127 73 2MB
English Pages 288 Year 2024
Oxford Studies in Normative Ethics
Oxford Studies in Normative Ethics Volume 13 Edited by
MARK TIMMONS
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 2023 The moral rights of the authors have been asserted All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2023941277 ISBN 978–0–19–889590–9 DOI: 10.1093/oso/9780198895909.001.0001 Printed and bound in the UK by Clays Ltd, Elcograf S.p.A. Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.
Contents Acknowledgments List of Contributors
Introduction Mark Timmons 1. Paradoxical Proposals and Consent Tom Dougherty 2. Manipulation, Disrespecting Autonomy, and Deliberative Projects Ron Aboodi and Shlomo Cohen
vii viii
1 9
31
3. Moral Indifference Monika Betzler and Jonas Vandieken
55
4. Buck-Passing and the Value of a Person Kyla Ebels-Duggan
77
5. Moral Theorizing and the Limits of Coherence Sarah Buss
99
6. A Contractualist Approach to Moral Uncertainty Michael Bukoski
126
7. The Ethics of Expectations Rima Basu
149
8. Satisficing Virtuously Earl Conee
170
9. Pro Tanto Rights and the Duty to Save the Greater Number Benjamin Kiesewetter
190
10. Can’t Kant Count? Innumerate Views on Saving the Many over Saving the Few Sergio Tenenbaum
215
vi
11. What Does Virtue Have to Do with Consequences? Iskra Fileva
235
12. Two Conceptions of Rights David O. Brink
253
Index
277
Acknowledgments Versions of the chapters in this volume were presented at the 2022 Arizona Workshop in Normative Ethics, January 13–15, 2022. I thank the Center for the Philosophy of Freedom and the Department of Philosophy at the University of Arizona for their generous financial support of the workshop. Of course, the views expressed in the volume’s chapters do not necessarily reflect the views of the Center or the Department. I am grateful to Carla Bagnoli, Paul Hurley, Jessica Fisher, Paddy McShane, Doug Portmore, and Michael Zimmerman for serving as referees for the 2022 program and to two anonymous referees for Oxford University Press who wrote excellent comments on the workshop papers and gave constructive advice to authors. Thanks also to Caleb Dewey who prepared the volume’s index, to Betsy Timmons who greatly facilitated conducting the workshop online, and to my editor Peter Momtchiloff for his support. Mark Timmons Tucson, AZ
List of Contributors Ron Aboodi is a Postdoctoral Fellow at Tel Aviv University. Rima Basu is Assistant Professor of Philosophy at Claremont McKenna College. Monika Betzler is Professor of Practical Philosophy and Ethics at Ludwig Maximilians University Munich. David O. Brink is Distinguished Professor of Philosophy at the University of California, San Diego. Michael Bukoski is Assistant Professor of Philosophy at Florida State University. Sarah Buss is Professor of Philosophy at the University of Michigan. Shlomo Cohen is Associate Professor of Philosophy at Ben-Gurion University of the Negev. Earl Conee is Professor of Philosophy at the University of Rochester. Tom Dougherty is Professor at the University of North Carolina at Chapel Hill. Kyla Ebels-Duggan is Professor of Philosophy at Northwestern University. Iskra Fileva is Associate Professor of Philosophy at the University of Colorado, Boulder. Benjamin Kiesewetter is Professor of Practical Philosophy at Bielefeld University. Sergio Tenenbaum is Professor of Philosophy at the University of Toronto. Jonas Vandieken is a Postdoctoral Fellow in Philosophy at Ludwig Maximilians University Munich and the University of Toronto.
Introduction Mark Timmons
Oxford Studies in Normative Ethics features new work on topics in normative ethical theory. This thirteenth volume features chapters on the following topics: paradoxical proposals and consent, manipulation and autonomy, moral indifference, the value of persons, moral theory and coherence, a contractualist approach to moral uncertainty, normative expectations and practical identity, satisficing and virtue ethics, pro tanto rights, problems with saving the many, virtue and consequences, and the nature of moral and political rights. A paradoxical proposal, as explained by Tom Dougherty in “Paradoxical Proposals and Consent,” is one in which (i) an agent wrongs a deliberator by proposing to refrain from performing an action A unless the deliberator chooses an option (ii) even though it would not be wrong for the agent to perform A. How is the deliberator wronged in such cases when the agent is not proposing to do something wrong should the deliberator refuse to choose the option? For instance, Citizen knows that Criminal carried out the heist and threatens to turn Criminal into the authorities unless Criminal pays Citizen off. The question Dougherty addresses in his chapter is which paradoxical proposals (if any) “constrain a deliberator’s choices in a way that invalidates consent that is given because of these proposals?” In some cases of concern proposals introduce a disincentive for the would-be victim, thus worsening the victim’s choice situation. If Citizen would not otherwise turn in Criminal, her proposal introduces a disincentive. In such cases consent is invalidated. However, as Dougherty points out, there are other paradoxical proposals that do not invalidate consent yet still wrong the recipient of the proposal by, e.g., taking advantage of the person’s vulnerability via independent disincentives, i.e., disincentives that are not brought about by the proposal. Dougherty holds that the two different kinds of paradoxical proposal call for a two-pronged approach. His approach to cases in which the proposal introduces a disincentive for the deliberator to comply is that their consent is invalidated because they have a valid claim not to have their Mark Timmons, Introduction In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Mark Timmons 2023. DOI: 10.1093/oso/9780198895909.003.0001
2
choice situation worsened. For the other type of case in which the deliberator’s consent is not invalidated, were the agent to act on the proposal they may still wrong the deliberator by wrongfully taking advantage of their situation, setting back an important interest of theirs. Since, as Dougherty points out, the two prongs of his approach are independent, either of them could be accepted while rejecting the other. Manipulation either relies on influencing someone’s non-deliberative psychological mechanisms (e.g., playing on someone’s guilt) or it relies on some means of trickery (e.g., lying). In “Manipulation, Disrespecting Autonomy, and Deliberative Projects,” Ron Aboodi and Shlomo Cohen explore the practical meaning of disrespecting a person’s autonomy by focusing on how such disrespect results from manipulation, with the particular aim of illuminating the notion of autonomy. They take the object of such disrespect to be what they call “the manipulee’s deliberative project” understood as one’s “interrelated ongoing efforts to manage her conduct by committing to normative stances, ideals, policies, plans, goals, and the like.” Disrespecting a manipulee’s deliberative project is how they propose to precisify disrespecting a person’s autonomy. They proceed by considering intuitions about cases of manipulation. Before turning to their own account, Aboodi and Cohen consider various accounts of the wrongness of manipulation in the literature and find them wanting as precisifications of disrespecting a manipulee’s autonomy in yielding intuitively incorrect verdicts when manipulation does disrespect a manipulee’s autonomy. Although, as they note, there are various ways of disrespecting someone’s autonomy, they focus on intentionally interfering with it (and risking such interference), arguing that whether paternalistically motivated or not, disrespecting someone’s autonomy is the most typical pro tanto wrong-making feature of manipulation. The authors conclude by considering both theoretical and practical implications of their view. Monika Betzler and Jonas Vandieken in “Moral Indifference” argue that such indifference should be recognized as a distinct species of interpersonal wronging. They distinguish occurrent, dispositional, and structural forms of such wronging, united by the victim being denied the kind of moral status constituted by having the authority to make and respond to moral claims and thus the standing to make a moral difference. In their example of occurrent moral indifference Mary inadvertently bumps into John, immediately apologizing, though John does not respond to her apology. Mary’s reaction to being ignored is that she “feels momentarily slighted as a fellow moral being who can participate in accountability practices and thus
3
legitimately expects uptake of her valid claims.” This kind of treatment, they observe, constitutes an implicit non-reciprocal kind of address that denies the victim basic moral status. To foreground the distinctive nature of moral indifference, they contrast it with treating someone merely as a means, contemning, and being a “moral asshole,” each of which at least recognizes the victim as having basic moral standing, denied by moral indifference. The distinctive nature of such indifference, Betzler and Vandieken explain, implies that there are different forms of recognition respect. In addition to the Strawsonian participant and objective stances one might take toward others, an “indifferent stance”—a kind of non-reciprocal denial of the victim’s agency—deserves recognition. They conclude with remarks on situations in which moral indifference may be instrumentally valuable in shielding one from irresolvable interpersonal moral conflicts. According to T.M. Scanlon’s buck-passing conception of value, for something to be valuable is for there to be reasons for everyone to respond to it in various ways. Reasons are understood to be provided by “ordinary” features of the valuable thing—features other than being valuable. This conception, then, represents a reduction of values to reasons. Applied to the value of persons, Scanlon’s view is that their value consists in the reasons one has to treat persons in accord with moral requirements—requirements which on Scanlon’s contractualist moral theory are expressed by principles for the general regulation of conduct that no one (properly motivated) has reason to reject. In “Buck-Passing and the Value of a Person,” Kyla Ebels-Duggan argues that Scanlon’s view (and buck-passing conceptions of value in general) fail to fully capture the value of persons. To make her case, Ebels-Duggan first considers relationships of love and the importance of distinguishing between genuinely meaningful loving reactions and merely responding to the reasons there are to value a person, which on Scanlon’s view, constitutes valuing the other even when whose actions are “empty gestures.” Genuinely meaningful loving reactions depend on an attitude of appreciation that cannot be identified with recognition of moral reasons to do or refrain from certain actions. One might suppose that Scanlon could restrict his claim about the value of persons and say that his view is meant to capture what it is for a person to have moral value, the proper response to which is respect. However, Ebels-Duggan argues that the valuing attitude of respect cannot be reduced to recognition of reasons one has to outwardly comply with moral requirements; one’s attitude toward others matters as well. One might then suppose that Scanlon’s recognition of reasons for attitudes as figuring in the value of persons could rescue the buck-passing view.
4
However, as Ebels-Duggan explains, careful attention to reasons for valuing leads to the view that understanding such reasons requires an independent grasp of valuing attitudes through experiencing them. One might suppose that a moral theory, to be plausible, ought to avoid endorsing incompatible moral commitments, indeed that taking the moral point of view in one’s thinking ought to involve endorsing a variety of moral commitments that may come into conflict but conflicts we can and should resolve. These related assumptions about moral theory and moral agency are challenged by Sarah Buss in “Moral Theorizing and the Limits of Coherence” in which she defends the idea that a kind of incoherence is not only unavoidable in theory and practice but desirable. Buss argues for this view partly by reflecting on being a person of moral virtue. The idea is that such individuals have ideals (e.g., being beneficent and being just) and are thereby invested in acting in ways that such ideals require. Inevitably, one will confront situations in which living up to one or another of one’s ideals will conflict with what one all-things-considered ought to do. To suppose that someone might craft her commitments so that all conflicts among ideals can be resolved results in an impoverished conception of one’s identity—empty of content and thus something which, even if possible, is not something to strive to achieve. A kind of incoherence is internal, then, to being a person of virtue. Buss also argues that the constraints that an ideal of coherence imposes on one’s understanding of substantive ideals reveals that the coherence of alleged instances of supererogation is illusory. Buss’s central claim about supererogation is that those who conceive of the supererogatory as actions that are supported by the strongest moral reasons yet not morally required are confused about what it is for a consideration to qualify as “moral.” Still, as she points out, being confused about supererogation need not be a failing. Incoherence, then, in our moral agency and reflected in our moral theories is arguably both unavoidable while yet desirable. The issue of moral uncertainty, which has received increasing attention in the ethics literature, is whether being uncertain about fundamental moral facts or principles has any effect on what one morally ought to do. Those who think it does have proposed moral uncertainty principles not associated with any particular moral theory, while there are those who deny that such uncertainty has any bearing on what one morally ought to do. A heretofore unexplored “middle” position agrees that moral uncertainty can affect one’s moral obligations, but instead of proposing “global” moral theory-neutral principles, this position settles for moral uncertainty principles that are internal to a moral theory. Such uncertainty principles would be “local” in
5
the sense of being first-order principles belonging to some moral theory. This option is explored by Michael Bukoski in “A Contractualist Approach to Moral Uncertainty” which, as the title indicates, considers how a contractualist moral theory (using Scanlon’s as an example) might approach the issue of moral uncertainty and do so in a way that avoids two principal objections that beset theory-neutral attempts to formulate moral uncertainty principles. The idea, then, is that from within a contractualist moral theory one can formulate moral principles that treat an individual’s moral uncertainty as a relevant factor in determining what one morally ought to do (regardless of one’s level of confidence in those very principles). Bukoski proceeds to defend a contractualist moral uncertainty principle that prescribes a moderate degree of moral caution in cases of uncertainty—one that does not prohibit all moral risk taking. And he argues that the best contractualist uncertainty principle will be “discriminating” in being sensitive to the content of the views about which one is uncertain. He illustrates this latter dimension addressing cases in which one is morally uncertain of contractualism itself and cases in which one has doubts about the range of beings with moral status. As Rima Basu explains in “The Ethics of Expectations,” the notion of expectation has received relatively little attention from philosophers, despite its ubiquity in everyday life. Basu’s chapter, then, explores the nature of expectation in its various guises with particular attention given to interpersonal expectations and the ethical issues they raise. Regarding the nature of expectations, Basu distinguishes multiple functional roles they play, including their predictive, prescriptive, and proleptic roles. A key to such expectations is their anticipatory function that Basu illuminates by the metaphor of a map, familiar from discussions of the ethics of belief. Mere predictive expectations about the future can, of course, simply be wrong, but other kinds of expectation can constitute wrongs themselves. For instance, expecting one’s child to come to have a morally malicious identity is inherently wrong—morally problematic as such. Furthermore, “weighty” interpersonal expectations can function proleptically as when one gives someone they are closely attached to (e.g., their children) directions on how to lead their lives and thereby structure their own life with those same expectations. Of course, in some such cases, these kinds of expectation can be enabling, but in other cases they are inhibiting, and one thereby wrongs the target individual by, for instance, attempting to impose upon them an identity that they cannot share. Basu elaborates these and other ethical issues that expectations pose, noting in conclusion that besides issues that interpersonal expectations
6
raise, there are also intrapersonal issues regarding what we expect of ourselves that remain largely unexplored. One reaction to various putative counterexamples to maximizing versions of consequentialism (MC) is to embrace satisficing consequentialism according to which (roughly) morally permissible acts are ones that are “good enough.” So, for instance, while maximizing versions fail to accommodate many cases of supererogation (classifying them as obligatory), satisficing versions comfortably handle them: when acts are ones that maximize value (are optimific) but come at a high personal cost to the agent, one is normally permitted on a satisficing view to perform a suboptimal act; doing the optimal thing is doing more than obligation requires in the circumstances and is thus supererogatory. Accommodating cases of supererogation is but one example of how satisficing consequentialism (SC) fares better than its maximizing cousin. However, as Earl Conee in “Satisficing Virtuously” points out, SC has its own set of difficulties in providing a non-arbitrary conception of what counts as “good enough.” After canvassing various SC proposals and finding all of them problematic, Conee explores the prospects of a hybrid view that disjunctively combines MC with virtue theory into a theory of moral permissibility. However, this view also succumbs to counterexamples. Conee concludes that a fully adequate theory of moral permissibility can be given solely in terms of virtue thus: “A is morally permissible iff A is virtuous.” In “Pro Tanto Rights and the Duty to Save the Greater Number,” Benjamin Kiesewetter defends “rights contributionism,” the thesis that claim rights can be overridden which he then employs in addressing the duty of saving the greater number in Taurek cases. In the first part of his chapter, Kiesewetter defends rights contributionism by: (1) articulating the properties of a species of moral reason that has what he calls a “directed deontic structure,” (2) correlating such reasons with moral rights, and (3) arguing that such rights while preemptive in the sense that they can trump certain competing reasons can nevertheless be infringed—that they are therefore contributory rather than absolute. The chapter’s second part considers so-called Taurek cases¹ in which one either can save some number of people or can save an even greater number of people (or do nothing). Such cases seem to be governed by a principle of saving the greater number: “Other things equal, agents have an overall moral obligation to save the ¹ From John Taurek’s paper “Should the Numbers Count?” Philosophy & Public Affairs 6: 293–316.
7
greater number in Taurek cases.” While this principle is easily accommodated by at least some versions of consequentialism, deontological theories have a harder time. Kiesewetter, however, defends a distinctively deontological approach that aims to vindicate the principle of saving the greater number based partly on an appeal to the pro tanto right to be saved. The strategy is to defend a “modest” reasons aggregation principle, according to which reasons aggregate by default although aggregation can be sometimes prevented by rights. This principle, together with the contributionist conception of rights including a right to be saved in Taurek cases, yields a duty (pro tanto) to save the greater number. Taurek cases are also a topic in Sergio Tenenbaum’s “Can’t Kant Count? Innumerate Views on Saving the Many over Saving the Few.” So-called innumerate views regarding Taurek and similar cases deny that the fact that one can either save the many or save the few creates an obligation (at a fundamental level) to save the many simply because it is better to save more lives. Kantian and other non-consequentialist views that put constraints on maximizing aggregate welfare seem committed to embracing innumeracy that conflicts with what Tenenbaum calls the “More the Merrier” principle (somewhat like Kiesewetter’s “Save the Greater Number” principle); principles that many find intuitively plausible. Tenenbaum’s project is to find a “middle way,” denying that in Taurek cases one is required to save the many while also avoiding reliance on consequentialist principles. The task, as Tenenbaum sees it, is to defend the obligation to save the many in the Taurek cases but do so without relying on (or vindicating) the More the Merrier principle. Tenenbaum’s middle way deploys a Kantian notion of the dignity of persons that explains why in Taurek-like cases one has no obligation to save the many, but that in other cases one does have an obligation to save the greater number. According to consequentialism about virtue, consequences are all that matter in determining which traits are virtues. Opposed to this position is motivationalism about virtue according to which virtue is solely a matter of internal states of a person and so consequences are irrelevant. In “What Does Virtue Have to Do with Consequences?” Iskra Fileva observes that both positions have their attractions and problems. While the consequentialist approach captures reasons for labeling some traits such as helpfulness as virtues and worthy of social encouragement, it violates intuitions about cases including those in which inner states seem morally good apart from consequences. Motivationalism captures intuitions about individual virtue, yet it does not give a plausible answer to the question of which traits should
8
be encouraged (and which discouraged), or else provides an intuitively wrong answer to the question. One might suspect that these views talk past one another: the consequentialist is concerned with what makes some dispositions virtues or vices while the motivationalist is concerned with what makes a person virtuous or vicious. However, Fileva seeks to defend a conciliatory position that illuminates and explains the relation between virtues and vices on one hand and motives and consequences on the other. On the view she proposes, consequences bear on the rightness of actions and thus which dispositions should be encouraged and which discouraged. Honesty and kindness, viewed in terms of their overall consequences are to be encouraged. However, consequences may not be the crucial determinant when it comes to virtue ascription to individuals. Rather we ascribe virtue based on the reasons for which someone does the honest or kind thing. A key idea in Fileva’s reconciliatory account is that a virtuous disposition requires suitable motivation, one that connects one’s motives with the reasons that make that type of action permissible or socially desirable and where permissibility and social desirability in turn depend on (or are sensitive to) the action’s consequences. After developing her view (beyond my bare description) Fileva proceeds to respond to likely objections from consequentialists and motivationalists. In the final chapter, “Two Conceptions of Rights,” David O. Brink distinguishes contributory rights from resultant rights. Contributory rights, when applicable, are defeasible (though presumptively decisive) and contribute to the resolution of moral, political, and legal controversies. Resultant rights, by contrast, are final and indefeasible, the result of adjudicating competing considerations bearing on some controversy (including those represented by competing contributory rights). After explaining the two conceptions of rights, Brink proceeds to make several observations about the significance of distinguishing them, including the following. First, insofar as moderate versions of deontology recognize rights, those rights must be contributory. Second, if one recognizes conflicts of rights, one is treating them as contributory rather than resultant. Third, conflicts of rights foreground a difference in their respective contents: contributory rights are general and easily summarized (e.g., right to freedom of speech) while resultant rights, as correct solutions to competing considerations, tend to be qualified, “fine print” rights. Fourth, constitutional rights should be understood as contributory. Fifth, in some cases, it is necessary to recognize new contributory rights. Sixth, because contributory rights differ importantly from resultant rights, appealing to them in arguing for resultant rights avoids circularity.
1 Paradoxical Proposals and Consent Tom Dougherty
1. Introduction If someone agrees to hand over money because a mugger threatens them with violence, then they are not giving valid consent—i.e. they are not giving consent that releases the mugger from their duty not to take the money.¹ This is plausibly explained at least partly by the fact that they would be wronged by the violence. But this explanation won’t apply to cases in which someone is proposing to act permissibly in the event that consent is withheld: Snitch. Citizen discovers Criminal committed the masked heist. Citizen threatens to tell the authorities unless Criminal pays them off. Sacking. Because business is bad, Employer must lay off a staff member. Worker is one of two equally eligible candidates for being fired. Employer says to Worker, “I will fire you if and only if you do not have sex with me.” Citizen is permitted to tell the authorities and Employer is permitted to sack Worker. Yet it is wrong for Citizen and Employer to make these proposals.
¹ Since there may be other moral considerations besides valid consent that bear on what it is permissible to do, valid consent is not sufficient for permissibility. Elsewhere, I defend a more complicated view that distinguishes “fully valid,” “partially valid,” and “fully invalid” consent. The extra complexities make no difference to our discussion in this essay. Tom Dougherty, “Sexual Misconduct on a Scale: Gravity, Coercion, and Consent,” Ethics 131(2) (2021): 319–44. The terminology of valid and invalid consent is standard in bioethics and common in normative ethics. Within the criminal law, the term “consent” is commonly used to refer to an act that is necessarily normatively efficacious. On that terminology, what I call “invalid consent” would be described as e.g. merely “prima facie consent,” the absence of consent, or “assent.” See, respectively, Heidi Hurd, “The Moral Magic of Consent I,” Legal Theory 2(2) (1996): 121–46; Larry Alexander, “The Moral Magic of Consent II,” Legal Theory 2(3) (1996): 165–74; Kimberly Ferzan and Peter Westen, “How to Think (Like a Lawyer) about Rape,” Criminal Law and Philosophy 11(4) (2017): 759–800. Tom Dougherty, Paradoxical Proposals and Consent In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Tom Dougherty 2023. DOI: 10.1093/oso/9780198895909.003.0002
10
With a nod to the literature on the so-called paradox of blackmail, I will call these “paradoxical proposals,” which I define as follows:² X makes a paradoxical proposal to Y (or X accepts a paradoxical proposal initiated by Y) if and only if (i) X makes a proposal to Y (or X accepts a proposal initiated by Y) (ii) a term of the proposal is that X will not Φ if and only if Y Ψs; (iii) because of this term, X wrongs Y by making (or accepting) the proposal; and (iv) X would otherwise not wrong Y by Φ-ing. The definition’s parentheses anticipate that we will discuss proposals initiated by the consent-giver, but this refinement need not detain us now. My goal is to investigate how these proposals affect someone’s consent from a moral point of view. To make our inquiry more tractable, I will make an idealizing assumption of relevant full information: both the consent-giver and the consent-receiver know the consequences of giving or withholding consent. This assumption often does not hold in the real world where people are uncertain of how each other will act, but I leave for another occasion how to generalize an account for cases involving uncertainty.³ With that assumption implicitly in mind, my central question is: which paradoxical proposals, if any, constrain a deliberator’s choices in a way that invalidates consent that is given because of these proposals? To answer this question, I will defend a two-pronged account that distinguishes paradoxical proposals according to whether the proposals introduce disincentives that a deliberator faces in the event that they withhold consent. For paradoxical proposals that introduce disincentives, the first prong of my account is a principle that implies that these proposals can invalidate a deliberator’s consent in virtue of worsening the situation ² The so-called paradox of blackmail is that it can simultaneously be permissible to request that someone performs an action A1, permissible to perform a separate action A2 oneself, but impermissible to request that someone performs A1 in return for one’s not performing A2. A2 could be the disclosure of information, but the Sacking case illustrates that the philosophically interesting phenomenon is not limited to disclosure. For excellent essays that respectively introduce the relevant jurisprudential and philosophical literatures, see Mitchell Berman, “Blackmail,” in The Oxford Handbook of Philosophy of Criminal Law, ed. John Deigh and David Dolinko (New York: Oxford University Press, 2011), 37–106; James Shaw, “The Morality of Blackmail,” Philosophy and Public Affairs 40(3) (2012): 165–96. ³ For an indication of how the generalization would go, see Tom Dougherty, “Coerced Consent with an Unknown Future,” Philosophy and Phenomenological Research 103(2) (2021): 441–61.
11
in which the deliberator makes their choice. The second prong of my account concerns paradoxical proposals that exploit disincentives that exist independently of these proposals. I argue that although these proposals do not invalidate a deliberator’s consent, they may wrong the deliberator in virtue of taking advantage of the deliberator’s vulnerability in order to secure their compliance to an arrangement that unjustifiably sets back their interests. I will proceed as follows. In section 2, I show why paradoxical proposals are not covered by a common principle that states a sufficient condition for coercion that invalidates consent. In section 3, I sketch my two-pronged account and introduce theoretical notions that are needed to make it precise. In section 4, I defend the first prong of the account. In sections 5 and 6, I defend the second prong of my account.
2. The theoretical challenge of paradoxical proposals To provide the context for my account, let us first discuss an influential way of thinking about how coercion invalidates consent. Consider: Theft. Bully proposes to steal Victim’s sunglasses unless Victim loans Bully the use of their phone. Victim agrees. Trade. Merchant proposes to trade their pair of sunglasses in exchange for using Contractor’s phone. Contractor agrees. Victim and Contractor are moved by the same consideration—a pair of sunglasses. Yet only Victim’s consent is invalid. Since it is natural to call Bully’s proposal “a threat” and Merchant’s proposal “an offer,” it is tempting to think that threats invalidate consent but offers don’t. But we should resist this temptation for two reasons. First, some threats are acceptable ways to obtain consent: Steroids. Coach proposes that unless Player agrees to take a steroids test, Coach will drop Player from the team. Whether or not Coach’s threat is coercive, it does not invalidate Player’s consent to the test. Second, insofar as we rely on linguistic intuitions about “offers” and “threats,” the distinction is fraught. Consider:
12
Return. Bully has already stolen Victim’s sunglasses. Bully proposes to return the sunglasses if and only if Victim lets them use their phone. Victim agrees. Ordinary English lets us say both that Bully is “offering” to return Victim’s sunglasses and that Bully is “threatening” not to return them. Either way, these linguistic facts tell us little about the normative features of Bully’s proposals in the Theft and Return cases: both proposals invalidate Victim’s consent. Accordingly, the dominant approach in the literature junks linguistic intuitions about offers and threats and instead focuses on the normative properties of the incentives or disincentives that someone imposes to induce someone to consent. As a representative and promising principle along these lines, consider: Sanction. An agent’s coercive announcement invalidates an individual’s consent if (i) the announcement attaches a sanction to an option in an individual’s deliberation such that the agent would wrong the individual with this sanction; and (ii) the individual consents to avoid this sanction.⁴ The Sanction principle is attractive because of what it implies about the cases that we have encountered. Since Bully would wrong Victim by taking or keeping possession of their sunglasses, the principle implies that Victim’s consent is invalid. But because Coach would not wrong Player by dropping them from the team, the principle does not imply that Player’s consent is invalid. The principle is also attractive because it illuminates why a victim has a complaint against an agent acting on their consent when it has been invalidated by wrongful coercion: the victim might retort, ⁴ Principles along these lines are common in the literatures on consent and coercion. See Alexander, “The Moral Magic of Consent II”; Thomas Scanlon, Moral Dimensions, (Cambridge MA: Belknap Press of Harvard University Press, 2008), 80–3; Japa Pallikkathayil, “The Possibility of Choice: Three Accounts of the Problem with Coercion,” Philosophers’ Imprint 11(16) (2011): 1–20; David Owens, Shaping the Normative Landscape, (Oxford: Oxford University Press, 2012); Shaw, “The Morality of Blackmail”; Tom Dougherty, “Why Does Duress Undermine Consent?” Noûs 55(2) (2021): 317–33; for classical kindred approaches that specify a “moralized baseline” defined in terms of how well off a deliberator would be, were an agent to discharge all their duties, see Robert Nozick, “Coercion,” in Philosophy, Science, and Method: Essays in Honor of Ernest Nagel, eds. Sidney Morgenbesser, Patrick Suppes, and Morton White (New York: St. Martin’s Press, 1969), 440–7; Alan Wertheimer, Consent to Sexual Relations, (Cambridge: Cambridge University Press, 2003).
13
“You had no right to act on consent that I gave because you illegitimately constrained my options.”⁵
3. Distinguishing independent and introduced disincentives While principles along the lines of Sanction are both appealing and popular, it is not hard to see why these principles don’t get traction with paradoxical proposals. For example, since Citizen is not proposing to wrong Criminal in the event that Criminal withholds consent, the Sanction principle does not imply Criminal’s consent is invalid.⁶ So to handle paradoxical proposals, we need to supplement a principle like Sanction with something else.⁷ To fill this gap, I will propose an account that distinguishes two classes of paradoxical proposals and analyzes each in turn. I will start by providing a rough sketch of a principle that governs an important species of paradoxical proposals before introducing distinctions that will make it precise. My account focuses on whether a deliberator is made worse off by the paradoxical proposal itself. This depends on what the agent is otherwise going to do. Consider two scenarios. First, suppose that Citizen would not otherwise inform on Criminal. Then by making their paradoxical proposal, Citizen puts Criminal in a worse position to decide whether to give Citizen money: Criminal would prefer to make this decision with no strings attached. I will argue that if Criminal consents because the proposal puts them in this worse position, then their consent is invalid. Second, suppose that independently of making the paradoxical proposal, Citizen does plan to inform on Criminal. Then Citizen’s proposal improves Criminal’s choice situation by giving them an option that they would otherwise lack. Accordingly, the proposal would not invalidate their consent. That is not to deny the existence of other reasons why it might be wrong for Citizen to
⁵ These complaints can be offered in interpersonal justification. See Thomas M. Scanlon, What We Owe to Each Other, (Cambridge MA: Harvard University Press, 1998); Dougherty, “Why Does Duress Undermine Consent?” ⁶ Since my primary goal is to develop a positive account of paradoxical proposals that does not rely on the Sanction principle, and I wish to keep this essay to a reasonable length, I will not discuss attempts, such as those of Shaw and Scanlon, to show that despite appearances that principle can handle cases like Snitch and Sacking. Scanlon, Moral Dimensions; Shaw, “The Morality of Blackmail.” ⁷ If a supplementary principle states a sufficient condition for invalid consent, then it is not a competitor with the Sanction principle that also states a sufficient condition.
14
take Criminal’s money. For example, Citizen could be wrongfully setting back Criminal’s financial interests by taking advantage of their vulnerability. To make this sketch more precise, let us introduce two distinctions. First, to cash out whether a proposal worsens or improves a deliberator’s choice situation, let us distinguish incentives or disincentives that the deliberator faces: Incentive. If a deliberator D has an option O with consequence C, then C is an incentive for D to choose O if and only if C makes D better off. Disincentive. If a deliberator D has an option O with consequence C, then C is a disincentive for D to choose O if and only if C makes D worse off. With these definitions, I am stipulating “incentive” and “disincentive” as theoretical terms for the purposes of this essay. Ordinary English is more permissive about how these words are used. Second, in the sketch, I stressed what the agent would otherwise do. Does Citizen otherwise intend to inform on Criminal? These questions focus our attention on whether the paradoxical proposal itself (as opposed to other behavior of Citizen’s) creates a new disincentive that otherwise would not exist. Accordingly, let us distinguish consequences according to whether they causally depend on a paradoxical proposal: Introduced Consequence. A consequence is introduced by an agent’s paradoxical proposal to a deliberator if and only if the paradoxical proposal creates a causal connection between one of the deliberator’s options and the consequence. Independent Consequence. A consequence is independent of an agent’s paradoxical proposal to a deliberator if and only if before making the paradoxical proposal it is already the case that the agent is poised to bring about this consequence. As with most causal notions, these notions face the problem of deviant causal chains. Accordingly, phrases like “causal connection” would have to be charitably interpreted as “the right sort of causal connection.” As I cannot provide an account of what constitutes the “right sort,” I rely on our intuitive ability to discern whether a causal chain is deviant or not, hoping that you share my confidence in knowing a deviant causal chain when we encounter one. Together, these distinctions allow us to characterize two classes of paradoxical proposals. The introduced/independent distinction applies to both
15
incentives and disincentives as consequences. Let us restrict our attention to the special case in which a deliberator chooses whether to consent to the agent performing an action (e.g. the agent taking money). Insofar as the agent is motivating the deliberator to consent, the agent is attaching either an incentive to consenting or a disincentive to withholding consent, and the deliberator is motivated to pursue the incentive or avoid the disincentive. Some paradoxical proposals introduce disincentives to withholding consent. Other paradoxical proposals introduce incentives to consenting against a backdrop of independent disincentives. Let us consider each type of paradoxical proposal in turn.⁸ When a paradoxical proposal introduces disincentives to withholding consent, the proposal worsens the deliberator’s choice situation. Consider a variant of an earlier case: Lazy Snitch. Citizen has discovered that Criminal was the masked person responsible for the bank heist. Criminal and Citizen both know that Citizen is lazy and hence would not otherwise publicize this information. However, Citizen threatens to publicize this information unless Criminal pays them off. Criminal and Citizen both know that because Citizen cares about being a person of their word, they always follow through on threats. Citizen’s proposal introduces a disincentive: because of the proposal, Criminal’s option of not paying Citizen will now lead to Citizen informing on them. But Citizen’s proposal does not introduce an incentive: since Citizen was not otherwise going to inform on Criminal, the proposal does not introduce a new consequence that Criminal welcomes. By contrast, when a paradoxical proposal exploits independent disincentives, the proposal itself improves the deliberator’s choice situation by introducing an incentive for them to consent. Consider: Proactive Snitch. Citizen has discovered that Criminal was the masked person responsible for the bank heist. Criminal and Citizen both know that Citizen is proactive and hence otherwise will publicize this information. ⁸ An anonymous reviewer raised the possibility that there is no fact of the matter as to whether the agent independently plans to impose a consequence on the relevant option of the deliberator. To handle this scenario, one possibility would be to introduce a third prong to our account. But rather than introducing this, I suggest that we say that when it is indeterminate whether the agent independently plans to impose the consequence in question, it is in turn indeterminate whether the deliberator’s consent is valid.
16
However, Citizen proposes not to publicize this information if Criminal pays them off. Citizen is imposing a disincentive on Criminal’s option of withholding consent: withholding consent will cause Citizen to inform on Criminal. However, this disincentive is already in place independently of Citizen’s paradoxical proposal. Therefore, although Citizen’s independent behavior worsens Criminal’s choice situation, this is not a scenario in which the paradoxical proposal is worsening Criminal’s choice situation. By attaching an incentive to consenting, the proposal improves Criminal’s choice situation. By distinguishing these classes of paradoxical proposals, we can analyze them separately. In what follows, I will defend a two-pronged account. The first prong concerns paradoxical proposals of the former class. For these, I will defend a principle that specifies when these proposals invalidate consent. The second prong concerns paradoxical proposals of the latter class. For these, I will argue that these proposals do not invalidate consent, even though they may lead to conduct that is wrong on other grounds.
4. My account’s first prong: the Worsening principle My first prong concerns paradoxical proposals that introduce disincentives. The introduction of such a disincentive is significant for how the proposal affects the deliberator’s “choice situation.” For our purposes, I will understand a deliberator’s choice situation as the combination of the deliberator’s options and the causal consequences of these options.⁹ With respect to the Lazy Snitch case, let us compare the following choice situations: (a) If Criminal gives Citizen money, then Citizen does not inform on Criminal. If Criminal does not give Citizen money, then Citizen does not inform on Criminal. (b) If Criminal does not give Citizen money, then Citizen informs on Criminal. If Criminal gives Citizen money, then Citizen does not inform on Criminal. ⁹ For other theoretical purposes besides, we may wish to generalize the notion of a choice situation so that it includes other morally relevant phenomena besides the consequences of the options. For discussion of a broader range of morally relevant features of choice situations, see Scanlon, Moral Dimensions, 77–80; Niko Kolodny, “What Makes Threats Wrong?” Analytic Philosophy 58(2) (2017): 87–118, 94–7.
17
In the absence of Citizen’s paradoxical proposal, Criminal would be in choice situation (a). This is the status quo that obtains by default. But Citizen’s paradoxical proposal puts Criminal in choice situation (b). Since (b) is a worse choice situation than (a), the proposal worsens Criminal’s choice situation. I endorse the following principle that states when such a paradoxical proposal invalidates consent: Worsening. An agent invalidates a deliberator’s consent if (i) the deliberator consents because the agent makes or accepts a proposal that moves the deliberator from choice situation 2 to choice situation 1; (ii) choice situation 1 is worse than choice situation 2; and (iii) the deliberator has an all-things-considered claim against the agent moving the deliberator from choice situation 2 to choice situation 1. The idea is that if the agent has illegitimately worsened the deliberator’s choice situation, then this deprives any subsequent consent of the moral significance that it would need to be valid.¹⁰ For example, if Criminal hands over money because Citizen has infringed a claim of Criminal’s not to worsen their choice situation, then Criminal retains a claim against Citizen taking the money. Citizen cannot deflect the latter claim by appealing to Criminal’s choice to hand over the money. This appeal would fail because the choice results from Citizen wrongfully having influenced Criminal to make this choice. In other words, since Criminal has a complaint against Citizen’s blackmail, Criminal retains a complaint against Citizen’s acting on a choice that Criminal makes because of the blackmail. The Worsening principle’s clause (i) concerns how a proposal moves a deliberator from one choice situation to another. Here, moving should be understood causally rather than counterfactually. That is, the principle concerns whether the proposal in fact causes the deliberator to transition from choice situation 2 to choice situation 1. The clause should not be understood counterfactually because of unusual cases of overdetermination. Suppose that were the agent not to make a paradoxical proposal, a third party would make an identical paradoxical proposal that the deliberator ¹⁰ For defense of this view, see Owens, Shaping the Normative Landscape; Dougherty, “Why Does Duress Undermine Consent?” For a nearby view that would license similar reasoning, see Pallikkathayil, “The Possibility of Choice.”
18
would accept. Then it would be overdetermined that the deliberator will move from one choice situation to another. Because of this overdetermination, it would be the case that counterfactually, were the deliberator to reject the agent’s paradoxical proposal, the deliberator would not end up in a better choice situation. But these counterfactual considerations are separate from considerations of causation: the overdetermination does not prevent it being the case that the acceptance of the agent’s paradoxical proposal causes the transition in the deliberator’s choice situation. Let’s move on to clause (iii). To see why it is significant that the Worsening principle only applies if this clause is satisfied, recall the Steroids case. By threatening to drop Player from the team, Coach puts Player in a worse choice situation: Player would prefer to choose whether to take the steroids test with nothing hanging on their choice. But Coach has a good reason for making the threat: the threat promotes fairness in sport. Because of this reason, Coach does not infringe a claim of Player’s not to have their choice situation worsened. Since clause (iii) is not satisfied, the Worsening principle does not imply that Coach’s proposal invalidates Player’s consent. Clause (iii) makes use of the idea of a deliberator having a claim against their choice situation being worsened by an agent. I am not able to offer an account of this idea. This constitutes an important limit to my overall account of paradoxical proposals. But it is not a reason to reject the Worsening principle. And even in the absence of this account, we can rely on intuition to recognize acceptable and unacceptable ways of worsening someone’s choice situation. Intuitively, it is acceptable to do so in legitimate market transactions. For example, a supplier might propose to a shopkeeper, “Buy widgets from me or I will sell to your competitor who will thereby gain a competitive advantage.”¹¹ If the supplier previously had no intention to sell to the competitor, then the proposal would worsen the shopkeeper’s choice situation. But intuitively the proposal would not worsen the choice situation in a way that infringes a claim of the shopkeeper. On the other hand, in Lazy Snitch, Citizen plausibly infringes a claim of Criminal’s by making the paradoxical proposal. This is presumably to be explained in terms of the unacceptability of blackmail as a way of negotiating with others. To bring out what is distinctive about my Worsening principle, let me compare Niko Kolodny’s analysis of a case like Lazy Snitch. Kolodny’s
¹¹ Thanks to Sumeet Patwardhan and Larry Alexander for pressing me on this point.
19
analysis makes use of the following two definitions. First, an agent “conditions a response” to a deliberator’s choice by imposing a penalty on the deliberator if they do not comply with a demand (or a benefit if they comply with it). Second, an agent “announces a response” when they communicate to the deliberator that they have conditioned this response.¹² Kolodny defends a principle that is superficially similar to my Worsening principle: Choice. An agent’s conditioning or announcing a response to a deliberator’s choice wrongs them when and because it leaves their choice situation worse than they are entitled to from the agent.¹³ The difference between Kolodny’s principle and mine can be illuminated by considering the Lazy Snitch case. We saw that my Worsening principle implies that Citizen’s consent is invalidated. But Kolodny denies that their Choice principle applies to the case.¹⁴ To see why, consider the aforementioned choice situation (a):¹⁵ (a) If Criminal gives Citizen money, then Citizen does not snitch on Criminal. If Criminal does not give Citizen money, then Citizen does not snitch on Criminal. Criminal is not entitled to be in this choice situation because Citizen has good reasons to snitch. Consequently, the Choice principle does not imply that Citizen’s proposal wrongs Criminal. To elaborate the difference between the Worsening principle and the Choice principle, we can characterize Kolodny’s Choice principle more precisely: Choice. An agent’s conditioning or announcing a response to a deliberator’s choice wrongs them when and because: ¹² Kolodny, “What Makes Threats Wrong?” 89. ¹³ Kolodny, “What Makes Threats Wrong?” 93. ¹⁴ Kolodny makes this point with respect to the following case: Car Wash. Employer has legitimate business reasons for sacking Worker. Employer proposes that Worker washes Employer’s car to escape the sack. Kolodny does not have a settled position about what to say about this case. Kolodny floats the idea that the proposal is an expressive wrong: Employer expresses that they endorse not standing in relations of social equality with Worker. Kolodny, “What Makes Threats Wrong?” 104–5, 108–12. ¹⁵ See Kolodny’s Hush case that lies outside of the scope of the Choice principle. About this case, Kolodny claims it is unclear whether a wronging takes place, stating “I don’t have any firm intuition” about this issue. Kolodny, “What Makes Threats Wrong?” 103.
20
(i) the conditioning or the announcing puts the deliberator in choice situation 1; (ii) choice situation 1 is worse than another choice situation 2; and (iii) the deliberator has a claim to the agent making it the case that the deliberator is in choice situation 2.
While Kolodny does not formulate the principle in this way, it must be interpreted this way in order to justify Kolodny’s position that the principle does not apply to the Lazy Snitch case: the key reason is that Criminal is not entitled to be in choice situation (a)—i.e. clause (iii) of the Choice principle is not satisfied. By contrast, my Worsening principle does not rely on the claim that Criminal is entitled to be in choice situation (a). Instead, my Worsening principle focuses on the fact that Criminal is entitled to Citizen not worsening their choice situation. If you like, my principle does not invoke Criminal’s claim to occupy a particular choice situation, but instead Criminal’s claim against a transition from one choice situation to another. While the Worsening principle and the Choice principle have this crucial difference, these principles state only sufficient conditions. As such, these principles are logically consistent with each other. Therefore, reasons to accept Kolodny’s Choice principle are not reasons to reject my Worsening principle, and vice versa. My point is not that we must reject Kolodny’s Choice principle, but rather that we should accept my Worsening principle.
5. My account’s second prong: taking advantage of vulnerability to unjustifiably set back interests The first prong of my account does not cover paradoxical proposals that do not introduce disincentives but instead get their motivational efficacy from the existence of independent disincentives. Against the backdrop of these independent disincentives, these paradoxical proposals introduce incentives for consenting. Our earlier example was the Proactive Snitch case in which Citizen was independently poised to inform on Criminal. Holding fixed this independent disincentive, Citizen’s proposal introduces an incentive for Criminal to hand over money to Citizen by making it the case that this payment will secure Citizen’s silence. For these paradoxical proposals, our account needs a second prong. Do these proposals invalidate consent? Ultimately, I will say that they do not invalidate consent. However, I want to acknowledge upfront that we
21
might be tempted to say that they do invalidate consent when we consider quid pro quo sexual harassment: Planned Sacking. Because business is bad, Employer must lay off a staff member. Worker is one of two equally eligible candidates for who should be fired. Employer decides to fire Worker and calls them into Employer’s office to deliver the news. As Worker enters the office, an idea occurs to Employer, who says to Worker, “I will fire you if and only if you do not have sex with me.” Worker agrees and they have sex. Holding fixed the independent disincentive that Employer plans to fire Worker, Employer’s proposal introduces an incentive for Worker to have sex: sex will now cause Employer not to fire Worker. If we are inclined to deny that Worker’s compliance amounts to valid consent, then we should allow that this type of proposal can invalidate consent. What principle might explain why Worker invalidly consents? A morally significant feature of the case is that Worker is in a vulnerable position. This suggests the following view: Vulnerable Consent. Y gives invalid consent to X if X is taking advantage of Y’s vulnerability to get Y to consent. This view would also govern cases like the following: Sex Tourism. In a developing country, Indigent has no other employment options apart from engaging in sex work with tourists. Sex work is against Indigent’s religious principles, and Indigent experiences shame when engaging in it. Out of economic desperation, Indigent offers Tourist sex in return for money and Tourist accepts. Plausibly, Tourist wrongs Indigent. Suppose we accept this as a datum that an account of consent needs to explain (while leaving open the question of the nature of the wronging—i.e. the respect in which Tourist wrongs Indigent). This datum would be explained by the Vulnerable Consent thesis. However, it does not seem plausible that in general a vulnerable person’s consent is invalid. For example, consider consent to enter a home: Lodger. In a developing country, Homeowner struggles to make ends meet and so they rent out their spare room to tourists who are attending a
22
local language school. Homeowner rents out the room to Tenant at a very low price. I leave open the possibility that because the amount of rent is so low, Tenant wrongs Homeowner. But if this is a wronging, then it should be characterized as Tenant’s failing to meet a moral claim of Homeowner’s to be paid more. We should not say that Tenant fails to meet a moral claim of Homeowner’s that Tenant not spend time in their home. That is, Tenant does not wrong Homeowner in the way that someone would wrong Homeowner by trespassing in their home without their valid consent. But the Vulnerable Consent thesis implies that Homeowner’s consent is invalid. Therefore, we should reject the Vulnerable Consent thesis. In the next section, I will offer a separate argument for holding that Worker validly consents in Planned Sacking. Here I will focus on showing how we can account for Employer’s sexual misconduct without appealing to invalid consent. Let us start with an account that Victor Tadros develops to explain wrongings in cases that are similar to Planned Sacking.¹⁶ Tadros characterizes these wrongings as follows: Setback.
X wrongs Y if X unjustifiably sets back an interest of Y’s.
Here is how Tadros would apply the Setback thesis to Planned Sacking. Given the legitimate business reasons to sack Worker, Employer would be justified in setting back Worker’s interest in employment. However, the fact that Worker prefers employment to sexual integrity does not make it the case that these business reasons justify Employer in setting back Worker’s interest in sexual integrity.¹⁷ And indeed neither business reasons nor any other reasons justify Employer in setting back this interest. Therefore, even though Worker prefers employment to sexual integrity, Employer is still unjustified in setting back Worker’s interest in sexual integrity.
¹⁶ Tadros discusses the following case: Cop Threat: Bobby, a police officer, has discovered compelling evidence that Han, Yolanda’s husband, has committed armed robbery, for which he will serve several years in prison if he is convicted. Bobby tells Yolanda that he will bury the evidence if she has sex with him. To avoid him doing this, she has sex with him. Victor Tadros, Wrongs and Crimes, (Oxford: Oxford University Press, 2016), 228–32. ¹⁷ In saying that Employer sets back Worker’s interest in sexual integrity, we need not thereby assume that Employer sets back Worker’s interests in the aggregate.
23
While it is plausible to think that Worker’s interest in sexual integrity is set back, Tadros does not tell us what this interest amounts to. On the assumption that we are looking for an interest that explains why quid pro quo sexual harassment constitutes a wronging, we can rule out certain candidates. By making admittedly artificial stipulations, we can imagine a case in which Worker lacks various interests that could be specified independently of their participation in the sexual encounter.¹⁸ For example, we can imagine that Worker does not suffer experiential distress and that they do not have personal values concerning their sexual life with which the encounter is inconsistent. Even if we stipulate that the encounter does not set back these interests of Worker’s, the encounter still amounts to sexual misconduct. So if Tadros is right that the encounter sets back Worker’s interest in sexual integrity, this interest cannot be identified with those other interests. What then could be the key interest that explains why quid pro quo sexual harassment constitutes a wronging? The most plausible answer is that Worker has an interest in avoiding sexual subordination in their workplace. This derives from Worker’s more general interest in standing in equal social relations with others. Sometimes, there can be a justification for limited forms of subordination that set back this interest. For example, business reasons justify Employer having the asymmetric power to sack Worker. But these business reasons do not justify Employer in making use of this power relation to subordinate Worker in other ways. By making sex a condition of keeping their job, Employer takes advantage of the power relation to get Worker to serve Employer’s sexual interests. This is an especially problematic form of subordination insofar as it is an affront to Worker’s dignity.¹⁹ Worker’s interest in avoiding subordination is not merely a matter of their personal preferences: even if an individual does not care about subordination or willingly takes part in it, there can still be a setback to their interest in equal social relations with others. Thus, even though Worker gives valid consent, Employer still wrongs them by sexually subordinating them.²⁰
¹⁸ Thanks to Pascal Brixel for pressing me with these points. ¹⁹ For discussion of dignitary theories of sexual harassment, and the critique that these theories need supplementation with other theories to capture the gendered dimensions of harassment, see Elizabeth Anderson, “Recent Thinking about Sexual Harassment,” Philosophy and Public Affairs 34(3) (2006): 284–311, 291–7. ²⁰ An anonymous referee has pointed out that in light of this point, we could streamline Tadros’s account so that it is not framed in terms of setting back interests but instead is framed simply in terms of the wrong of sexual subordination. One advantage of framing the account in
24
Even with this additional elaboration of the interest in sexual integrity, Tadros’s account needs further supplementation. The Setback thesis does not mention the key notion of taking advantage of vulnerability. As such, it does not distinguish Planned Sacking from the following case: Bonus. CEO has discretion about how to allocate annual bonuses on the proviso that the allocation should be recompense for each worker’s performance on the job. One of the firm’s partners, Partner, merits a bonus of $1 million on top of their generous base salary. CEO makes a proposal with the following terms: “If Partner has sex with CEO, then Partner receives a bonus of $2 million. If Partner does not have sex with CEO, then Partner receives a bonus of $1 million.” They have sex. CEO wrongs Partner by sexually subordinating them. But CEO wrongs Partner less gravely than Employer wrongs Worker in Planned Sacking. This is because Partner is not subject to economic duress that makes them vulnerable with respect to the proposal. To mark this difference, we should add: Vulnerability Aggravation. When a victim is wronged by a setback to an interest of theirs, this wronging is aggravated when the victim occupies a vulnerable position. By combining the Vulnerability Aggravation thesis with the Setback thesis, we can say that an especially grave wronging takes place when X wrongs Y by taking advantage of Y’s vulnerability in order to set back an interest of Y’s without adequately compensating Y for the setback to this interest. If exploitation consists in unfairly taking advantage of someone’s vulnerability, then this implies that sexual exploitation is an especially grave wrong.²¹ To apply the Setback thesis to Lodger, we can make use of the following thesis: Compensation. If a setback to an interest has not been fairly compensated, then the setback is unjustifiable. terms of interests is that this makes the Compensation thesis (mentioned later in the main text) a plausible addition to the account, and this thesis can help explain the moral features of certain cases. ²¹ For defense of this view of exploitation, see Mitchell Berman, “The Normative Functions of Coercion Claims,” Legal Theory 8(1) (2002): 45–89, 85.
25
Then we can say that Tenant sets back Homeowner’s interest in having the spare room free without fairly compensating Homeowner for this setback to their interest. To support this analysis, let us imagine varying the strength of this interest. Suppose that the spare room is otherwise unoccupied and Tenant can access it without intruding on Homeowner’s privacy or otherwise affecting Homeowner’s life. Then Tenant wrongs Homeowner less gravely. This would be explained by the view that the gravity of the wronging is explained by the size of the inadequately compensated setback to an interest. The Compensation thesis also gives us a handle on the Sex Tourism case. Two morally significant features of that case are that sex work is against Indigent’s religious principles and that Indigent experiences shame for engaging in it. These factors are morally significant because money does not adequately compensate someone for compromising their religious principles or for the experience of shame. Indeed, if we imagine away these costs, then our moral analysis of sex tourism should be different. Consider: Sex Tourism II. In a developing country, Indigent has no other employment options apart from sex work. Indigent has no special objection to sex work that they do not have against other forms of service work. For Indigent, it is simply another job alongside factory work or bartending. Out of economic desperation, Indigent offers Tourist sex in return for money and Tourist accepts. It is not the case that Tourist wrongs Indigent at least as gravely in Sex Tourism II as they do in Sex Tourism. If there is a wronging in Sex Tourism II, then this would be explained by the fact that Tourist is paying Indigent too little. As such, the Sex Tourism II case is analogous to Lodger. Let us consider how the account we have been developing applies to our earlier case: Proactive Snitch. Citizen has discovered that Criminal was the masked person responsible for the bank heist. Criminal and Citizen both know that Citizen is proactive and hence otherwise will publicize this information. However, Citizen proposes not to publicize this information if Criminal pays them off. I maintain that if Criminal accepts the proposal, then Criminal’s consent is valid. There remains a further question as to whether Citizen wrongs
26
Criminal by unjustifiably taking advantage of Criminal’s vulnerability to set back an interest of Criminal’s without fair compensation. It is difficult to determine whether this is so. Clearly, Criminal is vulnerable and there is a setback to their financial interest in keeping the money. But it is hard to say whether they are fairly compensated by Citizen’s not informing on them. Clearly, Criminal prefers Citizen’s not informing on them to Criminal’s returning the money. But that does not by itself show that the compensation is fair. My sense is that if we imagine that the sum is sufficiently low—say $5— then Citizen does not wrong Criminal. If so, then this is explained by the Setback thesis. Finally, let us bring the foregoing discussion to bear on the Planned Sacking case. I take it to be a datum that Employer gravely wrongs Worker in this case. Our discussion has furnished theoretical resources for explaining this datum. The Setback thesis is that an agent wrongs a victim if the agent unjustifiably sets back an interest of the victim. Meanwhile, the Vulnerability Aggravation thesis is that this wronging is aggravated when the victim occupies a vulnerable position. Since Employer unjustifiably sets back Worker’s weighty interest in avoiding sexual subordination in the workplace, and since Worker is in a vulnerable position, these theses imply that Employer gravely wrongs Worker. Thus, we have an explanation of the datum even if we grant that Worker’s consent is valid. And next I will offer an argument that implies that we should grant this.
6. A further argument for my account’s second prong: the symmetry of initiating and accepting proposals In the last section, I argued that Criminal validly consents in Proactive Snitch. To further defend this claim, I will now offer a separate argument with two premises. My argument’s first premise is a claim about a nearby case: Interception. Citizen has discovered that Criminal was the masked person responsible for the bank heist. Citizen is about to walk into the police station to pass on this information. On the doorstep, Criminal intercepts Citizen and proposes to pay them off if Citizen does not inform on them. Citizen agrees. I take it to be clear that in Interception Criminal bribes Citizen. As an instance of corruption, it is impermissible for Citizen to accept this bribe.
27
However, Criminal cannot complain that their money has been extorted from them and they cannot subsequently demand that the bribe is repaid. But if there is no extortion for which Criminal is entitled to recompense, then Criminal validly consents to the transfer of money. Therefore, Criminal validly consents to the transaction in Interception. That is my argument’s first premise. My second premise is that with respect to the validity of consent Interception and Proactive Snitch are morally similar cases. I imagine that this premise may initially strike many readers as implausible in light of the fact that the cases differ with respect to who is the source of the paradoxical proposal. In Interception, Criminal is the person who gets the deal going. If Criminal had not intervened on the doorstep of the police station, then there would have been no deal and Citizen would have informed on them. I anticipate that many people will be inclined to think that this makes an important moral difference.²² It is tempting to think that if Citizen is the source of the proposal, then they commit blackmail and wrongfully extort money from Criminal, but if Criminal is the source of the proposal, then they are offering a bribe with the consequence that they validly consent to the transfer of money.²³ If this moral difference is made by the identity of the source of the proposal, then we should reject my second premise. I deny that the identity of the proposal initiator matters in this way. Recall our paradigmatic case of wrongful coercion that invalidates consent: Theft. Bully proposes to steal Victim’s sunglasses unless Victim loans Bully the use of their phone. Victim agrees.
²² With respect to the law, the difference has struck Sidney DeLong as sufficiently important that it gives rise to the “second paradox of blackmail”: it is not unlawful for one who knows another’s secret to accept an offer of payment made by an unthreatened victim in return for a potential blackmailer’s promise not to disclose the secret. What would otherwise be an unlawful blackmail exchange is a lawful sale of secrecy if it takes the form of a “bribe.” Sidney DeLong, “Blackmailers, Bribe Takers, and the Second Paradox,” University of Pennsylvania Law Review 141(5) (1993): 1663–93, 1664. ²³ I imagine that this analysis will strike many as intuitive on its own terms. It can also be motivated on the back of some accounts of the wrongfulness of coercion. For example, Stephen White argues that an agent wrongfully coerces a victim in virtue of transferring onto the victim responsibility that the agent themselves ought to shoulder concerning their own moral agency. This view places importance on the identity of the person initiating a proposal because this determines whether responsibility is transferred. Stephen White, “On the Moral Objection to Coercion,” Philosophy and Public Affairs 45(3) (2017): 199–231.
28
This case can be modified so that Victim is the source of the proposal: Counter. Bully credibly announces that they are going to steal Victim’s sunglasses. Victim counters that they will loan Bully the use of their phone if Bully does not steal the sunglasses. Bully agrees. Even though Victim is the source of the proposal in Counter, Victim’s consent to the loan of the phone is invalid. From this, we should draw the following two morals. First, the fact that someone is the source of an agreement does not mean that they validly consent to the agreement. Accordingly, the fact that Criminal is the source of the proposal in Interception does not entail that their consent is valid. This suggests that the salient difference between Proactive Snitch and Interception—the identity of the initiator of the deal—is not a morally relevant difference. The second moral to draw from comparing Counter and Theft is that whether Victim’s consent is valid depends on the consequences of Victim’s not agreeing to the deal and the moral status of these consequences. Specifically, if Victim does not agree to the deal, then Bully steals their sunglasses—a theft that has the moral status of a wronging. We can also apply this point to Interception: whether Criminal’s consent is valid should be explained in terms of the consequences of refusing and the moral status of these consequences. In other words, the fact that Criminal validly consents is to be explained by the fact that if the agreement does not hold, then Citizen informs and Criminal has no complaint against this informing. This is also true of the Proactive Snitch case. Therefore, the Proactive Snitch and Interception cases are alike in the morally important respects. Let me add two further considerations in support of this claim. First, in general, contracting parties are equally authors of an agreement, and this agreement is equally an exercise of their agency, regardless of who initially proposed the agreement. Suppose that you and I make an exchange: your car for my money. I would be no less a party to the transaction were I responding to your advertisement that the car is for sale. You would not be less of a party were I to have approached you with an unsolicited cash offer. But if in general both contractors are equally parties to a deal, regardless of who initiates it, then we have little reason to expect that the identity of the initiator makes a moral difference with cases like Proactive Snitch and Interception. Second, we can imagine variants of our earlier cases in which Citizen and Criminal are holding back from being the initiator of the proposal. They are both aware that they have a mutual interest in
29
striking the deal. Likewise, both are aware that the other is aware of this mutual interest. Consequently, both anticipate that the other is willing to propose the deal if necessary. In order to avoid being the person who initiates the deal, they enter into a waiting game. It is hard to believe that anything morally significant hangs on who is the first to cave. In light of this defense of my second premise, let me make a brief aside. Insofar as we aim to capture the morally significant phenomena, we should be interested in proposals that have certain terms, whoever initiates the proposal. It is in light of this point that our initial definition of a paradoxical proposal covered both the making of a proposal and the acceptance of a proposal initiated by another party. Let us return to my central argument of this section. So far, I have defended two premises. My first premise is that Criminal validly consents to the paradoxical proposal in Interception. My second premise is that with respect to the validity of consent Interception and Proactive Snitch are morally similar cases. From these two premises, it is a straightforward inference to the conclusion that Criminal validly consents in Proactive Snitch. Since this argument does not appeal to the specific terms of the paradoxical proposal in Proactive Snitch (i.e. cash for silence), parallel arguments could be offered for other paradoxical proposals with independent disincentives. Therefore, we have discovered a more general argumentative strategy that supports the conclusion that consent is not invalidated by paradoxical proposals that introduce incentives in circumstances where there are independent disincentives. And this strategy would apply to the Planned Sacking case: if Worker initiates the paradoxical proposal, then they validly consent to the sexual encounter. Moreover, the identity of the initiator does not matter for the validity of the consent. So Worker validly consents in the version of the case in which Employer initiates. Of course, this conclusion must be contextualized by the preceding section’s conclusion that Employer nonetheless gravely wrongs Worker in virtue of considerations unrelated to the validity of consent.
7. Conclusion We have discussed which paradoxical proposals invalidate consent that a deliberator gives in response to these proposals. I have offered a twopronged account that distinguishes paradoxical proposals according to
30
their causal relationships to disincentives that deliberators face in the event that they withhold consent: some disincentives are introduced by paradoxical proposals while other disincentives are already in place independently of these proposals. The first prong of my account concerns the former class of paradoxical proposal. For these, I have argued that such a proposal invalidates the deliberator’s consent in virtue of infringing the deliberator’s claim not to have their choice situation worsened. This is the implication of my Worsening principle. The second prong of my account concerns paradoxical proposals of the latter class. I have argued that these proposals do not invalidate consent. However, if an agent acts in accordance with a paradoxical proposal, they may still wrong a deliberator on other grounds, e.g. by taking advantage of the deliberator’s vulnerability in order to unjustifiably set back an interest of theirs. These prongs are independent in the sense that we can accept either prong while rejecting the other. However, I hope to have offered persuasive reasons that we should accept both as doing so allows us to answer the question of which paradoxical proposals can invalidate someone’s consent.²⁴
²⁴ For helpful comments, criticisms, and discussions, I would like to thank Larry Alexander, Abdul Ansari, Pascal Brixel, Shlomo Cohen, Johann Frick, Simone Gubler, Jordan Mackenzie, Sarah McGrath, Sumeet Patwardhan, Gideon Rosen, Sergio Tenenbaum, anonymous referees, the participants in my “Coercion and Consent” graduate seminar at UNC Chapel Hill, the participants in Johann Frick’s graduate seminar at Princeton University, and audiences at Virginia Tech, the 2019 UNC PPE Retreat, and the 2022 Arizona Workshop in Normative Ethics.
2 Manipulation, Disrespecting Autonomy, and Deliberative Projects Ron Aboodi and Shlomo Cohen
1. Introduction Some attempts to influence seem morally problematic in virtue of disrespecting a person’s autonomy. This applies to many instances of coercion and manipulation. We shall try to gain insights into the practical meaning of “disrespecting a person’s autonomy” by focusing on manipulation. There’s much controversy, and room for further debate, about the concept of manipulation, but some things are safe to say. Manipulation is a form of influence that differs from coercion or straightforward rational persuasion. Manipulation typically works either by mobilizing non-deliberative psychological mechanisms—such as emotions—or by trickery that diverts the manipulee’s deliberative efforts.¹ Such trickery could target the manipulee’s reasoning or the information that feeds it: providing false, irrelevant, or misleading information; or concealing true and relevant information. Our argument will not rely on any further (positive) conceptual claims about manipulation. Some think that manipulation is categorically impermissible, no matter what countervailing moral considerations are present. Much more common, and relatively more plausible, is the weaker position that manipulation is inherently pro tanto wrongful, in the sense that the manipulativeness of an act constitutes a moral reason against performing it, but this reason may be outweighed by other reasons.² These positions don’t square with the fact that there are instances of manipulation that seem morally innocuous, as we shall illustrate. (It is not a conceptual truth that manipulation is morally problematic.) But manipulation typically raises at least a suspicion of pro tanto wrongness. ¹ In line with Noggle (2020).
² E.g. Noggle (1996, 2018, 2020).
Ron Aboodi and Shlomo Cohen, Manipulation, Disrespecting Autonomy, and Deliberative Projects In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Ron Aboodi and Shlomo Cohen 2023. DOI: 10.1093/oso/9780198895909.003.0003
32
Different instances of manipulation can have different wrong-making features.³ But arguably, the moral worry that manipulation raises most typically is that it disrespects the manipulee’s autonomy. This worry applies to a vast variety of manipulations: malicious as well as benevolent, deceptive as well as non-deceptive, pressure-based as well as trickery-based, etc. Wrongful manipulations whose intended (or perhaps merely foreseen) influence problematically violates, undermines, or diminishes the manipulee’s autonomy—typically involve disrespecting her autonomy. This applies also to manipulations that recklessly (or unjustifiably) risk such effects.⁴ “Autonomy” has been understood in many different ways. How should “autonomy” be precisified (made more precise) in this context, as the relevant object of disrespect? This is the question upon which we’ll focus. Unlike most of the literature that examines the concept of autonomy,⁵ we will not offer an account of this concept and then explain disrespecting autonomy derivatively. Rather, our starting points will be case-based normative intuitions about manipulation. The importance of focusing on manipulation—rather than coercion—is twofold. First, coercion has been given much more attention in the literature on autonomy. Second, coercion usually tramples autonomy so blatantly that it becomes difficult to discern the factors that render an attempt to influence disrespectful toward the person’s autonomy. We will focus on manipulations—often those that play at the borderline of ethical permissibility—in order to expose these factors. We will argue that, in order to capture these factors and avoid irrelevant ones, the object of disrespect should be characterized as what we call “the manipulee’s deliberative project.” A person’s deliberative project is, as we shall explain, her interrelated ongoing efforts to manage her conduct by committing and adhering to normative stances, ideals, policies, plans, goals, and the like. “Disrespecting the manipulee’s autonomy” can be precisified as “disrespecting the manipulee’s deliberative project,” at least for practical purposes. We find this proposed precisification illuminating and practically helpful, because “disrespecting X’s deliberative project” (as we shall explain this notion) is more specific than “disrespecting X’s autonomy.” While we ³ For a nice survey see Blumenthal-Barby (2014). ⁴ See Baron (2014) on reckless manipulation. ⁵ E.g. Christman (2009). Although Christman (2009: 135) recognizes that autonomy “is the locus of interpersonal respect operative in personal relations which, in turn, places limits on paternalistic interferences or manipulative dealings with them,” he doesn’t base his account on our intuitions about these limits.
,
33
think that our proposal has implications for the right account of autonomy in other contexts as well, we will not try to defend or undermine any such existing account, let alone to offer one ourselves. Nor will we discuss the deeper normative question of why autonomy matters. In section 2 we highlight the significance of the manipulee’s deliberative commitments, such as normative stances, ideals, policies, and plans. In section 3 we propose precisifying “disrespecting the manipulee’s autonomy” as disrespecting the manipulee’s deliberative project. Section 4 concludes with a brief preliminary discussion of the implications to applied and theoretical ethics.
2. The Significance of the Manipulee’s Deliberative Commitments 2.1 Rationality and Responsiveness to Reasons We’ll start with three prominent accounts from the literature on the wrongness of manipulation. Although these accounts are not trying (like us) to precisify autonomy as the relevant object of the disrespect, there are two reasons for examining whether they can be used for this purpose adequately. First, these accounts are trying to characterize the inherent—or at least typical—wrong-making feature in manipulation; and disrespecting autonomy is a central candidate for being the most typical wrong-making feature in manipulation. Second, the notions to which these accounts appeal— rationality or responsiveness to reasons—are “in the vicinity” of autonomy.⁶ According to Robert Noggle (1996, 2018), the wrongness of manipulation stems from inducing a mental state that falls short of ideally rational functioning. To illustrate, Noggle (2018) argues that it is necessarily pro tanto wrongful to place the healthier food at eye level in order to influence the customers to choose it, as in Thaler and Sunstein’s (2009) famous Cafeteria case. Noggle assumes that both the salad and the unhealthy food are equally available despite the fact that the unhealthy food is on a lower shelf. Hence, the higher salience that is achieved by placing the salad at eye level doesn’t indicate that the salad is more available. The higher salience of the salad is, in fact, irrelevant from the perspective of rational ⁶ Some have incorporated responsiveness to reasons into their account of autonomy (e.g. Sher 1997).
34
decision-making. Making the decision in an ideally rational manner, according to Noggle, requires disregarding such irrelevant information. He claims that being influenced by such irrelevant information amounts to a “mistake,” by which he means “a mental state that falls short of ideally rational functioning,” and causing such mistakes (intentionally) is inherently pro tanto wrongful (Noggle 2020). Jason Hanna (2015) argues that the (deontological) wrongness of manipulation stems from what he calls “worsening the manipulee’s deliberative position.” This amounts to causing the target to either (1) act on the basis of bad reasons, or (2) fail to adequately reflect on reasons, or (3) attend to irrelevant considerations, or (4) place more (or less) weight on certain considerations than those considerations properly merit (Hanna 2015: 630–31). Moti Gorin (2018) takes a similar position, arguing that the wrongness of manipulation stems from deliberately causing the manipulee to behave in a way that is detached from (or fails to track) the reasons that ought to govern her behavior. As Gorin conceives of this detachment, it doesn’t require acting contrary to those normative reasons; it suffices that these reasons don’t play an appropriate role in motivating the agent. The first problem we would face if we were to use any of these three accounts to precisify disrespecting the manipulee’s autonomy is that there are innocuous types of manipulation that such precisification would seem to render disrespectful. We’ll refer to such incorrect verdicts as “false positives.” To illustrate, think of using makeup to conceal unattractive facial features from one’s romantic partner. Or think of manipulations that aim to conceal a surprise party that you are organizing for a friend. In both cases, there is a sense in which the manipulation induces a mental state that falls short of ideally rational functioning, or worsens the manipulee’s deliberative position, or causes her to fail to track some reasons. Nevertheless, it seems clear that—in some possible versions of these examples—the manipulation would be morally innocuous, without any disrespect toward the manipulee’s autonomy.⁷ We cannot think of any plausible criteria based on Noggle’s (1996, 2018, 2020) account which would render some versions of the surprise birthday and the makeup manipulations morally innocuous, while at the same time rendering the Cafeteria case pro tanto wrongful.
