety of diferent sub-groups within this broader social group. We can now go on to consider what it is for a social group, so understood, to be subordinated to others. What does this involve? First, in most situations of social subordination, members of the subordinated social group have less power than members of other groups—not just less political power, but less social power as well, and not just less power in the sense of a diminished capacity to do certain things on their own, but less power in the sense of a diminished capacity to compel others to do what they want them to do. Tere are, of course, difcult questions here about how we are to conceive of power—whether it is relational or can be conceptualized as a kind of resource that could be distributed; whether it makes sense to analyze how much power particular individuals have or whether it must be analyzed structurally and systemically. But I do not think that my argument requires me to take a stance on these questions, so I shall leave them open. Subordinated social groups also generally have less de facto authority than others. Having de facto authority over others is diferent from having power over them. In order to get you into the place that I want you to be in, it is enough for me to have the power to move you there: all I need is a large and strong enough army of helpers and a means of confning you. But I can only get you to do what I want of your own volition if I have de facto authority over you. So de facto authority includes the power to get you, of your own volition, to obey me. It is common within political philosophy to think of de facto authority largely in terms of the power to secure others’ obedience. But within the context of social subordination, I think it is important for us to think of de facto authority as involving a broader set of powers, including the power to be listened to, to be taken seriously when one brings a complaint against another. One fascinating efect of the many recent successful complaints of sexual harassment against prominent flm producers such as Harvey Weinstein and prominent actors such as Bill Cosby is that they have made us collectively aware of one kind of authority that female actors, up until now, have lacked. For before these many successful complaints, some actresses did try, unsuccessfully, to bring complaints of sexual harassment. But many were just not believed: as women, they lacked the authority to speak and to be assumed to be telling the truth, and were too often assumed to have been overreacting in an emotional way or misinterpreting the meaning of men’s actions. In fact, until recently, female actors were in something

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

129

uncomfortably analogous to the position of women in those cultures in which the testimony of two to three women is legally required in order to equal the weight of the testimony of one man. Teir voices did not carry the same credibility as a man’s voice. So it is important not to have too narrow a conception of what de facto authority involves, when we think of the kind of authority that is lacking in cases of subordination. What is lacking is not just the power to get other people to obey you when you issue orders. Before you are even in a position to get other people to obey you, you need to have authority in the sense that other people are ready to listen to you; ready to assume that, absenting evidence to the contrary, you are telling the truth rather than overreacting or misreading events. And that prior authority is the kind of authority that members of subordinated groups quite often lack. Niko Kolodny has helpfully described a further feature of social subordination using the term “consideration.”17 In situations of social subordination, Kolodny argues, the group that possesses more power and de facto authority may be ascribed certain attributes or personal traits that, within that particular society, attract positive responses of deference and respect. Importantly, these responses are directed not just at these attributes or traits, but at the people who possess them: people with certain features are more likely to be shown greater deference and respect, and their interests are likely to be given greater priority, even in situations where they ought to be weighed equally with those of others. Moreover, the traits themselves are not just regarded as pleasing or as important (as, for instance, athleticism is in some circles, or intellectual acuity in others), but as traits that mark people out as in some sense belonging to a higher or better class of people. So when one shows deference to someone on the basis of such a trait, or excludes another person because she lacks it, one is contributing to a pattern of responses that mark some people out as higher, or lower, than others. It seems right, and deeply insightful, that in many cases of subordination, certain traits attract greater deference of just this sort. But I think it’s important to add that subordinated groups aren’t just perceived to lack such traits. Often, the subordinated group is defned in terms of a corresponding trait that comes to be regarded as worthy of censure, because it has been identifed with patterns of action or dispositions of behavior that are perceived as worthless, or worse, as vices. For instance, Muslims living in the United States at the moment don’t just sufer from a lack of deference or consideration, based on perceptions of their religion. Rather, this trait—their religion—is in certain social and political circles regarded as a sign that they are likely either to be terrorists or to be connected with terrorists or, at the very least, 17  See Kolodny, “Rule Over None II,” supra note 16 at 296–7. My analysis of subordination owes much to the basic framework that Kolodny sets up in this article.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

130

Sophia Moreau

to be unpredictable religious extremists. So the trait “being Muslim” functions in certain social circles to mark people out as deserving of condemnation and ostracism. When we think of subordination, then, we should think not just of the absence of consideration toward the disempowered group, but of the use of corresponding traits to condemn, publicly humiliate, or ostracize this group. I shall use the term “censure” to refer to these negative public attitudes. Kolodny writes as though some of the traits ascribed to certain social groups (he names race and lineage as examples) simply happen to attract greater consideration than the corresponding traits of other groups. So, for instance, being of Caucasian or European descent, or being perceived to be white, attracts greater consideration in our societies than being indigenous. For the purposes of Kolodny’s own argument, he does not need to say more than this. But if we are trying to develop a picture of social subordination that helps us understand how discrimination sometimes subordinates, I think it matters very much how certain traits come to attract this greater consideration. For of course, certain races and religions don’t just randomly or arbitrarily attract greater consideration while the corresponding traits of others happen to attract censure. Rather, particular traits come to be associated with dispositions to behave in certain ways, with certain talents or lack of talents, and with certain social roles. And it is through this association that the traits come to acquire greater consideration or greater censure. So, for instance, as I suggested above, in certain social and political circles in the United States, the Muslim religion has come to be associated with religious extremism and with a propensity to engage in terrorist activity. Tese associations of certain traits with particular dispositions, patterns of behavior, and roles are what we commonly call “stereotypes.” Stereotypes, as I understand them here, are generalizations about particular social groups that ascribe most of their members certain desires, dispositions of behavior, or obligations, simply because they possess whatever trait defnes that group, as a group: Muslims are thought more likely to be religious extremists, simply by virtue of being Muslim; women are held to be under an obligation to beautify themselves, because that’s what women are for.18 Some of these generalizations may be false; others may be true, and may still contribute to subordination. What is important about them, for the purposes 18  I have learned much about stereotypes from Rebecca Cook and Simone Cusack’s pioneering discussions of stereotypes and stereotyping in Gender Stereotyping: Transnational Legal Perspectives (Philadelphia, PA: University of Pennsylvania Press, 2010). For my own analysis of the role of stereotypes in cases of discrimination, see Sophia Moreau, “Equality Rights and Stereotypes,” in David Dyzenhaus and Malcolm Torburn (eds.), Philosophical Foundations of Constitutional Law (Oxford: Oxford University Press, 2016), pp. 283–304.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

131

of subordination, is that they serve to rationalize the diferences in the power and de facto authority given to the groups marked out by these traits, and the diferences in the consideration and censure they attract. By “rationalize” I do not mean that they actually justify it, but rather that they constitute the kind of proposed justifcation that is plausible enough that many people in fact accept it. And in some cases, stereotypes seem to work by making us think that there is no need to justify certain ways of treating others: they make the connection between a certain trait and a certain social role or a certain kind of treatment seem so obvious that we feel we do not need to justify placing someone with that trait in that social role or giving her that kind of treatment. So, whether by rationalizing certain acts or by apparently obviating the need to justify them in the frst place, stereotypes play an important role in the persistence of disparities of consideration and censure, and in the perpetuation of unequal power and de facto authority.19 If we were just to stop here—thinking of subordination in terms of disparities in the power and de facto authority held by certain social groups, and in the degree of consideration or censure they attract, based on certain traits—we would omit an important fact about social subordination. Diferences in power and de facto authority are not only held in place by habits of conscious or explicit consideration or censure or by the stereotypes that support such consideration and censure. Perhaps even more importantly—because more silently, and more insidiously—they are kept in place by apparently neutral policies, practices, and physical structures that privilege the interests of the dominant group, while overlooking those of the subordinate group. Particular such structures have been examined by legal scholars working on indirect discrimination, by feminists such as Rae Langton and Catharine MacKinnon working on pornography, and by critical race theorists and disability theorists trying to expose the ways in which apparently neutral policies and political concepts work to perpetuate the privileged status of

19  Must stereotypes be false, to contribute to subordination? I think not. Some of the gender stereotypes that contribute to women’s subordination, for instance, may be statistically true; but they are presented or used as though they are true as a matter of biology, when in fact they refect the ways in which women in certain communities are socialized and they limit the opportunities open to women and the careers that they are able easily to enter and to conceive of as possible for themselves. Similarly, “statistical discrimination”—that is, diferential treatment of diferent groups that results from using group averages to determine which policies are rational—is often based on statistically sound generalizations. What is problematic in these cases isn’t the truth value of the generalizations. It’s the fact that using the generalizations in these ways perpetuates diferences in the power and authority enjoyed by these diferent groups, results in undue deference being given to some and undue censure to others, and renders certain groups invisible in certain contexts.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

132

Sophia Moreau

certain groups and the disadvantaged status of others.20 But no one has, to my knowledge, developed a general philosophical theory of subordination across these diferent contexts that gives a place to these structures; and there is, quite strikingly, no general scholarly term for them. I shall call them “structural accommodations.” Tis term is intended to highlight two important facts about them. First, unlike consideration and censure, they are not attitudes or dispositions of behavior of the discriminator or the public at large: they are real structures in our social and physical environment. In some cases, as we shall see, they are literally physical structures. In other cases, they are structures in the sense that they are policies and practices that structure our workplaces, our homes, and our shared social environment. And second, they work by tacitly accommodating the needs or interests of one group, and overlooking those of others—with the result that the needs of the dominant group come to seem normal and natural, whereas the diferent needs of the subordinate group come to seem exceptional and even odd. Consider frst a very literal example of structural accommodation: certain standard features of the buildings in which we live and work. Most houses have a short fight of steps leading up to the front door, and most storefronts facing onto commercial streets standardly have a single step leading up to the door. Tis easily accommodates those of us arriving on foot, but poses obvious difculties for people in wheelchairs or for those with certain muscular difculties. Light switches are standardly placed four feet above the ground, and bathroom mirrors at a similar height—again, perfectly within reach of many adults, but out of reach, and out of sight, for people in wheelchairs. Tobin Siebers, a disability rights theorist, has written quite movingly about the ways in which such structures not only exclude people with certain disabilities from these spaces, but implicitly send a message about the normal human body who is expected to reside there and the normal guest or client who is welcome there.21 Our houses and our stores presuppose a certain kind of human body and tacitly invite inside those who share such a body, while not issuing invitations to those who do not share it. Tis ideal or normal body remains invisible until someone with a non-standard 20 See, for instance, Kimberlé Crenshaw, “Race, Reform and Retrenchment: Transformation and Legitimation in Antidiscrimination Law,” Harvard Law Review 101(7) (1988), pp. 1331–87. See also Langton, Sexual Solipsism, supra note 16; Catharine MacKinnon, Only Words, supra note 16; Catharine MacKinnon, Towards a Feminist Teory of the State (Cambridge, MA: Harvard University Press, 1991); and Catharine MacKinnon, “Diference and Dominance: On Sex Discrimination,” in Ann Cudd and Robin Andreasen (eds.), Feminist Teory: A Philosophical Anthology (Oxford: Blackwell, 2005), pp. 392–402. 21 Tobin Siebers, “Disability Studies and the Future of Identity Politics,” in Linda Alcof, Michael Hames-Garcia, Satya Mohanty, and Paula Moya (eds.), Identity Politics Reconsidered (London: Palgrave Macmillan, 2006), pp. 10–30. See also Tobin Siebers, Disability Teory (Ann Arbor, MI: University of Michigan Press, 2008).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

133

body appears. When that happens, we might add, it can look as though it is the person with the disability who requires some “special” accommodation. But this is only because houses and storefronts have already been built in such a way as to accommodate the needs of the rest of us. Siebers’ point is not that the construction of houses and stores is an act of deliberate exclusion or deliberate deference to certain body types and censure of others. He recognizes, quite rightly, that houses and storefronts are constructed this way because this answers to the needs of the majority of the adult population. For this reason, I don’t think this particular structural accommodation can be accurately re-described as just another form of “­consideration” in Kolodny’s sense. It is not an instance of conscious deference or respect for people without disabilities. It refects a quite neutral, pragmatic efort to build in a way that is efcient and in demand. But it contributes signifcantly to the subordination of those with disabilities— by making it physically impossible for them to enter into certain buildings, by thereby making it more difcult for them to enter into certain social and commercial relationships, and by making their bodies seem invisible and unnatural.22 Consider next the example of public washrooms that are segregated by gender, with washrooms for men bearing a large sign on the door that represents a man in trousers, and washrooms for women bearing a large sign on the door that represents a woman in a skirt. We are now aware of the ways in which such signs and practices marginalize transgendered persons and place them at greater risk of being taunted or bullied. But at the time when many such washrooms were built, most of the people commissioning them thought that it was normal and natural to segregate people in this way. Tis was not intended as an expression of respect or deference for people whose body fts their gender identity, or of censure for those who are not in this position; yet it has had the efect of normalizing the divide, and of rendering invisible those who do not ft on one side of it or the other. Tis is another example of what I am calling a “structural accommodation”—in this case, a 22  One might object here that the exclusion of people with disabilities from traditional buildings is not a true case of discrimination, so this is not a helpful example. In American law, the “failure to accommodate” certain disabilities or religions is treated as something distinct from wrongful discrimination—both are prohibited, but there are separate bodies of law that govern them. Tis distinction between failures to accommodate and wrongful discrimination has, however, been criticized. Legal scholars such as Samuel Bagenstos, Christine Jolls, and Sharon Rabin-Margalioth have argued that the obligations imposed on us by accommodation requirements are no diferent in kind or degree of onerousness from those imposed on us by anti-discrimination law. See Samuel Bagenstos, “Rational Discrimination, Accommodation and the Politics of (Disability) Civil Rights,” Virginia Law Review 89 (2003), pp. 825–923; Christine Jolls, “Antidiscrimination and Accommodation,” Harvard Law Review 115 (2001), pp. 642–99; and Sharon Rabin-Margalioth, “Antidiscrimination, Accommodation and Universal Mandates—Aren’t Tey All the Same?,” Berkeley Journal of Employment and Labor Law 24 (2003), pp. 111–52.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

134

Sophia Moreau

feature of our built environment that accommodates the needs of the majority and constitutes them as normal, while overlooking the needs of a less privileged social group. I have given two quite literal and physical examples of structural accommodations. But “structural accommodations” in my sense need not actually be physical structures, and they need not function to exclude the subordinated group in quite such a literal way. Tink of the many policies in your own academic faculty or department that accommodate the needs of the average male junior faculty member, while posing some obstacles for young female junior faculty members. A tenure clock that runs out four or fve years after one’s frst appointment is perhaps a good idea for someone whose wife can bear their children; but if you have to bear them yourself, and you have to do so within these particular fve years because you are getting older, it is more difcult. In some departments, faculty meetings run from 4–6 p.m., which means that a woman who has children, and who is responsible for picking them up from day-care, has no choice but to exit the meeting early, in full view of her colleagues, who know exactly where she is going and who sometimes view it as a sign that she isn’t able to be fully attentive to her work. When we invite guest speakers to give a talk, we often take them for drinks (if you were Ronald Dworkin, you would take them for raw oysters) and there is usually an expectation that each of us will have a drink or consume a few raw oysters, partly out of collegiality and partly to demonstrate our sophistication. Tis poses a dilemma for those women who are, or are trying to become, pregnant. Tey may not want to partake, and may not want to disclose why; yet if they don’t partake and don’t ofer some explanation, they appear at best less than collegial, and at worst provincial. None of these practices is designed to disadvantage women. Tey do not seem best described as expressions of greater consideration or deference toward men or censure of women. Tey just happen to accommodate the needs of men who either have no children or have a partner who can bear and take care of them, because this particular social group formed the majority of faculty members at the time that these practices were developed. So they are, in my sense, “structural accommodations”—features of our environment that tacitly accommodate the needs of certain groups, while also normalizing them and rendering the more marginalized group invisible or seemingly exceptional. I should also add that, as the day-care pick-up example shows, the needs that are accommodated do not need to be natural or biological needs: they can be needs that arise because of the social burdens that are placed on one group or another, the way women tend to bear more of the burdens of taking children to and from childcare. I hope I have given enough examples to explain why, in my view, states of social subordination need to be thought of not just as involving diferences

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

135

in power and de facto authority and lesser consideration or censure, but also as involving a variety of structural accommodations that both deny certain opportunities and resources to the subordinated group and serve to render their diferent needs invisible or abnormal. Because they serve this normalizing function, structural accommodations seem to me to stand in a special, supportive relationship to the other features of subordination. Because they help to constitute the needs of the superior group as normal and natural, they serve indirectly to rationalize the diferences in power and de facto authority between these groups and those that are subordinate to them, and also indirectly to provide further support for the various expressions of deference and consideration that are given to these groups in other contexts. If, as the gender-segregation and labeling of public toilets implies, it is normal and natural to be born one gender or the other and to have the gender identity that corresponds to the body you were born with, then those who don’t have this are unnatural—and perhaps they don’t deserve the kind of consideration given to the rest of us. If it is normal and natural for a smart, high-powered academic to produce a book within their frst few years, then it looks as though women who can’t manage this aren’t capable enough to hold power and don’t deserve as much deference. Tere is another reason why structural accommodations help to rationalize diferences in power, de facto authority, and consideration or censure. Tis is that they, just like the patterns of consideration and censure we examined earlier, are bound up with stereotypes about the subordinated group. One such stereotype is that when a woman has young children, she becomes unable to focus on anything except her children; whereas when a man has young children, he is able properly to compartmentalize them and remain a serious scholar. Because of this stereotype, the structural accommodation of holding meetings from 4–6 p.m. has particularly serious efects on women— because it colors our interpretation of what members of the subordinated group are trying to do, when they try to work around this particular accommodation. When a woman walks out of a departmental meeting at 5:50 p.m., she is not just a scholar leaving the meeting early, as her male colleague might be seen to do. She is much more likely to be seen as a mother abandoning her work for her children; and this in turn is often taken as evidence that she must not really have been fully focused on her work, even during the time when she was at the meeting. So the structural accommodation and the stereotype work together to paint her action in a particular light, to reinforce the stereotype, and to rationalize the diferences in power and de facto authority that put the subordinated group in this position to begin with. I have suggested that structural accommodations are bound up with stereotypes, with diferences in power and de facto authority, and with practices

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

136

Sophia Moreau

of censuring certain groups, or giving them less consideration, on the basis of certain traits. But it is worth noting that it is quite possible for a structural accommodation to be innocuous, if it is unconnected with these other features of subordination. So there is nothing inherently objectionable in structural accommodations per se. Tey become implicated in unjust subordination only because, and only to the extent that, they are bound up with stereotypes, diferences in power and de facto authority, and practices of assigning censure and lesser consideration to certain social groups. To see this, consider one structural accommodation: the fact that most stores are open during daylight hours and close at night, rather than being open all night and closed during the day. Tis is convenient for the majority of us, who are awake during the day and who sleep for some part of the night. But it adds hardship to the lives of those employees who work night shifts: if they need to make purchases at stores other than all-night convenience stores, or if they wish to shop together with friends, they have to disrupt their normal sleep time, which is during the day. And this disruption likely afects their bodies more than it would ours, given that their natural sleep rhythms are already disrupted. So the disadvantage they sufer as a result of this policy is more than trivial. It seems also worth noting that this is a group that is already disadvantaged, since night-shift work increases one’s risk of sufering from a host of health problems, such as high blood pressure and metabolic syndrome. So we have here a structural accommodation that imposes more than a trivial disadvantage on an already disadvantaged group. Nevertheless, in certain societies, this structural accommodation would not seem problematic. Suppose that the only people who worked night shifts in a particular society were people in relatively prestigious professions: emergency physicians and nurses at hospitals, lawyers who burned the candle at both ends, judges who were on call all night. Te fact that most stores were only open during the day would not then perpetuate practices of censure toward, or lesser deference toward, these night-shift workers; nor would it support stereotypes about them being less able to handle regular work or less well educated, nor would it perpetuate diferences in social or political power or de facto authority between this group and day workers. In fact, it might even have positive efects on how others viewed the members of this group by adding to the mystique and aura surrounding them: these professionals somehow still manage to get their groceries purchased even though most stores aren’t open while they are at work! By contrast, in a society such as our own, in which many night jobs involve menial labor, require little education, and have much less prestige attached to them—jobs such as janitorial work, cleaning, garbage collecting, security enforcing—and tend overwhelmingly to be held by immigrants who are already mistrusted because of their race, the shared practice of opening stores during daylight

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Discrimination and Subordination