⁷ Additional examples of innocuous manipulations can be found in Sarah Buss (2005), an article which undermines the view that all manipulations are pro tanto wrongful.
,
35
Hanna’s and Gorin’s accounts seem to be in a better position to solve this problem, because they evaluate manipulation by comparing its effect to the counterfactual scenario in which the manipulator hadn’t intervened. (Noggle often ignores this counterfactual scenario, viewing the influence as pro tanto wrongful regardless of the alternative.) Accordingly, it may be suggested that Hanna’s and Gorin’s accounts avoid the false positives in the innocuous manipulation cases, due to the fact that in the counterfactual scenario wherein the manipulator hadn’t intervened—the manipulee would have been even less responsive to reasons overall. In what follows we’ll use the locution “diminishing overall responsiveness to reasons” to denote either worsening the manipulee’s overall deliberative position in the sense that is relevant to this version of Hanna’s (2015) account, or increasing overall detachment from reasons in the sense that is relevant to this version of Gorin’s (2018) account. The problem with evaluating manipulations by whether or not they diminish overall responsiveness to reasons is that this might lead to “false negatives”—incorrect vindications of morally problematic behavior. Consider the following case: Mail₁ Jane₁ and Manuel₁ are roommates. Manuel₁ often checks the mailbox for both of them, placing Jane₁’s mail on her desk, with her consent. One day a contribution request letter arrives for Jane₁ from the University of Kaching (UoK), her alma mater.⁸ Manuel₁ correctly and justifiably believes that Jane₁ is in significant financial hardship, needing the money much more than UoK. Nevertheless, Jane₁ is likely to contribute to UoK if she reads the letter, because she falsely believes, quite strongly, that she should contribute to UoK whenever they ask her to do so, even under financial distress. Looking at the bunch of envelopes he is holding perpendicularly to Jane₁’s desktop, Manuel₁ realizes that if he places them tilting to the right then UoK’s letter would appear on top. Manuel₁ decides—in light of the aforementioned facts—to tilt to the left, so that the letter from UoK will be placed at the bottom of the pile. Knowing that Jane₁ often leaves the envelopes at the bottom of the piles on her desk untouched for a very long time, he is
⁸ We thank an anonymous referee for the idea to make the relevant organization Jane’s alma mater.
36
hoping that she will ignore the contribution request, for her own good. It works: Jane₁ doesn’t open the letter and doesn’t make the contribution. Whether or not it is justified all-things-considered, Manuel₁’s subtle manipulation seems to disrespect Jane₁’s autonomy. But without Manuel₁’s intervention, it is likely that Jane₁ would have wrongly decided to contribute to UoK. In comparison to this counterfactual, the overall influence of Manuel₁’s intervention seems to better accord with the relevant objective normative reasons. This shows that manipulations can disrespect the manipulee’s autonomy without diminishing her overall responsiveness to reasons.⁹
2.2 Normative Stances To expose the factor that renders Manuel₁’s manipulation disrespectful toward Jane₁’s autonomy, consider the following variations on Mail₁: Mail₂ is identical to Mail₁ except that Jane₂ has no clear stand on whether, all-things-considered, she should contribute to UoK. She has a soft spot for UoK, but after it leads her to contribute she sometimes thinks, regretfully, that her contribution was unwise. All this is known to Manuel₂ and plays an important role in his decision to manipulate Jane₂. Mail₃ is identical to Mail₂ except that Jane₃ clearly and correctly believes that she shouldn’t contribute to UoK as long as she is in financial hardship. But contrary to her own judgment, her soft spot, which she shares with Jane₂, often leads her—when she doesn’t take the time to think it through—to contribute to UoK. Her contribution is always followed by regret, when she has time to reflect. All this is known to Manuel₃ and plays an important role in his decision to manipulate Jane₃. Compare all three versions of the case (Mail₁ to Mail₃) against each other: In which version does it seem clearer that Manuel disrespects Jane’s autonomy?
⁹ Gorin (2018: 246) writes that “an influencer cannot be obligated to limit her means of influence to reason-tracking forms of influence when the agent she intends to influence is not responsive to the relevant reasons.” This increases the worry that his account will generate a false negative concerning Mail₁.
,
37
In which version does Manuel’s disrespect toward Jane’s autonomy—if it exists—seem stronger? In which version does the manipulation seem more morally problematic? We hope that the reader sees the correlation between the answers to all three questions, and agrees that Mail₁ is the case where the relevant disrespect most clearly exists, is strongest, and most morally problematic. Mail₂ seems to take second place in all three respects. It is questionable whether Manuel₂ and Manuel₃ disrespect Jane’s autonomy; and if they do, the disrespect seems weaker and less morally problematic than in Mail₁. The main factor that distinguishes these cases is Jane’s normative stance— a term we shall use to denote not only normative beliefs but any position to which the person is committed (at least tentatively or in a particular context) concerning a theoretical normative question. Accordingly, the comparison suggests that Manuel₁’s manipulation disrespects Jane₁’s autonomy by disrespecting her normative stance. In contrast, Manuel₂ and Manuel₃ don’t intend to undermine Jane’s fulfillment of her normative stance. This explains why their manipulation seems less disrespectful—if at all disrespectful—toward Jane’s autonomy. Manuel₂ and Manuel₃ both aim to prevent Jane’s non-deliberative inclination from influencing her decision; an inclination that Jane doesn’t seem to endorse, rely upon, or view as a positive factor. Mail₃’s unique feature is that the non-deliberative inclination to contribute is antithetical to Jane’s own normative stance. Unlike Manuel₁’s and Manuel₂’s manipulations, Manuel₃’s manipulation aims to help Jane₃ to fulfill her normative stance, defending it from interference by her recalcitrant non-deliberative inclination. This seems to explain why, out of the three versions of Mail, it is most questionable that Manuel₃ disrespects Jane₃’s autonomy. Hence, the best precisification of disrespecting the manipulee’s autonomy would render it sensitive to the manipulee’s normative stances (so that disrespecting them is at least one way to disrespect the manipulee’s autonomy).
2.3 Deliberative Commitments But there is more to respecting autonomy (in the relevant context) than respecting the manipulee’s normative stances. Consider Mail₁*, which is identical to Mail₁ except that Jane₁* is deliberatively committed to a policy
38
of prioritizing giving to UoK over her own financial needs; a policy that isn’t (in her mind) based on a normative stance. If Manuel₁* manipulates her in the same way to prevent her from fulfilling this commitment, his manipulation would seem similarly disrespectful of her autonomy (and similarly pro tanto wrongful). So not only normative stances but also other types of deliberative commitments—such as policies and plans—are relevant for the type of disrespect that we are trying to characterize. Our view on the nature of such deliberative commitments is inspired by Michael Bratman’s work.¹⁰ The defining feature of deliberative commitments is their primary function: to constrain the person’s practical deliberations by treating some practical questions as settled, at least tentatively. Sometimes a person treats a practical question as settled due to her normative stance (her position on a theoretical normative question). But this is not a necessary feature of ideals, policies, and plans, as we use these terms. The state of being committed to a policy, for instance, amounts to treating certain choices as settled in practical deliberation. Once a person commits to a policy, it acquires a “default” status of non-reconsideration in related deliberations. Nevertheless, most commitments are tentative in the sense of allowing reconsideration when there’s some justification to do so. All of this applies also to ideals, which have a higher level of generality than policies, and to plans, which (typically) have a lower level of generality than policies. To appreciate the relevant significance of the manipulee’s deliberative commitments, let us reflect on what versions of the case of concealing surprise party plans would come closer to disrespecting the manipulee’s autonomy. We can easily imagine versions that are not disrespectful by imagining the manipulator as knowing that the manipulee values having a certain kind of exciting or pleasurable experience—at least on the relevant occasion—more than he values what the manipulation undermines: being in control, or complete transparency in the relationship. Versions of the case that do not meet this condition seem to come closer to disrespecting autonomy. Consider a person who makes it very clear to his friends that, for him, excitement and pleasure are always secondary to being in control and having complete transparency in all his relationships. If his friends nevertheless secretly organize a surprise party for him, this will seem disrespectful toward his autonomy.
¹⁰ E.g. Bratman (2007).
,
39
Our emphasis on deliberative commitments brings to mind the “internalist” (or “coherentist”) accounts of autonomy or self-governance, which aspire to identify—in compatibilist terms—that core part of the self which governs in what is considered “self-governed behavior.” Several candidates have been offered for this core part of the self, such as Watson’s (1975) evaluational system, Frankfurt’s (1987) wholehearted identification, or Bratman’s (e.g. 2007) governing policies. But our theoretical task is not to be confused with that of such internalists. Regardless of which of the internalists’ candidates is indeed the relevant core part of the self, intentionally interfering with the functioning of any of these candidates usually seems disrespectful toward the person’s autonomy. Moreover, interfering with a person’s mundane policies and plans that are “distant” from this core can also amount to disrespecting her autonomy. Consider Mail₁**, which is identical to Mail₁*, except that Jane₁**’s policy is based on a factual mistake concerning the effectiveness of UoK’s use of money, a mistake that Manuel₁** does not make; and this is the only reason for the difference in their stances about whether Jane₁** should contribute to UoK. Nothing in the description of the case suggests that Manuel₁**’s manipulation disrespects Jane₁**’s evaluational system. Nor does it seem to disrespect her wholehearted identification or Bratman-type governing policies, or any similar internalist candidate. Nevertheless, it seems to disrespect her autonomy. In order to precisify autonomy as the relevant object of disrespect in the broadest manner, which captures all the typical ways in which manipulation can disrespect the manipulee’s autonomy—we need a comprehensive notion that encompasses all the “core of the governing self” candidates offered by internalists, and more.
3. Disrespecting the Manipulee’s Deliberative Project Section 2 identified various types of deliberative commitments that matter for the evaluation of whether and how much an instance of manipulation disrespects a person’s autonomy. These include fundamental values and goals as well as strategies to fulfill them; the relevant strategies being policies and plans (even mundane plans). But in order to precisify “autonomy” as the object of disrespect in typically wrongful manipulation, we need a holistic notion that includes—but is more than the sum of—the person’s actual, current deliberative commitments, read de re. In order to fulfill
40
certain deliberative commitments—especially ideals—one sometimes needs to change other deliberative commitments (not only mundane plans but even other ideals and fundamental normative stances). As we’ll illustrate further on, interfering with the dynamics of such changes may sometimes be disrespectful toward the person’s autonomy. This suggests that the object of such disrespect—which we call “the person’s autonomy”—is a dynamic entity. Which dynamic entity? Notice that it is not the content of the person’s deliberative commitments that seems to generate an autonomy-based reason to respect them. Rather, it is the fact that the person is committed and adheres to them. Accordingly, the dynamic entity we’re looking for is the activity or project of committing and adhering to ideals, policies, plans, and the like. This activity is sometimes called “self-governance,” “self-determination,” or—to come closer to the mundane parts of life— “managing one’s conduct.” Managing one’s conduct clearly differs from being moved non-reflectively by what we call non-deliberative motivation: merely habitual behavior, instinctive or uncontrolled reactions, visceral emotional mechanisms, or non-deliberative inclinations. Nevertheless, the project of managing one’s conduct relies on non-deliberative motivations. For example, without certain “habits of non-reconsideration” (Bratman 2007) it seems that deliberative commitments wouldn’t be able to fulfill their function of constraining deliberation. To bring a more mundane example: without automatic “system-1” mechanisms it doesn’t seem that we’d be able to drive safely. Any plan to drive anywhere relies on such habits. Accordingly, by influencing a person’s non-deliberative motivations one can influence this person’s project of managing her conduct. And some such influence could be disrespectful to the project.
3.1 Our Proposal In light of this analysis, we propose to precisify “disrespecting the manipulee’s autonomy” as “disrespecting the person’s deliberative project,”¹¹ where a person’s deliberative project is her interrelated ongoing efforts to manage her conduct by committing and adhering to deliberative commitments. ¹¹ The term “deliberative project” has been used in a different context in Enoch (2011). The pursuit of what John Rawls (1971) calls “a life plan” counts as a deliberative project. But we are not committing here to any of Enoch’s or Rawls’s conceptual requirements or assumptions. Our use of the term is in line with Aboodi (2021).
,
41
By “adhering to commitments” we mean trying to fulfill the commitments successfully. This implies that you can interfere with a person’s deliberative project by restricting her physically, blocking some of her options, etc.¹² Another implication is that interfering with non-deliberative motivations upon which the person’s policies or plans rely can amount to interfering with the deliberative project. For instance, a person may consider a particular habit as conducive to one of her goals and rely on the workings of this habit when forming policies and plans. In such a case, if we interfere with this person’s habit, we’d be interfering with her deliberative project. To take another example, a person may consider anger as justified in a particular type of situation. (When she reflects on this type of situation she envisions being moved by anger as supported by her ideals, beliefs, and goals.) As a result, in the relevant type of situation, calming down this person’s anger might interfere with her deliberative project. By contrast, when a person’s anger interferes with fulfilling her deliberative commitments—given that she hasn’t relied upon her anger when forming them—calming her down would probably be conducive to her deliberative project. One may disrespect deliberative projects in various manners.¹³ The obvious manner is intentionally interfering with it. Unjustifiably risking such interference is also typically disrespectful. We’ll focus on these two manners of disrespecting deliberative projects. But there are additional manners, such as neglecting to help a person to fulfill her deliberative project, neglecting to acknowledge a person’s deliberative project, belittling it, mocking it, etc.¹⁴ Whether a particular action or attitude is disrespectful toward another person’s deliberative project, and whether such disrespect is morally problematic, varies from case to case and depends on many factors that we won’t try to identify here. Our aim is to precisify “autonomy” in “disrespecting the manipulee’s autonomy,” not “disrespecting.” Our characterization of the notion of deliberative project purposely leaves room for further specification (in future research) concerning what unifies ¹² In line with Marina Oshana (1998: 82). ¹³ Note that we refer to disrespect toward the performance of a deliberative project, rather than toward the capacity to deliberate (well), or to form commitments and adhere to them (successfully). Manipulations whose aim is to improve such a capacity might still be disrespectful toward the manipulee’s autonomy if they are antithetical to the manipulee’s commitments and attempts to fulfill them. We return to this point in section 3.4. ¹⁴ Concerning educators (including parents), respect for the learner’s deliberative project should arguably include minimizing the hindering of its future development (at least in legitimate and reasonable directions). On the other hand, educators should also strive to make the learner’s future deliberative projects better, morally and prudentially. For an initial discussion of these issues (focusing on manipulation in education) see Aboodi (2021).
42
and how we should individuate deliberative projects. For instance, it might make sense to distinguish between the person’s current and future deliberative projects, in parallel to a distinction between the autonomy of the person’s current and future self.¹⁵ These matters should be settled in accordance with the non-trivial resolution of the relevant normative questions.
3.2 The Adequacy and Comprehensiveness of Our Proposal We believe that our proposal nicely coheres with our pre-theoretic intuitions concerning the cases we raised and others. The more the manipulation disrespects the deliberative project, the more it seems to disrespect the manipulee’s autonomy. This applies, for a start, to the surprise birthday party and makeup examples: the more the intended and foreseen influence of these manipulations is in tension with the manipulee’s deliberative project, the more it seems to disrespect her autonomy. The same applies to the Mail variations. In Mail₁, Mail₁*, and Mail₁** the manipulation hinders part of the manipulee’s deliberative project by hindering the fulfillment of her policy to contribute to UoK whenever UoK makes the request. This explains why the manipulation seems to be pro tanto wrongful for disrespecting the manipulee’s autonomy, despite being conducive to her well-being. In Mail₂ and Mail₃ the disrespect seems less significant, if it exists at all. This coheres with our account, because the aim of these manipulations is no longer antithetical to Jane’s policy to contribute. However, as Mail₂ and Mail₃ are described, they leave room for the suspicion that these manipulations are (or risk being) antithetical to other elements in Jane’s deliberative project. For example, some people value authoring their life on their own to such a degree that they prefer to make “their own” mistakes over being protected by subtle paternalistic manipulation. Similarly, some people value relationships that are safe from manipulation to such a degree that they prefer to lose all benefits that paternalistic manipulations could offer them (which can go beyond mere romantic excitement and surprises). If Manuel has reason to think that Jane is likely to fall under any of these categories (or, perhaps, if
¹⁵ Once this distinction is made, this article can be described as focusing on disrespect toward the autonomy of the manipulee’s current self, elucidating it by appeal to the manipulee’s current deliberative project.
,
43
he has no evidence to the contrary), this could render his paternalistic manipulation somewhat disrespectful toward her autonomy even in Mail₃. Notice also that coping with challenges, mistakes, and failures can be conducive to the development of a person’s capacity to manage her conduct in a better way, and paternalistic manipulations risk inhibiting this development. Taking such a risk on behalf of another person, without sufficient evidence that she would prefer that you do that, would sometimes be disrespectful toward her deliberative project. To illustrate, imagine Jane₂ in a scenario where Manuel never intervenes to prevent her from contributing to UoK. Suppose she makes the contribution each time it is requested of her, and as expected, this leads to further financial deterioration. The negative impact on Jane₂’s well-being might lead her to reflect on her conduct and internalize her financial problem and its prudential significance. One potential result could be settling on a policy to avoid contributions to UoK until her financial situation improves. In other words, if Manuel₂ doesn’t prevent Jane₂ from making a mistake, she might learn from her mistakes. So by intervening Manuel actually risks interfering with her process of learning from experience. Insofar as (the manipulator’s evidence suggests it is sufficiently likely that) the person sufficiently values such growth, inhibiting it could be somewhat disrespectful toward the person’s deliberative project. Arguably, a similar risk exists in Mail₃, though less significant. Manuel₃’s manipulation aims to prevent Jane₃’s anticipated executive failure: being moved by an inclination out of weakness of will. But Jane₃ has the potential to learn from such failures how to better manage her conduct: to resist such inclinations, or to find indirect means that would protect her deliberative project from them (such as asking Manuel to recycle such contribution requests until further notice). Manuel₃’s manipulation risks interfering with this process of learning from one’s failures. Perhaps Manuel₃ has some reason to think that Jane₃ values such growth so much that taking this risk would be disrespectful toward her deliberative project. Whether or not taking such risks amounts to disrespecting Jane₃’s deliberative project, these possibilities could at least explain why Manuel₃’s manipulation raises the suspicion for some readers that it disrespects Jane₃’s autonomy. Having said all that, to the extent that we can imagine a specification of Mail₃ where Manuel is totally respectful of Jane’s deliberative project, we think that his manipulation could be fully respectful of Jane’s autonomy. Remember that in all the Mail cases, Manuel doesn’t deceive Jane about the contribution request or restrict her options: if she only looks through her new mail she will easily find the contribution request and will be able to contribute. Nothing
44
obligates Manuel to tilt his hand to the right rather than the left whenever placing the envelopes on her desk. Indeed, he chooses the left as part of a plan that exploits Jane’s nondeliberative tendency to ignore the letters at the bottom. But such plans don’t necessarily disrespect the manipulee’s autonomy.¹⁶ As long as Manuel₃ has sufficient evidence that none of Jane₃’s stances or other deliberative commitments render his intervention problematic or unwanted, it seems that it would not be disrespectful toward her autonomy. Most of this paper focuses on problematic paternalistic manipulations (aiming at the good of the manipulee), because this is usually where concern for the manipulee’s well-being and respect toward her autonomy come apart. This helps us isolate disrespect toward the manipulee’s autonomy from other wrong-making features, which helps to minimize the contamination of the intuitions to which we appeal. But our proposal is a fortiori relevant for non-paternalistic manipulations as well.¹⁷ Whether paternalistic or not, manipulation typically raises a suspicion that it is disrespectful toward the manipulee’s deliberative project, because it typically works either by mobilizing non-deliberative psychological mechanisms—such as instinctual emotional reactions—or by trickery that diverts the manipulee’s deliberative efforts. Wrongful manipulations that mobilize non-deliberative mechanisms, such as playing on someone’s guilt or vanity (in a wrongful manner), typically disrespect the manipulee’s deliberative project by interfering with the deliberative processes that serve it, or with their causal efficacy. Wrongful manipulations that rely on trickery, such as false advertisement and fake news, typically disrespect the manipulee’s deliberative project by diverting these deliberative processes (for example, by feeding disinformation). Hence, our proposal coheres with the view that disrespecting autonomy is the most typical wrong-making feature in manipulation.
3.3 Comparison to Some Neighboring Accounts In what follows we will elucidate our proposal further and explain some of its advantages by comparing it to two alternative accounts “in the same neighborhood.”
¹⁶ Enoch (unpublished manuscript) might lead some to worry that such interventions necessarily disrespect the target’s autonomy. Ron Aboodi addresses such worries in a paper in production. ¹⁷ For example, Iago’s paradigmatic manipulation in Shakespeare’s Othello clearly disrespects Othello’s deliberative project. This fact shouldn’t be overshadowed by—and cannot be reduced to—other wrong-making features of Iago’s manipulation (such as intentional harm).
,
45
3.3.1. Undermining the Capacity to Manage One’s Concerns Eric Cave (2007) aims to explain the (pro tanto)¹⁸ wrongness of a type of manipulation that he calls “motive manipulation.” As Cave (2007: 131–32) defines motive manipulation, one necessary condition is that it works by mobilizing some “non-concern motive”—that is, a motive that either (a) doesn’t constitute a pro- or con-attitude, or (b) is inaccessible to consciousness. Cave (2007: 138) argues that the relevant wrongness stems from a principle which he calls Modest Autonomy: “individuals should refrain from activities that threaten to undermine the capacity of others to manage their concerns.” Whereas Cave (2007) restricts his inquiry to motive manipulation, we aspire to precisify autonomy as the object of disrespect that is typical in manipulations of all kinds. Manuel’s manipulation in Mail₁, for instance, doesn’t count as motive manipulation. So it is not surprising that Cave’s (2007) account won’t suffice for our purposes. Disrespecting the manipulee’s deliberative project cannot be reduced to threatening to undermine the manipulee’s capacity to manage her concerns. Indeed, undermining this capacity would often be disrespectful toward the manipulee’s deliberative project. But a manipulator could interfere with the manipulee’s fulfillment of her deliberative project merely by making it harder for her to fulfill her existing deliberative commitments, without threatening her capacity to change any of her motives. The last point also raises the worry that Cave’s (2007) account cannot adequately characterize the typical wrongness of motive manipulation unless amended to better cohere with our own proposal. Consider the way Cave (2007: 140–41) analyzes the following case: “Suppose your concerns incline you to appear solemn at department meetings, and knowing you defenseless against bad puns, I get you to laugh by working one into my latest proposal. In getting you to laugh, I have set back your concern to appear solemn, but one might doubt that I have interfered with your capacity to manage your concerns. Laughter is just not that debilitating. Here, one might think, we have motive manipulation, but no real threat to the target’s capacity to manage her own concerns. But if I get you to laugh by successfully appealing to a non-
¹⁸ Although Cave (2007) describes this wrongness using the term “prima facie,” the way he uses this term suggests that he has in mind pro tanto wrongness. This latter term is officially adopted by Cave (2014) in a related paper that aims to explain the wrongness of unsavory seduction along similar lines.
46
concern urge within you, then while you are in the grip of this urge, you cannot strive towards your concern to appear solemn. This is even clearer if you are terribly ticklish, and I get you to laugh by tickling you. By getting you to laugh in either of these ways, I interfere with your capacity to supplement, winnow, reorder, revise, or strive towards your concerns as you see fit, however briefly.”
According to us, the (seeming) fact that the manipulee values appearing solemn is what makes the manipulation seem problematic: the manipulation is disrespectful toward this aspect of the manipulee’s deliberative project. However, Cave (2007: 140–41) distinguishes between (a) setting back the target’s concern, and (b) putting her in the grip of an urge that briefly prevents her from striving toward her concern. Cave thinks that (b) is the main moral problem here. We find this position highly questionable. Consider a case where the target doesn’t have any concern for appearing solemn at department meetings, but rather values having a good laugh on such occasions. This doesn’t change the fact that by getting the target to laugh, the joke teller briefly hinders the target’s capacity to manage her concerns (or deliberatively strive toward them). Cave’s (2007) account renders such joke telling pro tanto wrongful, which amounts to a “false positive”: such joke telling is normally morally innocuous. In our opinion, the best way to defend Cave’s (2007) account would be to amend it in accordance with our diagnosis. In particular, the risk of setting back the manipulee’s concerns has a central role in the typical wrongness of manipulation. Typically, the manipulee’s capacity to manage her concerns matters only when, and because, it is important for managing and fulfilling them as she deliberatively sees fit. Plausibly, this generalizes to other capacities that are central to other accounts of autonomy.
3.3.2. Options and Endorsement Jennifer Blumenthal-Barby (2014: 126)—drawing on John Christman and Harry Frankfurt—suggests the following criteria for assessing the extent to which an attempt to manipulate is or is not compatible with preserving the manipulee’s autonomy: 1. The extent to which the attempt blocks or burdens options; and— 2. The extent to which the person influenced is aware of and endorses, or were he aware would endorse, the relevant attempt as a process by which his desires, decisions, or actions were formed.
,
47
Let us briefly examine these criteria. Although Blumenthal-Barby (2014) doesn’t spell it out, the general thought behind the first criterion is that the fewer options one has, the less autonomous one is. But while this may often be the case, it is not always the case. As Gerald Dworkin (1982) argues, we have many good reasons to limit our range of options; including the related decision costs, and the possibility of error and temptation. Enoch (2022) argues that sometimes it is irrationally fetishistic to pay a price to secure an option that one rightly considers as inferior to some other available option. These considerations challenge Blumenthal-Barby’s first criterion. But both of these challenges can be adequately met if this criterion is amended in accordance with our proposal. In order to assess whether and how much blocking or burdening options disrespects the person’s autonomy, what matters is whether and how much having these options is valuable according to the person’s deliberative project. Blocking or burdening bad options would seem disrespectful toward a person’s autonomy whenever she takes them to be good options and this stance guides her deliberative project— even if this stance is mistaken, as in Mail₁. And (as Mail₁* and Mail₁** illustrate) not only the normative stances matter; the person’s strategy to fulfill them is also part of her deliberative project, and it affects how much having the relevant option is valuable according to her deliberative project, and how disrespectful it would be to block or burden it. Even if a person values having a choice in a manner that seems irrationally fetishistic—such as willingness to pay a price to secure an option that she rightly considers as inferior to some other available option—this valuing should normally be respected, at least prima facie and to some extent. The general thought behind Blumenthal-Barby’s second proposed criterion is that desires, decisions, and actions, that are shaped by some influence that one wouldn’t endorse, would be less autonomous in virtue of such influence. Blumenthal-Barby does not fully spell out what sense of endorsement she has in mind, but she repeatedly refers to Christman (2009) and at some points seems to equate endorsement with Christman’s notion of lack of resistance. This makes sense because requiring active endorsement, which is more than indifference, would seem to generate false positives.¹⁹ However, if we merely require lack of resistance, without any idealizations of the motivation that leads to it, Blumenthal-Barby’s second criterion
¹⁹ For example, if someone is known to be indifferent to the possibility of being manipulated by his friends to conceal a surprise party that they are organizing for him, it seems that doing so wouldn’t disrespect his autonomy.
48
would generate some incorrect verdicts. It doesn’t seem that we should count, for example, hypothetical non-resistance motivated by an involuntary addiction (from which the manipulee would like to free himself). Christman (2009: 155) is aware of this caveat and adds the requirement that the relevant hypothetical lack of resistance would be motivated by “sustained critical reflection”—that is not constrained by distorting factors—on the effect of the manipulative influence “over a variety of conditions in light of the historical processes (adequately described) that gave rise to it.”²⁰ Additional elucidations of this hypothetical non-resistance test are needed to avoid complications, but we’ll focus on just one suggested specification. To prevent this idealized hypothetical non-resistance test from generating wrong verdicts on whether a manipulation disrespects a person’s autonomy, we suggest that the manipulee’s deliberative commitments at the time of the manipulation should be held fixed during her hypothetical critical reflection. This would allow the test to generate the intuitive verdicts in our previous examples: Jane₁ would have probably resisted the manipulation, and if Jane₃ has no general preference against such paternalistic interventions (as discussed in section 3.2) she would have probably not resisted. Once this suggested specification is accepted, (some version of) this test would track the extent to which the manipulation disrespects the agent’s deliberative project. And this tracking seems to us as the best justification for using (this version of) this test. In sum, our proposal offers a unified explanation (without invoking the notion of autonomy) for the justifiability of Blumenthal-Barby’s proposed criteria, correctly amended: they can be used as indicators of the extent to which the manipulation disrespects the manipulee’s deliberative project.
3.4 External Standards Are there standards such that, regardless of whether they are valuable according to the person’s deliberative project, it would disrespect her autonomy to diminish the conformity of her deliberative project to these standards?
²⁰ Christman (2009: 155) adds an additional complicated idealizing condition that is so strict that, if used in our context, it would surely generate false negatives.
,
49
Arguably, rationality and responsiveness to reasons would be the leading candidates to serve as such external standards. But in light of our discussion in section 2.1, we should be cautious about incorporating these notions into our precisification. First, we want to capture the disrespect toward the manipulee’s autonomy in some manipulations—such as Mail₁—that aim to increase the manipulee’s overall responsiveness to reasons. Second, we don’t want any “false positives” in cases such as concealing unattractive facial features or surprise party plans. So we don’t want to brand every manipulation that intentionally causes behavior (or functioning) that is not ideally rational or responsive to reasons as disrespectful toward the manipulee’s autonomy. This suggests that standards such as rationality or responsiveness to reasons matter for our particular purpose only when, where, and to the extent that the manipulee is committed to them. If so, they seem to matter only as “internal,” rather than “external” standards; their relevant significance is derived from the person’s deliberative commitments. Clearly not everyone values thin ideals such as rationality in every context and to the same extent. This seems to generalize to every candidate external standard. For any such external standard, it seems that we can construct examples that show the following: The extent to which diminishing the manipulee’s conformity to this external standard would disrespect the manipulee’s autonomy depends on the manipulee’s relevant deliberative commitments. This applies—somewhat paradoxically—even to the standard of being the author of one’s life story. Some religious devotee may value following the instruction of the Rabbi, for example, much more than being the author of her life story. In some such cases, the aim of freeing her from the Rabbi’s influence in order to increase the self-authorship of her life would be disrespectful toward her autonomy (at least the autonomy of her current self).²¹ Some might think that the fact that there are objective requirements for what counts as a deliberative project implies that respecting a person’s deliberative project calls for maintaining, or enhancing, or at least not diminishing the capabilities that underlie these requirements. But this doesn’t always follow. Most people’s deliberative projects would, indeed, value maintaining and even developing these capabilities, because they are conducive for a large variety of goals. However, there are self-destructive
²¹ Perhaps respecting the autonomy of her current self and respecting the autonomy of her future self come apart here. We have been focusing on the former, and perhaps the latter should also be taken into account. We leave these possibilities for future research.
50
goals for which these capabilities are not so conducive. Such goals can be part of a deliberative project despite the self-destructive nature of their content. A deliberative project that is self-destructive—based for instance on a nihilist view, or compatible with letting oneself get addicted to hardcore drugs—is still a deliberative project. Fully respecting such deliberative projects requires respecting their destructive aspects as well. Interventions that are antithetical to these aspects should not be justified by appeal to respecting the manipulee’s (current self) autonomy. (They could be justified by appeal to other moral considerations, such as improving the manipulee’s well-being.) This applies not only to deliberative projects that include explicitly destructive goals but also to more common deliberative projects that merely value the relevant capabilities less than they value some other goal.
4. Conclusion Our conclusion is that, insofar as the locution “the manipulee’s autonomy” functions to characterize the relevant object of disrespect, it can be precisified (at least for practical purposes) as “the manipulee’s deliberative project,” which includes committing and adhering to deliberative commitments. We argued that this proposal coheres with all of our case-based intuitions concerning the extent to which manipulation disrespects the manipulee’s autonomy, generating no “false positives” or “false negatives.” Our proposal seems to capture the typical ways in which manipulation can disrespect a person’s autonomy, more comprehensively and adequately than any alternative we can think of. Our notion of deliberative project is more comprehensive than any of the “internalist” (or “coherentist”) notions of self-governance, but it may be classified in the same family with them, enjoying their advantages of being consistent with compatibilism, and carrying minimal (if any) metaphysical load. Whereas we characterized the relevant notion of autonomy as a project or an activity, others (e.g. Gerald Dworkin, T. M. Scanlon, or, as we have seen above, Eric Cave) characterize it as a capacity. But the notion of disrespecting a capacity seems to stretch the ordinary meaning of disrespect and necessitates further substantiation before one can put it to practice.²² It
²² See Dean (2021).
,
51
seems relatively easier to understand what “disrespecting P” amounts to when “P” is a particular activity. Moreover, it seems that one general lesson that may be drawn from our argument in section 3.3.1 is that capacities matter (in the relevant context) only to the extent that they are important for the fulfillment of the person’s deliberative project. We argued against the view that it disrespects a person’s autonomy to diminish her deliberative project’s conformity to some objective normative standard (such as rationality) even when this standard is in tension with her deliberative project. This goes against a rationalist Kantian tradition that reaches all the way to Noggle’s account of the wrongness of manipulation. More generally, our argument implies that we cannot incorporate the external standards that feature in “externalist” or “objectivist” accounts of autonomy into the precisification of autonomy as the relevant object of disrespect. This applies, for example, to Christman’s (2009) causal-historical conditions for autonomy. Relatedly, if a normative principle of respecting how people choose to lead their lives would be constructed on the basis of our proposal, it would be quite wide and inclusive. At the same time, we don’t rule out the possibility that external standards affect how much a deliberative project should be respected. The matter depends on a deeper question that we shall now touch on.
4.1 Theoretical Implications We haven’t tried to answer the deeper question of why, when, and to what extent deliberative projects should be respected. Any answer will be controversial. Some consequentialists might explain the importance of respecting deliberative projects on the basis of some impersonal consequentialist principle.²³ Arguably, managing one’s conduct as one sees fit is an important aspect of leading a good life. Others would argue that the obligation to respect people’s deliberative projects is a basic deontological obligation, or stems from a different deontological obligation that is more basic. Some may take this more basic obligation to be what we owe to each other, or the obligation to respect persons. Since the starting points of our argument are
²³ In similarity to J. S. Mill’s defense of liberty on utilitarian grounds in his On Liberty. The central challenge for consequentialists here is to account for the deontological intuition that we should sometimes respect a person’s deliberative project even when this conflicts with the greater good.
52
case-based intuitions that any such ethical theory must cohere with (or at least justify ignoring), so is our conclusion. Some may try to argue that the obligation to respect deliberative projects stems from the obligation to respect people’s autonomy, in some different sense that is irreducible to their deliberative project. Having been focusing on the practical meaning of disrespecting the manipulee’s autonomy, we haven’t said enough to rule out such a view. But any such view would have to explain what is “the person’s autonomy” over and above “the person’s deliberative project,” and how this broader notion of autonomy illuminates the normative issues. Lastly, there are potential implications to theories about which person or choice counts as autonomous. Clearly, the ability of any autonomous person to author her life story, or to create her own evaluational system, is at best partial (pace Sartre). Hence, such theories must set a threshold for the sufficient level of self-determination that counts as autonomous. How is such a threshold to be set? In light of our proposal, one possibility worth exploring is that the relevant threshold should be set according to the minimal conceptual requirements for having a deliberative project. While these requirements need further specification, our argument narrows down the options. In particular, a choice can merit respect even when it fares poorly according to external standards such as “being the author of one’s life’s story” or responsiveness to reasons (see section 3.4). This would arguably set the relevant threshold lower than prominent alternatives.
4.2 Implications to Applied Ethics The central practical implication of our argument is that, in order to determine whether and how much a particular instance of manipulation is morally problematic, one of the most important factors to examine is whether and how much it respects the manipulee’s deliberative project. (We shouldn’t ignore personal deliberative projects and focus solely on ideally rational functioning—pace Noggle—or on the reasons that ought to govern the manipulee’s behavior, pace Gorin.) Despite the room we left for further specification of this formulation, we think that this formulation is already less ambiguous, more concrete, and easier to work with in comparison to the ambiguous notion of “respecting the manipulee’s autonomy.” Manipulations typically threaten, risk threatening, or convey disrespect for the manipulee’s deliberative project (as argued in section 3.2). In light of
,
53
this fact, there are contexts in which a general policy of treating every manipulation with suspicion would be justified, despite the fact that some manipulations are innocuous (pace Noggle). Although we have been focusing on interpersonal manipulation, we believe that our proposed precisification can also help in evaluating paternalistic nudges in the public domain. One of the strongest considerations against such nudges is that they disrespect people’s autonomy. This may amount to the claim that they disrespect deliberative projects, in light of our argument (notice that Manuel’s manipulation has all the features of a nudge). But in the public domain, it’s much harder for the nudger to know the particulars of the deliberative projects of the people she is influencing. The legitimacy of the nudge would turn on how such uncertainty should be handled, which is a question we haven’t tried to answer. Plausibly, the nudger should consider the probabilities that she is interfering or contributing to deliberative projects, as well as the magnitude of the interference or contribution. This opens the door to the view that the moral status of the same nudge can change from one group or society to another.²⁴
References Aboodi, R. (2021). What’s Wrong with Manipulation in Education? Philosophy of Education, 77(2), 66–80. Baron, M. (2014). The Mens Rea and Moral Status of Manipulation. In Coons, C., & Weber, M. (Eds.) Manipulation: Theory and Practice. Oxford University Press, 98–121. Blumenthal-Barby, J. (2014). A Framework for Assessing the Moral Status of “Manipulation.” In Coons, C., & Weber, M. (Eds.) Manipulation: Theory and Practice. Oxford University Press, 121–35. Bratman, M. E. (2007). Structures of Agency. Oxford University Press. Buss, S. (2005). Valuing Autonomy and Respecting Persons: Manipulation, Seduction, and the Basis of Moral Constraints. Ethics, 115(2), 195–235. ²⁴ This research was supported by the Israel Science Foundation (grant No. 1077/18). Previous versions of this paper were presented at the 2020 Israeli Philosophy Association conference, the 2020 Rocky Mountain Ethics Congress, the 2021 Junior Scholar Workshop at the Hebrew University’s Faculty of Law, and the 2022 Arizona Workshop in Normative Ethics. We thank participants in all these sessions for their feedback. We are also grateful to Hanoch Dagan, David Enoch, Zeev Goldschmidt, Alon Harel, Thomas Hurka, Elijah Millgram and two anonymous referees for this journal who provided valuable comments on earlier drafts, and to Avigayil Bar-Asher for proof-reading.
54
Cave, E. (2007). What’s Wrong with Motive Manipulation? Ethical Theory and Moral Practice, 10(2), 129–44. Cave, E. (2014). Manipulation and Unsavory Seduction. In Coons, C., & Weber, M. (Eds.) Manipulation: Theory and Practice. Oxford University Press, 176–201. Christman, J. (2009). The Politics of Persons: Individual Autonomy and SocioHistorical Selves. Cambridge University Press. Dean, R. (2021). The Peculiar Idea of Respect for a Capacity. In Dean, R., & Sensen, O. (Eds.) Respect: Philosophical Essays. Oxford University Press, 140–56. Dworkin, G. (1982). Is More Choice Better Than Less? Midwest Studies in Philosophy, 7(1), 47–61. Enoch, D. (2011). Taking Morality Seriously: A Defense of Robust Realism. Oxford University Press on Demand. Enoch, D. (2022). Autonomy as Non-Alienation, Autonomy as Sovereignty, and Politics. Journal of Political Philosophy, 30(2), 143–65. Enoch, D. (Unpublished manuscript). How Nudging Upsets Autonomy. Frankfurt, H. G. (1987). The Importance of What We Care About: Philosophical Essays. Cambridge University Press, 159–76. Gorin, M. (2018). Paternalistic Manipulation. In Grill, K., & Hanna, J. (Eds.), The Routledge Handbook of the Philosophy of Paternalism. Routledge, 236–47. Hanna, J. (2015). Libertarian Paternalism, Manipulation, and the Shaping of Preferences. Social Theory and Practice, 41(4), 618–43. Noggle, R. (1996). Manipulative Actions: a Conceptual and Moral Analysis. American Philosophical Quarterly, 33(1), 43–55. Noggle, R. (2018). Manipulation, Salience, and Nudges. Bioethics, 32(3), 164–70. Noggle, R. (2020). Pressure, Trickery, and a Unified Account of Manipulation. American Philosophical Quarterly, 57(3), 241–52. Oshana, M. (1998). Personal Autonomy and Society. Journal of Social Philosophy, 29(1), 81–102. Rawls, J. (1971). A Theory of Justice. Belknap Press of Harvard University Press. Sher, G. (1997). Beyond Neutrality: Perfectionism and Politics. Cambridge University Press. Thaler, R. H., & Sunstein, C. R. (2009). Nudge: Improving Decisions about Health, Wealth, and Happiness. Penguin. Watson, G. (1975). Free agency. The Journal of Philosophy, 72(8), 205–20.
3 Moral Indifference Monika Betzler and Jonas Vandieken
Introduction We all stand to be wronged as individual persons by the actions and attitudes of others. We are, as Michael Thompson puts it, “wrongable” beings.¹ In recent years, philosophers have tried to individuate various kinds of relational wrongs. These wrongs differ from violations of norms and principles in that they violate the valid claims of particular individuals, and thus constitute wrongings.² Yet one particular and more fundamental relational wrong has been unduly neglected, a relational wrong we call “moral indifference,” which morally disables another person. We believe that moral indifference is a pervasive phenomenon, albeit one that is difficult to grasp since it is not always easily gleaned from a person’s behavior. More specifically, moral indifference manifests itself in various modes and forms. Whereas other relational wrongs are often things we do to others, the morally indifferent person wrongs others by what she doesn’t do. The aim of this paper is to elucidate the nature of moral indifference and its disabling mechanism. We offer an analysis of moral indifference as a failure to accord others the ability to make a moral difference, which in turn risks undermining their sense of agency. In section 1 we introduce and locate moral indifference in its various forms. In section 2 we identify the target of moral indifference, before demonstrating in section 3 how moral indifference can wrong others in more fundamental ways than other relational wrongs. We then explain three implications of our account for ongoing debates about important ¹ Thompson 2004, 368. For other important work on the notion of wronging (of particular individuals) and its place in moral theory, see e.g. Darwall 2006, Cornell 2015, Wallace 2019, Zylberman 2021. ² Examples include various forms of bodily injury, disrespect, humiliation, doxastic wronging, gaslighting, and stereotyping. Monika Betzler and Jonas Vandieken, Moral Indifference In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Monika Betzler and Jonas Vandieken 2023. DOI: 10.1093/oso/9780198895909.003.0004
56
concepts in moral philosophy. We argue in section 4 that the concept of recognition respect needs to be disambiguated to capture the fundamental recognitional failure involved in moral indifference. In section 5 we show how Strawson’s distinction between the participant stance and the objective stance needs to be complemented by the indifferent stance. Finally, in section 6, we highlight a further, perhaps surprising implication: that in contexts where a particular moral claim is unjustified, moral indifference can be a legitimate way of protecting oneself and others from entering into a dynamic of further relational wrongs. Before proceeding, the following clarification should be noted as the term “moral indifference” is not widely used in philosophical discourse. Our interest lies in explicating the concept of a distinct kind of wronging and offering a unified account of what it is that the morally indifferent do not do to others. Our usage differs from a merely impersonal understanding of moral indifference. A person may well be considered morally indifferent if she acknowledges, but is not motivated by, the demands of impersonal morality; for example, if she judges that she is morally obligated to help yet is not motivated to.³ But we conceive of moral indifference as a distinctly relational phenomenon. It is about being indifferent to the valid claims of those with whom we have a moral relation. Though we do not endorse a particular relational or, broadly speaking, second-personal account of morality, we believe there are solid arguments for such an account, and indeed think that the interpretation of moral indifference we offer is evidence in favor of a relational account. According to relational accounts, “morality involves a set of requirements on action that are always already constitutively connected to claims that others have against us, just insofar as they are persons.”⁴ What settles the fact of obligation in a certain situation—what determines that I owe it to you, for example, not to step on your foot—is the fact that a particular individual holds a claim on another, for example not to have their foot stepped upon unbidden. On the resulting view, the entire set of moral obligations that we have to each other is best understood as a set of directed obligations, that is, as obligations that are always owed to particular individuals who stand to be wronged by one’s unexcused failure to comply with the relevant obligation.
³ See Milo 1981, 373–93. See also Arpaly 2011, 64f. ⁴ Wallace 2019, 1. Wallace’s work presents the most developed relational account of morality. See also Darwall 2006, and Zylberman 2018 and 2021.
57
Once we conceive of moral indifference as a distinct relational phenomenon, we can also explain why such indifference is not the same as the everyday indifference of not caring.⁵ You might simply be neutral about—not care— whether you have pasta or fish for dinner. But when persons and their claims are at stake, such neutrality is not only not appropriate, but simply not possible. Given that you occupy a distinctly relational standpoint that always already binds you to another, you wrong others if you do not respond at all to their (in principle) valid claims.
1. Making Space for Moral Indifference Moral indifference can manifest itself in a range of attitudes characterized by their blocking effect on the morally relevant agential capacities of those subjected to it. It can be occurrent, dispositional, and structural. Consider the following scenario: Apology Ignored. Mary, walking swiftly with her bike, inadvertently bumps into John. After realizing that she has bumped into John, Mary apologizes to him for doing so. John, however, does not respond to the apology but simply walks away. He does so despite being fully aware of Mary’s attempt to apologize and being capable of responding to it. In not receiving any response to her apology, Mary feels momentarily slighted as a fellow moral being who can participate in accountability practices and thus legitimately expects uptake of her valid claims. In Apology Ignored, we have a clear case of moral indifference that becomes manifest in response to a particular claim. For in willfully ignoring Mary and thus denying her uptake of her apology, John conveys to Mary that she is not worthy of his consideration. None of this need be dispositional in order to constitute moral indifference. We can stipulate that John is usually a very considerate person who is just having a bad day. What is important is that Mary’s apology to John and John’s awareness of her attempted apology give rise to a claim on Mary’s part against John that he respond to her apology, a claim that is itself grounded in the more basic claim we all have on each
⁵ See Lillehammer 2017, 17–35, who provides an overview of different usages of the term “indifference,” which mainly involve “not caring.”
58
other to be duly recognized as equally authoritative participants in the moral sphere and its associated practices.⁶ While Apology Ignored illustrates how moral indifference can become manifest in light of a claim that is addressed to another person at a particular point in time, we can now move on to consider cases where moral indifference is dispositional: The Silent Treatment. Amy is fifteen years old. Whenever she does something that does not meet her mother’s approval, her mother stops talking to her for at least a week. Yesterday, Amy forgot to water the flowers—a household chore she has been assigned by her mother. As a result, her mother treats her as if she were not there. She does not look at Amy or address her in any other way. She neither expresses blame nor responds to Amy’s desperate apologies. Amy feels erased and at a loss as to how to reconnect with her mother. Amy’s mistake—albeit a minor one—calls for some response from her mother, either a reproach or a display of understanding as to why Amy has not fulfilled the task assigned to her. The lack of any directed reaction from Amy’s mother, however, stems not from a belief that Amy’s faulty behavior is so minor as not to warrant mention. Nor does it stem from her belief that subjecting Amy to “the silent treatment” serves an educational purpose and thus, ultimately, from care for her daughter. Rather, her lack of response signals that she finds Amy’s forgetfulness so bad that she deems Amy unworthy of any further interaction. Importantly, moral indifference is here not simply manifest in occurrent behavior against the backdrop of a particular claim. Amy’s mother is disposed to not showing any reaction across various possible worlds, such as a world in which Amy apologizes, or one in which Amy is desperately awaiting a response, even a negative one. One might think that the mother is not actually indifferent to Amy, but is in fact blaming her. We believe that she does not, for what Amy’s mother is doing is refraining from engaging with Amy altogether. In blaming another—even in atypical ways that do not involve direct or expressed reproach—one engages with the other by asking them for uptake of and
⁶ To be clear, Mary does not have a claim on John to accept her apology, only on his not ignoring it. Admittedly, whether this is really a case of moral indifference is subject to interpretation, and Mary could be wrong in how she assesses John’s (willful or negligent) ignorance.
59
response to the negative reactive attitude of blame by accepting it and feeling guilt or remorse. Blame is an engaging attitude, one that calls on the blamed person to respond, and—even more importantly—acknowledges the other person’s standing and ability as an equally authoritative participant in moral discourse.⁷ But it is precisely this engagement and acknowledgment that Amy is denied by her mother. We therefore contend that Amy’s mother is indeed subjecting her to moral indifference and not to blame. To further buttress the intuition that moral indifference often manifests itself dispositionally, consider next the following scenario: Blanking the Homeless. Gregory P. Smith, who has been homeless for 25 years, states: every homeless person has one thing in common: they know how it feels to be an outcast [ . . . ]. I felt so invisible that sometimes being bashed was a relief: at least the thugs were engaging with me. [ . . . ] In a twisted way, I sometimes felt a bashing was preferable to being ‘blanked’ by the general public.⁸ Smith’s report describes the wrong of “blanking” someone, as he terms it. Not being addressed—that is, not being looked at, listened to, or recognized in some other way—is sometimes even worse, he says, than being addressed negatively, even wrongingly (in Smith’s example, being bashed). There are several ways in which moral indifference can manifest itself in this case. Some people may simply fail to notice Smith even though they should notice him. Others may notice him but decide not to engage. Regardless of the specifics, Smith feels “blanked” by his fellow moral beings, who could instead show compassion—or bash him for lingering in the streets, unable to make a living. To be sure, the people in question do not have a perfect duty to respond. Not everybody has a duty to show compassion with every homeless person on every possible occasion. Analogously, Amy’s mother does not have to show (dis)approval with regard to her daughter’s behavior all the time. What makes these scenarios cases of moral indifference is that Amy is confronted with her mother’s disregard for a considerable amount of time, and that Smith seems to be “blanked” all the time by the vast majority of people passing by no matter how miserable he looks. It is the continual indifference in both cases that wrongs Amy and Smith. This of course does not imply that either party would prefer being bashed or ⁷ See e.g. Darwall 2006 and 2013.
⁸ This example is taken from Brownlee 2020, 111.
60
otherwise negatively engaged with every time they are seen. But the fact that Smith regards “sometimes being bashed” as preferable highlights the distinct wronging of moral indifference. When moral indifference is shared by a significant portion of society and becomes a social practice, we are dealing with structural moral indifference. It is then the general public that is disposed to treat Smith with moral indifference. It is no coincidence that moral indifference is generated by and exacerbates asymmetric and hierarchical relations of power. Such relations are often conducive to domination and the illegitimate deprivation of moral authority. But even though systematically unequal relations of power facilitate and exacerbate moral indifference by entrenching structural moral indifference, these are not strictly necessary for moral indifference to occur.
2. The Target of Moral Indifference: The Ability to Make a Moral Difference We have specified the phenomenon of moral indifference in its three guises, and showed that it comprises a variety of more passive and more active attitudes. It ranges from negligence to willful ignorance, and encompasses not noticing when one should notice, and noticing but not engaging, both of which disable or deny the other’s morally relevant agential capacities. To uncover the wrong-making feature of moral indifference, we must now identify its target. This will also help us identify what unifies these multifarious attitudes and what marks them as distinct modes of the wronging that is moral indifference. What exactly is moral indifference directed at, then, such that the other feels disregarded, blanked, invisible, or closed off? And what exactly is it that is occurrently, dispositionally, or structurally disabled in others? Our answer turns on the observation that in each of the scenarios mentioned, it is conveyed to the recipients of moral indifference that they do not have the de facto authority to make a moral difference.⁹ Generally, making a moral
⁹ By de facto authority we mean the practical authority that one has as a matter of fact, legitimately or not. This is to be contrasted with de jure authority, by which we mean the legitimate authority one enjoys in principle, e.g. in virtue of being a member of the moral community. The latter is comparable to the legitimate authority one possesses simply in virtue of being a citizen of a given nation state and the rights one has by law, whether one can in fact exercise these rights or not, e.g. the right to vote. If a citizen successfully votes in an election, she has exercised both de facto and de jure authority. See e.g. Darwall 2006 and 2010.
61
difference consists in reciprocally participating in the exchange of (in principle) valid claims.¹⁰ In being able to make a difference, one is ultimately able to address, receive uptake of, and respond to, claims.¹¹ Importantly, this does not mean that one’s own claims must always be fulfilled, or that one must always fulfill the claims of others. One does not even need to implicitly or explicitly address claims to another in order to be subject to moral indifference. To illustrate this point, consider the example of someone blanking their partner for no real reason and without any attempt by the partner to make a moral difference. This example lets us make an important clarification regarding the de facto authority to make a moral difference at issue here. The latter need not necessarily be understood as an authority to implicitly or explicitly address a claim on another as an instance of some kind of communicative action, but rather as a kind of standing to (in principle) address claims and receive uptake. In turn, this goes to show that one can acknowledge or fail to acknowledge the standing of the other person to make a moral difference, irrespective of the other’s explicitly addressing any claims in the first place. What the person blanking her partner is doing is denying the partner the standing to make a moral difference. Ultimately, then, being able to make a moral difference means having, in principle, the standing to address and receive uptake of claims, a standing that registers as normatively significant to others as in principle meriting uptake and a response. In addition, it means having, at least in principle, the standing and the capacity to respond to the claims of others. Consequently, making a moral difference has various dimensions which include, but are not limited to, ex ante address of claims to uptake and ex post response to the claims of others. Going back to our examples, Mary, Amy, and Smith are all “sources of valid claims”¹² who are able to offer reasons for their claims, appreciate uptake, respond to the claims of others, stand up for themselves, and accept or reject criticism. In being subjected to moral indifference, their ability to make a difference is (willfully or negligently) ignored and they are thereby denied the possibility to use it. Moral indifference is thus a distinct kind of ¹⁰ There are certain characteristic features of claims that we are committed to. These include, but are not limited to, claims that are inherently directed or relational. Claims differ from (claim-) rights, which constitute a subcategory of claims, in that claims are less robust (e.g. we have no necessary discretion or power over them) and not necessarily unconditional, i.e. claims are defeasible. See e.g. Thomson 1990, and Wallace 2019, 6–9. ¹¹ See Darwall 2004, 46: Moral reasons get their force from a kind of authority that we have “to make claims and demands of one another as free, rational, and equal members of the moral community.” ¹² We borrow this term from John Rawls 1985, 242.
62
“claimant injustice.”¹³ Mary is denied uptake of her valid claim to receive a response to her apology. Amy is prevented from arguing with her mother, holding her mother accountable, or pointing to extenuating circumstances that might justify her forgetfulness. Smith is ignored and thereby disabled from participating as an equally competent and authoritative source of (in principle) valid claims to point to his difficult situation, ask for help, or simply make himself seen by others. Note that even if the other’s claims are rejected, there is still uptake. John could have angrily refused to accept Mary’s apology. Amy’s mother could have scolded her for her forgetfulness. Smith could have been blamed for lingering in the streets. These negative reactive attitudes would openly communicate to Mary, Amy, or Smith that they are recognized as having the ability to make a moral difference, even if the difference they are making is criticized. Whether or not this criticism is justified, it at least acknowledges the other as a source of (in principle) valid claims. Like these negative reactive attitudes, moral indifference serves a communicative function, even if it is largely unexpressed. Similarly to the explicit communication of negative reactive attitudes, moral indifference sends a message, and that implies a recipient. The success of moral indifference is thus predicated on uptake.¹⁴ But in contrast to these other negative reactive attitudes, moral indifference communicates implicitly and thus sends the message that the other’s de facto authority does not have to be taken into account, but is to be regarded as irrelevant.¹⁵ This message is meant to be communicative either occurrently, dispositionally, or structurally. The morally indifferent person thereby refrains from assigning de jure authority to the other as a source of (in principle) valid claims. Moral indifference thus targets the other’s ability to make a moral difference by (willfully or negligently) ignoring or denying that the other is a
¹³ See Carbonell 2019, 182. While Carbonell focuses primarily on social conditions, our focus is both broader (in that we attend to occurrent, dispositional, and structural ways of being morally indifferent) and narrower (in that we specify a distinct mode of denying that capacity to make a moral difference). ¹⁴ See Macnamara 2015, 250–5. Her work helps elucidate how moral indifference shares with reactive attitudes the function of evoking uptake, yet differs in how that function is fulfilled. ¹⁵ That moral indifference is indexed on the elicitation of uptake limits the cases in which we can plausibly be said to be speaking of moral indifference. A great many people have valid claims that we cannot all take into consideration. Take someone in Flint, Michigan, who has a valid claim to clean water. Are we, who are living in faraway Germany, indifferent to his claim because we are not doing anything about it? On our understanding of moral indifference we are not, for we do not communicate anything to him, thereby not intending to elicit uptake in the other.
63
source of (in principle) valid claims. But the way in which it wrongs the other can differ in scope. Ignoring or denying the other’s ability to make a moral difference does not necessarily amount to a dismissal of the other’s moral standing altogether. For example, it may well be that Amy’s mother would come and help her daughter if she fell into severe depression or had a terrible accident. Moral indifference can thus target the other’s ability to make a moral difference in context-specific ways. Its success conditions depend, at least in part, on how far-ranging the denied claims are. While Amy’s mother might deny Amy a certain standing within their motherdaughter relationship and with regard to certain behaviors, at least for some time, the general public might deny Smith his standing as a moral agent in a very broad range of cases. Furthermore, the way in which moral indifference constitutes a wronging may differ in seriousness and depth. How serious it is depends on the moral importance of a particular claim. In Apology Ignored, Mary might regard her claim to uptake of her apology as legitimate, yet not particularly important. This does not undo John’s wrong, but it highlights that less serious cases of wronging allow us more room to simply dismiss them, if only for prudential reasons. For example, it can be psychologically costly to worry about whether others meet one’s claims, and therefore prudent to disregard breaches of the less important ones. In serious cases, however, we would think that an agent lacked self-respect if she did not care about the violation of legitimate and important claims. These variations and complexities concerning the scope and depth of the wrong of moral indifference notwithstanding, moral indifference constitutes a distinct kind of wronging, whose wrong-making features and various modes we now want to explain more thoroughly.
3. The Distinct Wrong of Moral Indifference Moral indifference, in targeting the other’s ability to make a moral difference, is a “globalist response”¹⁶ in that it is directed at the other’s morally relevant agential abilities. Moral indifference disables the other person’s agential standing as a source of (in principle) valid claims, which makes moral indifference a grave and distinct kind of wronging. Consider Julia
¹⁶ We borrow this term from Bell 2013, 40.
64
Markovits’ observation that “Someone who kills out of ill will, say, must at least recognize his victim to be a center of subjectivity – a source of aims and desires of his own. Not a saving grace. But its absence leaves a distinctive chill of its own.”¹⁷ To bring out more clearly the globalist character of moral indifference, consider a “localist response” that is directed at a particular action:¹⁸ Amy’s mother could be resentful that Amy forgot to water the plants. In contrast to such localist responses, moral indifference is directed globally at the agent herself. To be clear, that does not mean that the other’s agency is always entirely disabled. Rather, and as we have highlighted, this disablement can be limited in scope. Amy’s mother, for example, is being morally indifferent to her daughter only when she does not fulfill a task assigned to her. While this illustrates that the scope of moral indifference can be limited, it is nevertheless a globalist response insofar as Amy’s mother responds to her daughter’s standing as an agent with the ability to make a moral difference.¹⁹ Ultimately, it is always the standing of the other as agent rather than just a specific action that the morally indifferent person responds to. To further explain how moral indifference differs in kind from other relational wrongings, it is helpful to compare it with other globalist responses that attack or undermine the agential standing of another as a source of (in principle) valid claims. This, in turn, will help further illustrate the different modes of wronging that are entailed by moral indifference. We focus on three such responses that share important qualities with moral indifference but do not possess its disabling character: using someone merely as a means, having contempt for someone, and being a moral “asshole.”²⁰ This helps us identify three characteristic wrong-making features of moral indifference: instrumentalizing objectification, withdrawal, and the vice of superiority. Why, asks Markovits, does there seem to be “something particularly abhorrent about such terribly wrong actions performed out of such total indifference – more abhorrent, in some ways, than the same actions performed out of ill will?”²¹ It may be, in part, because when someone is morally
¹⁷ Markovits 2012, 309. ¹⁸ See also Bell 2011, 452. ¹⁹ This highlights an important distinction between localist and globalist responses. Whereas the latter can be limited in scope, the former cannot. For whenever one responds to the specific action of another, the question of scope does not arise. There is one action to respond to, period. Not so with a globalist response, insofar as there are possibly several aspects of another’s agency and standing as someone to be engaged with that one can respond or fail to respond to. ²⁰ See James 2012. ²¹ Markovits 2012, 309.
65
indifferent she is more likely also to merely use others, objectify them, treat them as instruments rather than as persons. Moral indifference entails an inherently objectifying attitude that prevents the other from making a moral difference. Yet it is important to note that one need not be indifferent to another in order to objectify her. There is at least one sense of objectification where one must recognize the other as an agent and a source of (in principle) valid claims. Take the example of slavery: to the master, the slave is an instrument, for example for harvesting cotton. But crucially, the master cannot deny his slave the standing of an agent and, in turn, the standing of a source of (in principle) valid claims. As Corinna Mieth and Jacob Rosenthal state, “Cynical as it may sound, especially in the case of slavery: the mistreated person has to cooperate in a certain way, or otherwise the instrumentalization could not happen. [ . . . ] Instrumentalizing an agent as an agent [ . . . ] means that the agent has to do something specific, and [ . . . ] as long as there is agency proper the person can in principle refrain from doing this.”²² At least in one sense of objectification, the objectifier needs to regard the other as a source of (in principle) valid claims to which she must be sensitive in order to get the other to do as she wishes.²³ While this might not be possible without suffering grave consequences, one can in principle respond to these instances of objectification by refusing to do what the other demands. That is, one can in principle make a moral difference. The person subjected to moral indifference, however, finds herself in a normative void, unable to object or cooperate, for she is not deemed worthy of any kind of response, even a negative one. Another globalist response to a person’s agential standing as a source of claims is contempt. While contempt may lead to moral indifference, the two differ significantly. Both contempt and moral indifference are forms of withdrawal. According to Macalester Bell, to have contempt “is to negatively and comparatively regard or attend to someone who has fallen below some standard that is part of the subject’s personal baseline; this form of regard constitutes a withdrawal from the target of contempt.”²⁴ But contempt differs from moral indifference in that it communicates to the other person that she can be rightfully looked down upon. Even in withdrawing, ²² Mieth and Rosenthal 2022, 91–92. ²³ See Nussbaum 1995, 257, for seven different ways in which one can objectify another. As the very first notion of objectification, she mentions instrumentalization, which is compatible with having respect for the thing objectified and need not entail “lack of consideration for feelings and subjectivity” (ibid., 260). ²⁴ Bell 2013, 65.
66
contempt involves making a claim, and “this claim may be addressed through contempt’s characteristic withdrawal.”²⁵ The other is still regarded as someone who can, and in fact should, give uptake and a response to the claim implicit in contempt.²⁶ Contempt impairs the relationship with the other by downgrading her, but it involves the other as someone deserving of a “form of regard,”²⁷ even if it be ill will. Moral indifference by contrast amounts to withdrawal, period. The morally indifferent person still addresses the other, but she does so without making any claim. Like disgust, moral indifference simply aims to cut off reciprocal communication with the other.²⁸ As a form of non-reciprocal address, moral indifference thus precludes the possibility of a coherent response. And it is precisely this implicit non-reciprocal quality that makes moral indifference a distinct injury. Similarly, moral indifference shares certain characteristics of the so-called moral “asshole”:²⁹ assumed entitlement to special advantages and the vice of superiority. Both moral indifference and the attitudes exhibited by a moral asshole are globalist responses to another person, yet they exhibit different kinds of relational wrongs. The moral asshole sees himself as entitled to special advantages—say, queue-jumping—and as such regards himself as superior to others, thereby calling into question the equal moral standing of others as sources of (in principle) valid claims. But he still has to regard the other as a fellow participant in the moral sphere, for he is in principle willing to offer a justification for his conduct, namely that he is special and as such deserves the special advantages. That makes room for the other, at least in principle, to make a moral difference. Although the moral asshole shows ill will, he still reciprocally engages with the other, even if he sees the other as inferior. By contrast, the morally indifferent person does not engage in reciprocal exchange and sees no need to justify her actions or otherwise react to the other. Quite the contrary: the wrong of moral indifference essentially manifests itself in an attitude of exclusion and denial of the other’s standing as a source of (in principle) valid claims. In being subject to moral indifference, one is essentially walled off, suspended in the moral universe without any normative say. What we are arguing, then, is that the morally indifferent person denies the other her ability to make a moral difference. And it is this denial that distinguishes moral indifference from the other globalist ²⁵ Ibid., 187. ²⁶ See ibid., 187. ²⁷ Ibid., 56. ²⁸ See the brief discussion of disgust in ibid., 186. ²⁹ See James 2012.
67
responses discussed. Whereas the latter curtail the abilities of others to make a moral difference, the morally indifferent person denies these abilities altogether. What is important about using others as mere means, about contempt, and about the attitudes exhibited by the moral asshole, is that none of them violate the basic commitment of morality to stay in reciprocal communication with others. By contrast, the morally indifferent person does not engage in reciprocal address, uptake, and response of the relevant kind. Rather, she blocks the necessary exchange of (in principle) valid claims that is necessary for justifying her action to someone else or to react to the other person.³⁰ The grave injury resulting from moral indifference is a specific agential uncertainty. By not receiving uptake of and response to one’s own claims, one can become uncertain as to the status and validity of those claims, which can cause one to question one’s own worth and standing as an agent. As a result, one can, even if only momentarily and to a greater or lesser degree, become uncertain about one’s standing to press claims, make demands, or legitimately hold others accountable by exhibiting reactive attitudes.³¹ In contrast to other relational wrongs, moral indifference thus fundamentally undermines our ability to make a moral difference; that is, our ability to participate as agents and sources of (in principle) valid claims. It is thus a globalist response of a particular kind. Admittedly, that is not the only way one might become uncertain about one’s agential standing. For instance, victims of testimonial injustice, whose claims are heard but not taken seriously, can become uncertain of their standing as members of the moral community and come to question their sense of agency. The injury inflicted by moral indifference, however, leads to a distinct, because more fundamental, agential uncertainty. It is more fundamental because the other is denied the ability to, even in principle, make a moral difference. This goes back to the difference between someone curtailing one’s ability to make a normative difference and their denying that one has the ability to make a normative difference altogether. While testimonial injustice can certainly lead to agential uncertainty, it still leaves open the possibility of making a moral difference by, for example, objecting to the injustice suffered. What we are arguing, then, is that moral indifference distinguishes itself from
³⁰ See also Bell 2013, 184. ³¹ Mary’s situation in Apology Ignored is a case in point. Even she may experience the kind of agential uncertainty we identify here. But she may well recover quickly and go on with her day without having been unsettled in a deep and lasting way.
68
other relational wrongs in virtue of the fundamental kind of agential uncertainty it gives rise to.
4. Basic Second-Personal Regard Having identified the target and distinct nature of the wronging of moral indifference, we will now consider some possible implications of our analysis. The first pertains to our understanding of “recognition respect.”³² As sources of (in principle) valid claims and as equal members of the moral community, we demand recognition respect from each other: “interpersonal recognition is precisely what we are demanding of others when we assert our claims against them.”³³ We thereby regard ourselves as “entitled to have other persons take seriously and weigh appropriately the fact that [we] are persons in deliberating about what to do.”³⁴ There are various ways to fail to provide the relevant recognition. Moral indifference is distinct in that it withholds recognition by denying what we call basic second-personal regard. To see this, we need to distinguish between two kinds of recognitional failures entailed by the concept of recognition respect. One is a failure to recognize the (in principle) valid claims of others as giving rise to moral obligations and thus as constraints on one’s agency. The other is a failure to recognize others as sources of (in principle) valid claims in the first place.³⁵ While moral indifference clearly constitutes a failure to respect and recognize another, so too do other globalist responses. Consider again the moral asshole. In claiming special entitlements, he fails to accord others recognition respect. But the failure of the moral asshole does not prevent the other from making a moral difference. After all, one can object to or call out the moral asshole. He might react by pointing out his special entitlements. And this is important. For while the moral asshole fails to accord recognition respect to the other, he accords what we call basic second-personal regard to others by (in principle) enabling uptake of their claims and thus conceiving of the other as a partner in the reciprocal exchange of claims, even if those claims are treated unequally.
³² See in particular Darwall 1977 and 2006. ³³ Wallace 2021, 6. ³⁴ Darwall 1977, 38. ³⁵ See also Wallace 2021, 6.
69
Moral indifference, by contrast, is not simply a failure to accord recognition respect in that first sense. More fundamentally, the morally indifferent person fails to accord the other person basic second-personal regard: that is, recognition respect in the second sense. For the morally indifferent person not only sees the other’s claims as not significant enough to constrain her own agency, but moreover does not see the other person as a source of (in principle) valid claims to begin with, thus wronging the other in a distinct manner. Distinguishing between these two different failures to accord recognition respect is important, not least for adequately capturing the recognitional failures and subtleties involved in moral indifference. In fact, secondpersonal recognition respect—as the appropriate attitude to adopt when relating to others and deliberating about what we owe them in light of their particular claims—presupposes basic second-personal regard. We can only deliberate about which of others’ valid claims function as constraints on our agency if we regard them as sources of (in principle) valid claims.