137

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

hours only starts to look more problematic. In order to be implicated in social subordination, then, structural accommodations need to be supported by, and in turn perpetuate, stereotypes, habits of censure and consideration, and diferences in power and de facto authority between diferent social groups. Because structural accommodations, like the diferences in power and authority possessed by diferent social groups, can be innocuous or justifed, there is an important diference between these features of subordination and the expressions of consideration or censure mentioned in condition (ii). Consideration and censure involve taking the praise or criticism that is due to a certain trait and transferring it to the person in a variety of other contexts. So they are always unjustifed. Structural accommodations, by contrast, and diferences in power and de facto authority, and even stereotypes, may sometimes be innocuous. Tey become problematic only when they work together to consign certain social groups to a lesser status in society. Tus far, I have laid out a number of common and morally relevant features of social subordination. I have argued that one social group is unjustly subordinated to another when: (i) Te group has less relative social and political power and less relative de  facto authority over others, and (ii) Te group has or is ascribed traits that attract less consideration or greater censure than the corresponding traits of the empowered group, and (iii) Tese traits are the subject of stereotypes, which help to rationalize the diferences in power and de facto authority, the habits of consideration and censure, and the structural accommodations, and (iv) Tere are structural accommodations in place in society that tacitly accommodate the needs of a dominant group while overlooking the needs of at least some members of the subordinate group; and these accommodations work together with stereotypes to rationalize the diferences in power and de facto authority, and the diferences in consideration or censure. I ofer these four conditions as a set of common and morally salient features of situations involving the subordination of one social group by another, features that, as we will see in Section 4, are relevant in understanding when and why discrimination subordinates. All four conditions will often be satisfed when one social group is subordinated for some substantial period of time. But I do not think we need to suppose that they are individually necessary conditions, nor that they are jointly sufcient conditions. A complete philosophical account of subordination might require additional stipulations. And it seems quite possible that, at certain early stages in the

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

138

Sophia Moreau

subordination of a particular social group, some of these conditions could be satisfed but not others, even though over time all four will likely be satisfed.23 But it does not follow that these features of social subordination are unimportant. Tey are present in most cases of ongoing subordination. And as I shall go on to argue in Section 4, they give us a good basis for understanding how discrimination can sometimes subordinate particular individuals and groups, and why it is wrongful when it does. Tey also help us understand the diferences between the ways in which direct discrimination subordinates and the ways in which indirect discrimination subordinates.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

4.  HOW DIRECT AND INDIRECT DISCRIMINATION SUBORDINATE Now that we have an account of what social subordination involves, we can go on to consider how direct and indirect discrimination might contribute to such subordination. Let us look frst at direct discrimination. Recall that, according to our earlier defnition, a policy directly discriminates against a person, P, if the policy treats P less favorably on the basis of some trait, t, than it would treat those who lacked t. And recall that policies that directly discriminate either explicitly single out people with a certain trait that is a prohibited ground of discrimination, or single them out on the basis of some trait that is very closely connected to such a trait. In order to see how such policies can subordinate certain social groups, it helps to note an important fact about the prohibited grounds of discrimination. We hold, both as a matter of law and in our own moral thought, that not just any trait can constitute a prohibited ground of discrimination. Rather, those traits that are justifably treated 23  For instance, it may be that a certain structural accommodation, initially innocuous, comes gradually to support stereotypes about a certain group that rationalize excluding them from certain prestigious professions, and that over time these stereotypes, combined with the persistence of the structural accommodation, lead in turn to expressions of censure of this group. Or it might be the case that expressions of censure, without any accompanying structural accommodations, lead certain groups to become regarded as so inferior that they are efectively invisible in certain social circles or certain areas of life; and that this in turn nurtures stereotypes about them, and leads to structural accommodations that privilege the needs of others and fail to consider the needs and capacities of this group. In both of these examples, although all four conditions do come to be satisfed over time, there is an interim period in which one or two are not satisfed. During some of that time, we might want to say that there was no unjust subordination. But we might well conclude that there was unjust subordination of the group for some of the time, even in the absence of any censure of them, or even in the absence of structural accommodations excluding them.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

139

as prohibited grounds—race, gender, sexual orientation, and religion, for instance—are traits on the basis of which at least one and often quite a number of social groups have been denied equal power and de facto authority over others; have been subjected to greater censure or lesser consideration, in the sense that they have been condemned or thought of as less worthy of respect than others; have been stereotyped; and have had their needs overlooked by certain structural accommodations that cater to the needs and circumstances of more powerful social groups. To say this is not to claim that in any particular case of direct discrimination, the use of such a trait or its proxy will necessarily perpetuate all of conditions (i) through (iv). But it is highly likely to perpetuate a number of them, given the past history of these traits and the social uses to which they have been put. Consider, as an example, the Jim Crow laws briefy mentioned at the start of this chapter, which left blacks in the United States with separate and inferior schools, hospitals, prisons, washrooms, seating areas in public transit, and even water fountains. Tese laws used the trait “black” in order to accord blacks less power and de facto authority, and they used it in such a way as to ascribe to blacks a variety of undesirable traits, because of their alleged blackness—laziness, stupidity, incivility, uncleanliness, and so on. So they helped to perpetuate the disparities of power and de facto authority between blacks and whites and the stereotypes that held such disparities in place, and they thereby helped to rationalize the many structural accommodations that privileged the needs of whites over blacks. Our account of social subordination, then, helps us to understand that policies that are directly discriminatory against groups marked out by a prohibited ground of discrimination can play an important causal role in sustaining the four conditions of unjust subordination. But there is a second way in which direct discrimination can subordinate. It can also constitute an expression of censure, of the kind mentioned in condition (ii), a statement that a particular group is inferior and can justifably be treated as inferior. During the Jim Crow era, even water fountains were segregated. Te signs above white fountains read “Drinking fountain: Whites only.” Te sign above fountains for blacks read: “Drinking fountain: Colored.” Tese signs did not just function to tell people where to drink, nor did the water fountains just provide water. Perhaps more importantly, they marked out “Colored” as the inferior group. Tey did so partly because the term “only” was attached only to the sign for “Whites,” implying that no one would want to drink from the fountain for “Coloreds” if they were eligible to drink from the “White” fountain. But they also did so through their association with stereotypes such as “Colored people are unclean” and through their association with the many other separate and inferior public facilities which this group was assigned.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

140

Sophia Moreau

Our account of subordination, then, allows us to conceptualize two rather diferent ways in which direct discrimination can subordinate a certain social group. It can (a) play an important causal role in sustaining some or all of the four conditions of subordination. But in addition, it can (b) constitute an expression of censure of the subordinated group, or an expression of lack of deference toward them, a way that marks out this group as inferior. Consider, as another example, the gender-specifc dress codes I mentioned earlier, in Section 1. Recall that these codes explicitly prevent women from wearing certain allegedly “male” uniform options, and require them instead to wear tight, body-ftting, and revealing clothing. Tis perpetuates the stereotype that women are sexual objects without independent agency, and that part of their function, not just as waitresses but as women, is to be beautiful in the eyes of men. It thereby marks women out as inferior. Men, it says, have independent agency and need to dress as such. But women need to dress in such a way as to please men. Unlike the segregated water fountain example, the gendered dress code seems to be less a case of censure and more a case involving lesser consideration. But it is still a case of one social group being branded or stigmatized as inferior to another. What about cases of indirect discrimination or disparate impact? Tese are cases in which neither a prohibited ground of discrimination nor its proxy is explicitly used in the policy itself. But the policy has a disproportionately disadvantageous efect on a certain group because they possess a trait that constitutes a prohibited ground of discrimination. Indirect discrimination can seem puzzling, and its moral status unclear, partly because it is less easily interpreted as the kind of expression of censure or denial of equal consideration that is involved in direct discrimination. But my account of subordination has the resources to explain why indirect discrimination, too, can subordinate people. For my account of subordination focuses not just on the expressions of censure or lesser consideration mentioned in condition (ii), but on the “structural accommodations” in condition (iv) that work tacitly to disadvantage groups marked out by certain traits, and on the stereotypes that rationalize these accommodations and seem to rationalize our not looking for viable alternatives. As I shall now explain, many instances of wrongful indirect discrimination can be seen as structural accommodations—and moreover, as the kind of structural accommodation that is  problematically bound up with stereotypes, diferences in power and de facto authority, and practices of censure and lesser consideration of subordinated groups. As a frst example of wrongful indirect discrimination, consider the cases involving tests for promotion within a certain occupation, such as tests for frefghters or police—tests that do not draw any explicit distinctions along racial lines, but are failed in far greater proportions by blacks and Latinos

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

141

than by whites.24 In some cases of this type, the diferential results are due to prejudicial grading or “buddy systems” and networks of nepotism within the profession that give whites an edge. Tese cases look rather more like direct discrimination. So let us consider those cases in which only the tests themselves are responsible for the diference: the test questions use situations and analogies and bits of information that, in a particular community, whites are more likely to have encountered already. Tis is still, I take it, an example of the kind of indirect discrimination or disparate impact that many would fnd wrongful. My account of subordination allows us to explain why. Tese tests are an instance of a “structural accommodation” inadvertently given to white employees. Tey privilege the interests and knowledge of whites over those of blacks, and even though they do so completely unintentionally and without malice, they nevertheless serve to perpetuate diferences in power and de facto authority, and they work together with stereotypes about blacks (they are so lazy that these results must be accurate; they couldn’t be competent enough to do well on these tests anyway) to rationalize the persistence of these structural accommodations. Tis account of subordination, then, gives us at least one plausible way of understanding how indirect discrimination causally contributes to social subordination. Policies that discriminate indirectly can constitute the kind of structural accommodation that privileges other groups over a given group, reinforces stereotypes about that group, and indirectly rationalizes habits of censure and lesser consideration of them. So indirect discrimination, like direct discrimination, can play an important causal role in sustaining conditions (i) through (iv). It might seem, however, as though there is no analogue in the case of indirect discrimination to the capacity of direct discrimination to contribute to subordination in a further way, by marking out certain groups as inferior. After all, didn’t we see earlier that structural accommodations, unlike expressions of censure and lack of deference, are not inherently problematic? Tey only become problematic through their association with certain stereotypes, diferences in power and de facto authority, and practices of ascribing censure and consideration. Perhaps, on this account, all that we can say about indirect discrimination is that it plays some causal role in sustaining subordination—but, unlike direct discrimination, it does not literally mark out certain groups as inferior. Some scholars might be quite content to claim this. Indeed, many believe that indirect discrimination is signifcantly diferent from direct discrimination. Indirect discrimination is, on their view, either an injustice of a diferent and less serious kind, or it is not an injustice at all, but simply 24  See, for instance, Ricci v. DeStefano, 557 U.S. 557 (2009).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

142

Sophia Moreau

an unfortunate state of afairs for those who are disadvantaged.25 And if you take this view, you might think it is actually quite plausible to suggest that indirect discrimination doesn’t, in fact, mark out certain groups as inferior, but only indirectly contributes to states of afairs in which one social group is socially subordinated to another. But I want to resist this view. I think that indirect discrimination can be just as morally troubling as direct discrimination. And I think that my account of subordination gives us the resources to explain why indirect discrimination, too, does not merely play an indirect causal role in sustaining subordination, but can actually mark out a group as inferior. It is of course true that indirect discrimination does not explicitly classify subordinated groups using the traits that are the basis for lesser consideration or censure of them. But I shall try to argue in what follows that the structural accommodations that are at issue in many unjust cases of indirect discrimination serve in an important way to render subordinated groups invisible, and thereby to mark them out as inferior. To see this, let us turn back to my example of restaurant dress codes and the practices associated with them. One way to understand restaurants’ practices of stocking only uniforms for those who are not pregnant and do not have disabilities is as a “structural accommodation.” It is not done out of prejudice toward pregnant women or people with disabilities: rather, such uniforms are not easily and conveniently available, and most people who apply for waitressing jobs are not pregnant or disabled, so it isn’t economically efcient for restaurants to keep a stock of such uniforms on hand. But this structural accommodation works together with certain stereotypes about pregnant women and people with disabilities (Pregnant women aren’t able to work efciently or to focus on their work! People with disabilities are not beautiful, so who would enjoy being served by them?) to mark them out as inferior. We can say the same about the training manual that only contains pictures of svelte women wearing particularly feminine uniform options. Both the absence of such uniforms and the absence of pictures of people who look diferent and who make diferent choices seem to function to render these groups invisible, and to deny their claim to equal status in no less real and forceful a way than would a sign that read “Pregnant women and people with disabilities: Keep out!” In fact, in an interesting way, the absence of the uniforms and the absence of pictures of these people in the manuals seem to mark them out as inferior even more efectively than a sign would—and 25  See, for instance, Hellman, “Indirect Discrimination and the Duty to Avoid Compounding Injustice,” supra note 14; Eidelson, Discrimination and Disrespect, supra note 6, p. 39; John Gardner, “Liberals and Unlawful Discrimination,” Oxford Journal of Legal Studies 9 (1989), pp. 1–22; and John Gardner, “Discrimination as Injustice,” Oxford Journal of Legal Studies 16(3) (1996), pp. 353–67.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

143

even more efectively than the signs on the water fountains do, in my earlier example of direct discrimination. For a sign at least names the subordinated group and so calls attention to their existence. By contrast, the absence of the uniforms, and the absence of pictures of pregnant women or women wearing non-revealing uniforms, quite literally serves to render them invisible as potential candidates for the job of waitress. Tey simply do not exist in this particular part of our social world—and so neither do their needs. So indirect discrimination, too, can mark out a social group as inferior. It does so by working together with associated stereotypes and habits of censure or lesser consideration to render a group invisible. One might at this point object that there is something paradoxical, and therefore problematic, about my claim that indirect discrimination both renders a group invisible and marks them as inferior. How could a policy really do both of these things?26 In order to mark a group out as inferior, doesn’t a particular policy have to call attention to them in some way, the way that direct discrimination does? Or, otherwise put, if a structural accommodation really did render a group invisible to us, wouldn’t we simply stop seeing them, rather than see them as inferior? But the paradox here is only apparent. In societies ordered by social castes, the lowest caste, such as the Dalit caste in India, is both invisible and inferior. Indeed, the full extent of their inferiority is demonstrated by their invisibility. Although others “see” them in the sense that they see human bodies occupying a certain space, they do not “see” them in the sense of recognizing them as full citizens, capable of participating fully in society and deserving of all of the rights that others are given in that society. Nor do we need examples as extreme as caste systems in order to see that a social group can be at once invisible and marked out as inferior through that very invisibility. Tere is currently an artistic program supported by the Toronto Transit Commission, which commissions local artists to sketch anonymous people riding local subway trains and buses. Many copies of each sketch are made and put up inside the trains and buses, under the heading “Sketching the Line.” Although the majority of people who ride public transportation in Toronto are from visible minorities and many are women, the portraits that have been posted so far consist mostly of white men. Tere is only one drawing of a woman, and she is obviously Caucasian. When I telephoned the Director of the program to discuss the absence of visible minorities and women in these sketches, he replied, softly and slowly but with barely controlled exasperation: “Ma’am, the artists are just drawing what they see.” He was right, in a sense; though wrong in his assumption that this exonerated the program. Te artists were just drawing what they saw. It may seem 26  I am grateful to Cheshire Calhoun for this objection.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

144

Sophia Moreau

incredible that a person could ride a train crowded with women who look Chinese, Haitian, Korean, Anishnawbe, and Pakistani, and yet see only the one white male teenager sitting in the corner. But even though Toronto celebrates its multiculturalism, it is nevertheless true that members of visible minorities, and especially women from visible minorities, are still often not regarded as “normal” or “typical,” even when they outnumber white men in a particular place. So it seems quite likely that the artists just didn’t see them as the normal or representative public transit rider. Tey were, in this sense, invisible; and the “Sketching the Line” program unfortunately perpetuates their invisibility and thereby marks them out as inferior. I have now used my account of social subordination to suggest a number of ways in which direct and indirect discrimination can work to subordinate social groups. Direct discrimination can sometimes constitute an expression of censure toward, or lesser consideration for, a subordinated group, as is mentioned in condition (ii). And it can sometimes causally perpetuate the conditions described in (i), (iii), and (iv), sustaining diferences in power and authority between the subordinated group and more privileged groups, supporting stereotypes that in turn rationalize inferior treatment of the subordinated group, and keeping in place problematic structural accommodations. Indirect discrimination can sometimes contribute to subordination, similarly, by playing an important causal role in sustaining conditions (i) through (iii). And although it does not normally constitute an expression of censure, it can sometimes serve to render certain social groups invisible in certain contexts, thereby marking them out as inferior. One might at this point object that talk of “marking out people as inferior” and “rendering people invisible” sounds very much like talk of the expressive meaning of an act or policy. What is it, really, to “mark out as inferior” or to “render invisible” if it is not to send the message that a particular social group is inferior or invisible? So it might seem as though, in spite of my earlier criticisms of the expressivist view, my account has an important expressivist dimension to it, even though it goes beyond this and looks also at a broader range of efects of discriminatory acts. But although part of what it is to “mark out someone as inferior” is to send a message about this person, I do not think that this idea can be entirely parsed in terms of the social message that a policy sends. Rather, marking someone out as inferior or invisible (or both) involves doing things in the world to that person, altering their situation in certain ways, imposing additional costs on certain opportunities, creating certain disincentives that the rest of us do not have to worry about. When all storefronts have a step leading up to them and this renders Jean and his disability invisible, the step doesn’t just send a social message about Jean. It literally prevents him from accessing the store by himself. It reinforces our shared assumption that the normal shopper is

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Discrimination and Subordination

145

someone who is not in a wheelchair. Tis in turn creates signifcant incentives for Jean not to come forward and ask for the same opportunities as others, for if he were to do so, he would have to present himself as “abnormal.” It seems mistaken to suggest that all of these efects on Jean are simply aspects of the social message that is sent about him. So my account of subordination difers in several ways from the expressivist account. It looks at a broader range of efects of discriminatory acts when assessing whether they subordinate. It considers subordination to be something that happens to a person as a member of a group, because of a socially salient trait that they share with others. And it takes the social meaning of acts and policies to be but one component of a full analysis of the ways in which those acts and policies mark out some people as inferior or render them invisible. My account also enables us to see more clearly the ways in which diferent sub-groups within a particular social group can be subordinated in relation to each other, and subordinated in diferent respects and to diferent extents. On an expressivist account, all acts of discrimination that wrong the members of a particular group wrong them in the same way—by demeaning them as individuals. But my account of social subordination gives us a richer way of describing what is going on, and encourages us to think about the diferences between what happens to one sub-group and what happens to another. Tink back to my example of the training manual. Because the manual depicts only women with svelte fgures wearing conventionally feminine uniform options, it implies that women waitresses are objects of beauty to be enjoyed by men, and so marks out all such women as inferior to men. But, as we saw earlier, it also creates a hierarchy within the class of women, between those who measure up to conventional standards of beauty and those who do not—and it subordinates the latter in a further way. My account of subordination ofers us a rich set of concepts with which to analyze this nested form of subordination, and also with which to analyze the ambiguous position of the women who meet these standards and the precariousness of their status relative to the more subordinated women— which does not seem adequately described simply by saying that they are not demeaned relative to these other women but are demeaned relative to men. My account encourages us to explore the particular ways in which directly and indirectly discriminatory policies reinforce diferent patterns of consideration or censure toward diferent sub-groups, support diferent stereotypes about each sub-group, and rationalize diferences in power and authority between these sub-groups as well as between the group as a whole and other, more privileged groups. In this chapter, I have tried to show that part of the wrongness of discrimination can be traced back to the ways in which certain discriminatory acts

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

146

Sophia Moreau

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

and policies subordinate people. Tis is not, in my view, the only reason why discrimination can be wrong. Discriminatory acts and policies are often challenged as wrongful for other reasons: for instance, because they deny certain people the freedom to shape their lives according to their own values, or because they leave some people without access to the kinds of basic goods and social institutions that one must have access to if one is to be a full and equal member of a certain society. We can accept that discrimination is often wrong because it subordinates some people to others without having to deny that such other facts, too, may play a role in explaining why discrimination is wrong. So the account that I have proposed in this chapter is not ofered as a complete account. On the contrary, it is best understood as one part of a more extensive pluralist account of what makes discrimination wrong.27, 28

27  My own pluralist account of why discrimination is wrong is laid out in more detail in Te Many Faces of Inequality, forthcoming from Oxford University Press. 28  Earlier versions of this chapter were presented at the University of Copenhagen; the U.C.L.A. Legal Teory Workshop; the “Workshop on Political Equality” at N.Y.U. in 2017; the 5th Annual Workshop for Oxford Studies in Political Philosophy; the Gould School of Law’s “Law and Philosophy Workshop”; and the Faculty of Laws, University College London. I am especially grateful to Cheshire Calhoun, Kasper Lippert-Rasmussen, and Samuel Schefer for extensive comments on the chapter, and to Colm O’Cinneide, Barbara Herman, Greg Keating, Niko Kolodny, Jonathan Quong, Daniel Viehof, and Gary Watson for helpful suggestions.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

6

Protecting Vulnerable Languages Te Public Good Argument Alan Patten

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

1. THE PUBLIC GOOD ARGUMENT According to one estimate, a language dies out every 14 days.1 Even languages that are not on the brink of extinction often face marginalization. Tey are used less and less frequently in high-status contexts and are instead pushed into informal and private realms of communication. Faced with these tendencies, many speakers of vulnerable languages favor state policies designed to protect and preserve their languages. My topic is an important but understudied argument in favor of such policies. Te argument I have in mind is advanced by people who observe that the speakers of a language can face a collective action problem. Tey would prefer a situation in which their language is preserved at the cost of some constraints on their individual choices over a situation in which their choices are left unconstrained but the language is not preserved because of free-riding or failures of coordination. Te argument is that, where language preservation problems have this structure, it is permissible for the state to impose the constraints that are needed to preserve the language. Since this argument treats languages as public goods in something like the economist’s sense, I shall call it the “public good argument” for language preservation. Te public good argument is hardly the only argument to be advanced on behalf of preservationist policies. Some advocates of such policies stress the importance of language to “identity,” and ask, “After all, if we’re concerned with identity, then what is more legitimate than one’s aspiration that it never 1

  Russ Rymer, “Vanishing Languages,” National Geographic, July 2012.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

148

Alan Patten

be lost?”2 Others emphasize the connection between language and culture, arguing that individuals depend on a healthy culture for their access to an adequate range of life options.3 Yet another form of preservationism regards languages as intrinsically valuable. Rather like works of art, they are products of human creativity and ingenuity, and they ought not to be destroyed or undermined without a sufciently good reason.4 In my view, each of these other arguments contributes an important perspective to normative debates about language policy. Te connections between language, on the one hand, and identity, culture, and intrinsic value, on the other, point to plausible reasons why people might care about the health and prosperity of their language and/or their particular language community. Except in special circumstances, however, the fair way for public institutions to respond to these reasons is by ofering equal treatment or status to the various languages spoken by the citizens living under those institutions. Rather than elevating some particular language to a dominant ofcial status, public institutions should respond in a proportionate manner to demand for services (e.g. schooling) in a particular linguistic medium.5 Although this regime of equal treatment will sometimes be of great value to speakers of vulnerable languages (who might otherwise fnd their language excluded from public institutions altogether), it may be insufcient to preserve some particular language in the long run. Te equal treatment by public institutions of the several languages spoken in the community does not guarantee that those languages will be equally successful or even that they will all survive. Te inegalitarian pressures of the economy, of global culture, of demographics, and of other determinants of language use may swamp the preservationist tendencies fostered by equal treatment by public institutions. Tis possibility leads some observers to conclude that equal treatment is too formal an approach to language policy. Equal treatment entails giving the same public rights, benefts, and resources to all even when it is predictable that some will race ahead and others lag behind. By treating all languages the same, very unequal outcomes result. For those who fnd this possibility objectionable, it is tempting to emphasize once again the importance of language for identity, culture, and intrinsic value. But, if my earlier comments were on the mark, this simply leads back to a regime of equal treatment. On plausible assumptions, the claims relating to identity, culture, and value 2   Charles Taylor, “Te Politics of Recognition,” in Amy Gutmann ed., Multiculturalism and the Politics of Recognition (Princeton, 2004), p. 40. 3   Will Kymlicka, Liberalism, Community and Culture (Oxford, 1989) and Multicultural Citizenship (Oxford, 1995). 4   David Crystal, Language Death (Cambridge, 2000). 5   Alan Patten, Equal Recognition: Te Moral Foundations of Minority Rights (Princeton, 2014).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Protecting Vulnerable Languages

149

are fairly accommodated by equal treatment itself. Tere is no further right to a successful culture, identity, or realization of intrinsic value that would justify unequal treatment. Evidence for this comes from a comparison of language and religion. Many regard religion as an object of identity, a matter of culture, and a source of intrinsic value. But few would infer that there is a right to a successful religion; at most, these considerations bolster the argument for the equal treatment of religions. Te public good argument is worth considering because it suggests a way of defending further protections for a vulnerable language beyond those that are enjoyed as a result of equal treatment. Te public good argument points to one signifcant diference between the situations of dominant- and vulnerable-language speakers, which is that the latter but not the former face a collective action problem. It is this diference, so the argument goes, that gives the state additional license to intervene in particular ways on behalf of a vulnerable language, even if this means some departures from equality of treatment. Te public good argument also highlights a diference between religion and language. As Jean Laponce once put it, “I don’t have to pray with my neighbor, but I do have to talk with him.”6 Because language involves a level of social coordination that is not generally necessary for religion, it is more vulnerable to collective action problems than religion. Historically, something like the public good argument was an important part of the logic behind Quebec’s language legislation. We shall look at this example in the next section. In recent years, the argument has received renewed attention from several prominent political scientists and philosophers. David Laitin argues that language communities and policies should be considered as consumption goods.7 Groups may try to mobilize to get the state to provide this or that good, and beyond some bedrock liberal principles, the test of legitimacy is simply success in the democratic process. According to Laitin, “communities should be free to provide the public goods its taxpayers demand just so long as fundamental liberal principles are not violated.”8 A key argument of the 2011 book on linguistic justice by Philippe Van Parijs can also be understood as a version of the public good argument. Following the approach developed by Laponce, Van Parijs argues that, because language is a tool for communication, its use always involves a form of coordination between people. When a multilingual person attempts to communicate with some other person or group, she must think not just about her own preferences for language use but, crucially, also about the   Jean Laponce, Langue et territoire (Laval, 1987), p. 143.   David Laitin, Nations, States and Violence (Oxford, 2007), pp. 115–28. 8   Ibid., p. 117. See also Yael Tamir, Liberal Nationalism (Princeton, 1993), p. 54; and David Miller, On Nationality. 6 7

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

150

Alan Patten

language competences of her intended audience. She must look, in other words, for a lingua franca: a language that both she and the members of her intended audience can speak.9 Tis need to coordinate on a lingua franca can lead to collective action problems in which speakers of a vulnerable language fnd it personally convenient and advantageous to learn and use the more dominant language, even though if they could only coordinate their resistance to the dominant language the incentives would be reversed, dominantlanguage speakers would fnd it advantageous to learn the vulnerable language, and the vulnerable language would serve as the lingua franca. Van Parijs fnds a linguistic rights regime based on equal treatment to be objectionable when it applies to a situation in which the speakers of one language are plagued by these collective action problems and the speakers of the other are not. Te equal treatment approach leaves too much to be determined by the disaggregated, uncoordinated decisions of individuals. It is all too predictable that vulnerable languages will fare poorly in this environment, and that speakers of those languages may collectively produce an outcome that they disprefer to the outcome that would ensue if all (or at least most) of them stuck in the critical situations to the use and trans­ mission of their own language. Tere is a role to be played by the state, on this argument, in raising the status of the vulnerable language to the point where it becomes the preferred lingua franca.10 My aim in this chapter is to explore some of the contours of the public good argument. Although I maintain that the equal treatment model ought to be accorded a central place in an account of linguistic justice, I do think there is something to the public good argument and that such an argument can justify some genuine departures from equal treatment that are designed to better protect vulnerable languages. I shall argue against a more skeptical account of the public good argument, which suggests that the argument would hardly ever justify special government protections for vulnerable languages. At the same time, I doubt that the public good argument justifes a signifcant departure from the equal treatment model. Where background conditions are just, there are important limits on the reach of the public good argument. Te account to be developed below helps to identify and justify these limits. 2. LANGUAGE COMMUNITY AS A PUBLIC GOOD Let me start with some examples to motivate the public good lens that is being adopted here. In Montreal, in the 1960s, immigrants were given a   Philippe Van Parijs, Linguistic Justice in Europe and the World (Oxford, 2011).   Ibid., ch. 5.

9

10

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Protecting Vulnerable Languages

151

choice over the linguistic medium in which their children were educated in the public school system. Tere were public schools operating in both French and English, and parents could choose between them. For a variety of reasons, there was a pronounced tendency for immigrants to select an English-language option. Although Francophones represented a majority of more than 60% of Montreal’s population in the 1960s, around 90% of Allophone children (children whose parents were neither native French- nor native English-speakers) were enrolled in English-language schools.11 As the birthrate amongst Francophone Quebecers was also falling at this time, immigrant school choices contributed to a serious anxiety that it was only a matter of time before French-speakers were relegated to minority status in Montreal. With this anxiety in the background, there were renewed eforts to make French-language schools more open and accommodating to new immigrants. And legislation passed in the 1970s (Bills 22 and 101) included measures that compelled Allophone children (with limited exceptions) to attend public school in French. At the same time, another deeper source of anxiety concerned the school choices of Francophone parents. In the 1960s, a choice between French- and English-language public schools was given not just to immigrants and Englishspeakers, but to all Quebecers, including French-speakers. French-speaking parents could decide to educate their children in English schools if they so preferred. And while this was not an option that a great many ever took (in Montreal about 3% of Francophone children attended English-language schools in the 1960s and early 70s),12 there was some reason to fear that an increasing number of Francophone parents would avail themselves of this opportunity. Traditionally, profciency in English was not an essential requirement for employment in sectors where Francophone Quebecers tended to work. But with the economic and social changes of the 50s and 60s, Francophones increasingly sought out positions in white-collar positions in which English was an asset. Tese social and economic changes created an incentive for Francophone parents to try to give their children an extra boost by sending them to English-language schools. Te more that parents responded to this incentive, the more that French would have been pushed into the margins of life in Montreal. On plausible assumptions about the preferences of Francophones, the situation they were beginning to face had the structure of a prisoner’s dilemma (PD). Suppose to begin with that Francophone parents in this situation have two relevant goals. First, they care about the fate of their own linguistic community. Tey want to see it survive and fourish, in part because they   Marc Levine, Reconquest of Montreal (Temple, 1990), p. 56. See also p. 101.  Ibid.

11 12

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

152

Alan Patten

want to have the opportunity for themselves and for their children to use and enjoy the language in a wide range of contexts. Second, they care about the particular social and economic advantages of their own children. Tey want their children to be upwardly mobile, and they recognize that profciency in English is an important asset for this purpose. (Tey have a third goal too, which is that their children learn French to facilitate intra-family commu­ nication in French. But they expect this goal to be realized adequately within the family and thus it is not relevant to schooling choices.) Te PD structure arises when parents have the following preference ordering:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

1. Te French-language community survives and fourishes; they educate their children in English. 2. Te French-language community survives and fourishes; they educate their children in French (with English taught as a subject). 3. Te French-language community does not survive and fourish; they educate their children in English. 4. Te French-language community does not survive and fourish; they educate their children in French (with English taught as a subject). If parents are at liberty to make their own choices about the language of education, then a PD follows from these preferences. Parents will assume that whichever choice they make will have no perceptible impact on the survival or fourishing of their linguistic community. Given that assumption, it is rational for them to choose the option that advances their other relevant goal, namely conferring social and economic advantages on their own children. But once enough of them reason in this way, the cumulative efect is to jeopardize the survival and fourishing of the linguistic community. Here is a second example of the phenomenon, adapted from the Van Parijs argument described earlier. Dutch-speakers in 1950s Belgium are attempting to communicate in an interpersonal setting with both Dutch-speakers and French-speakers. French-speakers have some level of profciency in Dutch, and Dutch-speakers have some profciency in French, but the second-language profciency of Dutch-speakers is generally superior. Dutch-speakers in this context have two relevant goals: the survival and fourishing of their language, and efectively communicating with others. Suppose that the preferences of individual Dutch-speakers are as follows: 1. Teir language community survives and fourishes; and they use French in mixed interpersonal settings with Dutch- and French-speakers. 2. Teir language community survives and fourishes; and they use Dutch in mixed settings.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Protecting Vulnerable Languages

153

3. Teir language community does not survive and fourish; and they use French in mixed settings. 4. Teir language community does not survive and fourish; and they use Dutch in mixed settings. Again the preference ordering gives rise to a PD. Each will reason that their personal choice of language in a given communicative interaction will have no perceptible impact on the security of Dutch. So they will focus on their goal of efective communication, which, given the weaker secondlanguage profciency of French-speakers, implies opting for French. But to the extent that this pattern is repeated over and over again, it will ultimately jeopardize the security of Dutch. Dutch-speakers will have incentives to perfect their French further, while French-speakers will have no corresponding incentive to learn Dutch. One complication is that Dutch-speakers might adopt a more dynamic perspective in their reasoning about language use. Tey might expect that if enough of them insist on using Dutch in mixed settings then local Frenchspeakers will have an incentive to improve their level of profciency in Dutch. If that were to occur, then their frst and second preferences would be reversed, and the structure of the game would now be that of an assurance problem. If Dutch-speakers can be assured that enough from their ranks will insist on the use of Dutch, then there would (eventually) be no trade-of between the personal use of Dutch in a mixed setting and the security of the language. Te problem, however, is that, in the absence of such assurance, French remains the preferred choice for personal communication. It would be possible to construct many similar cases where language choices have these familiar PD- and assurance-game structures. What these cases share in common is that the survival and fourishing of the (local) community of speakers of the vulnerable language is assumed to be a public good. To be sure, the goods associated with a particular language and language community are fully available only to those individuals who are profcient in the language. Tey are goods derived from participating in the language community, which implies at least some minimal level of comprehension and conversational ability. In addition, some of the goods associated with a language community— e.g. much of its arts and entertainment—are accessible only to people who can aford to pay for the relevant cultural goods. In this sense, the benefts of a distinctive language community are not available to everyone; some people are excluded. In spite of these considerations, we can say that a language community is a public good for its members because it is characterized by what economists term “non-excludability”: if it is available for the enjoyment of anyone in the group who satisfes conditions like those mentioned above, then it would

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

154

Alan Patten

be unfeasible or unacceptably costly to exclude others in the group from enjoying the good, even if those others have not made contributions to the cost of supplying the good. For the purposes of the analysis here, the public good status of language communities is not meant as a conceptual or logical claim, but as one that is empirical and normative in character and that is, to some extent, a matter of degree. Te degree to which a good such as a language community ought to be considered public depends on the feasibility of restricting non-contributors from access to the good and on the moral char­ acter of any proposed mechanisms for restricting access (e.g. shunning or shaming non-contributors).13 A standard and well-known problem with many public goods is that they are underprovided, or not provided at all, by decentralized decision-making processes such as the market. Tis problem arises most acutely when a good is produced under conditions that exhibit three features. First, the good is produced (and maintained) by the contributions of many people. Te good is not naturally occurring, nor is it feasibly provided by a single person or a small number of people. Second, the relevant contributions are costly to those who make them. And, third, a contribution by each and every person who will enjoy the good is not strictly necessary for the good to be provided. So long as enough people make costly contributions, the provision of the good will be secured. Public goods with these features are underprovided when and because individuals notice that their particular contributions are costly but unlikely to make a diference to the provision of the good. Vulnerable-language communities ft the profle of public goods that are particularly vulnerable to under-provision. For a language community to survive and fourish, many diferent people need to use the language in a wide range of diferent contexts of communication (frst condition). Although many uses of the language will be costless—two speakers of a language using their shared language to communicate together—some of them may come with a cost (second condition). As we have seen, using the language in the schooling of one’s children, or in a business context, may mean forgoing an advantage that would be derived from the use of a more dominant language. And, while it is necessary for the survival and fourishing of a language that many people contribute, it is not necessary that everyone contribute (third condition). Some number of people could be passive consumers of the language who do not undertake costly eforts to support the language and still the language community might survive and fourish as long as enough other speakers are contributing. To the extent that contributing is costly, and that individuals judge that their particular contribution is unlikely to 13

  For some remarks about shunning, see Van Parijs, Linguistic Justice.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Protecting Vulnerable Languages

155

be pivotal, the language community will be under-supported in a regime where contributions are discretionary. Over time, with all else being equal, one would expect the language to be spoken by a dwindling number of people, and for it to occupy an increasingly marginal position in the society’s patterns of communication. Notice that this analysis need not overlook the importance of language for identity. Te logic of the argument has nothing to do with how or how much speakers of a vulnerable language value their language. Even speakers for whom the language is a core part of identity might be tempted to use a diferent more dominant language in key settings if doing so seems advan­ tageous to them and they do not expect their personal decisions to be consequential for the security of their language.14 To be sure, there are serious questions about how far this analysis, or anything like it, approximates the actual condition of real-world language communities. As I noted above, in the 1960s, when there was choice with respect to language of instruction, only a small minority of Francophone families in Montreal opted for English. Arguably, this outcome is consistent with the analysis just proposed, since the advantages of English may not yet have been salient for many people, and so the costs of sticking to French were relatively negligible. But it does raise the question of whether the public good analysis is more illuminating in theory than in practice. I won’t try to pursue this empirical issue here, but instead will content myself with the thought that if the public good analysis is applicable in actual cases, then an important normative question needs to be considered. Te normative question is this. Does the fact that, in some situations, a language can be considered an underprovided public good justify state intervention on behalf of that language? An “intervention” in this context means a state protection for a vulnerable language that goes beyond a baseline of equal treatment and freedom of linguistic choice. As we have seen, some commentators argue that intervention in this sense is justifable. Laitin suggests it would be justifable if democratically approved by a majority, so long as the intervention would not violate any basic liberal rights or freedoms. Van Parijs maintains that the collective action problems facing vulnerablelanguage speakers support the adoption of the territoriality principle. Tis principle, which was adopted in 1960s Belgium, establishes a territory within 14   Admittedly, linguistic identity could be defned so that possessing such an identity means being committed to using the language even when it is personally advantageous to use another and even when it makes no perceptible diference to the health of the language or language community. Short of this stipulation, however, the argument sketched in the text is compatible with a range of diferent kinds and levels of attachment to one’s frst language.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

156

Alan Patten

the state in which the vulnerable language is in the majority and is designated as the sole public language for ofcial and business purposes.15 Suppose that the stipulated preferences in the education example are universally shared amongst Francophones. With this special assumption granted, many observers would agree that the case for state intervention in the marketplace of decisions about language is very strong. A prohibition on education in English would make everybody better of, raising them from their third most preferred outcome to their second most preferred. It is true that there might be technical difculties in determining who is to be considered a Francophone for the purposes of such an intervention. In addition, the idea of distinguishing Francophones from non-Francophones and applying diferent rules to each might be considered invidious, divisive, or in some way unfair. I shall return to these concerns later on. But setting them aside for the moment, the case for state intervention seems very compelling. Tis analysis is obviously driven by the extreme assumption that all of the people who would be directly impacted by a prohibition—Francophone parents—have the same preference orderings with respect to language and education. It thus abstracts from the most difcult and controversial feature of policies to protect a vulnerable language: the fact that they typically impose costs and burdens on people who do not share the strong commitment to preserving and promoting the language that inspires those policies. If it is allowed that some Francophone parents may be strongly committed to giving their children an English-language education and/or relatively indiferent to the fate of the French-language community, then the argument I’ve just sketched is no longer available. For these parents, preferences 2 and 3 are swapped, and so it is not the case that intervention leaves them better of by their own lights. Tis sort of problem is even more apparent in the example of commu­ nication in a multilingual setting. In this case it cannot be presumed that all who would be directly afected by a state intervention would share the same preference orderings. Te structure of the problem guarantees that there will be winners and losers. Tere needs to be a common standard for all (a lingua franca) but diferent participants have diverging preferences about what that standard should be. Te collective action problem arises for speakers of the 15  Te regime favored by Van Parijs thus rejects equal treatment within linguistic territories. But since each language is to get its own territory, there is a sense in which equal treatment is realized at a higher level in his model. Insofar as some speakers are caught on the “wrong” side of the linguistic frontiers, however, these individuals do not enjoy equal treatment of their language by the public institutions under which they live. So in general I count the territoriality principle as a departure from the baseline of equal treatment, and thus as something that needs a special justifcation such as the public good argument considered in the present chapter.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Protecting Vulnerable Languages

157

weaker language, who are trying to more efectively compete with the stronger language. Tus, even if Dutch-speakers are assumed to have identical preferences, a policy to promote Dutch as the lingua franca in multilingual settings would have an impact on French-speakers, whose preferences will predictably not be the same. What we need, then, is some kind of framework for thinking about the legitimacy of state interventions to provide public goods in circumstances where people in the community have diferent preferences relating to those goods. Much of the rest of this chapter will be devoted to developing such a framework and applying it to the problem of vulnerable language protections. Te framework I shall describe consists of two central principles—the unanimity principle and the principle of correction. Tese two principles pull in opposite directions, but are based on a common value—the idea that each person should enjoy a fair opportunity for self-determination. I shall argue that, overall, the best way of promoting this value is by seeking a compromise between the two principles, a compromise that allows for some departure from unanimity in the provision of public goods. Applying this framework, we shall see that the public good argument does sometimes (but only sometimes) justify state interventions to protect vulnerable languages.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

3. THE UNANIMITY PRINCIPLE Te argument for the unanimity principle presupposes a distinction between two diferent categories of public goods, which I shall call “essential” and “discretionary” public goods. Essential public goods are public goods the support or provision of which by the state can be given a reasonable public justifcation that is independent of the fact that they happen to be valued by some citizens. For instance, education is a public good (insofar as we all beneft from a better-educated population), and state policies to ensure that all children receive an education (e.g. through public schools) can be justifed on grounds such as personal autonomy, equality of opportunity, democratic citizenship, and economic prosperity. Street lighting is a public good that helps to create a safe environment in which people can move around after dark. And so on. If someone says that they do not value or care about public goods of these kinds, there exists a public justifcation that can be ofered to explain why that person should still be expected to contribute to the provision of that good (e.g. by paying taxes). One can debate whether this or that public good ought to be considered essential, and also whether the category of essential public goods is a small one (as libertarians and classical liberals maintain) or a substantial one (as many liberal egalitarians believe). Te important point for our discussion is that essential public goods not be