5. The Indifferent Stance Our conception of moral indifference as a distinct kind of wronging also has important implications for an appropriate understanding of what it means to be regarded as an equal member of the moral community and the accountability practices tied to that. Following Strawson’s influential view, holding someone morally accountable implies acknowledging that she is an equal member of the moral community. She is thus eligible for reactive attitudes that express such acknowledgment.³⁶ When we see others as equal members of the moral community, we adopt what Strawson calls the participant stance: we see them as partners in moral reasoning, as addressees of claims via our reactive attitudes. These include not only blame, indignation, and resentment, but also praise, gratitude, and forgiveness. We thereby acknowledge the other’s ability to understand, accept, or reject such claims, and thus see ourselves and the other in a relationship of reciprocal claim-based address and exchange.³⁷ Holding others morally accountable for the way in which they (fail to) acknowledge these claims
³⁶ See Strawson 1962, 1–28. For other views that tie morality to accountability and its associated practices, see Darwall 2006, esp. Ch. 4, and Wallace 2019, Ch. 3. ³⁷ See Shoemaker 2007, 71.
70
amounts to responding to their underlying good or ill will via the corresponding reactive attitudes. The objective stance, in contrast, is one we adopt when we (momentarily or generally) do not regard others as equal members of the moral community by no longer holding them morally accountable for their actions as manifestations of their will. According to Strawson, there are two reasons for adopting the objective stance. One arises when the other person is momentarily excused from accountability while generally still retaining the capacity to engage in claim-based exchange. A person might have been in shock when acting (for example as a result of having witnessed an accident), and is therefore not held accountable for that particular action. Another reason for adopting the objective stance is that the other lacks the capacities necessary for moral address (for example, she may be comatose, incapacitated due to mental ill-health, a young child, or severely cognitively disabled). This exempts her from moral accountability altogether. In adopting the objective stance toward others, we relate to them (momentarily or generally) with detached attitudes and measures that treat their actions as caused events, not manifestations of their will. We see the other “as an object of social policy; as a subject for what, in a wide range of sense, might be called treatment; as something [ . . . ] to be managed or handled or cured or trained; perhaps simply to be avoided.”³⁸ Any communication that we engage in with others from the objective stance is importantly non-reciprocal, in that it does not regard them (momentarily or generally) as morally accountable agents and thus as equal members of the moral community. We will not resent someone toward whom we have adopted the objective stance, because the objective stance characteristically inhibits “ordinary interpersonal attitudes in general, and the kind of demand and expectation which those attitudes involve.”³⁹ It is important to note that we can still care for and love others to whom we have taken up the objective stance.⁴⁰ That is, we can still regard them as sources of (in principle) valid claims.⁴¹ ³⁸ Strawson 1962, 9. ³⁹ Strawson 1962, 13. ⁴⁰ See Bagnoli 2021, 650. ⁴¹ As a result, we can regard the comatose, the incapacitated due to mental ill-health, the very young, and those with cognitive or learning disabilities as sources of claims who cannot be reciprocally engaged with and who cannot in turn be held morally accountable. At the same time, they can still be viewed as members of the moral community, albeit unequal ones. If we were to address them with moral indifference, we would send the message that they are to be disregarded as sources of such claims altogether and wrong them subsequently, in at least two ways: first, by failing to acknowledge and sustain the capacities they may still have; and second, by entrenching a social practice of ignoring them as sources of (in principle) valid claims, thus contributing to their marginalization. The indifferent stance can thus be more injurious in these cases than the objective stance.
71
Sometimes, we can take up the objective stance as a “moral strategy” that deliberately excludes others from the moral community, which in turn deprives them of the chance to reciprocally engage with others as participants in a shared moral practice. One might therefore think that moral indifference is akin to the objective stance. But this alignment is merely apparent. We claim that moral indifference is a distinct kind of stance alongside the participant and objective stances: the indifferent stance. It shares with the objective stance the fact that the other is (momentarily or generally) not acknowledged as an equal member of the moral community by not being addressed as an appropriate target of the abovementioned reactive attitudes. This similarity becomes especially poignant when the indifferent stance is compared to the strategic use of the objective stance. But important differences remain. The strategic use of the objective stance is belittling, as it treats the other as if she did not have the requisite agential capacities. It is therefore a wronging kind of engagement, though an engagement nevertheless. For it leaves room for the other to refuse or criticize this belittling treatment. The indifferent stance, by contrast, willfully or negligently ignores or denies the other’s standing as a source of (in principle) valid claims by non-reciprocally communicating that the other should be so regarded. The wrong of the strategic objective stance is to consciously treat the other as something other than what they are, namely as someone with no agential capacities that qualify her as an equal member of the moral community. The wrong of the indifferent stance is to illegitimately communicate that the other’s capacity to engage in an exchange of (in principle) valid claims is to be ignored or denied altogether, even though such an exchange would be possible and is generally called for as excusing or exempting conditions do not apply to the other. She is dealt with neither as a subject to be treated nor as an object of compassion and concern. Instead—and this is what the indifferent stance shares with the participant stance—the other is addressed, yet in the indifferent stance this occurs in a way that blocks any response. Hence its non-reciprocal nature. The indifferent stance thus non-reciprocally communicates by way of occurrent, dispositional, or structural indifference and the associated attitudes that the other is not worthy of even basic second-personal regard. The indifferent stance thereby sends the message that—despite her equal membership in the moral community—the other’s doings or identity are regarded as not even deserving uptake. In contrast to the participant stance, the other is denied the possibility to make a moral difference and hence
72
disqualified as someone whose will counts. All of the above suggests that the indifferent stance is a distinct and peculiar kind of stance that can be conceptually located between the participant and the objective stances. The indifferent stance comprises the range of attitudes we have highlighted, including (willful or negligent) ignorance, disregard, and denial. These attitudes are reactive, yet non-reciprocal in the way they address the other. The indifferent stance is adopted because the other is thought, justifiably or not, to be too morally problematic for engagement, or to belong to a group that is beyond the pale of morality. In taking up the indifferent stance, one does take account of the other, but as someone not worthy of moral consideration. Admittedly, the participant stance, the indifferent stance, and the objective stance can intertwine and overlap. But despite that, moral indifference is a distinct stance in that it amounts to a specific kind of wronging. Individuating the indifferent stance helps us provide a more nuanced view of the way we regard or disregard others as equal members of the moral community. It allows us to capture a distinct nonreciprocal denial of others’ agency.
6. The Value of Moral Indifference A third interesting implication of our conception of moral indifference is that it can be of moral value under non-ideal circumstances. Consider the following case: Obtrusive Neighbor. Ben has been fired and has since faced financial hardship. Learning about his plight, Beth, a friendly and caring neighbor, lends him money a couple of times, but he never pays her back. Instead, Ben starts to approach her for money whenever she leaves the house in the morning, and he sometimes waits for her in the evenings to ask for more. This seems to be a clear case in which a person makes an illegitimate request. Ben’s request grounds his claim to be given money. That Beth has lent him money in the past does not make his claim to receive more money from her in the future valid. But Ben’s request-based claim puts Beth under normative pressure to respond. She is faced with the choice to either give in to his request or enter into a fruitless discussion with him. Suppose Ben is easily offended. If Beth says “no,” he will be unwilling to accept her refusal, call her a “nasty capitalist,” and start shouting at her. How should Beth respond to him?
73
The subtle and non-reciprocal communicative function of moral indifference can be morally valuable in cases where someone like Ben makes illegitimate claims on another. Instead of entering into a spiral of blame or— worse—mutual offenses, moral indifference allows Beth to respond without adding insult to injury. By disengaging from Ben, Beth can non-reciprocally communicate to Ben that his request-based claim to more money and his dismissal of her rejection of his request have convinced her that he is not worth engaging with anymore. Moral indifference is thus self-protective—it protects Beth from complicity in Ben’s illegitimate request and bars a threatening downward spiral of their relational dynamic. If she does not become morally indifferent, Beth cannot but continue to reject Ben’s request, which would lead him to further dismissal and offense. Moral indifference is also other-protective as it protects Ben from being confronted with further negative reactions by Beth. In cases where others make unjustified moral claims, moral indifference can be a less injurious way of keeping others from making a moral difference. This is not to say that moral indifference is intrinsically valuable, and it may continue to be pro tanto wrong.⁴² But under non-ideal conditions, moral indifference is a means to protect oneself as well as the other from the downward spiral of reproach, blame, and rejection, even if this might come at the price of ultimately severing the relationship altogether. Note, however, that the reason to be morally indifferent here is that the claim the other makes is illegitimate. Self- and other-protection is merely a downstream effect. Moral indifference can thus be instrumental in maintaining the value of our moral relations by shielding us from irresolvable interpersonal conflicts, even if this ends a particular relationship, or prevents it being formed in the first place. It might not be the only attitude that is called for in a particular case where illegitimate claims are being addressed, but it can be crucial in navigating a relationship characterized by such claims. While we should not cultivate moral indifference as a dispositional attitude, we can make use of it as a tool in particular instances where the relational wrong that it constitutes is a means to shield us from pointless conflict.
⁴² We leave open the question whether moral indifference is pro tanto wrong. It might be regarded as morally fitting in cases of illegitimate claims.
74
7. Conclusion We have tried to explain the phenomenon of moral indifference and show that it is a distinct kind of wronging, before discussing three important implications. In doing so, we hope to have made progress in advancing an accurate understanding of moral indifference specifically, and a more nuanced understanding of other concepts at the core of relational moral theorizing more generally. Undoubtedly, many questions still need to be answered. Among these are how the wrong of moral indifference compares to other relational wrongs, and which tactics are available for combating moral indifference in everyday life. For now, we simply hope to have put moral indifference on the map as a distinctively relational moral phenomenon. We believe that any adequate moral theory, especially in relational form, needs to account for moral indifference; for without it, a crucial element in our understanding of the moral domain will be missed.⁴³
References Arpaly, Nomy. 2011. Unprincipled Virtue: An Inquiry into Moral Agency. Oxford: Oxford University Press. Bagnoli, Carla. 2021. “Disclaiming Responsibility, Voicing Disagreements, Negotiating Boundaries,” in: David Shoemaker (ed.), Oxford Studies in Agency and Responsibility, Volume 7. Oxford: Oxford University Press, 283–306. ⁴³ We are grateful to Maike Albertzart, Macalester Bell, Jacob Leo Blitz, Sarah Buss, Ulrika Carlsson, Alex Edlich, Micha Gläser, Nora Heinzelmann, Sinja Hofmann, Richard Holton, Benjamin Kiesewetter, Brendan de Kenessey, Andrew Knox, Rae Langton, Andrew Lichter, Jörg Löschke, Guido Löhr, Anne Meylan, Sumeet Patwardhan, Herlinde Pauer-Studer, Julia Peters, Linda Radzik, Geoff Sayre-McCord, Sarah Stroud, Mark Timmons, Charlotte Unruh, and Ariel Zylberman for very stimulating written comments and extensive discussions. This paper was presented at the Arizona Workshop in Normative Ethics, at the Research Colloquium in Practical Philosophy at the University of Mainz, the Departmental Colloquium of the University of Tübingen, the Research Colloquium in Practical Philosophy at LMU Munich, the Conference on “Moral Vulnerability” at the University of Oslo, and at the Thumos Seminar at the University of Geneva. We thank these audiences for very valuable feedback. We are also indebted to the participants of the EthicsLab hosted by LMU Munich’s Center for Advanced Studies in connection with the research group on “Relationships in Transition,” and to two anonymous reviewers for their insightful comments. Research on this paper has been generously funded by the LMU Center for Advanced Studies. We also gratefully acknowledge funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 740922) and the German Research Foundation DFG (grant award No 442297699).
75
Bell, Macalester. 2011. “Globalist Attitudes and the Fittingness Objection,” Philosophical Quarterly 61 (244), 449–72. Bell, Macalester. 2013. Hard Feelings: The Moral Psychology of Contempt. Oxford: Oxford University Press. Brownlee, Kimberley. 2020. Being Sure of Each Other. An Essay on Social Rights and Freedoms. Oxford: Oxford University Press. Carbonell, Vanessa. 2019. “Social Constraints on Moral Address,” Philosophy and Phenomenological Research XCVIII, 167–89. Cornell, Nicolas. 2015. “Wrongs, Rights, and Third Parties,” Philosophy and Public Affairs 43 (2), 110–43. Darwall, Stephen. 1977. “Two Kinds of Respect,” Ethics 88 (1), 36–49. Darwall, Stephen. 2004. “Respect and the Second-Person Standpoint,” Proceedings and Addresses of the American Philosophical Association 78 (2), 43-59. Darwall, Stephen. 2006. The Second-Person Standpoint. Cambridge, MA: Harvard University Press. Darwall, Stephen. 2010. “Authority and Reasons: Exclusionary and SecondPersonal,” Ethics 120 (2), 257–78. Darwall, Stephen. 2013. “Being With,” in: Stephen Darwall (ed.), Honor, History, and Relationship: Essays in Second-Personal Ethics II. Oxford: Oxford University Press, 110–130. James, Aaron. 2012. The Moral Asshole: A Theory. New York: Doubleday. Lillehammer, Hallvard. 2017. “The Nature and Ethics of Indifference,” Journal of Ethics 21, 17–35. Macnamara, Coleen. 2015. “Reactive Attitudes as Communicative Entities,” Philosophy and Phenomenological Research XC, 546–69. Markovits, Julia. 2012. “Saints, Heroes, Sages, and Villains,” Philosophical Studies 158 (2), 289–311. Mieth, Corinna, and Jacob Rosenthal. 2022.“Blind Spots in the Formula of Humanity: What Does It Mean Not to Treat Someone as an End?” in: Christoph Horn and Robinson dos Santos (eds.), Kant and Absolute Value. Berlin: DeGruyter, 89–104. Milo, Ronald D. 1981. “Moral Indifference,” The Monist 64, 373–93. Nussbaum, Martha. 1995. “Objectification,” Philosophy and Public Affairs 24, 249–91. Rawls, John. 1985. Justice as Fairness. Political not Metaphysical. Cambridge, MA: Harvard University Press.
76
Shoemaker, David. 2007. “Moral Address, Moral Responsibility, and the Boundaries of the Moral Community,” Ethics 118 (1), 70–108. Strawson, Peter. 1962. “Freedom and Resentment,” in: Peter Strawson. 1974. Freedom and Resentment and Other Essays. London and New York: Routledge, 1–28. Thompson, Michael. 2004. “What is it to Wrong Someone?” in: R. Jay Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith (eds.), Reason and Value. Oxford: Oxford University Press, 333–84. Thomson, Judith J. 1990. The Realm of Rights. Cambridge, MA: Harvard University Press. Wallace, R. Jay. 2019. The Moral Nexus. Princeton, NJ: Princeton University Press. Wallace, R. Jay. 2021. “Recognition and the Moral Nexus,” European Journal of Philosophy 29 (3), 634–45. Zylberman, Ariel. 2018. “The Relational Structure of Human Dignity,” Australasian Journal of Philosophy 96 (4), 738–52. Zylberman, Ariel. 2021. “Relational Primitivism,” Philosophy and Phenomenological Research 102 (2), 401–22.
4 Buck-Passing and the Value of a Person Kyla Ebels-Duggan
Yesterday, I attended the memorial service for a friend and neighbor. She was someone that I liked a lot, but many of those in attendance knew her better than I did and certainly loved her. She was an English teacher at a local high school—clearly a favorite among students. She was a smart, highenergy person who was constantly arrested by the beauty of small and ordinary things. She loved words and books and poetry; she was passionate about justice; she was fiercely kind and wickedly funny. I knew all of these things about her from personal experience and came to understand them better through the stories recounted at her funeral. At this gathering her friends, her colleagues, her sister, and her teenaged son struggled to say who she had been and what she had meant to them. They did their best to talk about the value of her life and some of them successfully communicated quite a lot about this. This was the third memorial gathering I had attended in a month. (It has been a hard year.) I left each with a new appreciation of the value of the person whom we were grieving and remembering together, the value of those with whom the grief was shared, and even the value of unrelated people that I love and care about. Some philosophers hold that value can be reduced to reasons. On such a view, to say that something is valuable is just to say that there are reasons to respond to it in certain ways. And to value something is just to recognize that one has such reasons. T.M. Scanlon’s buck-passing account of value is a leading example of this sort of view.¹ In the important case of the value of a person, Scanlon suggests that the relevant reasons are moral reasons: when we talk about the value of a person, we are referring to the reasons to regulate ¹ See T.M. Scanlon, What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998). Chapter 2, pp. 78–107. Kyla Ebels-Duggan, Buck-Passing and the Value of a Person In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Kyla Ebels-Duggan 2023. DOI: 10.1093/oso/9780198895909.003.0005
78
-
our interactions with her in accord with moral principles and valuing a person is just a matter of recognizing that one has moral reasons with respect to them. In this paper, I explain why I find this view unsatisfying and lay the groundwork for a better account. In brief, my dissatisfaction with Scanlon’s buck-passing account of the value of a person can be put like this: you can’t take it to a memorial. It is inadequate to what we reflect on and try to communicate in this sort of setting—the value of a person that we appreciate when we love her. Scanlon’s identification of the value of a person with the narrowly circumscribed set of moral reasons is especially unsatisfying, but I argue that the problem is deeper. It persists even if we allow a wider set of reasons for action into the account, and even if we include, also, reasons for attitudes. The problem is that we cannot dispense with the concept of value or do without value talk. We need the concept of value to make sense of what we’re doing when we are valuing something, which, in turn, we need to make sense of our conviction that we have reasons to act in particular ways. We also need it to make sense of the content of certain beliefs, for example the belief that a person is valuable. After laying out Scanlon’s general buck-passing account, I reconstruct his argument that the value of a person is best understood in terms of moral reasons. I then argue that this could not account for the value that we see in those we love. Moreover, it cannot even account for the value to which we respond with moral respect, what we might call the moral value of a person. Finally, I ask whether the idea that value consists not just in reasons for action, but also reasons to have certain attitudes, can rescue the buckpassing view. I argue that it cannot. Value is more fundamental than reasons. We need to appeal to value first.
The Buck-Passing View of Value On Scanlon’s buck-passing view of value for something to be valuable or have value is for everyone to have reasons to act towards it in certain ways or take up certain attitudes towards it.² These reasons, in turn, are provided by properties of the object. Value is “not a property that provides us with reasons. Rather, to call something valuable is to say that it has other properties that provide reasons for behaving in certain ways with
² Scanlon, What We Owe to Each Other. p. 95.
-
79
regard to it.”³ For example, a bicycle might be valuable in virtue of being lightweight and having smoothly functioning gears and durable components. To say that it is valuable is to say that I have reason to act towards it in certain ways—to make use of it for transport, say, and to maintain it in good working order. The bicycle might also be aesthetically valuable in virtue of its retro paint job and sweet lugs. These features of the bike may give me reason to appreciate and admire it and to photograph it in circumstances that present its aesthetic features to best advantage. And it may be of sentimental value, due to the fact that I purchased it with my own funds at the age of 13 and have been riding it continuously ever since. These relational properties give me reason to be especially protective of the bicycle, locking it up safely, and to continue to use it in preference to possible alternatives. We can think of this account of value as passing the buck in two ways: first, it understands value wholly in terms of reasons and, second, it understands reasons as ordinary properties of the valuable thing. But there is an important difference between these two “passes”: the move from values to reasons is meant to be completely reductive. Scanlon holds that saying that something is valuable is just a short-hand way of talking about reasons. Nothing other than linguistic efficiency would be lost, and some clarity might be gained, if we substituted reasons-talk for values-talk across the board. By contrast, the second pass is not reductive: while the ordinary features of or facts about the object provide the reasons, that they are reasons is a further, distinctively normative claim, over and above the ascription of any merely descriptive properties.⁴ Nevertheless, there is no need to posit value as a part or property of things, distinct from ordinary properties, and no need to entangle ourselves in the heavy-duty metaphysics or epistemic puzzles that such a posit might suggest. Instead value is, according to Scanlon, a “formal higher order propert[y] of having some lower-order properties that provide reasons of the relevant kind.”⁵ There is an ambiguity here to which I will return. The buck-passer might claim that to say that something is valuable is to say that it has some
³ Scanlon, What We Owe to Each Other. p. 96. ⁴ Thus Scanlon affirms that Moore’s question is indeed open: for any descriptive property, X, we can meaningfully say: I know that this thing has X, but is it valuable? (Scanlon, What We Owe to Each Other. p. 97). And Cf. T.M. Scanlon, Being Realistic About Reasons (New York, NY: Oxford University Press, 2014) where he interprets a reason as a relation among a fact, an agent, the circumstances, and an action or attitude. p. 34. ⁵ Scanlon, What We Owe to Each Other. p. 97.
80
-
properties or other that provide reasons to act in a particular way. Alternatively, he may claim that to say X is valuable is to claim of some of its particular properties that they provide reasons to act in the specified way. Scanlon’s own account moves between these two approaches. As we will see, in the first instance he interprets the claim that a person is valuable as the claim that we have moral reasons with respect to her, taking no stance on what properties of a person provide them. Nevertheless, I will consider below whether moving to the view that valuing a person is a matter of regarding some specific considerations or properties of the person as moral reasons makes the buck-passing view more compelling. I will argue that it does not. Scanlon offers two arguments in favor of the buck-passing view. First, he reflects on how we talk about value. He suggests that when we call something “good” we are usually commending it in virtue of ordinary properties that it has. So, the discovery is good because it contributes to a cure for cancer; the resort is good because it’s on the beach, service is prompt, and the food is excellent. The food is excellent because of the fresh ingredients and creative combinations of flavors. There seems to be no additional work that either needs to be, or could be, done by adding “value” or “goodness” as an additional reason-giving property over and above these ordinary ones. It would be odd to say that in addition to featuring the beach and the service and the food, the resort is a good one. Instead, this last comment makes sense only as a kind of summary. Second, Scanlon appeals to the fact that many different kinds of things are valuable and that they are valuable in many ways. Elizabeth Anderson makes this case in more detail: a bedroom can be good in virtue of being cozy, a joke by being hilarious, a tool by being useful.⁶ Scanlon and Anderson agree that it’s unhelpful and obfuscating to suggest that all of these things share a single, identical property of goodness or value. The appropriate or warranted responses to something’s value vary as well. If value were a unified property this would be difficult to explain. But if value is just a way of talking about ordinary properties, suggesting that they call for some positive response, there is no mystery in the fact that these responses vary. Though presented as supporting the idea that we can reduce value to reasons, these arguments actually seem to lead to a more modest ⁶ Elizabeth Anderson, Value in Ethics and Economics (Cambridge, MA: Harvard University Press, 1993). pp. 1–4.
-
81
conclusion.⁷ They support the idea that value is not a further property, over and above an object’s more particular properties. They thus suggest that it is helpful to think of value as a higher-order property of some kind. But this falls short of establishing the first pass described above, the claim that this higher-order property can and should be understood just in terms of reasons. This is surprising, since the claim about the reduction of value to reasons is usually thought to be the most distinctive commitment of the buck-passing view. It is this reductive claim that is of most interest to me. A possible explanation for this lacuna in the argument is that, though the buck-passing view officially rejects the need for any concept of value, in fact Scanlon and Anderson have as their target a specific account of the priority of value to reasons. Their claim that valuable things call for a variety of responses is a rejection of the consequentialist position that production is the uniquely appropriate response to value, such that we should always seek to bring about, or maximize the occurrence of, value or of valuable things. Anderson and Scanlon counter that value may, instead, call for protection, respect, appreciation, consideration, awe, or love.⁸ The consequentialist position that what we have reason to do is to produce as much value as possible does entail that value is prior to reasons.⁹ But to conclude from the rejection of this view that we can do without the concept of value altogether would grant too much to the consequentialist. We should not accept even the conditional claim that, if value is prior to reasons, then what we have reason to do is just to produce or maximize value. Not every value-first view is a consequentialist view. There are other ways in which the good may be conceptually prior to or independent of the right or the practically rational.¹⁰ ⁷ Cf. R. Jay Wallace, “Scanlon’s Contractualism,” Ethics 112, no. 3 (2002). pp. 445–9. And Cf. Jussi Suikkanen, “Buck-Passing Accounts of Value,” Philosophy Compass 4, no. 5 (2009). pp. 768–79. ⁸ The buck-passing view itself doesn’t entail the denial that all values are to-be-produced and Scanlon acknowledges that someone could accept it while thinking that all of the relevant reasons are reasons to produce states of affairs. But he argues that this is implausible as a substantive matter. For example, he thinks that valuing friendship involves taking one’s self to have reason to be loyal to one’s friends, and friendship is valuable just in case people generally do have these sorts of reasons. Thanks to Rahul Kumar and Richard Kraut for pushing me on this point. ⁹ Cf. Philip Pettit, “The Consequentialist Perspective,” in Three Methods of Ethics (Malden, MA: Blackwell, 1997). pp. 92–174. Anderson also anticipates this argument. Value in Ethics and Economics. pp. 8–37. ¹⁰ Many of these are compatible with Rawls’ thesis of the priority of the right over the good. Cf. John Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971). p. 31. For other discussions of priority see: Talbot Brewer, The Retrieval of Ethics (New York, NY: Oxford University Press, 2009). pp. 192–235; Miriam Ronzoni, “Teleology, Deontology, and the Priority
82
-
There may be an additional way of motivating the buck-passer’s reduction of value to reasons—what I have called the first pass. We might think, as Scanlon apparently does, that reasons are less mysterious than value. Famously, he opens What We Owe to Each Other by announcing that he will treat reasons as the fundamental normative concept. In the first chapter of that work, he argues that perennial metaphysical and epistemological doubts about reasons are overblown and that we should dismiss the skepticism about practical reasons that sometimes follows in their wake. Reasons, he assures us, are something that we already understand; we know perfectly well what it is for a consideration to count in favor of an action or attitude and are often rightly confident that particular facts stand in this relationship to particular attitudes and actions. It is no more puzzling that some consideration counts in favor of doing something than it is that some consideration counts in favor of believing something. The buck-passing view would extend this anti-alarmism about reasons to dispel longstanding challenges to claims about value. Realism about values also provokes metaphysical and epistemological objections: What, exactly, could value be? How does it fit into a sensible naturalism about the world? Even if we allowed that the world contains this mysterious thing, how could we identify particular instances of it or be justified in concluding that any particular thing is of value? And why should we care about it so much that it makes sense to order our lives and choices around it?¹¹ If saying that something is valuable is nothing more than a way of asserting the existence of certain reasons, and Scanlon has shown that claims about what reasons we have are in good order, then these puzzles about value will likewise be dissolved. With the buck-passing view, then, Scanlon explicates the apparently mysterious concept of value in terms of the purportedly unmysterious concept of reasons. But I will argue below that we cannot make sense of the reasons the buck-passer needs without invoking prior claims about value.
of the Right: On Some Unappreciated Distinctions,” Ethical Theory and Moral Practice 13, no. 4 (2010). pp. 453–72; Jon Garthoff, “The Priority and Posteriority of Right,” Theoria 81, no. 3 (2015). pp. 222–48; Stephen J. White, “Standing Up for Neutrality,” in Responsibility and the Demands of Morality: The Collected Works of Stephen J. White, ed. Kyla Ebels-Duggan and Berislav Marušić (New York NY: Oxford University Press, forthcoming). ¹¹ Cf. Christine M. Korsgaard, “Realism and Constructivism in Twentieth-Century Moral Philosophy,” in The Constitution of Agency (New York, NY: Oxford University Press, 2008).
-
83
The Value of a Person The value of a person is an important test case for any theory of value. We need not think, as Kant may have, that persons are the unique bearers of value to agree that this value is a central part of an attractive worldview or ethical outlook. The consequentialist’s account of the value of a person is patently unattractive. Anderson’s and Scanlon’s insight that production is not always the appropriate response to value is at its strongest here. Though many would grant that persons have value, few think that our primary response to this value should be to produce it, and even fewer would endorse maximizing it. Persons are, as Kant says, existent ends. We respond to their value, not by making more of them, but by properly valuing—respecting or loving—each one.¹² In accordance with his buck-passing view, Scanlon understands the ascription of value to persons as a claim that we have certain reasons with respect to them. He initially proposes that valuing human lives is a matter of “seeing reasons not to destroy them, reasons to protect them, and reasons to want them to go well.”¹³ After some reflections on euthanasia and suicide, he modifies this formulation to privilege each person’s point of view on his own life: “From the point of view of the person whose life it is,” he tells us, “the value of a life may be identified with the reasons for living it.”¹⁴ A person who does not have adequate reason to want to go on living—perhaps because the remainder of his life promises nothing other than pain, severe cognitive decline, and the dissolution of personality—may not have reason to prevent his own death. Rather than preserving someone’s life over his own reasonable objections, we should share his aims.¹⁵ This can require allowing his life to end but may still be a way of valuing that life or, more to the point, valuing that person. Generalizing from this sort of reflection, Scanlon concludes that “recognizing the value of human life is a matter of ¹² Immanuel Kant, Groundwork of the Metaphysics of Morals, trans. Mary J. Gregor, Cambridge Texts in the History of Philosophy (New York, NY: Cambridge University Press, 1998). pp. 44–5 (Ak 4:437–8). ¹³ Scanlon, What We Owe to Each Other. p. 104. ¹⁴ Ibid., p. 105. One might well worry that this seems to support an inference from lesser well-being to lesser value of the person. We might have thought that it is one thing to say: a person born with such and such condition will be less well off and quite another to say a person with this condition is less valuable than one without it, but the identification that Scanlon endorses here makes it harder to see the daylight between the two positions. ¹⁵ See Kyla Ebels-Duggan, “Against Beneficence: A Normative Account of Love,” Ethics 119, no. 1 (October 2008). pp. 142–70, where I argue that even love is properly expressed by deferring to another’s aims in this way.
84
-
respecting each human being as a locus of reasons, that is to say recognizing the force of their reasons for wanting to live and wanting their lives to go better.”¹⁶ On this view the value of a life just is the force of these reasons and when we say that we should value other human beings, we mean nothing other than that we should be properly responsive to these reasons. To recognize the force of a person’s reasons is not merely to note that it makes sense for her to respond to them, but to allow these reasons to have the right sort of influence on one’s own practical thinking and decisions. When I am trying to determine what to do, then, among the considerations that I must take into account are the reasons that each person has to want to live and to want her life to go better. To properly take these reasons into account is, on Scanlon’s view, to value each person. But we cannot explicitly consider, much less positively respond to, all of the reasons that each person has each time we act. There are too many people, and each has too many reasons, for this to be a meaningful directive. So, we face a task of sorting among or composing all of these reasons in a way that allows for a practical response. We need principles to organize our practical thought. Scanlon holds that these would be principles for the general regulation of behavior that no one has sufficient reason to reject. And so, he concludes that “[r]especting the value of (rational) human life requires us to treat rational creatures only in the ways that would be allowed by principles that they could not reasonably reject insofar as they, too, were seeking such principles.”¹⁷ Of course these are just the contractualist principles that Scanlon takes to be constitutive of morality. So this view reduces the value of persons to our moral reasons. Valuing a person is just recognizing that one has reason to treat them as morality requires. And a person’s value is nothing other than the considerations that count in favor of treating them in accord with moral requirements.¹⁸
Valuing, Loving, and the Wrong Kind of Reasons There is a pleasing simplicity about this way of understanding the relationship between the value of a person and moral requirements. But, on ¹⁶ Scanlon, What We Owe to Each Other. p. 105. ¹⁷ Scanlon, What We Owe to Each Other. p. 106. ¹⁸ Though he hedges slightly, Scanlon concludes the account by affirming that “the idea of valuing human life and the idea of respecting one’s duties and other people’s rights ought to be closely related, if not the very same thing.” Scanlon, What We Owe to Each Other. p. 106.
-
85
reflection, the idea that a person’s value just consists in the moral reasons that we have with respect to her is not very plausible. The problem emerges most clearly when we think about those we love. Loving someone is an especially deep and profound way of valuing them. To love someone is to see value in her or see her as valuable. What you “see” here is not merely that you have reasons to treat her as morality requires. This is the Scanlonian idea that we can’t take to a funeral. Scanlon’s narrow focus on specifically moral reasons makes the view especially unattractive. But, in fact, the problem is deeper and will plague any attempt to understand valuing wholly in terms of acknowledging reasons to do, or refrain from doing, certain things and so to reduce value to practical reasons. Many accounts of love do try to understand it in something like this way. To love a person, some think, just is to take yourself to have special reason to promote his well-being, or to seek out his company. One might think that loving a person is a matter of recognizing reasons to do the dishes, buy him a gift, stay up late at night caring for her if she is sick, or more generally to promote his well-being or advance her ends.¹⁹ It is true that we often have these sorts of reasons with respect to those we love and plausible to suppose that love often partly consists in recognition of, and responsiveness to, some such reasons.²⁰ There is normally a real tension between loving a person and failing to understand or appreciate that one has reasons to act in certain ways with respect to them.²¹ But reductionist buck-passing views have to go further and claim that loving a person just is acting in some specified way or taking yourself to have reasons to do so.²² This further claim is implausible. Again, it is especially so if we identify the relevant actions as just those that are morally ¹⁹ Cf. Ebels-Duggan, “Against Beneficence: A Normative Account of Love.” ²⁰ But it may be better to say that many relationships that we stand in to those we love are constituted by responsiveness to these reasons, rather than that love itself is so constituted. ²¹ The tension in question is a rational one. It is not exactly the same as believing contradictory things, but it shares something in common with that sort of case. It is a matter of holding attitudes that are rationally incompatible with one another. ²² At least this is true so long as they agree with me that love is a species of valuing. Scanlon sometimes verges towards identifying love with feelings, regarded as arational brute or sensible responses (e.g., Scanlon, What We Owe to Each Other. p. 42). While that would allow some differentiation between loving a person and interacting with them on moral terms, it would be differentiation of the wrong kind. A simplistic empiricism about love and affection would contrast with Scanlon’s much more sophisticated and illuminating rationalism about desires. Cf. Brewer, The Retrieval of Ethics. pp. 62–5; Cf. Kant on practical and pathological love: Kant, Groundwork. 4:399. Immanuel Kant, The Metaphysics of Morals, trans. Mary Gregor (New York, NY: Cambridge University Press, 1996). 6:401–3. I have written on this distinction in Kyla Ebels-Duggan, “Love and Agency,” in
86
-
permissible—as on Scanlon’s official view. But even allowing for the wider range of actions referenced by the accounts of love that I’ve just mentioned, we can easily identify grounds for so acting, responsiveness to which would clearly not amount to love.²³ Perhaps one does the dishes just because one wants the kitchen clean or buys gifts as part of a program of manipulation. Or one might be moved by the idea of oneself as a loving person, or the wish to be perceived that way, rather than being moved by the person whom one purportedly loves. Someone acting on one of these grounds would take herself to have the reasons to act in the stipulated loving ways, but this would be compatible with failing to see any value in another person and so with failing to love them. Moreover, if love were reducible to the recognition that you have reasons to act in particular ways it would turn out to be an essentially calculative state. Learning to love a person would be nothing other than learning which moves to make when, which principles of action to apply in which situations. Even sophisticated machines could then perfectly well count as loving us. Of course, we can make progress in learning how to express our love well; we can get better at interacting with those whom we love. Learning this sort of thing is an important part of growing up. But learning how to make the appropriate moves is not the same as actually coming to value a person as we do when we love him.²⁴ A relationship in which one knew only that certain actions were appropriate, but had no independent grasp on the value that makes them appropriate, would be a failing relationship.²⁵ A distinction between meaningful loving actions and so-called “empty gestures” is internal to our understanding of loving relationships and interaction. But if we reduce the value of a person to the existence of reasons to act in some specified way, we will not be able to draw this distinction. Someone might suggest that this is unfair to the buck-passer. Above I acknowledged an ambiguity in Scanlon’s own account. Sometimes he suggests that saying something is valuable is claiming that we have reasons to act towards it in a particular way. Elsewhere, he suggests that it is to claim Routledge Handbook of Love, ed. Adrienne Martin (New York, NY: Routledge, 2019), and in Kyla Ebels-Duggan, “Beyond Words: Inarticulable Reasons and Reasonable Commitments,” Philosophy and Phenomenological Research 98, no. 3 (2019). pp. 623–41. ²³ Cf. David Velleman, “Love as a Moral Emotion” in self to self (New York, NY: Cambridge University Press, 2006). p. 89. ²⁴ Cf. Anderson’s discussion of an example from Star Trek: Anderson, Value in Ethics and Economics. p. 40. ²⁵ Cf. Thomas Scanlon, Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, MA: Harvard University Press, 2008). pp. 37–88. Brewer, The Retrieval of Ethics. p. 161.
-
87
of some of its particular properties that they provide reasons to act in the specified way. In this section, I have been construing his view of valuing someone in the first of these ways—as simply a matter of seeing that you have reasons.²⁶ But a buck-passing view might be on stronger ground if it resolved the ambiguity in the other way, attending more carefully to what I identified, above, as the second pass—from reasons to properties of the object. On this approach the reductive view would hold that valuing a person is not just seeing that one has reasons to act in particular ways, but of seeing some particular properties as providing these reasons.²⁷ But how, exactly, is this view supposed to go? Which particular properties provide reason to do the dishes, to visit you in the hospital, or to respect your rights?²⁸ When he introduces the buck-passing view of value, Scanlon puts it in terms of “natural properties.”²⁹ But the idea that we are going to be able to identify some natural property to serve as the correlate of each practical reason doesn’t seem very promising on its face. R. Jay Wallace argues that the restriction to natural properties is incompatible with Scanlon’s own examples.³⁰ Neither the pleasantness of a resort nor the illuminating nature of a scientific discovery are narrowly descriptive properties.³¹ He thus suggests that Scanlon should allow that thick evaluative properties can provide the reasons in question. Responding to Wallace, Scanlon immediately accepts this suggestion without further comment.³² He writes: My thesis was that goodness is not itself a property that provides reasons, not that the underlying properties that do this are always natural properties, and I should not have written in a way that suggested this.
²⁶ “To value something is to take one’s self to have reasons for holding certain positive attitudes towards it and for acting in certain ways with regard to it . . . To claim that something is valuable (or that it is ‘of value’) is to claim that others also have reason to value it, as you do.” Scanlon, What We Owe to Each Other. p. 95. ²⁷ Thanks to Sarah Buss and Benjamin Kiesewetter for pressing me to think more about this. ²⁸ Notice that one implication of this view is that we will not count as loving a person unless we get the answer to this question right. Apart from the other issues that I raise, this seems to be an inordinately demanding view of love. I suspect that Scanlon puts the view in the weaker way—just recognizing that there are reasons—because it avoids this implication. ²⁹ Scanlon, What We Owe to Each Other. pp. 96–7. ³⁰ Wallace, “Scanlon’s Contractualism,” Ethics 112, no. 3 (April 2002). pp. 441–49. ³¹ Wallace also suggests that the restriction to natural properties is incompatible with Scanlon’s account of practical reflection and his account of the epistemology of practical reasons in Chapter 1 of What We Owe to Each Other. ³² T.M. Scanlon, “Reasons, Responsibility, and Reliance: Replies to Wallace, Dworkin, and Deigh,” Ethics 112, no. 3 (April 2002). p. 513.
88
- [Wallace] is quite right that more specific evaluative properties often play this role.
This expansion of the set of reason-giving properties makes the view more attractive, but also much less distinctive. In fact, if the buck-passer appeals to evaluative properties, he hasn’t really passed the buck—at least he hasn’t executed the first pass, from value to reasons.³³ Understanding value in terms of more particular evaluative properties does amount to a rejection of what Anderson calls monism about value, that goodness is a single, unified property that provides us with reasons to act. But the opponent of monism need not be a buck-passer. Evaluative properties just are ways of having value or being valuable. To ascribe an evaluative property is not an alternative to recognizing something’s value but a way of recognizing it. To say that the properties that provide reasons are evaluative properties is to grant that we can’t really eliminate value talk and to give up the buck-passer’s game.
Respect and Moral Reasons So far, I have argued that the valuing attitude of love can’t be reduced to the recognition of reasons to avoid or perform some set of actions. It follows that the value of the person that one appreciates when one loves them cannot be reduced to a set of reasons for acting. This value may give rise to or ground a range of reasons for action, but it cannot simply be identified with such reasons. But one might wonder if this is just a point about love, which is, after all, a distinctively robust valuing attitude. Perhaps Scanlon appeals only to moral reasons because he’s not trying to account for the value we see in those we love. Perhaps he is just aiming to give a more narrowly targeted account of something that we might call a person’s moral value.³⁴ On this reading, he’s trying to make sense of what we attribute to a person when ³³ He could, for example, appeal to the property of having Kantian dignity, or of being an end in one’s self, as the basis of moral reasons. I am sympathetic with the idea that, in the end, we will need to abandon talk of “properties” altogether. Cf. Cora Diamond, “Eating Meat and Eating People,” The Journal of the Royal Institute of Philosophy 53 (October 1978). pp. 157–75; Cora Diamond, “The Importance of Being Human,” Royal Institute of Philosophy Supplement (1991). pp. 35–62. Cf. Raymond Gaita, A Common Humanity (New York, NY: Routledge, 2000). Kyla Ebels-Duggan, “Learning from Love: Reasoning, Respect, and the Value of a Person,” in Reconsidering the Value of Humanity, ed. Sarah Buss and Nandi Theunissen (2023). ³⁴ Thanks to Sarah Stroud and Rahul Kumar for pushing me on this point. Though Scanlon’s section does read as if he means to be talking about the value of a person—or the value of a life— full stop, the view might be more charitably understood in this more modest way.
-
89
we think that they have the sort of value that gives them “moral status” or “moral standing”—maybe this is what Kant means by dignity. His suggestion is that we understand these locutions as referring to the moral reasons that we have with respect to a person. I’m skeptical that it makes sense to think of a person as having one kind of value that we respond to when we love them and a different kind of value that we respond to in moral relationships.³⁵ But, setting that doubt aside, an argument parallel to the one about love shows that the buck-passing view would be unsatisfactory even for these more modest purposes. Even if we confine our attention to narrowly moral reasons and relationships, we cannot eliminate appeals to value. Above, I said that, in the context of a loving relationship, we care about the difference between loving actions and empty gestures. One might think that matters are different in the moral case. We have reason to want people to respect our moral rights and entitlements—to treat us in accord with principles that no one could reasonably reject—whether or not they harbor underlying attitudes of respect. And, on the flip side, it seems that we are entitled to demand a range of treatment, but aren’t, and couldn’t be, entitled to demand any particular attitude or regard from others. I think that we do, indeed, have reason to care about how others treat and interact with us without regard to the underlying attitudes that they have towards us. I further agree that there is an important sense in which we can be entitled to demand certain actions from others, but not to demand certain attitudes from them. But it does not follow that the valuing attitude of respect for a person is nothing other than the recognition that we have reasons to act in accord with the requirements of morality, or that the moral value or dignity of a person consists in just these reasons. Even in narrowly moral relationships, we reasonably care, not only about how others treat us, but also about the attitude towards us that motivates this treatment.³⁶ In mere passing interactions with strangers it can matter whether an offer of some
³⁵ Cf. Velleman, “Love as a Moral Emotion.” Ebels-Duggan, “Learning from Love: Reasoning, Respect, and the Value of a Person.” ³⁶ This is one insight that Kant was after when he argued that only actions done from duty have moral worth. Kant, Groundwork. 4:397–4:399. It is not enough that other people serve as the occasion for our moral performance. These actions have their special significance only if they are responsive to the claims that others have on us. It is also on this basis that Rawls distinguishes between a society in which there is an overlapping consensus about justice and one that constitutes a mere modus vivendi—why Rawlsian stability for the right reasons is a proper subset of stability. John Rawls, Political Liberalism (New York, NY: Columbia, 1996). pp. 133–72.
90
-
small assistance expresses appropriate concern about your needs or a condescending underestimation of your powers of self-sufficiency. And though we may want our neighbors to respect our rights in any case, it matters whether they do so only because they believe that they will otherwise suffer legal consequences or instead because they genuinely respect—that is, they genuinely value—us. This shows that valuing a person, even in the purportedly narrow sense at issue in moral relationships, outstrips recognizing reasons to act in given ways.³⁷ If we can distinguish between others treating us in the relevant ways for extrinsic reasons, on the one hand, or out of genuine moral respect, on the other, then we cannot reduce respect to recognizing reasons for action any more than we can reduce love to recognizing such reasons. That argument was given from the point of view of the moral patient, a person who has reason to care about how others treat and regard her. The irreducibility of the value of persons to reasons for treating them in certain ways also shows up from the point of view of the moral agent, a person considering how to interact with others. Suppose I ask why I should guide myself by moral principles or start to wonder about what makes sense of doing so.³⁸ Scanlon addresses this question of justification in his discussion of that aspect of the authority of moral reasons that he calls their priority. Here he is after an account of why it makes sense for us to treat moral reasons as overriding or silencing other sorts of reasons for acting. The account that he offers appeals to a value. It is not the value of a person, but the value of a certain kind of interpersonal relationship, which he calls “mutual recognition.”³⁹ We stand in this relationship just in case we interact on terms that no one could reasonably reject. Scanlon thinks that we have reason to care about whether we stand in this relationship to others. He does not offer an argument for this purporting to address someone who is skeptical about the value of this relationship, nor does he think that any such argument is possible. Instead, he reflects on the American social upheaval of the 1960s, diagnosing the debates and conflicts of that time as the expression of people’s investment in living together on
³⁷ In later work, Scanlon acknowledges this when he claims that moral permissibility governs only external action, but that the significance or meaning of permissible actions can vary with the attitudes underlying them. He agrees that we have reason to care about both of these aspects of an action. Scanlon, Moral Dimensions: Permissibility, Meaning, Blame. ³⁸ Christine M. Korsgaard, The Sources of Normativity (Cambridge; New York: Cambridge University Press, 1996). pp. 7–21. ³⁹ Scanlon, What We Owe to Each Other. p. 162.
-
91
fair terms. The civil rights movement, the anti-war protests, and other overlapping social movements confronted Americans with the possibility that the terms on which they were living could not be justified to all. Some reacted to this by joining in demands for major revision of these terms, while others defensively rejected the claim that the terms were unjust. Both reactions, Scanlon suggests, manifest an underlying concern with the relationship of mutual recognition. This is evidence that people do value standing in this kind of relationship. Scanlon holds that we are not wrong to do so and his reflections on this historical moment are meant to elicit our assent: mutual recognition is a relationship worth valuing. I find all of this both plausible and illuminating, and its renewed salience in the contemporary social-political context will be obvious to all. But what can it mean for the relationship of mutual recognition to be valuable? On a buck-passing view like Scanlon’s, it has to mean that we have reasons with respect to it. As with friendship, to which Scanlon explicitly compares the moral relationship, the most fundamental among these reasons will be those the recognition of which constitutes the relationship. Valuing friendship involves recognizing the reasons to be with one’s friend, advance their well-being, and share their ends—what Scanlon calls reasons of loyalty. Similarly, valuing the relationship of mutual recognition is a matter of recognizing that one has moral reasons. But so understood, there is an obvious problem with this account: it simply reasserts the existence of the moral reasons whose priority Scanlon aimed to defend, or at least to make intelligible. The value of the relationship cannot both ground the claim about the priority of moral reasons and also be wholly reducible to the moral reasons. Unless the value of the relationship is something over and above the existence of these reasons, the account is unsatisfyingly circular. So, here again, we see that we cannot do without a distinct concept of value. Thus, the lesson about love generalizes to the moral attitude, and indeed to all kinds of valuing attitudes. The underlying problem is the buck-passer’s attempt to reduce value to reasons.
The Value of a Person and Reasons for Attitudes So long as we try to understand value just in terms of reasons for actions we will face the obstacles above: wrong-kind-of-reasons problems and threats of circular justification. But Scanlon’s buck-passing view allows for another category of reasons—reasons to have certain attitudes towards a thing.
92
-
In this section, I explain why I think that this move also fails to rescue the buck-passing view. In his first-pass characterization of the reasons that constitute the value of human lives, Scanlon mentions three categories: reasons to protect these lives, reasons not to destroy them, and reasons to want them to go well. The first two of these are reasons bearing on action. They are reasons to do, or refrain from doing, things that we can choose to do or refrain from doing—things that are, in this sense, voluntary or under the control of the will. The last category is different. It is a set of reasons for an attitude, that of wanting. Adding recognition of reasons for just this one attitude to that of reasons for acting does not look like it would round out the valuing attitudes of either respect or love. But Scanlon seems to intend his list of reasons to be illustrative, not exhaustive, and once we allow that the reasons that comprise a person’s value can include reasons to take up certain attitudes towards her we are free to expand the set of those attitudes as we see fit. This may seem as if it would allow for a more adequate view of what it is to value a person, and in turn what it is for a person to have value. But, as we have seen, by the end of Scanlon’s discussion of the value of persons this third category of reasons falls away. He identifies valuing a person with recognizing reasons to treat her in accord with contractualist principles. These principles—“principles for the general regulation of behavior”—govern only action. They do not concern how we regard or think about one another. In Kantian terms, they set limits on our external freedom, our publicly observable acts.⁴⁰ I don’t think that this shift in Scanlon’s presentation of the view is merely accidental. In the first section I suggested that one of his central motivations for the buck-passing view may be that, as he sees it, reasons are better understood than values. But, in fact, we understand some reasons better than others: we arguably have a better grasp on reasons for action than we do on reasons for attitudes. Moreover, reasons for valuing attitudes raise special puzzles, as I will explain below. We can understand what reasons are in terms of the role that they play in thought and action. Let’s accept Scanlon’s account of reasons for actions: To take some consideration to be a reason for an action is just to regard it as counting in favor of doing that thing. It is, as Hieronymi puts it, a ⁴⁰ Scanlon, What We Owe to Each Other. p. 4 and Kant, Metaphysics of Morals. E.g. 6:224, 229.
-
93
consideration that bears positively on the question of whether to ф.⁴¹ This is perfectly ordinary and unmysterious. That there’s food in the fridge may be a reason for me to head to the kitchen. That it would make you happy might be a reason to buy a gift or to do the dishes. We can decide to act on the basis of any reason that seems to us sufficient and, absent external obstacles, straightaway act on this decision.⁴² To be a reason for action, then, is to function as a possible basis for choosing and performing that action. Reasons for attitudes function differently from reasons for actions because we do not adopt attitudes just by choosing to do so—attitudes are not under our voluntary control. Beliefs, for example, cannot be taken up at will. You cannot choose to have some belief on the basis of any consideration that counts in favor of having it. You might be wholly convinced that there are overwhelming pragmatic reasons to adopt a belief: you know that you will have a better chance of jumping the ditch if you believe you can do it, and you have to do it to escape some danger. But these sorts of considerations could, at most, lead to a decision to act in ways that might succeed in manipulating yourself into having the belief. Kavka’s famous toxin puzzle shows that intentions, somewhat surprisingly, work the same way.⁴³ Though we can act at will, we cannot intend at will. So reasons for these attitudes can’t be considerations on the basis of which we choose them.⁴⁴ But we can make sense of reasons for attitudes in a different way: as considerations that could figure in a line of reasoning that concludes with said attitude. Believing and intending are attitudes to which we can reason. If you reason through a sound argument to the conclusion that p, you ordinarily thereby come to believe that p. Similarly, a line of practical reasoning can conclude with the decision or intention to ф. While we cannot choose to believe p or choose to intend to ф, to be convinced that ⁴¹ Pamela Hieronymi, “The Wrong Kind of Reason,” The Journal of Philosophy 102, no. 9 (2005). pp. 437–57; Pamela Hieronymi, “The Use of Reasons in Thought (and the use of earmarks in arguments),” Ethics 124, no. 1 (2013). pp. 114–27. ⁴² Pamela Hieronymi, “Two Kinds of Agency,” in Mental Actions, ed. Lucy O’Brien and Matthew Soteriou (Oxford: Oxford University Press, 2009). pp. 144–5. We can even choose to act just on the strength of our conviction that there are some good reasons, without knowing or having a view about what these are. For example, I can trust someone’s report that I should take an alternate route home, and choose to do so, even if they don’t tell me the reasons for avoiding my normal path. ⁴³ Gregory Kavka, “The Toxin Puzzle,” Analysis 43, no. 1 (1983). pp. 33–6. ⁴⁴ With respect to attitudes like beliefs and intentions, we have two distinct sets of reasons. They are, in Hieronymi’s helpful terms, reasons for the attitudes, and reasons to manage the attitudes. Reasons for attitude management are just a subset of reasons for action, considerations on the basis of which we can choose to act. Pamela Hieronymi, “Controlling Attitudes,” Pacific Philosophical Quarterly 87 (2006). pp. 45–76.
94
-
we have sufficient reason to believe or to intend, or to be convinced by a line of reasoning, ordinarily amounts to having these attitudes. They are judgment sensitive in Scanlon’s sense and under our evaluative control in Hieronymi’s sense.⁴⁵ Not so for wanting, respecting, loving, or other valuing attitudes.⁴⁶ Reasons for valuing cannot be understood on either the model of reasons for actions that are under our voluntary control or the model of reasons for attitudes that are under our evaluative control. We cannot adopt a valuing attitude just by choosing to do so. Thus we cannot understand reasons for valuing as considerations that could bear on such a choice.⁴⁷ But neither can we reason to love, moral respect, or any other valuing attitude. There is no line of reasoning that we can think through that would conclude with these attitudes. This can be easy to miss, because we can reason to closely related beliefs. We can believe that there are considerations that warrant valuing attitudes, that make them, as some philosophers say, a fitting response to their objects. We can believe that something is worth wanting or worth loving. Or, more simply, we can believe that something is valuable. There is no reason to deny that, like other beliefs, these beliefs about value can, in principle, be conclusions of reasoning. We can then understand reasons for them as considerations that could figure in such reasoning.⁴⁸ But believing that something is worth valuing is not already to value it and these two attitudes can come apart. We can have the first without the second, and there is no rational failing involved in doing so—as when I judge that all children are worthy of love, though I do not love them all as I love my own.⁴⁹
⁴⁵ Scanlon, What We Owe to Each Other. pp. 18ff. Hieronymi, “Two Kinds of Agency.” p. 140. ⁴⁶ At least this is true if wanting picks out a valuing attitude: being inclined towards something, or feeling like doing it, or being in the grip of an attraction towards a thing. Cf. Brewer, The Retrieval of Ethics. Tamar Schapiro, “What are Theories of Desire Theories of?” Analytic Philosophy 55, no. 2 (2014). pp. 131–51, Tamar Schapiro, Feeling Like It (New York, NY: Oxford University Press, 2021). “Wanting” could also be used in something like what Schapiro calls the placeholder sense, where wanting to ф is interchangeable with thinking that I have reasons to ф. Cf. Thomas Nagel, The Possibility of Altruism (Oxford: Clarendon Press, 1970) on motivated desires pp. 29–30; Ebels-Duggan, “Beyond Words: Inarticulable Reasons and Reasonable Commitments.” ⁴⁷ This explains the tendency to think that moral reasons could regulate only actions, which can be demanded or required of us, and not attitudes, which cannot be. You cannot be required to do something that you cannot choose to do. ⁴⁸ Though there are important questions about how we can acquire the normative concepts that support that line of reasoning. Ebels-Duggan, “Learning from Love: Reasoning, Respect, and the Value of a Person.” ⁴⁹ Samuel Scheffler, “Valuing,” in Reasons and Recognition: Essays on the Philosophy of T.M. Scanlon, ed. R. Jay Wallace, Rahul Kumar, and Samuel Freeman (Oxford: Oxford
-
95
To see the distinctive nature of reasons for valuing attitudes in another way, consider Hieronymi’s characterization of attitudes like beliefs and intentions that are under our evaluative control and her understanding of the reasons for them. She holds that such attitudes embody our answers to a question. My belief that p just is my answer to the question whether p? and my intention to ф is my answer to the question whether to ф? But there is no question to which love, moral respect, or any other valuing attitude is an answer.⁵⁰ The best candidates for such a question—is he worth loving? Or perhaps does she have the moral status of a person?—would be answered by beliefs related to valuing of the sort that we just considered, not by valuing itself—and, again, one could answer the questions affirmatively without holding the corresponding valuing attitude. We can reason to such judgments and we can understand reasons for them as whatever could figure as premises in such reasoning. We can also choose, on the basis of these judgments, to treat a person in certain ways. But we can neither reason to, nor choose to see or experience or regard a person in certain ways. And the latter is essential to valuing them.⁵¹ Could the buck-passer take this insight on board and maintain that the reasons for attitudes that partially constitute values are just reasons for the beliefs that something is valuable or that it is worth valuing? I doubt he can do so while maintaining the important distinction between valuing and judging valuable. But there is a further problem as well: this version of the buck-passing view could give no account of the content of the belief that something is valueble and, in turn, would render it impossible to understand what it would be for there to be reasons supporting their belief. The idea is to understand the value, here the value of a person, in terms of reasons, among which are reasons to believe that he is valuable. But what does this belief come to? On the buck-passer’s view, it’s a belief that we have reasons with respect to the person. These may include reasons for action, but we’ve seen that it won’t work to University Press, 2011). We can also value something while not believing that it is worth valuing, or even while believing that it is not worth valuing. But these combinations of attitude do seem to be in rational tension. ⁵⁰ Another way to put the point is that valuing, unlike belief and intention, is provisional. It is not a mistake or confusion to get critical distance on it. Cf. Schapiro, Feeling Like It. pp. 31, 62. We need to be careful here since there do seem to be some cases (e.g., love of one’s children) in which it would be wrong to distance ourselves from the content of our valuing attitude. EbelsDuggan, “Beyond Words: Inarticulable Reasons and Reasonable Commitments.” But the error here is a substantive one. By contrast, believing and doubting p at the same time or intending to ф while also wondering whether to ф, is a formal error in reasoning. ⁵¹ Cf. Iris Murdoch, The Sovereignty of Good (New York, NY: Routledge and Kegan Paul, 1970).
96
-
stop there, so the suggestion is to include reasons to have the belief in question—whatever it is. But the buck-passing approach leaves us with nothing to fix this further content of the belief. To make sense of the idea that a consideration counts in favor of a belief we need some independent grasp on what it is that is believed. But, with respect to the beliefs in question, the buck-passing view denies that there is, as it were, anything independent here to grasp. The view does no better with the belief that the thing is properly valued, because it lacks an informative account of what valuing is, or what any of the more specific valuing attitudes are. We could run the argument in terms of love again, but instead consider Talbot Brewer’s discussion of admiration. To admire a person, is to see her as having a particular kind of value in virtue of her traits or achievements. Brewer names this value merit.⁵² Since merit is a species of value, on the buck-passing view we should be able to replace talk of merit with the citing of reason-giving properties. It looks like we can: If I admire you as a musician, I will think that you have merit in virtue of your ability to sight read and play pieces by ear, your technical skill with various instruments, and your sensitive interpretation of the music. But to say that these properties provide reasons is to affirm not merely that a person has them, but also that they call for a certain response.⁵³ Which response? What are these sorts of properties reasons for? The answer might again partly consist in actions, but we’ve seen that we can’t stop there. The suggestion is that they are also reasons to believe that the thing calls for admiration or is properly admired. But now the buck-passing account has it that the attitude of admiration includes recognizing reasons to take up the attitude of admiration—whatever it is. We seem to have a tight, and so uninformative, circle of concepts.⁵⁴ Without an account of the content of these attitudes, we can’t understand the judgment that a thing calls for one of them—that it is worth admiring, respecting, or loving. Just as the value of the relationship of mutual justification disappears into the reasons for treating others in certain ways, so the content of attitudes of valuing seems to disappear into
⁵² Brewer, The Retrieval of Ethics. p. 156; Cf. Stephen Darwall, “Two Kinds of Respect,” Ethics 88, no. 1 (1977). pp. 36–49. ⁵³ This is just to reiterate the point that the “second pass” of the buck-passing view, from properties of the thing to reasons, is non-reductive. ⁵⁴ Brewer, The Retrieval of Ethics. p. 155. Brewer suggests that we might rescue the buckpassing view from circularity by relying on a generic pro attitude but would pay an unacceptable price in terms of articulacy: cf. his museum example on p. 159.
-
97
convictions that we have reasons for acting, throwing us back into the wrong-kind-of-reasons problem that we have already discussed.⁵⁵ To summarize, reasons for valuing work differently, that is they have a different role in our practical thought and responses, than either reasons for actions or reasons for the attitudes that are under our evaluative control. When he suggests that valuing a thing might involve recognizing reasons to have certain attitudes towards it, Scanlon does not acknowledge or comment on these differences. But they create obstacles for applying our understanding of reasons for acting, or reasons for believing or intending, to reasons for having the sorts of attitudes that seem most likely to figure in an account of valuing. Pace Scanlon’s earlier claims, then, reasons for valuing are not something that we understand clearly, and so might use to illuminate the nature of value. This is not to say that we don’t have reasons for valuing. Valuing clearly is subject to some evaluative standards; it is something that we can do well or badly. It is, rather, only to observe that we can’t take these reasons for granted, and then make sense of value in terms of them, as the buck-passer suggests. Our grasp on reasons for valuing attitudes is no better than our grasp on value itself. It seems more promising to understand the former in terms of the latter than vice versa. On that sort of account, the value of what we appreciate gives us reasons for certain actions and attitudes, rather than being reducible to such reasons. We can and do judge that we have reasons both for acting in certain ways and for having certain valuing attitudes. When we ask after the basis on which we could reach these judgments about reasons, it seems to me that the best sorts of answers will appeal to what we learn from, or what is revealed by, our experiences of value—for example from what we see in a person when we love her. Loving a person is a matter of seeing him as having value. The value-first approach allows that there is something that we are seeing when we see a person this way, some content to what is seen. It is difficult—I have elsewhere argued that it can be impossible in principle—to articulate this content fully, such that we could communicate to another the entirety of what it is that we see. And of course sight itself is only a metaphor here. But, I have argued, it cannot be that what we see is merely reasons to interact on terms that no one could reject, or indeed reasons to act in any particular way. Nor can it be that it is just these reasons and, in addition, reasons to have some attitude.
⁵⁵ Cf. Brewer, The Retrieval of Ethics, on grief on p. 176 and also “tenderness,” pp. 162–9.
98
-
Valuing attitudes are attitudes that we reason from, not attitudes that we reason to.⁵⁶ They provide the content on the basis of which we reach conclusions about what is worth valuing and how best to express these values in action.⁵⁷ We grasp these reasons in experiences like wanting to be able to justify our interactions with others and, more fully, in experiences of loving individual persons. That is why when we try to express what a person means to us, the way we value her, the value of her life, we almost always talk about these experiences. We tell stories about our interactions with her, about our times together. We relate what a person said and did and try to convey what it is, or was, like to do things with them. These stories cannot fully capture the value that we experience, but they are nevertheless the best we can do by way of communicating something about the value of her life.⁵⁸
⁵⁶ Cf. G. E. M. Anscombe, Intention, 2nd ed. (Cambridge, Massachusetts: Harvard University Press, 1963). p. 66. Cf. Aristotle on reasoning about ends. Aristotle et al., The Nicomachean Ethics (New York: Oxford University Press, 1998). pp. 56–7. ⁵⁷ Here compare the relationship of perception of your surroundings to beliefs about them. ⁵⁸ Thanks to T.M. Scanlon, Richard Kraut, Sarah Buss, Sarah Stroud, Rahul Kumar, Benjamin Kiesewetter, and Sergio Tenenbaum for comments on and discussion of this paper.
5 Moral Theorizing and the Limits of Coherence Sarah Buss
The two most popular moral theories seem to get so many things wrong. But they do appear to have one advantage over the alternatives. If morality were as the consequentialists or the Kantians conceive it, there would be no incompatible moral requirements. On either of these accounts, an ideal moral agent might have many incommensurable values; she might have conflicting desires. But every specific permission or requirement would be justified in terms of a more general fundamental moral principle. And there would be no circumstances under which an ideal moral agent would be forced to violate a moral requirement. The Kantians and the consequentialists are united in assuming that an accurate account of the moral terrain resembles an accurate account of the physical terrain in being an account of something whose component parts fully cohere with each other. The alternative is “common sense,” “ordinary,” morality. As Shelly Kagan observes, “Despite the immediate appeal of our [widely shared, nonKantian] deontological moral intuitions, they cannot be incorporated into an adequate overall moral theory, and in this regard, at least, our moral intuitions are unreliable.” (Kagan 2001, 53)¹ I am interested in Kagan’s notions of “adequacy” and “reliability,” and in their relation to the ideal of coherence (and the related ideal of systematicity). I want to better understand the assumption that insofar as our moral convictions fail to cohere, they fall short of the truth. Should we endorse this assumption? I do not think the answer is obvious. We can, it seems to me, aspire to greater coherence in our moral lives and our moral theories without thinking that we are confused insofar as we endorse conflicting moral requirements and conflicting moral assumptions. ¹ See also, Barbara Herman (1993). Sarah Buss, Moral Theorizing and the Limits of Coherence In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Sarah Buss 2023. DOI: 10.1093/oso/9780198895909.003.0006
100
My aim in this paper is to explore the significance of moral incoherence. In the first half I will argue that, given the relationship between practical reason and substantive moral ideals—and given, correspondingly, the way human beings develop into morally virtuous beings—a certain incoherence is essential to being morally good. The necessity of incoherence is essential to the distinction between right and wrong; and this is because the distinction between right and wrong is a function of the interdependent relation between what is maximally indeterminate (responding appropriately to reasons) and what is more determinate (responding in ways that manifest various substantive moral and nonmoral virtues). This sort of incoherence differs from the sort that characterizes the “sophisticated consequentialist.” This person believes that her substantive moral commitments are justified by considerations that are not decisive from the point of view of the commitments themselves. Just as Kagan and other consequentialists take the lack of coherence in (nonKantian) deontological theories to be a reason to reject these theories, so too, many moral theorists reject consequentialism on the ground that its criterion of right action cannot serve as an action-guiding principle. In the paper’s second main section, I will argue that “supererogatory” actions (as they are typically characterized) have a similar structure: it is not possible to reconcile (i) the belief that a given action has more to be said for it morally than any alternative with (ii) the belief that the action is morally optional; one’s moral point of view is internally conflicted insofar as (i) one takes one’s action to be morally better than the alternatives even though (ii) one also believes that one is not morally obligated to act this way. Does this mean that this internal conflict is undesirable, all things considered? I will rather tentatively defend a negative answer to this question. There might, I suggest, be good reason to endorse a certain “sophisticated deontology,” according to which at least some agents are justified in occupying a practical point of view that cannot be reconciled with their own considered moral commitments. In exploring this suggestion, my main aim will be to prompt further reflection on the costs and benefits of moral incoherence, and on what these costs and benefits imply about the reasoning enterprises in which we are engaged as both moral theorists and moral agents.
Moral Virtue There are various widely shared assumptions about the traits that are characteristic of a good person. To keep things simple, let us focus on just
101
two of these traits: (i) a morally good person is beneficent and (ii) a morally good person is just. (What I have to say about these two virtues applies equally to the others.) As Aristotle notes, in order to develop virtues, someone must start with some basic raw material. She must begin with the natural correlate of each virtue (the “natural virtues” (Aristotle, 1941, Book VI, Chap. 13, 114b1-1145a12, pp. 1035–36)). Otherwise put, the commitment to treating people with concern and respect owes its determinacy to the substantive ideals that (as ideals) we take ourselves to have reason to live up to independent of any weighing of reasons for and against. We rely on these ideals, and the reasons that are apparent to us given that we have internalized these ideals, in order to determine what counts as “treating people with concern and respect.” In order to develop the moral virtue of beneficence, someone must have at least a weak disposition to help out when she sees someone in need, even though she has little understanding of when such responses are called for and when they are not. In order to develop the moral virtue of justice, she must be at least somewhat inclined to deal fairly with people and to expect them to treat her and others fairly too, even though she may see instances of unfairness where they do not really exist, is often overly harsh in responding to someone who fails to give others their due, and is even willing to do the wrong thing in the interest of justice. How does such a person become truly virtuous? Following Aristotle, I want to suggest that this is a process in which each of the natural dispositions is conditioned by her maturing sense of how it is appropriate to respond to her circumstances, where this gain in practical wisdom is, at the same time, conditioned by her natural beneficence and her natural commitment to justice. As a human being matures ethically, her natural virtues enable her to refine her understanding of what she has reason to do (and feel, and see—in what follows, I will focus exclusively on dispositions to act); and her developing capacity to reason enables her to modify these dispositions in such a way that they become different aspects of a single disposition to do what she has reason to do. Becoming truly beneficent involves becoming beneficent in such a way that one is just. Becoming truly just involves becoming just in such a way that one is beneficent. And, again, being beneficent-in-such-a-way-that-one-is-just-and-just-in-such-a-waythat-one-is-beneficent is being such that one is disposed to do what the circumstances call for, whatever these circumstances may be. If two virtues merge into one, then—trivially—there can be no conflicts among the responses they dispose us to have. If, moreover, these virtues were the only constituents of the practical point of view of the person who
102
had them, then there could be no conflict between what the agent regards as called for insofar as she has these virtues and what she regards as called for, all things considered. Of course, it is unlikely that anyone ever fully realizes this ideal of perfect unity. But I want to defend something stronger: this ideal cannot be fully realized, and it ought not to be fully realized, even if it could be. I will argue, first, that (i) a perfectly coherent moral point of view could not be more than a regulative ideal; the possibility of conflicting commitments is built into the moral point of view. I will then defend the stronger thesis that (ii) even if we could alter our ideals so as to avoid this deeper incoherence, it would be undesirable to do so. According to this argument, there are powerful reasons to doubt that one moral theory is superior to another insofar as it avoids endorsing any incompatible requirements.