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

158

Alan Patten

confated with discretionary ones. Te unanimity principle, as we shall see, applies only to discretionary public goods. Discretionary public goods are public goods that are not essential but that are valued by a signifcant number of people. As public goods, these goods are vulnerable to under-provision (for the reasons explored earlier) and so people turn to the state for support or provision. Since state intervention on behalf of discretionary public goods cannot be supported by a reasonable public justifcation of the kind described above (this is what makes them discretionary), all that can be said in favor of intervention is that people value them and they would be underprovided in the absence of intervention. Te unanimity principle was frst formulated by the Swedish economist Knut Wicksell and then defended by John Rawls in A Teory of Justice.16 According to the principle, it is legitimate for the state to support or provide discretionary public goods if and only if taxpayers agree to pay the taxes needed to cover the costs of state action “if not unanimously, then approximately so.”17 Rawls’s argument for the unanimity principle is very simple. He begins with the assumption that a just distribution of income and wealth has been established in some society. If there wasn’t justice in this sense, and state provision of the public good was relevant to achieving it, then presumably there would be a reasonable public justifcation for state intervention, and the good would not be discretionary. Rawls then asks whether further public expenditures on public goods would be legitimate beyond the expenditures that are needed to establish and maintain the just distribution. In response, Rawls claims that, in the context of a just distribution of wealth and income, it would be illegitimate for the state to tax away some of one person’s resources simply in order to subsidize another person’s discre­ tionary expenditures. It follows from this claim, and the assumption that the distribution of resources is actually just, that the question of the legitimacy of public expenditure on discretionary public goods can be reduced to a question of efciency. Taking into account the means of fnancing some specifc public expenditure proposal, would everyone be better of, or at least not worse of, than they would be without such an expenditure? To allow non-efcient expenditures on public goods (i.e. tax/expenditure proposals which leave at least one person worse of ) would, in efect, be to violate the requirement that one person’s justly held share of resources not be taxed away simply in order to subsidize the expenditures of others. As Rawls puts 16  Knut Wicksell, “A New Principle of Just Taxation,” translation by James  M. Buchanan, Richard A. Musgrave, and Alan T. Peacock (eds.), Classics in the Teory of Public Finance (New York: St. Martin’s Press, 1958), pp. 72–118. John Rawls, A Teory of Justice: Revised Edition (Harvard, 1999), pp. 246–50. 17  Rawls, A Teory of Justice, p. 250.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Protecting Vulnerable Languages

159

it, “there is no more justifcation for using the state apparatus to compel some citizens to pay for unwanted benefts than there is to force them to reimburse others for their private expenses.”18 Te underlying idea is that each individual should have a fair opportunity to pursue and fulfll the ambitions that she in fact has—a fair opportunity for self-determination.19 Part of what it means to have such a fair opportunity is to have a fair share of resources at one’s disposal. When the baseline distribution is just, and the state taxes somebody in order to subsidize the expenses of another, then the taxpayer is left without a fair opportunity to pursue her own ambitions. Tis is as much true for taxes to support unwanted discretionary public goods as it is for taxes to subsidize private expenditures by others on discretionary private goods. Reducing the problem of expenditure on discretionary public goods to a problem of achieving an efcient allocation of resources suggests that the unanimity principle is the correct principle to regulate the state provision of such goods. If the means of covering the cost of some proposed expenditure on a discretionary public good cannot be agreed upon unanimously (or “approximately so”), then this must be because some people are declaring that the tax/expenditure proposal would leave them worse of than they were before. It might appear that the application of the unanimity principle would rule out any kind of state intervention on behalf of a public good that is not independently required by justice when not everyone in the community regards the public good as worth paying for. I shall argue that the unanimity principle is too restrictive in certain cases, but it is worth noting that this last suggestion makes it out to be more restrictive than it actually is. Te suggestion presupposes that any costs and burdens that are entailed by state intervention to provide some good must be uniformly imposed on all members of the community. But nothing in the unanimity principle entails this assumption and, indeed, both Rawls and Wicksell seem to have imagined that the most broadly acceptable tax/expenditure proposals would operate according to a beneft criterion whereby those who beneft more shoulder a greater share of the costs and burdens.20 Applying this framework to the protection of vulnerable languages, an initial question is whether language preservation should be considered an essential or a discretionary public good. As noted at the outset, my assumption is that justice mandates equal treatment of languages but does not support 19  Ibid.   For discussion of this idea, see Patten, Equal Recognition, ch. 4.  Rawls, Teory of Justice, p. 250; Wicksell, “A New Principle,” p. 104. Tis point is also made by David Miller, “Justice, Democracy, and Public Goods,” in K. Dowding, R. Goodin, and C. Pateman (eds.), Justice and Democracy: Essays for Brian Barry (Cambridge, 2004), pp. 131–2. 18 20

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

160

Alan Patten

a stronger right or claim to language preservation. If there is a reasonable public justifcation for additional support for vulnerable languages, it is not directly based on justice. David Miller has suggested that there might be an identity justifcation that is indirectly based on justice. According to Miller, there are certain identity goods, including language, architecture, and patterns of landscape, that help to constitute a community. If someone insists that she does not personally beneft from or value these features of her community, Miller thinks she should be reminded that “she benefts from belonging to a political community that is constituted in part by the values in question.”21 If the goods were to diminish or disappear, “the community could not exist in its present form.”22 Te disappearance of the identity goods, Miller adds, might reduce the willingness of these people to support and uphold institutions that realize justice. But this kind of reason for treating language as an essential public good is not very compelling.23 If the point of describing language as an identity good is to highlight the great importance that language preservation can have for speakers of the language, then this consideration is already built into the assumption that justice requires equal treatment of languages. Tere isn’t a further public justifcation for state intervention based on identity in this sense. (Again compare the reasons the state has to intervene in order to preserve struggling religions.) Miller’s suggestion that support for justice-realizing institutions could be eroded by language loss is problematic for diferent reasons. Presumably it is wrong for people to condition their willingness to support justice-realizing institutions on getting their way in language policy. (Imagine a variation on Miller’s scenario in which speakers of the dominant language demand that even less be done for vulnerable languages.) Tere might occasionally be a pragmatic reason to give into unreasonable demands to avoid a greater evil. But there isn’t a general justifcation of language protection policies that emerges from this kind of argument. If language protection is indeed a discretionary public good, and the argument sketched by Rawls is sound, then any policies designed to protect a vulnerable language would have to satisfy the unanimity principle. Tis does not mean that everyone would have to share identical preferences regarding language. We are exploring what happens when this extreme assumption is relaxed. Rather, it means that the policies would have to be devised so that nobody ends up bearing a cost for whom there isn’t an ofsetting or greater 22   Miller, “Justice, Democracy, and Public Goods,” p. 138.  Ibid.   For further discussion of Miller’s view, see Alan Patten, “Public Good Fairness,” in Daniel Butt, Sarah Fine, and Zofa Stemplowska (eds.), Political Philosophy Here and Now: Essays in Honour of David Miller (Oxford University Press, forthcoming). 21

23

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Protecting Vulnerable Languages

161

beneft. Te costs and burdens of such policies must be borne by people whose preferences are such that, all things considered, they are happy taking on those costs and burdens in exchange for the level of linguistic protection that is provided in return. So, for instance, in our education example, the prohibition might be applied only to those families that have the propreservation preferences imagined in the example. For them, but presumably not for families with diferent preferences, the loss of liberty is worth it. Obviously, this is a very limited basis on which to rest a case for state protection of vulnerable languages. As we have seen, what makes such policies controversial in the frst place is the prospect of costs and burdens being imposed on members of the community who do not share the relevant preferences for the protection of the language. Te unanimity principle agrees that this prospect would make such policies objectionable, in efect limiting the set of permissible state interventions to cases where no cost or burden is imposed on people without pro-protection preferences. If the unanimity principle is the only relevant principle in this area, then it follows that state intervention to solve the collective action problem facing Dutch-speakers in our example involving communication amongst multilinguals would not be permitted, since this would impose costs on Francophones who are assumed to lack the relevant preferences. It would also follow that a general Frenchlanguage education requirement in the education case would be impermissible, since we are assuming that some Francophone families lack the relevant preferences. Te unanimity principle would permit a French-language edu­ cation requirement that exempted families who declared themselves to have divergent preferences. Since this is not much of a “requirement,” it is clear that the unanimity principle permits only a modest departure from the equal treatment view of linguistic justice. 4. THE PRINCIPLE OF CORRECTION I noted at the outset that the equal treatment model of linguistic justice was consistent with signifcant disparities in the security of the dominant- and the vulnerable-language communities. Since the unanimity principle ofers at best a weak instrument for further protection of the vulnerable language, it may do little to reduce the possibility of these disparities. But the disparities will likely lead members of the vulnerable community to suggest that there is something unfair about the way they are being treated. We saw in Section 3 that the unanimity principle is based on the idea of fair opportunity for self-determination. If background conditions are otherwise just, people shouldn’t be forced to contribute to the provision of discretionary public goods that others value but they do not. To force people in this way would

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

162

Alan Patten

leave them with less liberty, or fewer resources, with which to pursue their own ends. But perhaps members of the vulnerable linguistic community can claim that the unanimity principle (in conjunction with equal treatment) would deny them fair opportunity for self-determination. It is unfair that dominantlanguage speakers enjoy language security for free, while vulnerable-language speakers have to pay a stif cost for it or fnd it unrealizable at any cost. In response to this charge of unfairness it might be argued that the mere fact that one person’s (or group’s) preferences are costlier to satisfy than another’s does not imply that anyone is denied a fair opportunity for selfdetermination. People with a fondness for beachfront living will generally fnd that their preferences are much costlier than those of people who prefer living in remote inland locations. Tere is nothing unfair about this diference. Te higher price for beachfront properties refects the opportunity cost that others face when somebody ends up the owner of such a property. If others are expected to go without something that they really want, then it is right that the person who does get that thing should have to pay a price for it that refects the level of frustration of the preferences of others. Tat way the others at least don’t have to compete with as much of the owner’s purchasing power when they turn their attention to alternative ways of realizing their preferences. As Ronald Dworkin argues, this opportunity-cost account of fair prices suggests a special connection between fairness and market prices.24 In general, market prices are a gauge of opportunity costs: the more that people would be forgoing in not getting to be the owner of some resource, the more they will bid up the market price. When each person has a fair share of purchasing power, then the allocation of resources through the market leaves each person with a fair opportunity for self-determination, even when some people fnd that the market prices facing their preferred way of life are much higher than the prices facing others with diferent preferences. While this response is efective against some versions of the complaint that vulnerable-language speakers are treated unfairly by equal treatment, it fails against a version of the complaint that stresses the public good character of language protection. Te reason is that the opportunity-cost justifcation of market prices breaks down for public goods. For private goods, market prices do track opportunity costs. But this relationship between market prices and opportunity costs disappears for public goods that are subject to freeriding and/or assurance problems. Depending on the context, the “price” of a public good could be interpreted in several diferent ways. In some contexts, it could mean the “tax-price,” that is the tax that citizens would be charged to enjoy another unit of provision 24

  Ronald Dworkin, Sovereign Virtue (Harvard, 2000).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Protecting Vulnerable Languages

163

of the good. If the unanimity principle is adhered to, then such a tax would have to be voluntary: people who declare that they do not value the public good enough to pay the tax should not be charged. But then the “price” would be driven up, not just by the fact that some citizens genuinely don’t value the public good in question, but also by the fact that some citizens engage in free-riding or are driven to misstate their true preferences because of a lack of assurance. As a result, from the fact that the tax-price of a public good is high, it cannot be inferred that people who claim not to value the public good would be forgoing a lot if it were to be provided. Tey might actually beneft from the provision of the good. Tat is, there might be people who don’t declare that they beneft, either because they are free-riders or because they lack assurance concerning what other potential contributors will do. Tey judge that they would be net benefciaries of a scheme under which the public good is provided at a particular level and they contribute their fair share, but they decide not to contribute (by declaring that they would not beneft) because they think their own contribution would be inconsequential given the contributions made by others (free-rider problem) or because of doubts about whether enough others would contribute (assurance problem). In the absence of a voluntary tax scheme, the “price” of a public good could be interpreted diferently. It could be understood as the amount of the side-payments that individuals would have to pay to other individuals in order to secure provision of an additional unit of the public good. Suppose that, in the education case, if people are left to make their own choices, there would be widespread free-riding and French would (eventually) disappear. Tose who care intensely about French-language preservation realize, however, that if they ofer substantial enough side-payments to potential free-riders, and spend enough on monitoring the behavior of these individuals, then they will be able to secure their language after all. Tey can preserve their language, then, but the price is high. Anglophones, by contrast, are able to preserve their language for free: under a regime of equal treatment, enough people will decide to use English that the language will be secure without anyone having to incur additional costs. Te problem is that this price disparity cannot be justifed on opportunity-cost grounds. Te price of French-language preservation is not high solely because of people who genuinely prefer choice over preservation. It is high in part because of free-riding. By failing to capture the positive externalities enjoyed by the free-riders when the language is protected, the price overstates the costs that are imposed on others by adopting a policy of linguistic protection. Unlike high market prices for private goods—which on Dworkin’s argument track something of ethical signifcance—high prices for public goods deny people with a preference for those goods a fair opportunity for self-determination.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

164

Alan Patten

A very similar analysis can be given of the multilingual communication case. Dutch-speakers might be able to overcome the assurance problem they face by entering into assurance contracts with one another. But these contracts would be very expensive to arrange and to enforce, making language protection a costly proposition for Dutch-speakers. But the high price of Dutch preservation cannot be justifed on opportunity-cost grounds. As with the education case, the price overstates the costs that others would bear if the law were to simply mandate the use of Dutch in public contexts within a particular territory. Tere would be some people for whom such a law is genuinely costly (e.g. French-speakers) but others who would beneft from it even though they would not have independently chosen to use Dutch without assurance about the behavior of others. A fair scheme for pricing public goods like linguistic security would refect the true opportunity costs that devoting resources to those goods, or restricting liberty for their sake, imposes on the rest of the community, once the benefts to free-riders or those in need of assurance are netted out of actually observed prices.25 Te scheme must refect the value to the rest of the community of alternative uses of those resources or liberties. But it must also be attentive to the fact that the provision of a public good does not, in general, beneft only the individuals who would contribute in the absence of state intervention but has positive spillover benefts for those who freeride or lack assurance.26 In other words, if people with preferences for public goods (such as language protection) are to enjoy a fair opportunity for selfdetermination, then some kind of correction must be ofered to the outcome that would arise in the absence of state intervention. Following Ronald Dworkin, I call this the “principle of correction.”27 According to the principle, as Dworkin defnes it, “constraints on freedom of 25  On “true opportunity costs,” see Dworkin, Sovereign Virtue, pp. 156–7. True opportunity costs, Dworkin argues, are the costs that people would face in an auction characterized by perfect information and costless organizational transactions. 26   Te argument here is related to but narrower than the “fair-play” principle defended by H.L.A. Hart and John Rawls. In Hart’s original formulation, “when a number of persons conduct any joint enterprise according to rules and thus restrict their liberty, those who have submitted to these restrictions when required have a right to a similar submission from those who have benefted by their submission” (“Are Tere Any Natural Rights?” Philosophical Review (1955) 64: 174–91, at 185; for Rawls’s development of the principle, see Teory of Justice, p. 96). My notion of true opportunity costs does not suppose that individuals should pay their fair share of any beneft they receive that is produced by cooperative action. True opportunity costs factor in only those benefts for which a voluntary contribution would not be made because of free-riding or an assurance problem— that is, only those cases where the would-be contributor would decline to contribute because of a belief that his or her contribution would be inconsequential and/or because of doubts about whether other benefciaries would contribute. 27  Rawls, Teory of Justice, p. 155.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Protecting Vulnerable Languages

165

choice are required and justifed” if they would help “to achieve a genuinely equal distribution measured by true opportunity costs.”28 Dworkin ofers the example of a group of homeowners who want to maintain a uniform architectural style in their neighborhood. He thinks it would be legitimate for the law to restrict new construction so that it conforms with the generally preferred style. Even if the preference for the uniform style is not unanimous (which would be the case if someone proposed to build in a diferent style), Dworkin thinks that a local ordinance mandating uniformity would be justifed in order to protect others from the spillover costs associated with someone departing from the style.29 Applying the principle of correction to the problem of language protection, we fnd strong support for state intervention on behalf of languages that are vulnerable to collective action problems. Whenever prices are driven up by free-riding or assurance problems, the principle of correction provides a reason for intervention. In the absence of intervention, the vulnerable-language speakers would face prices for preserving their language community that do not refect the true opportunity costs that would be imposed on others by their managing to bring about language preservation. In situations of this kind, many others in fact want the language to be preserved, but are willing to free-ride on the eforts of others or lack assurance about the behavior of others. Tere is a sense, then, in which these prices are too high, and thus those with preservationist preferences would lack a fair opportunity for self-determination.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

5. FINDING A COMPROMISE Te principles formulated by Rawls and Dworkin highlight diferent considerations that need to be acknowledged when examining the state’s responsibilities to support or provide public goods. Te unanimity principle favored by Rawls is grounded in the thought that, in a context where everyone has a fair share of income and wealth, it would be unfair to impose burdens on individuals for the sake of providing public goods that they genuinely do   Ibid., p. 157.   Although he does not say so himself, I believe that the principle underlying Dworkin’s argument here is that each person ought to have a fair opportunity to realize the goals and ambitions that she in fact has. Putting the principle this way helps in fending of possible objections. For example, why does Dworkin’s argument not imply that people with a preference for racial segregation should be assisted by a “whites-only” municipal ordinance? Te answer is that nobody has a claim to anything more than a fair opportunity to realize their preferences. Given what we know about the evils of racial segregation, it could not be claimed that a refusal by the state to help with segregation (or for that matter an active efort by the state to desegregate) would deny anybody a fair opportunity. 28 29

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

166

Alan Patten

not value enough to make the burdens worth bearing. If compelled to contribute, those individuals would be denied a fair opportunity to pursue and realize their own preferences and ambitions. Dworkin’s principle of correction, in turn, is motivated by the idea that individuals should have the opportunity to translate their fair shares of income and wealth into goods needed for their vision of the good life on the basis of prices that are fair. When there are free-riders or problems of assurance, market prices (including voluntary tax rates and side-payments needed to induce free-riders to cooperate) cease to be fair because they overstate the costs that others would bear if the public good were provided. In efect, so long as there are problems of free-riding or assurance, leaving public good provision up to the market, or correcting only when unanimity can be secured, produces a bias against ideas of the good life that presuppose collective action and in favor of more individualistic visions of the good. People with cooperative preferences are denied a fair opportunity for self-determination. Tere is tension, then, between the two principles and thus between the two components of fair opportunity that underlie the two principles. Te tension arises because of free-riding and/or assurance problems. If everyone perfectly revealed their preferences—neither misrepresenting them for personal advantage, nor distrusting that others will act cooperatively if that is needed to secure their top preference—then the gap between the two principles disappears. Public goods could be provided on the basis of the unanimity principle, and nobody could reasonably complain that the resulting prices were unfairly high. With perfect preference revelation, those prices would track the net opportunity costs imposed on others and would thus be fair. Tere may perhaps be situations in which perfect preference revelation is possible. Tey would presumably involve smallish numbers of people, a limited number of decision points, and the capacity of the state to establish procedures that align truthful preference revelation with self-interest. However, the protection of vulnerable languages is not a case in which perfect preference revelation seems realistic. Te cases we are concerned with involve a large number of speakers making a huge number of decisions about language use in areas of life that are often removed from the gaze or the incentives of the state. So in the kinds of cases we are considering the tension between the two principles cannot be dispelled. In general, no unfairness-free policy is likely to be available: one or other of the requirements of fairness is likely to be ofended. Either some people will have to contribute to discretionary public goods that they do not value, or other people will fnd that their ends are more costly or difcult to pursue than they ought to be because of assurance

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Protecting Vulnerable Languages

167

problems and free-riding. Since both requirements of fairness can be traced back to the idea that each person should have a fair opportunity for selfdetermination, there is at least a common metric for thinking about trade-ofs. It would make no sense to recommend a major violation of one requirement for the sake of a minor improvement in the satisfaction of the other. A lop-sided trade-of would result in a net loss of fair opportunity. Instead, the best that can be done is to try to balance the competing demands in a reasonable way. Tis balancing process should be sensitive to several diferent factors. State intervention becomes preferable, 1. the greater the potential for free-riding and assurance problems; and 2. the more consequential that free-riding/assurance problems are for the costs of providing the good that face the individuals who value it and are willing to do their share. State intervention becomes less preferable,

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

3. the greater the number of people who attach little or no value to the public good in question; and 4. the more signifcant the costs that such non-valuers would be made to pay were the state to intervene. Let us see how this framework performs in the context of the examples introduced in Section 2. Running through the four elements of the framework, one could, I think, mount a fairly strong defense of Quebec’s requirement that Francophones send their children to French-language schools: 1. For the reasons outlined in Section 2, the theoretical potential for free-riding in this case seems quite high, even if widespread actual free-riding had not materialized in the years before the language legislation was adopted. 2. If the government were to forgo intervention (instead adopting a policy of equal treatment plus freedom of choice), and free-riding were to be widespread, French could quite quickly fnd itself in a marginal position, especially if other (arguably more controversial) components of Quebec’s language legislation were also to be dropped. 3. Among Quebec Francophones support for the education provisions has always been very high, and there has been relatively little political or legal contestation of those provisions. 4. Te fact that it would be reasonable only to ofer French-language public schooling options if not for the presence of the historic English minority suggests that the contribution that is forced upon Francophones with atypical preferences regarding their language should not be considered very substantial. (To see that this would indeed be reasonable, compare with a 100% Anglophone Canadian province or U.S. state. Few people would insist that such a province or state ought to provide French-language public schooling options just