The less-than-full coherence of the morally virtuous person When I say that it is not possible to realize the ideal of a perfectly unified practical point of view, I mean that it is not possible for each natural virtue to ground a distinct normative take on the world, while at the same time evolving into the sort of disposition that can be fully integrated into a point of view from which no response appears to be called for unless one takes oneself to have sufficient reason to respond in this way. Insofar as natural beneficence and natural justice are the raw materials from which we develop our capacity to appreciate what we have sufficient reason to do (and to respond accordingly), we cannot develop this capacity without being disposed to see our responses under two different substantial descriptions, neither of which owes its content exclusively to the content of the other, and neither of which is reducible to “whatever I have sufficient reason to do.” This means that in order to develop into a virtuous human being, we must rely on dispositions that are impediments to the unity of the virtues. This is not an observation exclusively about some early stage in our moral development. If the practical point of view of the mature moral agent is constituted by different substantive commitments (to being kind, generous, just etc.), then there is never a stage in the developmental process at which these commitments are irrelevant to what is at stake for this agent when she is assessing her reasons for action. To stick with the simple case in which a person begins her moral training with only two natural virtues: in wondering whether this person has made moral progress, we are necessarily wondering whether her present practical point of view is constituted by the
103
commitment to being both beneficent and just. We cannot answer this question without relying on some conception of what it is to be just and some conception of what it is to be beneficent, each of which is more substantial than the virtue of being responsive to reasons. The aspiration to be just and beneficent is the aspiration to be disposed to act justly and beneficently, where this is understood as the disposition to do what the situation calls for. It is this combination of the substantial and the insubstantial that makes it possible for the initial—“natural”— dispositions to develop into genuine virtues, dispositions that really are responsive to what one has reason to do under each distinct circumstance. Yet precisely because this process depends on substantive commitments that are not reducible to the commitment to doing what is called for, all things considered—precisely because without the substantive commitments, there would be no normative/evaluative material to work with—the integration of the natural into the practically wise can never be complete. It could not be complete unless the substantive ideals were to lose their substance. But without this substance, there would be no basis for determining what someone must do under any given circumstance in order to be responsive to reasons. In sum, each distinct ideal can (and must) be characterized in such a way that actions which satisfy certain descriptions just do not count as living up to this ideal, no matter how compelling the reasons may be for acting this way. Otherwise put: each substantive ideal has a purely descriptive aspect which limits the extent to which being governed by this ideal involves responding to reasons. The fact that our substantive ideals impose such limits is essential to the role they play in our reasoning. It is what makes it possible for our reasoning to get off the ground by making it possible for us to see certain facts as reasons without appealing to any additional considerations to justify assigning them this role. This does not mean that we could not possibly be mistaken about whether a given way of being—descriptively characterized—is a good way of being; it does not mean that we could never discover sufficient reason to conclude that we should reject the internalized ideal—and the normative appearances it makes possible—as confused. It does mean, however, that we could not make this discovery about all our ideals—or at least not all at the same time. More importantly for my purposes, given that our substantive ideals are not themselves entirely responsive to independent normative assessments, they cannot lose their credentials as ideals simply because there are circumstances under which the requirements they
104
underwrite conflict with what we have most reason to do, all things considered. No matter how compelling my reason for beating an innocent stranger may be (e.g., even if I must beat her in order to prevent her and many others from being killed by the psychopath who has credibly informed me that these are my options), in treating her this way, I violate the requirement to treat innocent human beings kindly. (In short, this requirement is not just pro tanto, even though it can conflict with what I am required to do, all things considered.) The limited incoherence thus tolerated can be compared to that which is highlighted by the Preface Paradox: (1) Insofar as I am beneficent and just, I am committed to responding beneficently and justly in every possible circumstance; (2) Given that the two virtues cannot fully merge into one, my present commitments are such that there are possible circumstances in which, as long as I maintain my interpretation of these commitments, I will face conflicting requirements. Accordingly, (3) I cannot respond in all possible circumstances in the way I am now committed to responding in all possible circumstances. This analogy might suggest that even if the potential for conflict among substantive ideals/commitments can never be eliminated, whenever an actual conflict arises, an agent can and should refine or abandon at least one of the conflicting ideals. According to this suggestion, though an adequate moral theory must make room for a latent incoherence in an agent’s commitments, it need not and should not endorse any transparently conflicting requirements. I believe, however, that moral agency requires a more deeply entrenched incoherence than this. I will argue that at least some of the conflicts that can arise, given our present commitments, are conflicts we cannot and should not resolve. According to this ideal of moral agency, an ideal human being is constituted by a wide range of heterogenous commitments. In fleshing out this ideal, I will, in effect, be explaining why a moral theory is inadequate if it has no place for less-than-fully coherent moral requirements.
The desirability of a less-than-fully coherent moral point of view The claim that incoherence is an irremediable aspect of the moral point of view reflects the assumption that the moral point of view is constituted by a heterogeneous collection of partly determinate ideals, each of which
105
incorporates the indeterminate ideal of responding appropriately to one’s circumstances. One way to resist this claim (a response that brings us closer to the conception of moral agency with which this paper began) would thus be to reject the underlying assumption, identifying the moral point of view with something extremely indeterminate—e.g., the ideal of acting only on principles that no one can reasonably reject (where, for all I am going to say here, these could be principles conformity to which maximizes [expected] utility). Even on this alternative account, however, a person’s practical point of view would be a site of conflict as long as she had any substantive commitments. Indeed, on this account, the moral point of view would owe its coherence to the fact that all substantive commitments were assigned to another location within the deliberating self, and that all conflicts were thereby limited to extra-moral conflicts among these commitments and to conflicts between these commitments and the single commitment constitutive of moral virtue. A further modification might seem to rule out such conflict: rather than being a collection of commitments (to act justly, to act beneficently, to act only in ways that others cannot reasonably reject, etc.), might not all substantive constituents of the practical point of view simply be dispositions to attribute a given salience and weight to certain considerations (under certain circumstances)? This conception of practical reasoning is compatible with the fact that practical reasoners like and want and value many different, even incommensurable, things. It is thus compatible with the fact that choosing one option from among others is rarely a simple matter of “weighing.” But though a rational agent so conceived would often regret having to forego one action in order to perform another, she would never find herself stuck with conflicting commitments. As long as she did what she took herself to have sufficient reason to do, she would not fail to “live up to” her ideals. She would be invulnerable to the possibility of witting self-betrayal. This is not a portrait of an ordinary human being. Because we have substantive ideals, we apprehend our circumstances as calling upon us to do certain things and not others. This is an essential aspect of our orientation toward the world; we cannot cease apprehending things this way whenever it also seems to us that these normative appearances do not perfectly conform to our assessment of how the reasons add up. To be sure, the apparent requirements do not rule out all (even most) exceptions. Nonetheless, they imply that which exceptions are justified is not simply determined by what we have most reason to do, all things
106
considered. (The same point applies to the not-quite-as-thin ultimate end of “flourishing,” or “living a good human life.”)² Given how integral some substantive commitments are to our identities, it does not seem as though we can abandon them at will. Nor does it seem that we can be sufficiently detached from them to be willing and able to adjust them in whatever way complies with our commitment to “responding to reasons.” But what if we could? Would this be an improvement? To the contrary, someone who satisfied this condition would have an extremely impoverished identity. Though she would value some things more than others, these would be the valuings of a connoisseur, an appraiser. She would not, as we say, be invested in anything. Or rather, she would be invested in only one thing: acting in a way that reflects/expresses her appraisals. In an obvious sense, the actions of such a person would indicate what she cares about. But—again, on the problematic assumption that it is possible to care about things without having substantive ends that are neither provisional nor derivative—there would be nothing she stands for in the sense that what someone stands for is integral to her identity. Because she would have various pro- and con-attitudes, she would also have interests; and so she could be hurt. But this sort of harm is more like the harm to which we are subject in virtue of our capacity to feel pain. It is not the harm of being forced to betray our commitments. It is not an instance of selfbetrayal. I hope the readers of this paper agree that I have described a highly unattractive human being. From what I can tell, even if it were possible for us to be more fully coherent than the person whose virtuous character evolves from her crude, natural understanding of how she ought to respond to her circumstances, the price would be far too high for any reasonable human being to pay.³ It seems, then, that the perfectly virtuous agent is a less-than-fully coherent agent. This is an observation about the necessary conditions of moral virtue. But it is also an observation about the possibility of clear-eyed incoherence: it is sometimes reasonable for someone to have a general end, believe that she cannot achieve this end without pursuing certain
² For a very interesting critique of Aristotle’s attempt to account for all practical ends as instrumental or constitutive means to this single end, see Lear (2002). ³ It is interesting to compare my remarks in the last couple paragraphs with Frankfurt’s remarks on “volitional necessity.” (Frankfurt 1988) According to Frankfurt, someone who is subject to volitional necessity is psychologically constrained by commitments and “cares” that are essential to her identity. Importantly, these commitments and cares can conflict with her normative verdicts—and she can be aware of this fact.
107
more specific ends, and yet wittingly fail to form the more specific intentions regarding the necessary means.⁴
A Second Form of Moral Incoherence? Throughout this paper’s first section, I have characterized the inner conflict that interests me as a form of “incoherence.” If some readers are disinclined to accept this characterization, this is, I suspect, because they endorse the very thesis I am challenging: they assume that incoherence is intrinsically, essentially to-be-avoided. I have argued that unless our practical point of view is empty of content, we have conflicting ends. And to have conflicting ends is one way of being incoherent. I have also argued that because these ends play a critical role in our apprehension of reasons and requirements, and because in this capacity, they are an important aspect of our identity, the fact that they are not fully coherent is an inadequate basis for criticism. Are there other desirable forms of moral incoherence? More generally, what distinguishes desirable from undesirable forms? My reflections in the preceding section suggest that the answer may lie in whether we can justify preserving the conflict. But this just seems to provide us with a different way of posing the same question: what sorts of justification are the right sort? The sophisticated utilitarian has moral commitments she thinks cannot be justified on their own terms. Yet she takes herself to be justified in maintaining these commitments—justified on utilitarian grounds. This justification is destabilizing precisely because it supports holding certain moral beliefs which are, according to this very justification, false. In contrast, the justification I have offered in the preceding section can be accepted by anyone who is conflicted in the relevant way because it is internal to the commitments themselves. What should we say about the sort of incoherence manifested by an agent who is a sophisticated utilitarian? To explore this question, I will turn my attention from (i) the constraints which our substantive ideals impose on the role that the ideal of coherence can and should play in our understanding of what we have reason to do to (ii) the constraints which the ideal of ⁴ This point is, of course, compatible with the fact that there is much we can do to minimize the incidence of such conflicts. Whether the ideal of motherhood is generally reconcilable with the ideal of engaging in meaningful work outside the home depends almost entirely on the political, social, and economic structures within which women are making their choices. For an elaboration on this point, see Marcus (1980).
108
coherence imposes on our (understanding of) our substantive ideals.⁵ I will argue that these coherence constraints provide the basis for a distinct critique of the possibility of supererogation. According to this critique, someone is confused about the moral significance of her action if she believes that what she is doing is morally superior to other things she is morally permitted to do under the very same circumstances. She fails to recognize that giving extra moral weight to certain considerations is not justified on her own terms.⁶ In assessing the moral significance of her action, the agent who regards her action as supererogatory resembles the akratic agent in counting the same consideration twice. Akrasia is a paradigm case of irrationality. Nonetheless, I will tentatively suggest that when the same sort of incoherence is manifest in cases typically characterized as “supererogatory,” it may well be justified; as the sophisticated consequentialist claims, a moral agent may sometimes be justified in occupying a practical point of view that cannot be justified on its own terms.
Akrasia By stipulation, the weak-willed agent acts contrary to her “best judgment.” One reason why this is possible is that it is possible to fudge the relationship between one’s evaluative judgments (e.g., “all things considered, it is good ⁵ One of the main lessons of this paper is that the content of what I am calling our “personal ideals” has a certain complexity. To shift from talk of “ideals” to talk of “thick ethical concepts”: (i) some of the descriptive content of the concepts I am discussing constrains what can count as being good in the relevant way (this is the anti-holist bit); but, at the same time, (ii) the rest of the descriptive content is constrained by the descriptive content of the other concepts (this is the holist bit). Points (i) and (ii) illuminate the widely shared intuition that it is not possible to “disentangle” the descriptive from the evaluative elements of such concepts. Point (ii) straightforwardly states this entanglement: the right description depends on the right evaluation (which depends on the proper application of other thick ethical concepts—so, to this extent, the entanglement of the evaluative and the descriptive is entangled with the entanglement of each thick ethical concept with the others). But point (i) suggests that there is more to the interdependence than this: given (i), if no instance of kindness is in this respect good, then there just are no instances of kindness. (To be kind is to be good in this respect. So if nothing with the same descriptive content can have this evaluative significance, then nothing can be kind.) ⁶ On the “paradox of supererogation,” see Sidgwick (1907). Harman (2016) distinguishes the “puzzle of supererogation” from “the paradox of supererogation.” The paradox asks how there can be actions that are favored by moral reasons but not morally required. Clearly, these two questions are closely related: if, as I will argue, supererogatory action is not conceptually possible, then it is a puzzle how so many people could think otherwise. I address this puzzle insofar as my dissolution of the paradox sheds light on how the relevant confusion might arise and insofar as I suggest that there might be good reasons to encourage it.
109
to limit my intake of sweets”) and one’s actions (e.g., this act of eating a second piece of pie). In short, one can almost always find a way to reassure oneself that one is not strictly intending to do what one is opposed to doing, all things considered. (“I’m going to eat just this one little sweet treat”; “I’m going to eat something sweet just this once”; and: “Since it is just this one, just this once, I am not really defying my own evaluative standpoint.”) Often one can avoid such rationalizations by simply not connecting the dots. Think, for example, of someone who believes that she should make significant sacrifices to slow global warming, yet flies all over the planet, buys food in plastic containers when there are readily available alternatives, etc. Such a person can continue to tell herself that she is going to make those sacrifices “sometime soon,” that “this is not the right time.” On many occasions, moreover, she can cover over the conflict—keeping it beneath the surface of conscious experience—without resorting to any form of rationalizing. She can simply “not notice” the apparent tension by limiting the scope of the reasons she considers in deciding what to do. As she is busy planning her international travel, she need not give the least thought to how this relates to rising global temperatures; and so, she need not give the least thought to how her travel plans relate to her evaluation of actions that contribute to rising global temperatures. Neither of the two sorts of cases I have just described involves selfconscious synchronic incoherence. In both cases, the conflict between one’s evaluation and one’s intention (and between one’s more general intention and one’s more specific intention) is not an avowed conflict. I will not repeat here the critique I have offered elsewhere of the possibility of clear-eyed akrasia.⁷ I do, however, want to note that rather than acting against her own best judgment by conveniently overlooking, or downplaying, certain facts she regards as counting against her action, an agent can accomplish the same result by regarding her all-things-considered evaluative judgment as the verdict of exclusively theoretical reasoning about what it would make sense for someone in her situation to do; she can evaluate her own agency in the way she evaluates the agency of others, with no intention of being constrained by her own self-appraisal. Insofar as someone maintains such a detached stance toward her own action, she can act contrary to “her best judgment” without experiencing any agential conflict.
⁷ See Buss (1997) and Buss (2020a).
110
As others have noted,⁸ the akratic agent, in effect, counts the same consideration(s) in favor of her action twice: once, in determining what she has sufficient reason to do in a wide range of circumstances; and a second time in determining what, under one of these very circumstances (in which she now finds herself), she has sufficient reason to do. If my reflections on moral virtue are correct, then if the rationalization or inattention or self-alienation that underwrite this sort of irrationality are appropriate targets of criticism, this cannot simply be because they are instances, or constituents, of one or more forms of incoherence. I have elsewhere offered an account of where the problem lies.⁹ Here, I want to focus on the two conditions mentioned above: (i) the akratic agent’s inclination to do something motivates her to greatly limit the extent to which she considers the reasons against doing what she is inclined to do and (ii) insofar as she attends to these reasons, she is moved to regard them exclusively from the stance of a detached observer who has no intention of treating them as constraints on what she decides to do. I want rather tentatively to suggest that one or both of these conditions is also a necessary condition of certain widely admired actions: so-called supererogatory actions. If I am right about this, then the relationship between coherence and moral virtue is even more elusive than I have thus far suggested.
Supererogation Of actions that are often characterized as “supererogatory,” there is nothing especially perplexing or noteworthy about (i) cases in which someone simply takes morality to be much more demanding than most of the rest of us do or (ii) cases in which (a) someone believes that she is morally permitted not to promote someone else’s interests in a certain way, yet (b) she acts this way in response to her personal cares and concerns. The interesting cases are (iii) those in which (a) someone believes that she is morally permitted not to do X and (b) she believes that the moral reasons in favor of doing X are stronger/more compelling than the moral reasons against—so that doing X has a greater moral value than not doing X.¹⁰ ⁸ See Davidson (1980). ⁹ See Buss (unpublished manuscript). ¹⁰ The classic paper on the topic is Urmson (1958). For some recent discussions of supererogation, see Portmore (2008), Horgan and Timmons (2010), Crisp (2013), Heyd (2015). (The
111
I want to suggest that this pair of beliefs is confused. It reflects a confusion about what it is for a consideration to qualify as “moral.” In discussions of supererogation, “moral” reasons are often identified with “other-regarding” reasons. But this cannot be right. For one thing, many people think that some moral considerations concern one’s relation to oneself. Less controversially, there are plenty of other-regarding considerations that do not qualify as moral: “I did it to make her cry”; “I did it to make her laugh”; “I did it to see how she would react”; etc. Of course, others might not be justified in demanding that someone do something she is morally obligated to do, and they might not be justified in criticizing someone for doing otherwise.¹¹ But, of course, such cases are not cases in which someone is morally permitted to do something she has stronger moral reasons to refrain from doing. One could try identifying the “moral” with a disjunction of certain substantive considerations, e.g., Xing would be beneficent, or would further a just cause (to be distinguished from doing what justice requires). But on this construal, there is no interesting sense in which “supererogatory” actions are morally superior to any other morally permissible actions. Nothing in the story of why the agent treated these substantive considerations as decisive would appeal to a distinctively moral reason to do so. In saying that someone did something “supererogatory,” one would simply be saying that she did something helpful or something that furthered a just cause, when she was not morally required to do this. One would not be saying anything about why this qualifies her as doing what she has most reason to do, all moral things considered. More importantly, as we have already noted, not all actions that satisfy the relevant substantive descriptions are properly regarded as responses to distinctively “moral” concerns. The moral point of view—the point of view of a virtuous person—is constituted by commitments to beneficence and justice that are as responsive as such commitments can be to whatever considerations are relevant to what we have reason to do. A virtuous person is someone whose commitments to beneficence and justice are last of these explicitly addresses the relation between supererogation and the virtues.) For some discussions on whether Kant acknowledges the possibility of supererogatory actions, see Baron (2016) and Hill Jr. (1971). ¹¹ I am here assuming that what we are morally obligated to do is not simply a function of what others can reasonably demand. For a couple of influential accounts of moral obligation that raise questions about this assumption, see Darwall (2009) and Scanlon (1998). Of course, even if the assumption is false, this does not provide any independent support for the intelligibility of actions which are not morally required and yet are morally favored.
112
commitments to responding appropriately, or reasonably. This means that some dispositions to help those in need, or to promote justice, which do not cause any harm or violate any rights are not constitutive of beneficence or justice. Of course, this does not mean that they are incompatible with beneficence or justice. The point is simply that if they are compatible, this is the full extent of their moral significance. If one concludes that a given action is morally permissible, then one cannot coherently insist that the moral considerations favor acting otherwise. Having explained how this follows from the role that coherence constraints play in our moral ideals, I will then consider whether at least some people may nonetheless be justified in interpreting the moral significance of their actions in a way that cannot be reconciled with their own moral commitments. My critique of the possibility of supererogation begins with the observation that though our substantive ideals impose normative constraints on our ability to realize the ideal of coherence, this does not prevent the ideal of coherence from making a very significant contribution to their content (substance). Short of the nonnegotiable constraints on reinterpreting our substantive ideals without which the inputs to our reasoning would have no normative significance as inputs, each of these ideals is sensitive to other, independent, normative assessments. After all, we think that living up to them is a way of being good/a good way of being. In acting under the guise of substantive ideals, we act under the guise of the good. Accordingly, what counts as living up to them under any given circumstance depends, in large part, on what acting (and feeling) appropriately involves under this circumstance. (This, again, is the aspect of our ideals that philosophers highlight when they speak of the “unity of the virtues.”) In short, though the determinacy of our ideals (and of the virtues, in particular) imposes a constraint on this unity, this is compatible with the fact that each substantive ideal is responsive to the regulative ideal of being integrated into a coherent set of ideals, and thus owes much of its determinacy to its relation to the other ideals in this set. Insofar as the dispositions associated with each of our ideals (the dispositions to see certain reasons and requirements and to respond accordingly) depend for their identity on their relation to our other dispositions, each of them is under pressure to merge into a single exquisitely sensitive disposition to feel, desire, and act appropriately under all circumstances. This means that when someone has internalized a given substantive ideal (e.g., the ideal of motherhood), the significance she attributes to any given fact (e.g., the fact that the person who is asking for an extra bedtime story is her child) partly depends on the significance this fact has, given her other
113
fundamental commitments (e.g, her commitments to being a good teacher, friend, neighbor, etc.). In short, there is—as holists about reasons have observed—an important sense in which we must take all relevant values into account when we are assessing what it takes to realize any given value under any given circumstance. Otherwise put, what it takes to live up to any given ideal depends in large part on what it takes to live up to the others. When one’s child asks one for an extra bedtime story, one must not only “weigh” this fact against other facts, like the fact that one has a pile of papers to grade. One must also rely on one’s other ideals to determine what a good mother would do. From this it follows that within the limits of its nonnegotiable determinacy, each ideal has no further independent content such that some ways of realizing it under these circumstances are more perfect realizations than others, from the point of view of this very ideal. Within the limits that determine what is not compatible with realizing a given ideal, what counts as realizing it—what a good X would do under the circumstances—depends on the relations among all the relevant considerations—i.e., all the relevant nonnormative facts and all the norms that are relevant to assessing these facts. Having determined that helping someone in this way, under these circumstances, would be incompatible with the demands of justice, one concludes that—given that one is not required to act this way, under these circumstances, in order to be beneficent—acting this way, under these circumstances, would not be beneficent, even though, again, it would clearly be helpful. So, too, having determined that offering help in this way, under these circumstances, would fall short of one or more of one’s other ideals (having determined, for example, that it would be incompatible with the requirements of motherhood), one concludes that, given that one is not required to act this way in order to be beneficent, helping in this way would not be appropriate, and so would not be beneficent. If there is nothing beneficence requires under the circumstances, then what a beneficent person would do depends on what a just person would do, and on whether the beneficent person is also a mother, teacher, friend, etc. Accordingly, under these circumstances, the significance of helping, given all one’s ideals, just is the significance of helping from the point of view of the ideal of beneficence itself. To be sure, if someone has internalized this ideal, then she is likely to regard as presumptively significant the fact that an action could help someone. But insofar as her ideals reflect her assumptions about what is good (and insofar as they are thus responsive to reasons), her ideal of beneficence is the ideal of responding to every presumptively significant consideration
114
appropriately. Again, insofar as one’s action does not violate the limits determined by the nonnegotiable descriptive content of an ideal, whether it is an appropriate response to one’s circumstances depends on how this ideal relates to the others. This, again, is what matters from the point of view of the ideal. The preceding reflections are rather abstract. Let us work our way back to the ideal of moral agency by first considering a more substantive ideal: the ideal of motherhood. Let us consider a version of this ideal according to which a mother fails to do what a mother is obligated to do if she never reads to her children, despite being a competent reader. Someone who has internalized this ideal would reasonably assume that this obligation is not the obligation to read a determinate number of books each day: a good mother (with the requisite degree of literacy) might read her children two books before bedtime, or three—or one. If this were how a mother understood the obligations associated with her role, she would not think that if she were to read her children three bedtime stories, she would be a better mother (better qua mother) than someone who reads her children only one bedtime story. Of course, there could be some respect in which the first situation is better for the children than the second. (Perhaps they will acquire a larger vocabulary.) The point is that in bringing about this better state of affairs, a parent would not be more perfectly manifesting the virtues of motherhood—even if, again, reading to one’s children is among the factors relevant to whether one is a good mother. Notice that these remarks do not rule out the possibility that someone could have a more fine-grained (more determinate) ideal of motherhood, according to which when there are no competing requirements, a better mother would read the extra book. Importantly, however, this judgment makes no sense unless it is the judgment that in such circumstances, the ideal of motherhood trumps one’s other ideals—of teacherhood, wifehood, daughterhood, citizenship, etc. If the judgment could not be so construed, then one’s commitment to this ideal would have no implication for what an even slightly better mother would do under these circumstances. Again, what a better mother would do under certain circumstances depends on whether a good mother would recognize that she has reasons at least as good not to read an extra story. Insofar as one believes that a better mother would read the extra bedtime story under certain circumstances, one believes that a mother who does not read the extra story under these circumstances would not live up to her commitments as well as she has overriding reason to do; i.e., one believes
115
that under these circumstances, a good mother would not recognize any contrary considerations as equally compelling; i.e., one believes that under these circumstances, refraining from reading an extra bedtime story cannot be justified (since if it could be justified, a good mother would not be a better, more exemplary, mother for thinking otherwise); i.e., one believes that if one is a mother oneself and does not read the extra bedtime story under these circumstances, then one can be criticized from the point of view of one’s own ideal, even if other people would not be justified in criticizing one for falling short of this ideal in this way. Something similar can be said of any other ideal. Consider, for example, the ideal of prudence. Even if prudence does not require me to do something—e.g., brush my teeth three times a day—I may have one or more reasons to do this. If I brush my teeth three times daily for these reasons, this need not be because I think I would be violating the requirements of prudence were I to brush less frequently. Indeed, the lesson of the preceding discussion is that, regardless of whether we characterize the considerations that move the super-dedicated tooth brusher as “decisive” or “compelling” (as opposed to merely “favoring”), they are not distinctively prudential considerations.¹² According to the stipulated circumstances, the only prudential significance of these considerations is that being moved by them is compatible with acting prudently. The same point applies to the ideal of treating people “with concern and respect.” And this brings me back to my critique of supererogation. Helping does not have a given moral significance in itself. Sometimes we are morally obligated to help. Sometimes we are morally obligated not to help. Sometimes we are morally permitted to help, though we are not morally required to do so. Again, in order to make the relevant determinations, we must rely on a conception of treating people with concern and respect that partly relies on our conception of what is required in order to live up to our other ideals. For my purposes here, I do not need to identify the nonnegotiable limits of what is involved in “treating people with concern and respect.” (Perhaps nothing more falls within these limits than the requirement to treat people in ways they cannot reasonably reject. Perhaps the ideal has a more substantive nonnegotiable content.) The important point is that within these limits, one’s assessment of what is morally permissible just is one’s assessment of the moral significance of helping; and this is because, within these ¹² For a defense of supererogation that appeals to the different “roles” reasons can play, see Horgan and Timmons (2010). See also Gert (2012).
116
limits, the moral significance of helping just is the moral significance of helping, given one’s other ideals. If we determine that we are morally permitted—but not morally required—to help someone by doing X, this exhausts our moral assessment of the circumstances and how we might respond to them. Again, this is no less true if—as is surely the case—whether an action would be helpful is relevant to determining whether in failing to act this way one would be failing to live up to the ideal of treating people with concern and respect. For again, one has already determined the moral significance of helping in determining how helping-by-doing-X-under-these-circumstances relates to one’s other ideals (and to the permissions and requirements associated with these ideals). Just as there is some respect in which it is better to read one’s child more bedtime stories, even though doing so does not make one a better mother, and just as there is some respect in which it is better to brush one’s teeth more than two times a day, even though this does not make one more prudent, so too, there is some respect in which helping a person is better than not helping him, even though this does not make one morally better. One is not a more perfect exemplar of the relevant ideal in each of these cases (the particular actions in these cases do not themselves more fully manifest the relevant ideal) even though one gives a decisive guiding role to the very sort of considerations that are relevant to what a good mother, or a prudent or beneficent person would do. Again, this point does not depend on obscuring the distinction between “justifying,” or “favoring,” or “merit-conferring” reasons, on the one hand, and decisive, or compelling, or “requiring” reasons, on the other. People willingly grant that it is sometimes wrong to do what one is usually permitted, and even obligated, to do. I am simply pointing out that the same considerations apply in the context of our evaluations of permissible actions: these actions need not be better morally just because they have features that are relevant to whether actions are morally permissible, and even obligatory. Just as we can act wrongly in promoting justice, so too, we can act virtuously for this purpose without acting more virtuously than we otherwise might. From the fact that there are circumstances under which an action is required because it is an act of helping someone in need it does not follow that when helping someone lacks the moral weight to ground a moral requirement, it nonetheless has an additional moral weight that renders it more than permissible. To give this fact this extra weight under these circumstances is to be guilty of double counting. Let me be clear: the considerations that count in favor of helping when it is easy to do so do not evaporate when the costs get very high. The point is that they acquire a different significance from the moral point of view. Again,
117
this is what we intuitively acknowledge when we acknowledge that we can know that A would be helping B if she were to do X without thereby knowing whether in doing X she would be doing the morally right thing. It is important not to underestimate the holistic character of such evaluations. It is not just that whether A’s helping B is morally permitted or required depends on other features of the circumstances. The point is that A’s helping B does not have a given moral significance independent of its relation to these other features. To see this, let us assume that, contrary to what I have just been arguing, it is possible to identify certain facts as having a certain moral significance, independent of identifying various other facts that are relevant to what one has reason to do, all things considered. On this assumption, someone could have two different views about the relationship between (a) what someone has most reason to do, all moral things considered, and (b) what she has most reason to do, all things considered: (i) she could think that the former judgments determine the latter (because moral reasons trump all other sorts) or (ii) she could think that she is sometimes justified in giving certain nonmoral reasons greater weight than the moral reasons. If she endorses the first view, then, by her own lights, she cannot justify refraining from engaging in a helpful action if this is what she has most moral reason to do. If, alternatively, she endorses the second view, then, by her own lights, when she chooses to do what she thinks she has most moral reason to do, this choice will not reflect a greater moral virtue than the alternative. Since, by her own lights, she has already identified all the moral reasons, by her own lights, there is no additional moral reason to favor the moral reasons over the nonmoral reasons. If she chooses to do what she takes herself to have more moral reason to do, this will, by her own lights, simply reflect her preference (on certain nonmoral grounds, or on no grounds at all—in this particular circumstance, or in all circumstances, or in all circumstances of this sort) for giving more weight to the considerations she takes to be morally relevant than she gives to whatever countervailing nonmoral considerations she takes there to be. If we share this person’s view about the relationship between moral and nonmoral reasons, then we agree that her choice does not reflect a stronger, more perfect, moral commitment than would be manifest had she chosen not to help; by our own lights, the fact that she assigns greater weight to the moral reasons is not, in itself, a morally relevant fact. Of course, this does not mean that we have no reason to praise her. Indeed, she is almost surely praiseworthy (i) for doing something which had (or was likely to have) certain desirable consequences and/or (ii) for having a disposition to do
118
what she is morally obligated to do, even at great cost to herself. The point is simply that it is no part of either of these assessments that helping under these circumstances is, in itself, a morally superior action.
The desirability of a less-than-fully coherent moral point of view (Take Two) There is nothing perplexing—and, in particular, nothing incoherent—about actions that involve the agent’s making a morally optional sacrifice in order to promote human well-being, or justice, or peace on earth. I have argued, however, that someone would be incoherent if she believed that the reasons favoring an action are reasons to think that acting this way is morally better than the morally permissible alternatives. To believe this would be, in effect, to occupy two incompatible points of view at once: (i) a point of view from which the moral significance of certain considerations is that one is morally permitted to do X or to refrain from doing X and (ii) a point of view from which the moral significance of these considerations is that they speak more strongly in favor of one’s doing X than refraining from doing X. To say that these points of view are incompatible is not merely to say that they are distinct—as would be the case if one of them were constituted by one’s commitment to promoting one’s own interest, or one’s commitment to doing whatever one is most strongly disposed to do in response to someone in need, while the other was constituted by one’s commitment to treating others with concern and respect. Rather, without abandoning the point of view from which the moral significance of another person’s needs depends on far more than the fact that he has them, one would, at the same time, be interpreting a person’s need (or one’s own disposition to help this person) as in itself morally favoring a certain course of action. In so doing, one would not merely mistake (i) the disposition to help someone when this is neither morally required nor morally forbidden for (ii) a moral virtue. One would also be obscuring (and in obscuring, defying) one’s own (hard-earned) understanding of the moral significance of human needs. I am tempted to think that it is only moral theorists who get hung up on categorizing as especially morally admirable someone who helps others when, by her own lights, she is not morally required to do so. Don’t ordinary folks regard such actions as “beyond the call of duty” without insisting that they are referring to what lies morally beyond? Yet further reflection leads me to wonder whether the theorists are channeling a deep fact about human
119
beings—a fact that grounds the incoherence I have just been discussing, and justifies it too. I want to explore this possibility, and what it implies, by considering what we should make of someone who is vulnerable to being incoherent in this way. The person who misunderstands her own stance toward her action when she exceeds the bonds of duty resembles the akratic agent as I have described her in being cloudy eyed about her own inner conflict. Like the akratic agent, her eyes could clear without any alteration in her motivational structure only if her assessment of the reasons for and against her action were limited to the purely theoretical assessment of a spectator who appraises her actions with no expectation of influencing them in any way. In this case, she would avoid a conflict among her commitments by regarding her judgments regarding what she is morally permitted to do as irrelevant to her moral agency. Without thus alienating her agency from her reasoning, the only way for a person to maintain the incoherence I have just described would be to obscure from herself the fact that she is, in effect, double counting certain considerations in response to her inclination to give these considerations more moral weight than they really have by her own lights. The case of akrasia suggests that a person stands in a criticizable relation to her own reason if she relates to her intentions in either of the ways just described. It seems to me, however, that we may well have good reason to cultivate in ourselves and others a disposition either not to notice the reasons we have for rejecting a conception of virtue as confused, or, on certain occasions, to dissociate our normative verdicts from our own agency. As I have noted, someone could be disposed to engage in helpful, justicepromoting/rights-protecting behavior without thinking that this is what she has most moral reason to do. She might simply care about certain things— including, importantly, the well-being of others—which make certain morally permissible helpful actions appear to be called for. Acting this way could be integral to her most important personal projects. Though this is possible, I am nonetheless inclined to think that many human beings would not be moved to make a big, or even a moderate, sacrifice to help strangers in need, or to promote justice in cases in which they are morally permitted not to, without being moved by the thought that this is what an especially virtuous person would do, regardless of what her personal preferences and projects happen to be—or the thought that this is what someone would do if she cared about others in the way that a morally virtuous person does. If this hypothesis is correct, then it seems that in order
120
to avoid incoherence, human beings who were disposed to make significant sacrifices to promote and protect the interests and rights of others would have to be convinced that morality is very, very demanding. They would have to assume that a morally virtuous person (i.e., someone who is just and beneficent and has whatever other traits of character are constitutive of moral virtue) would regard opportunities to promote and protect the interests and rights of others as requiring such a response, even when this involved making a very significant sacrifice. This is not the place to debate the plausibility of such a conception of moral requirements. I would like, instead, to suggest that if endorsing this conception is the only way for someone to avoid incoherence while taking herself to have overriding moral reason to behave in the way I have mentioned, then this fact may be a reason to endorse the relevant form of incoherence. We need to be open to the possibility that it is a good thing if at least some people are disposed to rightly believe that no one is morally required to help others in certain ways, even while they also mistakenly believe that the reasons to help in these ways are the reasons that someone with greater moral virtue would take to favor helping, all things considered. Perhaps there is something good about the sort of double standard that is tightly connected to this incoherent stance: it may well be desirable for most, if not all, of us to take ourselves to have all-moral-thingsconsidered reasons to bring about certain desirable results, even though we do not believe others would be morally mistaken to attribute less significance to these reasons, and even though we do not take there to be anything special about us that would justify this distinction. It may be that making an exception of oneself in this way is itself a distinct sort of virtue.¹³ Our discussion of weakness of will shows us how it is possible for someone to maintain the sort of incoherent stance I have been discussing here: she could simply avoid considering whether it is possible to reconcile (i) the assumption that the moral considerations favor doing X now with (ii) her other commitments. In thus managing to live with her incoherence, she would rely on the fact that any inquiry into what justifies her actions must stop somewhere—usually shortly after it starts. Postponing further inquiry into the adequacy of one’s own reasoning is the stuff of a human life. Only rarely do we follow our deliberations to the ¹³ In another recent paper I have explored a different sort of case in which this sort of double standard may be morally commendable. See Buss (2020b).
121
place where they force us to confront the sort of worries that trouble philosophers: “But what, ultimately, justifies my assumption that this fact counts as a reason? What sort of consideration could possibly do the trick?” For the most part, we take it for granted that we have sufficient reason to do what we are moved to take ourselves to have sufficient reason to do. Only very rarely outside the philosophy study do we bother to inquire into the credentials of such dispositions. Reflecting on weakness of will suggests that we can be criticized for exploiting our ability to do things without paying attention to the credentials of our reasons for doing them. But reflecting on supererogation suggests that under certain circumstances, a motivated failure to keep track of one’s own reasons need not be a failing. Indeed, the disposition to sustain a less-than-fully coherent position in this way can be a practical virtue—even a moral virtue. This suggestion brings me back to the case of the sophisticated consequentialist. This theorist suggests that human agents are more likely to do the right thing if they misunderstand what distinguishes right from wrong. She thus suggests that she herself is more likely to do the right thing if she ignores or forgets the implications of her own moral commitments. Similarly, I have just suggested that human beings are more likely to bring about certain desirable results (including morally desirable results) if they misunderstand the implications of their own moral commitments. Whereas according to the sophisticated consequentialist, moral confusion is an essential feature of moral agency, according to the sophisticated deontologist, moral confusion is an essential feature of certain actions that are admirable given the moral and nonmoral value of their (likely) effects. Given this point of agreement, we cannot favor either deontology or consequentialism on the ground that the other sort of theory is committed to the desirability of this sort of incoherence. We thus seem to have another reason to question the assumption that there is a straightforward relation between whether a moral theory is adequate and whether it tolerates incoherence.
Conclusion There are, it seems, at least two connections between our vulnerability to incoherence and the fact that our capacity to reason depends on prerational (natural) dispositions to regard facts as reasons and requirements:
122
(i) Because certain prerational dispositions to evaluate our circumstances in certain ways and to respond accordingly are a necessary condition for the development of the virtues, there can be no fully coherent conception of what we are morally required to do. (ii) Because whenever we form an intention, we have no choice but to rely on certain prerational dispositions to evaluate our circumstances in certain ways, it is possible (and, eventually, necessary) for us to ignore questions we might intelligibly raise about whether our normative assumptions can really be justified. In discussing the first connection in the first section of the paper, I suggested that incoherence may be an essential feature of an adequate moral theory—even if coherence is also a fundamental regulative ideal. In bringing this essay to a close, I want to make a few final, even more tentative, remarks about the challenges posed by the second constraint on the relation between moral agency and pure reason. To repeat: the incoherence that results from not probing one’s normative assumptions is sometimes a good thing. Sometimes, not noticing a problematic feature of one’s attitudes can play a key role in one’s disposition to sacrifice one’s own well-being (or be willing to do so) in order to bring about very desirable results. As the case of weakness of will indicates, however, it is far from clear that we can draw any general conclusions from such cases of virtuous incoherence. Among the many complicating factors, the fact that incoherence of the relevant sort plays a valuable and/or essential role in our moral agency provides us with additional opportunities for going astray: it is very tempting to treat this fact as a reason to engage in sloppy reasoning, or a reason not to bother probing our assumptions about what we have sufficient reason to do. Indeed, this is one of the (often unconscious) enabling rationales in many cases of weakness of will: “Given that my ability to make sense of the normative landscape is so limited, given that my normative assumptions do not form a coherent whole, given that gains in coherence need not be gains in my ability to appreciate what I have reason to do, and given, in any case, that I do not have much time to weigh most of the pros against most of the cons, why should I worry about whether I can find an adequate justification for doing what I am here and now inclined to do? Why should I worry whether I am really justified in acting this way by my own lights? I might as well scratch the itch now. (And isn’t this just what everyone does most of the time anyway?)”
123
One way to prevent lazy reasoning from dominating our individual and collective decision-making is to point out the incoherence it covers up. But if my reflections here are correct, this is not only unlikely to be an effective strategy; its ineffectiveness will reflect the fact that incoherence is a necessary and, in some of its manifestations, a desirable aspect of human life. There are days when it seems to me that the most important problems our country now faces can all be traced to the widespread suspicion of reason and reasoning—that these problems are symptoms of the view that all reasons are self-serving rationalizations and that public discourse is ultimately an expression of the will to power. In such moments, it seems to me that more of us who engage in normative inquiry for a living should be trying to find better ways to justify our faith in our attempts to root out the incoherence in our commitments. But if what I have said here is correct, then this defense of reason must, as Kant says, take the form of a critique of reason: in encouraging people to achieve greater coherence in their views about what is desirable and appropriate, and about what they are called upon to do in response to the circumstances of their personal and social lives, we must acknowledge the reasons why full coherence is not possible— and not desirable either. If this critique of reason’s drive to greater coherence is to be, at the same time, a defense of this drive, we must have a better story to tell about what the limits are, and what distinguishes the desirable from the undesirable cases. To this end, we need to explore the extent to which what appear to be cases of desirable incoherence are really just cases of irreducible plurality, and the extent to which what appear to be cases of irreducible plurality are really sufficient conditions for (the possibility of ) a deeper incoherence. We need to supplement this conceptual inquiry with an inquiry into the normative significance of the types of incoherence we value and of the reasons we value incoherence when it takes these particular forms.¹⁴
References Aristotle, 1941, Nicomachean Ethics. Ross, W.D. (Trans.). In McKeon R. (Ed.), The Basic Works of Aristotle, New York: Random House: 927–1112. ¹⁴ I would like to thank Roger Crisp, Elijah Milligram, Sarah Paul, Francois Schroeder, Mark Timmons, attendees of the 2022 Workshop in Normative Ethics, and two anonymous referees for comments on earlier drafts of this paper. I would also like to thank Mark for his helpful advice and thoughtful responses to my inquiries throughout the writing and editing process.
124
Baron, Marcia, 2016, “A Kantian Take on the Supererogatory,” Journal of Applied Philosophy 33(4): 347–62. Buss, Sarah, 1997, “Weakness of Will,” Pacific Philosophical Quarterly 78(1): 13–41. Buss, Sarah, 2020a, “Some Reflection on the Relation Between Reason and the Will,” in Routledge Handbook of Practical Reason, Ruth Chang and Kurt Sylvan (ed.), New York: Routledge: 196–213. Buss, Sarah, 2020b, “Some Musings About the Limits of an Ethics that Can Be Applied,” Journal of Applied Philosophy 37(1): 1–33. Buss, Sarah, unpublished manuscript, “Norms of Rationality and the Superficial Unity of the Mind.” Crisp, Roger, 2013, “Supererogation and Virtue,” in Oxford Studies in Normative Ethics (vol. 3), M. Timmons (ed.), Oxford: Oxford University Press: 1–24. Darwall, Stephen, 2009, The Second-Person Standpoint: Morality, Respect, and Accountability, Cambridge, Mass: Harvard University Press. Davidson, Donald, 1980, “How is Weakness of the Will Possible?” in Essays on Actions and Events, Oxford: Oxford University Press. Frankfurt, Harry, 1988, “The Importance of What We Care About,” in The Importance of What We Care About, Cambridge: Cambridge University Press: 80–94. Gert, Joshua, 2012, “Moral Worth, Supererogation, and the Justifying/Requiring Distinction,” Philosophical Review 121: 611–18. Harman, Elizabeth, 2016, “Morally Permissible Moral Mistakes,” Ethics 126: 165–85. Herman, Barbara, 1993, “Obligation and Performance,” in The Practice of Moral Judgment, Cambridge, Mass: Harvard University Press: 169–83. Heyd, David, 2015, “Can Virtue Ethics Account for Supererogation?” in Supererogation, Royal Institute of Philosophy Supplement (vol. 77), C. Cowley (ed.), Cambridge: Cambridge University Press: 25–47. Hill Jr., Thomas E., 1971, “Kant on Imperfect Duty and Supererogation,” KantStudien 62(1–4): 55–76. Horgan, Terry and Timmons, Mark, 2010, “Untying a Knot from the Inside Out: Reflections on the ‘Paradox’ of Supererogation,” Social Philosophy and Policy 27(2): 29–63. Kagan, Shelly, 2001, “Thinking about Cases,” Social Philosophy and Policy 18(2): 44–63. Lear, Jonathan, 2002, “Happiness,” in Happiness, Death, and the Remainder of Life, Cambridge, MA: Harvard University Press: 1–60.
125
Marcus, Ruth, 1980, “Moral Dilemmas and Consistency,” Journal of Philosophy 77: 121–36. Portmore, Douglas, 2008, “Are Moral Reasons Morally Overriding?” Ethical Theory and Practice 11(4): 369–88. Scanlon, T.M., 1998, What We Owe to Each Other, Cambridge, Mass: Harvard University Press. Sidgwick, Henry, 1907, The Methods of Ethics, London: Macmillan. Urmson, J.O., 1958, “Saints and Heroes,” in Essays in Moral Philosophy, A.I. Melden (ed.), Seattle and London: University of Washington Press: 198–216.
6 A Contractualist Approach to Moral Uncertainty Michael Bukoski
People are often uncertain about what they morally ought to do. This uncertainty results not only from uncertainty about morally relevant nonmoral facts but also from uncertainty about fundamental moral facts or principles, which determine what one morally ought to do in light of the non-moral facts. It is fairly widely agreed that factual uncertainty affects what one morally ought to do, at least in a “subjective” as opposed to “objective” sense of “ought”.¹ Does (fundamental) moral uncertainty also affect what one morally ought to do? Two positions dominate debate. One posits special moral uncertainty principles, unaffiliated with any particular moral theory, to govern action under moral uncertainty.² A common candidate is the principle of maximizing expected choiceworthiness (MEC), analogous to the familiar principle of maximizing expected utility in response to factual uncertainty. The other claims that moral uncertainty does not affect what one morally ought to do.³ If an action is otherwise permissible, it remains permissible even if the agent believes that it is or may be wrong; and if an action is otherwise wrong, it remains wrong even if they believe that it is or may be permissible. In this paper, I have two main aims. The first aim is to call attention to a neglected theoretical option: it may be that an agent’s moral uncertainty does affect what they morally ought to do, but that the relevant moral uncertainty principles are simply components of the true first-order moral theory, justified alongside and in the same way as the other principles making up that theory. This approach can capture some of what is attractive
¹ See, e.g., Zimmerman (2008). ² Early discussions include Lockhart (2000), Ross (2006), and Sepielli (2009). ³ See, e.g., Harman (2015) and Weatherson (2014, 2019). Michael Bukoski, A Contractualist Approach to Moral Uncertainty In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Michael Bukoski 2023. DOI: 10.1093/oso/9780198895909.003.0007
127
about the idea that moral uncertainty affects what one morally ought to do, while avoiding some of the objections to views that make moral uncertainty principles completely independent of first-order morality. The second aim is to explore this possibility in the context of a particular moral theory by developing a contractualist approach to moral uncertainty. These two aims are somewhat independent. Contractualists and those sympathetic to contractualism will find a contractualist approach to moral uncertainty interesting in its own right, as the extension of a compelling moral theory to a neglected topic. For others, the contractualist account will serve mainly to illustrate the plausibility of the more general approach to moral uncertainty that I describe. Of course, though I shall not attempt to defend contractualism against alternatives, a plausible account of the moral implications of moral uncertainty is a point in its favor, at least until other moral theories have been shown to provide at least equally plausible accounts. I proceed as follows. First, I introduce the issue of the moral significance of moral uncertainty and identify the kind of moral uncertainty principle that I want to defend in this paper. I distinguish between global principles, which are independent of and impartial between first-order moral theories such as contractualism and utilitarianism, and local principles, which are components of particular moral theories. Second, I argue that contractualist reasoning justifies a local moral uncertainty principle that prescribes a moderate degree of moral caution, neither prohibiting all morally risky action nor licensing a reckless disregard for moral risk. Third, I argue that the best contractualist moral uncertainty principle is discriminating rather than content-neutral, in the sense that its prescriptions differ depending on the content of the moral views in which a morally uncertain agent has credence, and I describe two ways in which it is discriminating. Finally, I show how the contractualist principle (and by extension other local moral uncertainty principles) can avoid, or at least mitigate, some important objections to global moral uncertainty principles such as MEC.
1. Moral uncertainty and its implications A number of philosophers have argued that moral uncertainty affects what one ought, in some sense, to do.⁴ A few distinctions will help to locate the
⁴ For a recent overview, see MacAskill et al. (2020).
128
kind of moral uncertainty principle that I want to defend. To begin, it seems clear that moral uncertainty can affect what an agent rationally ought to do in the instrumental sense of rationality, on which rationality is a matter of effectively pursuing one’s ends given one’s information. The end of doing what is morally right is no different from this perspective than any other end an agent might have, such as making money or achieving health or happiness. Suppose that someone believes that people probably have limited duties of aid to distant strangers but also has some credence in the view that people are morally required to donate most of their disposable income to charity and that failing to donate an amount that would save a life is morally on a par with failing to rescue a drowning child to avoid an equivalent cost, which is very morally wrong (Singer 1972). Depending on the value they place on doing what is right (or more closely approximating what is right) vis-à-vis their other ends, it may be instrumentally rational for them to donate more money than they otherwise would, or even to donate most of their disposable income, just in case it turns out to be morally required.⁵ Whether moral uncertainty affects what one ought to do in other senses of “ought” is a more difficult question. I will focus on whether an agent’s moral uncertainty affects what they morally ought to do. There is a familiar distinction between what one objectively ought to do and what one subjectively ought to do. Objective principles describe what one ought to do given the moral truth and the relevant non-moral facts. Subjective principles describe what one ought to do given one’s beliefs (or perhaps one’s evidence) about (at least) the non-moral facts. Whether an agent’s beliefs about fundamental moral facts or principles affect what they subjectively ought to do is more controversial. If they do—a view sometimes called moral uncertaintism—then it would likely be subjectively wrong for the person just described not to give additional money to charity. The details depend on which principle correctly describes what one subjectively ought to do in situations of moral uncertainty. A common candidate is the principle of maximizing expected choiceworthiness (MEC).⁶ Here, “choiceworthiness” (or “moral value”) is a measure of how much is at stake, morally speaking, in a given option, taking into account all the moral reasons for or against an ⁵ The same reasoning applies if the agent has more specific moral ends: not to do what is right as such, but to do what is just, or fair. It is a separate question whether one morally ought to have any of these ends. Smith (1994) famously describes the desire to do what is morally right de dicto as objectionably fetishistic; see also Weatherson (2014). ⁶ See MacAskill and Ord (2020) for a recent defense of MEC.
129
agent’s choosing that option. MEC prescribes choosing the option for which the sum of its choiceworthiness according to each moral theory weighted by one’s credence in that theory is greatest. It has significant practical implications. For example, if one believes that eating meat, or having an abortion, might be very morally wrong, whereas the alternatives are almost certainly permissible, then MEC may prescribe refraining from both even if one thinks them likely permissible.⁷ At this point, I want to distinguish between what I will call global and local moral uncertainty principles. Global moral uncertainty principles are not part of any particular first-order moral theory, such as contractualism or utilitarianism. They are justified on grounds that do not beg the question in favor of any particular moral theory. Philosophical discussions of moral uncertainty typically assume that the relevant moral uncertainty principles are global principles. This is evident in how the arguments offered in defense of moral uncertainty principles are typically not substantive moral arguments but instead are neutral between different moral theories. For example, one important argument is that moral uncertainty principles are necessary to provide adequate moral guidance for morally uncertain agents.⁸ Another is that moral uncertainty should affect what one ought to do in a way that is symmetrical with how factual uncertainty affects what one ought to do.⁹ By contrast, local moral uncertainty principles are part of particular moral theories and describe what, according to that theory, someone in a situation of moral uncertainty (subjectively) ought to do, just as other moral principles included in the theory describe what one ought to do in other circumstances. Local moral uncertainty principles can be justified by the same kind of substantive moral arguments as the other moral principles included in a moral theory. For example, a contractualist moral uncertainty principle can be justified using distinctively contractualist arguments, even though proponents of other moral theories, such as utilitarianism, would likely reject these arguments. Later, I will argue that a contractualist approach to moral uncertainty, and by extension many other local approaches to moral uncertainty, can avoid several important objections to global moral uncertainty principles. Here ⁷ See Moller (2011) on abortion and Guerrero (2007) on vegetarianism. MacAskill (2019) surveys a variety of putative practical implications of MEC in conditions of moral uncertainty and discusses complications resulting from moral uncertainty about multiple issues simultaneously. ⁸ E.g., Lockhart (2000: 8–9), Sepielli (2009: 8), MacAskill and Ord (2020: 330–2). ⁹ Weatherson (2019: 39–40) observes that this assumption is in the background of much discussion of moral uncertainty.
130
I want to forestall one concern. Moral principles are commonly thought to have two main roles or functions: to provide a standard of rightness by describing the moral status of actions (as required, permissible, prohibited, etc.) and to provide action-guidance by equipping people with a standard or procedure they can apply in order to act rightly. It may be thought that global moral uncertainty principles can provide effective action-guidance for morally uncertain agents because they do not beg the question in favor of any particular moral theory about which an agent might have doubts, but that local principles fail by this metric, since anyone who doubts the moral theory of which a particular local principle is a part will perforce also doubt whether that principle describes the correct response to moral uncertainty. Someone uncertain about the truth of contractualism would obviously have qualms about following contractualism’s moral uncertainty principles! But matters are not so simple. Moral uncertainty principles first of all provide a standard of (subjective) rightness, so a failure to provide effective action-guidance of this sort is not a reason to reject a local principle provided an alternative justification for it is available. People who doubt the truth of a moral theory will have qualms about following any of its principles; no one thinks that shows them all to be false. Second, local principles do provide effective action-guidance to some people, namely those who accept the moral theory in question but remain uncertain about its more detailed content or how to apply it to a particular case. Third, it is not clear that global moral uncertainty principles provide action-guidance of the sort in question, since someone may doubt whether any particular global principle, such as MEC, describes the correct response to moral uncertainty. Indeed, someone who has credence in a local moral uncertainty principle will ipso facto have doubts about the truth of global principles. If one thinks contractualism may be true, and that if so, one ought to act in situations of moral uncertainty in the way I will describe, then one will ipso facto doubt whether one ought to follow MEC or any other global principle when its implications diverge from what contractualism prescribes. Global principles are thus no more action-guiding than local principles. Finally, even if global moral uncertainty principles do ultimately determine what one ought to do in situations of moral uncertainty, local principles would remain important because they affect what one morally ought to do according to global principles. Failing to account for a moral theory’s own prescriptions about how to act in situations of moral uncertainty will result in a mistaken conception of the competing moral views that global principles such as MEC aim to take into account.
131
2. A contractualist approach to moral uncertainty principles Contractualist theories belong to the broader family of social contract theories, which characterize moral principles as the principles to which people would agree under specified hypothetical circumstances. In contractarian theories (e.g., Gauthier 1986), the contracting parties are motivated to satisfy their preferences, whatever they are; the basic thought is that by accepting some principles of mutual restraint and assistance, each person can fare better by their own lights. In contractualist theories, people are assumed to have specifically moral motives, such as the desire to treat others in ways that they could not reasonably reject if they were similarly motivated in finding terms for regulating their common life (Scanlon 1998). I will focus on Scanlon’s (1998) well-known contractualist account, but the main points generalize to other social contract theories. Scanlon (1998: 153) proposes a master principle that provides a general criterion for correct moral principles: An act is wrong if its performance under the circumstances would be disallowed by any set of principles for the general regulation of behavior that no one could reasonably reject as a basis for informed, unforced, general agreement.
Actions are right if they are not wrong. We apply the criterion roughly as follows. Given a candidate principle, some people may have reason to reject it. Relevant reasons must be “personal” as opposed to “impersonal,” but personal reasons are not limited to considerations of the objector’s wellbeing (Scanlon 1998: 218–23); they can include, for example, whether a candidate principle treats them unfairly by arbitrarily privileging others (1998: 212, 216). We consider objections to each candidate principle from each relevant standpoint without regard for the number of people who occupy that standpoint. A principle can reasonably be rejected if an alternative principle is such that the strongest objection to it is weaker than the strongest objection to the principle under consideration. In other words, the correct moral principle on some matter is the one to which the strongest objection is weakest. Though it may impose greater burdens on someone than an alternative principle would, we can justify the former principle to them by noting that, under the latter, someone else would face even greater burdens.
132
To illustrate, take Scanlon’s (1998: 235) example of Jones, who has become trapped under some electrical equipment in the television station broadcasting a World Cup match. Turning off the equipment to rescue him from the extremely painful electric shocks would mean disappointing millions of fans watching the game. Is it morally right to rescue him? We apply the reasonable-rejection criterion by weighing Jones’ objection to a principle permitting postponement of the rescue until after the game (he would suffer extreme pain in the meantime) against a fan’s objection to a principle requiring immediate rescue (it would disrupt their enjoyment of the game). The former objection seems clearly stronger, so Scanlon concludes that the principle requiring that Jones be rescued immediately cannot be reasonably rejected. We ought to rescue Jones even if we disappoint the fans. For another example, consider activities such as passenger air travel that are beneficial to the agent but impose risks on others. Ashford (2003) argues that Scanlon’s contractualism leads to the conclusion that air travel is impermissible. The would-be passengers can object to a principle forbidding air travel by pointing out the absence or inconvenience of alternatives, while the would-be victims of plane crashes can object to a principle permitting air travel that it will cost them their lives. The latter objection seems clearly stronger. However, Kumar (2015) defends contractualism against this apparent counterexample by arguing that the relevant objection comes from the ex ante rather than ex post standpoint of the victims—that is, the standpoint of someone on whom the very small risk of a plane crash is imposed, not someone for whom that risk is known to eventuate. Because the benefit is substantial and the risk is very slight, Kumar (2015: 49) concludes that a principle permitting air travel subject to appropriate safety precautions and provision of compensation to victims or their families could not be reasonably rejected.¹⁰ Scanlon’s reasonable-rejection criterion provides a tidy basis for justifying local moral uncertainty principles within contractualism. For we can ask: Are moral uncertainty principles among the principles that people could not reasonably reject? Note first that even committed contractualists may be uncertain about what they morally ought to do in particular contexts or ¹⁰ Whether the morally relevant standpoint in cases of risky action is the ex ante or ex post standpoint (or some combination) is a matter of debate among contractualists and their critics (see also, e.g., Lenman 2008, Fried 2012, Frick 2015, Rüger 2018). Scanlon (1998: 207, 236–7) originally endorsed the ex post approach but later (2013) endorsed the ex ante approach. Here I focus on local moral uncertainty principles in the context of ex ante contractualism. Ex post contractualism is a different moral theory, with different implications regarding moral uncertainty.
133
cases, for they may not know which principles could not be reasonably rejected or how best to apply a principle to a particular case. Even among professional philosophers, there is disagreement and debate about contractualism’s implications for a variety of issues, not only risky activities but also, for example, the nature and extent of our duties of aid (Kumar 1999, Ashford 2003). Scanlon himself (1998: 361) observes that although we should “strive to find terms of justification that others could not reasonably reject . . . we are not in a position to say, once and for all, what those terms should be. Working out the terms of moral justification is an unending task.” If philosophers are uncertain, ordinary people trying to apply contractualist principles in their everyday interactions should be still more so. Moreover, people recognize that their moral uncertainty is not idiosyncratic; others will be uncertain in similar ways. Second, some ways of dealing with moral uncertainty are morally better than others, as measured by the strongest objections to which principles permitting them are subject. Two standpoints are relevant: (1) the agent considering whether to perform some specified action, about the moral permissibility of which they are uncertain, and (2) the potential victim should it turn out that the action is morally wrong. Now consider four candidate principles describing what the agent ought to do in this situation: No-principle option: One’s moral uncertainty has no effect on what one ought to do. Conservative principle: Whenever one believes, to any degree (or perhaps any degree above a very low threshold), that an action might be wrong, one ought to refrain from performing that action. Permissive principle: Only when one is certain (or perhaps nearly certain) that an action would be wrong is one required to avoid it. If one has any (significant) positive credence in the view that the action is not wrong, then it is permissible. Moderate principle: All else equal, the higher one’s credence in the view that an action would be wrong, the more reason one has to refrain from that action, and the more likely it is that one ought to refrain. The no-principle option rejects the idea that in light of one’s moral uncertainty one subjectively ought to act in a way that potentially diverges from how one objectively ought to act. The conservative, permissive, and moderate principles accept that idea, but disagree about how one’s moral credences affect what one subjectively ought to do. Note that the moderate principle
134
is schematic; different versions of the principle fill out the details in different ways. Which principle does the reasonable-rejection criterion favor? From the standpoint of potential victims of inadvertent wrongdoing, the permissive principle is objectionable because it renders the protections accorded by other contractualist moral principles very fragile. For example, Kumar (2015: 35–9) argues that a principle permitting lethal medical experimentation on non-consenting people could be reasonably rejected even if the resulting discoveries would increase total happiness. A victim’s objection to the harms inflicted upon them and the violation of their bodily autonomy is stronger than the competing objection of a beneficiary of the experiments. But a principle forbidding nonconsensual medical experimentation would not adequately protect people from that threat if the experiments were subjectively permissible so long as the experimenters sincerely believed they might be permissible—for example, because they erroneously suspected that utilitarianism could not be reasonably rejected, or that the low odds of being selected for experimentation so weaken the objection from the ex ante standpoint of potential victims that the objection to the prohibitive principle from the standpoint of potential beneficiaries of experimentation is stronger (as in the air travel case). Thus, from the patient’s standpoint, the objection to the permissive principle is that it will very likely and very often permit agents to perform objectively wrong actions, thereby threatening the legitimate interests of the patient. By contrast, the conservative principle provides maximum protection to the interests of potential victims of action under moral uncertainty. However, the conservative principle is highly objectionable from the standpoint of morally uncertain agents. Given the pervasiveness of at least minor moral uncertainty, it would set back their own legitimate interests by requiring them to refrain from many activities that are objectively permissible and that they have genuine reason to pursue. For example, someone may believe (I will assume, correctly) that a principle permitting air travel for leisure purposes cannot be reasonably rejected, but retain some uncertainty about this conclusion (does a fondness for foreign travel really justify imposing even a very small risk of death on nonconsenting strangers?), in which case the conservative principle prohibits their traveling. Multiplied across contexts, the conservative principle severely restricts the scope and pursuit of people’s projects, and this result provides a strong objection against it. The moderate moral uncertainty principle fares better. Because it gives significant weight to the interests of both agents and potential victims in
135
situations of moral uncertainty, the objection to the moderate principle from each standpoint is weaker than the objections from the same standpoint to the conservative and permissive principles, respectively. Thus, the reasonable-rejection criterion favors the moderate principle over these alternatives. But what about the no-principle option? From the agent’s standpoint, it may be slightly preferable to the permissive principle because it slightly less often prohibits agents from engaging in objectively morally permissible actions, since it permits them even when the agent believes they are certainly or nearly certainly wrong. But it is highly objectionable from the patient’s standpoint. The problem is that morally uncertain agents cannot simply act on the correct moral principles under that description, but must choose what to do given their uncertainty about whether the action in question is morally wrong. Because the no-principle option does not prescribe how to make that choice, it fails to prescribe acting in a conservative or moderate way as opposed to a permissive way. But we have seen that from the patient’s standpoint, there is a strong objection to allowing people to deal with moral uncertainty in the way prescribed by the permissive principle. Because the no-principle option fails to guide agents away from those same actions (to which, given their interests, they will predictably be drawn), patients have a strong objection to that option—stronger than their objection to the moderate principle, which provides their interests with more effective protection against the vicissitudes of morally uncertain others. So, the reasonable-rejection criterion favors the moderate principle over the no-principle option.¹¹ One might think there is another candidate principle: the very principle describing what the agent objectively ought to do in their circumstances.¹² Whereas the no-principle option provides no guidance to an agent, a principle such as “nonconsensual medical experimentation is prohibited” clearly provides guidance and, if indeed a contractualist moral principle, cannot be reasonably rejected. But simply reintroducing at this stage the contractualist principles describing what an agent objectively ought to do is ruled out by the question at hand, which is whether contractualism should ¹¹ Note that this argument presupposes ex ante as opposed to ex post contractualism. Under the latter, we would have to weigh objections from the standpoint of agents who are prohibited from objectively permissible acts they have reason to perform (as opposed to merely facing the risk of such) and the standpoint of patients who are inadvertently treated objectively wrongly by morally uncertain agents. Full discussion is impossible here, but the obvious worry is that ex post contractualism would favor an excessively conservative or risk-averse stance here, just as in Ashford’s (2003) airplane case mentioned earlier. ¹² I’m grateful to Jamie Dreier and Benjamin Kiesewetter for this objection.
136
provide special (“subjective”) principles for acting in light of moral uncertainty. To answer that question, we assume that agents cannot follow the otherwise justified contractualist moral principle directly and consider whether the reasonable-rejection criterion justifies additional moral instructions for this scenario. I have argued that it does. The conservative, moderate, and permissive principles occupy different positions concerning one dimension of moral uncertainty: the agent’s degree of credence in the view that the action in question is morally wrong. MEC is also sensitive to another dimension of moral uncertainty: the moral stakes of the situation (i.e., the choiceworthiness of the agent’s options) according to each of the moral views in which the agent has credence. To see that the best version of a contractualist moral uncertainty principle is also sensitive to the moral stakes, compare two candidate principles: S1: All else equal, the worse a wrong that a moral view in which an agent has credence claims that a particular action would be, the more reason the agent has to avoid taking that action. S2: The moral stakes of an action according to the moral views in which an agent has credence make no difference to what the agent ought to do. Which principle does the reasonable-rejection criterion favor? From the agent’s standpoint, and taking the moderate principle as established, S2 requires the same level of moral caution regardless of the agent’s view about the moral stakes, whereas S1 requires a higher degree of caution in some cases (i.e., those in which the potential wrong is more severe) but a lower degree of caution in other cases. There seems no reason to think that either principle interferes more, on the whole, with the agent’s pursuit of their legitimate interests, so they are equally burdensome and equally objectionable. However, from the patient’s standpoint, there is a stronger objection to S2 because, for a given level of imposition on agents, S1 more effectively protects the legitimate interests of patients by focusing the agent’s exercise of moral caution on cases in which potentially significant moral wrongs are at stake rather than those in which more trivial moral wrongs are at stake.¹³ So, S2 can reasonably be rejected in favor of S1.
¹³ I assume the moral agents applying these principles are normally competent in the sense that, although they may often have credence in false moral views, their moral thinking is not wholly or radically unreliable. Thus, actions that agents believe would, if wrong, be very wrong,
137
So far we have focused on cases in which the agent’s interests are at odds with the patient’s interests. Similar reasoning justifies similarly impartial concern for the interests of those involved in other kinds of moral scenarios, such as those in which an agent must balance the competing claims of two or more patients, though I will not explore these separately here.
3. Content-neutral vs. discriminating moral uncertainty principles A principle is content-neutral if it holds that an agent’s moral credences determine what they ought to do in a way that takes impartial account of the moral views in which the agent has credence. MEC is content-neutral because it takes account of the choiceworthiness of an agent’s options according to each moral view in which the agent has credence, without bias in favor of some views or against others. A principle is discriminating if it is not impartial, by allowing only some kinds of moral views in which an agent has credence to affect what the agent ought to do, or weighting some kinds of views more heavily than others. A principle may discriminate along various dimensions depending on which features it claims justify giving more weight to some moral views than others. In this section, I argue that the best contractualist moral uncertainty principle is discriminating in at least two ways.
A. Uncertainty about the truth of contractualism One may be uncertain about the implications of contractualism: about which principles cannot be reasonably rejected, or how a principle applies to a particular case. One may also be uncertain about the truth of contractualism itself. A discriminating principle might treat these kinds of moral uncertainty differently. Using “affect what one ought to do” as schematic shorthand for what is described by the moderate moral uncertainty principle are more likely actually to threaten the legitimate interests of patients in a severe way than actions they believe would be less severely wrong. If their moral thinking were wholly unreliable in this regard, then S2 might better protect the legitimate interests of patients. The reasonablerejection criterion may thus justify different moral uncertainty principles for people with different degrees of moral competence. My thanks to Ron Aboodi for helping me see the need for this assumption.
138
plus any appropriate modifications or extensions (such as those described earlier and elsewhere in this section), we can consider two candidate principles: TC1: Moral uncertainty about the implications of contractualism and about which moral theory is true both affect what one ought to do. TC2: Moral uncertainty about the implications of contractualism affects what one ought to do, but moral uncertainty about the truth of contractualism does not. TC1 is content-neutral; TC2 is discriminating. To illustrate the difference, suppose that someone is fairly confident that contractualism is true and that it does not require donating most of one’s disposable income to charity to help those in need. However, they also have some credence in utilitarianism, which they believe does require donating most of one’s disposable income. Depending on the details of the case and the moderate moral uncertainty principle, TC1 might require the agent to donate at least some additional money. According to TC2, an agent’s uncertainty about which moral theory is true does not affect what they ought to do, so they are not required to donate more than contractualism requires. By contrast, if their uncertainty were about whether contractualism itself requires that they donate most of their disposable income to charity, then TC1 and TC2 would not differ in their implications. Note that “uncertainty about the implications of contractualism” should be understood in a broad way that does not require thoughts about the philosophical theory of that name.¹⁴ Scanlon (1998 155) argues that contractualism captures a central feature of ordinary moral thought and motivation, namely “the ideal of justifiability to others” or “the sense that others could reasonably object to what I do.” One need not be a philosopher to think of some actions as required on this basis, as opposed to because they make the world better, conform to nature, or fulfill God’s commands. Which principle does the reasonable-rejection criterion favor? From the patient’s standpoint, more conservative moral uncertainty principles are favored over more permissive ones, all else equal. That seems to favor TC1 over TC2, but the situation is more nuanced. The basis on which the patient can object to a more permissive principle is that it increases the risk that
¹⁴ I’m grateful to Kyla Ebels-Duggan for raising this issue.