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

168

Alan Patten

because some Anglophone parents might prefer such an education for their children.) Te balance of the argument seems quite diferent if the restriction to French-language public schooling swept beyond Francophones to include Anglophones as well. 1. On plausible assumptions about their preferences, Anglophones would not be free-riders on the protection of French. 2. Frenchlanguage education for Anglophones is not essential for securing French. 3. Many Anglophones would regard the elimination of English-language public schooling as a signifcant burden. 4. Arguably this reaction would be reasonable given that such a policy would greatly accelerate the demise of the historic Anglophone community in Quebec. As I mentioned earlier, some might fnd objectionable the very idea of distinguishing Francophones from non-Francophones and applying diferent rules to each. Tere are various reasons why this might seem problematic, not all of which I can explore here. It is perhaps worth recalling that, in most liberal democracies, the law makes distinctions between diferent groups of persons all the time. For example, veterans, or farmers, or the disabled, are treated by the law in ways that diverge from the treatment of everybody else. What is important is that diferences in treatment not be arbitrary, but refect genuinely diferent claims or burdens. Te analysis sketched above is intended to suggest that Francophones and Anglophones are in diferent situations. To this extent, the application of diferent rules is not arbitrary and the objection loses some of its force. It might still be objected that there is something especially unfair about diferences in legal treatment based on birth rather than individualized assessment. Tere is something to this objection, but it does not rule out demographic categories if they are a very good proxy for the sorts of individual-level diferences in needs and claims that are appropriately taken into account. If it were discovered that this is not the case, then defenders of Quebec’s legislation would need to reassess. Te other kind of case we have been considering involved communication in multilingual settings. In this case, two groups are competing to establish their language as the lingua franca. Te more dominant language (e.g. French in Belgium) has a default status because it is more widely or profciently spoken as a second language. But speakers of the vulnerable language (e.g. Dutch) would do better if they could coordinate their resistance to French and establish their own language as the common standard. In this scenario, the potential for coordination failures arises within the Dutch-speaking group, but not amongst Francophones, who can be assumed to have dif­ ferent preferences. However, unlike the education case, there is no policy option of limiting the cost imposition associated with promoting Dutch to Dutch-speakers. Promotion in this kind of case consists in giving some advantage to Dutch so that it can more efectively compete against French.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Protecting Vulnerable Languages

169

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

For instance, it might involve requiring that businesses over a certain size conduct their internal operations in Dutch, or that communications amongst professionals (e.g. medical records, legal documents) be written or conducted in Dutch. So it is predictable in this scenario that costs will be imposed on non-free-riders. Still, I think government action may sometimes be justifable in these cases. Suppose to begin with that the pro-Dutch policies are confned to the area of the country where Dutch-speakers are in the majority (i.e. Flanders). Ten, the burden imposed on non-free-riders (i.e. Francophones in our example) is not especially substantial. Te reason for this is that there is considerable utility in having some lingua franca or other in the sorts of situations we are considering (e.g. communication among professionals, within and among large businesses, etc.). In a linguistically diverse society, it is inevitable that some will have to accept that their language will not serve as the common linguistic standard. Te only question is whether this lingua franca will be determined by government policy or as the equilibrium outcome of uncoordinated decisions about language use and acquisition. Given that this burden should fall on somebody, it does not seem objectionable to use government policy to make it fall on the minority rather than the majority. To put this another way, the burden on the Francophone minority of making Dutch the lingua franca is not as great as it initially seems. It is true that they are required to use Dutch in certain kinds of contexts and this is burdensome to them. But they also beneft from the establishment of a lingua franca and so the net burden they face is less severe. In all this, there is a disanalogy with the education case, where there is not the same imperative to coordinate on a single linguistic standard. 6. CONCLUSION In general, then, the public good framework can help to justify certain language protections that go beyond a norm of equal treatment. Tese protections help speakers of the vulnerable language to overcome the potential for free-riding and assurance problems and thus leave them with a fair opportunity to realize their language-related preferences. At the same time, the framework explains why some restrictions that might be defended in terms of the public good argument would not in fact be justifed. Such protections impose excessive costs on non-benefciaries, costs that deny them a fair opportunity to pursue and realize their preferences. Commentators like Laitin and Van Parijs who make the public good argument for language protection thus raise an important point. To insist that linguistic justice is exhausted by equal treatment would be to overlook the

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

170

Alan Patten

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

genuine relevance for justice of the collective action problems facing speakers of vulnerable languages. Te claim that there is a tight analogy between linguistic and religious justice makes the same mistake. Unlike religion, language is rife with potential collective action problems.30 At the same time, Laitin and Van Parijs fail to theorize the limits of the public good argument. Tey don’t acknowledge that the same fundamental reason for thinking that justice requires some state intervention on behalf of underprovided public goods is also a reason for thinking that justice would prohibit such intervention when the benefts are slight or when the costs imposed on non-benefciaries are signifcant. Overall, the equal treatment approach should be qualifed by the public good argument but it remains at the core of how we should think about linguistic justice.

30   I do not dispute that religious believers might sometimes face a collective action problem. Consider, for instance, semi-devout Catholics who strongly value the continued existence of a Catholic cultural community but who also value secular activities that confict with (e.g.) attending Mass. Individuals with these beliefs and preferences might fnd themselves in a PD. Or imagine semi-devout Catholics who especially value the social benefts of religious worship and so are inclined to attend Mass only if they are assured that others with whom they would enjoy social interaction would attend. Still I think there is a diference in kind between the linguistic and religious cases. As religious believers, individuals have good reasons to participate in religious worship regardless of what co-religionists do. Language use, by contrast, is in many contexts largely a matter of coordinating behavior and so is more fundamentally dependent on what others are doing. It is coherent for a religious believer to observe her religion even when others are lapsing. It makes little or no sense for a speaker to insist on using her native language when none of the people she is addressing can understand that language. Moreover, even in cases where religious believers do face a collective action problem, my hunch is that the case for state intervention would generally be weaker than in the linguistic case because the value of allowing people to make their own decisions about religion is greater than the value of letting people make their own decisions about language use. I’m grateful to Kevin Vallier for correspondence about this issue.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

7

Localized Restricted Aggregation

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Victor Tadros

Restricted aggregationism is a general view in deontic theory that interpersonal aggregation of harm makes a diference to what a person should do in some but not all cases. Tere is a related view in axiological theory that is concerned with the relationship between aggregation and the goodness of states of afairs, but that is not our topic. To understand restricted aggregationism, focus on my decision which of two groups to save from harm, {A} and {B}, where I cannot save both. If the harm that members of {A} will sufer is signifcantly graver than the harm that members of {B} will sufer, and these groups are equal in size, and other things are equal, I should save {A}. Now suppose that {B} is larger (i.e. has more members) than {A}. Standard restricted aggregationist views imply that in some cases I should rescue {B} if {B} is sufciently large. In other cases I should rescue {A} no matter how large {B} is. On standard versions of the view, the frst judgment is true where the diference in the threat of harm faced by members of {A} and {B} is sufciently small and the second judgment is true where it is sufciently large. Restricted aggregationist views contrast with both pure aggregationist and pure anti-aggregationist views, and have at least some implications that are more intuitive than either alternative. To see this, consider: Case 1:  members of {A} will be killed if they are not saved whereas members of {B} will sufer severe injury. Contrary to pure anti-aggregationist views, it seems that I ought to save {B} if {B} is much larger than {A}. Restricted aggregationist and pure aggregationist views have this implication. Now consider: Case 2:  members of {A} will be killed if they are not saved whereas members of {B} will sufer only a broken fnger.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

172

Victor Tadros

Contrary to pure aggregationist views, it seems that I ought to save {A} no matter how much larger {B} is than {A}. Restricted aggregationist views and pure anti-aggregationist views have this implication. Tus of the three views, only restricted aggregationist views have intuitive implications in both cases, and this is some reason to believe that some such view is true. Of the other two views pure aggregationism seems more plausible. Te intuition that I ought to rescue {B} if {B} is very much larger than {A} is extremely powerful. And there are debunking explanations for our antiaggregationist intuitions in Case 2 that I will explore below. I will thus compare the merits of restricted aggregationism and pure aggregationism. My aim is to ofer a general theoretical foundation for thinking about restricted aggregationism, to set out a particular version of this view in the light of it, and to defend it both against challenges to restricted aggrega­ tionism in general, and to this view in particular. As the particular view I defend is difcult to summarize without some background, I will not do so here. I do not show that restricted aggregationism is true, only that it is reasonably plausible. Any theory comes with a price tag.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

I.  OUTWEIGHING AND DISABLING A natural, but rough, way to explain our contrasting intuitions in Case 1 and Case 2 is that some harms are too trivial to make a diference to my duty to save another person from a very serious harm such as death, whereas others are not. If some harm is too trivial to make a diference, an aggregation of people who will sufer these harms cannot make a diference either. Let us make this more precise. Deontic propositions are true when certain facts that are capable of grounding them obtain. A fact that is capable of grounding a duty is a duty-grounding type of fact. Such facts are typically facts about conduct. For example, the fact that my ving will prevent you from being harmed is a duty-grounding type of fact—it is the type of fact that can ground a duty on me to v. If I have the duty to v, the fact that my ving will prevent you from being harmed grounds my duty to v. Similarly a fact that is capable of grounding a permission is a permission-grounding type of fact. For example, the fact that my ving will cost me my arm is a permission-grounding type of fact—it is the type of fact that can ground my permission not to v. If I have the permission not to v, where I would be required to do so were it not for the fact that doing so would cost me my arm, the fact that my ving will cost me my arm grounds my permission not to v. A fact may obtain, and be a duty-grounding type of fact, without grounding any duties. In the simplest case duty-grounding and permissiongrounding types of fact confict. For example, suppose that I can rescue

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

173

either A from death or B from severe injury. Preventing harm to B is the kind of fact that can ground duties, but I lack the duty to prevent harm to B because then I cannot save A, whose plight is even worse. My duties and permissions depend on a resolution of conficts between duty-grounding and permission-grounding types of facts. When a duty-grounding type of fact obtains, but does not ground a duty, it is defeated. Now contrast two ways in which a permission-grounding fact might defeat a duty-grounding fact: either by outweighing or disabling its duty-grounding force.1 In the case we have just considered the permission-grounding force of saving A outweighs the duty-grounding force of saving B. But permissiongrounding facts sometimes disable the force of duty-grounding facts, rather than merely outweighing that force. To see that this is intuitive, compare the role of consent in a doctor’s decision whether to operate on a patient to cure a debilitating illness. If the patient consents, let us suppose, the doctor has a duty to operate but he must not do so if she does not consent. Facts about the illness are duty grounding. Tis is obviously true because a doctor is not required to operate on a patient who consents to the operation when she is not ill. Tus the fact that her illness can be cured grounds the duty on the doctor to operate on her, and not simply the fact that she consents. If the patient does not consent, her lack of consent is permissiongrounding—it makes it true that the doctor lacks a duty that he would otherwise have. But this is not because her lack of consent outweighs the duty-grounding force of curing the illness. Without consent, these facts lack duty-grounding force altogether. Tus facts about saving the person from the illness are of a duty-grounding type—they are the kind of thing that can ground a duty to save the person from illness. But they do not have duty-grounding force in this case. Tus, the doctor’s permission (and duty) not to treat the non-consenting patient is not explained by the fact that his reason to cure the illness is outweighed by non-consent. Here is an explanation why the illness lacks duty-grounding force if the patient does not consent: the fact that she is an autonomous agent makes it appropriate that she controls whether facts about her welfare have dutygrounding force, because it is valuable for autonomous agents not only to be in control of what happens to them, but also in control of the way in which their welfare fgures in the practical reasoning of others. Respect for 1   We fnd the same contrast in the case of reasons. See especially J.  Dancy, Ethics Without Principles (Oxford: OUP, 2004) ch. 3. Dancy explores this view as part of his more radical particularist view about morality, which rejects moral principles altogether. Tis, though, is not an implication of accepting this contrast. For discussion, see S. McKeever and M. Ridge, Principled Ethics: Generalism as a Regulative Ideal (Oxford: OUP, 2006).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

174

Victor Tadros

autonomy is thus an attitude that we have when we appropriately respond to the autonomy of others by deferring to their decisions about how their welfare fgures in our practical reasoning. Autonomy thus sets the appro­ priateness conditions for the welfare of others having duty-grounding force. Appropriateness conditions of this kind determine whether some valuable fact, such as the person being cured, has duty-grounding force. Where they lack this signifcance, they still have other evaluative properties. For example, even though her welfare is not duty-grounding, it is still good for her to be cured. Tus, where the patient does not consent, it is wrong for the doctor to perform an operation that is good for her. Tis is one way in which deontic and axiological views about aggregation might come apart—the fact that it is better that trivial harms do not occur need not imply that these harms have deontic signifcance in a decision whether to prevent them, or to prevent much graver harms to others. Te idea that appropriateness conditions determine when a dutygrounding fact has duty-grounding force is intuitive in other cases not involving consent. For example, suppose that I am at a funeral. Te person next to me has a rash, and I have some cream that could cure it. Te fact that I could give her the means of curing the rash is a duty-grounding type of fact—were I not at a funeral I would have a duty to ofer her the cream because of it. It is plausible that the duty-grounding force of curing the rash is disabled rather than outweighed in this case by the inappropriateness of  ofering her the cream at a funeral. Again, this is a case where it is inappropriate to do something good for the person, and thus that the goodness of curing the rash is not duty grounding. Or consider blaming a person for committing a wrong that I have entrapped her to commit. It may be good to blame the person, because it would help her to respond appropriately to her wrongdoing. And had I not have entrapped her I might have had the duty to blame her for this reason. But this fact may lack duty-grounding force because I lack the standing to blame her. Of course, not everyone will fnd all of these cases compelling. I only present them as plausible illustrations of the idea that the dutygrounding force of some facts can be disabled. II.  APPROPRIATENESS, HARM, AND AGGREGATION Restricted aggregationists may make a similar claim about the permission not to save a person from trivial harm where doing so is incompatible with saving another person from very serious harm.2 Recall Case 1. Suppose that 2   Some rely on a related contrast between whether claims are outweighed or canceled by competing claims. See, for example, A. Voorhoeve, “How Should We Aggregate

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

175

{B} is much larger than {A}, and I thus have a duty to rescue {B}. Te aggregate duty-grounding force provided by preventing harm to the members of {B} outweighs the aggregate duty-grounding force provided by preventing harm to the members of {A}. Terefore, I am required to rescue {B}. Contrast Case 2. Te duty-grounding force of preventing harm to members of {A} disables the duty-grounding force of preventing harm to each member of {B}. Tere is thus no force of the latter kind to aggregate and I am required to rescue {A} no matter how large {B} is. Tis assumes that the facts that disable the duty-grounding force of preventing harm to one member of {B} disable their duty-grounding force in the aggregate. Tis need not follow, but it is plausible in this case, which explains why it seems wrong to save {B} no matter its size. Tis idea draws on the fact that appropriateness conditions deprive preventing the broken fnger of duty-grounding force. Suppose that I can rescue either A from death or B from a broken fnger. Te broken fnger is a duty-grounding type of fact—if I could prevent B sufering the broken fnger at no cost to anyone, I would be required to do this. But (roughly speaking) it is inappropriate that this consideration has duty-grounding force where A’s life is at stake. Tis idea naturally explains some further judgments about Case 2—for example, why it is not only inappropriate for B to demand that she be saved, but also why it is inappropriate for B to demand that her broken fnger be taken into account in the decision whether to save A.  Such a demand is inappropriate because it implies that her broken fnger is the kind of thing that weighs against the duty to rescue a person from death, and it does not. Furthermore, making this demand seems inappropriate because it would be disrespectful to A to make this demand, given A’s plight. But this is only disrespectful if it is inappropriate for me to judge that B’s broken fnger counts against my deciding to save A. And this is inappropriate only if B’s broken fnger lacks duty-grounding force. We can now meet a familiar argument ofered by pure aggregationists. Tey typically accept that it is intuitive that I ought to save {A} in Case 2 no matter how large {B} is. Tey respond with debunking explanations of that intuition, the most familiar of which is our putative inability frmly to grasp harm to large numbers.3 Tis debunking explanation, though, is not very Competing Claims?” (2014) 125 Ethics 64. Tis is related to the more general contrast between outweighing and disabling. I prefer to avoid the language of claims, as that language is most comfortably used in rights-based or contractualist theories. 3   See, for example, M. Huemer, “In Defence of Repugnance” (2008) 117 Mind 899; J. Broome, Weighing Lives (Oxford: OUP, 2004) 56–7; A. Norcross, “Comparing Harms: Headaches and Human Lives” (1997) 26 Philosophy and Public Afairs 135; J. Horton, “Aggregation, Complaints, and Risk” (2017) 45 Philosophy and Public Afairs 54. Voorhoeve

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

176

Victor Tadros

plausible. We not only have intuitions about what I should do, but also intuitions about whether some consideration is relevant to what I should do, so that it might make a diference in the aggregate. We have the intuition that I ought to save {A} no matter how large {B} is because it follows from the more basic intuition that the harm that {B} will sufer is irrelevant given the plight of members of {A}.4 In the light of this, a better debunking explanation is that ignoring minor harms leads to better and more efcient decision-making—when faced with the opportunity to save a person from being killed, it is often safe to ignore minor harms to others on the grounds that it is normally impossible to establish that enough people will be saved from minor harms to justify refraining from saving a person’s life given epistemic and time constraints on decision-making. Our intuitions that minor harms are irrelevant might be explained by the fact that this heuristic device for determining what to do tends to lead to good decisions given that we are fallible decision-makers even if, in principle, interpersonal aggregation is always warranted in determining what to do. To use Joseph Raz’s term, these harms are thus excluded reasons.5 Tis might also explain why it is wrong for members of {B} to demand that their broken fngers should be taken into account in my decision whether to save {A}: they ought not to do so, because their doing so tends to make me go wrong. But though this is an improvement, it is also unsatisfying, for this explanation suggests that the broken fnger does provide me with a reason for action, though one that I should exclude from my deliberations in practice. And that does not capture the more basic intuition that it is wrong for me to take the broken fnger into account because doing so is inappropriate given the plight faced by members of {A}. Tis draws on the general lesson that in refective equilibrium, we consider both intuitions about the implications of theories, but also intuitions about the explanations of those intuitions. I have relied on the idea that the harm that {B} sufer lacks dutygrounding force because certain appropriateness conditions are not satisfed. Can we provide a reductive analysis of these conditions? I doubt it. Alex Voorhoeve makes the best attempt to do something like this. To reconstruct his view slightly, he argues that a person’s claim to be saved is disabled by the requirement not to prefer the state of afairs where they are saved rather than others. Tus he thinks that there is a relationship between personal suggests that it is difcult to fnd conclusive evidence for this explanation in “Balancing Small Against Large Burdens.” 4   See also Voorhoeve, “How Should We Aggregate Competing Claims?” 75–6. 5   See J. Raz, Practical Reasons and Norms (Oxford: OUP, 1999) ch. 1.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

177

prerogatives and the canceling of claims, where the idea that claims are canceled is similar to the idea that facts that ground those claims lack dutygrounding force.6 Tis draws on the more general idea that we have limited agent-relative prerogatives to value our own welfare over the welfare of others, and to take that into consideration in deciding what to do. Tis idea has some appeal, but I doubt that it is right. First, it is hard to explain the relationship between personal prerogatives and the canceling of claims. Suppose that I am required to rescue a person from death even if this will cost me a broken fnger. Te fact that I will break my fnger surely weighs against my being required to save the person. It is a consideration that is outweighed rather than canceled in determining what I am required to do. But if so, it is hard to see why it is canceled rather than defeated in the decisions that others make about whether to save a person from death or save me from sufering a broken fnger. Second, although the signifcance of a person’s duty to bear a certain cost correlates closely with the harm that it is permissible to infict on a person in the course of using that person for the sake of some goal, it does not correlate closely with the harm that it is permissible to infict on a person where the person is not used. It is often permissible to infict greater costs on others than they would be required to bear as a side-efect of saving others.7 As we are not concerned especially with using cases in this context, there is less reason to think that the duties that a person has to bear costs are relevant in determining what it is permissible to do for the sake of others. Tird, the fact that one person has a duty to bear a cost for the sake of some goal does not imply that many people together are each required to bear the same cost for the sake of that goal. For example, it may be valuable to the people in my town that a bridge is built over a river. If I alone could build the bridge in a day, it is plausible that I would be required to do this, even if this bridge does not beneft me. It does not follow that any number of people must together work for a day to build a bridge that will not beneft any of them. Some might think that this is so because there is a group of people who are the benefciaries of the bridge being built. But the point generalizes beyond this case. For example, suppose that I am watching television and I see a person struggling to get his shopping home. I might be required to stop watching and help. It does not follow that any number of people who   See Voorhoeve, “How Should We Aggregate Competing Claims?”.   See, further, V. Tadros, Te Ends of Harm: Te Moral Foundations of Criminal Law (Oxford: OUP, 2011) ch. 6; A. Walen and D. Wasserman, “Agents, Impartiality, and the Priority of Claims Over Duties: Diagnosing Why Tomson Still Gets the Trolley Problem Wrong by Appeal to the ‘Mechanics of Claims’” (2012) 9 Journal of Moral Philosophy 545. 6 7