139
morally uncertain agents will treat them in ways that are objectively morally wrong, violating their legitimate interests as determined by contractualism’s other moral principles. Because the patient is not entitled to the benefits of what non-contractualist moral theories falsely claim the agent ought to do, the proper question is not whether TC1 or TC2 is most favorable to the patient without qualification, but rather which better protects the patient specifically from objective moral wrongdoing as described by contractualism. TC2 is likely to protect a patient’s legitimate interests better than TC1, since the latter differs only by giving weight to an agent’s credence in non-contractualist moral theories, which may prescribe actions that contractualism regards as wrong or fail to prescribe actions that contractualism regards as required. Of course, on some possible occasions an agent would more effectively avoid objective wrongdoing by following TC1. This result could obtain if the agent has credence in contractualism but mistaken views about its implications, and also has credence in a non-contractualist moral theory that happens to prescribe the same action that contractualism in fact prescribes. Suppose that someone disinclined to help others mistakenly believes that contractualism requires making no more than slight sacrifices to save an imperiled person, but their credence is divided between contractualism and utilitarianism, and they must now choose whether to make a moderate sacrifice to save someone’s life. They would more effectively avoid objective wrongdoing by following TC1. But on average, an agent’s views about the implications of contractualism will likely do better than their views about the implications of other moral theories in which they have credence at approximating the true implications of contractualism, so allowing credence in non-contractualist moral theories to affect what one ought to do less effectively protects the legitimate interests of patients. From the agent’s standpoint, more permissive moral uncertainty principles are favored over more conservative ones, all else equal, because they less often require refraining from objectively permissible actions that the agent may have reason to perform. Because TC1 gives weight to uncertainty about which moral theory is true as well as uncertainty about the implications of contractualism, it requires agents to sacrifice their legitimate interests out of moral caution on more occasions, so agents have a stronger objection to TC1 than TC2. To be sure, on some occasions an agent would more effectively avoid objective wrongdoing without shortchanging their own legitimate interests by following TC1. Someone might mistakenly
140
believe that contractualism requires donating most of their disposable income to charity, and have credence in both contractualism and a non-contractualist moral theory according to which only moderate efforts to help others are required. In this case, TC1 better protects the agent’s legitimate interests. But as before, an agent’s views about the implications of contractualism will likely do better on average than their views about the implications of other moral theories in which they have credence at approximating the true implications of contractualism. Because TC1 is thus more objectionable from both agent and patient standpoints, the reasonablerejection criterion favors TC2.
B. Uncertainty about which kinds of beings have moral status According to the reasonable-rejection criterion, a principle is justified if the strongest objection to it is weaker than the strongest objection to any alternative principle. Objections are raised from relevant standpoints in light of the personal reasons of someone occupying that standpoint. Now, even someone confident that contractualism is true might be uncertain about which kinds of beings have standpoints relevant for applying the reasonable-rejection criterion. I will describe this as uncertainty about which kinds of beings have moral status. A content-neutral moral uncertainty principle such as MEC takes no account of whether an agent’s moral uncertainty is of this sort, but a discriminating principle may treat it as relevant. Consider the example of non-human animals. Taking for granted that rational beings have moral status, we can identify two candidate moral uncertainty principles: MS1: Moral uncertainty about whether animals have moral status affects what one ought to do. MS2: Moral uncertainty about whether animals have moral status does not affect what one ought to do. MS1 is content-neutral; MS2 is discriminating. Which principle does the reasonable-rejection criterion favor? Applying the criterion requires assumptions about which standpoints are relevant for evaluating candidate principles, so which principle for dealing with uncertainty about moral status is justified may depend on which kinds of beings
141
actually do have moral status. Supposing that includes both rational beings and animals, two standpoints are relevant: (1) rational beings and (2) animals. From the standpoint of animals, MS1 is obviously highly objectionable; it threatens the legitimate interests of animals by prescribing that agents who have doubts about the moral status of animals give less weight to their interests. One might think that from the standpoint of rational beings, MS1 is preferable for the same reason: it permits discounting the interests of animals for the benefit of rational beings. But the interests that underwrite objections to candidate moral uncertainty principles are the legitimate interests of those who occupy each standpoint, as determined by contractualist moral principles. If animals have moral status, then benefits to rational beings that would be justified only if animals did not have moral status do not represent legitimate interests of the former and do not constitute reasons for objecting to a candidate principle. So, from the standpoint of rational beings, MS1 and MS2 are equally preferable. Thus, if animals actually do have moral status, the reasonable-rejection criterion favors MS2. Now suppose that animals do not have moral status. In that case, only the standpoint of rational beings is relevant. Rational beings can now object that MS1 will sometimes require compromising the legitimate interests of rational beings for the sake of animals. From the agent perspective, it may require morally uncertain agents to refrain from objectively morally permissible actions that they may have reason to perform, such as eating meat. From the patient perspective, it may require agents to trade off the interests of other rational beings with the interests of animals. For example, it may require that a morally uncertain scientist not perform medical experiments on animals for the sake of benefiting human beings. So rational beings have a stronger objection to MS1 than to MS2, which rules out moral concern for animals regardless of an agent’s beliefs about their moral status. Thus, if animals do not actually have moral status, the reasonable-rejection criterion favors MS2. The overall result is that the reasonable-rejection criterion favors MS2 regardless of whether animals, like rational beings, have moral status. But the implication of that principle depends on the truth of the matter: it requires showing moral concern for animals if they do have moral status, and not showing it (at the expense of rational beings) if they do not, regardless of an agent’s degree of credence in each possibility. This result contradicts one of the standard examples in discussions of moral uncertaintism. It is often claimed that if one is uncertain about whether eating meat is
142
morally wrong, but confident that vegetarianism is permissible, then one ought to be a vegetarian in order to avoid the moral risk of eating meat. The basis for thinking that eating meat may be wrong is typically taken to be the possible moral status of animals, as opposed to potential risks to rational beings (e.g., if harming animals made people more likely to harm humans). The argument thus assumes not only that moral uncertainty affects what one ought to do, but that uncertainty about moral status in particular affects what one ought to do. That assumption would be unproblematic if the relevant moral uncertainty principle were a content-neutral principle such as MEC. But according to contractualism’s local moral uncertainty principle, it is mistaken. If animals do not in fact have moral status, then eating meat is permissible even if one believes that it is or may be morally wrong. And if animals do have moral status, then eating meat is wrong (assuming that other contractualist principles would then prohibit it) even if one believes that it is or may be permissible. One might worry that a discriminating principle such as MS2 does not provide adequate guidance to morally uncertain agents, since whether animals actually do have moral status is precisely what they do not know. However, the ultimate justification for MS2 is not that it provides guidance to morally uncertain agents but that the reasonable-rejection criterion favors it. Considerations about whether a moral uncertainty principle provides effective guidance are important insofar as they bear on the objections that can be raised to candidate principles from relevant standpoints. But objections can be raised on other grounds as well, and there is no guarantee that the principle to which the strongest objection is weakest will be one that morally uncertain agents can effectively apply and follow in every situation. That said, discriminating principles such as MS2 still provide a kind of guidance to morally uncertain agents. We can distinguish between positive guidance, which tells an agent what they ought to do in terms they can put into practice, and negative guidance, which rules out some answers to the question of what one ought to do or some arguments in favor of particular answers. Though MS2 does not provide positive guidance, it provides negative guidance by ruling out MS1. Someone who knows that MS2 is true knows that they are not required to refrain from eating meat (etc.) simply on account of their uncertainty about whether animals have moral status, which is an important contribution to their moral deliberation even if it does not settle the issue on its own. Without positive guidance, the agent is left to their own devices to decide how to act. They might decide to
143
follow the moral view in which they have most credence. If they value doing what is right, it may be instrumentally rational for them to refrain from eating meat just in case it is wrong, even though this precaution is not morally required if animals in fact lack moral status. The earlier argument generalizes to principles that encompass other kinds of beings. Candidate principles may say (like MS1) that uncertainty about the moral status of a kind of being affects what one ought to do, or (like MS2) that it does not. From the standpoints of the kinds of beings that actually do have moral status, principles of the first sort are objectionable because they prescribe discounting their interests insofar as the agent believes that they may lack moral status and/or diluting concern for their interests with concern for the interests of kinds of beings that the agent mistakenly believes may have moral status. So the reasonable-rejection criterion favors principles of the second sort, according to which uncertainty about whether a kind of being has moral status makes no difference to what one morally ought to do. Thus, this reasoning applies straightforwardly to another standard example in the literature on moral uncertainty: uncertainty about the permissibility of abortion grounded in uncertainty about the moral status of the fetus (e.g., Moller 2011). If the fetus in fact lacks moral status and abortion is objectively permissible, it remains permissible for the morally uncertain agent. Similar reasoning also justifies giving no weight to credence in moral views on which some beings have a lower degree of moral status than they actually have, in the sense that objections from their standpoint are discounted relative to objections from other standpoints. So, credence in sexist or racist views, or others of the sort, does not subjectively permit discounting the interests of other people on such grounds.
4. Some advantages of local moral uncertainty principles Philosophical discussions of moral uncertainty principles have typically assumed that they must be global principles, and some central objections to moral uncertaintism trade on the distinctive features of global principles. In this section, I show how the contractualist approach avoids two such objections. Similar responses are likely available on behalf of local moral uncertainty principles within other moral theories.
144
A. The regress objection A common worry is that moral uncertaintism leads to a vicious regress of higher-order moral uncertainty principles.¹⁵ Just as people can be uncertain about first-order moral issues such as vegetarianism and abortion, they can also be uncertain about the morally appropriate response to moral uncertainty, or in other words, about which (if any) moral uncertainty principle is true. If uncertainty about first-order moral issues justifies a moral uncertainty principle that describes what an agent ought to do given their uncertainty, then uncertainty about moral uncertainty principles seems to justify a higher-order moral uncertainty principle that describes what an agent ought to do given their uncertainty about that issue. But people are also likely to be uncertain about which second-order moral uncertainty principle is true, which would justify a third-order moral uncertainty principle, and so on, apparently without end. The result is that, if moral uncertaintism is true, there seems to be no determinate answer to the question of what one ought to do in situations of moral uncertainty. Therefore, the objection goes, we should reject moral uncertaintism. The threat of regress arises because typical justifications for global moral uncertainty principles, such as the idea that there should be principles that agents can apply given their actual (imperfect) epistemic situation, seem to justify higher-order moral uncertainty principles as well. Because local approaches can appeal to the full resources of a substantive first-order moral theory, they have more options for justifying first-order moral uncertainty principles in a way that does not also justify endlessly higher-order moral uncertainty principles. Different first-order moral theories might accomplish this task in different ways, but the contractualist can avoid the regress in the following way. People cannot reasonably reject the first-order moderate moral uncertainty principle because it is a definite improvement compared to the no-principle option; the strongest objection to the firstorder moderate principle is weaker than the strongest objection to the noprinciple option. However, an infinite set of higher-order moral uncertainty principles is not an improvement over a second-order no-principle option, so it can reasonably be rejected. It is not an improvement because, since it fails to provide determinate guidance about what to do in situations of moral uncertainty, it less effectively protects the interests of potential victims of
¹⁵ See, e.g., Weatherson (2014: 155–7).
145
decision-making under moral uncertainty than first-order moral uncertainty principles alone.¹⁶ It is therefore more objectionable from the patient’s standpoint. It is also more objectionable from the standpoint of agents, who can object to bearing the burden of attempting to apply such a complex set of principles, especially if it does no better at protecting the interests of potential victims and thus burdens them to no point. By contrast, a firstorder moderate moral uncertainty principle does not seem forbiddingly complex or demanding. It is the same sort of standard of appropriate caution that contractualism already requires of people in the case of risky activities such as driving and air travel.
B. The content objection Because people can have credence in morally awful views, moral uncertainty principles such as MEC entail that people sometimes morally ought to do morally awful things, which seems false. For example, Harman (2015) describes the case of a father who has credence in conservative moral views that forbid him to teach his daughter to drive. She argues that, nevertheless, it would be wrong for him to refuse to teach her. Similarly, one of Weatherson’s (2019) chief objections to moral uncertaintism is that it justifies morally abhorrent behavior from people with false moral views. He (2019: 1–2) writes: What is important is doing the right thing, being a good person, and having rational beliefs. If one has misguided views about the right, the good, and the rational, then there is nothing good about conforming to those misguided views. And this matters, because many people have views about the right, the good, and the rational, that are very misguided indeed.
He considers the example of Robespierre, who ordered the executions of many innocent people in the name of what he sincerely believed was justice, and concludes: “It’s not good, in any way, to be that kind of ideologue” (2019: vii).
¹⁶ Possibly the addition of second-order or finite higher-order moral uncertainty principles would serve the interests of patients and/or agents better than first-order moral uncertainty principles alone. I will not investigate this possibility here, but it poses no problem for the account.
146
The content objection arises because of the content-neutral nature of global moral uncertainty principles: they give weight to an agent’s credence in a moral view no matter how awful it is by the standards of the true moral theory. The discriminating character of some local moral uncertainty principles gives them the potential to avoid, or at least mitigate, the content objection. To illustrate, consider the contractualist approach I have described. In Weatherson’s example, Robespierre has little or no credence in contractualism or its underlying moral ideal of treating people in accordance with principles they could not reasonably reject. He endorses a different, and terribly mistaken, conception of morality. But the contractualist principle specifies that only uncertainty about the implications of contractualism affects what one morally ought to do. Robespierre’s credence in abhorrent moral views is irrelevant; what he ought to do is simply to act in accordance with the correct contractualist moral principles, which prohibit (e.g.) having people executed for the reasons he did. The contractualist can thus agree that Robespierre’s actions were in no way morally right. Applying the contractualist principle to Harman’s example requires further specifying the grounds of the father’s doubts about the permissibility of teaching his daughter to drive. If they result from credence in noncontractualist moral theories, then his case is like Robespierre’s. If they result from credence in the view that women lack moral status or have lower moral status than men, then again, the contractualist principle specifies that they make no difference to what he ought to do. These are the likely explanations. It is difficult to see how someone could mistakenly believe on other grounds that contractualism justifies a prohibition on teaching women (but not men) to drive, but if the father has somehow acquired credence in that view, then the contractualist’s moderate moral uncertainty principle leaves open the possibility that he acts rightly by refusing to teach her. But the implausibility of the scenario means that this result is no bullet for the contractualist to bite in practice. Moral uncertaintism will inevitably prescribe some deviation from what would otherwise be right, so its proponents need not worry about this result as long as they can avoid the most serious counterexamples.
5. Conclusion My aim has been to explore the general structure of the most plausible contractualist account of what people morally ought to do in situations of
147
moral uncertainty. This inquiry demonstrates the plausibility of treating moral uncertainty principles as local components of first-order moral theories rather than freestanding global principles. It also contributes to the development of contractualism by applying it to the topic of moral risk, an underexplored issue in contractualist theorizing. A question for further exploration is whether other moral theories can also justify local moral uncertainty principles and how those principles might resemble or differ from the contractualist principles that I have described. Similar arguments apply, at least prima facie, in the case of contractarian moral theories. The underlying idea is that people have reason to prefer that those with whom they interact respond to moral uncertainty in some ways rather than others. If morality is a kind of contract negotiated and agreed upon by individuals (or their hypothetical proxies) in order to better advance their own interests, as on a contractarian view, then people have reason to prefer a contract that includes specific sorts of instructions about what to do in light of uncertainty about the content of the contract or how to apply it to particular cases, since a contract with these provisions will better promote their interests. However, structurally more distant moral theories, such as consequentialism, other forms of deontology, and virtue ethics, must await consideration elsewhere.¹⁷
References Ashford, Elizabeth. 2003. “The Demandingness of Scanlon’s Contractualism.” Ethics 113(2): 273–302. Frick, Johann. 2015. “Contractualism and Social Risk.” Philosophy and Public Affairs 43(3): 175–223. Fried, Barbara. 2012. “Can Contractualism Save Us from Aggregation?” Journal of Ethics 16: 39–66. Gauthier, David. 1986. Morals by Agreement. New York: Oxford University Press. Guerrero, Alexander A. 2007. “Don’t Know, Don’t Kill: Moral Ignorance, Culpability, and Caution.” Philosophical Studies 136: 59–97.
¹⁷ For helpful feedback on the ideas in this paper, I am grateful to Samantha Yuan, two anonymous referees, audiences at the University of Florida, the 2021 Pacific APA (with Rahul Kumar as commentator), the 2022 Arizona Workshop in Normative Ethics, and participants in a reading group at Florida State University.
148
Harman, Elizabeth. 2015. “The Irrelevance of Moral Uncertainty.” In Oxford Studies in Metaethics, edited by Russ Shafer-Landau, vol. 10, 53–79. New York: Oxford University Press. Kumar, Rahul. 1999. “Defending the Moral Moderate: Contractualism and Common Sense.” Philosophy and Public Affairs 28(4): 275–309. Kumar, Rahul. 2015. “Risking and Wronging.” Philosophy and Public Affairs 43(1): 27–51. Lenman, James. 2008. “Contractualism and Risk Imposition.” Politics, Philosophy and Economics 7(1): 99–122. Lockhart, Ted. 2000. Moral Uncertainty and Its Consequences. New York: Oxford University Press. MacAskill, William. 2019. “Practical Ethics Given Moral Uncertainty.” Utilitas 31: 231–45. MacAskill, William, Bykvist, Krister, and Ord, Toby. 2020. Moral Uncertainty. New York: Oxford University Press. MacAskill, William, and Ord, Toby. 2020. “Why Maximize Expected ChoiceWorthiness?” Nous 54(2): 327–53. Moller, D. 2011. “Abortion and Moral Risk.” Philosophy 86(3): 425–43. Ross, Jacob. 2006. “Rejecting Ethical Deflationism.” Ethics 116(4): 742–68. Rüger, Korbinian. 2018. “On Ex Ante Contractualism.” Journal of Ethics and Social Philosophy 13(3): 240–58. Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Harvard University Press. Scanlon, T. M. 2013. “Reply to Zofia Stemplowska.” Journal of Moral Philosophy 10: 508–14. Sepielli, Andrew. 2009. “What to Do When You Don’t Know What to Do.” In Oxford Studies in Metaethics, vol. 4, edited by Russ Shafer-Landau, 5–28. New York: Oxford University Press. Singer, Peter. 1972. “Famine, Affluence, and Morality.” Philosophy and Public Affairs 1(3): 229–43. Smith, Michael. 1994. The Moral Problem. Malden, MA: Blackwell Publishing. Weatherson, Brian. 2014. “Running Risks Morally.” Philosophical Studies 167: 141–63. Weatherson, Brian. 2019. Normative Externalism. New York: Oxford University Press. Zimmerman, Michael J. 2008. Living with Uncertainty: The Moral Significance of Ignorance. New York: Cambridge University Press.
7 The Ethics of Expectations Rima Basu
She was folding laundry when I told her I was gay. I’ll never forget the look of sadness mixed with betrayal that fell across her face. I had seen my mother cry before, but never quite like that. I had hurt her in a way that I still struggle to make sense of, and she had hurt me too in a way that I still struggle to articulate. So, I wrote this paper instead of going to therapy. Although that last sentence is meant in jest, it serves to highlight a mismatch between the central significance of parental expectations in our lives and the little philosophical analysis such expectations have received. At least with regard to my mother’s beliefs, I can pinpoint where she goes wrong because we have an ethics of belief. That is, we know what belief aims at, we know the norms governing belief formation, and perhaps we even know when beliefs can wrong. Expectation, however, is trickier to get a handle on. The first part of the paper will be devoted to getting a rough handle on the target phenomenon, i.e., the nature of expectations. To get a broad account of the attitude of expectation on the table I begin by comparing expectation to more familiar attitudes that, when taken together, encompass the multiple roles that expectations seem to play. These attitudes include belief, including credences and normative beliefs and second-order beliefs about what other people believe, desire, aspiration, and interpersonal hope. Taking such a wide view will enable us to account for not only the different things we do when we expect but also the multiple ways what we expect can be inappropriate or wrong. That is, once we get clear on the nature of expectations, or at least, unravel its multiple threads, we can turn to the second part of this paper: the wrongs of expectations. In the second part of this paper, I argue that given the multiple roles played by expectations there will be multiple norms governing different kinds of expectations. In turn, there will be multiple ways that expectations can not only go wrong but also wrong. Of particular interest to me, given the opening example, are the ways that expectations can wrong the subjects of Rima Basu, The Ethics of Expectations In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Rima Basu 2023. DOI: 10.1093/oso/9780198895909.003.0008
150
the expectations. Although not all of us know what it’s like to come out to our parents, we do each know what it’s like to let them down and the hurt and disappointment that follow. Worse still, some expectations can be deeply alienating and the hurt that accompanies such expectations cuts us in deeper ways. I must however warn the reader expecting a full accounting of precisely how and when expectations wrong, that such an account is not forthcoming. The sheer range of types of expectations and the different functions expectations play in our lives leaves me skeptical there is a truly general account that can be offered. My goal in the second part of this paper is thus humbler: simply to canvas the available options for capturing the wrongs of expectations and leave it to future work to see if a unifying story can be told. Any unifying account must be able to capture not only the variety of expectations but also the variety found in ways that expectations can wrong.
1. The Nature of Expectations As little consensus as there is with regard to the nature of belief, expectations are trickier still. Whatever it means to say that beliefs aim at the truth, most agree that something has gone right when beliefs are true and something has gone wrong when they’re false. To say the same for expectations, i.e., when they go right and when they go wrong, we need to get clear on the functional role that the attitude of expectation plays. That is, what are we doing when we expect? Sometimes expectations have a functional profile that is belieflike, i.e., like a theoretical attitude, but sometimes expectations act in different ways altogether, i.e., more like practical or evaluative attitudes. To see this, let us start by considering the following expectations. (1) (2) (3) (4) (5) (6)
I expect that it’ll rain tomorrow. I expect the meeting will end by noon. I expect we’ll run into each other at the party. I expect you to do the reading before class. I expect you to succeed at everything you set your mind to. I expect you to be on your best behavior.
On a first glance, there is something (1)–(6) have in common: they each contain a predictive element about what the world will be like. It might be tempting to suggest that the rationality of expectations is akin to the
151
rationality of beliefs. That is, the rationality of expectations is dependent on whether you have sufficient evidence to make the prediction. For example, it’d be inappropriate to expect that it’ll rain tomorrow if there’s only a 5 per cent chance of rain. It would be inappropriate to expect that the meeting will end by noon unless you have good reason to think so. If you’re not going to the party it’d be odd to tell someone you expect to run into them there. So, let us start by considering this connection between expectation and belief in more detail.
1.1 Expectations as predictions Perhaps expectations are simply future-directed beliefs, that is, expectations are our predictions about what we think will happen. If this is right, expectations would also be weaker than belief because although expectations aim at getting things right, there’s an inbuilt threshold of forgiveness for getting things wrong given uncertainty about the future. For example, if the meeting runs past noon because of an unforeseen urgent agenda item or because someone was uncharacteristically late, you weren’t wrong to have expected the meeting to end by noon and to have planned your day accordingly. We might then expect expectations to be more akin to a report of our credences than a report of our outright beliefs. There are, however, two reasons to doubt this initial analysis. First, expectations are not necessarily future-directed.¹ Rather, some expectations can be about the past. Second, expectations do not merely express our estimations or best guesses.² Rather, sometimes expectations express our wishes or desires for what we hope the world will be like and in performing this second function, expectations more closely resemble desires than beliefs. On the first point consider the following past-directed expectations: (7) I expect the dog didn’t eat her breakfast. (8) I expect it rained last night. Although these expectations are past-directed, they still seem akin to the best guess gloss on expectations we began with. That is, the rationality of (7) and (8) seems to depend on whether you have good evidence for what you ¹ Thanks to Gabbrielle Johnson for pushing me to say more on this point. ² For an account of thinking (and in turn, believing) in terms of guesses, see Holguín (2022).
152
expect to find out was the case. This can be accommodated under a prediction gloss on expectations by noting that these expectations are aiming at discovering something about the world. That is, these expectations are still predictions. Further, they involve some element of discovery, of something not yet known or anticipation of figuring something out. For example, when we think about when someone would utter (7) or (8), it is because there is something in need of explaining, e.g., that the dog keeps begging for food or that the driveway is wet. Thus, although not futuredirected, the expectations are still operating like predictions. These expectations are guesses about what the world must have been like in order to explain what the world is like. However, not all expectations are predictions. Consider the expectation that your students will do the reading before class, i.e., (4). It’s felicitous to say that you expect your students to do the reading but that you believe otherwise and will be lesson planning accordingly. Similarly, you can make plans with your perpetually tardy friend and expect to see them at 9 a.m. but fully believe that they’ll be late or simply not show because they slept in. That you have strong evidence on the contrary does not make it any less appropriate to have some of the expectations listed above. It does, however, sound odd to say that you expect that it will rain tomorrow but you don’t believe that it will. That is, we can contrast the following: (9) I expect students to do the reading, but I don’t believe they will. (10) #I expect it will rain tomorrow, but I don’t believe it will. Expectation is functioning differently in (9) than in (10). Expectation is playing a predictive role in (10) and that results in the Moore-paradoxical character of that sentence. You cannot predict that p, but also predict that not-p. The fact that (9) does not have the same Moore-paradoxical character suggests that “expect” is playing a different role. Thus, sometimes expectation has a different functional profile than belief. To explore this difference in functional profile I turn now to two possible explanations of why expectation behaves differently in (9). First, in section 1.2 I explore whether expectations are normative beliefs, beliefs about what should happen. That is, students should do the reading, but I don’t believe they will. This normative use of expect contrasts with the epistemic or predictive use of expect in (10). Alternatively, in section 1.3 I explore whether expectation is better glossed as a kind of aspiration, some hope you’ve invested in your students. That is, I hope students will do the reading,
153
but I don’t believe they will. This would similarly give us a normative use of expect where expectations express proleptic reasons. Despite these different roles, in section 1.4 I argue that these different roles of expectation can be brought together under a unifying account of what expectations do for us by deploying a familiar metaphor more common to discussions of the nature of belief: expectations provide us with maps. The nature of expectations is to give different types of guidance.
1.2 Expectation as prescription Perhaps expectations are simply normative beliefs about what we think should happen or would like to happen. That is, perhaps expectations have a direction of fit more like desires than belief. Standardly, discussions of direction of fit begin with Anscombe’s (2000, p. 56) grocery shopping case and the contrast between two kinds of lists: one compiled by a shopper and one compiled by a detective following the shopper around the grocery store. The list compiled by the shopper is an expression of their desires. If what the shopper has in their cart doesn’t match the list, then there’s no mistake in the list but rather a mistake in the shopper’s performance. However, if what is in the shopper’s cart doesn’t match the list compiled by the detective, then there is a mistake in the detective’s list. The detective’s list is subject to revision when there’s a mismatch between the contents of the shopper’s cart and the detective’s list, whereas the shopper’s list is not similarly subject to revision. The lists relate to the world in different ways, just as desire and belief relate to the world in different ways. With this in mind, consider again the expectation that students will do the reading but the belief that they won’t. Is it a mistake to expect that students will do the reading but not believe that they will? No, not if the expectation is expressing a desire. Beliefs function to track the world; desires express our wishes for what we’d like the world to be like. Expectations, however, are not merely desires. As we’ve seen, they also contain an element of prediction which complicates the story. The expectation that students will do the reading is not the same as the expectation that you’ll get milk at the store. When you expect to get milk at the store but fail to put the milk in your cart that’s a failure of execution, a failure in your performance. When your students fail to do the reading, what kind of failure is that? It can’t be simply that your expectation communicates your desire for students to do the reading because they have no reason to care about your desires or whether
154
your desires eventuate. As frustrating as it is to say the following vaguery: the expectations that go beyond mere predictions seem to be doing something more than just adding a prescription to the picture. It is this something more that we need to get a handle on before we can make progress on outlining how expectations can wrong. An account that could help us get clearer on this additional function of expectation is Cristina Bicchieri’s (2016) account of expectations according to which expectations are a certain class of beliefs, beliefs about what is going to happen or what should happen. However, as I noted earlier, expectations can also be beliefs about what has happened. Nonetheless, this account can capture the two roles that we’ve seen expectations play: a predictive role (what has or is going to happen) and a prescriptive role (what should happen or should have happened). The expectation that your students will do the reading is a belief about what should happen not how things will in fact play out. This is why there’s no contradiction, nothing Mooreparadoxical, in expecting students to do the reading but believing they won’t. Furthermore, the expectation contains an implicit prescription in the form of evaluating such a future state of affairs as a good state of affairs. That is, the students should do the reading because it would be good for them and that is why they should want to. Further, consider the following two expectations: (5) I expect you to succeed at everything you set your mind to. (6) I expect you to be on your best behavior. Both of these expectations involve predictive and prescriptive elements and something more. What it means “to succeed” is normatively-laden. For example, my parents and I disagree about what’s essential to success. Similarly, we have disagreements about “best behavior.” These expectations go beyond mere predictions or prescriptions in that these expectations are heavier. We often talk metaphorically of the weight of expectations or of expectations being a burden, and that metaphor of weight shouldn’t be dismissed as mere loose talk. This way of talking about expectations directs us to look for what more these expectations seem to include that makes them feel so weighty. Returning to Bicchieri (2016), perhaps what’s going on here is that these expectations are communicating not only our own predictions and prescriptions but also our beliefs about other people’s personal normative beliefs. Namely, that expectations express second-order beliefs, beliefs about what
155
other people believe. For example, returning to Anscombe’s grocery store, if we were to try to explain the behavior of the shopper and why they move through the store as they do, queue where they do, pay what they do, etc., it would be useful to explain all of that in terms of the shopper’s beliefs and their beliefs about what other shoppers believe and what the store workers believe. Group action is a coordination problem and it’s helpful to know what an agent believes and what they believe others they’re attempting to coordinate with believe. Although this example does not feel as weighty as the expectations in (5) and (6), some normative social expectations might simply serve to ease and enable this kind of social coordination. Expectations help structure these interactions so they go more smoothly.³ However, not all of the cases of expectations are on the same level with respect to their weightiness. That is, some of these expectations perform a function that goes beyond the easing or enabling of social coordination. That is, some expectations also engage in the practice of shaping or of influence. Consider again my mother’s expectations of me. We can make some progress on understanding the case by analyzing it in terms of my mother’s belief “that other people believe (and will continue to believe) that certain behaviors are praiseworthy and should be carried out, while others should be avoided” (Bicchieri 2016, p. 12). When my mother told me she never wanted life to be difficult for me, she was expressing her belief that other people believe that the life I had “chosen” for myself was not a praiseworthy one, that I ought not to have chosen such a path and made life hard for myself. However, the depth of her disappointment can’t be captured by beliefs about what the Chatterjees, Dasguptas, or Goswamis would think or even the prescription that I should act in accordance with prevailing social or cultural norms of what’s expected of a good Indian daughter. What is missing from the explanation of the weightiness of some expectations is whatever mechanism is responsible for the obligation-generating nature of such expectations. Some expectations are not just predictive or prescriptive or second-order beliefs about what other people believe. Some expectations also express relationships of dependence and reliance. We depend on people, rely on people, and feel let down by one another when we can no longer depend or
³ For another example, see Breakey’s (2022) discussion of expectations and how expectations give rise to obligations. In common with Bicchieri, for Breakey focusing on expectations as predictions helps to explain why expectations give rise to obligations, i.e., such expectations help us navigate the world by making it more predictable.
156
rely on those same people. I feel let down when my mother doesn’t accept me in a way that’s different from when other people don’t. My mother’s expectations—more generally, the kinds of expectations our parents, teachers, coaches, advisors, lovers, and other attachment figures have of us—expresses this kind of dependence and reliance. She was depending on me and now she can’t depend on me anymore. I upended the stable conditions under which she could think of me, and in doing so, she became unmoored. However, I too was depending on her, I needed her, and in her rejection of me, I too become unmoored. To make sense of this kind of unmooring what we need to add to our account of expectations is a story about how some normative expectations play a role in shaping who we are.
1.3 Expectations as aspirations To understand so-called weighty expectations we can start by noticing that there is a class of expectations that have an aspirational character in addition to a predictive or prescriptive character. That is, some expectations function in the same way as proleptic reasons. Proleptic reasons are reasons that are not yet our own but we expect will be ours in the future. For example, although your child may not currently show any interest or aptitude in mathematics it may nonetheless be appropriate to proleptically engage with them as though they will come to be interested in the subject. As Agnes Callard (2018, p. 43) explains, “proleptic reasons are provisional in a way that reflects the provisionality of the agent’s own knowledge and development: her inchoate, anticipatory and indirect grasp of some good she is trying to know better.” As Mark Schroeder (2020, p. 83) explains, proleptic reasons are reasons that “get a little bit ahead of themselves.” Proleptic reasons are central to the kinds of relationships we have with our parents and other guardian figures like teachers, coaches, advisors, and mentors; relationships that are characterized by attachment. As Monique Wonderly (2017, p. 242) explains, attachment involves: a set of evolutionarily adaptive behaviors that serve to provide the infant with a sense of security. The attached infant attempts to remain in close proximity to her primary caregiver, treats her as a “secure base” from which to safely explore unfamiliar surroundings, seeks her out for protection as a “safe haven” when threatened or hurt, and protests separation from her—for example, via clinging, crying, and other displays of distress.
157
In her work, Wonderly argues that the promise of an attachment account of love is that it can make sense of central features of love that are hard to capture on other accounts. Our interest is not in accounting for love, but in accounting for weighty expectations, though the two may be intertwined. For our specific purposes here this account of attachment can help explain how expectations function in the context of loving relationships.⁴ Consider, for example, love’s depth. When we lose what we love “we tend to feel as though we are ‘less together,’ on unstable ground, no longer ‘all of a piece,’ and so forth” (Wonderly 2017, p. 243). Compare this with the feeling of being unmoored that comes when weighty expectations are dashed. Our attachment figures provide a kind of stability and security. The functional role of their expectations is to provide that kind of stability and security in our lives. As we grow into the people we become, their expectations provide a kind of affectively charged scaffolding, their expectations shape us just as proleptic reasons anticipate who we will become. If that metaphor of scaffolding strikes you as familiar, that is because it is how Victoria McGeer (2008) describes interpersonal hope. Another common metaphor for hope is that of investment (see Martin 2020). These metaphors are useful for several reasons. First, we recognize the metaphor of scaffolding in the kind of structure that exceptions impose. Expectations help us grow by giving us direction. Second, we recognize the metaphor of investment in the way that some expectations involve a creation of debts and how that gives rise to the obligation-generating feature of those expectations.⁵ Often these debts are forgiven. Sometimes investments are made with no expectations attached. Other times, however, they can be held over you and when held in that way they can feel like an unwelcome burden. Think again of my mother. She had made sacrifices and invested her hopes and dreams into me. Perhaps her anger and disappointment in my coming out can be explained by my reneging on some implicit agreement that she thought we had. Reneging on some of the expectations that come with her investment in me.⁶
⁴ Thanks to Adrienne Martin for this pointer. ⁵ cf. Breakey’s (2022, p. 2801) discussion of expectations as hope-casts, that is, “[a] hope-cast is like a forecast, but where its holder also hopes for the predicted event.” These hope-casts can also be morally loaded, which explains why we often feel disappointment when the world (and the people in it) don’t match these expectations. As Breakey (2022, p. 2802) writes, “If X hopecasts that Y will A in S, then X will think Y ought to A in S. If his hope-cast is disappointed, the violation will be arousing and upsetting.” ⁶ Although not my focus here, it can be illuminating for criticism to compare my discussion with Chenyang Li’s (1997) discussion of filial duties and criticisms of accounts of filial obligation
158
Parents, and attachment figures more generally, are well-positioned to offer this kind of expectival scaffolding and investment, and this fits neatly with the proleptic role of expectation I’ve been trying to illuminate. If we think of these expectations in the same vein as interpersonal hope, then we are also in a position to recognize the following point from Martin (2020, p. 230): that “when we invest hope in people, we hope to create a certain intertwining of agencies. Put suggestively, the feeling that a person has let you down marks a hit on your agency.” Similarly, as Wonderly (2017, p. 244–5) notes: it is through our attachments with others that “they have the power both to help shape our agency (e.g., we function better with them; they play stabilizing, balancing, and corrective roles in our lives) and to cripple us in deep and devastating ways.” Weighty expectations shape who we are. The proleptic role of such expectations explains why they can feel intrusive and unwelcome. When we think of teenage angst it is directed at these weighty expectations; we are upset at our attachment figures for becoming too involved.
1.4 Expectations as maps So, what are expectations? We’ve seen that expectations play several multifaceted roles. Some expectations are predictive, like belief and credence, some are prescriptive, like desire and normative beliefs, some are proleptic, like aspiration and interpersonal hope, and some particularly weighty expectations emerge from our closest relationships because of how we depend upon, rely on, and influence one another. Some of the key features we’ve seen in expectations is that although they often play multiple roles, generally to expect is to anticipate.⁷ Even past-directed expectations involve anticipation. To capture this central feature of expectations it is illuminating that ground the obligations in either the contribution that parents make to the self, prudent investor accounts, and more generally, accounts that are based on the special relationship between parents and children. Thanks to Kenneth Silver for this pointer. ⁷ A reader here has asked whether this gloss on expectations can be right because it doesn’t sound right for prescriptive expectations. For example, one could expect their students to do the reading before class while at the same time anticipating that they won’t. I wonder, however, to what degree one can expect their students to do the reading without any anticipation at all that they will. We generally don’t expect things we don’t also anticipate to at least some degree. Perhaps we simply anticipate more strongly that they won’t have done the reading and thus we have lesson planned accordingly? It strikes me as infelicitous to in some sense say that you expect students to do the reading if every lesson plan is premised on no students having done the reading. Some degree of anticipation must be present.
159
to deploy a metaphor more common to discussion of the nature of belief: the metaphor of maps. Frank Ramsey (1990, p. 146) once remarked that beliefs subserve the function of navigation, that is, beliefs are the “maps by which we steer.” Expectations too provide maps by which we steer. Expectations are maps we draw of what we expect to find, sketches of what we think will be there, outlines of who we think we’ll be when we get wherever we’re going. But it does not follow from this metaphor that our expectations are simply our sketchy credences. Rather, the map metaphor can accommodate all the functional roles of expectation that we’ve outlined thus far. Maps, after all, can have many dimensions and overlays. Maps can tell you not just where things are located, but the relationships between them, and with respect to topological overlays, the shapes of the things themselves. We draw maps not only of things that are already there but of where things have been and where they will be. The metaphor of maps, thus, seems rich enough to capture the various roles that expectations play.⁸ The map provided by our attachment figures functions like a blueprint for our lives. The blueprint guides how we move through the world; it gives us a restricted but safe space in which to explore. Parents, and our attachment figures more generally, play a role in fixing a map for our lives and these maps contain implicit prescriptions about what ought to be there, what they expect to see, what other people expect to see, and their best guesses of how the map will be filled in as we use it to navigate. That map is then handed to us for us to fill in. Understanding expectations as maps can also help us to grasp another function that expectations play. Navigating via a map is an act of trust. We trust that the map portrays things accurately. Trust isn’t mere reliance; trust is marked by dependence. As we’ve seen with weighty expectations, they occur within relationships marked by dependence. When you are trusted, you are being counted upon. Karen Jones (1996) goes as far as to give an account of trust in terms of expectations. As Jones (1996, pp. 5–6) argues, to trust someone is “to have the confident expectation that, when the need arises, the one trusted will be directly and favorably moved by the thought that you are counting on her.” Furthermore, in a manner similar to the way ⁸ What provides this metaphor with the depth needed to capture the various roles of expectations is how maps can overlay one another and provide different dimensions and different ways of looking at the same thing. This is how, as a reader asks, we can keep consistent the idea that the maps we operate with not only reflect what is there but also what normatively should be there.
160
in which some particularly weighty expectations can feel coercive and intrusive, Jones notes that trust can feel the same. That is, sometimes we don’t welcome trust; sometimes trust can feel coercive.⁹ We trust maps, we depend upon maps, we rely on maps, we represent the world through maps, we use maps to navigate, maps impose structure on the world. Expectations function similarly. As tricky as it is to get a rough handle on expectations, thinking of expectations in terms of maps gives us something to grasp. A further upshot of the map metaphor is that thinking of expectations as maps can also help develop an ethics of expectations. After all, there are lots of ways a map can be wrong or used incorrectly. For example, to understand the disagreement between me and my mother the map metaphor can explain how we might be talking past each other. In coming out, in upending her expectations of me, she might think that I was tearing up the map that she had given me. The lack of a map to understand what was happening is why she feels unmoored, lost at sea. However, there is something else that could be happening. With the map metaphor we can make a familiar distinction between parents having expectations regarding ends versus expectations about the means to those ends. My mother’s end for me is to be happy and secure, and she had given me a map to ensure such a life. A map that marked out a path of all the conventional things. In rejecting those conventional things I’m not necessarily rejecting the map. Rather, I could just be showing her that there is a different path to that same goal. What she couldn’t yet see was that I could still end up where she wanted me to end up, i.e., living a happy life. My path would just be different. This now brings us to the wrongs of expectations. That is, if expectations are maps, we can ask about the right and wrong ways to draw and use maps. In some cases, the norms themselves map onto the functional role that the expectation is playing. However, when expectations play multiple roles things get trickier.
2. The Wrongs of Expectations To start, it might be tempting to simply extend the work done on the wrongs of belief to develop an account of the wrongs of expectation. However, too
⁹ See also Darwall (2017) for discussion on trust and expectation.
161
narrow a focus on belief will obscure just how treacherous the waters are. Instead, the route I will take here is to explore how the different functional profiles of expectations give rise to different kinds of mistakes, and in some cases, different wrongs. Before we dive in let me first note that I will not aim to give a unifying account of these wrongs here. Perhaps there is a general story to tell, but that goes beyond the scope of this paper. Sometimes expectations wrong in virtue of being false, sometimes in virtue of being morally offensive, sometimes in virtue of being alienating, sometimes in virtue of all three and more. I leave it as a project for future work on expectations to see if there is some general unifying story that can be told about how expectations wrong; the goal here is just to show that they can wrong and canvas some plausible explanations for how they might do so.
2.1 A mistaken view of the evidence The most obvious way that expectations can fail in performing their functional role is due to ignoring the evidence or an evidential mismatch between the expectation and the world. If we think back to the predictive function of some expectations, such expectations can suffer from a purely epistemic fault. That is, just as something goes wrong with belief when beliefs are insensitive to the evidence, something goes wrong with predictive expectations when they don’t match the world in the right way. Although we should expect some expectations, in particular, the expectations that function more like prescriptions or proleptic reasons, to have some resistance to counter-evidence, expectations cannot be completely untethered from the world. Expectations that are unmoored from reality wouldn’t serve the function of maps at all.
2.2 A mistaken view of morality Sometimes expectations can also involve a mistaken view of morality. That is, just as the world places constraints on what kinds of expectations are appropriate or inappropriate, so too might morality. For example, Dembroff and Saint-Croix (2019) argue that although there is a general duty to recognize and respect another person’s identity, such a duty can be defeated when the identity in question is a morally malicious identity. As Dembroff and Saint-Croix (2019, p. 590) note, “agential identities are (in part) ways of
162
being in the world. And we think that there are better and worse ways of being in the world.” The externalization of identities that can cause undue harm, they argue, should be neither encouraged, recognized, nor respected. Here we can use the phenomenon of doxastic wronging to make this point more forcefully. Doxastic wronging is the thesis that we can wrong others in virtue of what we believe about them (Basu and Schroeder 2019, Basu 2019 and 2021). Extending this thought to expectations, some expectations may be wrong in virtue of being morally problematic expectations in and of themselves. The argument for doxastic wronging begins with the recognition that beliefs are committal mental states. Notice that the same can be said for expectations because expectations also commit you to a particular representation. Any mental attitude that is a committal mental state is capable of wronging by relating the subject to the content of that attitude and thereby contributing to a morally harmful narrative. Notice that this applies equally to expectation as much as it does to belief.
2.3 A mistaken view of guidance To understand when proleptically functioning expectations not only go wrong but also can wrong we need to first understand why our attachment figures feel let down when those expectations are thwarted. That is, understanding why they feel let down can help to pinpoint what went wrong in holding that expectation. When we proleptically hold expectations of others we are doing two things. First, we are giving them directions of how to live their lives. Second, we are structuring our own lives with those same expectations. The expectations we have of others are part of our own maps, maps we use to help us make sense of the world and plan for the future. The weighty expectations that were the focus of section 1.3 involve our attachment figures extending their agency through us in this way and being disappointed when that extension fails. We can see the dissolution of this failure of extension in many stories of coming out. Our parents imagine various milestones we might meet based on societal expectations of success at those various stages, e.g., prom, graduation, marriage, children, etc. Our parents build plans around us, and they rely on us to successfully execute those plans. When we disrupt that, the future no longer contains what our parents desired and this naturally leads to feelings of disappointment, of feeling let down. Many parents successfully
163
reimagine a future for their children under new expectations, that is, they can redraw their maps. Many, however, fail. This kind of failure of reimagination or reconceptualization marks an expectation that is inappropriately held. Some expectations are also potentially alienating.¹⁰ Some expectations risk cutting at the very core of who we are and telling us that we’re unworthy. When expectations take this form, they can go wrong in a manner similar to how predictive expectations go wrong: these expectations are not, nor could they ever be, connected in the right way to who you are. The ends provided by these expectations could never be an end under which you could act. Potentially alienating expectations function similarly. The map not only fails to fit the world, but to make it fit we would have to change something fundamental about ourselves. It especially hurts when the expectations of our loved ones take this form because we depend on them as the first people through which we understand ourselves. To make this point about being connected in the right way more precise we can draw on Hilde Lindemann’s (2016) account of holding. What our attachment figures do is they hold us in personhood through the beliefs and expectations they have of us. As Lindemann argues, we become persons through our interactions with other persons. There are stories that get told about us that shape who we become. Our identities are, at least in part, narrative constructions made up of these beliefs and expectations others have of us. Holding, when it is done well, “supports an individual in the creation and maintenance of a personal identity that allows her to flourish personally and in her interactions with others” (Lindemann 2016, p. x). Our parents, and attachment figures more generally, engage in this practice of shaping and enabling our agency; they set ends for us before we’re capable of setting those ends ourselves.¹¹ When holding goes well we can flourish, when done poorly it can be destructive.¹² So, what are some ways in which this setting of ends, this holding, can be destructive? One way is when they involve an insidious co-opting of the
¹⁰ I use “alienation” here on purpose to draw an analogy to alienation as it appears in Williams’ (1973) critique of consequentialism, i.e., that consequentialism requires a kind of alienation from ourselves, that is, if we were to understand ourselves in the way consequentialism requires, we risk losing our grip on ourselves. See also Railton (1984). ¹¹ Compare also to Ebels-Duggan’s (2018) account of the liberal dilemma of childrearing. ¹² Compare this to what we saw in section 1.3 with Wonderly’s account of attachment. Namely, that it is through our attachments with others that those others have the power to shape our agency. Just as Wonderly’s account of attachment begins with our earliest attachments, e.g., the infant–primary caregiver relationship, in Lindemann’s account of holding our parents are the first to engage in this kind of holding. As I’ve noted, when holding goes well, we
164
mechanism of holding that undermines what is morally valuable about holding. We hold others in personhood to enable them to grow. When holding is done poorly it is restrictive and inhibiting. An example of insidious holding is when we instrumentalize others.¹³ As Quill Kukla (2020, p. 15) notes, there is a pernicious idea that the role of parents is to “create a specific type of person.” This way of thinking treats children as products, as objects that can be crafted and according to which our own worth and success can be measured. This way of thinking encourages morally unacceptable behavior like coercion and surveillance. Under such a picture of parenthood a good parent would take any means necessary to craft the perfect child. When we instrumentalize instead of providing scaffolding what is provided is more akin to internal fixation, i.e., a mechanical device made of metal plates, pins, rods, wires, or screws that is designed to fix broken bones and fractures. When expectation functions more like internal fixation than scaffolding the expectations aim at correction not guidance. Such expectations impose a structure on the agency of another, a structure that simply does not fit unless you break things first. We are vulnerable to a lot of narratives about our lives, but we are especially vulnerable to the narratives our parents tell about our lives. To end on a positive note, let me say something in the direction of how to avoid these wrongs by identifying a second way such expectations wrong: by not being justifiable to their subjects. To avoid these wrongs, we can employ a familiar Kantian distinction between treating someone as a means versus treating them merely as means. We are permitted to treat others as means, but not as mere means. The key difference turns on whether the other person can share in your ends, share in your goal, whether the other person could choose your end as her end. When we deceive and when we coerce, we treat people as mere means because we act under ends they could not agree to. My suggestion is that the good weighty expectations can be distinguished from the bad along similar grounds. That is, the appropriateness of interpersonal expectations rests on whether we can justify our expectations to one another. The ends need to be such that there’s the possibility that they could become coauthored. That is, we are not wronged when we find
can flourish, and when done poorly, it can be destructive. Similarly, our attachment relationships can either play stabilizing and balancing roles in our lives, or they can “cripple us in deep and devastating ways” (Wonderly 2017, p. 245). ¹³ This is of course not to say that instrumentalizing is always and in every case wrong.
165
ourselves in the following position with regard to those who hold us in personhood: if we can understand ourselves the way the other person understands us without risk of losing our grip on ourselves. I can’t imagine myself how my mother imagines me, and she can’t imagine me the way I imagine myself.
3. The Ethics of Expectations The route I took in this paper was to cash out an ethics of expectation in terms of the different functional profiles of expectations. We’ve seen that expectations play several roles and involve a rich constellation of attitudes. Key to expectations is that they have an anticipatory function and I cashed this function out with a familiar metaphor common to belief: the metaphor of a map. The map metaphor helped us not only to get a better grasp on the phenomenon of expectations but also to understand the various ways that expectations can not only go wrong but also wrong in and of themselves. That is, we now hopefully have a clearer picture of both what expectations are and how they can lead us astray, but perhaps not what unifies all these uses of “expectation” and the wrongs they’re capable of. But perhaps that’s not surprising. Given the multiple roles that expectations play, we’ll eventually find ourselves facing a dilemma. That is, sometimes the norms governing expectations will conflict. In this regard we find ourselves no worse off than we are in the case of the ethics of belief. It is commonly accepted that there are at least two competing norms governing belief—believe truth! shun error!—and we can’t fulfill them both. Just as we make a choice with regard to belief to either err on the side of believing even if there’s a chance of error or err on the side of not believing if the risk of error is too high, it seems we must do something similar with regard to expectation. For example, compare the requirement to avoid alienating a loved one through your expectations and the requirement to not hold a loved one in a morally malicious identity. What if it is really important to the person you love that they have this morally malicious identity, i.e., that they just are a bad person? In such a case, if you expect them to be otherwise, if you expect them to be better than they are, then you risk alienating them.¹⁴ We cannot simply stop expecting things of others because of the risks
¹⁴ See Yao (2020) for more cases that are like this.
166
involved, and the task of figuring out not only if there’s a unifying story of expectations but also how to balance the risks present in expecting is the task of an ethics of expectations. Although I hope we are now in a better position to understand the nature of and the wrongs of expectations, what I’ve said in this paper barely scratches the surface of what there is to be said about the ethics of expectations. For example, there are other routes available for developing an ethics of expectation. So, in this final section I wish to end by gesturing at other directions for fruitful research into the topic. First, how would our analysis change if we started by asking not what is the functional role of expectations, but by asking instead when is an expectation fitting? Second, the discussion of this paper has focused on interpersonal expectations, but we expect things not only of others but also of ourselves. Succeeding or failing at the expectations we set for ourselves is at least in one important sense up to us and this is different from the position we find ourselves in with regard to interpersonal expectations. Thus, it’s unclear to what extent what’s said here will generalize to the intrapersonal case. One might also worry about the language of wronging itself. For example, some expectations may not rise to the level of wrongs, but instead may be, in the language offered from Julia Driver (1992), suberogatory. That is, perhaps some expectations may be morally offensive but not themselves wrong. That is, some expectations may be rude, impolite, and violate other interpersonal norms without rising to the level of wronging.¹⁵ Relatedly, one might also question whether expectation ought to be treated as a deontic category in and of itself. As Horgan and Timmons (2022) note, the Greek “deon” from which “deontic” and “deontological” derive, simply means “that which is binding.” And expectations, as we have seen, are certainly binding. In their discussion of gratitude, Horgan and Timmons note that in gratitude we often feel the urge to return the favor. However, in such cases, there’s no real obligation that we must do so: one cannot demand that another return a favor, but often it is something that’s nonetheless expected, and it can have a binding force on our will. So, in gratitude there is anticipation of a future opportunity to, in some sense, pay back the kindness that one was shown. And in that anticipation there is something like a moral expectation. And finally, when thinking about the nature of expectations the general notion of expectations can be multifaceted in more ways than outlined in
¹⁵ See also Calhoun (2016) and Martin (2019, 2021).
167
this paper. For example, Rebecca Keller (manuscript) offers an account how expectations function in perception. Keller cashes out perceptual expectations in terms of the functional role they play in our visual system whereas here I have been focused on how expectations function within our interpersonal relationships. While tangential to the kinds of expectations under consideration in this paper, any full account of the general phenomena of expectations would need to consider not only how expectations function interpersonally and intrapersonally but also in other aspects of our lives. All of these questions, and likely more, fall under the broad topic of the ethics of expectations. But now you might begin to wonder why am I telling you about all the things I could have told you about, but didn’t. My reason is to demonstrate just how complicated the question we set out to answer is. Expectations operate in multifaceted ways and it is not always clear if some of the ways we talk about expectation are at all similar to other ways we talk of expectations. I hope to have made some progress in untangling part of the mess, but there is still much more to do. In closing I should return to my mother. We’re doing better now. She complimented my rice and recently added me to the extended family WhatsApp group.¹⁶
¹⁶ My brother has been trying to get everyone to switch to Signal, but to no avail. Jokes aside, I have a lot of people I need to thank. I first started thinking about this paper after my interview at CMC where I was asked a question about why my dissertation was narrowly focused on belief when there are many other attitudes that seem capable of wronging. It was a good question that deserved a better answer than whatever I managed at the time and I’ve been unable to stop thinking about it ever since. I have to thank the Gould Center for Humanistic Studies for funding a humanities lab on the topic of expectations which allowed me to start to work out some of these ideas, and the students who participated in that lab for their feedback. This paper has been presented under a variety of titles, including “The Weight of Expectations” and “The Parent Trap: Does Doxastic Wronging Start at Home?”, and this current version has benefited from feedback from audiences at the Washington University St Louis Workshop in Ethics, PeRFECt4 at the University of Pennsylvania, Indiana University, MIT, the Social (Distance) Epistemology series, the Oxford lockdown epistemology group, Cal Poly Pomona, Queen’s University, the Central APA, the Center for Ethics at the University of Toronto, Oakland University, Trinity College Dublin, UT Austin, University of St Andrews, Yale, and the Arizona Workshop in Normative Ethics. Also a special thanks to two anonymous readers for this volume and Cory Davia, Jenna Donohue, Maegan Fairchild, Amy Flowerree, Georgi Gardiner, Laura Gillespie, Sukaina Hirji, Paul Hurley, Liz Jackson, Gabbrielle Johnson, Zoë Johnson King, Renée Jorgensen, Adrienne Martin, Michael McKenna, Mukasa Mubirumusoke, Mark Timmons, Sumeet Patwardhan, Cat Saint-Croix, Kenneth Silver, Drew Schroeder, Mark Schroeder, Dion Scott-Kakures, Julie Tannenbaum, and Peter Thielke for taking the time to talk through many of these ideas with me and in some cases immensely helpful written feedback as well.
168
References Anscombe, G. E. M. (2000). Intention. Harvard University Press. Basu, R. (2019). The Wrongs of Racist Beliefs. Philosophical Studies, 176 (9):2497–515. Basu, R. (2021). A Tale of Two Doctrines: Moral Encroachment and Doxastic Wronging. In Lackey, J., editor, Applied Epistemology, pages 99–118. Oxford University Press. Basu, R. and Schroeder, M. (2019). Doxastic Wronging. In Kim, B. and McGrath, M., editors, Pragmatic Encroachment in Epistemology, pages 181–205. Routledge. Bicchieri, C. (2016). Norms in the Wild: How to Diagnose, Measure, and Change Social Norms. Oxford University Press. Breakey, H. (2022). When Normal is Normative: The Ethical Significance of Conforming to Reasonable Expectations. Philosophical Studies, 179:2797–821. Calhoun, C. (2016). Common Decency. In Moral Aims: Essays on the Importance of Getting it Right and Practicing Morality, pages 103–22. Oxford University Press. Callard, A. (2018). Aspiration: The Agency of Becoming. Oxford University Press. Darwall, S. (2017). Trust as a Second-Personal Attitude (of the Heart). In Faulkner, P. and Simpson, T., editors, The Philosophy of Trust, pages 35–60. Oxford University Press. Dembroff, R. and Saint-Croix, C. (2019). “Yep, I’m gay”: Understanding Agential Identity. Ergo, 6. Driver, J. (1992). The Suberogatory. Australasian Journal of Philosophy, 70:286–95. Ebels-Duggan, K. (2018). Freedom and Influence in Formative Education. In Schmidtz, D. and Pavel, C. E., editors, The Oxford Handbook of Freedom, pages 309–404. Oxford University Press. Holguín, B. (2022). Thinking, Guessing, and Believing. Philosopher’s Imprint, 22(6):1–25. Horgan, T. and Timmons, M. (2022). Expressing Gratitude as What’s Morally Expected: A Phenomenological Approach. Ethical Theory and Moral Practice 25:139–55. Jones, K. (1996). Trust as an Affective Attitude. Ethics, 107(1):4–25. Keller, R. (Manuscript). Perceptual Expectation. Kukla, Q. R. (2020). Taking Children’s Autonomy Seriously as a Parent. APA Newsletter Feminism and Philosophy, 19(2):14–17.
169
Li, C. (1997). Shifting Perspectives: Filial Morality Revisited. Philosophy East and West, 47(2):211–32. Lindemann, H. (2016). Holding and Letting Go: The Social Practice of Personal Identities. Oxford University Press. Martin, A. (2019). Obligations of Gratitude: Directedness without Rights. In Roberts, R. and Teloch, D., editors, The Moral Psychology of Gratitude, pages 59–82. Rowman and Littlefield. Martin, A. (2020). Interpersonal Hope. In Bloser, C. and Stahl, T., editors, The Moral Psychology of Hope, pages 229–48. Rowman and Littlefield. Martin, A. (2021). Personal Bonds: Directed Obligations without Rights. Philosophy and Phenomenological Research, 102(1):65–86. McGeer, V. (2008). Trust, Hope and Empowerment. Australasian Journal of Philosophy, 86(2):237–54. Railton, P. (1984). Alienation, Consequentialism, and the Demands of Morality. Philosophy and Public Affairs, 13(2):134–71. Ramsey, F. (1990). General Propositions and Causality. In Mellor, D. H., editor, F. P. Ramsey: Philosophical Papers. Cambridge University Press. Schroeder, M. (2020). Treating like a Child. Analytic Philosophy, 63:73–89. Williams, B. (1973). A Critique of Utilitarianism. In Utilitarianism: For and Against. Cambridge University Press. Wonderly, M. (2017). Love and Attachment. American Philosophical Quarterly, 54(3):232–50. Yao, V. (2020). Grace and Alienation. Philosopher’s Imprint, 20(16):1–18.
8 Satisficing Virtuously Earl Conee
1. Introduction Here is a standard act consequentialist theory of morally permissibility: (MC) Alternative A is morally permissible iff the outcome of taking A would have no less moral value than would the outcome of taking any alternative to A. MC asserts a maximizing requirement for permissibility. Maximizing is far from inevitable. Generally, permissibility does not require a comparison, much less a maximum. No maximizing is required for something to be permissible as a matter of etiquette, or grammar, or game playing, or legality. Concerning specifically moral evaluations, Kantian moral theories do not require maximizing for permissibility, with no appearance of a resulting incoherence. These observations about permissibility give us a reason to consider versions of act consequentialism that do not require maximizing. Satisficing act consequentialism (SC) is the prominent non-maximizing form of act consequentialism. SC theories count alternatives as morally permissible because they are good enough. Versions of SC differ on what makes alternatives good enough. SC gains considerable further interest from some objections to MC.¹ The objections support the permissibility of sorts of conduct that seem good enough. Our main question here will be about how best to respond to those objections: What is the best theory that overcomes them? We should distinguish between two projects. One of them is to employ the objections that suggest satisficing to find the best act consequentialist ¹ Most of the objections to be discussed here are familiar from the literature, including (Bradley, 2006), (Hurka, 1990), (Mulgan, 2001a, 2001b), (Rogers, 2010), (Slater, 2020) and (Vallentyne, 2006). Earl Conee, Satisficing Virtuously In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Earl Conee 2023. DOI: 10.1093/oso/9780198895909.003.0009
171
moral theory. The other project is to seek the best moral theory that we can develop, when we set out with the aim of finding a plausible way to overcome the objections to MC that suggest a satisficing response.² We will undertake the second project here. A theory emerges that adequately avoids the objections that suggest satisficing. But we will find that we can do better. The theory has two sufficient conditions for moral permissibility. That makes trouble. The trouble is avoided by dropping one of the conditions.
2. Problems for MC a. One well-known problem for a maximizing theory is that it makes no room for the morally supererogatory. MC implies that not taking the best alternative is never permissible, however much the best requires of the agent in effort, risk, cost, sacrifice of what is most treasured, and anything else. The obligatory alternatives are the ones that it is impermissible not to take. So, MC implies that the best alternative is always obligatory. This classification by MC is not indefensible. The obligatory is distinct from the praiseworthy. It can be argued that in the cases where what is obligatory is heroic, the conduct becomes so praiseworthy as to be easily confused with being supererogatory. Also, the morally obligatory is distinct from what is obligatory to fulfill some role, like that of a good citizen. It can be argued that doing better than is required by the duties of good citizenship, good parenting, or the like, can be easily confused with doing something morally supererogatory. These ways to defend MC are not nearly conclusive. There is heroic conduct that seems quite clearly to be better than what is moral necessity. Among the clearest examples are some besieged battlefield rescues of wounded soldiers. The courageous exertions of the rescuers are aweinspiring. In many cases, calling for evacuation support instead would clearly have been safer and entirely honorable, though greater suffering ² There are other potentially worthwhile projects in the vicinity. For instance, we could try to find a consequentialist theory that did not invoke satisficing but responded to objections in response to which it has been invoked. For example, we might seek a suitable version of rule utilitarianism. An initial worry about the prospects of that project is that it must answer the objections in a way that requires maximizing some moral value under some circumstances, if the theory is not to employ satisficing. Yet as will soon be clear, some reasons invoked to satisfice seem to go directly against any moral need to maximize, no matter how indirectly or hypothetically it occurs. See notes 3 and 4 below.
172
would have resulted from the delay. Counting acts like these recues as obligatory, rather than supererogatory, entirely on the grounds of their having the best available outcomes, seems to be just a mistake. For one thing, it belies the manifest humility of the heroes who say that they “had to do it.” It is not credible that their acts were morally necessary. b. There are possible cases in which the agent must make some major sacrifice if the agent is to gain a best outcome, while that outcome is very little better than the agent could achieve with no sacrifice by anyone. The outweighing value might be spread as thinly as the value allows. Those who lose out on the greater value need never be troubled; indeed, they need never have a clue that they might have been better off. For one example, we might find ourselves confronted with a stark choice that cannot be modified or postponed. The choice is to impoverish ourselves now by donating nearly all we have to a worthy cause, or to commit now to donating our remaining financial resources later, as a bequest. If we choose neither, no good will come of that. We foresee the following about our donating now: The worthy cause would distribute our donation very widely, making a small positive impact on each of many lives, and our impoverished life would be very difficult for us. We foresee the following about our making the bequest: Our remaining life would be financially comfortable, the bequest would be a smaller amount, the worthy cause would use the bequest to give smaller though still noticeable benefits to the same people, the bequest would have a slightly worse outcome overall, and those who received the smaller benefits would not have any reason to believe that they might have gotten larger ones. It is clearly permissible to take the maximizing alternative for the sake of some such slightly better outcome when it has some such costly impact upon ourselves. This self-sacrifice is supererogatory, though. Yet MC does not permit us to take the next best alternative in this sort of case (or any other). To think that morality does not allow the next best choice, even in the most extreme versions of the sort of case just described, seems to depend on overestimating the moral significance of moral value. An SC theory can count maximizing in this sort of case as going beyond good enough and consequently as morally supererogatory. c. There are cases of everyday good conduct that is thoroughly nice and harmless. Such conduct seems morally permissible when it is not quite the best that could have been done. For instance, suppose that a newcomer to our group is warmly welcomed by us. The newcomer appreciates our hospitable treatment. Our warm welcome seems clearly to be morally
173
permissible, even when a still more enthusiastic welcome would have required our taking just a little more trouble and would have been a bit better overall. Yet according to MC only the more enthusiastic alternative was permissible for us.³ Again, an SC theory can count the warm welcome as good enough. An SC theory enables us to count the maximizing as mildly supererogatory and not morally obligatory. d. MC implies that it is morally obligatory for agents to optimize when this slightly furthers their own interests over otherwise equally good alternatives. Yet it does not seem always to be morally obligatory to do as well as one can for oneself, if that is what makes an alternative morally best by some small margin. (Rogers, 2010) For example, suppose that we could have minimally enhanced our level of well-being further by having choosen a slightly tastier dessert, and this difference is all that makes that choice morally best. It would have been a bit nicer for us, but it does not seem to be morally mandatory.⁴ The problem cases of this sort do not depend on value theories that boost alternatives by insignificant agent enjoyments. Insignificant agent costs make the same point. Suppose that employing relaxation method R1 would make us good at disengaging from some source of psychological discomfort, say, some awkward social setting that arises from time to time. Employing method R2 would be a bit easier for us and otherwise just as good. But learning method R2 would be marginally more difficult and does not appeal to us. All in all, we would be slightly better off to learn and employ R2. Suppose that knowing all of this, we choose R1. Though this choice would be slightly imprudent, it does not seem to be any sort of moral failing. Yet the choice does not maximize moral value. So, by MC it is morally impermissible. The general point made by such cases is that the moral value that results from serving one’s own well-being seems to be of limited moral significance, especially when it has negligible impact on others. Perhaps helping oneself considerably is worth something morally, as are considerable costs to oneself. But when a choice maximizes partly because of a tiny boost to the agent’s own well-being, and that further gain is of no concern to the agent, it
³ This is one sort of objection to MC that appears to oppose any moral need to maximize moral value, even indirectly or hypothetically. ⁴ This is another objection to MC that is difficult to reconcile with any sort of requirement to maximize.
174
does not seem that maximizing is morally obligatory. By implying this to be obligatory a moral theory like MC seems to deprive autonomy of even marginal moral significance, not for good reason, but just for the sake of tiny differences in value that do not concern their potential recipients.⁵ e. A further attraction of SC is its application to cases where value cannot be maximized because the agent has ever-better alternatives.⁶ Having infinitely many alternatives seems to exceed physical limits of the likes of us. Minds like ours seem unable to distinguish actionably each of infinitely many choices. But some possible agents have this capability and can have this choice. In such cases, there are some clear moral boundaries. MC’s implication for such cases is that no choice is permissible, since each has an outcome with less moral value than that of an alternative (in fact, each outcome has less value than that of infinitely many alternatives). MC’s implying the impermissibility of every choice places MC on the wrong side of the clear moral boundaries. It would be permissible for someone who had such a choice available to make an earnest effort, within whatever was the temporal limit of the choosing, to do as much good as the person found some way to do, even though any such choice would do infinitely less good than was available. This permissibility is affirmed by versions of SC according to which making such an effort is a good enough.
3. Satisficing Proposals Despite such attractions of satisficing theories, the attempts to specify acceptably what is good enough have run into serious trouble. Here is a rapid review of some proposals, largely from the literature, followed by problems for them, also largely from the literature. a. The good-enough threshold can be one-and-the-same outcome value for any situation, that is, the threshold can be some good-enough condition of the world after the act. Or the good-enough threshold can be specified as a condition after the act that is determined by some minimal level of well⁵ This permissibility requires less of a departure from MC than would be needed to make room for deference to an agent-centered prerogative, and it requires much less than would be needed to make room for deference to agent integrity. The modesty of the permitted losses to maximizing by allowing some narrowly limited leeway for agent values enhances the plausibility of this sort of reason not to require maximizing. ⁶ Such cases are described in (Bradley, 2006).
175
being of everyone who might be affected by any of the agent’s alternatives. As Ben Bradley has argued (Bradley, 2006), any such threshold permits gratuitous value obstruction. In fact, when the world, or those who can be affected, are already above the threshold, the theories permit gratuitous and viciously harmful reductions in welfare down to the threshold. Those permissions seem both drastically mistaken about morality and drastically contrary to the consequentialist perspective on what makes acts moral. When the fixed outcome value is unattainable, no alternative is good enough and so none is permissible. That too seems drastically wrong. b. The good-enough threshold can be set at some positive level of contribution by the act. But then, as Bradley has argued (Bradley, 2006), an agent who makes that positive contribution might thereby block a much better outcome. The satisficing theory counts this as acting permissibly even when the obstruction is done out of ill will. For example, suppose that X tells Y a good story and this adds sufficient positive value. Still, if X tells the story maliciously to distract Y from catching Y’s flight, at great cost to Y, then the added value does not make X’s act permissible. Contributing some amount of positive value does not imply permissibility. c. The good-enough threshold can be set at some level of burden to the agent. The burden might be determined by costs to the agent’s well-being, or by costs to the agent’s priorities, or by the needed effort. First let’s consider the burden as a reduction of the agent’s well-being below some specified level. This sort of SC permits the agent not to maximize when that would have an outcome for the agent that goes below the specified level.⁷ (Turri, 2005) This version of SC has trouble no matter where the fixed level is set. If it is set at mild cost to the agent’s well-being, then it permits not acting to prevent an open-ended amount of harm, whenever so acting would have any slightly greater cost to the agent’s well-being. That seems to ask too little of the agent. The required cost to the agent’s well-being that permits not maximizing might be set very severely. It might be required that maximizing would have the agent’s well-being so damaged that it might be thought that no cost to others is enough to make it morally obligatory for the agent to take an alternative that is that self-destructive. If satisficing is allowed only when the agent cost is that great, then maximizing will be required whenever the agent ⁷ Turri holds that only such sacrifices are permissible alternatives to maximizing, but perhaps not all of them.
176
cost is slightly less. Yet it can happen that the benefits of a still-great agent sacrifice that is required to maximize are minimally greater than the benefits from not sacrificing at all. When things are like that, requiring maximizing seems to go beyond what is morally necessary. The threshold cost to the agent of maximizing could be set somewhere in between these extremes. But then in some cases agent costs just above the moderate level would spare so much harm to others that maximizing seems morally obligatory. In other cases, maximizing at an agent cost just below the moderate level would make a difference to others that was too small to be worth the cost to the agent. Morally requiring the cost would be overly burdensome. In short, no one inflexible level of cost to well-being is plausibly thought to be the cut off for all choices. Allowing sub-optimal choices when there is some set cost to the agent’s priorities is similarly problematic. Small costs to what the agent values seem not to be enough to render permissible not maximizing when the maximum prevents much harm. Large costs to the valued divide into two types: those that render very low the agent’s level of well-being and those that do not. We have already discussed a problem for a threshold of high cost to agent wellbeing, whether from a loss of valued things or otherwise. Large costs to the valued that do not make for high cost to well-being include things that are unreasonably valued and costs that do not impact the agent because the agent is unaware of them or otherwise unaffected by them. It seems clear that none of those costs justify not maximizing. If things that are reasonably valued by the agent are destroyed or prevented by the maximizing act, but not to the extent of giving the agent a very low level of well-being, then there will be cases where this asks too much of the agent because of the small gain of maximizing over a cost-free alternative. There will be cases where this asks too little of the agent because the moderate costs are needed to accomplish a maximum without which there would be terrible costs for many. d. Richard Chappell advocates a version of satisficing that permits not maximizing when it would cost the agent too much exertion of “willpower” in a technical sense of the term. Chappell’s “willpower cost” is a function of two factors: the agent’s level of altruistic concern and the extent of mental effort required to perform an act. The altruistic concern is a matter of valuing others’ welfare. Chappell assumes that we have some independent understanding of how much altruistic concern is required for someone to act in accordance with having adequate moral concern for the situation. The
177
notion of “the effort ceiling” is explained in terms of this adequate moral concern. When an act is at the effort ceiling, the act is like this: it requires as much mental exertion as does one of the minimally mentally effortful acts that would implement an adequate moral concern in the situation. Here is my formulation of Chappell’s willpower satisficing consequentialism: (WSC) Taking some alternative, A, is morally permitted exactly when A’s outcome is at least as good as a best alternative that is at the effort ceiling. Thus, by WSC a sub-optimal alternative is permitted whenever it is as good as some alternative that expends enough mental effort to implement adequate moral concern in the situation. (Chappell, 2019) A problem for WSC arises because the content of moral concern is taken to be limited to valuing others’ welfare. This misclassifies cases where an agent is unconcerned about other morally important factors. For instance, if A’s concern for others’ well-being is unaccompanied by a concern for justice, A is permitted by WSC to serve others’ well-being at a great cost to justice, when it would take effort a bit above the effort ceiling for A to take a justice-enhancing alternative that resulted in slightly less gain to others’ welfare. WSC can be adjusted to consider moral factors that need not enhance welfare, such as justice and honoring commitments. The adjustment would have to be drastic, though. WSC implies the permissibility of accomplishing the best among any non-maximally morally valuable alternatives that are at an effort ceiling. Taking separate account of moral factors that are not contributors to moral value would require that the effort ceiling not be set by this value comparison. Rather than a best outcome, some new moral basis would have to be found for specifying the effort ceiling. It is not clear what that might be. In any event, there is a different problem for WSC. Its sub-optimal permissions are awarded on the basis of a ceiling that is set by mental effort. This is a strange candidate for a burden to the agent that justifies satisficing. On its face, undergoing some mental strain to make a moral choice seems to be a light burden. For example, a brief mental strain in deciding to do something generous does not seem to be much of a burden. If nothing else troubling to the agent is needed to accomplish the best, the mental effort seems a small cost rather than a burden that is permissibly avoided. Extreme exertions by the agent, mental or physical, are more plausible candidates for allowing sub-optimal alternatives. Yet such a basis for
178
permitting sub-optimal conduct allows agents who are little troubled by such exertions to neglect much better outcomes that require going above the threshold. More plausible still is a threshold that is set by a drastic reduction in well-being.⁸ Chappell observes that a hybrid ceiling might be constructed by combining mental effort with cost to well-being. He rightly notes the difficulty in finding a convincing way to do this. The more basic doubt remains, though, that effort plays any separate role in rendering permissible sub-optimal alternatives. e. It can be proposed that the good-enough threshold for permissible sub-optimal conduct is doing something that avoids blameworthiness. This could not be quite right. Even fully excusable moral wrongs are not permissible, though they are not blameworthy. So, avoiding blameworthy conduct is not always enough to avoid wrongdoing. In any event, the candidate threshold is that of avoiding moral blameworthiness. Using an explicitly moral evaluation of alternatives in a theory of what determines the moral evaluations of alternatives is entirely uninformative on the very topic on which we seek information. It gives us no information about what makes for the morality of an evaluation. f. A good-enough threshold can be described so that an alternative, A, is permissible if A is like this: any morally better alternatives to A are insufficiently better in comparison to their costs to the agent, this insufficiency being determined in a way that can vary with the context of the alternative.⁹ One notable potential asset of this proposal: The contextually determined variation in what is insufficiently better might avoid the problems discussed
⁸ Chappell poses objections to having agent well-being costs be the burden that justifies satisficing. A main reason he offers for counting mental effort as the justifying burden is that a mental effort account explains how prudentially beneficial alternatives can be burdensome to adopt. (254) But although a mental effort burden does explain what is burdensome about some alternatives that are known to be prudentially better, mental effort is not a good excuse for neglecting a better alternative. The need for mental effort—“It’d be very hard to get myself to do it”—does not give a good reason to neglect alternatives that the subject knows to be more prudent (allowing exceptions when the needed effort is seen to be so difficult as to override the greater prudence of taking the alternative). The burden imposed by the concerted mental effort needed to take morally better alternatives seems similarly inadequate. “It’d be very hard to summon up the great resolve that I’d need to get myself to do better”—does not seem to be a good reason not to do it. So, in neither case does the burden seem to justify sub-optimizing. At most, the mental effort required would justify taking a sub-optimal alternative only when substantially less effort would result in a small reduction in the moral value of the outcome, and not clearly even then. ⁹ This is my wording of a proposal by an anonymous reviewer. I am grateful for the suggestion.
179
in section c above for permitting sub-optimal conduct that does not exceed some fixed level of burden to the agent. An insufficiently-better proposal, with varying insufficiency, points to a type of satisficing theory of permissibility. It is the type in which for an alternative to exceed some varying difference between the moral value of the alternative’s outcome and its cost to the agent is enough to keep the alternative from being obligatory. But what determines where that threshold is? No basis for permissibility is stated until the factor is specified that determines the insufficiency in any given context. That specification has not been done. We do not yet have all the ingredients of the theory.¹⁰ In any event, no agent-cost ceiling on maximizing would solve all the discussed problems that suggest satisficing. An agent-cost ceiling does not imply the permissibility of routine instances of our being very nice, when our being even nicer would not be costly enough to us to hit the ceiling. It does not imply the permissibility of our doing slightly less well by ourselves than we could. It does not imply the permissibility of our making some possible choice in the ever-better alternative cases where doing better would not be costly to us.
4. Taking Stock Notably, among the sub-optimal thresholds that we have reviewed only the initial candidates award permission entirely because of the values of outcomes. The other satisficing theories depart from pure consequentialism. They limit the requirement to maximize because of costs to the agent that are already counted as diminished moral value by a maximizing theory. The impurity of the consequentialism is a just classification though, not an objection. In fact, the departure opens the way for a more promising ¹⁰ The reviewer adds this: “something grounds this permission - but thinking about what’s sufficiently costly is an ideal permissibility indicator.” This does not tell us what we need for us to have the moral factors in the theory. It tells us that the threshold’s identity is conceived in some thinking about a “sufficient cost.” But we still do not have an explicit identification of the sufficiency factor. What variation makes for the permission? It is possible that the sufficiency in the thinking can be made fully explicit like this: “an alternative is sufficiently costly in comparison to its outcome value that the alternative is not morally obligatory.” But if that is the thought, then a moral evaluation that is not informatively different from the theory’s topic is explicitly made in the thought. We seek a statement of a satisficing theory of moral permissibility. As a result, to use this expression of the thought content to specify within the theory the insufficiency that is asserted by the theory would make the theory uninformative. We remain in need of some usable identification of what determines the contextually varying insufficiency.
180
candidate. The threshold problem for specifying an acceptable SC admits of a general diagnosis. Plausibly it arises because moral permissibility is intuitively implied by morally decent conduct. This does not translate readily into any one sort of value threshold or burden threshold. We could set the permissibility level at moral decency itself, stated in just those terms. So specified, though, the permissibility condition is decidedly vague. Also, a vicious circularity threatens. A “morally decent” specification of permissibility raises the suspicion that “morally decent” here is short for something like “morally sensitive enough to be permissible.” It would be better to use more illuminating notions to identify a threshold. If, as it seems, permissibility is implied by moral decency, then maybe a successful and illuminating way to state the threshold will cast light on moral decency.¹¹
5. Virtue to the Rescue The most promising approach to sub-optimal permissibility appeals to virtue.¹² More specifically, the most promising theory finds permissibility in the alternatives that a virtuous person might take. The nature of being a virtuous person can get a fairly standard understanding. A virtuous person is someone with no vices who has some complete range of classic virtues. For definiteness we can assume that those virtues are honesty, moderation, justice, courage, compassion, benevolence, generosity, prudence, and practical wisdom. It is not crucial for present purposes exactly which virtues a virtuous person has. It is crucial that they include some primarily ¹¹ This is a genuine “maybe.” It may be instead that a conceptually revealing threshold is a sufficient condition for permissibility, but it is not an explanation of moral decency. See the end of this section and the beginning of section 8, where moral decency is further discussed. ¹² Kantian versions of moral decency can be constructed. Kantian moral views do not require value maximizing. So, they are not immediately excluded from accounting for reasons to satisfice. A serious investigation of Kantian views to accommodate reasons to satisfice would be a very extensive undertaking. There is an initial concern worth mentioning about their prospects. Kantian morality includes absolutes. Core Kantian ideas include absolutely prohibiting acting on a maxim that cannot be universalized, and absolutely prohibiting a mistreatment of a rational being. Moral absolutes like these are not in keeping with the reasons to permit satisficing. Broadly speaking, the reasons argue for moderation and compromise among morally significant factors. Going all out is almost never required; self-sacrifice is only sometimes required; variously strenuous efforts are sufficient for various choices, and so forth. The concern is that the Kantian absolutes will turn out to block the variations that the reasons call for. For instance, it seems possible for morality to require mistreating the rational being who is oneself, to prevent sufficient evil, even though avoiding serious self-harm is usually permissible at some cost to overall welfare. A Kantian absolute prohibition on such mistreatment obstructs accounting for this possibility.
181
self-regarding virtues, such as prudence and moderation. Traits like these temper value-maximizing inclinations from some primarily other-regarding virtues of the virtuous, such as benevolence and generosity. This inclusion is not skewing the virtues to serve present purposes, though. Michael Slote has argued with great plausibility that such self-regarding traits are among the virtues, by the test of their being admirable.¹³ Borrowing from virtue theories of right action, we can count alternatives as virtuous by the following specification: (AV) Alternative A is virtuous iff a virtuous person might characteristically choose A from A’s alternatives.¹⁴ Now we are ready for a theory that has both the consequentialist permissibility condition of maximizing moral value, and a satisficing permissibility condition of taking some virtuous alternative. Here is virtuously satisficing consequentialism: (VSC)
A is morally permissible iff A maximizes moral value or A is virtuous.
VSC overcomes the problems that we have seen for satisficing theories, while accommodating our initial reasons to allow satisficing. First, VSC provides for supererogatory acts (Problem 2a above). The virtuous would not be negligent or cowardly. But they would not always be heroic. Sometimes in a crisis a virtuous person would not go beyond making a minimally virtuous response. A virtuous person could take a responsible and helpful course of action out of a benevolence that is tempered by prudence and moderation. Some who are virtuous would be heroic, acting in ways that were primarily exercises of courage and compassion. Such acts are enough better than the minimally virtuous response to be supererogatory.
¹³ See Sections 12–15 in Slote’s contribution to (Baron, Pettit, and Slote, 1997). ¹⁴ AV is somewhat simplified. Three details are worth mentioning here. First, the “might” allows a multiplicity of permissible alternatives, on the assumption that additional and varying features of the virtuous, other than their virtues, enable them to decide differently from among the same alternatives. Second, the virtuous need not actually, or even possibly, have the alternatives. It is enough that someone virtuous could be in the epistemic position of someone who did have them and that someone virtuous in that position could choose the alternative. And third, a choice is made “characteristically” by someone virtuous, whether or not the choice exercises virtue, if the choice does not neglect to exercise virtue. For more details, see (Hursthouse, 2000) and (Conee, 2014).
182
Next, VSC allows everyday cordial conduct to be permissible when it is sub-optimal (Problem 2c above). In our example of welcoming a newcomer to a group, a virtuous participant in the welcome might be quite nice without being maximally good. Next, VSC does not require maximizing up to some fixed value, either for the outcome as a whole or for the affected individuals (Problem 3c above). Sometimes maximizing choices are highly imprudent with only a minor gain in resulting moral value. In such cases a virtuous person might not maximize. So, by VSC not maximizing would be permitted. Furthermore, VSC does not allow gratuitous value obstruction above its threshold. It could not be virtuous to prevent something better gratuitously. Moreover, VSC does not permit neglect of a much better alternative at anything less than a large cost to the agent’s well-being. The virtuous would incur modest costs, for the sake of doing much better, when this could be judiciously determined. If the cost to agent well-being of an alternative were huge, with little gain in overall moral value, at least sometimes the virtue of prudence would have someone virtuous not taking the alternative. For the sake of outcomes that are plainly much better, the generosity of the virtuous might have them sacrificing treasured things. Still, sacrifices of such things would not always be made. Instead, some who are virtuous would respond to opportunities to gain slightly better outcomes, but at the cost of losing things that they value highly, by prudently passing up those opportunities. For the sake of outcomes that are plainly much better, the generosity of the virtuous would have some of them making mentally or physically strenuous efforts. At the same time, some who are virtuous would respond to some chances to gain slightly better outcomes, but only from strenuous efforts, by prudently passing up those chances. A tiny benefit to the agent can be all that is better about the best alternative (Problem 2d above). The virtue of moderation would have some who are virtuous sometimes declining the benefit. So, VSC does not require maximizing here. There is the problem for MC that arises in cases where an agent has infinitely many alternatives and none is best (Problem 2e above). The virtuous would choose generously. Whatever they did, there would be infinitely many better alternatives that they could have chosen. That is an unavoidable feature of the choice. The virtuous would not be stymied by that, since neglect to choose generously would be an inexcusable failure of virtue. The virtuous would avoid any alternative that did not have an outcome that was of large value by any reasonable standard. In the most challenging versions of infinite
183
choice situations, where the potential recipients are morally equal, being unfair and arbitrary might be inevitable. Again, the virtuous would be undaunted. The virtuous could do no better than to avoid choices that were irresponsibly unfair or arbitrary. In the most challenging versions of the infinite choice, it would be enough for the choice to be permissible that it met the constraints of being generous, making a large contribution by any sensible measure, and not being irresponsibly unfair or arbitrary. Again, the notion of being morally decent seems to give us a sufficient condition for what is morally permissible. VSC’s virtue disjunct seems to provide an informative elaboration on what we conceive of as moral decency. The elaboration on this notion that VSC suggests is that exercising moral decency is constituted by engaging in virtuous conduct. To the extent that this is a plausible account of moral decency, the sufficiency of the decency for permissibility supports the sufficiency of VSC’s virtue disjunct. In any case, unlike moral decency conceived as such, the virtue condition of VSC clearly avoids vicious circularity.
6. Simplifying Advantageously The goal so far has been to find a moral theory that can be well defended against certain objections. It plausibly avoids objections to MC that suggest the permissibility of some sort of satisficing. It has just been argued that VSC is like that. If so, then we have found what we have been looking for. We would not do well to stop with VSC, though. The theory is disunified. This can be avoided by a simpler theory. Also, the remaining consequentialist disjunct in VSC has troubling implications that the simpler theory avoids. The maximizing disjunct of VSC affects what the theory counts as permissible just when a maximizing alternative is not something that someone virtuous might do. There are such cases. In some such cases someone acts viciously, and the conduct maximizes accidentally or incidentally. The vicious agent might even attempt to avoid the benefits of the maximizing outcome of the deed, out of sheer malevolence, but happen to fail in this attempt. Counting this conduct as permissible is a problem for VSC. Such alternatives seem wrong. We should note that there are somewhat similar cases of unintended maximizing with a different intuitive moral status. Someone does something routine that is intrinsically benign, such as knocking on a door or turning on a light. It is done with no moral concern, and it happens to maximize. This is typically permissible, as the maximizing disjunct of VSC allows.
184
Such conduct satisfies the virtue disjunct of VSC too, though, if the virtuous might have characteristically done so under the circumstances. And they might have. Conduct by the virtuous is characteristic if it does not neglect anything that virtue requires. Given that there are cases of appallingly vicious conduct that accidentally or incidentally maximize, though, the universal permissibility of taking any maximizing alternative depends on maximizing being morally acceptable, no matter what else is involved. There are familiar objections to maximizing always being permissible. They include cases of betrayals and injustices with a minimal gain in value over benign alternatives. Of special interest concerning VSC are problem cases in which the maximizing is the farthest thing from acting virtuously. Possible cases exist in which a thoroughly vicious act happens to maximize. For instance, suppose that a sadistic horrendous killing induces subsequent changes in security procedures. These changes prevent many lesser harms, so many that the act has a better effect on balance than the killer could otherwise have accomplished, though this is an effect that the killer would have opposed if it had been anticipated. The killing maximizes. Yet it seems clearly to be morally impermissible. Apart from their force as counterexamples to the sufficiency of maximizing, such cases show that the two disjuncts of VSC induce a severe incongruity in the theory. VSC counts as permissible both some entirely virtuous conduct that happens to have the worst consequences, permissible because of its relation to virtue, and some entirely vicious conduct that happens to have the best consequences, permissible because of its relation to moral value and despite its relation to virtue. It is obscure how it could be in the nature of acting permissibly that such disparate conditions are sufficient. A more unified account of moral permissibility is highly desirable. This unity can be achieved with no apparent cost by dropping the maximizing disjunct. Before doing so we should consider a special type of case in which it might seem that the maximizing disjunct is needed. It is a type of case in which a permissible alternative fulfills the maximizing disjunct of VSC, but seemingly not the virtue disjunct. These are examples in which a greatly more beneficial maximizing requires such a self-destructive sacrifice that doing so might seem incompatible with the prudence and moderation of the virtuous.¹⁵
¹⁵ The donation example in section 2b is not an instance of this type of choice. In this type of choice, the self-sacrifice is a more extreme loss of well-being and the gain in value from the selfsacrifice is much greater.
185
It is plausible, though, that some who are virtuous would make such sacrifices. It is plausible that some virtuous decisions that are strongly influenced by compassion and generosity, while exercising practical wisdom, conclude that it is worth entirely sacrificing one’s own well-being. In some possible cases, only by the agent so doing can many others be spared that same level of destruction. On the highly credible assumption that some who are virtuous would make such choices, the virtue disjunct itself permits this self-destructive maximizing. The maximizing disjunct is not needed. Plainly the virtue disjunct also permits maximizing in many less extreme circumstances. A virtuous person whose friend is having an emergency will nearly always try to do the best that can be done in the situation. More generally, benevolence and compassion often face little or no opposition from other virtues in making a practically wise decision to maximize, when modestly self-sacrificing maximizing is available. The virtue of moderation limits deferring to other-regarding inclinations. It also limits deferring to prudent inclinations. A minor inconvenience or disappointment cannot make practically wise the neglect by the virtuous of a significantly better outcome. Moderation also allows a tiny gain or loss to agent well-being to be disregarded. The maximizing disjunct in VSC does not seem to do the theory any good. A simpler and more unified theory results from dropping it and having a pure virtue theory: VT A is morally permissible iff A is virtuous.¹⁶ The claim being made for VT here is that it gives an informative necessary and sufficient condition for the alternative’s being morally permissible.¹⁷
¹⁶ Two of the other three main moral evaluations of alternatives can be derived from VT in the usual way: An alternative is morally obligatory iff it is uniquely morally permissible among its alternatives; an alternative is morally wrong iff it is not morally permissible. As for being morally right, this status implies being morally permissible. But something beyond permissibility seems to be required. This further factor in being morally right seems to be that the alternative is morally justified by moral features that it has, rather than just being rendered permissible by not having moral faults. Spelled out in terms of virtue, the idea might be that a right alternative is one that could be characteristically chosen by the virtuous acting on otherregarding virtue, rather than just being something that some who are virtuous could choose without going out of character. ¹⁷ See section 8 below for a discussion of what makes acts moral.
186
7. Some Assets of VT For moral permissibility to relate to virtue in the VT way seems to ensure that the permissible acts are morally decent. The permissible acts of VT are practically wise choices that take into account the interests of others, employing traits like benevolence and compassion, while not neglecting the interests of the agent, employing traits like moderation and prudence. Catering to greed, inconsiderateness, sloth, or any other vice, plays no role. Acts that could be chosen virtuously with practical wisdom are never deplorable or inexcusable. Thus, VT does seem to require choices that are at least morally decent. VT makes room for supererogation. VT does not permit vicious maximizing. VT does not set an inflexible value threshold for permissibility. VT permits alternatives with sub-optimal outcomes at great agent cost with little overall gain. VT permits limited witting losses to agent well-being. Generally, VT allows conduct that the reasons for satisficing argue to be good enough, including making generous choices in ever-better cases. At the same time, the theory’s self-regarding virtues encourage giving some attention to self-interest and the theory does not require acting ideally well. These factors allow agents to permissibly avoid the severe burdens that seem to support permissibly satisficing. VT does these things by permitting whatever the virtuous might do in the situation without acting out of character. Thus, the reasons not to require maximizing, together with the objections to other attempts to allow satisficing, turn out to favor an unalloyed virtue theory.
8. Ultimate Explanation There is a familiar powerful objection to virtue theories. The objection argues that a virtue theory cannot be correct about what finally explains moral permissibility. Applied to VT, the critic’s charge can be elaborated like this: Any alternative that might be chosen characteristically by someone virtuous would be chosen on a certain basis. By VT, alternatives chosen on those bases are always moral. Granting that this is so, why is it so? The most plausible explanation is that the virtuous are ideal permissibility detectors. Their virtuous deliberations are just right for identifying permissibility-making
187
features. Guidance from inclinations toward virtue has the virtuous seek factors that make for permissibility (though they have no need to conceive of what they seek as morality-constituting or otherwise apply any moral concept in their choice). The virtuous find such factors and base their choices on them. Those factors give the more fundamental explanation of the permissibility of the virtuous choices. What determines an alternative’s moral status is something the virtuous detect in their bases for choice, rather than the potential to be selected by them. Analogy: A clay sphere is not made spherical by the fact that under ideal conditions the clay would appear spherical to an ideal perceiver. Rather, such perceivers are ideal sphere detectors. The subjunctive conditional perceptual fact is explained mainly by the clay’s having an intrinsic shape that makes such perceptual objects appear spherical to such perceivers. Likewise, assuming that VT has a correct extension, the fact of being a potential choice of the virtuous is an intermediary fact. Again, the role of the virtuous is to be ideal permissibility detectors. The critic concludes: Thus, it is not VT’s virtue condition that finally accounts for an alternative’s being morally permissible. Rather, its permissibility is ultimately explained by something in the basis on which an alternative might be virtuously selected. The conclusion of this objection can be sensibly disputed. It is not unreasonable to think that the moral permissibility of an alternative is finally explained by the alternative having a distinctively moral sort of choiceworthiness, rather than by having some distinctively moral basis for choice. Specifically, it is not unreasonable to think that eligibility for virtuous choice is what ultimately explains moral permissibility. This view can be supported by a competing analogy. In this view an act’s moral permissibility is analogous to a food’s deliciousness. The ultimate explanation of what makes food delicious is something like this: A food is delicious in virtue of its taste being intensely enjoyable to those with suitable palates (they would be those of us who have a typical capacity to taste food, or something like that). Although certain gustatory qualities make food enjoyable, what finally explains a food’s being delicious is its susceptibility
188
to being enjoyed for its taste, not its having the actually enjoyable gustatory qualities.¹⁸ Morality is not a matter of taste. If the choices of the virtuous are to determine morality, they must differ appropriately from the enjoyment of tastes. The appropriate difference might be this: The alternatives are chosen with full sensitivity to moral reasons. The alternatives are moral because of their eligibility for such choices. However, there is no need here to deny the conclusion of the objection. Suppose that the VT virtue condition is not an ultimate explanation of moral permissibility. This does no damage to the present aims for VT. The theory is offered as a statement of an informative necessary and sufficient condition for moral permissibility. To be right about this, VT need not give an ultimate explanation. If VT is otherwise correct, then we would still get from it the highly informative implication that the morally permissible alternatives are the ones that the virtuous might have chosen.¹⁹ We would still get the explanations offered here of how the reasons that oppose maximizing and support satisficing turn out to be correct about what is morally permissible. On the supposition that VT does not give an ultimate explanation, important theoretical work remains to be done. A theory is needed that gives an illuminating identification of what is shared by the intrinsic natures of the bases on which the virtuous might choose. Whether or not VT is an ultimate explanation of moral permissibility, it enhances our understanding of how alternatives that do not maximize can be moral.
References Baron, M., Pettit, P., and Slote, M., 1997. Three Methods of Ethics: A Debate. Hoboken: Wiley-Blackwell. Bradley, B. 2006. “Against Satisficing Consequentialism.” Utilitas 18: 97–108. Chappell, R.Y. 2019. “Willpower Satisficing.” Nous 53: 251–65.
¹⁸ The credibility of this view of deliciousness can be seen by thinking of possible variations in our taste in food. A food’s being delicious under other possible conditions is settled by how our typical taste in food would respond to the food under those conditions, not by whether the gustatory qualities that delicious foods actually have would be present. ¹⁹ Philosophers who defend satisficing consequentialism might benefit from this information. Thinking about the choices of the virtuous might be especially useful in identifying the cases in which a correct satisficing condition holds. I am grateful to a reviewer for suggesting this.
189
Conee, E. 2014. “Moral by Virtue of Virtue.” In M. Timmons, ed., Oxford Studies in Normative Ethics, Volume 4. Oxford: Oxford University Press. 53–78. Hurka, T. 1990. “Two Kinds of Satisficing.” Philosophical Studies 59: 107–11. Hursthouse, R. 2000. On Virtue Ethics. Oxford: Oxford University Press. Mulgan, T. 2001a. The Demands of Consequentialism. Oxford: Oxford University Press. Mulgan, T. 2001b. “How Satisficers Get Away with Murder.” International Journal of Philosophical Studies 9: 41–6. Rogers, J. 2010. “In Defense of a Version of Satisficing Consequentialism.” Utilitas 22: 198–221. Slater, J. 2020. “Satisficing Consequentialism Still Doesn’t Satisfy.” Utilitas 32: 108–17. Turri, J. 2005. “You Can’t Get Away With Murder That Easily: A Response to Timothy Mulgan.” International Journal of Philosophical Studies 13: 489–92. Vallentyne, P. 2006. “Against Maximizing Act-Consequentialism.” In J. Dreier, ed., Contemporary Debates in Moral Theory. Hoboken: Wiley-Blackwell. 21–37.
9 Pro Tanto Rights and the Duty to Save the Greater Number Benjamin Kiesewetter
The distinction between contributory and overall normative concepts is an important reference point in contemporary moral theory. Overall concepts include “ought”, “wrong”, “required”, “prohibited”, and “permitted”, while the paradigm contributory concept is the concept of a “reason”—the notion of a factor that supports an action only to some extent (“pro tanto”), and stands in competition with other contributory factors. A natural way to illustrate the distinction appeals to the inputs and outputs of practical deliberation: overall concepts can be used in a verdict that concludes practical deliberation, while judgements employing only contributory concepts can merely serve to express premises of such deliberation.¹ How does the notion of a moral claim-right relate to this distinction? According to what I will call rights contributionism, rights are competing factors that can be overridden. Judith Jarvis Thomson advocates this view when she argues that one can infringe a right without violating it, where “violation” refers to the impermissible infringement of a right.² In contrast, according to rights absolutism, moral rights correlate with overall moral obligations, which cannot permissibly be breached. This view is often ascribed to Robert Nozick’s conception of rights as “side constraints”, and it has been explicitly defended by Russ Shafer-Landau and, more recently, Jay
¹ As I use the term “overall” in this paper, an overall concept need not be domainindependent (like the all-things-considered ought), but may instead be domain-relative (like the moral or the prudential ought). Consequently, an overall judgement might conclude moral deliberation without thereby concluding other kinds of deliberation. ² Thomson (1976, 40). See also Feinberg (1978) and Thomson (1990, Chs. 3–6). The notion of a pro tanto right differs from the epistemic notion of a prima facie right (cf. Montague 1988, 360); it refers to a full-fledged right rather than an indicator of such a right. Benjamin Kiesewetter, Pro Tanto Rights and the Duty to Save the Greater Number In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Benjamin Kiesewetter 2023. DOI: 10.1093/oso/9780198895909.003.0010
191
Wallace.³ To illustrate these two views, consider the following example by Joel Feinberg: Suppose that you are on a backpacking trip in the high mountain country when an unanticipated blizzard strikes the area with such ferocity that your life is imperiled. Fortunately, you stumble onto an unoccupied cabin, locked and boarded up for the winter, clearly somebody else’s private property. You smash in a window, enter, and huddle in a corner for three days until the storm abates.⁴
Rights contributionists, like Thomson and Feinberg, hold that the backpacker infringes the cabin owner’s property rights, but does so permissibly. By contrast, rights absolutists must hold either that the backpacker acts wrongly or that she does not infringe a moral right by breaking into the cabin. Insofar as they want to allow for the plausible verdict that the backpacker acts permissibly, rights absolutists have to deny that the cabin owner has a right against the unauthorized use of his property under the circumstances depicted by the example. On a specificationist construal of this view, the scope of the right against the unauthorized use of property has to be specified by way of an exceptional clause that excludes the circumstances of the example (as well as other circumstances that permit the unauthorized use of property). On a particularist construal, truths about rights are always relative to particular circumstances, and relative to the circumstances depicted by the example, the cabin owner has no right against the unauthorized use (and the same is true for other circumstances that permit the unauthorized use of property).⁵ This chapter has two aims. In the first part (sections 1–4), I argue that the notion of a moral right is to be understood as a contributory rather than an overall normative notion. Rights are competing factors that can conflict and have to be weighed against each other; they do not entail by themselves that an action is impermissible. The argument is an inference to the best explanation. I argue that (i) there are contributory moral factors
³ See Nozick (1974, Ch. 3), Shafer-Landau (1995), Wallace (2019, esp. 170–6). Note that I do not assert that Nozick really held this view; there are reasons to doubt that he did. ⁴ Feinberg (1978, 102). See also Thomson (1976, 40) for an example of the same structure. ⁵ See Shafer-Landau (1995, 213), who adopts a specificationist approach. Shafer-Landau (1995, 214) claims that only rights absolutism is compatible with particularism, but it is difficult to see why this should be so. In my view, the distinction between contributionism and absolutism is orthogonal to the distinction between particularism and generalism.
192
that contrast with standard moral reasons by way of having a number of formal properties that are characteristic of rights, even though they can be overridden, and (ii) that this is best explained by a view that takes these factors to correlate with rights. In the second part of the chapter (sections 5–6), I aim to show that the truth of rights contributionism matters for normative ethics. More specifically, I argue that rights contributionism clears the way for deontologists to justify the pre-theoretically plausible verdict that we ought to save the greater number in so-called Taurek scenarios—scenarios in which we have to choose between saving either a greater or a smaller number of different people. I offer a novel and distinctively deontological explanation of this verdict that is based on the assumption that everyone has a pro tanto right to be saved in a Taurek scenario. Before getting started, a word of clarification on the relation between rights, duties, wrongness, and wronging will be useful. Following Hohfeld’s seminal analysis of legal concepts (Hohfeld 1913), moral philosophers widely agree that claim-rights correlate with so-called directed duties— duties or obligations that are held specifically towards the claimholder. In line with this, I will assume that rights contributionism entails contributionism about directed duties. However, it is worth noting that rights contributionists can accept that the terms “duty” and “obligation” can also be used, in another sense, to refer to the notion of an overall moral requirement, which has to be distinguished from the notion of a directed duty.⁶ Thus, in response to the putative platitude that a breach of duty is morally impermissible,⁷ rights contributionists can (and should, in my view) hold that the assumption is a platitude only if “duty” refers to an overall moral requirement but not if it refers to the notion of a directed obligation. Moreover, since rights contributionists believe that rights infringements can be permissible, they must reject either the assumption that rights infringements necessarily wrong claimholders or the assumption that wronging entails wrongdoing. However, they can maintain that rights infringements wrong claimholders, or that wronging is wrong, whenever the right that is being infringed is undefeated.
⁶ This follows already from the plausible assumption that not all moral requirements are directed. But even if we agree with Wallace (2019) in rejecting this assumption, contributionists might hold that the term “duty” is ambiguous between a contributory and an overall notion. ⁷ Montague (1988, 348) and Shafer-Landau (1995, 209–10) both argue against rights contributionism on the basis of this assumption.
193
1. Directed deontic structure Some moral reasons have distinctive features that are characteristic of a right/duty-nexus. When a reason has these features, I will say that it has directed deontic structure.⁸ The basic point of the first part of this chapter is that directed deontic structure is located on the contributory level, i.e., it obtains regardless of whether the reason in question is defeated and thus regardless of whether or not the action in question is morally required or permitted. In this section, I start by describing the structure in question using the example of a promissory reason. I then argue that this structure is common for some but not all moral reasons and that it is characteristic of rights (section 2). Subsequently, I argue that reasons have directed deontic structure regardless of whether or not they defeat competing reasons or are defeated by them (section 3), and propose an abductive argument for rights contributionism based on the assumptions argued for (section 4). Suppose that Alan promises Belma that he will help her move, thereby coming to have a moral reason to do so. This reason has a number of interesting features. To begin with, it is a directed reason, which is to say that it is a reason conformity to which is owed to another person, namely Belma.⁹ As is known from discussions of the problem of third-party beneficiaries, this aspect of directedness cannot simply be identified with the fact that the action supported by the reason benefits Belma.¹⁰ Belma may be the beneficiary of an action that Alan has reason to perform without its being the case that conformity is owed to her, and Alan may owe a person conformity with a reason in support of an action that does not benefit her. For example, if Alan promises Carlos that he will help Belma move, conformity with this reason may benefit Belma, but is not owed to Belma, and it may not benefit Carlos, but is still owed to him. A second point, which is closely related to directedness, is that failure to conform with a promissory reason calls for amendments or moral repair. ⁸ This terminology is inspired by Wallace’s term “deontic structure” (Wallace 2013). But while there is significant overlap between Wallace’s notion of deontic structure and my notion of directed deontic structure, they are not equivalent. For one, directedness is not (as far as I understand it) part of Wallace’s notion of deontic structure, but rather part of its explanation. For another, what Wallace calls “lack of discretion to ignore” seems to be central for his notion, while one can consistently hold that lack of discretion to ignore is not a distinctive feature of directed deontic structure in my sense. ⁹ For present purposes, a directed reason may be understood as the reason that is constitutively related to a directed duty, but see section 3 for a characterization that does not appeal to the notion of a directed duty. ¹⁰ See esp. Hart (1955, 180–1).
194
If Alan breaks his promise, this fact gives him reason to apologize to Belma and to offer her compensation for costs that she might have incurred or for the missed opportunity, for example by offering to help her on another day. There are of course circumstances that justify the breaking of a promise (a possibility that I will discuss in more detail in section 3), and insofar as we take an apology to involve the admission of a wrongdoing, it seems too strong to say that promise-breakers have reason to apologize in such cases. But even when there was no morally preferable solution in the situation, promise-breakers have reason to regret that they did not do what they owed to the promisee; they have reason to express these regrets by saying sorry (if this is possible); and they have a special responsibility to help to alleviate the harm that their promise-breaking might have caused for the promisee (if there is any harm). Notably, the kind of regret that is called for in such cases is essentially first-personal and relational; it is not merely an attitude towards a regrettable state of affairs that anyone could hold, but a de se attitude about the agent’s own failure to do what he or she owed to another person. I will call this relational agent-regret.¹¹ A third feature of Alan’s reason worth highlighting is its peremptoriness, by which I mean the conditional property that a reason has when it makes the action morally required in the absence of sufficient competing reasons. Since Alan is morally required to conform with his promissory reason unless there is a countervailing reason with sufficient strength, this reason is peremptory. Fourthly, Alan’s reason is pre-emptive, by which I mean that it is capable of defeating certain competing reasons in a special, pre-emptive way. What exactly pre-emptive defeat amounts to is a contentious issue, but I hope that the following three points provide at least a rough idea without generating too much controversy. The first point is that pre-emptive defeat contrasts with (at least typical forms of ) both outweighing and undermining defeat. That is, a pre-emptive reason wins the competition against pre-empted reasons not simply because it is stronger (as any ordinary reason could in principle be stronger than a competing reason), nor does it figure as
¹¹ Williams famously discusses agent-regret in the context of moral luck cases, in which it involves “the wish, all things taken together, that one had acted otherwise”, whereas the regret appropriate in cases of justified non-conformity with a directed reason merely involves “a wish that things had been otherwise” (cf. Williams 1976, 31). But Williams is explicit that only the latter and not the former kind of wish is necessary for agent-regret, precisely because he thinks that there are “cases of conflict . . . , where either course of action, even if it is judged to be for the best, leaves . . . agent-regrets” (1976, 31).
195
an undercutting defeater (or “disabler” or “cancelling condition”) that undermines their status as reasons. A natural metaphor to use in this context is that of a trump in a card game, which overrides other cards independently of their count value. Hence, pre-emptive reasons can be said to override pre-empted reasons by trumping them rather than outweighing (or undermining) them. The second point (which is really just the other side of the same coin) is that pre-emptiveness entails a certain kind of protection, namely protection from being outweighed by pre-empted reasons. If one reason pre-empts another, then adding another reason of the same kind and strength as the pre-empted reason will not make a difference to the overall verdict of the case: whatever trumps the first reason will also trump the second one. This is crucially different from standard outweighing defeat. If one reason outweighs another, then adding another reason of the same kind and strength as the outweighed reason often tips the balance against the first reason and makes all the difference for the overall verdict. The third point is that pre-emptive reasons seem to provide a vindicating explanation of the phenomenon of deliberative bracketing. As several authors have remarked, it can be appropriate and characteristic of moral conscientiousness to give certain moral reasons a special treatment in deliberation, which involves disregarding certain kinds of competing reasons or excluding them from the process of weighing reasons.¹² This is naturally explained in terms of pre-emptive defeat: if one reason defeats another by trumping rather than outweighing it, then the second reason need not be weighed against the first, and it can appear inappropriate to give it serious consideration in deliberation. It is natural to ask what the deeper mechanism behind pre-emptive defeat is. According to Raz’s influential account, pre-emptive reasons defeat competing reasons because they are combinations of first-order reasons and second-order “exclusionary reasons” against acting for certain competing first-order reasons, and an exclusionary reason against acting for a reason defeats this reason in a way that does not amount to outweighing or undermining it. However, both the claim that there are reasons against acting on other reasons, as well as the claim that such reasons are capable
¹² See e.g. McDowell (1980, 17). Raz’s notion of an “exclusionary reason” (see esp. Raz 1990, 35–48 and 178–99) is naturally associated with this phenomenon as well. Strictly speaking, however, Razian exclusionary reasons only count against acting on certain first-order reasons, not necessarily against including them in deliberation.
196
of defeating the first-order reasons, can and have been questioned, and neither of these claims is entailed by the idea of pre-emptive defeat itself.¹³ So there is no need to presuppose them here. The notion of a pre-emptive reason I will work with is the notion of a first-order reason that defeats certain competing reasons in a special “trump-like” way (distinct from standard forms of outweighing and undermining), which entails that the reason is generally protected from being outweighed by reasons of this kind and makes deliberative bracketing appropriate. Back to Alan. If Alan is morally conscientious, he will not treat certain reasons that conflict with his promise as serious contenders, including, for example, reasons to engage in certain valuable activities, or standard welfarist reasons to promote his own or somebody else’s well-being. For example, if Alan is asked whether he would like to attend a theatre performance, but he cannot do this without breaking his promise, he will not take the reason to do this as a respectable competitor that has to be weighed against his promissory reason. This does not change if he learns that the performance is to be followed by another one that is equally attractive. This suggests that his promissory reason is pre-emptive; it defeats certain competing reasons in a special “trump-like” way and is in turn protected from being outweighed by reasons of this kind. A fifth feature of Alan’s reason is its waivability, by which I mean the property that a reason has when someone has the power to cancel the reason by declaration. Promissory reasons are a paradigmatic example of waivable reasons, as promisees generally have the power to cancel promissory reasons by releasing promisers from their promises. For example, if Belma consents to Alan’s going to the theatre performance rather than helping her, she thereby releases Alan from his promise and makes it the case that Alan’s promise is no longer a (promissory) reason to help her. Finally, promissory reasons are voluntary; they exist as a result of a voluntary, communicative act of commitment. By promising Belma that he will help her, Alan communicates an intention to undertake a commitment to help her, and it is a result of this voluntary act that he comes to have a new reason to do so. To sum up, promissory reasons are (i) directed, (ii) essentially related to moral repair, (iii) peremptory, (iv) pre-emptive, (v) waivable, and (vi) voluntary. These features do not co-occur contingently, but they are part ¹³ Recent criticisms of Razian exclusionary reasons include Whiting (2017) and Adams (2021).
197
of a common structure that is shared by a distinct class of moral reasons, which I call directed deontic structure. I take (i)–(iv) to be individually necessary and jointly sufficient for this structure. Features (v) and (vi) are arguably not necessary, but there is reason to think that they are sufficient, and they in any case indicate directed deontic structure.
2. A distinctive structure characteristic of rights In this section, I will first illustrate my claim that directed deontic structure is a distinctive property of a certain class of moral reasons and then argue that it is characteristic of rights. With regard to the first point, consider the moral reason to refrain from unauthorized use of somebody else’s property. This reason seems both pre-emptive and peremptory: it is protected from being defeated by standard welfarist reasons or reasons to engage in valuable activities and it makes acts impermissible in the absence of sufficient countervailing reasons. Moreover, conformity with it is owed to the property owner and non-conformity calls for repair (i.e. for apology, compensation, or at least the expression of relational agent-regret). It is also waivable, and at least typically generated by a voluntary act of acquiring ownership. So reasons to respect other people’s property have directed deontic structure. Next, consider the reason to refrain from harming others. This reason is not voluntary and perhaps not fully waivable, but it is peremptory, pre-emptive, directed, and repair-related, and so has directed deontic structure in the sense defined above. Similar points could be made about reasons against deceiving people, reasons for repaying one’s debts, or reasons to respect another person’s bodily integrity. Yet not all moral reasons have directed deontic structure— there is at least one important class of moral reasons that lack this structure. This class may be called the class of standard altruistic reasons—reasons that are provided by the fact that an action (or omission) would make some other person happy, or benefit her, do her a favor, etc. To illustrate, consider a case in which Belma’s neighbor Daria has a standard altruistic reason to help Belma move. She sees that Belma could need a hand, and that helping her would reduce her stress, but she hasn’t committed herself to helping, nor is this a case of emergency, in which her help is necessary or sufficient for preventing a great harm. That helping her would reduce Belma’s stress is, I take it, a reason to do this, and unless we presuppose a particularly restrictive, theoretically laden notion of morality, it seems to count as a
198
moral reason. Yet this reason lacks every single feature that I have highlighted as characteristic of promissory reasons and other reasons with directed deontic structure. It is not directed, as conformity with it is not owed to another person (recall that directedness is not to be confused with being related to someone else’s welfare). It is also not repair-related, as failure to conform with this reason does not give rise to reasons to apologize or compensate. Neither is it peremptory, for even if Daria lacks sufficient competing reasons, it still seems morally permissible for her to refrain from helping Belma. To be sure, it will then be morally better to help her, but this is plausibly a case of supererogation rather than a case of a moral requirement. Daria’s reason to help Belma also lacks pre-emptive force. Moral conscientiousness is fully consistent with weighing this reason against other welfarist reasons (including self-interested ones) or reasons to engage in valuable activities such as a theatre performance, and it is not protected from being outweighed by such reasons. Daria’s reason is also not waivable. While it may be possible for Belma to act in ways that make the reason disappear (by changing the fact that Daria’s help would reduce her stress), this must be distinguished from the power to cancel a reason by declaration—a power that Belma lacks in this case. Finally, it is clear that Daria’s reason is not voluntary, as it isn’t generated by an intentional action. In sum, directed deontic structure is a property of a distinct subclass of moral reasons. I will next argue that this structure is characteristic of rights. I hold that (i)–(iv) are essential (and thus necessary) features of all moral claim-rights, while (v) and (vi) are essential features of at least some rights. As I mentioned in the beginning of this chapter already, claim-rights correlate with so-called directed duties—duties owed to the claim holder. It follows from this that all moral claim-rights are directed and peremptory: conformity with a moral claim-right is always owed and since infringing a right involves a breach of duty, doing so cannot be morally permissible in the absence of sufficient reasons. It also seems clear that all such rights are pre-emptive. Rights are often described as trumps.¹⁴ And while this idea can be understood in ways that are arguably too strong or restrictive, the interpretation according to which rights are capable of defeating certain other considerations in a pre-emptive fashion is not. In fact, it is difficult to see how the notion of a moral right could have a point unless one assumes
¹⁴ See esp. Dworkin (1984).
199
that rights involve a special protected status, which is not vulnerable to all the standard forms of defeat. Further, it seems plausible to think that all moral rights are related to moral repair in the sense that right infringements call for apology or the expression of relational agent-regret, and in case it goes along with harms, also for compensation. Will theorists of rights hold that waivability is the core element of all rights. But even those who think that there are unwaivable or unalienable rights can and should accept that waivability is an essential feature of many rights. Finally, while clearly not all rights are the result of voluntary decisions, it seems essential for a certain class of rights that they can be granted in voluntary transactions, such as contracts.
3. Defeat Directed deontic structure, I will now argue, is located on the contributory level: reasons have this structure regardless of whether or not they defeat all competing reasons or are defeated by competing reasons themselves. To substantiate this claim, let me modify the original example of Alan and Belma, supposing that Alan’s child falls ill and has to be taken care of at home, which provides Alan with a conclusive reason to break his promise to help Belma (assume, for the sake of simplicity, that he is unable to reach her and ask to be released from his promise). Evidently, this does not change the fact that Alan has a pro tanto reason to keep his promise, that there is still something to be said in favor of doing so. That a reason is defeated by a competing reason does not mean that it is cancelled. Moreover, this reason still seems directed towards Belma. Directedness is not the kind of property that a reason loses by being defeated. It may be argued that keeping a promise cannot truly be owed if there is stronger reason to break it. But it is difficult to see why this should be so. Surely, I can owe you money even if I have stronger reason to feed my hungry child than give it back to you. If I can owe you money under such circumstances, why couldn’t I owe you the act I promised even when I have stronger reason to refrain? More importantly, if we cannot owe conformity to defeated reasons, this would only mean that the property of directedness has to be distinguished from the property of being a reason with which we can owe conformity. For example, there is a clear sense in which a reason is directed to S if S can
200
cancel the reason by declaration (i.e. if the reason is waivable).¹⁵ This property does not depend on whether or not it is outweighed: if Belma hears about Alan’s sick child, she can release him from the promise, thereby cancelling a reason that was overridden but still had normative weight. Since overridden reasons can be waivable, and waivability entails directedness, overridden reasons can be directed as well. This is further supported by the fact that failure to conform with an overridden promissory reason still calls for repair. It seems very natural for Alan not only to explain himself to Belma but to express relational agentregret that he broke his promise. It also seems that Alan owes Belma something like compensation.¹⁶ For example, should Belma ask him to promise to help her on another day, it seems plausible to think that the fact that he failed to keep his earlier promise gives Alan an additional reason to offer her another promise—perhaps even a peremptory one. And if Belma incurred expenses because of the broken promise, this seems to give Alan agent-relative reasons to assume these costs or contribute to them, even though he did nothing wrong. Alan’s reason also remains pre-emptive, which is revealed by two considerations. First, a very general point: pre-emptiveness is a property that a reason has with respect to certain but not necessarily all other reasons. For instance, promissory reasons plausibly are pre-emptive with respect to reasons to do others a favor, but not with respect to reasons to care for one’s children. It is therefore unclear why, if a reason is pre-emptive with respect to reasons of kind K, the fact that this reason is defeated by a reason of a different kind K* should undermine this reason’s pre-emptiveness—a property that holds with respect to reasons of kind K and not of kind K*. Secondly, defeated promissory reasons are still protected from being outweighed by pre-empted reasons. In other words, there are modal truths about such reasons that hold in virtue of the fact that they pre-empt other ¹⁵ Another plausibly sufficient condition is that an interest of S grounds a pre-emptive reason. Moreover, it seems to me plausible to think that a reason is directed if, and only if, at least one of these conditions hold. ¹⁶ This point about compensation is already highlighted by Feinberg (1978) and Thomson (1980) in support of rights contributionism. Both of them seem to presuppose, however, that rights are needed to explain why compensation is owed, which can be questioned (cf. ShaferLandau 1995, 216–17). By contrast, the argument of this chapter does not presuppose that rights are explanatorily prior to reasons for compensation. The view that non-conformity with defeated reasons could give rise to reasons for compensation is also sometimes denied (e.g. Montague 1984). I don’t think that the overall case for contributionism presented in this paper hinges essentially on this claim; readers unwilling to accept it are invited to bracket this feature and focus on the others.
201
reasons. For example, the pre-emptiveness of Alan’s defeated promissory reason explains why this reason would be conclusive in the absence of a reason to take care of his child even if he could do a lot of people a favor by breaking his promise. Pre-emptiveness can have this explanatory power only if it is not undermined by countervailing defeat. It is also clear that an outweighed promissory reason is still peremptory, which is to say that it still has the conditional property of making conformity morally required in the absence of sufficient countervailing reasons. Finally, that the waivability of a reason does not depend on whether it is overridden has already been argued, and that the same holds for voluntariness is immediately evident. At this stage of the argument, right absolutists might feel inclined to reject my earlier claim that a defeated promissory reason is still a reason.¹⁷ Could this be a plausible view? In my example, the promissory reason is most plausibly seen as being pre-empted by a duty of care, and although the orthodox view distinguishes pre-emptive from undermining defeat, there are also conceptions that take pre-emption to be a special kind of undermining defeat.¹⁸ We can circumvent this issue by considering an example in which the promissory reason is outweighed rather than pre-empted. Suppose, for example, that Alan has the opportunity to engage a babysitter, but feels that his presence at home would comfort his child. In such a case, it does not seem inappropriate to weigh his commitment to Belma against the reason to care for his child, and it is still possible that the latter reason prevails. Or suppose that Alan has a weighty reason to break his promise because doing so is the only way for him to meet an old friend who is in town for just one day.¹⁹ It seems to me implausible to think that the value of a friendship could never outweigh a promissory reason, but there is no reason to think of this case as one of pre-emptive defeat. In order to reject my thesis that directed deontic structure obtains on the contributory level, absolutists would have to maintain that every case of defeat of a reason with directed deontic structure is also a case of undermining defeat. This, however, is a highly dubious claim. First of all, it is ad hoc. Outweighing does not usually involve undermining, and so it is unclear why this should be so in the case of reasons with directed deontic structure. Secondly, it is plainly implausible to deny that there are outweighed reasons with directed deontic structure. This may be clearest in the case of reasons ¹⁷ Wallace (2019, 173) comes close to asserting this. ¹⁹ Cf. Wallace (2019, 176).
¹⁸ See Adams (2021).
202
against harm. Surely something counts against harming another person even if it is overall morally permitted. Similarly, something counts against stealing, deceiving, or breaking valid promises even if exceptional circumstances justify these acts. This point is, thirdly, reinforced by the phenomenology of conflict that is characteristic for outweighing defeat but absent from undermining defeat. Why would Alan feel conflicted about meeting his friend if his reason to do so would make it the case that he no longer has any reason to keep his promise? Fourthly, the view seems to have the false implication that the outweighed reason cannot join forces with another reason to defeat the defeater, as an undermined reason has no normative force that can be conjoined. For example, if Alan learns that he will reunite with an equally important friend if he participates in Belma’s move, this fact should together with the promise tip the balance in favor of doing so. This, however, presupposes that the promise still counts in favor of participating in the move, and is not undermined by the competing reason. I conclude that the point stands: directed deontic structure obtains on the contributory level.
4. An inference to the best explanation I have argued that some moral reasons exhibit a distinctive directed deontic structure that is characteristic of rights, and that they do so regardless of whether they are defeated by competing reasons. The very features that are characteristic of rights thus obtain on the contributory level. These results support rights contributionism, or so shall I argue in this section. That some moral reasons have a special directed deontic structure that is characteristic of rights calls for explanation. What relation do these reasons bear to rights that could explain this fact? Three possibilities come to mind. The first is that the relevant reasons are grounded in rights. On such a view, by promising Belma that he will help her, Alan grants Belma a moral right to his help, and this right provides Alan with a reason to help Belma. This delivers a straightforward explanation of why Alan’s reason has a rightscharacteristic structure. The second possibility is that rights do not ground reasons, but are themselves grounded in reasons. On such a view, rights are constituted by reasons that have some of the features that are distinctive of Alan’s promissory reason. By promising, Alan generates a reason that is directed, repairrelated, pre-emptive, etc., and that (or part of that) grounds Belma’s having a
203
right against Alan. This would also provide a straightforward explanation of why Alan’s reason has the distinctive features characteristic of rights. A third possibility, which would also provide the required explanation, is that rights are identical to reasons with directed deontic structure and neither takes metaphysical priority. There is no need to decide between these views here. All three of them entail that reasons with the distinctive features characteristic of rights necessarily correlate with rights, and on the assumption that reasons exhibit these features on the contributory level, this entails rights contributionism. This argument may be summarized as follows: 1. Some moral reasons have a distinctive directed deontic structure that is characteristic of rights. 2. The best explanation of this entails that these reasons correlate with rights. 3. Therefore, reasons with the relevant features correlate with rights (from 1 and 2). 4. But reasons have the relevant features even in cases in which it is morally permissible or even morally required not to comply with them. 5. Therefore, in some cases it is morally permissible or even morally required not to comply with a right (from 3 and 4). Premises (1) and (4) have been argued for above. I supported premise (2) by outlining three views that entail rights contributionism and that seem to provide excellent explanations of (1). Like other claims to a best explanation, (2) is defeasible by providing an equally good alternative explanation. Could rights absolutists provide such an explanation? Consider, first, a picture like Wallace’s, according to which a right/dutynexus is not constituted by reasons, but is a fundamental normative entity that provides reasons, while at the same time necessitating the moral impermissibility of non-compliance. It’s very difficult to see how one can explain that a rights-specific structure obtains at the level of morally outweighed reasons on such a view. Its proponents seem committed to either denying the existence of contributory reasons with directed deontic structure or leaving unexplained the strong analogies that these reasons bear to rights. Things seem to look better for an absolutist view that allows for an analysis of rights in terms of a conglomerate of a contributory normative notion and a no-defeater condition. For example, absolutists might suggest
204
distinguishing between rights and claims, where claims are taken to be contributory. On the basis of such a distinction, absolutists might hold that the occurrence of contributory reasons with rights-specific characteristics can be explained by the facts that (i) the relevant reasons correlate with claims (as they constitute claims or are provided by claims) and (ii) rights just are undefeated claims. My worry about this reply is that the view described seems to be a terminological variant of rights contributionism. It does not deny that there are contributory rights, it just calls them “claims” rather than “rights” and reserves the term “right” for absolute rights. This may be a legitimate terminological choice, but it does not support a substantive alternative to contributionism. Could absolutists avoid this complaint by analyzing rights directly in terms of undefeated reasons of a certain kind without granting the contributory property of having a claim? I’m not sure. First of all, the right kind of reasons for such an analysis are reasons with directed deontic structure, and it is an open question whether this structure can be spelled out in terms of reasons alone.²⁰ Secondly, supposing that this challenge can be met, it is still unclear to me how the view avoids the concession of contributory claims and thus the charge of collapsing into a terminological variant of contributionism. The view at issue admits that there is a unified normative property on the contributory level, which has to be distinguished from the property of being a (standard) moral reason but involves a reason with a directed and pre-emptive structure. It’s not clear to me what it means to deny that the instantiation of this property entails a claim over and above the terminological decision to use this term more restrictively. In any case, the substantive point that I have argued for is that there is a distinctive and unified normative structure that obtains on the contributory level—directed deontic structure—which is more complex than a simple moral reason and very naturally described in terms of a right/duty-nexus. Everyone is free to use the term “right” in a more restrictive way, as applying only in case the moral reason contained in the structure is undefeated, and use some other term for this phenomenon—but this doesn’t change the fact that the structure itself exists and unfolds its normative significance on the ²⁰ Note that I am presently concerned with an analysis of rights in terms of reasons for compliance, not in terms of other reasons, such as reasons for rejecting certain moral principles (cf. Scanlon 1998). This is because I am here looking for an absolutism-friendly explanation of (1), and it is specifically an analysis in terms of directed-deontic reasons for compliance that would seem to provide such an explanation.
205
contributory level. This substantive point has important upshots for both normative and meta-normative theorizing, no matter whether one uses the term “right” or the term “claim” or some other term to refer to it. Consider, for example, the view that deontic concepts or properties can be explained in terms of reasons. It is remarkable that so-called deontic buckpassers have hitherto focused almost entirely on overall concepts.²¹ If the argument of this chapter is correct, however, they face the additional and neglected challenge of accounting for pro tanto rights and duties in terms of reasons. Moreover, one of the implications of the argument presented is that pro tanto rights and duties differ in several crucial respects from standard moral reasons, and (as pointed out above) this means that meeting this challenge will be far from trivial. Obviously, this challenge cannot be avoided by restricting the term “right” to undefeated reasons with directed deontic structure, as it is this structure itself that deontic buck-passers are committed to explaining in terms of reasons alone. These meta-normative matters deserve to be discussed and explored in more detail, but that will have to wait for another occasion. In the remainder of this chapter, I will focus on rights contributionism’s relevance for normative ethics.
5. Taurek cases In his article “Should the Numbers Count?”, John Taurek (1977) considers a number of cases in which an agent has the choice between (a) doing nothing, (b) doing something that will save a number of people, and (c) doing something that will save a greater number of different people. I will call cases that instantiate this structure Taurek cases. A famous example that Taurek adopts from Philippa Foot (1967) is one in which six patients will die unless they are treated with a life-saving drug. While five patients need only one fifth of the drug, the other patient needs all of it to survive—the owner of the drug thus has to choose whether to let five die or one. As Taurek remarks, many take it to be plausible or even obvious that in cases like this, agents have a duty (in the sense of an overall moral obligation) to choose the option that saves the greater number of people. As everyone agrees, this duty has to be qualified by a ceteris paribus clause:
²¹ See e.g. Stratton-Lake (2002); Bedke (2011).
206
there can be relational duties and other factors that can make a difference to what may permissibly be done in Taurek cases. But if we abstract away from such situation-specific features, the requirement to save the greater number will strike many as plausible. Let’s call this: The Principle of Saving the Greater Number: Other things equal, agents have an (overall) moral obligation to save the greater number in Taurek cases. This principle is clearly not uncontroversial—Taurek himself famously rejects it, and a number of deontologists have followed him in doing so.²² But it still seems fair to say that it is prima facie plausible. The Principle of Saving the Greater Number is easily vindicated by standard consequentialism, as following it maximizes aggregated welfare, which is what standard consequentialism tells us to do. But it is not so easily vindicated by deontological theories. This is because deontologists are committed to the view that maximizing aggregated welfare is subject to deontological constraints and in many cases impermissible. A famous example that illustrates this point is Thomson’s transplant case, in which a surgeon has the opportunity to harvest the organs of a healthy person, thereby killing her in order to save five other patients each of which urgently needs one of the healthy person’s organs in order to survive.²³ While secretly sacrificing the healthy person would maximize aggregated welfare, and thus be required by the lights of standard versions of consequentialism, deontologists widely agree that doing so is impermissible. Given their rejection of interpersonal aggregation in this and other cases, it is not obvious how they can justify a duty to save the greater number in Taurek cases. Indeed, deontologists who have argued for the Principle of Saving the Greater Number, such as Frances Kamm and T. M. Scanlon, have been accused of relying implicitly on principles of interpersonal aggregation that are in tension with their deontological commitments.²⁴
²² Including Timmermann (2004) and Munoz-Dardé (2005). ²³ See e.g. Thomson (1990, 135). A case with the same structure is already discussed by Foot (1967, 24). ²⁴ See e.g. Otsuka (2000), Timmermann (2004), and Doggett (2009) for this charge against Scanlon (1998, 229–41) and Kamm (1993, Ch. 6). Doggett (2009, 14) concludes that no nonconsequentialist explanation of the Principle of Saving the Greater Number can be given. There are parallels between Scanlon’s “tie-breaking argument” and my own proposal below, which also entails that the additional reasons for saving the persons in the bigger group break the tie. But while Scanlon aims to derive this conclusion from contractualist assumptions
207
In what follows, I will argue that there is a plausible and distinctively deontological vindication of the Principle of Saving the Greater Number, which appeals to the contributionist assumption that in Taurek cases, everyone has a pro tanto right to be saved—a right that conflicts with and thus has to be weighed against the rights of the others. Independently plausible principles of how the aggregation of reasons interacts with rights can explain why reasons for saving different people allow for aggregation in Taurek cases but not in examples like the transplant case.
6. A deontological vindication of the Principle of Saving the Greater Number Consider the following principle of reasons aggregation: Reasons Aggregation: If two reasons provide independent support for a response, then the combined support for this response is stronger than the support that each reason provides individually—unless this is prevented by something. A couple of remarks are in order. Firstly, the principle only applies to reasons that are independent.²⁵ To illustrate, consider the following three reasons to use a certain toothpaste: (a) that it prevents cavities, (b) that it protects dental enamel, (c) that it prevents toothaches. While (a) and (b) are independent reasons to use the toothpaste, (c)’s being a reason is part of the explanation of why (a) is a reason, and so (a) and (c) are not independent. Reasons Aggregation entails that (a) and (b) are stronger in combination (if nothing prevents this), while (a) and (c) are not—which is evidently the right result. A second point to note is that the principle does not assume that reasons aggregation is necessarily additive. Consequently, it does not presuppose that the strength or weight of reasons can be represented by numerical values on an interval scale. Thirdly, Reasons Aggregation is here understood as relating to reasons in one and the same context. It does not entail that for (unsuccessfully, as I agree with his critics), I derive it from a formal principle about reasons, together with substantive assumptions about rights and how they prevent aggregation—the two arguments are thus very different. ²⁵ See e.g. Schroeder (2007, 125–6), Kiesewetter (2018, 111–12). For an epistemic case of dependency, see Kiesewetter (2020, 339–40).
208
two independent reasons R1 and R2, if R1 supports ϕ-ing in a context C1 (in which R2 is absent), and R2 supports ϕ-ing in a context C2 (in which R1 is absent), then adding either reason to the other context increases the support for ϕ-ing. R1 and R2 may undermine or attenuate each other when they obtain simultaneously, to the effect that the overall support for X is not increased. This is what happens in the joke about the restaurant that is said to have two flaws: that the food is terrible and that the portions are too small.²⁶ But examples of this structure do not cast doubt on Reasons Aggregation understood as a claim about the aggregation of reasons in one and the same context. Reasons Aggregation is modest in another respect as well, as it allows, fourthly, that aggregation can be prevented. It says no more than that reasons aggregate per default, and that a deviation from this rule needs some kind of explanation. It is important to note that Reasons Aggregation is a formal principle about the logic of reasons and as such should be seen as independent of firstorder normative disagreements such as the one between consequentialist and deontological views. The principle claims to be true of epistemic reasons for belief no less than of practical or moral reasons. For example, it seems plausible to think that you have more reason to believe p if p is supported by two independent sources of information rather than just one of them. Reasons Aggregation provides a powerful explanation of all the undoubtedly many instances of the aggregation of reasons, which constitutes a strong reason for accepting it. Those who reject this principle face the enormous challenge of providing alternative explanations.²⁷ Critics of certain forms of aggregation are well-advised to focus instead on the explanation of non-aggregation—a possibility that the stated principle explicitly allows for. In line with Reasons Aggregation, deontologists who sympathize with Taurek’s view might hold that—perhaps due to some requirement to respect persons and their separateness—the aggregation of reasons is prevented whenever two reasons are based in different persons’ welfare. Although it is beyond the scope of this chapter to rule out this view, it strikes me as unnecessarily strong: If I can give pleasure either to one stranger or to a
²⁶ See Dancy (2004, 16), who ascribes this joke to Jerry Dworkin. Kagan (1988, 14–23) and Nair (2016) discuss structurally analogous examples as challenges for the aggregation of reasons. ²⁷ Munoz-Dardé (2005, 200–8) rejects the idea that a duty to save the greater number can be vindicated by an appeal to the aggregation of reasons. It is less obvious that she also rejects Reasons Aggregation, as her main worry seems to be that a preponderance of reasons does not entail a duty (see also below). Insofar as she rejects Reasons Aggregation, she faces the challenge of explaining why, as she also concedes, reasons “sometimes combine” (2005, 200).
209
billion other strangers, isn’t there more to be said in favor of the second option? A better approach, I suggest, says that what prevents reasons from aggregating are rights. We have already seen that a right/duty-nexus involves a pre-emptive reason—a reason that is protected from being outweighed by certain other considerations because it defeats them in a special way I described as “trump-like”. I now want to suggest that part of what this means is that pre-emptive reasons prevent certain weaker competing reasons from aggregating and outweighing the pre-emptive reason in combination.²⁸ Consider, for example, the right to bodily integrity. If the healthy patient in Thomson’s transplant case has such a right, this means that the surgeon has a preemptive reason not to remove her organs without consent. And that this reason is pre-emptive partly means that it is protected from being outweighed by any number of organ beneficiaries.²⁹ One might wonder why it should be part of pre-emptive defeat to prevent the aggregation of pre-empted reasons. Couldn’t we let these reasons aggregate and still consider them pre-empted? The answer depends on whether we should allow weaker reasons to pre-emptively defeat stronger reasons (or stronger sets of reasons). While this possibility is suggested by some accounts of pre-emptive defeat, I believe that it creates trouble and is better avoided. Most strikingly perhaps, it entails that one sometimes ought to act against one’s strongest reasons, which rules out widely shared and independently attractive views about the connection between what one has strongest reason to do on the one hand, and what one ought to do or is rationally permitted to do on the other. While I cannot discuss this question here in any more detail, I will proceed on the assumption that a weaker reason cannot defeat a stronger one (other than by undermining or attenuating it and thus rendering it false that the defeated reason is stronger), and that the same is true for sets of reasons that are stronger in combination. This motivates the idea that the point of pre-emptive defeat consists at least partly in a mechanism that prevents weaker reasons from gaining combinatorial weight through aggregation.
²⁸ It is not part of this view that pre-emptive reasons necessarily prevent all weaker reasons from aggregating. There may be relevant weaker reasons that are not-pre-empted and thus not prevented from aggregating (cf. Voorhoeve 2014). ²⁹ I assume that the reason against killing the healthy patient pre-empts the reason for saving the suffering patient. While this is a substantive normative assumption, it strikes me as dialectically unproblematic, as it is very likely to be shared by those who think that the healthy patient has an undefeated moral right not to be killed in the first place.
210
The idea that rights prevent the aggregation of reasons that are subject to pre-emption is, I think, a distinctively deontological one that is compatible with Reasons Aggregation, which underlines the point that Reasons Aggregation is neutral with respect to the dispute between consequentialism and deontology in ethics. Taken together with Reasons Aggregation, it can be employed on behalf of a deontological vindication of the Principle of Saving the Greater Number. Start with the assumption that everyone who is in need of help in a Taurek case has a pro tanto right, against agent A, to being saved. This entails that for every person in need, A has a pre-emptive reason to help her. That this reason is pre-emptive means, in part, that it prevents certain other potential reasons from aggregating. But note an important point. A reason cannot pre-empt another reason that is of the same kind and strength. If it were to do that, the pre-empted reason would likewise pre-empt the preemptive reason and the reasons would mutually pre-empt each other. But this is conceptually impossible. Pre-emption is a form of defeat: to say that a reason pre-empts a competing reason implies that it wins that competition. Winning, however, is asymmetric, and just as two competing reasons cannot both win against each other, or prevail over another, they cannot mutually pre-empt each other. It follows from this that the ban on aggregation that is involved in each of A’s reasons for saving a person does not apply to the other pre-emptive reasons for saving a person. These reasons aggregate, according to Reasons Aggregation, simply because nothing prevents them from doing so. Since they aggregate, there is overall more reason to save the greater number,³⁰ which means that the rights of those in the smaller group are defeated by stronger competing reasons, while the rights of those in the larger group are undefeated.³¹ As we are morally required to comply with
³⁰ This is on the assumption that the reasons for saving each are equally strong. One might instead hold that they are incomparable (or “on a par”) and that, since incomparable reasons are insensitive to certain kinds of improvements, aggregation does not necessitate the conclusion that there is most reason to save the greater number. However, the kind of improvements incomparable reasons are insensitive to are small improvements, while the improvement in a Taurek case is a highly significant one. This suggests that aggregation supports the conclusion that there is most reason to save the greater number even if the reasons in question are incomparable. ³¹ I assume that a right or duty is defeated iff for every set of reasons S that contains the reason that is constitutively related to the right or duty, there is a set of competing reasons S* that is at least as strong as S. Since the set of reasons for saving the people in the smaller group is not stronger than any set of competing reasons, the rights of these people are defeated, while the same cannot be said about the rights of the people in the larger group.