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

178

Victor Tadros

are watching television must all stop to help if that is the only way to relieve the person’s burden. Overall, I think that the contextual interaction which determines whether a duty-grounding type of fact has duty-grounding force in a particular case is morally basic in just the same way that it is basic that duties are grounded in duty-grounding facts. We should not attempt a reductive explanation of appropriateness. Rather we should explain it by showing that restricted aggregationism is a plausible instance of the more general phenomenon of contextual interaction in the grounding of reasons and duties, as I did earlier.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

III.  LOCAL RELEVANCE I have so far been relying on the rough idea that major harms, such as death, disable the duty-grounding force of minor harms, such as a broken fnger. Tough this has not received much attention in the literature, diferent restricted aggregationist views difer about the purposes for which the duty-grounding force of a claim is disabled. For example, it is plausible that saving B from a broken fnger lacks duty-grounding force when it would otherwise weigh against the decision to save A’s life, but not when it would otherwise weigh in favor of saving A’s life. Te fact that saving A’s life will also save B’s fnger from being broken is an extra reason to do so. Furthermore, suppose that {A} contains one person facing a threat of death and many people facing a threat of a broken fnger, whereas {B} contains many people facing a threat of severe injury. Suppose that death disables the duty-grounding force of a broken fnger, but serious injury does not. Te broken fngers in {A} have duty-grounding force, in that they can ground a duty to save {A}, even if the death in {A} would be insufcient on its own to outweigh the threats to severe injury in {B}.8 Tis suggests that a better way of understanding Case 2 is that a broken fnger cannot ground a permission not to save a person’s life. But even this may not be accurate. Suppose that {A} contains one person facing a threat of death whereas {B} contains many people facing a threat of severe injury, but the number is just too small to outweigh the signifcance of the death, and so I ought to save {A}. Now add billions of people to {B} whose fngers will be broken if I save {A}. It is plausible that the threats added to {B} can tip the balance, given that the signifcance of the death has been almost completely counterbalanced by facts in {B} that weigh against it.9 8   For decisive arguments for the views in this and the previous paragraph, see P. Tomlin, “On Limited Aggregation” (2017) 45 Philosophy and Public Afairs 232, 244–7. 9   I am especially grateful to Larry Temkin for discussion here.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

179

So  strictly speaking, it is not true that threats of broken fngers cannot ground a permission not to save a person from death. Rather, they can­ not do so without being added to some other fact that appropriately counterbalances the signifcance of death. Here is a further distinction between views that will occupy us a bit more. On one version of restricted aggregationism, when saving B’s broken is inconsistent with saving A from death, B’s broken fnger lacks dutygrounding force in the overall decision whether to perform an act that saves A from death. Call this view “Global Relevance.” On an alternative view, B’s  broken fnger lacks duty-grounding force in counterbalancing the signifcance of A’s death, but may retain such force for other aspects of the overall decision whether to save A.  For example, it may counterbalance other duty-grounding facts. Call this view “Local Relevance.” To see the contrast between these views more clearly, and why Local Relevance is the better view, consider Patrick Tomlin’s case:10

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Adding People: Stage 1:  I can save {A} or {B} but not both. {A} has one dying member (X). {B} has ten members, each facing severe injury. Stage 2:  10,000 people facing threats of broken fngers are added to both groups. Let us plausibly suppose that Global Relevance requires me to save {B} at Stage 1 because X’s claim to be saved does not cancel the claims of the ten to be saved. As Tomlin rightly argues, it is not plausible that things are diferent at Stage 2, but if we assume, as I have argued that we should, that broken fngers can provide reasons in favor of saving a person from death, Global Relevance implies otherwise. For X’s death disables the duty-grounding force of the broken fngers of the 10,000 in {B} whereas the severe injury of those in {B} does not disable the duty-grounding force of the broken fngers of the 10,000 in {A}. Tus Global Relevance implausibly implies that I am required to decide diferently, and to save {A} at Stage 2. Tis reveals a general problem with Global Relevance.11 But Local Relevance does not face this problem. On that view, the right question to ask about the 10,000 people that are added to {B} is not whether their injuries lack duty-grounding force altogether, but rather whether they weigh against the duty-grounding force of those facing death in {A}. If their 10   “On Limited Aggregation,” 240. I have adjusted the numbers and kinds of harm, but the case is identical in what it shows. 11  Tomlin considers versions of Global Relevance that do not have the problematic implication, but he powerfully demonstrates that these versions have even worse problems.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

180

Victor Tadros

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

signifcance is disabled for this purpose, they may nevertheless play a role in outweighing other considerations that they compete with. Consider the people who are added to {B} at Stage 2 who face the prospect of a broken fnger. Tese injuries lack duty-grounding force for the purpose of assessing whether X’s death decisively grounds a duty to save {A}. But that does not show that they are irrelevant in the overall assessment of whether I am required to save {A} or {B}. Tey are sufciently close (indeed they are identical) to injuries of those added to {A}. Tey thus counterbalance the signifcance of those injuries and I ought to choose in the same way at Stage 2 as at Stage 1. Tis view is also theoretically attractive. Te prospect of a person facing a broken fnger seems irrelevant when deciding whether the prospect of a person’s death compels us to save her, at least if nothing more signifcant has partially or completely counterbalanced the threat of death. But it is relevant in an assessment whether other considerations for and against performing the life-saving act are defeated. It would not show inappropriate regard for the signifcance of X’s plight to treat the claims of those facing a broken fnger as having duty-grounding force for the purpose of counterbalancing signifcance of similar prospects faced by members of {A}. Indeed, it is appropriate to do this. For suppose that this is not true. If so, it is not the fact that X’s plight is so grave that makes the decisive diference between Stage 1 and Stage 2; it is the fact that those facing the prospect of broken fngers have been added to X’s group. It is difcult to see how to argue for the view that it is appropriate to weigh the broken fngers in {A} whilst denying that these broken fngers cannot be counterbalanced by the broken fngers in {B}. Tus, Local Relevance is better than Global Relevance. IV.  ADDING INTERMEDIATE GROUPS Here is a related objection to restricted aggregationist views, but one that seems to count against Local Relevance. Te general thrust of the objection is as follows. Suppose that we must save one group of people over another because the duty-grounding force of the plight of those in the second group is disabled by the plight of members of the frst group. When equal numbers of people sufering an intermediate threat are added to each side, we can be required to switch between groups. Indeed, we can sometimes be required to switch between groups even when a smaller number of people facing an intermediate threat are added to the second group than the frst. Tis may seem counterintuitive.12 12   A version of this objection was developed by Joe Horton in his excellent “Always Aggregate” (2018) 46 Philosophy and Public Afairs 160 in response to an earlier draft of

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

181

Here is a forceful illustration of this objection:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Adding People II: Stage 1:  I can save {A} or {B} but not both. {A} contains one member, X, who faces death. {B} contains 1 billion members, each facing a broken fnger. Stage 2:  10,000 people facing severe injury are added to {A} whereas ten people facing severe injury are added to {B}. Here is the puzzle. According to restricted aggregationist views, I ought to save {A} at Stage 1. But this may not be so at Stage 2. At Stage 2, the dutygrounding force of the severe injuries is not disabled by the fact that X’s life is at stake, and these injuries are sufcient to outweigh death, we have been assuming. Furthermore, the duty-grounding force of the broken fngers is not disabled by the fact that 10,000 people sufer severe injury in {A}, and 1 billion broken fngers, let us suppose, outweighs 10,000 severe injuries. Tus it seems that I should save {B} at Stage 2. But then, strikingly, it seems that I should switch from {A} to {B} at Stage 2, even though far fewer people have been added to {B} than {A} at Stage 2, where each person added faces a threat of the same harm. Local Relevance would not have this implication if we understand that view in the following way: the duty-grounding force of the ten added to {B} is counterbalanced by the duty-grounding force of ten of the 10,000 added to {A}. Tis leaves the signifcance of X’s death untouched, and X’s death is decisive, given that the only competing facts that remain are the threat to the fngers of the 1 billion, and the duty-grounding force of these facts is disabled. But even if this response has intuitive implications,13 it seems unprincipled for this reason. Te basic intuition that makes restricted aggregationist views initially attractive is that certain facts that can sometimes ground duties seem too trivial to count against our duty to save others from very serious harm. Considering these trivial facts seems to give rise to an inappropriate justifcation of the failure to save people from serious harm. But no such inappropriateness is involved in justifying the decision to save {B} at Stage 2. I justify the decision not to save X by pointing to the value of the severe injuries that the ten will otherwise sufer in {B}, and I justify the decision not to save the 10,000 by pointing to the value of the broken fngers that the 1 billion will otherwise sufer in {B}. So I have a complete and appropriate justifcation for saving {B} over {A}. this chapter—I thought it best to respond to that here to allow readers to fully evaluate the views without hunting for further papers. I think I have both strengthened and simplifed the objection here. 13   And Horton, in “Always Aggregate,” shows that it fares no better.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

182

Victor Tadros

If an appropriate justifcation for saving some group over some other group can be given, the fact that diferent reasoning in favor of saving that group would be inappropriate just seems irrelevant. In other words, we should give things of value the weight in our practical reasoning that their importance warrants where this can be done appropriately. As the fngers of the 1 billion can be given weight in our practical reasoning without implying that their importance could be sufcient on their own to outweigh the signifcance of X’s death, they should be considered when determining whether to save {A} or {B}. Terefore, the best version of Local Relevance, I think, justifes saving {B} over {A}. But then Local Relevance, like many other restricted aggregationist views, does have the striking implication noticed earlier. However, whilst this implication seems at frst blush strongly counterin­ tuitive, I ultimately don’t fnd it so. It initially seems counterintuitive because when we consider the ten and the 10,000 who will sufer serious injuries alone, the only question that we have in deciding which group to save concerns the weight that their injuries have in our reasoning, and the injuries that the 10,000 will sufer have much greater weight in total than the injuries to the ten. No question of appropriateness arises. We are thus led to focus on only on the fact that these injuries have a certain moral weight in our practical reasoning, and then it seems puzzling that adding similar facts with less weight {B} than to {A} can make it right to switch from {A} to {B}. But the ten who are added to {B} do not only have signifcance for the weight that their plight has in our reasoning. Prior to their being added, no one in {B} faced a plight that was sufciently signifcant to weigh against the decision to save X from death. Teir introduction to the case thus introduces facts that it is appropriate to consider in deciding whether to save X, and they count decisively against doing so. Facts are added to {A} that have even greater weight, but this is not decisive because there are other facts about {B} that have a great deal of potential weight, and that it is appropriate to weigh against the facts added to {A}. Once the case is understood in this way, we can see both why it initially seems implausible that I should switch from {A} to {B}, but also why this is ultimately unproblematic. To reinforce this way of understanding the case, notice that our reason against saving {B} at Stage 1 is not that the combined total weight of dutygrounding types of facts provided by members of {B} at that State is insufcient to outweigh the signifcance of X’s life. Teir total weight would easily be sufcient to do so were it appropriate to consider their plights. Tus, where we introduce facts to {B} that are sufcient on their own to justify not saving X’s life, there are very powerful reasons against doing so—the fact that doing so would result in our not saving 1 billion people

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

183

from a broken fnger. Given that, it is not surprising that introducing something of very considerable value to {A} is insufcient to justify saving that group. Tis is in line with the general idea behind Local Relevance—the fact that the broken fngers in {B} cannot play a role in the local decision whether X’s life is decisive does not imply that they are irrelevant in the overall decision whether to save the group containing X. Tey can still play a role in counterbalancing the harm that other members of {A} will face. Of course, not everyone will believe that weight is not all that matters, but that appropriateness also determines what we should do, in aggregation cases. However, we have already seen that it is highly plausible that weight and appropriateness are both morally signifcant in practical reasoning. Once we see this, we should not be surprised that there are cases where the weight of some facts seems to count powerfully in favor of saving some group, but this turns out not to be decisive. V.  WHEN OTHER THINGS ARE EQUAL

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Local Relevance might be thought at a disadvantage when compared with Global Relevance in cases where people have equal claims to be saved from a great harm, but saving one person will also save a person from a minor harm. As Frances Kamm notes, in such cases, the minor harm seems irrelevant, and it seems plausible that we ought to use some randomizing procedure to decide whom to save. Consider her case:14 Sore Troat:  I have enough medicine to save either A or B from a lethal illness. If I save B I will have a little medicine left which I can use to cure C’s sore throat. Otherwise, C will sufer the sore throat for a week. Kamm plausibly argues that saving B rather than A in order to prevent C’s sore throat is wrong. Rather, I ought to use some randomizing decision procedure, such as fipping a coin. She thus claims that the sore throat is an irrelevant utility, or irrelevant good. Tis is similar to the idea noted earlier: curing the sore throat achieves a good, but it is normatively inert, and it thus cannot ground a permission not to save A. If Kamm’s intuition is sound, it seems to favor Global Relevance over Local Relevance. Here is why. If I could cure C’s sore throat at no cost, that is what I ought to do. V thus has a claim to be cured of a sore throat.   Morality, Mortality vol. 1: Death and Whom to Save From It (Oxford: OUP, 1993) 146.

14

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

184

Victor Tadros

According to Local Relevance, C’s sore throat cannot outweigh A’s death, but it may be relevant more generally to what I am permitted or required to do. A’s claim to be cured from the lethal illness is exactly counterbalanced by B’s identical claim. Terefore, there is nothing left for C’s claim to outweigh. Terefore, Local Relevance implies that I ought to cure B and C rather than fipping a coin, which seems wrong.15 Global Relevance does not have this unwelcome implication because on that view, the dutygrounding force of C’s sore throat is generally disabled. Two ways of developing Local Relevance avoid this unwelcome implication, though. Te frst relies on the idea that if I choose to save B because I can cure C’s sore throat, C’s sore throat grounds a permission not to save A from death where otherwise I would be required to give him an equal chance of being saved. Friends of Local Relevance might claim that this is objectionable—C’s sore throat cannot ground a permission not to give A a chance of being saved from death, either on its own, or in conjunction with other permissiongrounding types of facts. It can only outweigh duty-grounding types of facts that do not disable its duty-grounding force. Here is the second. As A and B are equally placed, there is nothing to choose between them. Were C not present, it seems, I would be required to fip a coin. Tis seems required rather than picking some other method of selecting—say, a method that gives A a higher chance of survival than B, such as a weighted coin. Tis is so even though there is nothing to choose between the only two acceptable outcomes—A being saved or B being saved. One argument for that conclusion draws on the idea that a chance of being saved has value independently of whether a person is actually saved; another is that we have respect-based reasons to give others a chance of being saved where we can. Tus even though there is no diference in the value of the outcome whether the standard or the weighted coin is used, there is a diference in the value of the way that this outcome is reached. In order to justify curing B and C in Sore Troat, the duty-grounding force of C’s sore throat must be sufciently strong to defeat the dutygrounding force of this value. It is plausible that it is not, and that explains why I must fip a coin to decide whom to save. Tis would also nicely explain why we are required to use a randomizing procedure to determine whom to save in cases where we face two options that are morally close to each other in other respects, but not where we face two options that are morally distant from each other in other respects. For example, it seems that I ought to fip a coin, or use a weighted lottery, to determine whom to save in Sore Troat, but not in a case where C will sufer a much graver harm such as death or grave disability. In the latter case, it seems wrong to use a 15

  Tomlin suggests this concern in “On Limited Aggregation,” 259–60, fn. 30.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

185

weighted lottery. Tis is plausibly explained by the fact that the value of death or grave disability to C defeats or disables the duty-grounding force of A’s chance of survival.16 Tose who hold this view might then claim that the value of A and B having equal chances of survival either outweighs or disables the curing of C’s sore throat in Sore Troat. If it is outweighed, it makes a diference how many sore throats I will cure, but not if it is disabled. I have some sympathy for the former of these last two views. Tis is in part because I doubt that there is much value in having a chance of being saved that is independent of actually being saved—where I don’t actually get saved, it is either not better, or not much better, to have had a chance of being saved. It is in part because it has implications about aggregation that I fnd more plausible. Suppose that if I save B there will be a little bit of medicine left which will give me an opportunity to fnd a cure for the common cold. I will thus save billions of people from sufering sore throats. I then fnd it intuitive that I ought to save B rather than A. And this seems true even if each person only ever gets one cold, and these colds never result in anyone being harmed beyond the efect of the cold itself. Tis suggests that the duty-grounding force of C’s sore throat is not disabled. Te justifcation of lotteries is a large and difcult topic, so I am not confdent about the last conclusion. But we can at least conclude that there is a range of plausible responses for friends of Local Relevance to make to handle cases where other things are almost equal.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

VI. INTRANSITIVITY Suppose that there are competing duty-grounding and permissiongrounding facts with respect to some action. Restricted aggregationism, as I have developed it, is the view that whether the latter facts either weigh against, or disable, the duty-grounding force of the former facts depends on the comparative importance of these facts. Tis view has implications for our judgments in successive pairs of cases, and in multi-group cases, that some fnd troubling.17 16   Te view that this set of verdicts is most plausible is also defended in C.  Piller, “Treating Broome Fairly” (2017) 29 Utilitas 214, 226–32. Piller explores an alternative explanation why weighted lotteries should be used only in cases where diferences in the claims that the lottery will decide between are not too great, drawing on John Broome’s theory of fairness. 17  Larry Temkin has done most to highlight this problem, and explore possible responses over many years. See, especially, Rethinking the Good: Moral Ideals and the Nature of Practical Reasoning (Oxford: OUP, 2012).

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

186

Victor Tadros

Te frst implication to consider is familiar: the view implies a kind of intransitivity. Recall: Case 1:  members of {A} will be killed if they are not saved whereas members of {B} will sufer severe injury. And: Case 2:  members of {A} will be killed if they are not saved whereas members of {B} will sufer a broken fnger. Now consider:

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Case 3:  members of {A} will sufer severe injury if they are not saved whereas members of {B} will sufer a broken fnger. If death disables the duty-grounding force of a broken fnger, but severe injury does not, I ought to rescue {B} rather than {A} in Case 1 if {B} is much larger than {A}; {B} rather than {A} in Case 3 if {B} is much larger than {A}; but {A} rather than {B} in Case 2, however much larger {B} is than {A}. Tis might strikingly imply, for example, that I ought to rescue 1,000 people from severe injury rather than one person from death; 1,000,000 people from broken fngers rather than 1,000 people from severe injury; but one person from death rather than 1,000,000 people from broken fngers. One response is that there is no real intransitivity in these cases, because the salient quality of an option that a person has can depend on the other options available.18 Tus, I do not have the same option “to save {A} from death” in Case 1 and Case 2. Rather I have the option “to save {A} from death rather than saving others from severe injury” in Case 1 whereas I have the option “to save {A} from death rather than saving others from broken fngers” in Case 2. Transitivity relations are relations between options in diferent circumstances that have identical salient properties. If options have salient properties in virtue of the presence or absence of other options, transitivity is not at issue. Restricted aggregationism is thus consistent with transitivity, strictly speaking. It might be argued that the strict transitivity principle is practically useless.19 But it is at least useful in this way. Suppose that the following seems true of some set of options: I ought to choose 1) {o1 over o2}; 2); {o2 over o3}; and 3) {o3 over o1}. Strict transitivity implies there must be some diference between seemingly similar options across these sets. Tere 18  For further discussion of this general approach, see J.  Broome, Weighing Goods (Oxford: Blackwell, 1991) 100–7 and Temkin, Rethinking the Good, 457–65. 19   For further discussion, see ibid.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

187

must be some diference between o1 in 1) and o1 in 3), or o2 in 1) and o2 in 2), or o3 in 2) and o3 in 3). Whether what seems to be true is true, then, depends on whether there is some diference of this kind, and we might fnd it more or less plausible that there is depending on the case. Furthermore, in abstract cases that we have been describing, where we stipulate that other things are equal across the options in our cases, our intuitions are correct only if there is some relational or contrastive fact that is morally signifcant that makes the seeming intransitivity compatible with the strict transitivity principle. In other words, in our cases there is an intransitivity of a kind—intransitivity of “contrast-free” options. Tose who fnd this kind of intransitivity troubling might do so because they do not think that contrastive facts can be morally signifcant in cases like this. But, of course, restrictive aggregationism, as I explored it earlier, just is the view that there are contrastive morally signifcant facts of this kind. Te whole point of the view is that certain options disable the duty-grounding force of facts in alternative options. And, we saw, this idea has signifcant intuitive plausibility in itself. Intransitivity of contrast-free options is simply an upshot of the view that the deontic force of some facts irreducibly depends on the moral qualities of the facts they compete with.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

VII.  MORE OPTIONS Te detour through intransitivity helps to highlight a more difcult question: how to determine what I am required to do where a wider range of options is available. Consider: Case 4:  I can save one person from death, 1,000 people from severe injury, or 1,000,000 people from a broken fnger. One concern is that restricted aggregationist views imply that anything that I choose to do in cases like this is wrong.20 Suppose that death disables the duty-grounding force of a broken fnger, but not of severe injury, and that severe injury does not disable the duty-grounding force of a broken fnger. Ten, it seems, it is wrong to save one person from death rather than 1,000 people from severe injury, and it is wrong to save 1,000 people from severe injury rather than 1,000,000 people from a broken fnger, but it is wrong to

20

  See D. Parft, “Justifability to Each Person” (2003) 16 Ratio 368, 384.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

188

Victor Tadros

save 1,000,000 people from a broken fnger rather than one person from death. Terefore, anything I do is wrong, which cannot be true.21

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

i)  Te Signifcance of More Options Restricted aggregationist views have this unwanted implication only if we believe that adding an option that I ought not choose can make no diference to what I ought to do. To see this, compare Case 1 and Case 4. In Case 1 I ought to save 1,000 people from severe injury rather than saving one person from death. Tis might seem inevitably to lead to the conclusion that we ought to rule out saving the one from death in Case 4. If so, restricted aggregationism implies that I cannot act permissibly. When we compare Case 2 and Case 4 we rule out saving the 1,000. When we compare Case 3 and Case 4 we rule out saving the 1,000,000. As we have ruled out saving anyone in Case 4, saving anyone is wrong, which must be false. In the light of this, friends of restricted aggregation should reject the view that the addition of an option that I will not choose cannot make a diference to how I ought to choose between my other options.22 When comparing Case 1 and Case 4, they ought to deny that the fact that I ought to save the six rather than the one in Case 1 necessarily implies that I ought not to save the one in Case 4.23 Whilst this is somewhat surprising, it is not implausible. Indeed, the reason to accept this conclusion is exactly the same as the reason to accept intransitivity of contrast-free options. We have already seen that it is highly plausible that the moral signifcance of an option may depend on relational facts between that option and other options. Te moral qualities of an option are thus not completely internal to that option. From this, the introduction of new options may alter the moral qualities of one of our existing options but not another, and this will lead to a switch between these options. Some claim that this leads to a further problem: it compels me to seek seemingly irrelevant information.24 To see this, suppose I have two options of which I should pick some option (o1) over another option (o2). We have 21   For similar concerns about a related issue in the context of afrmative action, and some suggestions for how to respond, see Temkin, Rethinking the Good, 197–201. 22  See also Kamm, Intricate Ethics, 486; Voorhoeve, “How Should We Aggregate Competing Claims?,” 78–9. 23   See also Voorhoeve, “How Should We Aggregate Competing Claims?,” 78–9 to see this spelled out more formally. 24  Halstead, Ethics; and compare Temkin’s joke: “John goes into an ice cream store and tells the server, ‘I’ll have strawberry, unless you have vanilla—in which case I’ll have chocolate’,” Rethinking the Good, 388.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

189

seen that if there is some two-option case where if I am presented with a third option (o3), I ought not to choose that option, but I ought to switch between the two options that I had and choose o2. Tis also implies that if I am unsure whether I have o3, I ought to fnd out, not because I might take that option, but because whether I have it determines whether I should choose o1 or o2. But whether this is irrelevant information depends on whether contrastive facts are morally salient. If they are, the information that I seek is not irrelevant, because whether I have o3 really does make a diference to which of my other options I should choose. Some will think that cases where we seem to have no reason to seek certain kinds of information strengthen the confdence that we have in our intuitions that the thing about which we would seek information is irrelevant. I agree that this is so.25 And so I agree that cases where we lack information about o3 may strengthen the challenge somewhat. But they do no more than that.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

ii)  Reliance and Defeat Tis argument does not determine what I ought to do in Case 4. So what should I do? Alex Voorhoeve argues that restricted aggregationist views have clear implications in this case on the following basis. Te duty-grounding force of the broken fngers is disabled by death. Te duty-grounding force of the severe injuries is not disabled, and it outweighs the duty-grounding force of death. Terefore, I ought to save the 1,000 from severe injury.26 But this argument relies on an assumption that I have argued against— that if death disables the duty-grounding force of a broken fnger for the purposes of determining whether it is permissible not to save a person from death, the same thing is true for all other purposes in any decision where the broken fnger competes with death. Te analysis of Adding People shows that this assumption is false. Te fact that the claims of the 1,000,000 are disabled with respect to determining whether the death of the one is decisive does not imply that they are disabled with respect to determining whether other duty-grounding types of facts are defeated. Once we see this, we cannot safely draw the conclusion that I ought to save the 1,000. It might seem that Local Relevance implies that I ought to save the 1,000,000 on the following basis. Te claims of the 1,000,000 are disabled with respect to the question whether death is defeated. But the duty-grounding 25   Indeed, I use this strategy in a diferent context in “Permissibility in a World of Wrongdoing” (2016) 44 Philosophy and Public Afairs 101, 127–8. 26   See also F. M. Kamm, Intricate Ethics, 485.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

190

Victor Tadros

force of death is outweighed without these facts—the duty-grounding force of the severe injuries that the 1,000 will face outweighs it. Te dutygrounding force of the broken fngers is not disabled with respect to determining I ought to save the 1,000, and they defeat these claims. Terefore, I ought to save the 1,000,000. Tis, though, is also too quick. Tere is a problematic implication of restricted aggregationist views for Case 4 that has not yet been noticed. Suppose that I ought to save one of the three groups and a restricted aggregationist view is true. If so, facts about those who I ought to save both outweigh or disable the duty-grounding force of some competing facts about others, but also rely on the duty-grounding force of those facts. Tere is something troubling about this idea—to put it crudely, a group benefts from the plight of some other group, but at the same time they ensure their plight is not addressed. Tis is true whichever group I ought to save. For example, suppose that I ought to save the one. If this is true, this is so because the death disables the duty-grounding force of the broken fngers but the duty-grounding force of the broken fngers outweighs the duty-grounding force of the severe injuries. Tus the one both relies on the duty-grounding force of the broken fngers to ensure that the 1,000 are not saved rather than him, whilst at the same time his plight disables that force to ensure that he is saved rather than the 1,000,000. Similarly, suppose that I ought to save the 1,000. If this is true, this is so because the plight of the 1,000 outweighs the plight of the one, but the plight of the one disables the duty-grounding force of the 1,000,000. Tus the 1,000 both ensure that the plight of the one is not met, and at the same time rely on that plight for the purpose of disabling the duty-grounding force of the broken fngers. Or suppose that I ought to save the 1,000,000. If this is true, this is so because the signifcance of their injuries outweighs the signifcance of the severe injuries, and the severe injuries outweigh the signifcance of death, ensuring that death does not disable the duty-grounding force of the broken fngers. Tus the 1,000,000 rely on the fact that their plight outweighs the plight of the 1,000, and at the same time rely on the plight of the 1,000 to outweigh the duty-grounding force of the death that the one will face. On the non-aggregationist view, then, some people will at the same time rely on the plight of some other group, and yet ensure that this plight is not met. Te idea that this is troubling can be seen as part of a more general concern that is increasingly familiar in non-consequentialist ethics: the concern with the permissibility of harming a person whilst at the same time relying on that person to prevent oneself from being harmed. Suppose that I can prevent a group of people, {A}, from being harmed. If I do so I will harm another group {B}. A familiar view is that it is easier to justify harming

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

191

{B} if {B} would otherwise have made {A} worse of. It is harder to justify harming {B} if {B} would otherwise have made {A} better of. Tere is a stronger expectation on people to internalize the costs they infict on others, even involuntarily, than to beneft others. For example, one consideration in favor of a permission to divert a trolley from a track with fve people onto a track with one person on it is that if the rights of the one make it wrong to do this, the one being on the track will have disadvantaged the fve. In contrast, one consideration against a permission to throw one person in front of a trolley to save fve people from being killed by it is that if this is permissible, the one will have made the fve better of.27 I don’t claim that this is the only morally signifcant fact that explains these cases, but that it is plausibly part of a more complete explanation of them. Defending this view, and developing the best version of it, would take us too far afeld. Let us just work with the rough idea that harmfully benefting from a person is harder to justify than harming the person without benefting from them, where, roughly speaking, one harmfully benefts from a person where her presence benefts one, and one’s presence harms the person. If restricted aggregation is true, whatever I am required to do in Case 4, someone will beneft from someone else’s presence whilst making them worse of. Given that this is so whatever I do, it might be thought not to afect what I should do. However, whilst the same kind of duty-grounding fact counts against each of my options in Case 4, that fact may not count against each of my options with equal force. Te duty-grounding force of harmfully benefting from someone may depend on a range of factors, plausibly including the magnitude of harm and the magnitude of the beneft. Fully examining the implications of this fact about Case 4 would require more space than I have here. But because of this additional consideration, it is unsafe to conclude that Local Relevance implies that I ought to save the 1,000,000. To determine what to do, we would need to do more work on the factor that I have just alluded to and determine its signifcance compared with the considerations about weight and appropriateness that I have been considering so far, something that I cannot do here.

27   For diferent versions of roughly this idea, and its relationship to the wrongness of using others, see, for example, G.  Øverland, “Moral Obstacles: An Alternative to the Doctrine of Double Efect” (2014) 124 Ethics 481; A. Walen, “Transcending the Means Principle” (2014) 33 Law and Philosophy 427 and “Te Restricting Claims Principle Revisited” (2016) 35 Law and Philosophy 211; V. Tadros, “Wrongful Intentions Without Closeness” (2015) 43 Philosophy and Public Afairs 52.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

192

Victor Tadros VIII. RISK

Perhaps the most difcult objection to respond to concerns risk.28 I frst set out the argument from risk in a powerful form, and then suggest some ways to respond to it.

i)  Te Argument from Risk Consider: Risk 1:  X is unconscious. I can either a) save X’s fnger with a one in 1 million chance of killing him; or b) refrain from saving X’s fnger. It seems that I should choose a). Now compare: Risk 2:  One million people are unconscious. I can either a) save the fngers of all 1 million people in successive operations where each will sufer a one in 1 million chance of death; or b) refrain from saving the fngers of the 1 million.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Tose who accept that I should choose a) in Risk 1 might think it obvious that I should choose a) in Risk 2. It seems that if I should choose a) in Risk 1 I should choose a) in each iteration of Risk 1. Risk 2 is just a series of iterations of Risk 1. Now compare: Risk 3:  One million people are unconscious. I can either a) start a robot that will perform a series of operations on 1 million people to save one fnger of each person who survives, where there is no risk to anyone who lacks a certain genetic disorder, but which will certainly be lethal to anyone with such a disorder. I know that only one of these people has the genetic disorder, but I don’t know whom; or b) refrain from saving the fngers of the 999,999 remaining people. Some might think that Risk 3 is just like Risk 2, for in Risk 3 there is a one in 1 million chance that any person that I operate on will be killed, just as there is in Risk 2. 28  Something like the objection outlined here is in J.  Horton, “Aggregation, Complaints, and Risk” (2017) 45 Philosophy and Public Afairs 54. For more discussion of the problems that anti-aggregationist views face with respect to risk, and suggestions of how to deal with them, see R. Kumar, “Risking and Wronging” (2015) 43 Philosophy and Public Afairs 27; J. Frick, “Contractualism and Social Risk” (2015) 43 Philosophy and Public Afairs 175.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

193

Now compare: Certainty:  One million people are unconscious. I can either a) start a robot that will operate on all 1 million where there is no risk to anyone who lacks a certain genetic disorder. I know that X has this disorder; or b) refrain from saving the fngers of the 999,999 remaining people. Te only plausible moral diference between Risk 3 and Certainty, it might be argued, is that I know the identity of the person who will be killed in the latter case but not the former. But even this diference lacks moral signifcance, so if I ought to perform the operations in Risk 3, I ought to start the robot in Certainty. Restricted aggregationists claim that it is wrong to kill one to save the fngers of any number of people in Certainty. If this is true, either the intuitive view that I ought to choose a) in Risk 1 is wrong, or there is some moral diference between these cases. None of these options seems palatable. Te argument from risk concludes that we should reject restricted aggregationism, and accept that it is permissible to choose a) in Certainty.

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

ii)  Interpersonal and Intrapersonal Morality I begin with a general response to this argument—one that does not rely on any particular view about risk. Restricted aggregationist views are concerned with interpersonal morality. Tey compare the signifcance of the distribution of a total magnitude of harm across many diferent lives with the concentration of a smaller total magnitude of harm across fewer lives. For example, in Case 2, where {A} is smaller than {B}, my choice is to determine whether fewer people sufer a much graver harm, or more people sufer a much lesser harm. Tis creates a dilemma where the aggregation of the lesser harm is greater than the aggregation of the greater harm. Te intuition that I ought to save {A} is plausibly explained by the fact that we do not simply aggregate harms across lives. If the moral signifcance of harm simply aggregated across lives, restricted aggregationism is almost bound to fail, because the aggregation of harms in {B} will be greater than the harm in {A} if {B} is sufciently larger than {A}. Even if we have some antiaggregationist intuitions in the intrapersonal case, they are much more powerful in the interpersonal case. For example, it may be rational to have a temporarily debilitating operation when I am young to prevent many back spasms that will otherwise occur throughout the course of my life, even if the whole harm that is caused by the back spasms is no greater than the sum of the parts. But it seems much more difcult to justify harming me at this moment to prevent

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

194

Victor Tadros

very many other people sufering one back spasm each. Te separateness of persons, this suggests, plays a central role in explaining anti-aggregationist intuitions; a view that has a long history. Furthermore, many familiar non-consequentialist moral principles restrict promoting the good. As is well known, such principles make our prospects worse in conditions of uncertainty. But this does not seem a decisive reason to reject them.29 For example, compare a duty to rescue one’s own child rather than two stranger children with a duty to rescue the greater number of children, even at the expense of one’s own child. In conditions of uncertainty, where there is an equal chance of my child being any of these children in such rescue situations, and people do what they are required to do, the second principle improves the prospects of my child being rescued when compared with the frst. Whilst some may fnd this a powerful reason to prefer the second duty,30 the argument does not seem decisive, partly because interpersonal features of rescue cases are eliminated in conditions of uncertainty. Te argument from risk makes a series of subtle changes to a clear intrapersonal case—Risk 1—ending up with a clear interpersonal case— Certainty. Te cases in between—Risk 2 and Risk 3—are less clear in this respect. Tese cases press us either to abandon the view that interpersonal tradeofs are morally signifcant, or to develop a theory about how they interact with risk. Te argument from risk efectively suggests that we take the frst option, but given that it is highly plausible that the separateness of persons sometimes makes interpersonal tradeofs inappropriate where intrapersonal tradeofs are appropriate, we should explore other options. I will mainly focus on two possibilities, and I do so only tentatively. My aim is not to ofer decisive support for either possibility, but rather to show that each has some promise.

iii)  Interpersonal Tradeofs Tis subsection focuses on diferences between Risk 2, Risk 3, and Certainty. For that purpose, let us suppose that if it is permissible for me to act in Risk 1, it is also permissible for me to iterate similar acts any number of times in Risk 2. 29   Although some, of course, think so. See, for example, C. Hare, “Should We Wish Well to All?” (2016) 124 Philosophical Review 451. 30   See, especially, D. Parft, Reasons and Persons (Oxford: OUP, 1986) 55–66. For related discussion about the relationship between agent-relative/agent neutral theories and problems of aggregation in the light of Parft’s challenge, see Temkin, Rethinking the Good, 87–95.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

195

One feature that distinguishes Risk 2 from Certainty is that if I save those with the broken fngers in Certainty, there is a single indivisible act that involves an interpersonal tradeof between the life of one person and the smaller beneft to many others. Conduct that benefts {A} makes {B} worse of; conduct that benefts {B} makes {A} worse of. If interpersonal tradeofs are morally signifcant, we should expect that these tradeofs have implications for what I should do. Tis is not clear in Risk 2. We are supposing that the moral signifcance of the set of operations I perform in Risk 2 is reducible to the aggregate of the moral signifcance of each operation taken individually (an assumption that I will explore in subsection iv). As there are clearly no interpersonal con­ siderations that bear on Risk 1, it follows that there are no interpersonal considerations that bear on Risk 2. Te case that forms a bridge between these—Risk 3—is less clear in this respect. Whether interpersonal tradeofs are involved depends on whether the appropriate perspective to assess such tradeofs, assuming they exist, is the ex ante or the ex post perspective. From the ex ante perspective, each person has better prospects if I act than if I don’t, so there is no interpersonal tradeof involved. From the ex post perspective, one person is much worse of if I act than if I don’t, whereas many others are much better of if I act than if I don’t. So from this perspective the case involves interpersonal tradeofs. Tis highlights a concern about the argument from risk. Risk 3 is like Risk 2 in one respect—ex ante, each person has a small risk of being killed and a large risk of being benefted if I operate. It is like Certainty in another respect—ex post, one person is certain to be killed and many others benefted if I act. Now notice that there is a powerful intuitive diference between Risk 2 and Certainty. It seems intuitive that I should not act in Certainty but that I should perform all of the operations in Risk 2. But it is not clear whether I should act in Risk 3. One way to vindicate this set of intuitions is that there is an important moral diference between Risk 2 and Certainty, but we are unsure which of these cases Risk 3 is most like morally speaking. It may be just like Risk 2, or just like Certainty, or it might be an intermediate case. To put it crudely, Risk 3 is a distraction. It causes us to downplay the signifcance of interpersonal tradeofs, because we are uncertain whether they are engaged in the case. Given this difculty, we should be more confdent in the judgment that there is a diference between Certainty and Risk 2 than we should about our judgments about the relationship between Risk 3 and either of the other cases. Whilst there is some initial intuitive plausibility in the view that Risk 2 is no diferent to Risk 3, and that Risk 3 is no diferent to Certainty, the powerful intuition that there is a diference between Risk 2 and Certainty should make us less confdent in this view.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

196

Victor Tadros

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

iv)  An Objection Tis argument has some force, but some may deny that it is sufciently powerful to defeat the argument from risk, because even if there are diferences between Risk 3 and either of the adjacent cases, they are not plausibly sufciently important to support the anti-aggregationist view. First focus on the view that Risk 2 and Risk 3 are morally diferent. Suppose that there is a one in 10,000 risk of X being killed, and that this risk is small enough that I am permitted to operate. Now suppose that I am considering two populations of 1 million people, who have broken fngers that I could repair. Population 1 includes people just like X in Risk 1. If I operate on this population I expect to cause around 100 deaths. Population 2 includes one person who is certain to be killed, but I don’t know who that person is. If there is no diference between Certainty and Risk 3, and restricted aggregationism is true, it is wrong for me to operate on Population 2. But if there is no diference between Risk 1 and Risk 2, I am permitted to operate on Population 1. Tus I am permitted to save fewer fngers, causing more deaths in one population where I could operate on another population, saving more fngers and causing fewer deaths. And as we increase the size of these populations even further, the diference in the number of deaths increases. Multiply the size of these populations by 1 million and then it is wrong to perform an act that will save the fngers of all 1 billion people except one, who will be killed, but it is permissible to perform an act on another population that will save the fngers of 999,999 fewer people, causing 1 million deaths. Even those who agree that there is some diference between Risk 2 and Risk 3 might fnd it implausible that the diference between these cases is so signifcant that we should tolerate the deaths of almost 1 million extra people to do less good in virtue of this diference. Now consider the view that Risk 3 and Certainty are morally diferent. Tis view faces a similar objection. Suppose that I am permitted to operate on any number of people in cases like Risk 2, and Risk 3 is morally identical to Risk 2. I am then permitted to operate on 1 billion people to protect their fngers, where one distinct person is certain to be killed, but I don’t know whom. Now compare a similar population, where I know the identity of the person who will be killed. Now multiply the size of these populations by 1 million. In that case, it is wrong for me to operate on 1 million billion people in a way that is certain to kill one identifable person, but permissible for me to operate on 1 million billion people in a way that will certainly kill 1 million people who I cannot identify. Te diference between being able to identify those who will be killed, on that view, can result in a very large increase in the number of deaths that I am permitted to cause—almost a million extra deaths in this case.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

197

Given that a central motivation of anti-aggregationism is the special signifcance of grave harms in restricting us from acting, there is an unpleasant irony in the implication that it can lead us to act in a way that increases the number of people who sufer such harms very signifcantly when compared with alternatives. I am not sure whether this objection is decisive. Perhaps we should accept these results. First note that the application of non-consequentialist principles to large numbers generally results in a great deal of extra harm. For example, the Doctrine of Double Efect implies that it is wrong to intentionally kill 1 billion people to save 5 billion, resulting in 4 billion extra deaths, and some may balk at the extraordinary number of extra resulting deaths. But I doubt that this is sufcient to show that the Doctrine of Double Efect is false. Furthermore, we earlier explored the idea that our inability to imagine large numbers of harms and goods makes our judgments in these cases unreliable. Tis argument may have some force here—cases involving large numbers of deaths make us less sensitive to other morally signifcant facts. Alternatively, it might be argued that these implications can be avoided. One way to avoid them is to rely on interaction efects between options. For example, suppose that I am permitted to aggregate in cases like Risk 3, but not in cases like Certainty. Now suppose that I have two options concerning two populations, where one is like Risk 3 and the other is like Certainty, and the death toll is greater if I take the frst option. I may be permitted to take the second option, even if it would have been wrong to perform a similar action were I not to have the frst option. In such a case, my permission to kill a person is justifed not simply by the broken fngers that I will save, but to avoid the greater number of deaths that I will otherwise permissibly cause. Again, this result is somewhat counterintuitive, but that result needs to be compared with the rejection of intrapersonal tradeofs, and rejecting such tradeofs seems radical. From our discussion, we can see both that it is difcult to respond to the argument from risk by distinguishing Risk 2, Risk 3, and Certainty, but also that such a response is not without promise. Still, as we are running into some difculties, let us explore an alternative.