211
undefeated rights, A is morally required to save all of those who are part of the larger group. This justification of the duty to save the greater number is genuinely deontological, as it does not base this duty on any value ascribed to states of affairs, but on (i) individual rights that correlate with directed reasons, (ii) a formal principle about the aggregation of normative reasons in general, and (iii) a general rights-based account of the limits of aggregation, which explains why aggregation is prevented in the cases that are uncontroversial among deontologists. These assumptions are not uncontroversial, of course, and they are in need of more defense than I can give them here. But neither are they ad hoc or pre-theoretically implausible, and together they make for an attractive view on the matter that merits serious consideration. One might wonder why rights contributionism is an essential part of this story. Couldn’t we vindicate a duty to save the greater number just by appeal to aggregating reasons for saving lives without assuming that these reasons correspond to rights?³² I don’t want to insist that this is impossible. But critics of such an approach have pointed out that there is a considerable gap between judging that the balance of reasons favors an action and judging that there is a duty to perform it.³³ By contrast, the rights contributionist argument for the duty to save the greater number does not rely on any assumption that duties can be derived from the balance of reasons. It derives this duty from the fact that the people in the larger group have an undefeated right to be saved. In this argument, Reasons Aggregation is needed to explain why the duties to those in the smaller group are defeated, while the duties to those in the larger group are not—but there is no assumption at work that there is a duty to save the greater number because there is stronger reason to do so. The appeal to contributory rights in Taurek cases thus makes a significant dialectical difference.
7. Conclusion Let me briefly wrap up this chapter. In sections 1–4, I argued that pro tanto reasons can have directed deontic structure—a structure that is characteristic of and plausibly entails a right/duty-nexus. In sections 5–6, I turned to possible implications of this result, focusing on the numbers debate in ³² Some incidental remarks of Parfit (2003, 388) suggest that he was sympathetic to this view. ³³ See esp. Munoz-Dardé (2005, 202–8).
212
normative ethics. I argued that deontologists can provide a vindication of the duty to save the greater number that appeals to pro tanto rights to be rescued and an independently attractive account of how and when rightscorrelating reasons aggregate or are prevented from doing so. As noted already, more needs to be said to defend the suggested approach to rescue conflicts, especially against views according to which arbitrary choice or randomized procedures are permissible (or even required) in Taurek cases. I did not have space to discuss either the anti-aggregationist worry that fairness requires holding a lottery rather than saving the greater number, or the threshold-deontological worry that the present view cannot accommodate certain allegedly plausible trade-offs (in particular trade-offs between a pre-emptive reason and a huge number of pre-empted reasons). And while the correlation between rights and reasons with directed deontic structure argued for in sections 1–4 provides support for the assumption that there are rights to be rescued, I have not defended this assumption here against contrary views.³⁴ I hope to rectify these omissions on another occasion. Here I limit myself to the more modest conclusion that rights contributionism clears the way for an interesting, prima facie attractive, and distinctively deontological justification of the duty to save the greater number.³⁵
References Adams, N. P. 2021. “In Defense of Exclusionary Reasons”. Philosophical Studies 178 (1): 235–53. Bedke, Matthew S. 2011. “Passing the Deontic Buck”. Oxford Studies in Metaethics 6: 128–52.
³⁴ See e.g. Foot (1977, 44–9), Thomson (1990, 160–3). ³⁵ For valuable feedback on earlier versions of this chapter, I am very grateful to Rowan Cruft, Philip Fox, Jan Gertken, Tim Henning, Felix Koch, Kirsten Meyer, Phil Montague, Thomas Schmidt, Julie Tannenbaum, Mark Timmons, Andreas Vassiliou, Jay Wallace, Jack Woods, two anonymous referees, and audiences at Bayreuth, Berlin, Bielefeld, Erlangen, Heidelberg, Lillehammer, Oxford, Southampton, Stuttgart, Zurich, and the Arizona Workshop in Normative Ethics 2022. Work on this chapter has been funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation, Centre for Advanced Studies “Human Abilities”, project no. 409272951) and the European Union (Horizon Europe, ERC Grant 101040439, REASONS F1RST). Views and opinions expressed are however those of the author only and do not necessarily reflect those of the European Union or the European Research Council Executive Agency. Neither the European Union nor the granting authority can be held responsible for them.
213
Dancy, Jonathan. 2004. Ethics without Principles. Oxford: Oxford University Press. Doggett, Tyler. 2009. “What Is Wrong with Kamm and Scanlon’s Argument against Taurek”. Journal of Ethics and Social Philosophy 3 (3): 1–16. Dworkin, Ronald. 1984. “Rights as Trumps”. In Theories of Rights, edited by Jeremy Waldron, 153–67. Oxford: Oxford University Press. Feinberg, Joel. 1978. “Voluntary Euthanasia and the Inalienable Right to Life”. Philosophy & Public Affairs 7 (2): 93–123. Foot, Philippa. 1967. “The Problem of Abortion and the Doctrine of the Double Effect”. In Virtues and Vices, 19–32. Oxford: Oxford University Press (2002). Foot, Philippa. 1977. “Euthanasia”. In Virtues and Vices, 33–61. Oxford: Oxford University Press (2002). Hart, H. L. A. 1955. “Are There Any Natural Rights?” The Philosophical Review 64 (2): 175–91. Hohfeld, Wesley Newcomb. 1913. “Some Fundamental Legal Conceptions as Applied in Judicial Reasoning”. The Yale Law Journal 23 (1): 16–59. Kagan, Shelly. 1988. “The Additive Fallacy”. Ethics 99 (1): 5–31. Kamm, Frances Myrna. 1993. Morality, Mortality: Death and Whom to Save from It. Vol. 1. New York: Oxford University Press. Kiesewetter, Benjamin. 2018. “Contrary-to-Duty Scenarios, Deontic Dilemmas, and Transmission Principles”. Ethics 129 (1): 98–115. Kiesewetter, Benjamin. 2020. “Rationality as Reasons-Responsiveness”. Australasian Philosophical Review 4 (4): 332–42. McDowell, John. 1980. “The Role of Eudaimonia in Aristotle’s Ethics”. In Mind, Value, and Reality, 3–22. Cambridge, MA: Harvard University Press (1998, 2nd ed. 2002). Montague, Phillip. 1984. “Rights and Duties of Compensation”. Philosophy & Public Affairs 13 (1): 79–88. Montague, Phillip. 1988. “When Rights Are Permissibly Infringed”. Philosophical Studies 53 (3): 347–66. Munoz-Dardé, Véronique. 2005. “The Distribution of Numbers and the Comprehensiveness of Reasons”. Proceedings of the Aristotelian Society 105 (1): 191–217. Nair, Shyam. 2016. “How Do Reasons Accrue?” In Weighing Reasons, edited by Errol Lord and Barry Maguire, 56–73. New York: Oxford University Press. Nozick, Robert. 1974. Anarchy, State, and Utopia. New York: Basic Books. Otsuka, Michael. 2000. “Scanlon and the Claims of the Many versus the One”. Analysis 60 (3): 288–93.
214
Parfit, Derek. 2003. “Justifiability to Each Person”. Ratio 16 (4): 368–90. Raz, Joseph. 1990. Practical Reason and Norms. Oxford: Oxford University Press (2nd ed., repr. 1999). Scanlon, T. M. 1998. What We Owe to Each Other. Cambridge, MA: Belknap Press. Schroeder, Mark. 2007. Slaves of the Passions. Oxford: Oxford University Press. Shafer-Landau, Russ. 1995. “Specifying Absolute Rights”. Arizona Law Review 37 (1): 209–25. Stratton-Lake, Philip. 2002. “Introduction”. In Ethical Intuitionism: ReEvaluations, edited by Philip Stratton-Lake, 1–28. Oxford: Clarendon Press. Taurek, John M. 1977. “Should the Numbers Count?” Philosophy & Public Affairs 6 (4): 293–316. Thomson, Judith Jarvis. 1976. “Self-Defense and Rights”. In Rights, Restitution, and Risk, 33–48. Cambridge, MA: Harvard University Press (1986). Thomson, Judith Jarvis. 1980. “Rights and Compensation”. In Rights, Restitution, and Risk, 66–77. Cambridge, MA: Harvard University Press (1986). Thomson, Judith Jarvis. 1990. The Realm of Rights. Cambridge, MA: Harvard University Press. Timmermann, Jens. 2004. “The Individualist Lottery: How People Count, but Not Their Numbers”. Analysis 64 (2): 106–12. Voorhoeve, Alex. 2014. “How Should We Aggregate Competing Claims?” Ethics 125 (1): 64–87. Wallace, R. Jay. 2013. “The Deontic Structure of Morality”. In Thinking about Reasons. Essays in Honour of Jonathan Dancy, edited by David Bakhurst, Brad Hooker, and Margaret Olivia Little, 137–67. Oxford: Oxford University Press. Wallace, R. Jay. 2019. The Moral Nexus. Princeton: Princeton University Press. Whiting, Daniel. 2017. “Against Second-Order Reasons”. Noûs 51 (2): 398–420. Williams, Bernard. 1976. “Moral Luck”. In Moral Luck, 20–39. Cambridge: Cambridge University Press (1981).
10 Can’t Kant Count? Innumerate Views on Saving the Many over Saving the Few Sergio Tenenbaum
1. The Specter of Numeracy Let us start with a version of a well-known vignette from the ethical literature: Basic Case I am in my boat in a bay which forks into stream A and stream B. I just received a distress signal from two different boats: one carrying two complete strangers (whom I name “Larry” and ”Mary” for easy reference) and another carrying one other person, also a complete stranger (whom I name “Jeri”). If I row to stream A, I can save Larry and Mary, but Jeri will die; if I row to stream B, I can save Jeri, but Mary and Larry will die. It is a popular view that in such a situation, I am under obligation to save Larry and Mary. Generalizing the point, we can formulate the following principle: The More the Merrier In a situation in which you can either save a group of n strangers or m strangers at no great effort to yourself, in the absence of any partiality or other morally relevant consideration, you ought to save n strangers if n > m. Maximizing consequentialists have an easy time explaining this principle: assuming that that the expected value of saving the life of a person is the same (at least as long as I know no other relevant information about them), the expected value of saving two lives is greater than the expected value of saving one life. In fact, this rationale might seem pretty compelling: it is just Sergio Tenenbaum, Can’t Kant Count? Innumerate Views on Saving the Many over Saving the Few In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Sergio Tenenbaum 2023. DOI: 10.1093/oso/9780198895909.003.0011
216
common sense that it is better to save more lives over fewer. Other types of ethical theories have a much more difficult time defending the principle; in fact, some philosophers who reject impartial consequentialism are willing to accept that in Basic Case, it is permissible to save Jeri, or that the right action is something different from choosing the outcome in which I save more lives. Famously, John Taurek (1977) argues that there is no obligation to save the many against the few, suggesting that he would flip a coin if he were caught up in this situation. Taurek rejects the central idea that he takes to ground The More the Merrier; namely, that there is an impersonal concept of “good simpliciter” that applies to outcomes or states-of-affairs, such that the more lives we save, the better the outcome or the state-of-affairs is. Saving Larry’s life is better for Larry and saving Mary’s life is better for Mary, but there’s no such thing as something being better simpliciter or better-for-Larry-and-Mary-and-Jeri, so that we can say that saving Larry and Mary is better than, say, saving Jeri. But this kind of “Geachean” skepticism about “good simpliciter” is not the only view that finds it difficult to vindicate The More the Merrier. Kantian views committed to the unconditional value of rational agents seem equally inhospitable to the principle, and contractualist views also face difficulties in generating principles that would deliver the same verdict as The More the Merrier. All these non-consequentialist views seem committed to the claim that in situations in which our options are to bring about that either n lives are saved or that m lives are saved, the fact that n is greater than m does not, at the most fundamental level, create an obligation to save n lives over m lives; at least, not on the simple grounds that it is better to save more lives. In other words, they are committed to what Parfit calls “innumeracy”. Innumeracy is not strictly incompatible with The More the Merrier, since it only makes a claim about what is true at the most fundamental level, but the tension is clear. Taurek is happy to reject principles like The More the Merrier, but others have been less cavalier about it; many find it extremely implausible that we should not favor saving the many over the few. However, given the tension between the principle and innumeracy, it is tempting to see non-consequentialist attempts at vindicating it as ad hoc maneuvering to accommodate what are essentially consequentialist intuitions. My aim here is to propose a middle way between Taurek’s bold denial that it is preferable to save the many in our Basic Case and endorsing a (seemingly) consequentialist principle. Although, philosophers generally reject Taurek’s conclusion that it is permissible to save the few, I will first present a series of cases in which no
’ ? 217 significant partiality is involved, but in which there seems to be nothing wrong with saving the few.¹ If this is correct then we should not grant that The More the Merrier is true in full generality. Yet, I think there is a very general argument in favor of accepting that we ought to save Larry and Mary in Basic Case that neither depends on, nor vindicates, The More the Merrier. In other words, I’ll argue that we need not sneak in any form of consequentialism to conclude that we ought to save the many, at least for an important subset of cases. Before we move on, a couple of caveats. Contemporary work on non-consequentialist aggregation has also considered questions of permissible trade-offs in cases in which the harm that I can prevent is different for different people.² These are interesting and important issues that need to be addressed by the non-consequentialist but I cannot do justice to them within the scope of this paper. Similarly, I’ll leave aside questions about how to extend any of the results here to contexts of risk and uncertainty.³ Finally, although I will mostly assume a Kantian framework for ease of presentation, the argument is compatible with different versions of non-consequentialism; it should thus be greeted as good news by a large sector of the nonconsequentialist population. In fact, the variations on Basic Case I’ll present should give pause to any non-consequentialist (or anyone at all, I think) who would be willing to endorse The More the Merrier. If I am right that this principle leads us astray in such cases, then other non-consequentialists should prefer my proposal for handling Basic Case over any view that is committed to The More the Merrier.
2. Saving the Few Most philosophers agree that if Jeri is your beloved daughter or spouse, you are permitted to save her rather than two strangers.⁴ Similarly, if Larry, Mary, and Jeri find themselves in this bind due to Larry’s and Mary’s ¹ Taurek says that he would flip a coin if faced with a situation where he could give a lifesaving pill to one person or five persons, but he does not seem to deny that it would also be permissible to save just one person. My argument is that in many circumstances we are permitted to save the few, and not just to flip a coin, and that in the cases that Taurek looks at we are obligated to save the many. So I’ll not be discussing, except in passim, questions about the fairness of this lottery as opposed to others. See Saunders (2009) for a discussion of these ideas. ² See Tomlin (2017) and Horton (2018) for skepticism about “limited aggregation” in these scenarios; for a recent positive proposal, see Steuwer (2021). ³ For my own views on this matter, see Tenenbaum (2017). ⁴ But they disagree on what types of partiality permit you to save the few. On Setiya’s view (Setiya, 2014), since we do not need a reason to love someone, there is no limit on how wide the
218
perfidious actions, I might be permitted, or even obligated, to choose to save Jeri. But are situations involving partiality (or involving other independently weighty considerations such as liability) the only ones in which I am permitted to save the few? Let us look at a variation of Peter Singer’s (Singer, 1972) famous case of saving a child at the expense of valuable clothing: Rescuing Baby Chris Sam is near a pond and her phone rings. A remotely located lifeguard says: “Our cameras show you near El Pond. This is fortunate; you can rescue Baby Chris! Chris will soon drown unless you come to his aid. I know you can’t swim, but the water is not too deep for you, so you can easily walk and fetch little Chris. Just make sure to keep your shoes on, otherwise you won’t be able to make it through the rocks to save the toddler.” Sam locates Chris, waves at him, he smiles back and she feels very happy she can save this baby. Of course, Sam does not care that she is going to wreck her brand new very expensive shoes; she just goes into the water. Sam starts wading towards Chris when the lifeguard calls her again: “You won’t believe this; there are two babies drowning in a pond not far from you. You don’t have much time left; leave Chris behind and go save the two babies.” Sam refuses to leave Chris behind, and continues in her rescue. Had Sam known before she started her rescue mission that there were two babies in an adjacent pond (while only Chris was to be saved in this one), she would certainly, pace Taurek, have gone to the adjacent pond. But it now seems permissible that Sam refuses to change course, and insists on rescuing Chris. Sam entered the pond to save Chris and now having settled on this course of action, it seems reasonable that she is not willing to leave him behind. Of course, as a virtuous agent, she must regret that she cannot save the other babies. But she can only do that if she abandons Chris to his death, and she now (reasonably) feels she must ensure that Chris is safe, even if she could save two other babies instead. Two observations are important here. Sam’s relationship with Chris is quite minimal. If one accepts that Sam is permitted to continue,⁵ it is not
partiality permission is, so his view “collapses” into Taurek’s (But Setiya (2023) defends a view closer to my view here). Parfit (1978) rejects the claim that casual acquaintances can ground these types of permissions. ⁵ Some philosophers with whom I discussed this case think you are obligated to save Chris. For my purposes, it is enough to accept a permission.
’ ? 219 because of any kind of partiality towards Chris, but simply because Sam set up to save Chris in particular. This conclusion seems to fit well with the Kantian view: on the Kantian view, a person has “dignity”, rather than a “price”. At the core of the distinction is the idea that you can “trade up” when things have a price, but not when they have dignity. Suppose we had instead a less virtuous person in the pond’s vicinity (Sammy); Sammy would like to save Chris, but Sammy can’t bear the thought of wrecking such expensive shoes. Suppose now that noticing Sammy’s hesitation, the lifeguard offers Sammy identical shoes and a significant sum of extra money. Sammy would probably be irrational if he were to turn down this offer, even ignoring the importance of saving Chris; after all, Sammy would be just trading up. But when Sam is offered the opportunity to rescue two toddlers, she would not be trading up if she accepted the deal; that which has dignity, “cannot be brought into comparison or competition at all, without assaulting its holiness” [4:435]. On the other hand, Sam’s decision to save Chris does not sit well with numerate views; the fact that Sam is on her way to save Chris, the fact that Chris smiled at Sam, or the fact that Sam is thirty feet closer to Chris: none of these properly compensate for the fact that the number of lives to be saved is cut in half. Someone could protest that the judgment that it is permissible to save Chris rests on an illusion; perhaps the inclination to judge it permissible to save Chris is an instance of some kind of “sunk cost fallacy”.⁶ But here lies a dangerous path. Is it essential to the scenario that all these toddlers are threatened by drowning and will be saved by Sam’s personal efforts? Would it be permissible for Sam not to give her expensive shoes to someone else who needs them to barter for a canoe that will allow them to save two other toddlers? It is hard to see how the claim that it is impermissible that Sam save Chris is compatible with our ordinary judgments about the duty of rescue in Singer’s original example. After all, one can always sell the clothing that will be wrecked in the rescue process and save more lives with the proceedings; it might be difficult to give up the idea that Sam can save Chris without radically changing our views about more ordinary rescue cases.⁷ ⁶ But note that the “sunk cost fallacy” explanation would predict that we would think that Sam is justified in continuing even before she had any acquaintance with who was being saved in the pond. I think this prediction is false; had she found out that she could save two babies instead halfway to the pond, the situation would be indistinguishable from our Basic Case. ⁷ It is quite difficult to delineate precisely the difference between the duty of rescue and other duties of aid (see, for instance, Igneski (2006)). My point is that we are already committed to reject in other situations that being able to save more lives always trumps the duty to save this life here and now even when all involved are strangers.
220
Of course, none of this is conclusive. However, at the very least, in such cases one cannot accuse the Kantian position of offending against common sense for allowing us to save Chris even if one were inclined (correctly to my mind) to regard it as an embarrassment if it implied the permissibility of saving the few in Basic Case. But I think this setup also helps us display the power of the Kantian position. Taurek asks whether it is permissible to save David, my casual acquaintance,⁸ instead of saving more people. From a perspective prior to engaging in any action, it does seem too flimsy a relation to override my reason to save one more person.⁹ However, things look different if I am already engaged in the action of saving the one. On the Kantian view, the reasons that Sam has to save two toddlers do not defeat the reason Sam has to save Chris. The lifeguard has not given Sam a reason that obligates her to stop what she is doing. Of course, the same relations of defeat apply before Sam starts engaging in the action (if, say, Sam knew all along that these two options were open to her). But when Sam is deciding between saving one or saving two, some other considerations may apply. Sam might need to act on a principle that is justifiable to everyone involved, or find a reason, or a selection criterion, that is fitting to the momentous decision she faces. Or, as I will argue, there might be something essentially different about the structure of the agent’s choice in a case like Basic Case. Of course, this all needs further grounding and explanation. In the next sections I try to explain more precisely how the Kantian can distinguish between Basic Case and Rescuing Baby Chris.
3. Kantianism, Incomparability, and Choice Principles According to Kant, a person has “dignity, that is, an unconditional, incomparable worth” (4:436). In the kingdom of ends: everything has either a price or dignity. What has a price can be replaced by something else . . . What . . . is raised above all price and therefore admits of no equivalent has dignity. (4:434)
Of course, I will not attempt here a detailed interpretation of Kant’s notion of dignity. But in saying that there is no equivalent to the value of a human ⁸ Setiya (2014) helpfully describes David as the barista at my local cafe in order to illustrate the nature of the relation. ⁹ More on this issue later.
’ ? 221 being, to their dignity, I think Kant is committed to denying that the value of two human beings is equal, at least in an important sense.¹⁰ This is exactly what distinguishes dignity from price: things that have a price can be replaced by things of equal value, but the “unconditional worth” of human beings does not allow for similar trade-offs. Obviously one person has neither more nor less value than another. But it seems that Kant also wants to rule out that they are of “equal value” if we understand as consequence of this evaluation that one could be replaced by another. It is notoriously difficult to determine what the dignity of each person implies in terms of the permissibility of various actions. But I think we can sketch some basic consequences of Kant’s notion of dignity for our purposes. Suppose I am about to save one person, and you point out that I might save two people instead. “Why would the fact that I can save two people create an obligation to change what I was planning to do?”, I ask. You answer as follows: whatever value there is in saving the life of one person, there will be significantly more value (twice as much?) in saving the lives of two people. Setting aside partiality, or special circumstances, one must choose the option of saving two lives. This kind of reasoning seems to go against Kant’s claims about the special value of human beings, against the idea that they have dignity. Dignity rules out this kind of calculation of value; the above reasoning treats human beings as if they had a price that could be matched and surpassed by the price that the lives of two human beings fetch. Given that human beings have dignity, the reason to save a human life cannot be defeated by the fact that one could save two lives instead. Kant seems committed to the following thesis: No Defeat Principle The reason I have to save one person is not defeated by the (supposed) greater value of saving two lives. The No Defeat Principle explains why the lifeguard’s appeal does not necessarily defeat the reason that Sam must save Chris. And it is worth mentioning that virtually any innumerate view seems committed to the No Defeat Principle. Contractualist views rule out adding the justification of two ¹⁰ Of course, Kant does put immense importance in a principle of equality among human beings, especially in his political philosophy. But this is not to say that people have equivalent values in the sense presented here.
222
independent people in the manner presupposed by simply adding the values of two lives. In refusing to allow a value that is not a value for someone, Taurek’s view is also committed to the No Defeat Principle. Thus if we can show how commitment to this principle is compatible with providing an adequate rationale for saving the many in Basic Case in terms that are clearly acceptable to all these theories, we will have provided a powerful rationale for saving the many that is compatible with a wide array of views that reject (foundational) numeracy. Of course, this is not all that the Kantian view has to say about our relationship to the lives of others. The Kantian view, like most ethical theories, also imposes a duty to aid. For Kant, benevolence is an imperfect duty, and thus it is a duty of wide obligation. It “has in it a latitude for doing more or less, and no specific limits can be assigned to what should be done” (6:393). This latitude is not unconstrained and at least in some contexts the duty to aid imposes obligations to perform particular acts. For the purposes of this paper, I’ll just assume that among these more specific duties, we have a duty of easy rescue that I will formulate as follows: Minimal Duty to Aid In situations in which you can directly save the life of a human being at (nearly) no effort or sacrifice to yourself, you have a duty to save their life. A few caveats about Minimal Duty to Aid. First, what counts as “nearly no effort or sacrifice to ourselves” is left open; obviously, the more we allow as “nearly no effort or sacrifice”, the more demanding the duty will be. Secondly, Peter Singer’s famous argument (Singer, 1972) seems to show that a principle like Minimal Duty to Aid creates a massively demanding duty to aid requiring that relatively affluent people make huge sacrifices for the sake of those in desperate straits elsewhere. One way of blocking this argument is to challenge Singer’s (apparent) assumption that we determine what counts as “nearly no effort” by looking at each action separately rather than examining a pattern of repeated actions. Perhaps what counts as “nearly no effort” depends not only on whether it is “nearly no effort or sacrifice” on this occasion but also on whether a pattern of performing such actions at each occasion will require great sacrifices from me.¹¹ My end here
¹¹ For responses to Singer’s argument roughly along these lines, see Thomson (2022); Timmerman (2015).
’ ? 223 is not to assess such a view, but to preempt some possible objections to the following principle: Pareto Principle If you can save A and B instead of just A at nearly no effort or sacrifice to yourself, then you should save A and B. This principle follows from Minimal Duty to Aid; in fact, it is just an instance of Minimal Duty to Aid applied to the case in which it is already given that you will save one person. But, of course, we need to be careful not to let this consequence necessarily commit ourselves to accepting Singer’s argument. So if we block the argument by accepting a version of the above proposal, then we’ll need to understand the extra effort or sacrifice not only by its measure on a particular occasion but also by what it demands of us if we act in similar ways in similar occasions. Other ways of blocking the argument will restrict the principle in different ways. Whatever view one favors here does not matter for my argument. The instances of the Pareto Principle that will be relevant for our purposes are rather uncontroversial. For instance, I might need just to extend my arm one more time to save the second person. Some of the cases involve literally no extra effort or sacrifice to save the second person; Taurek himself seems to accept some version of the Pareto Principle. In Basic Case, I have a reason to rescue Jeri given that her life is in danger. But I also have a reason to rescue Larry (alone), a reason to rescue Mary (alone), and a reason to save Jeri, and a reason to save both Mary and Larry. The Pareto Principle ensures that my reason to save Mary and Larry defeats any reason I have to save just Larry or just Mary. No Defeat Principle ensures, however, that my reason to save Larry and Mary does not defeat my reason to save Jeri or vice-versa. Since we have undefeated reason to act on either option, neither seems impermissible. So it seems that any view committed to No Defeat Principle, including the Kantian view, must accept that there is no obligation to save the many. We could preserve innumeracy and yet not be committed to Taurek’s view of Basic Case if we can argue that the Pareto Principle applies to Basic Case. But this might appear to be ruled out by the No Defeat Principle. After all, the Kantian view on the dignity of persons, at least as we interpreted it, seems to rule out exactly this kind of comparison between one and two lives. More generally, allowing the Pareto Principle to be used in this manner
224
seems indistinguishable from accepting that the value of two lives is greater than the value of one life; it would be just a form of ushering numeracy in through the back door. However, appearances here are misleading. We can use the Pareto Principle to deliver an obligation to save the many in the relevant cases. Or so I will argue.
4. Sweetening and No Defeat So far the combination of No Defeat Principle and Pareto Principle seem to imply the following: 1. If Sam has the option of saving baby Chris or baby A and baby B, Sam may save Chris (the reason to save Chris is not defeated by the reason to save baby A and baby B) (given No Defeat Principle). 2. If Sam has the option of saving Chris or saving Chris and baby A (or baby B), Sam must save both Chris and baby A (or baby B) (given Pareto Principle). Of course, there is nothing special about baby Chris or lacking with baby A or baby B. Either of the above claims would be true if we replace the babies that appear in each position. So for instance the following is also true: 3. If Sam has the option of saving baby A or Chris and baby B, Sam may save baby A (the reason to save baby A is not defeated by the reason to save Chris and baby B). These defeat relations are very similar to those familiar from the literature on incommensurability and incomparability. Take, for instance, Ruth Chang’s understanding of the “parity” (Chang, 2002) relation; if A and B are on a par, then it is rational to choose A or to choose B even though A and B are not of equal value or indifferent. Options on a par, unlike indifferent ones, resist “sweetening”. That is, if option A and option B are on a par, typically, option A is still on a par when we switch option B with a mildly improved version of it. So if going to restaurant a and going to restaurant b are on a par, they will still be on a par if we make a small improvement to one of the options (by, say, adding an offer for a free amuse bouche to restaurant a or a $1-off coupon to restaurant b). Of course, not all the defeat relations are the same, and the Kantian position is significantly
’ ? 225 different from the claim that the values of the lives of two people are on a par. In the parity cases, typically significant amounts of “sweetening” will change the relation: so, if restaurant a offers the meal for free or if they offer me an annual subscription to the restaurant for the price of the meal, it will probably suffice to make the option of going to restaurant a better than the option of going to restaurant b. But we cannot assume that the same is true for the Kantian view. The Kantian view might conclude that there is nothing that you can add to the option “saving another baby” that will defeat the option “saving Chris”.¹² However, the overlap on defeat relations is enough to learn an important lesson from a variation on the typical cases of choices among items on a par. In particular, looking at how choices between items on a par behave in certain cases of ignorance will help us see how the Pareto Principle applies to Basic Case.
5. Sweetening in the Dark As we said above, the resistance to “sweetening” in the case of alternatives that are on a par parallels the structure of the choices among saving lives in Kantian views. Hare (2010) points out that, in the case of parity, some complications arise if we lack knowledge of the options. Let us look at a slightly modified version of Hare’s central example: My house is on fire and I can direct the firefighters to either rescue my rare and unique Fabergé egg or to save my wedding album. I find it difficult to decide between them; neither seems to be more valuable to me than the other, but they are also not of equal value to me. These value relations remain unchanged if the firefighters can also rescue a $100 dollar bill if, and only if, they rescue the wedding album (though, of course, I must prefer that they save the wedding album and the $100 dollar bill over saving just the wedding album). Hare considers an interesting variation of this classic type of example. Suppose that I keep these valuables in identical boxes and I can’t remember which box holds which prized possession; they are equally likely to be in
¹² Although this might seem plausible when the numbers are small, many will balk at this consequence when the numbers get large enough (say, if we could save millions of babies instead of baby Chris). On my view, in order to accept this “intuition”, the Kantian needs to show that at some point the difference in numbers makes a qualitative difference; or, that if the numbers are high enough that the nature of the available actions change. A proper examination of these cases, though, is beyond the scope of this paper (for a related approach for cases involving risk, see my (Tenenbaum, 2017)).
226
either box. The firefighters can see that there is a $100 bill next to one of the boxes, and they ask me if I would rather that they save this box. So my options are roughly as follows: Table 1: Sweetening in the Dark Box A
Box B
.5 (album, egg)
.5 (album, egg); $100
In such a case it seems rational for me to ask the firefighters to save Box B. After all, it seems that Box B prospects are superior no matter how the album and the egg compare to each other. But something seems strange here: no matter what is in Box A, if I knew what was in Box A, I would be rationally permitted to choose Box A. Hare imagines a better-informed self who knows what is in each box. I know that my better-informed self is permitted to choose it. But then why am I not permitted to choose Box A? Hare himself thinks that there are two plausible decision theories at conflict here: prospectism and deferentialism; roughly, a clash between a principle that tells you to choose the best prospect given the probabilities and the values of each option, and a principle that allows you to choose any option that you know would be permissible if you had all the relevant information. Although Hare favors prospectism, he concedes that he has no conclusive argument against deferentialism. I think Hare’s characterization of the choice situation in terms of the decision theory framework obscures some features of decisions with options on a par. To my mind, there is no question that Box B is the only rational option here. Imagine trying to justify a decision to take Box A. Suppose someone said, “I took Box A because it might have had something that I would have chosen over the contents of Box B even if I were to get an extra $100.” However the same is true of Box B, except that if you choose Box B, you get an extra $100. Nonetheless I agree that we need an explanation of why my knowledge that my better-informed self might choose Box A is not a consideration that permits me to choose Box A. Let us look at my choice when I know that Box A has the Fabergé egg, and Box B the wedding album, and neither option affords me an extra $100. I appreciate the unique beauty and the historical significance of a Fabergé egg. This is clearly sufficient reason to choose the egg. On the other hand, the wedding album’s value has a radically different source; its relation to my life
’ ? 227 with my beloved, as well as the memories of that wonderful day, also constitute sufficient reason to choose the wedding album. In such a case if I choose Box A, I choose it because it contains the Fabergé egg (or perhaps, because it instantiates the value-constituting properties of a Fabergé egg). I’m not indifferent between these two options; we should not represent this choice as a comparative choice between two values whose quantitative nature is such that it permits me to choose one or the other. Rather, the fact that Box A contains the Fabergé egg is a good reason to choose Box A and this suffices to make choosing Box A permissible, as long as there is nothing about Box B that defeats this reason. Of course, the same story mutatis mutandis would explain the permissibility of my choosing Box B if I were to choose Box B. And, on the assumption that the relevant facts are known, adding $100 to either option is not enough to change any of these facts. It is worth noting that in the case in which I know what each box contains and I can’t make up my mind, it is permissible to randomize my choice by, say, flipping a coin. Moreover, in the absence of a coin to flip, I could use various procedures in lieu of a more strictly randomizing procedure; there would be nothing wrong in picking the property of “being the longer box” as settling which one I choose. But it would be wrong to think that my reason to choose Box A is that it is the one that contains the Fabergé egg and it is the longer box; in fact, if this were the reason that would be something wrong with how I was making my choice. And this can be seen by the fact that “being the longer box” cannot play the same role in the case in which we sweeten the opaque box. If it is rationally impermissible to pick Box A in such a case, it is also impermissible to pick it once I realize that Box A is the longer box. Relatedly, what allows me to rely on the Pareto Principle in the case of ignorance is not that I have no way to pick out uniquely the object that is in Box A (which I know to be on a par with object B). In fact, the description “The object in Box A” picks out the Fabergé egg and is available to me. But since ‘being in Box A’ is irrelevant to my reasons for choosing either option, my choice situation is still the one described above. In sum: In the scenario in which I know what each box contains, the reasons to choose Box B do not defeat the reasons to choose Box A (or viceversa). But the same is not true if I do not know what is in the boxes. I cannot choose Box A because the Fabergé is there, because that fact is not accessible to me. I could choose Box A because there might be a Fabergé egg there, but this possibility is there on Box B too. With respect to the Fabergé egg and the wedding album, each box offers me the same reason to choose it: it
228
might contain (it has .5 chance of containing) a Fabergé egg; it might contain (it has .5 chance of containing) the wedding album. Box B on the other hand offers something that Box A does not offer: $100. So Box B is favored by an Extended Pareto Principle: In situations involving a choice between φ-ing and ψ-ing, where φ-ing does better than ψ-ing in realizing at least one value, and does better or equal in realizing all other values involved in the choice, a rational agent chooses to φ, if she chooses to φ or ψ. It should be clear by now why the fact that my better-informed self would be permitted to choose Box A carries no weight: I simply have no access to the reason they have when making their decision. The permissibility of an action, at least its subjective permissibility, depends on which reasons are accessible to me, or, in other words, which reasons I can act on.¹³
6. Kant to the Rescue A similar reasoning applies to Basic Case, or so I will argue now. Suppose I am told that I have the option to save one stranger’s life (A) or the lives of two strangers (B and C). Each person in the universe has dignity and if I was saving one of them, the No Defeat Principle would apply and I would not be required to abandon saving the one for the sake of saving the two. But this is not my situation when I am deciding what to do. I know nothing of the persons I might save and thus my position is a choice in which each person that I would save could be anyone, but in one option I get to save one person who could be anyone (just like in the choice of one person), and also an opportunity to save another person who could also be anyone, except not a person identical to the first one. In other words, the choice is between saving a person and saving a person plus yet one more person (See Table 2 for a rough version of the choice). I do have available to myself descriptions that ¹³ I am assuming that there is a legitimate notion of subjective permissibility (whether or not there is also a notion of objective permissibility), both in the case of morality and in the case of instrumental rationality, and that this is the relevant notion from the point of view of the agent making a decision. Even if one were to reject such a notion, one would still need to say something about principles that guide rational or virtuous agents making decisions under limited information, and these are the principles that are relevant here.
’ ? 229 Table 2: Choice Among Strangers Option A
Option B
1/n(person₁ , . . . , personn)
1/n(person₁ , . . . , personn); 1/n
1(person₁, . . . , personn-1)
will successfully pick out person A (like “the person who I can save but only singly”), but since these predicates are irrelevant for my reason to engage in either action, they do not change my choice situation. But exactly what is relevant? Well, dignity is a value a person has, on Kant’s view, simply in virtue of being a rational agent. So it seems that no property of the person will be relevant for my reasons to save them, and we would end up with the surprising conclusion, if welcomed by some, that the Pareto Principle always applies and one should always save the many over the few. However, the matter is not so obvious. First, there are at least some people who stand in certain relations to me that are obviously relevant and that can potentially provide reasons that defeat my reason to save a person. As we said above, partiality (or liability) might provide reasons to save my relation (or the non-liable person) that defeat the reason to save a person. More importantly for our purposes, dignity attaches not to a generic property or to a trope, but to the particular person who has the property in question. Though a particular agent has intrinsic value in virtue of possessing this property, the value attaches to them in particular, rather than to the state-of-affairs such that the property of personhood is instantiated. So as long as my reason to act picks out the particular person I am about to save rather than a person, the Pareto Principle will not apply to it, and my reason to save this person will be undefeated by my reason to save the many. In the case of baby Chris, this is exactly what happens: as Sam is already engaged in saving baby Chris, her thought is already directed at baby Chris himself and thus her saving him is undefeated by the opportunity to save two babies. This would be true even before she enters the water: at the moment she sees baby Chris she has settled on saving baby Chris himself, and her reason is a reason to save this person in particular. In fact, had she received the update just as she started going in the direction of the pond, but before she makes any acquaintance with baby Chris, it would be difficult to justify not turning around and saving the two children instead. And even if you thought that a mere acquaintance is not enough to generate a proper
230
reason of partiality to save David, the barista of my local cafe,¹⁴ in another scenario this might end up sufficing to allow me to continue in my endeavor to save David. For instance, suppose I were at a beach and learned that a large ship had sunk and people needed to be saved. This is a dangerous task, and there are already enough people who are taking it upon themselves to save those shipwrecked, so I decide not to take the risk. But then I spot David barely keeping himself afloat, and I think “I need to get David to the shore”. As I swim towards David, I see that I can save two people instead if I turn in the other direction. Am I allowed to continue and save David? I think so. Of course this reasoning can be extended significantly beyond cases in which I am already in the midst of, or even settled upon, a particular course of action. After all, if what makes a difference is that the person herself appears in my reason, the person could also appear in my reason at the time of the decision. If I decide to save David himself, my reason is undefeated by the prospect of saving two people. This might appear counterintuitive, but I think it is a welcome result. Let us go back to our scenario of David being one of the victims of a shipwreck. Let us assume that at first I do not join the rescue efforts, and let us assume that even though this will result in fewer people being saved (there are not nearly enough people on the beach to save everyone who is drowning), it is permissible for me not to join (even though I am certain to save those whom I attempt to rescue, there’s a significant chance that I’ll die as a result of it due to my fish allergies). But then I remember that David was in the ship. I also know he was the only single person in a cruise for newlyweds (he was the hired barista for the cruise) and that he’ll be the only person by himself on the sea (the newlyweds keep very close to each other and far from others, and this configuration remained after the shipwreck). Since each person on the beach will have time to go only on one rescue mission, they are all facing a situation like Basic Case (except that there is more than one set of two people that each of us can save instead of saving David), and thus it is quite clear to me that David will not be saved. So, I remember the times when David and I were bantering while he was serving me coffee and I think to myself “I’ve got to save David”, and I then jump into the water in order to save him. It doesn’t seem that I act badly now when I choose to save David even if I could save two other people instead. Now, this is a case of supererogation. But, first, in other cases of supererogation, I seem to be under the same obligations as in an “obligatory”
¹⁴ Following Setiya’s suggestion mentioned above in footnote 8.
’ ? 231 rescue case once I decide that I’m making the necessary sacrifice; if as I am saving David I see that I can easily save another person with no extra effort, I would be then obligated to save the other person as well. Moreover, it is not hard to think of similar cases in which I have a clear duty of rescue; in fact, I do not think it would make any difference to our verdict if I had no allergies (and was thus obligated to rescue someone), and, I was about to join the rescue efforts, I realized that, given the configuration of the newlyweds, unless I were to save David, David would not be saved. Note that I am not claiming that “being a particular person” is a morally relevant consideration; after all, what kind of consideration this would be?¹⁵ Rather, what is relevant for this case is the fact I am saving David in particular is a different kind of reason than the fact that I am saving someone. Taurek thinks that if weak forms of partiality can ground duties of partiality that override the duty to save the many, the duty to save the many cannot be very stringent. But on our account, the issue is not that some kind of partiality, a mere casual or even newly acquired acquaintance, overrides the value of saving an extra life; rather, having access to the reason “saving David himself ’ ” makes the Pareto Principle inapplicable. But what exactly makes it possible for me to have a reason to save this particular person, rather than just a reason to save a person, or, to put it slightly more controversially, when is it possible for person herself to figure in the content of my practical reasons? At the very least the agent must be able to pick out the person, so the agent must be at least capable of having singular thoughts about them. But this might not be sufficient to allow a particular person to figure in my reason for action. Suppose I catch a glimpse of Jeri at a distance for a few seconds. Can the thought “Jeri is drowning” figure as a reason to save Jeri that is relevantly different from the reason given by the thought “someone is drowning”? It seems that the answer is “no”, and perhaps practical acquaintance requires more; my own view is that practical acquaintance requires a more robust engagement with my practical thought. Cases like Chris’ and David’s involved much more substantive agential engagement with the rescue target than a mere passing glimpse. Of course, this is just a promissory note, and to make good on it we need a more precise account of “robust engagement with my practical thought”. But however we account for what makes this kind of practical acquaintance ¹⁵ But see Setiya (2023). Setiya talks about “personal acquaintance”, which overlaps, but does not perfectly match, what I call “practical acquaintance” below. Setiya does think that personal acquaintance is morally relevant in similar cases.
232
possible, I have argued here that it makes available to us a reason that changes the character of our choice situation. This observation lets us also answer a potential concern: namely, that the structure of dignity makes Basic Case radically different from instances of “sweetening in the dark”. We already know that each person has dignity; while each box in Hare’s example is literally just a vessel for the different values it carries, human beings are not just vessels of some unknown value. Couldn’t I just permissibly take stream A so as to save someone who has dignity, no matter who that person turns out to be? Indeed, this is a good reason to take stream B, but this is also the very same reason I have to save the “first” person in stream A. It is exactly the fact that persons are not mere vessels of dignity that makes Basic Case structurally similar, in the relevant way, to the cases of sweetening in the dark. Note that in Hare’s example, I do know that Box A contains something whose value would give me a sufficient reason to choose it (or, to put it somewhat controversially, I know that the value of Box A’s contents is incommensurable with the value of Box B’s contents). However, this reason is not accessible to me; I can only act on the reason that the box has either the Fabergé egg or the wedding album. Similarly, each person has dignity, and thus I know that if I steer to stream B there will be someone there whose value would give me sufficient reason to save them. But we are not just vessels of dignity; thus the reason to save this particular person is not available to me. I can only act on the reason that I can save a person here. This reason, however, is identical to the reason I have to save Larry (or Mary) and thus the Pareto Principle kicks in and forces me to go to stream A. It is also worth noting that how I involve others in my agency can be morally relevant whether or not I am practically acquainted with them. Doubtless, if I find a way to push a person off a footbridge in order to save five without gaining acquaintance with my victim, I am still using this person as a means. But nothing in my argument jeopardizes this platitude. You cannot treat anyone merely as a means, and so you cannot treat someone as a means even when you don’t know who that person is. My reason for saving the person is defeated by the prohibition to use any human being in this manner. Here the comparison with cases of incommensurability might also be helpful. If someone offers to pay me $60 for an object, and all that I know about this object is that its value is incommensurable (or on a par) with a Fabergé egg, I can easily conclude that I must turn down the offer. After all, my reason to make $60 is defeated no matter what the object turns out to be. What follows from my argument,
’ ? 233 both in the “prudential” and the “moral” case, is that ignorance (or lack of practical acquaintance) potentially changes the nature of my reasoning, not that it changes the nature of the value in question. Finally, it is worth noting that this kind of reasoning, based on the Pareto Principles above, might have further implications. For instance, it can help the Kantian, and other deontologists, explain the plausibility of a different kind of verdict, in a different context. Let us look briefly at a consideration favoring effective altruism: whether or not you are required to donate to global charities, if you do decide to donate to these organizations, you should donate to the more effective ones. This kind of consideration seems overwhelmingly plausible if we are deciding between two charities investing in similar efforts, but one of them is significantly more effective than the other. If “Against Malaria” and “Somewhat Against Malaria” are both fighting malaria but the former saves many more people than the latter with the same amount of money, we seem to have decisive reason to prefer the former even if the people saved are not the same. However many would want to resist this conclusion if the charities are committed to different issues (say, one of them is fighting malaria while the other is trying to help refugees), or even if I am particularly committed to those who live in the “Somewhat Against Malaria” area (an area not covered by “Against Malaria”). In such cases, it seems plausible that I do no wrong if I prefer the less efficient charity. Nonconsequentialists might want to explain the permission to donate to “less efficient” charities in terms of the fact that the claims of different people cannot be compared in the way suggested by the effective altruist, but this leaves us wondering why the same reasoning does not apply to the first type of case. Perhaps the Extended Pareto Principle (or some similar principle) grounds the different deontic verdicts in these two types of cases. So can Kant count? Well, no, but Kant can do all that counters can and much that counters can’t.¹⁶
References Chang, R. (2002). The possibility of parity. Ethics, 112(4):659–88. Hare, C. (2010). Take the sugar. Analysis, 70(2):237–47.
¹⁶ For very helpful comments, I owe many thanks to Sarah Buss, Kyla Ebels-Duggan, Thomas Hurka, Jennifer Nagel, Julia Nefsky, two anonymous referees, and audiences at the 2021 Workshop in Normative Ethics and the Department of Philosophy at Bilkent University.
234
Horton, J. (2018). Always aggregate. Philosophy & Public Affairs, 46(2):160–74. Igneski, V. (2006). Perfect and imperfect duties to aid. Social Theory and Practice, 32(3):439–66. Kant, I. (2012). Groundwork of the Metaphysics of Morals, trans. Mary Gregor. Cambridge University Press. Parfit, D. (1978). Innumerate ethics. Philosophy and Public Affairs, 7(4):285–301. Saunders, B. (2009). A defence of weighted lotteries in life saving cases. Ethical Theory and Moral Practice, 12(3):279–90. Setiya, K. (2014). Love and the value of a life. Philosophical Review, 123(3):251–80. Setiya, K. (2023). Other people. In Buss, S. and Theunissen, N., editors, Rethinking the Value of Humanity. Oxford University Press. Singer, P. (1972). Famine, affluence, and morality. Philosophy and Public Affairs, 1(3):229–43. Steuwer, B. (2021). Aggregation, balancing, and respect for the claims of individuals. Utilitas, 33(1):17–34. Taurek, J. M. (1977). Should the numbers count? Philosophy and Public Affairs, 6(4):293–316. Tenenbaum, S. (2017). Action, deontology, and risk: against the multiplicative model. Ethics, 127(3):674–707. Thomson, J. A. (2022). Relief from rescue. Philosophical Studies, (179):1221–39. Timmerman, T. (2015). Sometimes there is nothing wrong with letting a child drown. Analysis, 75(2):204–12. Tomlin, P. (2017). On limited aggregation. Philosophy and Public Affairs, 45 (3):232–60.
11 What Does Virtue Have to Do with Consequences? Iskra Fileva
Some wise person says, “There is always two reasons for doing a thing: one is a good reason and the other is the real reason.” Ms. Walter B. Helm, quoted in The Rockford Daily Register-Gazette, February 11, 1905.
1. Introduction In a bold and provocative book, Julia Driver once argued for a consequentialist view of virtue—a view she labels pure evaluational externalism.¹ On this view, virtues are dispositions to produce good consequences systematically. Importantly, they do not require good internal states of the agent such as motives and intentions, and vices do not require bad internal states.² Driver’s argument proceeds by an appeal to what she labels virtues of ignorance. She takes humility, blind charity, and acting rightly but with the belief you are doing something wrong, as Twain’s Huck Finn arguably does, as examples of virtues of ignorance. The idea is that in all these cases, we have agents who act without knowledge of what is right and so without a good internal state. ¹ Julia Driver, Uneasy Virtue (Cambridge: Cambridge University Press, 2001). See also Robert Hartman’s “Utilitarian Moral Virtue, Admiration, and Luck,” Philosophia 43 (2015): 77–95 for an updated version of a consequentialist view of virtue, of a utilitarian variety. ² Note: Driver allows for the possibility that internal states may be necessary for some particular virtues but argues that they are not necessary for virtue in general, writing, “[P]articular virtues such as generosity can be analyzed in terms of specific psychological states that characterize them. For generosity the typical motive may be something like a desire to benefit others and the typical intention one of aiding them. While this is true of the virtue of generosity, it is not true of virtue across the board,” (Ibid., 62). Iskra Fileva, What Does Virtue Have to Do with Consequences? In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © Iskra Fileva 2023. DOI: 10.1093/oso/9780198895909.003.0012
236
Several objections have been raised to Driver’s argument for consequentialism, which proceeds by an appeal to cases. I take the most important one to be the following: the virtues of ignorance she lists only seem to be virtues if we imagine the agent manifesting them to also possess a good disposition of some sort.³ For instance, blind charity is only a virtue if the blindly charitable are moved by something like a tendency to see the best in everyone, which is arguably a good internal disposition. Similarly, Huck Finn, in choosing not to disclose Jim’s identity, is displaying a virtue only if he is motivated by a desire to help his friend or an inability to betray his friend and not, say, by a desire to spite his aunt. I do not think a good response on Driver’s behalf to this objection is available, though I will not try to show that here as there is a more important point I wish to make: the success of the objection—assuming, as I believe, that it succeeds—would not spell doom for consequentialism about virtue. This is because there is a better argument for consequentialism. A version of that argument was once advanced by David Hume. The argument asks us to consider the question of why it is good to be helpful to others, and why helpfulness is a disposition we ought to cultivate in ourselves and inculcate in our children. The answer is likely to appeal to consequences: helpfulness is instrumental in relieving distress and making life better for those who receive help. The answer is unlikely to appeal to motives, e.g., “Because helpfulness is motivated by a good state of the agent.” But virtues, whatever else they may be, would have to be good dispositions that we ought to cultivate. So they must be traits that have good consequences. This, I take it, is the crux of Hume’s point in the famous “monkish virtues” passage: Hume argued that in deciding what behavior to incentivize and label “virtuous,” we have to look at the consequences of said behavior.⁴ The motive behind the monkish virtues Hume decried may well have been a good one, and Hume does not criticize the motive. His insight was that however this may be, if the dispositions the motive gives rise to benefit no one, then they can hardly be called virtues.⁵ ³ Michael Slote, “Driver’s Virtues,” Utilitas 16–11 (2004): 22–32. For a detailed defense of the importance of motives, see Slote’s Morals from Motives (New York, NY: Oxford University Press, 2001). ⁴ David Hume, Enquiries Concerning Human Understanding and Concerning the Principles of Morals, edited by L. A. Selby-Bigge 3rd edition, revised by P. H. Nidditch (Oxford: Clarendon Press, 1975), 270. ⁵ It must be noted, however, that Hume is not, at the end of the day, a consequentialist in the relevant sense. Consider, for instance, the following passage: “It appears, therefore, that all virtuous actions derive their merit only from virtuous motives, and are considered as merely
? 237 Forgetting this point can be dangerous. The danger is that we may unmoor ourselves from consequences and take what we intuitively feel to be the purity of our own motives as a guide to action. The problem with this is not only—and perhaps not even primarily—that we have a strong selfserving bias and are bad at divining the nature of our own motives, but that even if we can say with complete certainty that our motives are good, that is insufficient to show that the actions that spring from those motives are right or justified. One may, for instance, out of compassion—a perfectly good and laudable motive—end up enabling one’s sister’s addiction by giving her money, or else spend an inordinate amount of resources on a tragedy with a known victim rather than a problem with a greater number of unknown victims. More generally, even if the motivation behind a given disposition is morally exemplary, if we have a reason to think that it will lead to suboptimal social consequences, we should generally discourage it and perhaps, pronounce it a spurious virtue at best.⁶ Call the intuition behind this point the consequentialist intuition. This intuition, as some of the considerations I just made in response to Driver suggest, has its contrary. In order to see that, suppose we approach the problem of virtue in another way and ask not what makes a disposition such as kindness or helpfulness a good candidate for inculcation, but rather, when and under what conditions we can ascribe virtue to some particular person. We can now elicit an anti-consequentialist intuition by imagining cases in which inner dispositions are divorced from action consequences, either because the dispositions lack consequences altogether or because there is a valence mismatch between internal states and consequences, such that when one is good the other one is bad and vice versa. Consider the first possibility first. There appear to be stable dispositions without consequences that may nonetheless constitute virtues and vices, for instance, secret jealousy, schadenfreude, or the good will and compassion of a person with locked-in syndrome.⁷ Or consider cases of systematically signs of those motives,” A Treatise of Human Nature, edited by Ernest Mossner (London: Penguin Classics, 1986), 530. Driver says that she intends her account to be no more radical than Hume’s, but her account actually is more radical than Hume’s since Hume accepts the role of motives. See Driver’s “Response to My Critics,” Utilitas 16–11 (2004): 33–41, 34. If the account I am about to offer succeeds, it will help explain how Hume may have thought consequences paramount without embracing consequentialism about virtue. ⁶ This is basically the argument psychologist Paul Bloom makes in Against Empathy: The Case for Rational Compassion (New York, NY: Ecco, 2016). It is also the type of argument that motivated the Effective Altruism movement. ⁷ Nicolas Bommarito, Inner Virtue (Oxford: Oxford University Press, 2017) offers a thoughtful, book-length defense of this view.
238
mismatched valence. Imagine, for instance, that the devil is doing his best to hurt Sarah, but God intervenes and thwarts the devil’s intentions every time, making Sarah better off, instead. The devil keeps hoping he’ll outsmart God on the next try, but he fails every time. The devil in this case seems to be acting viciously even though Sarah consistently benefits from his actions. Finally, think of cases in which people are disposed to consistently do what is right but for utterly selfish reasons: doing what is right would benefit them personally. In all of these cases, we are tempted, in assessing an agent’s character, to weigh heavily the motives rather than the consequences of the resulting actions. Call the intuition which inclines us to say that neither the devil nor thoroughly selfish people could be virtuous whatever good they may produce and however systematically, the motivationalist intuition. If we lean solely on the consequentialist intuition, we end up with the view, defended in a different way (that is, by an appeal to virtues of ignorance) by Driver, that consequences are the only thing that matters, that is, with consequentialism about virtue. If, on the other hand, we focus exclusively on the motivationalist intuition, we are led to conclude that virtue and vice are solely a matter of internal states of the agent while consequences are irrelevant to virtue and vice ascription, a view we may, for parity’s sake, label motivationalism about virtue.⁸ This clash of intuitions makes for a fragmented landscape. It is as though we were estimating probabilities, and some people’s estimate of the likelihood of an event’s occurrence was 1 while that of others was 0. Something seems to have gone wrong. What should we say about all this? Perhaps, we can argue that the two sides are not so much disagreeing as talking past each other, each focusing on a different issue. I have a good deal of sympathy with this type of response. But I think also that we cannot stop here. This is because it would be surprising if it turned out that what makes some dispositions virtues and vices in general is utterly divorced from what makes any given person virtuous or vicious. If it is the case, we have to explain how that can be. If it is not, we need an account that illuminates the connections between virtues and vices on the one hand, and consequences and motives, on the other. By the same token, a simple addition of the two approaches in one conciliatory—ecumenical—strategy that pronounces both consequences ⁸ Todd Calder, “Against Consequentialist Theories of Virtue,” Utilitas 19–12 (2007): 201–19; John Skorupski, “Externalism and Self-Governance,” Utilitas 16–11 (2004): 12–21.
? 239 and motives necessary for virtue while holding that neither is, on its own, sufficient would not do.⁹ Such a strategy, of course, won’t satisfy either motivationalists or consequentialists since the motivationalist intuition tells us not simply that internal states matter but that they are the only thing that matters, while the consequentialist intuition tells us that all that matters, in the end, are consequences. A theoretical kumbaya which pronounces both motives and consequences necessary while denying that either is sufficient is likely to be seen as a bad compromise by both warring parties. This is not the main problem, however, since it could be that motivationalists and consequentialists do not realize that they are talking past each other. The real problem is that a viable conciliatory approach—or “mixed” strategy,¹⁰ as Driver calls this type of approach—must either explain why the basis on which a disposition is to be labeled “virtue” is unrelated to the basis on which virtues are ascribed to individual people, or else it has to explain what the connections between the two things are. In what follows, I will pursue the second strategy. I will argue that virtue requires motivation that’s in harmony with the reasons that make a disposition socially desirable. In addressing my task, I proceed as follows: I begin by laying out my view (section 2) and then proceed to consider responses on behalf of consequentialism (section 3) and motivationalism (section 4) respectively before I offer a summary and make some additional points (section 5).
2. What non-viciousness teaches us about virtue I wish to begin with an observation: what people do is important, perhaps more important from a societal point of view than why they do what they do. Thus, we want people to do their jobs well. We hope that grocers will be honest and researchers will work hard. We want scientists to discover cures for diseases and architects to envisage nice buildings. We desire these things independently of why a given person may or may not do them. We desire them so much, in fact, that we are more than willing to amply reward Dr. Burning Ambition if he finds a cure for cancer even if his primary motivation is the utterly selfish one of proving himself better than his
⁹ This type of strategy, it must be noted, has appeal among non-philosophers. See Adam Feltz & Edward Cokely, “Virtue or Consequences: The Folk Against Pure Evaluational Internalism,” Philosophical Psychology 26–5 (2013): 702–17. ¹⁰ Driver ascribes it, plausibly, to Aristotle.
240
colleague Dr. Ellen Enviable, who has a more prestigious research position and whom he’d always envied since she appears to be a “golden child” to whom success came easily. Relatedly, we regard a person’s actions as right or justified when there are good reasons to do those actions even if the agent’s own reasons have little to do with what justifies her actions.¹¹ This is, presumably, what Mill had in mind when he said, “He who saves a fellow creature from drowning does what is morally right, whether his motive be duty or the hope of being paid for his trouble.”¹² Kant noted, in a similar vein, that the honest grocer who is honest only because it is good for business is acting in accordance with duty. While Kant’s main point was that such a grocer, not acting from the motive of duty or out of respect for the moral law, is not behaving in a way that is morally worthy, he suggests also, or at least implies, that the grocer’s behavior, being in conformity with duty, is perfectly justified at least in the sense of being permissible. If this line of reasoning is on the right track, we can say that consequentialists about virtue are mistaken about the proper object of consequentialist treatment. Consequences, we might say, are key to determining whether some type of action is right and should be rewarded and encouraged. This explains why, as we saw earlier, consequences are essential in determining whether a disposition to perform actions of a certain type—save drowning people, say, or be honest to one’s customers—ought to be incentivized. But consequences may not be the key determinant when it comes to individual virtue ascription. We should encourage people to save drowning children, find cures for diseases, and be honest in their business dealings. Perhaps, we can even praise them if we expect that praise would incentivize the desired behaviors. But we cannot ascribe virtue merely on the basis of the fact that a person does these things unless she does them for certain reasons. This is what I wish to argue for now. In making my case, I will begin by considering a person such as Kant’s grocer. What are we to say of the grocer’s character? Intuitively, Kant seems right in claiming that the grocer’s (honest) actions lack moral worth. Is it,
¹¹ W. D. Ross, relatedly, distinguishes between the Right and the Good, arguing that actions can be right or wrong but not good or bad while motives can be good or bad but not right or wrong. See W. D. Ross, The Right and the Good (Oxford: Clarendon Press, 1930). For counterarguments, see Michael Stocker, “Act and Agent Evaluations,” The Review of Metaphysics 27–1 (1973): 42–61. ¹² John Stuart Mill, Utilitarianism, edited by Roger Crisp (Oxford: Oxford University Press, 1998), 109.
? 241 perhaps, that the disposition to be honest to one’s customers for self-interested reasons is a bad and vicious one? The answer, I think, depends on whether the grocer accepts certain moral constraints on his pursuit of his own self-interest. If the grocer is committed to promoting his own self-interest subject to constraints, namely, he is committed to promoting it in a way that either benefits or at least does not harm others, then his disposition is a non-vicious one. Non-vicious dispositions need not involve deep concern for the wellbeing of others. For instance, a person may be disposed to be honest or to work hard on her scientific projects for the sake of reputation or fame. Such dispositions (probably highly common) are non-vicious. But there are two important points to note. First, non-viciousness does not amount to virtue. Nonvicious dispositions are permissible but not particularly praiseworthy (though as I noted earlier, there may be consequentialist reasons to praise them, for instance, in order to encourage them). By contrast, virtue is praiseworthy. Consequently, more is required for virtue. Virtues require suitable motivation: a kind of connection between the agent’s motives and the reasons that make her actions permissible or socially desirable. For instance, a disposition to be helpful is a virtuous one when it disposes one to act for the reasons that make helpfulness a good thing and something we ought to encourage, that is, the consequences for other people. Second, it is important that consequences do suitably constrain the disposition: if Jane is disposed to pursue her interests whether or not that would harm others, then her disposition is a vicious one, a kind of callousness, even if, coincidentally, it leads to good consequences most or even all of the time. Jane may get morally lucky and never do anything wrong, because her interests always happen to align with what is independently desirable from a societal point of view, but if her disposition is to pursue her interests regardless of moral constraints, then her disposition is in fact a vicious one though the viciousness remains masked. We can now begin to appreciate the respective roles that consequences and motives play in virtue. Roughly speaking, consequences help delineate the class of permissible actions. Non-viciousness requires a commitment to staying on the right side of the permissibility/impermissibility boundary. No disposition can be a virtuous one without being at least permissible. This explains why attempts to decide whether some type of action is right by scrutinizing a person’s motives—whether our own or those of others—are generally misguided. The purity of a person’s heart alone cannot make her actions right, and it cannot make the disposition to perform actions of a
242
certain type virtuous.¹³ This explains also why Hume’s monk or the person who enables her own sister’s addiction by compassionately providing financial support cannot be virtuous in acting as they do. Virtue becomes possible only after this minimum threshold has been cleared. But virtue requires more than clearing that threshold. It places stricter motivational constraints on agents and requires that an agent be directly motivated by some of the reasons that make her actions right. I must now complicate things somewhat before placing my account in a broader context. There are two different dimensions along which theories can be classified that ought to be distinguished: one is the objective/subjective dimension and the other is the consequentialist/non-consequentialist one. On objectivist accounts, what matters to virtue are objective facts rather than factors internal to agency, while the reverse is true of subjectivist views. Both consequentialist and non-consequentialist accounts can be objectivist as well as subjectivist. One can, for instance, have an objectivist but nonconsequentialist account on which virtues are dispositions to do what one has an objective duty to do and perhaps, to go beyond the call of duty, no matter for what reasons one acts. More familiarly, one can have a subjectivist consequentialist account on which virtues are dispositions to bring about good expected consequences, where expected consequences may differ— possibly radically—from objective consequences (since we can be mistaken about what would in fact bring about good consequences). In speaking about consequentialism here, I have in mind objectivist versions of consequentialism such as Driver’s. Of the two elements— consequentialist and objectivist—the objectivist element is the crucial one. My claim is that whether some disposition is a good candidate for the label “virtue” in general is a matter of objective considerations. I happen to think also that among objective considerations, consequences would be key, but that is a secondary point. I do not insist on a consequentialist account of rightness. I spoke in broad strokes of consequences for other people. I intend my view to be compatible with an account of rightness that accommodates partiality constraints, whether or not those can rightly figure in a consequentialist theory. The view is really one about the connection between the reasons that make actions right, the reasons we have to cultivate dispositions
¹³ Skorupski, in his “Externalism and Self-Governance,” suggests that on Driver’s view, a person may not be able to determine what the right thing to do is through deliberation. Driver, “Response to My Critics,” 38 responds, correctly, that this is a feature and not a bug of her account.
? 243 to perform actions of a certain type, and a virtuous person’s reasons for action. A person may be morally lucky when, in the absence of any commitment to curb the pursuit of her own goals, she consistently does what there is an objective moral reason to do, say because society is structured in such a way that this is not only possible but likely. (People have an incentive to do the morally right thing in almost every case.) She is minimally decent when she accepts certain constraints: like the cancer researcher in pursuit of fame, it is really her own good that she is after, but she only pursues her good in ways that align with the social good. She is virtuous when she possesses dispositions whose motivational component is in line with the objective reasons that make those dispositions desirable from a social point of view. Let me now place this account in a broader context. As a society, we recognize that the desires and interests of different people often fail to align. In societies at a certain stage of moral development, clashes of interests are not resolved by brute force but in one of two ways: either through moral behavior (one or both parties show willingness to act in ways that do not promote their own interests) or by making changes to the system so that the interests of an individual person align with those of other people, as when we offer rewards for socially desirable behaviors and for outcomes such as finding cures for diseases or saving people in burning buildings. Promoting the good of others while at the same time promoting one’s own good is in general always permissible, and the disposition to do so is not vicious. Promoting the good of others without an incentive or a self-interested reason is better than permissible—it is virtuous. Since virtue requires a particular kind of motivation, there is a danger of losing sight of what else it might require. This is how Hume’s monk may be led to believe that he is acting virtuously: he is doing something that requires a cost to oneself, a kind of self-sacrifice, without a tangible reward. The monk may, thus, feel virtuous in the act of self-flagellation, because selfflagellation is costly and unpleasant. Something similar is true of a person whose compassion for a sibling addicted to heroin enables the addiction: the enabler may not experience her action as costly, but she likely experiences her motivation as caring and compassionate. Actions that are not justified, however, cannot be virtuous, and justification is not guaranteed by the nature of one’s motives. On the other hand, since consequences for others are of great import, it may seem that a disposition to bring about good consequences is a virtue, however motivated. That’s the consequentialist proposal. I argued that this proposal is not right as it stands but that it gets something important right.
244
This is the view in a nutshell, then: virtues are dispositions with a motivational component that’s closely aligned with the objective reasons that recommend a certain type of behavior. There are some complications I shall return to, but first, I wish to address possible responses on behalf of motivationalists and consequentialists.
3. The consequentialist’s retort Consequentialists may argue that we don’t really need a commitment to moral constraints, much less intentions to act for moral reasons. It only seems that we do, because without such a constraint, we are unlikely to bring about good consequences reliably. But it is really the systematicity that matters in the end, not the motivation. This is Driver’s view. The problem here is that people can produce good consequences systematically yet for utterly selfish reasons. Intuitively, dispositions to do so do not amount to virtues. Julia Driver considers such dispositions. Initially, she gives a kind of bullet-biting response, suggesting that they may be virtues: It is conceivable, à la Mandeville, that a motive of selfishness could produce a great deal of public good (Mandeville 1970). The motive here is selforiented, though the good actually produced could be solely the public good. Good motives are important, so being motivated by a desire to see others happy usually produces good, because motives give rise to intentions that in turn typically give rise to actions guided by the intentions (and intentions can in turn give rise to motives). But it may be that a motive of self-interest produces good for others as well – and, if so, that motive forms part of the virtue. The virtue would be a disposition cluster consisting of tendencies to be motivated and forming beliefs in certain ways that produce good action.¹⁴
Later, however, in response to an objection from Onora O’Neill to the effect that on Driver’s view, Ayn Rand may turn out to be right that selfishness is a virtue,¹⁵ Driver says:
¹⁴ Driver, Uneasy Virtue, 91. ¹⁵ Onora O’Neill, “Consequences for Non-Consequentialists,” Utilitas 16–11 (2004): 1–11.
? 245 O’Neill cites the case of Ayn Rand, and her view that selfishness is indeed a virtue. But of course it is not enough to show that sometimes selfishness has good effects. I believe Rand to be quite mistaken. The good effects must be systematic and outweigh the bad effects also produced by the trait. So, while it is indeed possible that selfishness could turn out to be a virtue, the empirical claims necessary to support this seem wildly implausible to me.¹⁶
There are two main points Driver makes, then. The first is that a selfish disposition to behave well cannot, as an empirical matter, be expected to reliably lead to good consequences, so if we define virtues as stable and reliable dispositions to produce good, a selfish disposition that leads to socially beneficial consequences will not be a virtuous one, which is presumably the conclusion we want. However, a selfish disposition may well be reliably beneficial if society is structured in such a way that self-interest and the social good are closely aligned, as when people can expect to get rewarded for acting well and punished for acting badly. A thoroughly selfish agent who is good at promoting her own long-term interests would then act well consistently. Second, a selfish disposition subject to the motivational constraints I outline here is reliable as well, perhaps even more so (since it is likely to endure even when external inducements do not properly incentivize good behavior and disincentivize bad behavior). What makes it reliable—or more reliable than external inducements alone—are the motivational constraints. This disposition is not thereby virtuous, though it is not a vicious one either. I would like to say it is minimally decent. Mere reliability, then, does not suffice. Indeed, reliability produced via endorsement of moral constraints does not suffice for virtue either but only for lack of vice. But why think that I am right, and Driver is wrong? What reason do we have to resist thorough-going consequentialism? Maybe, we intuitively shrink from it, because we assume, a philosopher’s stipulations notwithstanding, that good behavior cannot be reliably produced without good motives, and we don’t want people to believe otherwise. Yet suppose for the sake of argument that reliably good behavior can be produced without good internal states. If we still wish to insist on the claim that internal states
¹⁶ Driver, “Response to My Critics,” 34.