v)  Acts and Conjunctions of Acts Tat alternative is that it is wrong to conclude from the permissibility of operating in Risk 1 that it is permissible to operate on any number of people in cases like Risk 2. Let us begin with the general issue. Tere are some morally relevant facts that obtain when a conjunction of acts is performed,

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

198

Victor Tadros

but no individual act in the conjunction has that moral property at all. In such cases, each individual act may seem wrong (or permissible) because it lacks the moral property where the conjunction of acts seems permissible (or wrong) because it has the moral property. Tis is familiar in cases where there is an interaction efect between facts that, taken alone, are morally inert, but that interact in a way that makes them morally signifcant. For example, the fact I give X a certain level of welfare might be inert on its own, as might the fact that I give Y a certain level of welfare, but an act that gives rise to both facts may be wrong because it involves bringing about an objectionable inequality between X and Y. Now suppose that certain acts are wrong because of the interpersonal tradeofs between harm and beneft that result. Tis fact cannot make it wrong to operate in Risk 1. But this does not imply that performing a series of identical acts in Risk 2 is permissible. Te interpersonal tradeofs that the conjunction of these acts gives rise to may violate a principle that requires me not to harm some severely for the purpose of benefting many others to a lesser degree. If I operate in Risk 1, there is no chance that my acts together will beneft some to a smaller degree whilst harming others to a much greater degree. If  I operate 1 million times in Risk 2, in contrast, I should expect my conjunction of acts to cause one death, benefting many others to a much smaller degree. And where I perform 1 billion similar operations, it is almost certain that my conjunction of acts will have this result. For this reason, it is unsafe to conclude from the permissibility of my acting in Risk 1 that my conjunction of acts is permitted in Risk 2. Now let us explore an objection. It might be argued that if a conjunction of acts performed in a series is wrong, whether one act in the series occurs or not must make a diference whether some morally salient quality occurs or not, either because of its intrinsic or relational properties. Tat individual act is wrong because it gives rise to the morally salient fact, and that explains why performing the conjunction of acts is wrong. Te morally salient fact may arise due to interaction efects between the later act and earlier acts, which explains cases involving interaction efects. For example, suppose that I perform two acts, one afecting X and the other afecting Y, that give rise to an objectionable inequality between X and Y. If the frst act would be permissible without the second, and the second would be permissible without the frst, the wrongness of the conjunction of these acts act is explained by the interaction between them. Where this is so, something morally signifcant occurs when the second act is performed that gives rise to this interactive efect. If I afect X frst, and there is no objectionable inequality, the act that gives rise to the objectionable inequality is that which afects Y, and that second act must then be wrong because of

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

199

the inequality that it brings about. Tis second act is wrong only if I have frst afected X. But the principle against objectionable inequalities makes my second act wrong when I have performed the frst. Now compare Risk 2, and suppose that restricted aggregationism is true. No act in the series can be wrong because it gives rise to a violation of the restricted aggregationist principle. Here is why. Restricted aggregationism is a principle that restricts the benefts to others that can justify an act of killing. But the fact that I have benefted others in the past, or will do so in the future, cannot make performing an operation on any particular person in the set wrong. Any individual operation performed in the set only involves the person operated on, and is done solely for her expected beneft. I thus don’t need to appeal to the benefts to others to justify operating on  any individual person. Each operation is justifed, if it is, simply by the beneft that the particular patient that I operate on is likely to receive. So no operation, taken individually, can be wrong because of restricted aggregationism. But, the argument goes, if restricted aggregationism does not make any individual act in the set wrong either taken on its own, or in virtue of its interaction with other acts that I perform, it cannot make the conjunction of acts wrong either. To see how forceful this objection is, suppose that I have performed 999,999 operations, and each has gone well. I am now deciding whether to operate on the last person, Y. If I operate on Y, and Y dies, I will have acted wrongly in the fact-relative sense. If restricted aggregationism is true, Y’s death will not be justifed by the beneft that the 999,999 have received. But, it seems, restricted aggregationism cannot make performing the operation wrong. If Y is killed, the operation is not plausibly wrong because I have previously benefted other people. Restricted aggregationism claims that certain benefts to others cannot justify killing a person. It does not ridiculously claim that benefts to others make killing wrong. Given this, my operation on Y can no more be wrong because of restricted aggregationism than the frst operation I performed, which did not implicate that principle at all. Te same thing is true of any act in the series, and therefore each act in the series is permissible. It might be thought that things are diferent if an earlier operation goes wrong. But they are not. Suppose that I operate on the 999,999th person, and that person is killed. I am now deciding whether to operate on the last person, Y. My operation on Y may be wrong because I have already killed a person. But it cannot be wrong because I will violate any restricted aggregationist principle. Tis is so for the reason ofered before. Te purpose of the operation is not to beneft others at all; it is to beneft Y. So it cannot be wrong because others have received insufcient benefts to justify my performing it.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

200

Victor Tadros

But although this seems to militate in favor of performing all of the operations if I perform one, things are not so simple. We can see this by focusing not on the decision whether to perform any particular act in the set of acts in Risk 2, but whether to perform all of these acts. Tis conjunction of acts is almost certain to kill a person. And then I must ask what can justify performing a conjunction of acts that will kill a person. Te answer must appeal to some good that the conjunction of acts will achieve. But if restricted aggregationism is true, these benefts do not justify the harm that I will cause. So the conjunction of acts does seem to implicate the antiaggregationist principle. It might be objected that the good that I aim at by performing this conjunction of acts is not to cure the fngers of many people, but rather to give each person a chance—a very high chance—of having her fnger cured. And the bad that I aim to avoid is a risk to each person—a very small risk— of being killed. If this is right, we can accept that performing all of the operations in Risk 2 is permitted whilst denying that I am permitted to act in Certainty, for in Certainty I don’t give anyone either a chance of being cured or a risk of being killed. I doubt that this is the right way to think about the conjunction of acts, though. I noted earlier that the chance of having one’s fnger cured is not a signifcant beneft independently of one having one’s fnger cured. Te proper way to think about the conjunction of acts is to think about the range of outcomes that the conjunction of acts might bring about, the moral signifcance of each possible outcome if it occurs, and the probability of each outcome occurring. If I perform one operation, there is a very high chance that I will cure a broken fnger without causing a death, and a very small chance that I will cause a death without curing a broken fnger. We evaluate these two outcomes, and the probability they will occur, and that determines what I should do. If I perform 1 billion operations, in contrast, I will almost certainly bring about an outcome where I cure some number of fngers, and kill some number of people. All of these states of afairs violate restricted aggregationism. Tere is only a miniscule probability that my conjunction of acts will result in my achieving an outcome where restricted aggregationism is not violated—the probability of killing no one. If this is the right way to assess conjunctions of risky acts, three possible responses to this problem that do not involve rejecting restricted aggregationism remain. First, it might be argued that if I perform all of the operations, it is permissible to perform each operation, but wrong to perform the conjunction of the operations. Tis view faces a decisive objection—it can make it impossible to avoid acting wrongly. To see this,

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.



Localized Restricted Aggregation

201

suppose that I am a doctor, and I am not only permitted but also required to perform operations that are in the best interests of my patients. In that case, I am both required to perform each operation, and yet required not to perform the conjunction of operations. It follows that before I have performed any of the operations, it is impossible for me to ensure that I don’t act wrongly. Either I violate my obligation to perform each operation, or I violate my obligation to perform all of the operations. We should avoid this radical implication if we can. A second possibility is that the permissibility of performing each operation makes the performance of the conjunction of operations permissible. On this view, although my conjunction of acts makes it almost certain that I will bring about a state of afairs where one person is killed and others are benefted, and I justify the death that I will almost certainly cause by appealing to the benefts to others, the fact that each of my individual acts is permitted makes performing this conjunction of acts permissible. If this view is right, it has no implications for Certainty, for in the latter case there is no way of dividing the larger act into permissible parts. A third possibility is that the wrongness of the conjunction of acts makes it wrong to perform some of the individual acts. On this view, there is nothing that makes any individual act wrong other than that it is part of a conjunction of acts that is wrong. Earlier, I suggested that the last act in the set in Risk 2 cannot make it true that the anti-aggregationist principle is violated. And this is right. But, on the view under consideration, the fact that the conjunction of acts violates the anti-aggregationist principle can make this act wrong. Pure aggregationists might argue that any principle that gives rise to such unpalatable options must be false. Tis argument, though, is too quick. Tere are other cases where it seems powerfully intuitive that there can be a set of acts where the value of each act in the set warrants performing that act, it seems wrong to perform all of the acts, and there are no interaction efects between acts in the set that explain this result. Consider: Life Choices: If I do nothing, everyone in the problem will immediately die. A)  I can avoid this by doing one of two acts: 1)  Act in a way that will shorten the life of one person by one year; or 2) Act in a way that will shorten the life of 2 billion people each by a 1 billionth of a year. It seems that I clearly ought to do 2), no matter how long the lives of the one or the 2 billion will be. Tis is so as 2) makes no diference to anything

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

202

Victor Tadros

that anyone has reason to value, given that the diference between living 1 billionth of a year longer or shorter is imperceptible, or close to imperceptible. B)  Now suppose that I can either: 1) Act in a way that will shorten the life of 1 billion people by one year; or 2) Act in a way that will shorten the life of 2 billion diferent people by one year. It seems clear that I ought to do 1).

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

C) Now suppose that I must make 1 billion choices, where each choice will either: 1)  Shorten the life of one diferent person each time by one year; or 2) Shorten the life of one group of 2 billion people each time by 1 billionth of a year. In this case, each choice is like the choice in A). But my conjunction of choices achieves an outcome which is identical to choosing 2) rather than 1) in B), which is surely the wrong result. No doubt, this case raises many difcult questions. But this is at least a plausible analysis of the case. Each act in the conjunction of acts in C) is like A). Taken independently of the conjunction of acts, for each act I ought to choose 2). But where we consider a conjunction of acts, I ought not to choose 2) each time. And this is so simply because the conjunction of acts directly afects the deontic status of members of the conjunction. No doubt, this is an unpalatable result, but the alternatives seem no better.31 Overall, then, I think that the view that it is permissible to perform some set of operations in Risk 2 and then stop, whilst initially very implausible, remains worth considering. I accept, of course, that there is something strongly counterintuitive about this result, but that is true of all of the views that we are considering. CONCLUSION Restricted aggregationism in general, and Local Relevance in particular, has a number of surprising implications, and raises a range of problems. Pure aggregationism is a much simpler theory. Some might think this some reason to prefer pure aggregationism. I don’t think that this is true. Simple 31   See, also, the brief discussion in D. Parft, On What Matters, Vol. 3 (Oxford: OUP, 2017) 431.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -



Localized Restricted Aggregation

203

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

moral theories do not do justice to the conficts we experience in everyday moral life, and in evaluating complex cases. Tey are for this reason unsatisfying. Tat being said, restricted aggregationism has some theoretical implications that are counterintuitive. Tis is especially true where risk is concerned. Overall, though, I doubt that these implications are so severe that we should prefer pure aggregationism, given that it is powerfully intuitive that some harms are too trivial to make a diference to the permissibility of saving a person’s life. No doubt Local Relevance needs further development, but I hope at least to have shown it to have some promise.32

32   I am grateful to the Leverhulme Trust for a Major Leverhulme Research Fellowship that aforded me the time to work on this chapter. Tanks to Tom Dougherty, Korbinian Rueger, Larry Temkin, Patrick Tomlin, Alex Voorhoeve, and Alec Walen for excellent comments on earlier drafts and related papers, and to audiences in Oxford, Arizona, St Andrews, and LSE and UCL. Special thanks to Doug Portmore and Joe Horton who were my commentators in Arizona and St Andrews respectively, and to Seth Lazar, who provided extensive comments as a reviewer for Oxford Studies in Political Philosophy.

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved. Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Index academic departments  134–5 aggregationism anti-aggregationism  171–2, 193–4, 197, 201 pure aggregationism  171–2, 175–6, 201–3 see also restricted aggregationism Anderson, Elizabeth  8, 21n33 anti-discrimination law  117–20 Apartheid regime  69–70 autonomy  32n50, 173–4

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Belgium see language preservation Boettcher, James  50 caste society  9, 11, 143–4 Chiang Kai-Shek  100 civil disobedience  90–6 conscientiousness and  108–9 nonviolence and  105–6 “Orthodox View” and  91–2, 97, 102, 110–12 publicity and  107–8 respect for law and  106–7 special features of  105–9 civility  90–1, 109–10 civil rights movement  69–70, 102 collective action problems  147, 149–50 see also language preservation comprehensive liberalism  39–40, 44–7, 51–2, 56, 59 legitimacy and  46–7 mutual moral authority and  57–8 non-public moral reasons and  52 public reason and  48–51 stability and  53 Dauenhauer, Bernard  52–3 Declaration of Independence  68 democracy 3–7 democratic authority  3, 24–5 citizen control of  104 value of  24

deontology canceling considerations and  176–7 defeating considerations and  180, 189–90 disabling considerations and  173–5, 178, 185–90 duty-grounding facts  172–3 duty-grounding force  173–81, 184–5, 187–91 outweighing considerations and  173–5, 177, 179–80, 182–4, 190 permission-grounding facts  172–3 permission-grounding force  172–3 discrimination 117–20 as demeaning  123–6, 145 direct discrimination  120–2, 125–6, 138–40, 142, 144 disability and  132–3 expressivist accounts of  119, 123–6, 144–5 indirect discrimination  120–3, 125–6, 140–4 invisible groups and  125, 133, 142–5 power diferentials and  124 promotion tests and  140–1 social groups and  124–8 store hours and  136–7 wrongness of  121n8, 122n10, 123–4, 145–6 see also social subordination disobedience  25, 90 see also civil disobedience, uncivil disobedience Doctrine of Double Efect  197 dress codes  121–3, 125, 140, 145 duty of civility  48–50 duty to obey the law  90–2, 111–12 procedural notion of  92–4, 102–5 substantive notion of  68, 97–102 duty to rescue  92–3, 98 Dworkin, Ronald  134, 162, 164–6 see also Principle of Correction Eberle, Christopher  54–5 ecotage  94–7, 99, 103–4

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

206

Index

equality see political equality, relational egalitarianism, social status inequality Estlund, David  6–7 fair play principle  164n26 friendship  4–5, 9–10, 13n23, 25–31 non-instrumental value of  31, 33–5 see also relational egalitarianism Fugitive Slave Act of 1850  91n4 Gaus, Gerald  39, 41, 43, 55–9 Greene, Amanda  87–8 hactivism  94–5, 97–9, 103 Hellman, Deborah  119, 123–6 Hobbes, Tomas  69–70

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

injustice general delegitimating efect of  70, 77–80, 85 legitimacy and  67–71, 73–4, 78–83, 86, 88 personal delegitimating efect of  68–9, 77, 85 international law  69–70 Jim Crow laws  117–18, 139 justice as module  42 legitimacy and  85 political relations and  31–2 see also injustice, legitimacy, libertarianism, linguistic justice justifcation correctness-based  40–1, 44, 54–5 legitimacy and  75–6, 81–3, 88 see also social justifcation Kamm, Frances  183–4 Kant, Immanuel  80 Kolodny, Niko  9, 16n29, 21–2, 24–5, 129–30 Laitin, David  149, 155–6, 169–70 language see language preservation language preservation  147–9 as Assurance Game  153 as consumption good  149

as Prisoner’s Dilemma  151–3 Belgium and  152–3, 155–6, 168–9 culture and  147–9 identity and  147–9, 155, 160 preference orderings and  156–7, 160–1 Québec and  149–52, 155, 167–8 state intervention and  155–7, 165, 167 value and  148–9 see also public goods, Public Goods Argument Larmore, Charles  60 legitimacy  41–7, 65 democratic theory of  87 dynamism and  71, 88–9 “Evidential Gambit” and  81–2, 86 history and  67, 79–80, 87–9 justice and  42–3 liberal principle of  40–3, 87 multipolarity of  80, 83, 85–6 non-moral conditions for  44 “Other-Regarding Gambit” and  79–81, 86 “Permeability Gambit” and  82–4, 86 “Right of Resistance Gambit” and 84–6 scalar notion of  86 structure and  67, 79–80, 87–9 threshold notion of  83–4 see also injustice, reparations liberalism see comprehensive liberalism, public reason liberalism libertarianism justice and  47 natural rights libertarianism  47, 55 public justifcation and  47 linguistic justice equal treatment and  148–50, 159–62, 169–70 opportunity for self-determination and  157, 159, 162–4 territory and  156n15 see also language preservation, public goods Lister, Andrew  48–50 Locke, John  82–4 luck egalitarianism  55 marriage 26–9 method of avoidance  42, 60–1

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

Index 207

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

Miller, David  9, 160 Mills, Claudia  52–3 moral responsibility  56 patient consent  173–4 peonage 14n24 perfectionism and public justifcation  47–8 political equality  3–5, 8–11, 21–2 by-product view  6 ideals and  7 moral baseline view  6–7, 23–4 see also political power, relational egalitarianism, social status inequality political liberalism  44–5, 48–9 political power  8, 10–11, 19–20, 23, 25–6 equal power ideal  29–36 exit options and  22, 36 inequality of  20–2 infuence and  5n3 political representation principal/agent relation and  22–3 Principle of Correction see public goods public goods assurance problem  162–4, 166–7, 169 discretionary public goods  158–9 essential public goods  157–8 free-rider problem  162–3, 166–7, 169 non-excludability and  153–4 preference revelation and  166 price and  162–4 Principle of Correction  157, 164–6 provision of  154 Unanimity Principle  157–62, 165–6 see also language preservation Public Goods Argument  147–50, 153–7, 159–62, 169 see also language preservation, public goods public justifcation  39 convergence and  49–50 mutual moral authority and 56–8 principle of  43, 52 respect for persons and  54–6 theoretical role of  58–60

value of  50 see also legitimacy, public reason, Rawls public reason  48–51 as public point of view  50–1 public reason liberalism  39 sectarianism and  60 public washrooms  133–5 punishment 112n35 Québec see language preservation Quong, Jonathan  60–1 Rawls, John civil disobedience and  93–4 legitimacy and  82, 87 public goods provision and  158–9, 165–6 public justifcation and  39–45, 48–9, 55, 58–61 Raz, Joseph  52–3, 92n6, 93 relational egalitarianism  3–5, 7–11, 18–19, 21–2, 34–7 anti-caste paradigm  10–11, 23–5, 33 friendship conception  25–34 reparations 71 general legitimating efects of  73, 78, 83 legitimacy and  65–6, 72–4, 78, 86–9 personal legitimating efects of  72–3, 78, 83 republicanism 10–11 respect for persons and basic moral rights  55 see also public justifcation, stability restricted aggregationism  171 argument from risk and  192–5 conjunction of acts and  197–202 contrastive facts and  187 Global Relevance and  179–80, 183–4 interpersonal morality and  193–4 interpersonal trade-ofs and  194–5, 198 intransitivity and  185–7 intrapersonal morality and  193–4 irrelevant information and  188–9 Local Relevance and  179–85, 189–91, 202–3 signifcance of more options  188–9 see also deontology

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -

208

Copyright © 2019. Oxford University Press USA - OSO. All rights reserved.

right of resistance  84–6 right to rule  65 Rousseau, Jean-Jacques  10–11 Schefer, Samuel  8 Sea Shepherds Campaigns  104n30 self-defense  96–7, 100–1, 105–6 Simmons, A. John  65–6, 82 justifcation and  75–6 legitimacy and  74–5, 79, 85 particularity and  75–6, 89 voluntarism and  76–8, 81–3, 86–7 social class  12n20 social justifcation  17–19, 24 social status diferentiation among  15–16 political representation and  22–3 social status hierarchy  9, 11–21, 23–4, 34 social status inequality  16–17, 34–5 expressivist dimension of  19n32 social subordination  126–41, 144–5 censure and  129–30, 135, 137–9 consideration and  133–5, 137–9 de facto authority and  128–32, 134–9 power and  128–31, 134–9 stereotypes and  130–2, 135–9 structural accommodations and  119–20, 129, 132–7, 139, 142–3 see also discrimination societal norms  13–14, 16

Index stability factors contributing to  52–3 public justifcation and  51–3 respect for persons and  54 stereotypes see social subordination subordination see social subordination, discrimination Tomlin, Patrick  179 Toronto Transit Commission  143–4 Unanimity Principle see public goods uncivil disobedience  90 conscientiousness and  108–9 costliness of  96–7 proportionality and  96–7 publicity and  108 respect for law and  107 unfavorable conditions and  110–11 vandalism 100 Van Parijs, Philippe  149–50, 155–6, 169–70 violence  94–5, 97, 99–103 state monopoly on  106 wrongness of  105–6 Voorhoeve, Alex  176–7, 179, 189 Wall, Steven  5, 40–1 war (ethics of )  96–7 Wellman, Christopher  93–4, 93n7 whistleblowing 94–8 Wicksell, Knut  158–9

Oxford Studies in Political Philosophy Volume 5, edited by David Sobel, et al., Oxford University Press USA -