246
are necessary for virtue, we have to give a reason, a reason that points to a source of value independent of consequences. There is, I think, such an independent source of value. In fact, Driver herself acknowledges, possibly unwittingly, this source, writing: [E]ven if it turns out that the good intentions ensure better consequences with respect to all human characteristics, this would not be sufficient to show that the traits that will still produce good without the good intentions and motives are not moral virtues. People who regard only the best traits as moral virtues are maximizers, and subject to all the problems that a maximizing requirement places on a theory.¹⁷
In this passage, Driver contaminates, so to speak, objectivist consequentialism by implicitly acknowledging that an independent source of value relevant to virtue exists. If, as Driver says, dispositions would in fact be better (“best traits”) when they not only produced good consequences but were motivated in a particular way, then motivation matters to virtue. But it seems to me that there is no good reason to go for this type of compromised consequentialism. Pure consequentialism, while it may sacrifice some of our intuitions about virtue, promises a kind of appealing theoretical purity. Compromised consequentialism sacrifices our intuitions (intuitively, Kant’s grocer is not virtuous though he is not vicious either) without a corresponding theoretical benefit such as explanatory simplicity. If we accept that good motives make already virtuous dispositions morally better (that is, more virtuous), why not accept that good motives make non-virtuous dispositions virtuous? Let us suppose, however, that in this passage, Driver has slipped, and that she actually intends to deny a second source of value relevant to virtue and wishes to accept, instead, pure consequentialism of an objectivist variety. Pure consequentialism promises a kind of theoretical neatness but at a steep price: at the cost of sacrificing our intuitions about what matters morally. We do not, in general, suppose that we can know whether people have some virtue or not unless we have some sense of what motivates them, what they think and feel. It simply wouldn’t do for the consequentialist to say that the only reason we care (or should care) about those things is fear of unreliability. Dispositions to behave reliably may well exist in the absence of moral
¹⁷ Driver, Uneasy Virtue, 58.
? 247 motivation. As I see things, the main problem for the consequentialist here is that there is no theoretical advantage to saying that such dispositions would be virtues. The claim that they would is simply an implication—and not a fortunate one—of the view. What is plausible about consequentialism is captured by my view. And what is plausible is the observation that the way to determine whether some type of behavior should be promoted is to look at its consequences, not its motives. If, by contrast, we want to know whether the disposition to exhibit said behavior—saving drowning children, for instance—is a virtue in a particular person or not, we have to know how that disposition is motivated. There is a final point about consequentialism I wish to make. It is unclear how a pure consequentialist account can explain why virtues can only be possessed by rational agents and not, for instance, by unthinking robots, lacking consciousness. Driver rules out this latter possibility by simply stipulating that virtues are dispositions of character and require the ability to form intentions.¹⁸ But this stipulation is ad hoc.¹⁹ There is certainly no consequentialist reason to make it. One wonders why it matters that an agent possesses the ability to form intentions if the nature of the intentions— whether they are good or bad—doesn’t matter. Earlier, we were told that the nature of the intentions doesn’t matter, because good consequences could be produced without a good intention. But of course, good consequences can be produced with no intention at all, and the only reason to rule out the possibility of a virtuous unthinking robot is a non-consequentialist one: this possibility is implausible and would make the view implausible. Once again, in order to make consequentialism palatable, we have to compromise it.
4. The motivationalist’s retort A committed motivationalist, on the other hand, can say two things. First, she can appeal to cases in which virtues and vices can be ascribed to a person solely on the basis of motivation and contend that motivational states ¹⁸ Driver, “Response to My Critics,” 35. ¹⁹ One could, perhaps, say that Driver must do this, because the word “virtue” in English refers to a type of character disposition involving the ability to form intentions. But this response will not do. For the word “virtue,” as ordinarily understood, includes particular kinds of motives also— virtuous ones. Virtues are, typically, seen as incompatible with acting compulsively or for bad reasons. If one wishes to defend a revisionary account, one better be a consistent revisionist. And a consistent revisionist would have to say that the ability to form intentions is unnecessary (though it may be correlated with virtue since it may help ensure the reliability of dispositions).
248
alone suffice for vice and virtue. Second, she can argue that even when inner states have consequences, it is not the consequences per se that matter but the motivation to bring them or not to bring them about. More generally, the motivationalist can claim that what makes our actions right or justified is, in the end, the motives. Perhaps, the motivations of Hume’s monk or the compassionate person who ends up enabling her sister’s addiction are not unassailable, and the resulting dispositions are flawed to the extent the underlying motivation is flawed. I will take these points in order. Earlier, I mentioned the case of a person with locked-in syndrome who cannot do anything but whose inner states may nonetheless be subject to moral evaluation, a case I borrow from Nic Bommarito.²⁰ We can think of other cases along those lines. For instance, in 1984, George Orwell’s character Winston describes a scene from a war film he’d seen. In the scene, a mother attempts to protect her young son from a bomb explosion though she knows she cannot. Winston says that she was: [P]utting her arms around him and comforting him although she was blue with fright herself, all the time covering him up as much as possible as if she thought her arms could keep the bullets off him. then the helicopter planting a 20 kilo bomb in among them terrific flash and the boat went all to matchwood . . .²¹
Though the style of this description is stream of consciousness, with no capital letters and periods, the meaning is clear: the mother attempts to protect her son in a way that is bound to be ineffectual. Her arms won’t fend off bullets, much less a bomb. Yet the mother does just what we expect a good parent to do. But both Orwell’s ineffectual mother and Bommarito’s good locked-in person examples involve situations in which the minimum threshold has actually been cleared: while no good consequences are expected in either case, no bad ones are expected either. (In fact, arguably, the action of Orwell’s character is not completely without good consequences, it is just that the good consequences are very short-lived—the child likely is comforted for a few seconds.) So virtue is incompatible with aiming to bring about bad consequences. What about vice? Could dispositions be vicious without aiming to bring about bad consequences? Could the schadenfreude of a person with locked-in ²⁰ Bommarito, Inner Virtue. ²¹ George Orwell, 1984 (New York, NY: Signet Classics, 1966), 8–9.
? 249 syndrome or the secret envy of an ordinary person amount to vices despite being, as it were, consequentially impotent? One possible response we can give is to say that motivational patterns of this sort that have no consequences are not really vices. This is Thomas Nagel’s view. Nagel (1998) argues that our private mental lives considered in isolation are not a proper subject of public moral standards.²² I have some sympathy with this line of argument, but I believe Nagel takes it too far. Sometimes, it is quite appropriate to morally evaluative our own psychological and motivational states independently of consequences.²³ Another, and in my view more promising, response is to introduce an asymmetry of sorts between virtue and vice and say that while virtue requires at minimum a commitment to avoiding bad consequences, vice can exist without disregard of bad consequences, let alone a desire to bring such consequences about. The secretly envious person need neither bring about nor aim to bring about bad consequences in order to have a deficient character. She may, in fact, be committed to not bringing about bad consequences. However, we must note also that such impotent vices—assuming they are vices at all—would pale in comparison to vices that come without a neutralizing commitment to avoiding bad consequences. The motivationalist can, for all that, insist that thorough-going motivationalism remains undefeated. This is her second point. She can argue that consequences don’t really matter to virtue, but that at most, a commitment to bringing about one or another set of consequences matters, and this commitment is an internal state. One can resist this line of argument by saying that the motivation to bring about or not to bring about certain consequences matters only because the consequences matter. Consider, for instance, what H. W. B. Joseph says about W. D. Ross’s attempt to sever the link between the right and the good: Why . . . ought I to do that, the doing which has no value (though my being moved to do it by the consciousness that I ought, has), and which being done causes nothing to be which has value? Is not duty in such a case irrational?²⁴ ²² Thomas Nagel, “Concealment and Exposure,” Philosophy & Public Affairs 27–1 (1998): 3–30. ²³ I argue for this point in my “Envy’s Non-Innocent Victims,” Journal of Philosophy of Emotion 1–1 (2019): 1–22. ²⁴ H. W. B. Joseph, Some Problems in Ethics (Oxford: Clarendon Press, 1931), 26 quoted at Jonathan Dancy, “Should We Pass the Buck?”, Royal Institute of Philosophy Supplements 47 (2000): 159–173, 160.
250
Joseph points out something important. A virtuous person must hold that consequences matter. But to say this is not enough since consequences may very well matter without mattering to individual virtue. All that matters to individual virtue, one may insist, is motivation. A motivationalist who offers this response but wishes to hold also that Hume’s monk or the compassionate sibling who serves as an addiction enabler do not, in the end, possess virtuous motivation can argue that, perhaps, the motives of those agents are, while not bad, somehow defective. I don’t think this response succeeds. While it is true that in some cases, a failure to properly consider the full range of an action’s consequences may be due to deficient motivation, in other cases, it may not be. It may, instead, be due to a purely epistemic failure on the agent’s part or even to ignorance about facts without epistemic failure. Worse still, it may not be due to anything properly described as a failure, whether cognitive or motivational. Beating one’s children with the aim of bringing them up as good citizens was never virtuous, even though when corporal punishment was the norm, the disposition to beat one’s children may have involved no failure on an individual agent’s part, either cognitive or motivational. One can, in fact, imagine a parent who had no desire to beat her children at all but who thought she should, because this is what responsible parents did. Such a person’s disposition to beat her children would not be a virtuous one though it may be one that is not blameworthy.²⁵ Last but not least, there are at least some virtues that involve an element of skill and are not plausibly seen as a matter of motivation alone. Consider tactfulness: no one can be tactful if she keeps making blunders, no matter how good her motivation is. To insist that necessarily, there would have to be something wrong with the motivation of the person who, a bit like Mrs. Malaprop, keeps making faux pas and says, for instance, “I am not prolix like you” when she means “I am not prolific like you” , is to defend the indefensible. Such a person may be tactless even though she is fully motivated to be tactful.²⁶
²⁵ It may be that when a person is blameless, as this imaginary parent arguably is, she cannot possess a vice. But she does not have a virtue either despite best intentions. ²⁶ Perhaps, one can say that traits such as tactfulness that have a strong skill component are not really moral virtues. But I don’t think this response succeeds. Kindness, for instance—an indisputably moral virtue—has a skill component as well. Consider the misguided attempts to comfort her described by physician Azra Raza. After she lost her husband, her friends, in an attempt to offer consolation, said all manner of inappropriate things, for instance, one friend told her not to worry as she would soon be reunited with her husband, which of course, implies
? 251
5. Conclusion I promised to develop an account that does justice to both motivationalist and consequentialist intuitions, one that illuminates the connections between virtues and vices on the one hand, and consequences and motives, on the other. I have now accomplished my task. I argued that in principle, objective reasons—perhaps, primarily consequences though not only—are what make a type of behavior worthy of the label “virtue” in general, not the motive. Objective reasons may also make it so that a well-intentioned and blameless agent—such as the parent described at the end of the last section—cannot possess a virtue. However, individual virtues and vices depend not simply on what one does or how consistently, but on why one does it, that is, they depend on motives, and a person is virtuous to the extent that her motivation is in line with the objective reasons. She lacks virtue altogether if there is no motivational connection, even when she is morally lucky and ends up doing what there is reason to do purely because of external inducements. She is minimally morally decent when she accepts certain constraints on her behavior. She is virtuous when her motives are not simply constrained by but in line with the objective good. I consider my strategy to be an improved version of Aristotle’s strategy, one that aims to explain what is plausible about both consequentialism and motivationalism.²⁷ I do not wish to engage in extended interpretation of Aristotle here, but I will note that Driver is right about the following: an account of virtue that involves such a strong knowledge requirement that it makes it impossible for us to ascribe any virtue to Huck Finn is flawed. To that extent, Aristotle’s view is flawed. But the objection does not apply to my view. I spoke broadly about one’s motives and intentions being “in line with” the objective moral reasons in part because I wanted to rule out a knowledge requirement. Huck is virtuous to the extent that he is motivated by the objective reasons
that she would die soon. See “Azra Raza on The First Cell,” EconTalk, March 23, 2020 available at: https://www.econtalk.org/azra-raza-on-the-first-cell/. In all likelihood, the person who said that was trying to be kind and helpful. ²⁷ It is worth noting that Aristotle may not have had a conception of moral virtue as we understand it as he may not have had a conception of morality. David Wolfsdorf, “Morality and Aristotelian Character Excellence,” Questions of Character, edited by Iskra Fileva (New York, NY: 2016), 19–32 makes a strong case for this view, but for present purposes, this doesn’t matter. It is a disagreement about what moral reasons there are. The issue that interests me is premised on the assumption that that can be determined. Once it is, we must then ask how virtue relates to those reasons, whatever they are, on the one hand, and to motives and intentions, on the other.
252
to act as he does, even if he does not appreciate the normative force of those reasons.²⁸ There is a final point I wish to make before closing this discussion. I believe that objective consequences are key not simply in finding out what dispositions ought to be promoted and considered virtuous but in determining the extent to which conflicts of interests between people or between all of us and the universe, so to speak, should be resolved through the cultivation of virtue. It is often possible to improve the world and the social system in ways that make virtue less necessary. I take it that laborsaving devices are just such an improvement: one that makes life better despite, perhaps, making us lazier than we would otherwise be. I do not wish to take this point to an extreme. There may well be value in the cultivation of strengths of character such that even if we lived in a paradise in which no one needed to work, it would be better to do some work. My point is simply that this value need not be paramount. It is certainly not so important that we ought to refrain from improving the world for fear of worsening our characters.²⁹
²⁸ He is motivated by something like a desire to help Jim escape slavery though he doesn’t believe helping people escape slavery is the right thing to do. On this point, see also Julia Markovits, “Acting for the Right Reasons,” Philosophical Review 119–12 (2010): 201–42. ²⁹ It is consequentialists, I think, who are best positioned to accommodate this point. Motivationalists need not demur, however. It is just that nothing follows from their account as to the question of how important virtue is compared to other values. That question is separate from the question of what virtue is.
12 Two Conceptions of Rights David O. Brink
Rights play an important role in moral, political, and legal debate, as illustrated by conceptions of them as constraints on pursuit of the good and as trumps in relation to considerations of majority preference and collective goals. But we can and do think of rights in two quite different ways, even if we don’t always fully recognize or attend to the difference. On one conception, rights are important inputs to normative debate that often, but not always, determine the outcome of that debate. Rights, on this conception, are contributory, contributing important pro tanto reasons to a debate requiring adjudication. The contributory conception of rights represents them as potentially defeasible moral factors, especially important and presumptively decisive pro tanto reasons for the resolution of a practical question. Another conception represents rights as the outcome of debate about the entitlements of parties to the debate. On this conception, rights are those entitlements that deserve to win the debate. Rights, on this conception, are resultant, rather than contributory. Unlike contributory rights, resultant rights are verdictive and indefeasible. Some philosophers are monists about rights, recognizing only contributory or resultant rights. But the ways in which we think, talk, and argue about rights suggest that we can and should be pluralists about rights, recognizing both contributory and resultant rights. The important point is not to confuse them. Some philosophers make claims about rights that are not easily reconciled with their monism. If rights are only resultant, it’s hard to recognize conflicts of rights and rights that may be overridden. If rights are only contributory, it’s hard to recognize rights that are verdictive. Once we see the difference between contributory and resultant rights, we should expect their content to diverge. In general, we will have more contributory rights than resultant rights, and some contributory rights will not be resultant rights. Moreover, though some normative contexts presuppose resultant and indefeasible rights, several important claims about rights
David O. Brink, Two Conceptions of Rights In: Oxford Studies in Normative Ethics Volume 13. Edited by: Mark Timmons, Oxford University Press. © David O. Brink 2023. DOI: 10.1093/oso/9780198895909.003.0013
254
.
and their significance presuppose the contributory and defeasible conception. For this reason, the contributory and defeasible conception of rights might be regarded as primary and the resultant and indefeasible conception of rights as derivative. My discussion will have the following structure. After a few preliminaries (section 1), I will provide a fuller account of the contrast between these two conceptions of rights (section 2). I go on to explain how some adjudicative contexts employ a conception of rights that is resultant and indefeasible (section 3). However, many other contexts employ the contributory conception. First, moderate deontology treats rights as contributory and defeasible (section 4). Second, the existence of conflicts of rights presupposes the contributory conception in which rights are defeasible. Basic human rights are contributory rights that can conflict, though the correct resolution of those conflicts allows us to identify fine print rights that are verdictive and indefeasible (section 5). Third, constitutional rights are best understood as contributory rights. In particular, the conception of individual rights ingredient in Equal Protection and Substantive Due Process analysis requires the defeasible conception of rights (section 6). Moreover, when existing contributory rights are inadequate to resolve normative conflicts, we may be led to recognize new contributory rights, as illustrated in the constitutional recognition of rights to association, privacy, and same-sex marriage (section 7). Finally, the appeal to rights as a justification of adjudicative outcomes presupposes the contributory conception of rights, inasmuch as the resultant conception would render such appeals circular (section 8). In fact, we should view resultant rights as the proper weighing and outcome of potentially conflicting contributory rights and other morally relevant factors.
1. Preliminaries First, my main focus here is on the metanormative question of what rights are, as distinct from the normative question of what rights we have and the deliberative question of how rights function in normative reasoning and debate. Nonetheless, I don’t know how to make progress on the issue of what rights are without making assumptions, if only defeasibly, about what rights we have and how they function in normative reasoning and argument. While our assumptions about what rights we have and how they function in normative reasoning are revisable, our conception of what rights are should be broadly faithful to familiar and plausible views about their content and role.
255
Second, rights are potentially heterogenous inasmuch as rights can be conceived as claims, permissions, authorities, and immunities, and rights can be held by individuals and groups. Though the issues I want to raise here are quite general, applying to any of these kinds of rights, it will simplify an already complex discussion to focus on individual rights in the form of claims against others that generate duties for others to act or refrain from acting in specified ways toward the right holder.¹ Third, so understood, rights are directed claims—claims directed at other individuals or institutions that the right holder be treated in some ways and not in others. Rights can include negative demands for noninterference with a person’s liberties and opportunities or positive demands for individuals or institutions to provide her with opportunities, resources, or services. Finally, I am interested in the nature of rights in multiple normative contexts—moral, political, and legal. We can distinguish moral, political, and legal rights by the source of the right. Moral and political rights are rights that individuals have against others or the state antecedent to positive law, whereas legal rights are rights that individuals have only as the result of positive law. Though we can distinguish moral, political, and legal rights by their sources, we need to understand what their common structure is. This metanormative question can and should be distinguished from the normative question of what rights we have. But my discussion will sometimes draw on common assumptions about which moral and political rights we have and which existing legal rights are morally and politically defensible.
2. Resultant and Contributory Conceptions of Rights Rights are important claims in moral, political, and legal reasoning and debate. In particular, individual rights are important moral factors that constrain the operation of other moral factors. Robert Nozick articulates this idea when he contrasts goals and constraints and understands rights as side-constraints on the pursuit and promotion of valuable goals.² In the
¹ Following Hohfeld, legal theorists sometimes analyze rights into four main kinds— privileges, claims, powers, and immunities. See Wesley Newcomb Hohfeld, Fundamental Legal Conceptions (New Haven: Yale University Press, 1920). For useful discussion, see Michael Moore and Heidi Hurd, “The Hohfeldian Analysis of Rights” The American Journal of Jurisprudence 63 (2018): 295–354. My focus here is on claim-rights and the associated duties they impose on others. ² Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), pp. 28–33.
256
.
absence of rights, promoting the good is a reason for action. However, rights constrain the permissible pursuit of good consequences, such that it is wrong to violate rights, even if that produces good consequences.³ Similarly, Ronald Dworkin regards individual rights as trumps on appeals to majority preference and collective goals.⁴ In the absence of rights, majority preference or collective good might be good reasons for action. But rights trump these other kinds of reasons, making it impermissible to violate rights for the sake of majority preference or collective goods. But these claims about the importance and role of individual rights in relation to other moral factors are compatible with two conceptions of rights. In particular, we might conceive of rights as important contributory moral factors that have defeasible moral significance or as indefeasible moral factors that carry the day in moral argument and debate. On the one hand, we might conceive of rights as conclusions of debates about the entitlements of individuals. The correct resolution of such a debate would settle how individuals may act and what they are owed from others. These verdicts would be all-things-considered assessments. As such, the resulting rights would be final and indefeasible. On this conception, rights are also inviolable. In A Theory of Justice John Rawls claims that rights and justice constrain and defeat claims of general utility and connects this claim about rights with inviolability. [W]e distinguish as a matter of principle between the claims of liberty and right on the one hand and the desirability of increasing aggregate social welfare on the other; . . . we give a certain priority, if not absolute weight, to the former. Each member of society is thought to have an inviolability founded on justice or, as some say, natural right, which even the welfare of everyone else cannot override. Justice denies that the loss of freedom for some is made right by the greater good shared by others. The reasoning which balances the gains and losses of different persons as if they were one person is excluded. Therefore in a just society the basic liberties are taken for granted and the rights secured by justice are not subject to political bargaining or to the calculus of social interests.⁵
³ It’s not clear to me whether Nozick thinks that rights defeat pro tanto reasons to promote the good or disable them. ⁴ Ronald Dworkin, Taking Rights Seriously (London: Duckworth, 1977), esp. ch. 7. ⁵ John Rawls, A Theory of Justice (Cambridge: Harvard University Press, 1971), pp. 27–8.
257
Here, Rawls ascribes inviolability, in the first instance, to persons. But if personhood is the ground of rights, it seems that the rights of persons should be inviolable as well.⁶ On the other hand, one might conceive of rights as important inputs into normative debates that often, but not invariably, determine the outcome of those debates. On this conception, rights are pro tanto contributory factors. One might view individual rights the way W.D. Ross thought of prima facie—or better, pro tanto—duties.⁷ They are moral factors that make an invariant contribution to the moral valence of the situations in which they occur, giving parties in those situations pro tanto reasons for action. Absent interference from other morally relevant factors, contributory factors determine what one ought to do all-things-considered (or, as Ross said, “sans phrase”). But where a situation is morally complex, involving multiple moral factors, the force exerted by a given moral factor F1 might be overridden by other moral factors F2–F3. This shows that contributory moral factors generate pro tanto moral reasons that are defeasible. Of course, some contributory moral factors create modest reasons for action that are easily overridden by other moral factors in a situation. For instance, beneficence is a morally relevant factor that gives me reason to do things that benefit others. But at least some reasons of beneficence create modest reasons for action that are easily defeated by considerations of rights, justice, or the cost to the agent. If we want to capture the idea that rights constrain or trump the pursuit of other contributory factors, the contributory conception cannot understand rights as just one minor contributory moral factor among many. Instead, the contributory conception of rights must understand them as especially important contributory factors that typically control the resolution of the practical situation in which they are present. On this contributory conception, rights are presumptively decisive moral factors about the important claims of individuals against other individuals and institutions that bear on the resolution of practical questions about people’s entitlements.⁸ ⁶ Of course, one might try to ground rights in personhood without the assumption of inviolability. ⁷ See W.D. Ross, The Right and the Good (Oxford: Clarendon Press, 1930), esp. ch. 2. Ross discusses rights in the first appendix to ch. 2, but that discussion does not engage his distinction between prima facie duties and duties sans phrase. ⁸ In “Pro Tanto Rights and the Duty to Save the Greater Number” (this volume) Benjamin Kiesewetter defends a conception of pro tanto rights that has much in common with my contributory conception. However, he imposes constraints on pro tanto rights that I do not accept for contributory rights. Moreover, he is a monist insofar as he thinks that rights are pro tanto claims, whereas I am a pluralist insofar as I recognize both contributory and resultant rights.
258
.
3. Adjudication and Resultant Rights It is possible to view rights as the outcome of correctly adjudicating competing interests and claims. This perspective on rights requires the resultant conception. One illustration of this conception of rights is John Stuart Mill’s conception of rights in Chapter V of Utilitarianism.⁹ There, he develops a conception of justice and rights that is part of an indirect utilitarian conception of duty. In particular, Mill links duty and sanctions. For the truth is, that the idea of penal sanction, which is the essence of law, enters not only into the conception of injustice, but into that of any kind of wrong. We do not call anything wrong unless we mean to imply that a person ought to be punished in some way or other for doing it — if not by law, by the opinion of his fellow creatures; if not by opinion, by the reproaches of his own conscience. This seems the real turning point of the distinction between morality and simple expediency. [CW X 246]
Here, Mill defines wrongness and, by implication, duty, not directly in terms of the nature of the action or its consequences but indirectly in terms of appropriate responses to it. This commits him to claiming that one is under an obligation or duty to do something just in case failure to do it is wrong and that an action is wrong just in case some kind of external or internal sanction—legal punishment, social censure, or self-reproach—ought to be applied to its performance. Sanctions determine when conduct is wrong, which allows Mill to say that an act is one’s duty just in case its omission would be appropriate to sanction. In this way, the sanction test distinguishes duty from expediency (CW X 246–48). Not all suboptimal or inexpedient acts are wrong, only those in which one ought to apply some sort of sanction (at least, self-reproach) to them. Justice is a proper part of duty. Justice involves duties that are correlated with rights (CW X 246–48). Justice implies something which it is not only right to do, and wrong not to do, but which some individual person can claim from us as a matter of right. [CW X 247]
⁹ John Stuart Mill, Utilitarianism in The Collected Works of John Stuart Mill, 33 vols., ed. J. Robson (Toronto: University of Toronto Press, 1965–91), vol. X.
259
An act is unjust just in case it is wrong and violates someone’s rights (CW X 249–50). Someone has a right just in case she has a claim that society ought to protect by force of law or public opinion. When we call anything a person’s right, we mean that he has a valid claim on society to protect him in the possession of it, either by the force of law, or by that of education and opinion. If he has what we consider a sufficient claim, on whatever account, to have something guaranteed to him by society, we say that he has a right to it. If we desire to prove that anything does not belong to him by right, we think this is done as soon as it is admitted that society ought not to take measures for securing it to him, but should leave it to chance, or to his own exertions. [CW X 250]
In conceiving of rights as claims that society ought to protect and enforce, I assume that Mill uses in the verdictive or all-things-considered sense. If someone has a right to something, it would be wrong for society not to protect and enforce that claim. Notice that these relationships among duty, justice, and rights do not yet introduce any utilitarian elements. But Mill does think that whether sanctions ought to be applied to an action—and hence whether it is wrong—and whether society ought to enforce an individual’s claim—and hence whether she has a right—both depend upon the utility or expediency of doing so (CW X 250). He does not say precisely what standard of expediency he has in mind. To fix ideas, let us assume that something counts as wrong just in case it is optimal to sanction that conduct in some way and that someone has a right to something just in case it is optimal for the state or society to enforce that claim. Whether Mill’s attempt to provide a utilitarian foundation for rights is successful is an interesting and complex issue, beyond the scope of this essay.¹⁰ For present purposes, I am interested in the part of his analysis of rights that is not specifically utilitarian—his claim that rights are claims of individuals to liberties, opportunities, or resources that society ought to enforce. This link between rights and enforceability is an ecumenical claim that might appeal to non-utilitarians, as well as utilitarians. It’s a substantive,
¹⁰ I discuss the adequacy of Mill’s various resources for providing a utilitarian foundation for rights in David O. Brink, Mill’s Progressive Principles (Oxford: Clarendon Press, 2013), esp. ch. 9 and “Mill on Justice and Rights” in The Blackwell Companion to John Stuart Mill, ed. C. Macleod and D. Miller (Oxford: Blackwell, 2016).
260
.
rather than trivial, claim, because one might want to recognize rights that ought not to be enforced in some circumstances. But it’s nonetheless a common claim about rights. This link between rights and enforceability reflects a resultant conception of rights. Rights are just those claims that society ought to enforce all-things-considered. The way in which an adjudicative perspective presupposes the resultant conception of rights is also reflected in Dworkin’s conception of the role of judges and other legal interpreters in ascertaining the rights of litigants, especially in hard cases.¹¹ Hard cases are potentially controversial cases in which litigants advance competing claims about who is entitled to what sort of treatment from whom. Dworkin’s rights thesis is that in most cases, even in hard cases, one litigant is entitled to a decision in her favor as a matter of pre-existing right. The rights thesis tells us to identify the rights of litigants with the best or most justified resolution of their competing claims. Thus, the rights thesis also links rights and enforceability. In this way, the rights thesis assumes the resultant conception of rights, treating rights as settled by the conclusions of adjudication about which moral claims the legal system should enforce. Dworkin is writing about legal rights and the theory of adjudication. But he is not a legal positivist. He conceives of legal rights as claims that society ought to enforce as a matter of political morality. But, without wading into contested issues in general jurisprudence about the relation between law and morality, we can easily enough adapt Dworkin’s rights thesis from a legal perspective to a moral and political perspective. For we can think of moral and political disputes arising from the conflict of contributory moral and political factors, and we can identify the rights or entitlements of parties to those disputes as those that are part of the best resolution of those conflicts. On this view, rights are consequential on the correct balancing and adjudication of competing moral and political claims. These rights will be verdictive. Whether they are also enforceable depends on whether every all-thingsconsidered moral or political duty is enforceable. The resultant conception of rights may be reflected in other claims about rights, as well. The resultant conception fits Joseph Raz’s claims about rights, in particular, his assumption that rights are duties that are grounded in the interests of individuals and defeat contrary moral claims and his conception of rights as claims that have pre-emptive force in normative reasoning
¹¹ Dworkin, Taking Rights Seriously, ch. 4, esp. pp. 81, 87.
261
and debate.¹² If rights pre-empt all other normative reasons, then they are resultant.¹³ John Oberdiek adopts something very much like Dworkin’s rights thesis, offering a striking statement of the resultant conception of rights. The central role of rights in normative argument is conclusory; people argue toward rights, not from them . . . . The way that one proceeds in moral argument, therefore, is to marshal normative reasons of various kinds that purportedly establish that some right exists. One does not start with rights.¹⁴
Here, Oberdiek not only accepts the resultant conception but also rejects the contributory conception.
4. Moderate Deontology and Contributory Rights Resultant rights are final and indefeasible. Contributory rights are not. Contributory rights are part of the toolkit of moderate deontology. To understand moderate deontology, we need to understand the contrast between agent-neutral and agent-relative reasons. An agent-neutral reason is one that can be specified without any essential reference to the agent who has it, and, as a result, all else being equal, every agent has the same agentneutral reasons. By contrast, agent-relative reasons do make essential reference to the agent who has them, as a result of which different agents have different agent-relative reasons. A reason to promote happiness generally or minimize harm is an agent-neutral reason, whereas a reason to promote my own happiness or minimize my own pain is an agent-relative reason. It is common to recognize agent-relative constraints on the agent doing harm to others, even if this is necessary to minimize total harm, and to recognize agent-relative options to perform suboptimal actions, out of special concern for the agent’s own interests or the interests of others who stand to him in special relationships, such as his loved ones and friends. ¹² See Joseph Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986), ch. 7, esp. pp. 181–6. ¹³ It’s less clear if Raz is committed to a resultant conception of rights if rights pre-empt some, but not all, other normative reasons. ¹⁴ John Oberdiek, Imposing Risk: A Normative Framework (Oxford: Clarendon Press, 2017), p. 115.
262
.
A sensible agent-relative morality employs constraints and options that are moderate, rather than absolute.¹⁵ Constraints are absolute if it is always wrong to treat others in certain ways (e.g. violating their rights), no matter how much good could be achieved or harm avoided by doing so. By contrast, constraints are moderate if there is some sufficient amount of good to be produced or harm to be avoided that would make it permissible to violate the constraint. A moderate constraint could explain why it would be wrong for a surgeon to kill one patient to save two without implying that one innocent could never be killed to save very large numbers of innocents. Likewise, options are absolute if they permit the agent to prefer her own good or that of a loved one to the good of others, no matter how great the cost to others. By contrast, options are moderate if there is some amount of good that could be achieved, or harm prevented, that would defeat the agent-relative option. A moderate option could explain why I am permitted not to sacrifice significant goods in my life for the sake of marginally greater goods to strangers without implying that I could refuse to save the lives of many when I could do so at little or no cost to myself. Some understand moderate agent-relative morality in terms of thresholds of opportunity costs, below which constraints and options must be respected and above which they should not. But thresholds are normatively arbitrary. Besides the difficulty of specifying a Goldilocks threshold, thresholds perversely attach enormous significance to small differences just below and above the threshold and no significance to large differences below or above the threshold. Though there could be some precise point at which constraints and options give way to the increasing agent-neutral opportunity costs of respecting them, there is no reason to expect that there is some precise point on the consequentialist scale where constraints and options lapse, any more than there must be some precise point at which accumulating grains of sand constitute a heap. It could be indeterminate at what point moderate constraints and options lapse or are overridden.¹⁶ Moderate deontology does not imply threshold deontology.
¹⁵ For discussions of moderate agent-relative morality, see, e.g., Thomas Nagel, “War and Massacre” reprinted in Thomas Nagel, Mortal Questions (New York: Cambridge University Press, 1979); Michael Moore, Placing Blame (Oxford: Clarendon Press, 1997), pp. 721–5; Shelly Kagan, Normative Ethics (Boulder: Ridgeview Press, 1998), p. 79; and Larry Alexander, “Deontology at the Threshold” San Diego Law Review 37 (2000): 893–912. ¹⁶ For present purposes, we can remain agnostic about whether moderate deontology can be given a utilitarian foundation, as Mill thought.
263
It is common to treat rights as agent-relative constraints on the pursuit of the good. But a moderate deontology treats constraints as presumptively decisive pro tanto moral factors that are nonetheless defeasible when the moral opportunity costs of respecting them are sufficiently high.¹⁷ This requires the contributory conception of rights, which understands them as potentially defeasible, rather than the resultant conception of rights, which treats them as final and indefeasible.¹⁸ We could identify rights only with the outcomes of moderate deontology, but this would be a revisionary modification of the common view that rights are moderate constraints.
5. Conflicts Among Contributory Rights If we recognize conflicts of rights, we are treating them as contributory, rather than resultant, rights. It’s common to recognize the possibility of conflicts of rights. For instance, rights to religious liberty or freedom of association can conflict with anti-discrimination norms and the rights to equality that those norms protect. A right to a fair trial might conflict with a right to privacy, when a defendant’s right to the participation of material witnesses requires them to disclose personal information about themselves. Of course, when rights conflict, they can’t all be respected. Perhaps some conflicts of rights have no non-arbitrary resolution. But let’s assume, if only for simplicity, that in most cases, even hard cases that are controversial, there is a unique non-arbitrary resolution of the conflict in which one right should prevail in those circumstances over the other right. Both rights are important pro tanto contributory factors, but one factor will be outweighed and defeated by the other factor. Of course, after the fact, we can represent the prevailing factor as the resultant right. But the prevailed factor cannot be a resultant right. And we shouldn’t assume that the prevailing factor in these circumstances prevails in all other circumstances. So it’s appropriate to treat ¹⁷ Nozick mentions, without endorsing, this moderate interpretation of rights as sideconstraints. See Nozick, Anarchy, State, and Utopia, p. 30n. ¹⁸ Judith Thomson distinguishes between infringements of rights that fail to respect those rights and violations of those rights that impermissibly infringe those rights. She then characterizes absolutism about rights as the thesis that all infringements of rights are violations of them. She rejects absolutism, concluding that rights can be infringed without being violated. See Judith Thomson, The Realm of Rights (Cambridge: Harvard University Press, 1990), p. 122. My view of contributory rights is like her view in some respects, though I would eschew her distinction between infringements and violations and simply say that contributory rights may sometimes be permissibly violated. Absolutism is plausible, I think, but only about resultant rights.
264
.
the rights that pose the conflict requiring resolution as both contributory and potentially defeasible. A resultant conception of rights treats them as absolute and compossible, whereas a contributory conception does not. If freedom of association claims are limited by anti-discrimination claims, then there can be no resultant right to freedom of association per se. A resultant right to freedom of association would have to be qualified. Indeed, it would have to be qualified by all the conditions in which freedom of association would be limited. The same is true about a right to equal treatment that prohibited unjustified discrimination and, indeed, about all rights. But then our actual rights, on the resultant conception, are highly circumscribed in ways that most of us may not understand. The resultant conception avoids conflicts of rights by making the content of rights potentially esoteric. By contrast, the contributory conception can recognize familiar claims about the content of rights but will treat rights as potentially conflicting. Insofar as we recognize conflicts of rights, this is evidence that we are thinking of them in a contributory fashion. It is sometimes true that when rights conflict and we respect the right that is in the circumstances weightier, permissibly infringing the less weighty right, we leave a moral residue or remainder that may call for some form of moral repair. This may not be true in all cases of rights that are permissibly infringed, but it is true in some such cases. For instance, justifications, such as self-defense or necessity, involve circumstances where rights may be permissibly infringed.¹⁹ In some cases, the permissible violation of a right leaves a moral residue that calls for repair. One such case is a necessity or lesser evils justification for infringing property rights.²⁰ Suppose that we are hunting in a remote area and you accidentally incur a self-inflicted gunshot wound and that I reasonably believe that the only way to get you timely medical attention is to call Emergency Medical Services using the phone in the only cabin within miles. The cabin is unoccupied. My trespass, though otherwise unlawful, is justified in order to avoid the greater evil of loss of life. Indeed, we might represent the situation as a conflict between property rights and the right to life. Though I permissibly infringe the cabin owner’s property rights, you and I owe an explanation and apology and
¹⁹ For discussion, see David O. Brink, Fair Opportunity and Responsibility (Oxford: Clarendon Press, 2021), ch. 8. ²⁰ Joel Feinberg discusses a similar case in “Voluntary Euthanasia and the Inalienable Right to Life” Philosophy & Public Affairs 7 (1978): 93–123.
265
compensation, if we damaged her property. It is a virtue of the contributory conception that we can explain why there is normative residue and a duty of repair.²¹ The residue and duty of repair are consequential on the breach of the contributory right. A right is breached on the contributory conception but not on the resultant conception that recognizes the way in which the content of a property right is limited by necessity.²² Conflicts of rights and their resolution bring out a difference in the content of contributory and resultant rights. The content of contributory rights that produce conflicts of rights are often easily summarized—for instance, freedom of speech, religious conviction and practice, association, privacy, due process, equal treatment, just compensation, and fair opportunity. Because these rights are familiar rights on lists of basic rights and are easily stated in general terms, we might call them large print rights. But it is clear that large print rights can conflict and won’t always prevail. This is why we must think of the rights that generate conflicts of rights as contributory rights that are potentially defeasible. We can adopt a resultant conception of rights that identifies rights with the correct resolution of a conflict among contributory rights. But, strictly speaking, the content of the resultant right will be distinct from the content of the contributory rights. This is because the contributory right that wins some conflicts is likely to lose other conflicts. At least, this will be true as long as no contributory right wins all contests with other rights. But then the right that wins a particular conflict must be a qualified or fine print right. The larger the number of conflicts that a given contributory right loses to other rights, the more qualifications and fine print must be built into the resultant right. Plausible resultant rights will be esoteric and contain too much fine print to be easily summarized or to be of much use in political debate or education. For this reason, political debate and education will rely on contributory rights that are understood to be important and presumptively decisive moral factors that are nonetheless potentially defeasible. Here, it’s worth noting that resultancy will be a matter of degree and be relative to context. For a given claim to win one normative contest, it may ²¹ This virtue of the contributory conception of rights is a special case of the virtue of the Rossian analysis of moral conflict. See David O. Brink, “Moral Conflict and Its Structure” Philosophical Review 103 (1994): 215–47. ²² The inability to offer this explanation of residue and repair seems to be an unwelcome feature of the resultant conception’s denial of conflicts of rights. For discussion, within a resultant conception, see Russ Shafer-Landau, “Specifying Absolute Rights” Arizona Law Review 37 (1995): 209–26. Whereas the contributory explanation of residue is straightforward, I don’t see a satisfactory resultant explanation of residue.
266
.
have to be qualified, introducing some fine print. This qualified right might be a resultant right relative to that contest. But that same right might lose in other contests, giving way to other moral considerations, either other rights or sufficiently great moral opportunity costs. But then the right that was resultant relative to the first contest will not be fully resultant, because, unless further modified, with additional fine print, it won’t win the second contest. The specification of a right is more fully resultant the more normative contests it wins. But a right will only be fully resultant or resultant simpliciter if it wins all contests. Fully resultant rights are likely to contain a great many qualifications and much fine print and, as a result, be esoteric. Contributory rights can be understood to generate hedged generalizations about resultant rights. Other things being equal, contributory rights are resultant rights.²³ For instance, other things being equal, freedom of speech should be honored. But other things are not equal if the moral opportunity costs of respecting those contributory rights are sufficiently great or if those contributory rights compete with other rights that, in the circumstances, are morally more important. Indeed, when one contributory right conflicts with another one that is in the circumstances weightier, that entails that the moral opportunity costs of respecting the defeated contributory right are too great. If so, the same defeasibility that applies to rights on a moderate deontology applies to rights in cases of conflicts of rights. We express this defeasibility in the hedged relation between contributory and resultant rights. Contributory rights can be large print rights but only if they are hedged. Indeed, if the qualifications necessary for specifying all the conditions under which contributory rights defeat other claims in normative debate are complex enough and defy codification, then even fine print rights will need to be hedged.²⁴ ²³ Hedged generalizations are common outside of ethics. Water boils at 100 degrees Celsius, but only other things being equal. This is true only of samples of water that are pure, that contain normal levels of dissolved air, and at normal levels of atmospheric pressure. For instance, water will boil at a lower temperature in Denver than in San Diego. If hedged generalizations are not going to be vacuous, we need to be able to say something independently plausible about which conditions are interfering and defeating and why. These claims can’t just be post hoc ways of preserving the truth of the hedged generalization. For helpful discussion, see Paul Pietroski and Georges Rey, “When Other Things Aren’t Equal: Saving Ceteris Paribus Laws from Vacuity” British Journal for the Philosophy of Science 46 (1995): 81–110. ²⁴ These issues about whether rights are large print or fine print and whether they must be hedged are connected to debates between generalism and particularism in ethics. My own view, for which I cannot argue here, is a form of generalism that combines contributory, fine print, and hedged generalizations. Normative principles can be thought of as statements of pro tanto or contributory normative factors (e.g. good-making or right-making features) that make an invariant contribution to the overall normative valence of the act or situation in which they occur. Contributory principles have enabling conditions that must be met if that factor is to make its
267
6. Constitutional Rights as Contributory Rights Familiar conceptions of constitutional rights must treat them as contributory rights. I will focus on the case of constitutional rights in the United States, the case with which I am most familiar. The United States Constitution recognizes individual rights that constrain legislative action primarily in the Bill of Rights (the first nine amendments) and in the Due Process and Equal Protection clauses of the Fourteenth Amendment. These constitutional rights should be understood as contributory rights that establish a strong pro tanto but defeasible reason to treat legislation infringing those rights as constitutionally impermissible. Explaining this verdict requires providing context for these constitutional claims. The United States is a constitutionally limited democracy in which constitutionally protected individual rights constrain legislative interference with those rights. It is the institutional role of the judiciary to interpret and enforce these constitutional rights by exercising judicial review and declaring invalid legislation that impermissibly infringes those rights. The Bill of Rights recognizes various individual rights affording protection against federal interference. On their own, these rights provide no protection against state and local interference. So, for instance, the First Amendment itself provides no protection against interference with expressive liberties by state or local authorities. What makes the First Amendment applicable to state and local government is the doctrine of Selective Incorporation, which incorporates fundamental rights ingredient in the idea of the rule of law in the Bill of Rights into the Due Process guarantee of the Fourteenth Amendment. The Bill of Rights recognizes individual rights against federal action; the Due Process clause of the Fourteenth Amendment recognizes individual rights against state and local government. Selective Incorporation refers to the gradual, piecemeal, and selective process by which the most fundamental interests and liberties recognized invariant contribution (e.g. only free and voluntary promises are pro tanto binding). In principle, enabling conditions could be folded into the specification of the normative principle, resulting in a fine print principle. If enabling conditions cannot be finitely or conveniently specified, then even contributory principles must be formulated in a way that is hedged, containing an “other things being equal” clause. Many normative situations involve multiple normative factors, and an overall normative verdict depends on which factors are most important or prevail. Because normative principles state only contributory factors, their contribution to the normative valence of an act or situation is subject to interference and possible defeat from other normative factors in the situation. If potential interference and defeat from other factors cannot be finitely or conveniently specified, then normative principles, understood as claims about overall normative valence, must always be hedged, containing an “other things being equal” clause.
268
.
in the Bill of Rights have been treated as part of the Due Process guarantee in the Fourteenth Amendment. Selective Incorporation is part of so-called Substantive Due Process. In effect, Substantive Due Process recognizes individual rights to fundamental interests, liberties, and opportunities, treating these rights as imposing constraints on democratic legislation that require special judicial scrutiny. This claim about the concept of Substantive Due Process has been reasonably constant. But the content of Substantive Due Process—in particular, which rights are recognized as fundamental—has evolved over time. The Lochner era, epitomized by Lochner v. New York (1905), accorded heightened protection to economic interests and freedom of contract.²⁵ The New Deal conceived of economic interests and liberties as conditioned by considerations of equality and fair opportunity and so treated liberty of contract as non-fundamental.²⁶ For the most part, modern Substantive Due Process accepts the New Deal reinterpretation of economic rights but privileges selected personal and political liberties. It embraces Selective Incorporation of fundamental provisions enumerated in the Bill of Rights, including rights to freedom of expression and religion, rights against unreasonable searches and seizures, rights to due process and fair trials, rights to just compensation for property appropriated by the state, and rights against cruel and unusual punishment.²⁷ Due Process rights require that legislation affecting them survive a heightened standard of review. The distinction between different standards of review was implicit in Lochner. Whereas a Court majority rejected New York legislative restrictions on the working hours of bakery employees, employing a very demanding standard of review, the dissent voted to uphold the labor regulations, employing a more deferential standard of review. But the differentiation between these two standards of review was only explicitly formulated later in Equal Protection, rather than Due Process, analysis.²⁸
²⁵ Lochner v. New York, 198 U.S. 45 (1905). ²⁶ New Deal Substantive Due Process jurisprudence is reflected in a series of cases, including Nebbia v. New York, 291 U.S. 502 (1934); West Coast Hotel v. Parrish, 300 U.S. 379 (1937); and Williamson v. Lee Optical, 348 U.S. 483 (1955). ²⁷ See, e.g., Palko v. Connecticut, 302 U.S. 319 (1937). ²⁸ In United States v. Carolene Products Company, 304 U.S. 144 (1938) the Court acknowledged a general presumption in favor of the constitutionality of democratically enacted legislation, but acknowledged three situations that would call for more careful scrutiny of legislation: (1) when the legislation seems on its face to violate specific guarantees in the Bill of Rights or the Fourteenth Amendment, (2) when legislation compromises democratic processes, and (3) when legislation discriminates against “discrete and insular minorities.”
269
For instance, in Korematsu v. U.S. (1944),²⁹ which was an equal protection case concerning the internment of Japanese Americans during WWII, the Court determined that racial classifications were “suspect” and so triggered “rigid scrutiny” of legislation, requiring a showing of “pressing public necessity.” Regrettably, a Court majority determined that the internment program met this stricter scrutiny. Since Korematsu, the Court has evolved a basically two-tier system of review, including more and less deferential standards of review. The more deferential standard of review is known as rational basis review. Rational Basis Review: Legislation is constitutionally permissible if it pursues a legitimate governmental interest in a reasonable manner. A legitimate governmental interest is one that is not constitutionally proscribed, and a reasonable manner of securing or promoting that interest is one that might plausibly be thought to advance that objective. Though it is possible for legislation to violate rational basis of review, this standard of review sets a comparatively low threshold of scrutiny. This comparatively deferential standard of review can be contrasted with strict scrutiny. Strict Scrutiny: Legislation is constitutionally permissible if and only if it pursues a compelling state interest in the least restrictive manner possible. To survive strict scrutiny, legislation must be pursuant to an extremely important, and not merely legitimate, interest. Moreover, it must do so in the least restrictive manner possible, that is, in a way that does least violence to the individual interest or liberty in question, compatibly with securing the compelling interest. Though legislation does not automatically fail strict scrutiny, as the Korematsu case demonstrates, it does set a much higher threshold for legislation to pass. Strict scrutiny is a comparatively demanding and less deferential standard of review.³⁰
²⁹ Korematsu v. United States, 323 U.S. 214 (1944). ³⁰ Gerald Gunther famously described strict scrutiny as “strict in theory, fatal in fact.” See Gerald Gunther, “Foreword: In Search of an Evolving Doctrine on a Changing Court: A Model for a New Equal Protection” Harvard Law Review 86 (1972): 1–48. However, this assessment is not accurate. An empirical study of federal cases found that a full 30 per cent of cases decided by appeal to strict scrutiny satisfied that standard. See Adam Winkler, “Fatal In Theory and Strict In Fact: An Empirical Analysis of Strict Scrutiny in Federal Courts” Vanderbilt Law Review 59 (2006): 793–871.
270
.
With some exceptions, this two-tier system of review now frames the interpretation of the Bill of Rights and both Due Process and Equal Protection analysis.³¹ That means that the recognition of constitutional rights is committed to this pattern of analysis. In particular, constitutional rights are individual claims to fundamental interests, liberties, and opportunities that constrain democratic action, at both federal and state levels, by requiring legislation affecting these rights to survive strict scrutiny. This conception of constitutional rights fits the contributory model. Constitutional rights are not absolute and don’t settle the question of the permissibility of legislation infringing those rights. Rather, constitutional rights create a defeasible presumption that legislation infringing those rights is impermissible, but that presumption is rebuttable if the legislation passes strict scrutiny—that is, if it pursues a compelling state interest in the least restrictive manner. Though strict scrutiny is a legal doctrine that emerged gradually from diverse contingencies and pressures of constitutional history, it is a remarkably good fit with a conception of constitutional rights as important but defeasible constraints on the pursuit of collective goods. Like moderate deontology and a conflict of rights, the constitutional perspective presupposes the contributory conception of rights as presumptively decisive but defeasible inputs into normative debate and the adjudication of claims. Though I have focused on constitutional rights as they have developed in the United States, I believe that this analysis of constitutional rights is reasonably robust within a comparative perspective.³²
7. New Contributory Rights The rights that often figure in moral, political, and legal debate are typically contributory rights—familiar large print rights that can conflict with each ³¹ The Court’s treatment of commercial speech, under First Amendment jurisprudence, and gender classifications, under Equal Protection jurisprudence, are exceptions to this rule, insofar as the Court subjects restrictions on commercial speech and regulations distributing social benefits and burdens by gender to an intermediate standard of review that conditions the constitutionality of legislation infringing protected interests on the underlying governmental interest and the fit between legislative ends and means both being substantial. ³² For instance, the rights recognized in the Canadian Charter of Rights and Freedoms (1982) are treated as defeasible claims under the “reasonable limits” test in §1, as analyzed under the Oakes test developed in R v. Oakes (1986), 1 SCR 103. Similarly, private law rights in Japan have been recognized to be defeasible since the important case of Japan v. Shimizu, Imp. Ct., 3 March 1919, 25 Minroku 356. The case is translated and analyzed in C.D.A. Evans and J. Mark Ramseyer, “Japan v. Shimizu: Negligence and Abuse of Rights in Early 20th Century Japan” Rechtsprechung, Case Law 51 (2021): 313–28.
271
other and that provide presumptively decisive but nonetheless defeasible factors in normative debate. In cases of conflicts of rights, not every right can be honored, revealing the defeasible nature of contributory rights. But in some cases it may be true that no previously recognized right fully illuminates and resolves the conflict. New conflicts and further reflection on previous conflicts may lead us to recognize new rights, not contained in our existing juridical toolkit. In this way, new rights can emerge from recognition of the limitations of previously recognized rights. The evolution of constitutional rights provides some examples. Modern Substantive Due Process came to see the need to recognize fundamental non-enumerated rights, such as rights of association and privacy. Though the First Amendment defends a right to peaceful political assembly, it does not explicitly mention a general right of personal association. Nonetheless, in a series of cases the Supreme Court has recognized a right of association as implicit in First Amendment rights to expressive liberties and political assembly. Initially, association was limited to official public organizations with a political mission, such as the NAACP, but eventually the Court came to recognize associational interests and rights in private associations, such as the Jaycees and the Boy Scouts.³³ Similarly, the Court came to see the need to recognize a right of privacy that was implied by individual liberties against governmental interference explicitly recognized in the Bill of Rights. Privacy was interpreted as a kind of personal autonomy and led to recognition of more particular rights, including a circumscribed right to get an abortion free from governmental interference, a right to same-sex consensual intimacy, and a right to same-sex marriage.³⁴
³³ The evolution, sometimes uneven, of freedom of association can be seen in, e.g., NAACP v. Alabama ex rel. Patterson, 357 U.S. 449 (1958); Bates v. City of Little Rock, 361 U.S. 516 (1960); Heart of Atlanta Motel, Inc. v. United States, 379 U.S. 241 (1964); Roberts v. United States Jaycees, 468 U.S. 609 (1984); Board of Directors of Rotary International v. Rotary Club of Duarte, 481 U.S. 537 (1987); New York State Club Ass’n v. New York City, 487 U.S. 1 (1988); Hurley v. Irish-American Gay Group, 514 U.S. 334 (1995); Boy Scouts of America v. Dale, 530 U.S. 640 (2000); and Masterpiece Cakeshop, Ltd, Et Al. v. Colorado Civil Rights Commission, Et Al., 584 U.S. ___ (2018). ³⁴ The evolution of a constitutional right to privacy can be seen in, e.g., Griswold v. Connecticut, 381 U.S. 479 (1965); Roe v. Wade, 410 U.S. 113 (1973); Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833 (1992); Lawrence v. Texas, 539 U.S. 558 (2003); and Obergefell v. Hodges, 576 U.S. 644 (2015). Of course, the recent decision in Dobbs v. Jackson Women’s Health Organization, 19–1392, 597 U.S. __ (2022) overturns both Roe and Casey, abandoning fifty years of settled interpretation of constitutional privacy rights in cases involving abortion. This is not the place to discuss the Dobbs ruling, which I regard as unsound, or its larger implications for constitutional jurisprudence.
272
.
But these new constitutional rights are themselves contributory rights. This follows from the fact that like all Due Process and Equal Protection rights they are presumptively decisive but defeasible when legislation restricting them satisfies strict scrutiny. So, for example, a strong case can be made that the associational rights of members of quasi-public associations can and should be restricted when those associations violate antidiscrimination norms and so infringe the rights to equal opportunity of those excluded from the quasi-public association. For instance, the Court was willing to uphold anti-discrimination laws that limit freedom of association by recognizing the legitimacy of anti-discrimination laws in Heart of Atlanta Motel, Inc. v. United States (1964). The Heart of Atlanta Motel refused to rent rooms to Black people in violation of Title II of the Civil Rights Act of 1964, which prohibited discrimination “on the basis of race, color, religion, or national origin” in public accommodation pursuant to Congressional authority to regulate interstate commerce. The Court held that the motel owner’s rights of freedom of association were limited by antidiscrimination norms and the associated rights of Black Americans. In this way, recognition of the inadequacy of existing conceptions of contributory rights can lead to the recognition of new rights, which themselves should be understood as contributory rights. When these newly recognized rights help illuminate and resolve normative conflicts, they are part of juridical progress, helping us better map the landscape of contributory rights.
8. The Explanatory Role of Contributory Rights As we have seen, contributory rights serve as important presumptively decisive but potentially defeasible inputs to normative debate about people’s entitlements, whereas resultant rights represent the correct resolution of those debates and, hence, an indefeasible statement of people’s entitlements. Because resultant rights figure as conclusions of normative debate, it must be contributory rights that play the familiar dialectical role of premises in these debates. We often appeal to people’s rights as a reason or justification for permitting them to do something, for protecting their liberties or opportunities, or for requiring others to treat them in particular ways. In other words, we appeal to individual rights in support of conclusions about people’s entitlements. In these contexts, we must be understanding rights as contributory,
273
rather than resultant rights. We can’t appeal to resultant rights as a reason for a conclusion about people’s entitlements, because that is just another way of stating the conclusion, not an independent premise in an argument for this conclusion. We couldn’t expect interlocutors looking for a defense of an assignment of entitlements to accept an appeal to resultant rights as evidence for this verdict. But contributory rights can play this dialectical role. To appeal to contributory rights in support of claims about resultant rights is not circular reasoning, because contributory rights are potentially defeasible and don’t entail resultant rights. Moreover, since contributory rights will be large print rights, they will often be familiar and common ground among interlocutors. These characteristics of contributory rights lend themselves to their dialectical role in normative debate. Their defeasibility gives them enough separation from resultant rights to avoid circularity in the justification of people’s entitlements. Though defeasible, contributory rights are positively relevant to the determination of resultant rights. After all, the resultant rights in a situation are just those contributory rights that have not been defeated by other contributory rights or the opportunity costs of respecting them. Because contributory rights are important presumptively decisive moral factors, they provide good, but defeasible, evidence for the assignment of resultant rights. Appeal to contributory rights should carry the day in normative debate unless it can be shown that their moral opportunity costs are too great or that they conflict with even weightier rights.
9. Concluding Remarks Rights are important factors in normative debate. On familiar views, rights are side-constraints on the pursuit of the good and trumps on appeal to majority preference or collective good. But these familiar claims paper over the distinction between contributory and resultant rights. That distinction is crucial, but often overlooked or misunderstood. Both kinds of rights are legitimate and, indeed, important. Resultant rights are indefeasible. They are what we aim at in normative debate and various kinds of adjudicative contexts. Because resultant rights are the outcome of balancing various kinds of conflicting normative factors, resultant rights will typically be highly qualified factors that can only be fully specified with a lot of fine print. For example, the full specification of a resultant right to freedom of speech will have to state all the conditions
274
.
under which free speech can be permissibly restricted. This might include restrictions on incendiary speech that presents a clear and present danger; restrictions on low-value speech (e.g. libel and fighting words); time, manner, and place restrictions on speech; restrictions on speech to protect captive audiences; restrictions on speech to prevent workplace harassment; and restrictions on hate speech. This is only some of the fine print that might be necessary to specify the contours of a resultant right of free speech. Indeed, the exact content of a resultant right to free speech may be very hard to state and is potentially esoteric. However, many familiar claims about rights and their dialectical significance presuppose that rights are contributory and potentially defeasible, rather than resultant and indefeasible. First, rights are part of moderate agent-relative morality. But moderate deontology implies that rights are defeasible when the moral opportunity costs of respecting them are great enough. Second, the possibility of conflicts among large print rights also implies that even if rights are normally decisive, they are nonetheless defeasible moral factors. Indefeasible rights will have to be fine print and won’t play the same role as defeasible rights in normative reasoning and debate. Third, constitutional rights of the sort involved in the Bill of Rights and Due Process and Equal Protection analysis demand strict scrutiny of legislation infringing these rights. Though strict scrutiny establishes a presumption of protection, that presumption can be rebutted if the state has a compelling interest that it pursues in the least restrictive manner. This makes constitutional rights contributory rights—important and presumptively decisive constraints that are nonetheless defeasible. Fourth, only contributory rights can provide a non-circular justification for conclusions about people’s enforceable entitlements. Contributory rights provide independent but defeasible evidence for verdicts about those entitlements. Contributory rights can be large print, and they are potentially defeasible. Of course, the fine print resultant right could figure as an input into moral and political debate. If so, it could function as a contributory right and would not be defeated. But contributory rights are not in general indefeasible. As we have seen, many presumptively decisive contributory factors are defeasible, and most large print contributory factors are defeasible. The fine print necessary for most resultant rights is usually identified only as the result of moral, political, and legal debate and adjudication. So although a fine print factor could in principle be an input to debate and adjudication, and would then be indefeasible, most contributory rights are large print and potentially defeasible. This means that even if fine print contributory rights
275
could also be resultant rights, the content of contributory rights and resultant rights will generally diverge. It also means that there will be many more contributory rights than resultant rights, because many contributory rights are not only defeasible but will be defeated in the course of normative debate and adjudication. On the one hand, attention to these familiar claims about rights and their significance presupposes the potentially defeasible character of rights. This testifies to the importance and centrality of the contributory conception of rights. On the other hand, both conceptions of rights have legitimate and, indeed, complementary functions. A determination of resultant rights is the ultimate aim of much normative debate in moral, political, and legal contexts. But it’s the interaction and proper weighting of contributory rights and other morally relevant factors that produce this determination. We make best sense of the interpretation of rights as side-constraints or trumps by understanding them as contributory rights that state important pro tanto and presumptively decisive moral factors that are nonetheless defeasible. Only by recognizing and weighting these contributory rights can we arrive at a proper appreciation of our resultant rights. The result of this analysis is a kind of pluralism about what rights are that recognizes both contributory and resultant rights. Failure to distinguish these two kinds of rights can produce confusion. For instance, it is problematic to claim that rights are absolute or inviolable but also to recognize conflicts of rights and rights that may be overridden. This would either be inconsistent or, perhaps more charitably, involve a tacit switch between resultant and contributory conceptions of rights. To be clear, I am not arguing that there is something incoherent about rights monism or that one couldn’t be a consistent resultant monist. One can maintain resultant monism if one is prepared to deny enough common and familiar discourse about rights as mistaken or misleading. A resultant monist must deny that moderate deontological constraints are rights, that rights can conflict, that infringements of constitutional rights are permissible if they satisfy strict scrutiny, and that large print rights play an explanatory role in determining people’s entitlements. This requires a sort of skepticism or error theory about contributory rights. They are not genuine rights, but only sources of rights.³⁵ But this is unnecessarily revisionary. We can respect and explain ordinary and theoretical claims about rights and their role in ³⁵ Insofar as resultant monism is skeptical about ordinary claims about rights, treating contributory rights as mere sources of rights, it is like rule-skepticism about the law that treats
276
.
normative argument by accepting pluralism and recognizing both contributory and resultant rights. On this view, contributory rights turn out to be explanatorily primary insofar as resultant rights are explained by the interaction of contributory rights and other pro tanto moral factors. Contributory rights can be explanatorily primary in this way, even if normative analysis is only complete when it establishes resultant rights.³⁶
constitutional provisions and statutes, not as laws, but as sources of law that is ultimately determined by the final judgments of courts. See John Chipman Gray, The Nature and Sources of Law, 2d ed. (New York: Macmillan, 1921). For discussion, see H.L.A. Hart, The Concept of Law (Oxford: Clarendon Press, 1961), ch. 7 and David O. Brink, “Legal Interpretation, Objectivity, and Morality” in Objectivity in Law and Morals, ed. B. Leiter (New York: Cambridge University Press, 2001). ³⁶ This essay has benefitted from a 2021 ethics workshop at the University of California, San Diego, the 2022 Workshop in Normative Ethics in Tucson, AZ, a 2022 UCSD political theory workshop, a 2022 Cambridge University Law and Philosophy Forum, and a 2022 ethics workshop at Cornell University. I’d like to thank Dallas Amico, Abdul Ansari, Dick Arneson, Robert Audi, Saba Bazargan-Forward, Reuven Brandt, Cheshire Calhoun, Aaron Chip-Miller, Kathleen Connelly, Tom Dougherty, Emma Duncan, C.D.A. Evans, Micha Gläser, Tom Hurka, Sean Ingham, Benjamin Kiesewetter, Tom Kirkpatrick, Matthew Kramer, Andy Lamey, Matt MacTravers, Dana Nelkin, Zeynep Pamuk, Doug Portmore, Sam Rickless, Gila Sher, Houston Smit, Holly Smith, Horacio Spector, Evan Tiffany, Mark Timmons, Manuel Vargas, Shawn Wang, David Wiens, Monique Wonderly, and two anonymous readers for helpful comments.
Index For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. action guidance 129–30, 142–3 akrasia as double-counting 110 as incoherence 108–9 Anderson, Elizabeth 80 Aristotle 101 attachment 156–8 autonomy and deliberative project 32–3, 50–2 and nudges 33–4, 53 and options 46–8 coherentist conceptions of 39 Bell, Macalester 65–6 Bicchieri, Cristina 154–5 Blumental-Barby, Jennifer 46–8 Bradley, Ben 174–5 Bratman, Michael 38–9 Brewer, Talbot 96–7 bribe 26–7 buck-passing accounts of value 77–82 and consequentialism 81, 83–4 and dignity 88–9 and love 85–6 and naturalism 82 and reason-giving properties 86–8 Callard, Agnes 156 Cave, Eric 45–6 Chang, Ruth 224–5 Chappell, Richard 176–7 commitments deliberative commitments 38 self-destructive 49–50 non-deliberative motivation 40 normative stance 37 compensation 24–6
consent and coercion 11, 27–8 and paradoxical proposals 29–30 conditions of invalidation 17–19 consequentialism incompatible moral requirements 99, 121 maximizing act 170–4 and demandingness 172 and infinite alternatives 174 and supererogation 171–4 objection from virtue theory 183–5 satisficing act 170, 174–9 and effort ceiling 176–8 and gratuitous value obstruction 174–5 and threshold inflexibility 175–6 and virtuous satisficing 180–5 saving the many 215–16 contractualism 84 and moral uncertainty 133–7 deviant causal chains 14 direction of fit 153–4 Driver, Julia 166, 235 Dworkin, Ronald 255–6, 260 expectations and wrongs 160–5 as aspirational 156–8 as expressive 154–6 as maps 158–60 as predictive 150–2, 161 as prescriptive 152–4 disunity of 165–7 Feinberg, Joel 190–1 Frankfurt, Harry 39
278
Kantianism dignity 229–30 incompatible moral requirements 99, 121 innumeracy 216 miminal duty to aid 222–3 no defeat principle 217–24 price versus dignity 220–1 saving the few 217–20, 229–33 saving the many 228–9 Kolodny, Niko 18–20 Kukla, Quill 163–4
moral indifference 57–60 and claimant injustice 61–2 and indifferent stance 71–2 and moral assholes 66–7 and protection 72–3 and reactive attitudes 61–2 and recognition respect 68–9 and withdrawal 65–6 globalist versus localist 63–4 relational versus impersonal 56–7 structural 60 target of 60–3, 67–8 moral repair 264–5 moral uncertaintism content-neutral versus discriminating 137–43 content objection 145–6 global versus local 129 maximizing expected choiceworthiness 126, 128–9 regress objection 145–6
Lindeman, Hilde 163 love 85 and admiration 96–7 and attitude entitlements 89–90 versus mutual recognition 90–1
Nagel, Thomas 249 non-human animals and moral status 140–3 and vegetarianism 141–2 Nozick, Robert 255–6
Mandeville, Bernard 244 manipulation 31–2 and capacity 46 and deliberative project 39–42 and disrespecting autonomy 32, 44 and endorsement 46–8 and motivation 45–6 and normative stance 37 and paternalism 42–4 and rationality 33–4, 48–9 and reasons responsiveness 34–6, 48–9 Markovits, Julia 64–5 McGeer, Victoria 157 Mill, John Stuart 240, 258–60 moral incoherence 102 and akrasia 108–10 and lazy reasoning 122–3 and supererogation 112 and the virtuous person 102–7 and weighing reasons 105 versus moral indeterminacy 102–5
ought subjective versus objective 128–9
Gauthier, David 131 Harman, Gilbert 145 Hieronymi, Pamela 92–3, 95 Hume, David 236 incommensurability 224–5 interests setting back versus infringement 18–20
paradoxical proposal 9–10 Pareto principle 223–4, 227–8 permissibility and moral decency 179–80 proposal incentive versus disincentive 14 introduced versus independent 13–16 initiating versus accepting 26–9 threat versus offer 11–13 Ramsey, Frank 159 Rawls, John 256–7 Raz, Joseph 260–1 reasons agent-neutral vs. relative 261–3 and aggregation 207–11 and defeat 199–202
and directed deontic structure 193–7 and moral repair 193–4, 200 and uncertainty 225–8 directed 193, 199 for attitudes 93–7 peremptory 194 pre-emptive 194–6, 200–1 proleptic 156 respect recognition 68–9 rights and reasons 202–3 as directed 255 as directed deontic structures 197–9 as non-aggregative 205–7 as side-constraints 255–6 constitutional 255, 260, 267–72 contributory 190–1, 202–5, 253, 257, 272–6 and conflict 263–6 monism vs. pluralism 253 resultant 253, 256–7, 273–5 Ross, W. D. 257 Scanlon, T. M. 77–82, 131–3, 138 Schroeder, Mark 156 Singer, Peter 127–8 sexual harassment 20–1 and sexual integrity 22–3 and subordination 23–4, 26 stance indifferent 71–2 objective 70 participant 69–70 Strawson, P. F. 69–70 suberogation 166 supererogation and moral ideals 112–16 and moral reasons 110–12
as double-counting 119 as non-morally significant 117–18 Taurek, John 205–7, 215–16, 220 Thompson, Michael 55 Thomson, Judith Jarvis 190–1 value of persons and ending life 83–4 and reasons for attitudes 91–8 as irreducible 97–8 virtue characterization of 103–4 disunity of 102–3 virtue theory and consequences 239–42, 252 and motivations 243 and supererogation 186 as non-ultimate explanation 186–8 consequentialist 236–7, 244–5 hybrid 238–9, 243–4, 246–7 motivationalist 237–8, 247–50 objectivist 242–3, 246–7, 251 vulnerability 20–6 Wallace, R. Jay 68, 87 Watson, Gary 39 Weatherson, Brian 145–6 Wonderly, Monique 156–8 withdrawal and contempt 65–6 and disgust 66 and moral indifference 65–6 self- vs. other-protective 73 wronging 55 doxastic 162 instrumentalizing 64–5 objectifying 64–5
279